Bug 114921

Summary:	NVIDIA 7676 Driver == dead system
Product:	[openSUSE] SUSE LINUX 10.0	Reporter:	Joachim Reichelt <Joachim.Reichelt>
Component:	X11 3rd Party	Assignee:	Stefan Dirsch <sndirsch>
Status:	RESOLVED WORKSFORME	QA Contact:	Stefan Dirsch <sndirsch>
Severity:	Critical
Priority:	P5 - None	CC:	aritger
Version:	Beta 4
Target Milestone:	---
Hardware:	Other
OS:	All
Whiteboard:
Found By:	Other	Services Priority:
Business Priority:		Blocker:	---
Marketing QA Status:	---	IT Deployment:	---
Bug Depends on:	114895
Bug Blocks:
Attachments:	tar of nvidia-bug-report.sh results for 7167 und 7676 driver

Description Joachim Reichelt 2005-09-02 09:12:27 UTC

I installed the NVIDIA driver for my MX460 card.
startx will make my system nearly unaccessible: No keyboard (but: NUMLOCK switch
works), NO Display at all, but one static underscore in the top left.
All I can do is push the reset button for a shutdown.
(I do not have a second system to test remote access)

Comment 1 Stefan Dirsch 2005-09-02 09:27:08 UTC

Please start in runlevel 3 and attach results of nvidia-bug-report.sh.

Comment 2 Susan Calvin 2005-09-02 11:52:14 UTC

Same thing happens on my system. The computer is unresponsive, but after a
looong time, X comes up and things go on as normal. See bug 114895 (I have put
in a dependency below). I also see that static top left underscore you are
reporting.

I do not have problems like these with the same NVIDIA binary drivers on another
system where I compile my own 2.6.13 vanilla kernels.

Comment 3 Joachim Reichelt 2005-09-02 12:08:18 UTC

There seems to bee serious NVIDIA problems on other Linuxes too:

http://www.nvnews.net/vbulletin/showthread.php?s=57b02bae16d57863c6a01b625e6ccffb&t=54863

e.g.

 Default  Re: 1.0-7676 for Linux x86 released
Has anyone who'd been experiencing the 99% XOrg CPU usage freeze-ups tried this
one? Does it fix the issue at all?

(I'm on Debian and would rather wait for it to get packaged -- or do it myself,
assuming this does stop the random freezes... but there's not much point for me
if it's still broken...)

For what it's worth, I'm using a GeForce 2 MX400, and I get the following line
in my kernel log just before every freeze: :-/
Quote:
NVRM: Xid: 6, PE0000 043c 00efefef 00000000 00adadad 00efefef

Hi!
I'm working with slackware 10.1, with a by-hand compiled kernel (2.6.12.3).
I've just tried this new Nvidia module and it doesn't work: I still have crashes
with messages like:
Code:

Aug 9 20:01:48 localhost kernel: NVRM: Xid: 6, PE0000 051c 0099ffff 0000fd38
00999bff 0099ffff


(see this thread)
I'm sure it is related to Nvidia module: when I switch to "nv" no crashes occur.

Any idea to bypass this problem?

Regards
qed is offline   	Reply With Quote

Comment 4 Joachim Reichelt 2005-09-03 20:38:56 UTC

Created attachment 48715 [details]
tar of nvidia-bug-report.sh results for 7167 und 7676 driver

There is a BIG Kernel problem with 7676 on my system.
Creation of file:
tar cvf NV-bug.tar nvidia-* NV-messages-7676.txt
driver 7176 seems to be o.k.
driver 6xxx does not compile at all.

Comment 5 Stefan Dirsch 2005-09-04 07:32:26 UTC

Andy, any ideas? I'll also give it a try on a GeForce 4 MX.

Comment 6 Stefan Dirsch 2005-09-05 14:39:41 UTC

This is sth. I could not reproduce with my GeForce 4 MX. I'm sorry.