Kubuntu Jaunty 64 bits hangups often with nvidia-glx-180

[Iñaki Baz Castillo]

Binary package hint: nvidia-glx-180

My system:

- AMD Turion(tm) 64 bits X2 Mobile Technology TL-52 (2 cores)
- Kubuntu Jaunty 64 bits
- nVidia Corporation G72M [Quadro NVS 110M/GeForce Go 7300] (rev a1)
  - nvidia drivers 180.44-0ubuntu1, 177.82-0ubuntu0.1

Using linux-image-2.6.27-11-generic or linux-image-2.6.28-11-generic I have a mostly stable system if I use nvidia driver nvidia-glx-177, no hang-ups.
But since I upgraded to the latest nvidia 180.44-0ubuntu1 my system hangs-up very very often (i.e., after 20 minutes from swith on my laptop). This is, it gets completely frozen and I must restart it by pressing the power button for 4 seconds.

I've tryed dissabling GLX effects but it doesn't matter.
Unfortunatelly I have no idea of which the cause could be. Any suggestion on how to inspect it is really welcome.

PD: I've changed the title and description of this bug since I thought is was related to kernel version (linux-image-2.6.28-11-generic) but I've confirmed that the hangups occur with any kernel when using nvidia 180.

[lspci]
00:00.0 RAM memory [0500]: nVidia Corporation C51 Host Bridge [10de:02f7] (rev a2)
     Subsystem: ASUSTeK Computer Inc. Device [1043:1367]
02:00.0 VGA compatible controller [0300]: nVidia Corporation G72M [Quadro NVS 110M/GeForce Go 7300] [10de:01d7] (rev a1)
     Subsystem: ASUSTeK Computer Inc. Device [1043:13a2]

[Jeremy Sequoia (jeremyhu)]

Is it completely hung? Can you ssh in from another box? I noticed a problem with the nvidia drivers resuming from suspend, but I could still ssh in from another box to debug. X was consuming all my CPU and it looks like it was spinning in the nvidia drivers with this backtrace:

(gdb) bt
#0 0x00007f8c7b3a5ad6 in _nv001302X () from /usr/lib/xorg/modules/drivers//nvidia_drv.so
#1 0x00007f8c7b646e6a in ?? () from /usr/lib/xorg/modules/drivers//nvidia_drv.so
#2 0x00007f8c7b647312 in ?? () from /usr/lib/xorg/modules/drivers//nvidia_drv.so
#3 0x00007f8c7b647976 in ?? () from /usr/lib/xorg/modules/drivers//nvidia_drv.so
#4 0x00007f8c7b3a5310 in _nv000939X () from /usr/lib/xorg/modules/drivers//nvidia_drv.so
#5 0x00007f8c7b6617d0 in ?? () from /usr/lib/xorg/modules/drivers//nvidia_drv.so
#6 0x00007f8c7b3a6ebd in _nv001769X () from /usr/lib/xorg/modules/drivers//nvidia_drv.so
#7 0x00007f8c7b63ed59 in ?? () from /usr/lib/xorg/modules/drivers//nvidia_drv.so
#8 0x00007f8c7b63ff8c in ?? () from /usr/lib/xorg/modules/drivers//nvidia_drv.so
#9 0x00007f8c7b625c2e in ?? () from /usr/lib/xorg/modules/drivers//nvidia_drv.so
#10 0x000000000048c962 in ?? ()
#11 0x000000000049dc59 in ?? ()
#12 0x000000000048624c in xf86Wakeup ()
#13 0x0000000000451e6b in WakeupHandler ()
#14 0x00000000004ef5ef in WaitForSomething ()
#15 0x000000000044e020 in Dispatch ()
#16 0x0000000000433d8d in main ()

[Iñaki Baz Castillo (ibc-aliax)]

I've not tried to access via SSH to my computer when it hangups due to this probem, but under my experience (I know how a complete crash look like) I would say that the computer gets completely frozen (i.e, the laptop leds don't react, the power button doesn't work except if I press it for 4 seconds...).
Anyway I will try to test it, but I'm 99% sure that I couldn't access via SSH.

[Thomas Meixner (tom-meixner)]

I have a similar problem on my Dell e6400 with an NVidia NVS 160m graphics card.

Depending on how hard the crash is I can still use the system or switch to a console. Here is what I have in my Xorg.0.log.old after a reboot. It looks to me the NVidia Driver crashes.

(II) NVIDIA(0): Initialized GPU GART.
(WW) NVIDIA(0): The NVIDIA X driver has encountered too many errors. Falling
(WW) NVIDIA(0): back to write-back cached memory.

Here is a screenshot of what the screen looks like when it crashes.
http://www.avecstyle.ca/screenshots/2.jpg

[Patrick Leliveld (p-leliveld)]

I have the same problem with my Compal IFL90 with an Nvidia 8600 VGA card. It freezes on random occasions.It is then impossible to ping the machine or to SSH to it. It is also impossible to switch to a local text console. The only solution is to use the power button.

[Iñaki Baz Castillo (ibc-aliax)]

Patrick, does it occur to you just with nvidia 180.44? For me, using nvidia 177 works ok while using 180.44 hangs my system really often.

[Thomas Meixner (tom-meixner)]

I just realized while going through my Xorg.0.log file that the RAM of my GfX Card (NVidia NVS 160m Quattro with 256MB RAM) is reported incorrectly:

(II) NVIDIA(0): NVIDIA GPU Quadro NVS 160M (G98) at PCI:1:0:0 (GPU-0)
(--) NVIDIA(0): Memory: 524288 kBytes

Would this cause problems if the NVidia Driver tries to access additional memory which isn't there? I don't know whether this might be related but I can crash KWin when I load a 200em Header in a browser (huge but works on other systems) .

[Bryce Harrington (bryce)]

Hi ibc-aliax,

Please attach the output of `lspci -vvnn`, and attach your /var/log/Xorg.0.log (and maybe Xorg.0.log.old) file from after reproducing this issue. If you've made any customizations to your /etc/X11/xorg.conf please attach that as well.

[This is an automated message. Apologies if it has reached you inappropriately; please just reply to this message indicating so.]

[Iñaki Baz Castillo (ibc-aliax)]

Ok, I attach /var/log/Xorg.1.log which is the log from some time ago when I used nvidia-driver 180 (using 173 I have not this issue).
I also attach the output of `lspci -vvnn` and my current xorg.conf (created with nvidia-setting with no customizations).

[aurelieng (aurelien-grosdidier)]

Parial crash here, with a scrambled screen with the following lines in /var/log/Xorg.0.log:

(WW) NVIDIA(0): The NVIDIA X driver has encountered too many errors. Falling
(WW) NVIDIA(0): back to write-back cached memory.

Configuration:

01:00.0 VGA compatible controller: nVidia Corporation Quadro FX 2700M (rev a1)
        Subsystem: Dell Device 019b
        Control: I/O+ Mem+ BusMaster+ SpecCycle- MemWINV- VGASnoop- ParErr- Stepping- SERR- FastB2B- DisINTx-
        Status: Cap+ 66MHz- UDF- FastB2B- ParErr- DEVSEL=fast >TAbort- <TAbort- <MAbort- >SERR- <PERR- INTx-
        Latency: 0
        Interrupt: pin A routed to IRQ 16
        Region 0: Memory at f5000000 (32-bit, non-prefetchable) [size=16M]
        Region 1: Memory at e0000000 (64-bit, prefetchable) [size=256M]
        Region 3: Memory at f2000000 (64-bit, non-prefetchable) [size=32M]
        Region 5: I/O ports at df00 [size=128]
        [virtual] Expansion ROM at f6e00000 [disabled] [size=512K]
        Capabilities: <access denied>
        Kernel driver in use: nvidia
        Kernel modules: nvidia, nvidiafb

See attachments for details.

[Michael Marley (mamarley)]

"Would this cause problems if the NVidia Driver tries to access additional memory which isn't there?"

No, it wouldn't. Your card is new enough to support TurboCache, and it therefore shares some RAM with the system.

[Thomas Meixner (tom-meixner)]

I have reinstalled my kubuntu jaunty 64 Bit installation recently. This time I selected the 177 driver and since then I don't have problems anymore.
I couldn't help but notice that all problems reported here are also for the Quattro or NVS Series of NVidia Drivers for laptops.

[Patrick Leliveld (p-leliveld)]

I was not able to test my machine with Nvidia 177, but instead I've used the NV driver for a few days. To my suprise I experienced a few of the same hangups as well! So the crashes are probably not related to the Nvidia-driver.

It reminds me of the crashes I had during the first two months with Ubuntu 8.10. They were caused by a kernel bug, so I'm considering falling back to the intrepid kernel.

[Michael Marley (mamarley)]

You might also try the 2.6.30 kernel, and file a bug report with the kernel developers if this kernel has the bug, too.

[Thomas Meixner (tom-meixner)]

There has been a driver update to the 180 series of drivers.
Amongst the other issues mentioned I thought this was interesting:

Fixed stability problems with some GeForce 6200/7200/7300 GPUs on multi-core/SMP systems.

More information here:
http://www.nvidia.com/object/linux_display_amd64_180.60.html

[Iñaki Baz Castillo (ibc-aliax)]

@Thomas: Thanks for so good information. I will check it (first I must check if the new revision of driver 180 is updated in Jaunty repository). Thanks a lot.

[Iñaki Baz Castillo (ibc-aliax)]

Unfortunatelly Jaunty still offers nvidia driver 180.44-0ubuntu1 (while the fix Thomas mentioned is 188.60). So I must wait :)

[Thomas Meixner (tom-meixner)]

I have manually installed the nvidia drivers for 180.60 and my system is stable since. I upgraded recently to KDE 4.3 Beta 2 and tried at the same time 180.44 again which resulted in 2 crashes in one afternoon (pixelated screens etc ...)

I will keep testing this driver release from NVidia and report back since it has been only 1 day.

Is there any word on when and if the 180.60 Release will be available on Jaunty? I only found this release targeted for Karmic but not Jaunty:
https://launchpad.net/ubuntu/karmic/amd64/nvidia-180-kernel-source/180.60-0ubuntu1

 If it is a driver issue with stability issues on NVidia mobility cards which I'm pretty convinced now we have people with this configuration running into a problem atm. What can we do about it?

The 180.60 NVIDIA driver I'm using atm can be found under the link below. As a word of warning using the NVIDIA installer I messed up my system quite badly (installing several driver versions above each other) until I found this post on the nvidia forum which worked for me. If you are not comfortable in repairing your system from the command line (no X) just in case something goes wrong I would not recommend it:

general install howto on Nvidia forum:
http://www.nvnews.net/vbulletin/showthread.php?s=ce6d3ece98bfa6045bb2eb53a2a7d43d&t=72490
driver download:
http://www.nvidia.com/object/linux_display_amd64_180.60.html

[Iñaki Baz Castillo (ibc-aliax)]

Thomas, thanks a lot for your testing.
I hope 180.60 NVIDIA driver is ported to Jaunty.

[Thomas Meixner (tom-meixner)]

I've been running the NVIDIA180.60 which I installed manually for a week now and just experienced my first crash with the same symptoms: Psychedelic pixelated screen. The crashes on 180.44 were far more often though (once a day at least) and this time I was also still able to shut down the System since the Desktop was still usable. 173 was still the most stable for me without a crash I can recall.

Here the errors from the Xorg.0.log during the crash
The errormessages are the same we saw before:
....
(II) NVIDIA(0): Initialized GPU GART.
(II) NVIDIA(0): Initialized GPU GART.
(WW) NVIDIA(0): The NVIDIA X driver has encountered too many errors. Falling
(WW) NVIDIA(0): back to write-back cached memory.
(II) NVIDIA(0): Initialized GPU GART.
....

I'm taking the issue to the NVIDIA forum for now and see how far I get there. Btw are you all KDE users or does this problem happen on Gnome too?

[Bryce Harrington (bryce)]

I've posted a new version of the -nvidia driver to our xorg-edgers PPA,
would you mind testing it either on Jaunty or Karmic and see if it
resolves this bug?

Get nvidia-graphics-drivers-180 - 185.18.14 here:

  https://edge.launchpad.net/~xorg-edgers/+archive/ppa

[Iñaki Baz Castillo (ibc-aliax)]

I'm already using the 185.18.14 nvidia driver (in my 64 bits laptop).
I will comment soon about its stability.

[Iñaki Baz Castillo (ibc-aliax)]

Two days using 185.18.14 nvidia driver with no crash. I think it works nice.

[trelayne (trelayne)]

Hi all, I continue to get freezing even after manually upgrading
from the default ubuntu 180.44 to the 185.18.14 nvidia driver .

I have:

- Jaunty 32-bit version running on a Dell Precision 4400 laptop
- Quadro FX 770M (514 MB)
- running compiz

Compiz graphical maneuvers (workspace switching, cube rotate
, etc.) work well initially. Then the screen freezes in the middle of
an effect, losing the mouse and keyboard. I'm usually able to
recover after 30 seconds, but often I'm left with remnants of
window images for a background, and some icons are missing
from the Firefox toolbar. And on more rare occasions, the X server
crashes and sends me back to the login screen.

Here are parts of the Xorg.0.log file:

. . .
(II) NVIDIA(0): Initialized GPU GART.
(II) NVIDIA(0): Initialized GPU GART.
(WW) NVIDIA(0): The NVIDIA X driver has encountered too many errors. Falling
(WW) NVIDIA(0): back to write-back cached memory.
(II) NVIDIA(0): Initialized GPU GART.
(II) NVIDIA(0): Initialized GPU GART.
. . .

I've attached the full log. But note that
the problem starts near the spot where you see
the above text.

I'm considering downgrading the nvidia driver.
Appreciate any insight.

[Thomas Meixner (tom-meixner)]

I've been using 185.18.14 (manually compiled from NVIDIA website) on my dell latitude e6400 64Bit for almost a month and haven't experienced a single crash since.

[Iñaki Baz Castillo (ibc-aliax)]

After 6 days using 185.18.14 I've experimented 2-3 crashes. Not sure if they occur due to nvidia driver however, but I assume yes.

[trelayne (trelayne)]

@Thomas,

when you say manually compiled, I assume you mean the procedure where you do:

sh NVIDIA-Linux-x86-185.18.14-pkg1.run ?

[Thomas Meixner (tom-meixner)]

@trelayne

>when you say manually compiled, I assume you mean the procedure where you do:
>sh NVIDIA-Linux-x86-185.18.14-pkg1.run ?

yes, I used the installer and the instructions found on the NVIDIA Forum. I posted some more details in my post on 2009-06-12.

It's strange - this bug seems quite picky since many people with the same driver version still report problems. For the record I'm also running KDE 4.3 RC1.

[trelayne (trelayne)]

[SYSTEM: Ubuntu Jaunty 32-bit, 4GB RAM, nVidia Quadro FX 770M -- 512MB ]

Hi again all,

after some research I'm wondering whether I might be dealing with two problems,
both fixable by tampering with /proc/mtrr :

(1) the way main memory is allocated to my graphics card:

     http://ubuntuforums.org/showthread.php?t=1130582 (Intel specific, but may apply for other
     manufacturers)

     Look at the MTRR stuff specifically. A technique for fixing /proc/mtrr settings
     is found here (i.e. using "write-combining"):

     https://bugs.launchpad.net/ubuntu/+source/xserver-xorg-video-intel/+bug/314928

(2) following material above, I found that the memory reported by lspci is different
     from what I see in /proc/mtrr (256MB versus 512MB). Could this be related to my problems?

---------------------------
# lspci -vvnn | less
        01:00.0 VGA compatible controller: nVidia Corporation Quadro FX 770M (rev a1)
        Subsystem: Dell Device 0250
        Flags: bus master, fast devsel, latency 0, IRQ 16
        Memory at f5000000 (32-bit, non-prefetchable) [size=16M]
        Memory at e0000000 (64-bit, prefetchable) [size=256M]
        Memory at f2000000 (64-bit, non-prefetchable) [size=32M]
        I/O ports at df00 [size=128]
        [virtual] Expansion ROM at f4000000 [disabled] [size=512K]
        Capabilities: <access denied>
        Kernel driver in use: nvidia
        Kernel modules: nvidia, nvidiafb
---------------------------

# cat /proc/mtrr
reg00: base=0x000000000 ( 0MB), size=32768MB, count=1: write-back
reg01: base=0x0e0000000 ( 3584MB), size= 512MB, count=1: uncachable

[Lauri Niskanen (ape3000)]

I have this bug on Jaunty 64-bit using GNOME. I have Geforce GTX 285 and 185.18.14 drivers.

The problem occurs quite rarely. Usually after my computer has been on for the whole day. The system freezes randomly for three seconds and I get a lot of graphical glitches. These messages are spammed on Xorg.log:

...
(II) NVIDIA(0): Initialized GPU GART.
(II) NVIDIA(0): Initialized GPU GART.
(II) NVIDIA(0): Initialized GPU GART.
(II) NVIDIA(0): Initialized GPU GART.
(WW) NVIDIA(0): The NVIDIA X driver has encountered too many errors. Falling
(WW) NVIDIA(0): back to write-back cached memory.
(II) NVIDIA(0): Initialized GPU GART.
(II) NVIDIA(0): Initialized GPU GART.
(II) NVIDIA(0): Initialized GPU GART.
(II) NVIDIA(0): Initialized GPU GART.
...

[Tomáš Myšík (gapon)]

The same for me, it was OK on Jaunty but happens quite a lot with Karmic :/
(Sony Vaio VGN-Z21XN, nVidia GeForce 9300M GS, nvidia-glx-185 [185.18.36-0ubuntu9])

Very annoying, hopefully it will be fixed soon.

[Tomáš Myšík (gapon)]

It seems that the latest nvidia driver fixes the problem.
(Using nvidia-glx-190 [190.42-0ubuntu1~karmic~nvidiavdpauppa3] from nvidia-vdpau PPA.)

[Sasquatch (sasquatch)]

I got this issue last night on my PC. I got this in my logs:
Xorg.0.log:
(WW) Apr 07 21:48:49 NVIDIA(0): The NVIDIA X driver has encountered too many errors. Falling
(WW) Apr 07 21:48:49 NVIDIA(0): back to write-back cached memory.
(EE) Logitech USB-PS/2 Optical Mouse: Read error: No such device

Syslog:
Apr 7 21:48:48 Lain kernel: [44952.795391] NVRM: Xid (0005:00): 13, 0000 e0015f00 0000009f 00000308 000d0071 00000008
Apr 7 21:48:48 Lain kernel: [44953.405094] NVRM: Xid (0005:00): 13, 0000 e0015f00 0000009f 00000308 000d0071 00000008
Apr 7 21:48:48 Lain kernel: [44953.428150] NVRM: Xid (0005:00): 13, 0000 e0015f00 0000009f 00000308 000d0071 00000008
Apr 7 21:48:48 Lain kernel: [44953.448307] NVRM: Xid (0005:00): 13, 0000 e0015f00 0000009f 00000308 000d0071 00000008
Apr 7 21:48:48 Lain kernel: [44953.474177] NVRM: Xid (0005:00): 13, 0000 e0015f00 0000009f 00000308 000d0071 00000008
Apr 7 21:48:48 Lain kernel: [44953.495281] NVRM: Xid (0005:00): 13, 0000 e0015f00 0000009f 00000308 000d0071 00000008
Apr 7 21:48:48 Lain kernel: [44953.517083] NVRM: Xid (0005:00): 13, 0000 e0015f00 0000009f 00000308 000d0071 00000008
Apr 7 21:48:48 Lain kernel: [44953.555508] NVRM: Xid (0005:00): 13, 0000 e0015f00 0000009f 00000308 000d0071 00000008

I left my computer around 21:30. Usually, I leave xscreensaver running to lock my system, but as I was re-encoding some files, I didn't activate it (in fact, I disabled it). When no screensaver runs, X would enable DPMS after 900 seconds (that's 15 minutes). Taken into account that the system was running, I'd have to guess that this problem occurred within 4 minutes after DPMS activation. This is also the first time I encounter this problem on my PC. X was taking all the CPU, messing up pulseaudio that was playing a stream from Audacious.
Also, both the keyboard and mouse did not respond to any input. Trying to change the LED for numlock and capslock failed. I had to SSH to the system and shut it down from there.

The system I'm using has a GeForce 7800 GTX with 256 MB of VRAM. Since it's a PC, there is no shared memory. I'm running Jaunty with Xfce4. nVidia driver 195.36.15 from the nVidia VDPAU PPA.

Aside from this nVidia issue, I've had the same kind of issue with my laptop that uses an Intel graphics card running Ubuntu Karmic with Xfce4. I would lock the system using xscreensaver and when it turns off the monitor, there is a fairly big chance it won't respond to anything and I have to force a power down or SSH to it and try to fix it that way. Sending the system to suspend or hibernate usually fixes it and allows me to resume work. I don't have any logs of that unfortunately. When I get the problem again, I will gather the logs for it too, so it can be compared to this issue, it might be related.

[Sasquatch (sasquatch)]

I got this similar issue on my laptop that I spoke about in the previous reply. I checked syslog and it did not show the same message as in this report. X didn't consume all CPU either. All I saw whas that the driver crashed and a reboot or standby/hibernate fixes it for the time being.
So, Intel and nVidia cards not giving video after some time or at certain circumstances: not related.

[Flames_in_Paradise (ellisistfroh-deactivatedaccount)]

Thanks to all of you for reporting this bug to Ubuntu. Jaunty Jackelope reached EOL on Oct.,23. 2010
See this document for currently supported Ubuntu releases:

https://wiki.ubuntu.com/Releases

Please upgrade to the latest version and re-test.
If the bug is still reproducible, increase the verbosity of the steps
to recreate it so we can try again.

Do feel free to report any other bugs you may find.

This bug-report will be closed by a bug-bot in a reasonable time.

[Launchpad Janitor (janitor)]

[Expired for nvidia-graphics-drivers-180 (Ubuntu) because there has been no activity for 60 days.]