Bug 47288 - System hangs with GeForce 6150SE nForce 430 ("EQ overflowing")
Summary: System hangs with GeForce 6150SE nForce 430 ("EQ overflowing")
Status: RESOLVED INVALID
Alias: None
Product: xorg
Classification: Unclassified
Component: Driver/nouveau (show other bugs)
Version: 7.6 (2010.12)
Hardware: Other All
: medium normal
Assignee: Nouveau Project
QA Contact: Xorg Project Team
URL:
Whiteboard:
Keywords:
Depends on:
Blocks:
 
Reported: 2012-03-13 13:29 UTC by Frank Schaefer
Modified: 2013-09-27 02:03 UTC (History)
1 user (show)

See Also:
i915 platform:
i915 features:


Attachments
Xorg.0.log (39.60 KB, text/plain)
2012-03-13 13:30 UTC, Frank Schaefer
no flags Details
Updated Xorg.0.log (146.50 KB, text/plain)
2012-07-17 21:08 UTC, Frank Schaefer
no flags Details
kernel messages (920.33 KB, text/plain)
2012-07-17 21:08 UTC, Frank Schaefer
no flags Details
gdb "bt full" and "info frame" (13.57 KB, text/plain)
2012-07-22 19:33 UTC, Robert-André Mauchin
no flags Details

Description Frank Schaefer 2012-03-13 13:29:05 UTC
My system (GeForce 6150SE nForce 430) hangs up from time to time under KDE 4.7.4. All I can do then is to move the mouse pointer.
Clicks have no effect and I can not switch to the console. I even can't shut down the system pressing the power button.
It seems I can trigger this by increasing the workload, e.g when maximizing/minimizing lots of windows.

Xorg.0.log shows the following messages (full file attached):
...
[   771.443] [mi] EQ overflowing. The server is probably stuck in an infinite loop.
[   771.443] 
Backtrace:
[   771.456] 0: /usr/bin/Xorg (xorg_backtrace+0x37) [0x80a86b7]
[   771.456] 1: /usr/bin/Xorg (mieqEnqueue+0x1d1) [0x80a2841]
[   771.456] 2: /usr/bin/Xorg (xf86PostMotionEventM+0xb0) [0x80c9d70]
[   771.457] 3: /usr/bin/Xorg (xf86PostMotionEventP+0x6f) [0x80c9ebf]
[   771.457] 4: /usr/lib/xorg/modules/input/evdev_drv.so (0xb63a7000+0x2dde) [0xb63a9dde]
[   771.457] 5: /usr/lib/xorg/modules/input/evdev_drv.so (0xb63a7000+0x3f0d) [0xb63aaf0d]
[   771.457] 6: /usr/bin/Xorg (0x8048000+0x6ea11) [0x80b6a11]
[   771.457] 7: /usr/bin/Xorg (0x8048000+0x129da2) [0x8171da2]
[   771.457] 8: (vdso) (__kernel_sigreturn+0x0) [0xffffe400]
[   771.457] 9: /lib/libc.so.6 (ioctl+0x14) [0xb73e7114]
[   771.457] 10: /usr/lib/libdrm.so.2 (drmIoctl+0x34) [0xb7198f04]
[   771.457] 11: /usr/lib/libdrm.so.2 (drmCommandWrite+0x3c) [0xb719ba9c]
[   771.457] 12: /usr/lib/libdrm_nouveau.so.1 (0xb777c000+0x2dff) [0xb777edff]
[   771.457] 13: /usr/lib/libdrm_nouveau.so.1 (nouveau_bo_map_range+0x10a) [0xb777f4ea]
[   771.457] 14: /usr/lib/libdrm_nouveau.so.1 (nouveau_bo_map+0x32) [0xb777f5e2]
[   771.457] 15: /usr/lib/xorg/modules/drivers/nouveau_drv.so (0xb7159000+0x5497) [0xb715e497]
[   771.457] 16: /usr/lib/xorg/modules/libexa.so (0xb7114000+0xbe93) [0xb711fe93]
[   771.457] 17: /usr/bin/Xorg (0x8048000+0x16ad63) [0x81b2d63]
[   771.458] 18: /usr/bin/Xorg (0x8048000+0xa9986) [0x80f1986]
[   771.458] 19: /usr/lib/xorg/modules/libexa.so (exaGetPixmapFirstPixel+0x94) [0xb71287d4]
[   771.458] 20: /usr/lib/xorg/modules/libexa.so (0xb7114000+0x109e0) [0xb71249e0]
[   771.458] 21: /usr/bin/Xorg (0x8048000+0xe426b) [0x812c26b]
[   771.458] 22: /usr/bin/Xorg (CompositePicture+0x21e) [0x8120b6e]
[   771.458] 23: /usr/bin/Xorg (0x8048000+0xde66c) [0x812666c]
[   771.458] 24: /usr/bin/Xorg (0x8048000+0xd9461) [0x8121461]
[   771.458] 25: /usr/bin/Xorg (0x8048000+0x2d9d7) [0x80759d7]
[   771.458] 26: /usr/bin/Xorg (0x8048000+0x207cc) [0x80687cc]
[   771.458] 27: /lib/libc.so.6 (__libc_start_main+0xf3) [0xb7334003]
[   771.458] 28: /usr/bin/Xorg (0x8048000+0x20ae1) [0x8068ae1]

The kernel log shows the following suspicious messages:
...
[  765.361432] [drm] nouveau 0000:00:0d.0: fail pre-validate sync
[  765.361442] [drm] nouveau 0000:00:0d.0: validate vram_list
[  765.361452] [drm] nouveau 0000:00:0d.0: validate: -16
...
 
I'm using openSUSE 12.1 (kernel 3.1.9, Xorg 7.6). I also tried kernel 3.3.0-rc7.
Comment 1 Frank Schaefer 2012-03-13 13:30:02 UTC
Created attachment 58394 [details]
Xorg.0.log
Comment 2 Johannes Obermayr 2012-03-13 13:43:32 UTC
Please install drm-nouveau-kmp-{flavor}, libllvm, libdrm, Mesa and xorg-x11-driver-video from home:jobermayr together with kernel from openSUSE:12.1:Update. Only bugs present in them are valid.
Comment 3 Frank Schaefer 2012-03-13 14:07:44 UTC
Ok, I will try that...
The kernel is already up-to-date (3.1.9).

Please confirm that the following changes are correct:

drm-nouveau-kmp-desktop: currently not installed (?!) => 20120309.0650_k3.1.9_1.4-1.1
libllvm: currently not installed (?!) => 3.0.99.svn20120312
libdrm: 2.4.26-15.1.2 => no update available. Do you mean libdrm2 (2.4.99git20120310.1721-2.1) ?
Mesa: 7.11 => 8.0.99
xorg-x11-driver-video: 7.6 => 7.6.99

What about xorg-x11-driver-nouveau ?
Comment 4 Johannes Obermayr 2012-03-13 14:20:54 UTC
Yes. Forgot to mention new libxcb package.
(Versions may change in a few hours ...)
Comment 5 Frank Schaefer 2012-03-16 02:13:06 UTC
Got the same hang a few minutes ago with updated packages. :-(
Comment 6 dacer 2012-04-09 03:41:00 UTC
seems similar to mine, not sure if it is the same problem, bug 48434.

Suggest, login with ssh from another machine (I used to use android phone) and run as root "killall -s 9 -I xorg", this kill x and restart gnome display manager (gdm3) and then I can login again. With this method you avoid to reset and damage your filesystem.
Comment 7 Frank Schaefer 2012-07-17 21:05:54 UTC
Still happens (even more often than before) with

kernel 3.4.5
Mesa 8.0.99.git20120707
libdrm 2.4.99.git20120706
libllvm 3.1.99.svn20120708


The kernel logs now show much more of these lines:

[drm] nouveau 0000:00:0d.0: fail pre-validate sync
[drm] nouveau 0000:00:0d.0: validate vram_list
[drm] nouveau 0000:00:0d.0: validate: -16
[drm] nouveau 0000:00:0d.0: cal_space: -16
[drm] nouveau 0000:00:0d.0: cal_space: -16
[drm] nouveau 0000:00:0d.0: cal_space: -16
...


This bug is really easy to trigger, all you need to do is to enable KDE standard desktop effects => lockup in less than 5 minutes !
Comment 8 Frank Schaefer 2012-07-17 21:08:04 UTC
Created attachment 64332 [details]
Updated Xorg.0.log
Comment 9 Frank Schaefer 2012-07-17 21:08:44 UTC
Created attachment 64333 [details]
kernel messages
Comment 10 Robert-André Mauchin 2012-07-22 19:31:18 UTC
Hello,

I'm using the same hardware (Geforce 6150/nForce 430) and experiencing the same problem since months, at least since GNOME 3.0 release with compositing enabled, and then with KDE 4.7 with compositing enabled also.

Currently I'm running Fedora 17 with these:
kernel 3.4.6-2
libdrm 2.4.33
mesa-dri-drivers 8.0.3
xorg-x11-drv-nouveau 37.20120306gitf5d1cd2.fc17

The problem seems to occur randomly, but often while typing text in a Firefox window.

Following the Debugging documentation, I have been able to log with SSH and use gdb. From here, I can see that Xorg does not crash, the display and input just seem "frozen".

Xorg log gives the following backtrace:

[ 10897.517] [mi] EQ overflowing.  Additional events will be discarded until existing events are processed.
[ 10897.539] 
[ 10897.540] Backtrace:
[ 10897.769] 0: /usr/bin/X (xorg_backtrace+0x36) [0x4650a6]
[ 10897.769] 1: /usr/bin/X (mieqEnqueue+0x26b) [0x55111b]
[ 10897.769] 2: /usr/bin/X (0x400000+0x48062) [0x448062]
[ 10897.783] 3: /usr/lib64/xorg/modules/input/evdev_drv.so (0x7f69244da000+0x60e4) [0x7f69244e00e4]
[ 10897.783] 4: /usr/bin/X (0x400000+0x80557) [0x480557]
[ 10897.783] 5: /usr/bin/X (0x400000+0xa4730) [0x4a4730]
[ 10897.783] 6: /lib64/libpthread.so.0 (0x3ce5000000+0xefe0) [0x3ce500efe0]
[ 10897.783] 7: /lib64/libc.so.6 (ioctl+0x7) [0x3ce48ea5f7]
[ 10897.783] 8: /lib64/libdrm.so.2 (drmIoctl+0x28) [0x7f6926b78548]
[ 10897.783] 9: /lib64/libdrm.so.2 (drmCommandWrite+0x1b) [0x7f6926b7a77b]
[ 10897.784] 10: /lib64/libdrm_nouveau.so.1 (0x7f6926533000+0x3085) [0x7f6926536085]
[ 10897.784] 11: /lib64/libdrm_nouveau.so.1 (nouveau_bo_map_range+0x103) [0x7f69265366b3]
[ 10897.784] 12: /usr/lib64/xorg/modules/drivers/nouveau_drv.so (0x7f6926739000+0x6718) [0x7f692673f718]
[ 10897.784] 13: /usr/lib64/xorg/modules/libexa.so (0x7f6925ae7000+0xb007) [0x7f6925af2007]
[ 10897.784] 14: /usr/bin/X (0x400000+0x1600f3) [0x5600f3]
[ 10897.785] 15: /usr/bin/X (0x400000+0xc99f0) [0x4c99f0]
[ 10897.785] 16: /usr/bin/X (0x400000+0x3161a) [0x43161a]
[ 10897.785] 17: /usr/bin/X (0x400000+0x3444a) [0x43444a]
[ 10897.785] 18: /usr/bin/X (0x400000+0x23485) [0x423485]
[ 10897.785] 19: /lib64/libc.so.6 (__libc_start_main+0xf5) [0x3ce4821735]
[ 10897.785] 20: /usr/bin/X (0x400000+0x2375d) [0x42375d]
[ 10897.785] 
[ 10897.785] [mi] These backtraces from mieqEnqueue may point to a culprit higher up the stack.
[ 10897.785] [mi] mieq is *NOT* the cause.  It is a victim.

I don't know if what I did with "gdb" was the correct way but the output next.
Comment 11 Robert-André Mauchin 2012-07-22 19:33:16 UTC
Created attachment 64507 [details]
gdb "bt full" and "info frame"
Comment 12 Ilia Mirkin 2013-08-25 06:27:01 UTC
This basically means that your card has hung. The EQ overflowing thing is just X noticing that its event queue is filling up because the DDX is stuck. The kernel messages indicate various validation/space/vram errors. Does this still happen on the latest kernels? If so, could you provide an updated log that includes the part that happens when the hang happens?
Comment 13 Michal Suchanek 2013-08-26 13:37:47 UTC
Since the desktop efects are probably GL/Mesa based you can also try upgrading Mesa and libdrm
Comment 14 Ilia Mirkin 2013-09-27 02:03:21 UTC
No response to re-test request in a month. Closing as invalid.


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.