Bug Report: Hard crash (freeze) with nVidia GTX 460

[elvanor]

Hello, I am experiencing hard freezes (eg, computer totally hung up, impossible to kill X or do whatever, have to reboot) with an ASUS nVidia GTX 460.

I am using kernel-2.6.37-rc5 (vanilla) on Gentoo, nVidia drivers 206.19.29. The problem also appeared using any combination of kernel and drivers since I installed the card (I tried 2.6.33, 2.6.35, 2.6.36 at least, with older nVidia drivers as well).

This happens about once a day (maybe once every two days) and seems to be totally random. Eg, sometimes I just move my mouse on the screen, and bam, it hangs; sometimes I leave my computer running during the night (doing nothing), I return next morning only to find it has frozen.

I am sure the culprit is the nVidia binary drivers; if I remove them and install Nouveau drivers, no more crashes (I did that and it ran for 3 weeks without problems - I reinstalled the binary drivers 4 days ago and already suffered from 3 hard freezes).

I attach the log requested but it seems to me it does not contain any useful info - maybe the script had a problem?

How can I further help nVidia troubleshoot and fix this *extremely painful* problem? It prevents from using my computer in normal conditions...

[Licaon]

Originally Posted by elvanor

I attach the log requested but it seems to me it does not contain any useful info - maybe the script had a problem?

To make sure this log file includes as much relevant information as possible, please start the X server with `startx -- -logverbose 6` and run `nvidia-bug-report.sh` after the problem has occurred. from If you have a problem, PLEASE read this first

[elvanor]

Here is the new log (made with -logverbose 6). I realized the first one was OK in fact, I just did not decompress it with the right tool.

Also, impossible to launch the script after the problem occurs, as the problem is a hard freeze ...

[artem]

1) Test your GPU using Video Memory Stress test (alas you need Windows for that)
2) Run memtest86+ for at least three hours.

[elvanor]

Hello, I did that last night, no problems reported.

I really doubt it is a hardware problem because:

1) I have no problems at all under Linux with Nouveau drivers;
2) I have no problems at all under Windows 7 with nVidia drivers.

So, Linux nVidia driver is probably the culprit.

[artem]

elvanor,

can you try some less exotic distros, like Fedora, Ubuntu, Suse, Arch?

No one knows which GCC compilation flags you used to build the world, only God knows what kernel options you activated so, given so many volatile variables, no one can guarantee NVIDIA module will work reliably for you.

(and get a life, please, Gentoo doesn't have all the advantages you imagine - of course, it's highly customizable, but you pay quite a steep price by facing the bugs no one else is experiencing)

[elvanor]

Hi, I could try another distro. However, I am a fairly experienced Gentoo user (actually a Gentoo developper) so I built my system sanely (sane C flags, kernel options, etc).

The problem is not with Gentoo, it's with closed source software installed on Gentoo ^^

[artem]

If you were a high profile customer who owned hundreds of NVIDIA based systems and they all had the mentioned bug, then your problem could have been solved or at least payed attention to.

Right now you can expect next to zero help - I haven't seen NVIDIA Linux developers in these forums for the past two weeks.

[Deanjo]

Originally Posted by artem

If you were a high profile customer who owned hundreds of NVIDIA based systems and they all had the mentioned bug, then your problem could have been solved or at least payed attention to.

Things always have typically slowed during the holiday season. No need for alarm.

Right now you can expect next to zero help - I haven't seen NVIDIA Linux developers in these forums for the past two weeks.

Not for two weeks eh?

http://www.nvnews.net/vbulletin/showthread.php?t=158141

http://www.nvnews.net/vbulletin/showthread.php?t=158147

http://www.nvnews.net/vbulletin/showthread.php?t=157952

[artem]

I meant their activity dwindled almost to nothing Probably you're right, we are due to the holidays season.

[gio___]

I've read something similar here:
http://www.nvnews.net/vbulletin/showthread.php?t=155342

I think the freeze are caused by the powemizer switch among frequencies... Have you tried to change the powemizer policy from "adaptive" to "maximum performance"?

Gio

[elvanor]

I am not experiencing X lockups but complete system freezes (or could this be the same? how can I tell?)