X lockups when using maya on FC4
 

Hiya, we're getting X lockups a few times a day whilst using maya 7.0 on Fedora FC4.

In the messages log we get:

Jun 28 15:03:01 licorice kernel: NVRM: Xid (0005:00): 13, 0000 01013900 00000039 00000328 00000000 00000800
Jun 28 15:03:01 licorice kernel: NVRM: Xid (0005:00): 28, L1 -> L0


We're using 1.0-8762, but the problems occur with 1.0-8128 as well.

Anyone know what the above error means and how to fix it?

This is on a Dell 670 running FC4 2.6.12-1.1456_FC4_dneg2smp

Thanks, Evan.

Re: X lockups when using maya on FC4
 

Please generate and post an nvidia-bug-report.log, along with details on how this problem can be reproduced.

Also, please verify whether the problem persists with the latest FC4 kernel, as the one that you referenced is rather old.

Thanks,
Lonni

Re: X lockups when using maya on FC4
 

The log generated was too large to upload, so here it is on my website.

http://barkboy.com/test/nvidia-bug-report.log

As far as steps to reproduce the problem go, its happening about three times a day just from using the maya interface switching between windows etc.


Thanks, Evan.

Re: X lockups when using maya on FC4
 

Oh, I forgot to mention that the above nvidia-bug-report is from while the system was running the latest Fedora Core 4 kernel update (2.6.17)

Evan.

Re: X lockups when using maya on FC4
 

To confirm, you're stating that just starting maya, without any datasets, and then switching to other windows triggers the X hang?

If so, what other applications are you running when this X hang occurs, and how did you determine that Maya was causing the X hang?

I have a few additional questions:
0) I see that you've disabled acpi at boot. Can you elaborate on why you've done this, and whether keeping it enabled has any impact on the problem?
1) It looks like you're using a Dell 670 system with an older BIOS. Does updating to the latest (A07) BIOS have any impact?

Thanks,
Lonni

Re: X lockups when using maya on FC4
 

Hiya, Cheers for looking into this.

The crashes are usually happening while the artist is switching between dialog boxes inside maya. He does have a scene loaded and by this time will have done a fair bit of interaction with the maya UI. Apple Shake and Science-D-Visions 3DEqualizer are also quite often open at the time, but not always. Unfortunately we've not yet been able to pin it down to any particular steps to make it easily reproduceable.

To answer your additional questions, I disabled ACPI on boot hoping to solve the problem after reading some other posts in this forum, but it has made no difference. I will upgrade the bios and see if that helps.

Thanks again, Evan.

Re: X lockups when using maya on FC4
 

If the BIOS update doesn't help, I'll need you to provide me with a means of reproducing this crash.

Thanks,
Lonni

Re: X lockups when using maya on FC4
 

The system just crashed while running the latest bios (A07). I'll keep working on finding a way to reproduce the problem.

Evan.

Re: X lockups when using maya on FC4
 

Another crash, this time it was while the artist was using scrubbing the timeline in 3DEqualizer. This is actually the fifth crash today, but I thought I would mention it as it wasn't within maya.

Evan.

Re: X lockups when using maya on FC4
 

does running glxgears for a while (1h+) result into a crash too?

Re: X lockups when using maya on FC4
 

@evanjfraser: Does disabling "Memory Remapping" in the BIOS help?

Explanation:
You are running with PAE Memory above 4GB (see "reg03" in /proc/mtrr).
We found 32bit Linux OS (RHEL3, RHEL4 tested) will show some kinds of weird
instabilities which have been resolved by setting this BIOS option.

Re: X lockups when using maya on FC4
 

Hi, I've picked this up from Evan.

We are still getting this problem with the latest drivers. The crashes are at random and in most cases switching to an older (8178) driver has fixed the problem. The bios on the Dell 670s has no Memory Remapping options, so we are unable to disable that. Is there a way to work around this at all?

I have included another bug report from a crashed machine using the latest drivers.

Thanks

-
Aaron