NVRM: Xid Errors

[catpantsnv]

Hi, I know that this particular issue has been posted many times before and left unresolved because the NV folks haven't been able to reproduce it. I too am experiencing this problem and believe I have some additional info.

First, let me describe the problem. It does not happen for me as often as it does for many of the other people who have experienced it; maybe once every couple of hours or so, and only when I am doing something graphically intensive (playing World of Warcraft via Cedega, in my case). The display seems to lock up for a considerable period of time, sometimes a few seconds, sometimes a minute or two. After that, things go back to normal and a tail of dmesg reports something akin to:

NVRM: Xid (0002:00): 13, 0003 00000000 00005097 000015f0 001a0024 00000100
NVRM: Xid (0002:00): 30, L1 -> L0

It is noteworthy that if WoW is running in D3D mode with Cedega when this occurs, I am able to Alt-Fkey onto another virtual desktop, and everything else is still running fine. If I am using Wine with WoW in OpenGL mode, I am not able to do this--it seems as if the Xserver itself has frozen.

As you can see in the attached nvidia-bug-report.log, I am currently using a PCIe 8800GTS. I recently upgraded from a 7600GT which also exhibited this problem. However, I was able to workaround the problem by underclocking the 7600GT slightly. My 7600GT is a PVT73GUGD3 (590MHz GPU, 800MHz memory), and I underclocked it to match the PVT73GUGF3 (570MHz GPU, 725MHz memory), after which I never experienced the problem again with that card. I have not yet attempted to underclock the 8800GTS, but will post the results here when I do.

I hope that this information is useful in tracking down and eliminating this problem, or at least helpful to other users.

[mooninite]

I know this is hard to do, but what are your card temps when it gives you Xid errors? Card temps at idle?

I've only ever seen Xid errors with heat related problems. People will say I'm an idiot, but I'm telling you my first hand experience.

[catpantsnv]

Had another Xid last night, card temp was 64-65 Celsius according to nvidia-settings. I've seen it peak out around 70 without issue, so I'd be disinclined to believe that the problem is temperature-related.

[chunkey]

Nvidia's Linux Driver FAQ

Q: My kernel log contains messages that are prefixed with "Xid"; what do these messages mean?
A: "Xid" messages indicate that a general GPU error occurred, most often due to the driver misprogramming the GPU or to corruption of the commands sent to the GPU. These messages provide diagnostic information that can be used by NVIDIA to aid in debugging reported problems.

So, without any knowledge about the inner life of the silicon chips or driver, it's pretty hard to give a hint or a conclusion what's wrong...

[mooninite]

Heat causes any complex microcircuit to error. Ignoring heat as a cause is ridiculous in the computing world. That's why I firmly believe my heat issues were causing my Xid errors, and eventually killed two of my nVidia cards in the past year.

65+ temps were consistent to when I was seeing Xid's. Think about how hot that is. 65 degrees Celcius is about 150 degrees Fahrenheit. That's hot. The Linux driver is not as tolerant as the Windows driver when it comes to heat in my experience.

Of course, I could be wrong, and I will end my contribution to this thread here, but I am not trying to give you misinformation.

[madman2003]

65 degrees C should be well within limits. In the past i've seen my card run at 100-120 degrees C without any problems (passively cooled 6600gt, so no i don't have a broken fan). My card idles above 65 degrees C, so i would find it strange if it produced errors at this temperature.

[helmotz]

Here using nvidia driver 185.18.36 on fedora 11.
I'm experiencing random lookups with NVRM: Xid errors in dmesg.
Nothing (underclocking, kernel options,...) helped so far.
I'm experiencing this problem since a year.
Can at least nvidia say what these errors mean?

Attaching the related bug report.

[piotrq__]

What's your PSU catpantsnv ?

[Nv5-SMURF]

Maybe its hardware related or not but i got the Xid errors now 5 times in last 3 months with different distros and kernels , never it shows up but now 3 times when scrolling in a PDF in kpdf and adobe-reader even twice (Weird never had it) a reboot , other 2 times X died but got back to login again and try it again next session no probs
And worse part is cant reproduce it at will , reading-pdf in same time gaming ,watch movie on 2e screen but all is OK
( in Console-mode prolly never will happen even with X started but still checking)
and my 260-216 wont reach 40 degrees C in 2D - 63 in 3D, PSU@600 , Mem checked OK

But still we wait for a Xid to happen