Date: Wed, 25 Feb 2004 09:56:42 -0600 (CST)
From: Eric Sandeen <sandeen@sgi.com>
To: Bela Babik <bela@colorfront.com>
cc: Chris Green <greenc@fnal.gov>, <linux-xfs@oss.sgi.com>
Subject: Re: XFS patched RHELv3 kernels now available - please test
In-Reply-To: <403CFB75.8050501@colorfront.com>
Message-ID: <Pine.LNX.4.44.0402250954100.26069-100000@stout.americas.sgi.com>
MIME-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII
Sender: linux-xfs-bounce@oss.sgi.com
Errors-to: linux-xfs-bounce@oss.sgi.com
Precedence: bulk

I would suggest submitting this to the bugs list at
nvidia, along with pointers to where they can get the
kernel & the source.  For obvious reasons we
cannot debug it.

They may be interested in the datapoint, though - I've
seen someone else with an RH9-based sgi-modified kernel
where the nvidia module oopses the box consistently,
although stock RH9 appears ok.  *shrug* - we just
added devfs, kdb, and VFS stuff for xfs....

-eric


On Wed, 25 Feb 2004, Bela Babik wrote:

> Hi,
> 
> thanks for the answer.
> > The module needs to be recompiled on your machine with the kernel source 
> > being properly configured (make dep after initializing the config). It's 
> > possible that when you re-installed the NVidia module, the NVidia package 
> > found a close-match (but incorrect) binary on its FTP site and that's why 
> > you're hanging. You need to give the -n (I think) switch to the .run 
> > script to prevent it downloading a binary module and force it to compile 
> > its own. You may need to play with the options but you'll get there 
> > eventually.
> I have a configured kernel and i have skipped the precompiled serch step 
> in the installation.
> 
> The server starts, and the NVidia logo appears and then hangup.
> The newer drivers are a bit better, because it begin to load
> the desktop and hangup befire it can finish.
> 
> Bela
> 
>