Re: when will nvidai-glx and nvidia-kernel-source hit etch? {Scanned} {Scanned} {Scanned}
On Tue, 2006-10-24 at 12:19 -0500, Henry Hollenberg wrote:
> Henry Hollenberg wrote:
> > I wrote:
> >
> >> OK, I added some unstable lines to my sources.list and ran apt-get
> >> update:
> >>
> >> deb http://cudlug.cudenver.edu/debian/ testing main contrib non-free
> >> deb-src http://cudlug.cudenver.edu/debian/ testing main contrib non-free
> >>
> >> deb http://cudlug.cudenver.edu/debian/ unstable main contrib non-free
> >> deb-src http://cudlug.cudenver.edu/debian/ unstable main contrib non-free
> >>
> >> then:
> >> apt-get -f install
> >> which installed:
> >> ii linux-image-2.6.18-1-amd64 2.6.18-2
> >> Linux 2.6.18 image on AMD64
> >>
> >> apt-get -t unstable install nvidia-kernel-common
> >> apt-get -t unstable install nvidia-kernel-source
> >> apt-get -t unstable install module-assistant
> >> m-a a-i nvidia
> >> apt-get -t unstable install nvidia-glx
> >>
> >> And somehow the 2.6.18 kernel headers got installed not sure when or how:
> >> ii linux-kernel-headers 2.6.18-3
> >> Linux Kernel Headers for development
> >>
> >>
> >> I checked /etc/X11/XF86Config-4 and it looked ok according to Andrew
> >> Schulman's instructions.
> >> I also added my user to group video in case that was needed also
> >> according to Schulman.
> >>
> >> I am still gettting a gdm startup error:
> >>
> >> Failed to start the X server....
> >>
> >> GDM: Xserver not found: /usr/X11R6/bin/X :0....
> >> Error: Command could not be executed!
> >> Please install the Xserver or correct the GDM configuration and
> >> restart GDM
> >
> >
> > Found something screwy:
> >
> > andy:/usr/X11R6# ls -l
> > total 28
> > lrwxrwxrwx 1 root root 6 2006-10-16 08:24 bin -> ../bin
> > drwxr-xr-x 3 root root 4096 2006-10-16 08:33 include
> > drwxr-xr-x 4 root root 4096 2006-10-16 08:33 lib
> > drwxr-xr-x 3 root root 4096 2005-12-23 14:54 lib64
> >
> > So /usr/X11R6/bin is rediricted to /usr/bin
> >
> > Which package is setting up this symlink?
> >
> > I think it is a very bad thing as that is how I ended up deleting my
> > whole /usr/bin
> > contents when this etch upgrade started....all to get my hands on
> > openoffice.
> >
> > So where is /usr/X11R6/bin's files?
> >
> > Look to be intermixed with /usr/bin's files now.....that sure seems screwy.
> >
> > I sure don't recall creating that symlink:
> > andy:/var/log# history | grep ln
> > 510 history | grep ln
>
> So gdm is looking for /usr/X11R6/bin/X but my xserver now is:
> /usr/X11R6/bin/Xorg
>
> when I try and run this binary by hand I see these errors:
>
> (==) Using config file: "/etc/X11/xorg.conf"
> Error: API mismatch: the NVIDIA kernel module has the version 1.0-8774, but
> this X module has the version 1.0-8776. Please make sure that the kernel
> module and all NVIDIA driver components have the same version.
> (EE) NVIDIA(0): Failed to initialize the NVIDIA kernel module! Please ensure
> (EE) NVIDIA(0): that there is a supported NVIDIA GPU in this system, and
> (EE) NVIDIA(0): that the NVIDIA device files have been created properly.
> (EE) NVIDIA(0): Please consult the NVIDIA README for details.
> (EE) NVIDIA(0): *** Aborting ***
> (EE) Screen(s) found, but none have a usable configuration.
>
> Ok, I have a Leadtek PX6800GT video card installed in this system so I don't think the
> supported NVIDIA GPU line is the issue must be the API mismatch....not sure how
> that would have happened. Think I'll cc Randall Donald the nvidia maintainer....
>
> andy:/usr/X11R6/bin# lspci
> 0000:01:00.0 VGA compatible controller: nVidia Corporation NV40 [GeForce 6800 Ultra/GeForce 6800 GT] (rev a2)
>
> I'll give kdm a whirl and see if it does better than gdm...
/usr/X11R6/bin was dumped and symlinked with the XOrg transition about
18 months ago. Nothing should be using /usr/X11R6, period. Especially on
an Etch system.
The error message you're getting is right in front of your face:
"the NVIDIA kernel module has the version 1.0-8774, but this X module
has the version 1.0-8776" means your nvidia kernel module has version
1.0-8774 but your nvidia-glx has version 8776. Don't try and mix and
match versions.
And for reference, you can't "just" add lines to unstable. Apt will
automatically use the most recent mirror it sees for *ALL* packages,
unless you prepare accordingly. "-t unstable" is redundant unless you
prepare apt for another default. As such, large chunks of your system
are likely to have been moved to unstable.
Let's try this again.
Installing Nvidia:
1) Nvidia is made from two main components - a binary driver
("nvidia-glx") and a kernel module (package name varies a lot). The
kernel module must be built against your EXACT, RUNNING KERNEL; the two
components must MATCH PRECISELY.
2) To install the kernel module, you must compile it. This requires
kernel headers for your running kernel ("linux-headers-$(uname -r)") and
the module source code ("nvidia-kernel-source").
3) To compile it, you can mess around by hand, or use the
"module-assistant" tool. "m-a a-i nvidia" will retrive a C compiler (if
needed), kernel headers for your running kernel (if needed) and the
nvidia-kernel-source package (if needed), AS LONG AS THEY ARE AVAILABLE
FROM YOUR SOURCES.LIST. If not, then you'll need to install the above by
hand.
4) When module-assistant has built and installed your kernel module,
install the "nvidia-kernel-common" package (useful Init scripts to make
nvidia behave) and "nvidia-glx" package (the actual driver). At this
point, you can reconfigure your X server to use "nvidia" rather than
"nv".
5) If given packages are unavailable on your release (e.g. Etch), then
you *might* be able to get away with downloading the .deb files from
Unstable and installing them with "dpkg -i". However, adding unstable to
sources.list is a sure-fire way to move to unstable, unless significant
time, care & attention is taken.
Reply to: