[PLUG] nvidia weirdness

[Derek Loree]

Hi all,

This is one of those "can anyone throw me a clue" weirdnesses.

I've got the 6111 version of the nvidia-kernel-source package from
Debian Unstable and compiled the new Nvidia drivers by executing

make-kpkg modules_image

from within the source directory of a working 2.6.4 kernel, I then
installed the .deb that was created by executing

dpkg -i nv*6.4*6111*

which was successful.  I then executed

/etc/init.d/gdm restart

and this was successful, the X-server went down, and came back up,
everything was fine and dandy.

Until I rebooted.  The Nvidia module failed to load and no X-server was
started.  I then executed

dpkg -i nv*6.4*6111*

which was successful.  I then executed

/etc/init.d/gdm restart

and this was successful as well, the X-server started right up, and ran
along just fine.

Sooo... I reboot, just to see what would happen.  No X-server.  Execute
the above incantation and it fires up just fine.  This is getting
annoying (and long winded).  

I can't think of what would cause this except some sort of race
condition as the kernel starts, so I get the current unstable kernel
source: 2.6.8.  I configured, compiled, installed, rebooted, (no
X-server, but that was expected) made the nvidia .deb, installed it and
rebooted again.  No X-server. Now I'm bumming.  It's not a big deal to
install the package each time it boots, but I usually like to run 4
X-servers and when they don't start on boot it takes a long time to
process the errors.

Anyone got a clue?  Is this a race condition in the kernel start up, or
just a configuration that isn't taking?

TIA,

Derek Loree