Xorg hangs and/or restarts

Bug #694611 reported by Chris on 2010-12-26
22
This bug affects 3 people
Affects Status Importance Assigned to Milestone
NVIDIA Drivers Ubuntu
Undecided
Unassigned
nvidia-graphics-drivers (Ubuntu)
Undecided
Unassigned

Bug Description

Having a latop with an nVidia Quadro FX 2800M and troubled when using the proprietary driver. :(

I've been using the "Additional Drivers" tool to select the recommended "current" driver. The result were frequent (~every few hours) freezing screens or resets/restarts.
As this was no way to work, I've used the older binary driver of the "Additional Drivers" tool. The few days that I used that driver, everything seemed to be stable - although sometimes quite slow and not smooth (the CPU is an Core i7 so that behaviour is unexpected an unacceptable) during easy tasks like web surfing and moving a window around...
So I've added the PPA to use the latest nVidia binary driver.

The system seemed more stable now and it's fully responsive, as expected.

But now I had a freezing screen again. :(

At least I could log in remotely now and see that Xorg used 100% CPU (of one of the cores, that is) and I could save the log files that showed a clear error now. (-> cf. attachments)

Chris (mail-christianmayer) wrote :
Chris (mail-christianmayer) wrote :
Chris (mail-christianmayer) wrote :
Chris (mail-christianmayer) wrote :

For the statistics:

* Distribution is Kubunutu 10.10

* First driver version was: 260.19.06 (I guess. The splash screen also showed "Beta Driver" although it's the normal 10.10 driver)
* The it was: 173.14.28-0ubuntu1
* Now I'm using "nvidia-current" in version 260.19.29-0ubuntu1~xup~maverick3

Chris (mail-christianmayer) wrote :
Download full text (3.3 KiB)

I just had another crash.

The last lines in the kern.log before the restart were:

Dec 28 18:11:14 obiwan kernel: [28453.648355] dell-wmi: Received unknown WMI event (0x11)
Dec 28 20:14:36 obiwan kernel: [35843.150921] dell-wmi: Received unknown WMI event (0x11)
Dec 28 20:43:51 obiwan kernel: [37595.833980] dell-wmi: Received unknown WMI event (0x11)
Dec 28 21:12:38 obiwan kernel: [39319.716735] NVRM: Xid (0001:00): 6, PE0003
Dec 28 21:14:48 obiwan kernel: [39449.831650] NVRM: Xid (0001:00): 6, PE0001
Dec 28 21:14:56 obiwan kernel: [39457.838446] NVRM: Xid (0001:00): 6, PE0001
Dec 28 21:14:57 obiwan kernel: [39458.944552] dell-wmi: Received unknown WMI event (0x11)
Dec 28 21:15:03 obiwan kernel: [39460.624556] NVRM: Xid (0001:00): 6, PE007e
Dec 28 21:15:03 obiwan kernel: [39460.626992] NVRM: Xid (0001:00): 6, PE007e
Dec 28 21:15:03 obiwan kernel: [39460.629315] NVRM: Xid (0001:00): 6, PE007e
Dec 28 21:15:03 obiwan kernel: [39460.631649] NVRM: Xid (0001:00): 6, PE007e
Dec 28 21:15:03 obiwan kernel: [39460.633939] NVRM: Xid (0001:00): 6, PE007e
Dec 28 21:15:03 obiwan kernel: [39462.632943] NVRM: os_schedule: Attempted to yield the CPU while in atomic or interrupt context
Dec 28 21:15:03 obiwan kernel: [39464.629645] NVRM: os_schedule: Attempted to yield the CPU while in atomic or interrupt context
Dec 28 21:15:43 obiwan kernel: [39504.519426] dell-wmi: Received unknown WMI event (0x11)
Dec 28 21:15:44 obiwan kernel: [39505.112253] mtrr: base(0xf3000000) is not aligned on a size(0xe00000) boundary
Dec 28 21:16:16 obiwan kernel: [39537.790124] mtrr: no MTRR for f3000000,e00000 found
Dec 28 21:17:33 obiwan kernel: Kernel logging (proc) stopped.

And the deamon.log (the last entry before those lines were over ten hours ago, i.e. irrelevant):
Dec 28 21:14:59 obiwan kdm[1310]: X server for display :0 terminated unexpectedly
Dec 28 21:14:59 obiwan acpid: client 1338[0:0] has disconnected
Dec 28 21:14:59 obiwan acpid: client 1338[0:0] has disconnected
Dec 28 21:14:59 obiwan acpid: client connected from 8827[0:0]
Dec 28 21:14:59 obiwan acpid: 1 client rule loaded
Dec 28 21:15:19 obiwan kdm[1310]: X server startup timeout, terminating
Dec 28 21:15:39 obiwan kdm[1310]: X server termination timeout, killing
Dec 28 21:15:43 obiwan kdm[1310]: Failed to start X server. Starting failsafe X server.
Dec 28 21:15:43 obiwan acpid: client 8827[0:0] has disconnected
Dec 28 21:15:43 obiwan acpid: client connected from 8852[0:0]
Dec 28 21:15:43 obiwan acpid: 1 client rule loaded
Dec 28 21:16:17 obiwan acpid: client 8852[0:0] has disconnected
Dec 28 21:16:17 obiwan acpid: client connected from 8882[0:0]
Dec 28 21:16:17 obiwan acpid: 1 client rule loaded
Dec 28 21:16:37 obiwan kdm[1310]: X server startup timeout, terminating
Dec 28 21:16:57 obiwan kdm[1310]: X server termination timeout, killing
Dec 28 21:17:07 obiwan kdm[1310]: X server is stuck in D state; leaving it alone
Dec 28 21:17:07 obiwan kdm[1310]: Failed to start X server. Starting failsafe X server.
Dec 28 21:17:33 obiwan acpid: client 8882[0:0] has disconnected
Dec 28 21:17:33 obiwan acpid: client connected from 8915[0:0]
Dec 28 21:17:33 obiwan acpid: 1 client rule loaded

Due to the X server restart I do...

Read more...

Chris (mail-christianmayer) wrote :

The crashes are continuing and I have saved log files from two more :(

Can someone help me to sort this issue out, please?!?

Chris (mail-christianmayer) wrote :

It's quite frustrating, to see that I'm left alone with my crashing system...

To continue the monologue:

Following that suspected culprit of http://www.nvnews.net/vbulletin/showthread.php?t=155342 I've set the performance setting of the GPU to a fixed value as it might be the dynamic setting of the powermizer that's one part of the problems.

As the slowest setting of the GPU also caused one lock up, I've chosen the second slowest - and the system didn't show a lock up since then.

The relevant part of my xorg.conf now looks like:

Section "Device"
        Identifier "Device0"
        VendorName "NVIDIA Corporation"
        Driver "nvidia"
        Option "NoLogo" "True"
        Option "RegistryDwords" "PowerMizerEnable=0x1; PerfLevelSrc=0x2222; PowerMizerLevel=0x2; PowerMizerDefault=0x2; PowerMizerDefaultAC=0x2"
EndSection

Chris (mail-christianmayer) wrote :

With those settings the system didn't have any hang ups any more.

=> This is a stable work around for me

BUT: This is not a final solution, as I'm wasting GPU power that I payed for and that I sometimes need!

(Hm, still noone interested?)

Changed in nvidia-graphics-drivers (Ubuntu):
status: New → Confirmed

My KDE crashes, too, after upgrading to natty. I use:
Kubuntu 11.04
Kernel : 2.6.38-11-generic #49-Ubuntu SMP Mon Aug 29 20:47:58 UTC 2011 i686 i686 i386 GNU/Linux
nVidia Corporation G98 [GeForce 8400 GS]

KDE hangs with following log:
Sep 13 14:21:32 taurus4 kernel: [87015.147079] NVRM: Xid (0000:04:00): 13, 0001 00000000 00005097 000015e0 00000000 00000100
Sep 13 14:22:26 taurus4 kernel: [87068.448068] NVRM: Xid (0000:04:00): 8, Channel 00000001
Sep 13 14:22:28 taurus4 kernel: [87070.448031] NVRM: os_schedule: Attempted to yield the CPU while in atomic or interrupt context
Sep 13 14:22:43 taurus4 kernel: [87085.448072] NVRM: Xid (0000:04:00): 8, Channel 00000001
Sep 13 14:22:45 taurus4 kernel: [87087.448034] NVRM: os_schedule: Attempted to yield the CPU while in atomic or interrupt context

thanks for any help

Knut (knutjorgen) wrote :

I have similar problem with Quadro NVS 4200.

To post a comment you must log in.
This report contains Public information  Edit
Everyone can see this information.

Other bug subscribers