| View previous topic :: View next topic |
| Author |
Message |
cfgauss
l33t
Joined: 18 May 2005
Posts: 740
Location: USA
|
 Posted: Wed Oct 19, 2016 11:55 pm Post subject: [SOLVED] Nvidia graphics card freeze |
 |
|
For the past week my ancient Nvidia GeForce 8600 GT graphics card using nvidia-drivers-340.98 has frozen once per boot. This last freeze was the most serious and didn't recover: I had to hard reboot. Here are the NVRM messages in /var/log/messages:
| Code: | Oct 19 18:30:15 localhost kernel: NVRM: GPU at PCI:0000:01:00: GPU-9c0eedb0-fcaa-d6b7-691a-b3598b871b17
Oct 19 18:30:15 localhost kernel: NVRM: Xid (PCI:0000:01:00): 69, Class Error: ChId 000c, Class 0000502d, Offset 00000250, Data ffffffff, ErrorCode 0000000c
Oct 19 18:30:15 localhost kernel: NVRM: Xid (PCI:0000:01:00): 69, Class Error: ChId 000c, Class 0000502d, Offset 00000250, Data ffffffff, ErrorCode 0000000c
Oct 19 18:30:16 localhost kernel: NVRM: Xid (PCI:0000:01:00): 13, Graphics Exception: ChID 000c, Class 00008297, Offset 00000e08, Data 04380000
Oct 19 18:30:16 localhost kernel: NVRM: Xid (PCI:0000:01:00): 69, Class Error: ChId 000e, Class 0000502d, Offset 00000250, Data ffffffff, ErrorCode 0000000c
Oct 19 18:30:17 localhost kernel: NVRM: Xid (PCI:0000:01:00): 13, Graphics Exception: ChID 000c, Class 00008297, Offset 00000e08, Data 04380000
Oct 19 18:30:17 localhost kernel: NVRM: Xid (PCI:0000:01:00): 69, Class Error: ChId 000c, Class 0000502d, Offset 00000250, Data ffffffff, ErrorCode 0000000c
Oct 19 18:30:18 localhost kernel: NVRM: Xid (PCI:0000:01:00): 13, Graphics Exception: ChID 000c, Class 00008297, Offset 00000e08, Data 04380000
Oct 19 18:30:44 localhost kernel: NVRM: Xid (PCI:0000:01:00): 69, Class Error: ChId 000e, Class 0000502d, Offset 00000250, Data ffffffff, ErrorCode 0000000c
Oct 19 18:30:45 localhost kernel: NVRM: Xid (PCI:0000:01:00): 8, Channel 00000007
Oct 19 18:30:47 localhost kernel: NVRM: os_schedule: Attempted to yield the CPU while in atomic or interrupt context
Oct 19 18:30:49 localhost kernel: NVRM: os_schedule: Attempted to yield the CPU while in atomic or interrupt context
Oct 19 18:30:49 localhost kernel: NVRM: GPU at 0000:01:00.0 has fallen off the bus. |
Any debugging hints will be gratefully received. In particular, I'd like to know if I need a new graphics card.
[SOLVED] I had an extra, identical video card but this one produced the same freezes and NVRM messages as the old one. In fact, I'm running VMware and my vmware.log complained of a bug in the 340.98 drivers and strongly suggested that I upgrade to newer drivers! So, I conclude that the NVRM messages listed above are due to buggy drivers rather than buggy hardware (as suggested by Hu, below). I bought a newer card which uses the latest drivers, 375.26, and have had no problems nor NVRM messages for three months. [/SOLVED]
Last edited by cfgauss on Tue Jan 10, 2017 2:49 am; edited 2 times in total |
|
| Back to top |
|
 |
Hu
Administrator
Joined: 06 Mar 2007
Posts: 23360
|
| Posted: Thu Oct 20, 2016 1:17 am Post subject: |
|
|
| As I read the eix output, that is an odd choice of version for x11-drivers/nvidia-drivers. It is marked as testing on both x86 and amd64, but there are newer stable and testing versions available. I suggest upgrading to the latest stable or testing version (as appropriate for your system) to see if the problem has been fixed. Before replacing the card, I would also try using the open source drivers. |
|
| Back to top |
|
|
cfgauss
l33t
Joined: 18 May 2005
Posts: 740
Location: USA
|
| Posted: Thu Oct 20, 2016 3:14 pm Post subject: |
|
|
| Hu wrote: | | As I read the eix output, that is an odd choice of version for x11-drivers/nvidia-drivers. It is marked as testing on both x86 and amd64, but there are newer stable and testing versions available. I suggest upgrading to the latest stable or testing version (as appropriate for your system) to see if the problem has been fixed |
For my ancient GeForce 8600 GT card, 340.98 is the latest driver.
. | Hu wrote: | | Before replacing the card, I would also try using the open source drivers. |
This is a good idea. Thanks. |
|
| Back to top |
|
|
Section_8
l33t
Joined: 22 May 2004
Posts: 627
|
| Posted: Thu Oct 20, 2016 5:52 pm Post subject: |
|
|
| I've had several video cards die in the past because the fan stopped working an the card would overheat and lock up at random times. Have you checked that out? |
|
| Back to top |
|
|
Hu
Administrator
Joined: 06 Mar 2007
Posts: 23360
|
| Posted: Fri Oct 21, 2016 1:27 am Post subject: |
|
|
Based on years of seeing weird nVidia driver bugs, I would suspect the proprietary nVidia driver of having a bug before I would suspect that the card has died. If newer versions have lost the ability to drive your card, that is unfortunate, but also very typical for nVidia as a company. They have a nasty habit of retiring driver support long before the cards die of old age, which is one of several reasons that I usually push people not to use the nVidia drivers.
As Section_8 suggests, you should definitely do some basic hardware checks. |
|
| Back to top |
|
|
Roman_Gruber
Advocate
Joined: 03 Oct 2006
Posts: 3846
Location: Austro Bavaria
|
| Posted: Fri Oct 21, 2016 8:15 am Post subject: |
|
|
The support has ended ~4 years ago or so, the reason why I sold my notebook with 9800m gts card.
Afaik patches + some kernel version + binary old driver worked.
My only advise, stay on the oldest available kernel in portage + the patches from this forum whne its needed to get the nvidia-drivers running.
2nd get a second hand gpu and replace it. in my area sometimes guys sell old gpus very cheaply or give them away also
or use the open source drivers... |
|
| Back to top |
|
|
Aiken
Apprentice
Joined: 22 Jan 2003
Posts: 240
Location: Toowoomba/Australia
|
| Posted: Fri Oct 21, 2016 9:02 am Post subject: |
|
|
| Code: |
NVRM: os_schedule: Attempted to yield the CPU while in atomic or interrupt context
|
I was getting a lot of this message with a pair of machines that each have 9600GT in them. Both had failing fans and were overheating. If it is something as simple as that another fan to blow air on the video cards heat sink may do the trick.
_________________
Beware the grue. |
|
| Back to top |
|
|
snkmoorthy
Guru
Joined: 19 Nov 2002
Posts: 376
|
|
| Back to top |
|
|
i4dnf
Apprentice
Joined: 18 Sep 2005
Posts: 271
Location: Bucharest, Romania
|
| Posted: Sat Oct 22, 2016 8:43 am Post subject: |
|
|
| Hu wrote: | | They have a nasty habit of retiring driver support long before the cards die of old age, which is one of several reasons that I usually push people not to use the nVidia drivers. |
With all due respect, if you care to check here (devtalk.nvidia.com) you'll see that both 304.xx and 340.xx legacy series of drivers have received updates as recent as 26.09.2016 which would hardly qualify as "retired driver support". That makes 10yo cards actively supported, cards that aren't even supported by other sw nowadays and you have to fallback on software rendering (which can hardly be said about other manufacturers, which dropped support even for 2yo cards at some point).
They are not supported by the "mainline" drivers because they lack the features the new drivers rely/improve on.
_________________
"The only difference between me and a madman is that I am not MAD" (SALVATOR DALI) |
|
| Back to top |
|
|
Hu
Administrator
Joined: 06 Mar 2007
Posts: 23360
|
| Posted: Sat Oct 22, 2016 5:55 pm Post subject: |
|
|
| That site presently returns 504 Gateway Time-out after a very long delay. Your claim would be nice if true, but is at odds with every other report I have ever read about the nVidia proprietary drivers and their support for legacy cards. |
|
| Back to top |
|
|
jburns
Veteran
Joined: 18 Jan 2007
Posts: 1227
Location: Massachusetts USA
|
|
| Back to top |
|
|
asturm
Developer
Joined: 05 Apr 2007
Posts: 9378
|
| Posted: Sun May 14, 2017 8:39 am Post subject: |
|
|
| i4dnf wrote: | | That makes 10yo cards actively supported, cards that aren't even supported by other sw nowadays and you have to fallback on software rendering (which can hardly be said about other manufacturers, which dropped support even for 2yo cards at some point). |
Two other well-known GPU chip manufacturers officially support open source drivers with support going back at least 10 years. You don't even have to worry about the latest kernel or xorg-server version, because it will just work. |
|
| Back to top |
|
|
|
Desktop Environments
