GPU lockup (EQ overflow) with nouveau experimental 3D drivers

[Erik B. Andersen]

It seems that Xorg will eventually lock up, using 100% CPU and not updating the screen, after the computer is running for a couple of hours. The only way to get out of it is kill Xorg with a signal 9.
I'm using the Nvidia Experimental 3D drivers from Jockey.
Let me know if there is any way I can help debug this.

[ 34127.580] [mi] EQ overflowing. The server is probably stuck in an infinite loop.
[ 34127.616]
Backtrace:
[ 34128.011] 0: /usr/bin/X (xorg_backtrace+0x26) [0x4a1586]
[ 34128.012] 1: /usr/bin/X (mieqEnqueue+0x1f4) [0x4a0854]
[ 34128.012] 2: /usr/bin/X (xf86PostMotionEventM+0x97) [0x47bcc7]
[ 34128.012] 3: /usr/bin/X (xf86PostMotionEvent+0xd0) [0x47bef0]
[ 34128.016] 4: /usr/lib/xorg/modules/input/synaptics_drv.so (0x7f0456299000+0x314f) [0x7f045629c14f]
[ 34128.016] 5: /usr/lib/xorg/modules/input/synaptics_drv.so (0x7f0456299000+0x4ed6) [0x7f045629ded6]
[ 34128.017] 6: /usr/bin/X (0x400000+0x6ac07) [0x46ac07]
[ 34128.017] 7: /usr/bin/X (0x400000+0x12303e) [0x52303e]
[ 34128.017] 8: /lib/libpthread.so.0 (0x7f045bac6000+0xfc80) [0x7f045bad5c80]
[ 34128.017] 9: /lib/libc.so.6 (ioctl+0x7) [0x7f045aae8597]
[ 34128.017] 10: /lib/libdrm.so.2 (drmIoctl+0x28) [0x7f045909a058]
[ 34128.017] 11: /lib/libdrm.so.2 (drmCommandWrite+0x1b) [0x7f045909c30b]
[ 34128.017] 12: /lib/libdrm_nouveau.so.1 (0x7f0458a57000+0x2b87) [0x7f0458a59b87]
[ 34128.017] 13: /lib/libdrm_nouveau.so.1 (nouveau_bo_map_range+0xfe) [0x7f0458a5a19e]
[ 34128.017] 14: /usr/lib/xorg/modules/drivers/nouveau_drv.so (0x7f0458c5d000+0x5bbe) [0x7f0458c62bbe]
[ 34128.017] 15: /usr/lib/xorg/modules/libexa.so (0x7f0457dda000+0x5e7f) [0x7f0457ddfe7f]
[ 34128.017] 16: /usr/lib/xorg/modules/libexa.so (0x7f0457dda000+0x863a) [0x7f0457de263a]
[ 34128.017] 17: /usr/lib/xorg/modules/libexa.so (0x7f0457dda000+0xf632) [0x7f0457de9632]
[ 34128.017] 18: /usr/lib/xorg/modules/libexa.so (0x7f0457dda000+0x101be) [0x7f0457dea1be]
[ 34128.017] 19: /usr/bin/X (0x400000+0xdd2c3) [0x4dd2c3]
[ 34128.017] 20: /usr/lib/xorg/modules/libexa.so (0x7f0457dda000+0x11248) [0x7f0457deb248]
[ 34128.017] 21: /usr/bin/X (0x400000+0xd81e1) [0x4d81e1]
[ 34128.018] 22: /usr/bin/X (0x400000+0x2da41) [0x42da41]
[ 34128.018] 23: /usr/bin/X (0x400000+0x21abe) [0x421abe]
[ 34128.018] 24: /lib/libc.so.6 (__libc_start_main+0xfe) [0x7f045aa2fd1e]
[ 34128.018] 25: /usr/bin/X (0x400000+0x21669) [0x421669]

ProblemType: Bug
DistroRelease: Ubuntu 11.04
Package: xserver-xorg-core 2:1.9.99.901+git20110131.be3be758-0ubuntu3
ProcVersionSignature: Ubuntu 2.6.37-12.26-generic 2.6.37
Uname: Linux 2.6.37-12-generic x86_64
Architecture: amd64
CompizPlugins: No value set for `/apps/compiz-1/general/screen0/options/active_plugins'
DRM.card0.LVDS.1:
 status: connected
 enabled: enabled
 dpms: On
 modes: 1280x800 1024x768 800x600 640x480 720x400 640x400 640x350
 edid-base64:
DRM.card0.VGA.1:
 status: disconnected
 enabled: disabled
 dpms: Off
 modes:
 edid-base64:
Date: Mon Feb 7 17:13:17 2011
DistUpgraded: Fresh install
DistroCodename: natty
DistroVariant: ubuntu
ExecutablePath: /usr/bin/Xorg
GraphicsCard: Subsystem: Hewlett-Packard Company Device [103c:30b7]
InstallationMedia: Ubuntu 11.04 "Natty Narwhal" - Alpha amd64 (20110127)
Lsusb:
 Bus 002 Device 001: ID 1d6b:0001 Linux Foundation 1.1 root hub
 Bus 001 Device 001: ID 1d6b:0002 Linux Foundation 2.0 root hub
MachineType: Hewlett-Packard Presario F500 (GF606UA#ABA)
ProcEnviron:
 LANG=en_US.UTF-8
 PATH=(custom, no user)
ProcKernelCmdLine: BOOT_IMAGE=/boot/vmlinuz-2.6.37-12-generic root=UUID=f1add3f0-0a81-4f8c-ac34-9dd6917e8bfb ro quiet splash vt.handoff=7
ProcKernelCmdLine_: BOOT_IMAGE=/boot/vmlinuz-2.6.37-12-generic root=UUID=f1add3f0-0a81-4f8c-ac34-9dd6917e8bfb ro quiet splash vt.handoff=7
Renderer: Unknown
SourcePackage: xorg-server
dmi.bios.date: 12/05/2007
dmi.bios.vendor: Hewlett-Packard
dmi.bios.version: F.1F
dmi.board.name: 30D3
dmi.board.vendor: Quanta
dmi.board.version: 65.3A
dmi.chassis.type: 10
dmi.chassis.vendor: Quanta
dmi.chassis.version: N/A
dmi.modalias: dmi:bvnHewlett-Packard:bvrF.1F:bd12/05/2007:svnHewlett-Packard:pnPresarioF500(GF606UA#ABA):pvrRev1:rvnQuanta:rn30D3:rvr65.3A:cvnQuanta:ct10:cvrN/A:
dmi.product.name: Presario F500 (GF606UA#ABA)
dmi.product.version: Rev 1
dmi.sys.vendor: Hewlett-Packard
version.compiz: compiz 1:0.9.2.1+glibmainloop4-0ubuntu9
version.libdrm2: libdrm2 2.4.23-1ubuntu3
version.libgl1-mesa-glx: libgl1-mesa-glx 7.10-1ubuntu1
version.xserver-xorg: xserver-xorg 1:7.6~3ubuntu3
version.xserver-xorg-video-ati: xserver-xorg-video-ati 1:6.13.2+git20110124.fadee040-0ubuntu4
version.xserver-xorg-video-intel: xserver-xorg-video-intel 2:2.14.0-1ubuntu6
version.xserver-xorg-video-nouveau: xserver-xorg-video-nouveau 1:0.0.16+git20110107+b795ca6e-0ubuntu4

[Bryce Harrington (bryce)]

Hi Erik,

I hate to say this but we don't provide development support for the 3D nouveau drivers in Ubuntu. I would redirect you to upstream, but they don't accept bug reports about the 3D drivers either (in fact, this is the primary reason we have not yet enabled it.)

So, if the drivers work for you that's great, if they don't, I'm afraid that's the end of the line. The next action would be to hack on the driver and send patches upstream, if you're technically inclined and have some time.

[Erik B. Andersen (azendale)]

Ok, well I guess I'll just let this bug be until they decide that they are ready to take bug reports.
To tell the truth, I tried these drivers about 6 months ago and I was impressed by how well they work now. So far, this bug was the only thing wrong with them. The bug happens so rarely that the other benefits (high resolution TTYs, nice boot screen) outweigh the problems (at least on a testing machine). So I'm definitely not complaining. I just though that maybe since the drivers were in Jockey, that maybe upstream was taking bugs.

[fejes (anthony-fejes)]

Figure I may as well comment and say I'm seeing the same thing, with a slight twist, since I'm using compiz as my window manager. In this case, compiz takes 105% cpu (slightly more than one of the 2 cores), and I can restart KDE by killing compiz and restarting it. Anyhow, given the above, I'm not expecting motion on this bug, but if it is picked up at some point, It might help to have a second report.

Thanks.

[fejes (anthony-fejes)]

I've switched back to the nvidia drivers and still have this behaviour, so it doesn't seem to be a part of the experimental 3d drivers.

[harrier (harrier77)]

I am having the problem with nvidia current on Natty;
4269.669] (II) XKB: reuse xkmfile /var/lib/xkb/server-0A497B65DAE1EE8CD2CCDAAA62D221D67888DBAC.xkm
[ 4738.273] (II) Open ACPI successful (/var/run/acpid.socket)
[ 4738.361] (II) NVIDIA(0): Setting mode "nvidia-auto-select"
[ 4740.001] (--) SynPS/2 Synaptics TouchPad: touchpad found
[ 4747.085] (II) XKB: reuse xkmfile /var/lib/xkb/server-0A497B65DAE1EE8CD2CCDAAA62D221D67888DBAC.xkm
[ 4762.450] (II) XKB: reuse xkmfile /var/lib/xkb/server-0A497B65DAE1EE8CD2CCDAAA62D221D67888DBAC.xkm
[ 4811.820] (II) XKB: reuse xkmfile /var/lib/xkb/server-0A497B65DAE1EE8CD2CCDAAA62D221D67888DBAC.xkm
[ 5371.763] (II) Open ACPI successful (/var/run/acpid.socket)
[ 5371.853] (II) NVIDIA(0): Setting mode "nvidia-auto-select"
[ 5373.582] (--) SynPS/2 Synaptics TouchPad: touchpad found
[ 5377.634] (II) XKB: reuse xkmfile /var/lib/xkb/server-0A497B65DAE1EE8CD2CCDAAA62D221D67888DBAC.xkm
[ 5385.219] (II) XKB: reuse xkmfile /var/lib/xkb/server-0A497B65DAE1EE8CD2CCDAAA62D221D67888DBAC.xkm
[ 5401.172] (II) XKB: reuse xkmfile /var/lib/xkb/server-0A497B65DAE1EE8CD2CCDAAA62D221D67888DBAC.xkm
[ 5805.453] (WW) NVIDIA(0): WAIT (2, 6, 0x8000, 0x0000f620, 0x000022a8)
[ 5811.027] [mi] EQ overflowing. The server is probably stuck in an infinite loop.
[ 5811.029]
Backtrace:
[ 5811.106] 0: /usr/bin/X (xorg_backtrace+0x3b) [0x80eab2b]
[ 5811.107] 1: /usr/bin/X (mieqEnqueue+0x1ab) [0x80e9d0b]
[ 5811.107] 2: /usr/bin/X (xf86PostMotionEventM+0xbf) [0x80c4b6f]
[ 5811.107] 3: /usr/bin/X (xf86PostMotionEvent+0x9a) [0x80c4d9a]
[ 5811.107] 4: /usr/lib/xorg/modules/input/synaptics_drv.so (0x97b000+0x4964) [0x97f964]
[ 5811.107] 5: /usr/lib/xorg/modules/input/synaptics_drv.so (0x97b000+0xad75) [0x985d75]
[ 5811.107] 6: /usr/lib/libutouch-grail.so.1 (grail_pull+0x2ec) [0xb573cc]
[ 5811.107] 7: /usr/lib/xorg/modules/input/synaptics_drv.so (0x97b000+0xab31) [0x985b31]
[ 5811.107] 8: /usr/lib/xorg/modules/input/synaptics_drv.so (0x97b000+0x6722) [0x981722]
[ 5811.108] 9: /usr/bin/X (0x8048000+0x6b41f) [0x80b341f]
[ 5811.108] 10: /usr/bin/X (0x8048000+0x12ddb6) [0x8175db6]
[ 5811.108] 11: (vdso) (__kernel_sigreturn+0x0) [0xe71400]

[Erik B. Andersen (azendale)]

@ harrier
What drivers are you using?

@fejes
Do you have a synaptics touchpad? The machine I had this bug occur on has a synaptics touch pad and harrier's backtrace appears to mention some synaptics driver stuff.

[fejes (anthony-fejes)]

@Eric

No, this is happening on a desktop machine without a touch pad, although it's possible that synapics drivers are being loaded at some point, unintentionally. I'll take a look to see if that's the case next time I'm on that computer. My workaround was to drop compiz, at which point the problem seemed to go away. (It's a shame, however, because I won't be able to upgrade any of my other computers until this bug has been fixed.)