[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: Using 4GB with amd64 - problems with NVidia graphics kerneldriver

> Does it work fine with the nv driver?
No, the 6600GT is not downward compatible. This was the first thing
I checked, some weeks ago.

In the meantime I have tried the nvidia installer for the latest
version, 7667, but got the following error message when loading the
NVRM: loading NVIDIA Linux x86_64 NVIDIA Kernel Module  1.0-7667 Fri Jun 17 07:
+14:03 PDT 2005
NVRM: RM/client version mismatch!!
NVRM:    aborting to avoid catastrophe!

For now it seems the only way to make it work is the debian way
via nvidia-kernel-common, nvidia-kernel-source, recompiling the
module, make-kpkg, and dpkg -i ... but you made me look closer at
the NVidia homepage, more specifically the forum, where I found

>The attached patch addresses an X startup problem on Linux/x86-64
>em64t systems with 4GB of RAM (or more, i.e. SWIOTLB
>using SuSE Linux 9.3 (e.g. Linux 2.6.5-7.147) and Linux 2.6.8
>kernels. To patch 1.0-71714 and install the updated driver, please
>follow the instructions below.

Unfortunately, even with the patched 7174 driver I get exactly the
same oops. I've tried to patch the driver towards 7667 and get an
even weirder error message than with the installer:
nvidia: Unknown symbol rm_is_legacy_device

This is definitely a bug, I'll have to notify Nvidia and
hope for the best.

I've noticed that sk98lin (I've patched my kernel with the non-GPL
version) also has a problem with 4GB (i.e. memory hole switched on):

sk98lin: Network Device Driver v8.16.2.3
(C)Copyright 1999-2005 Marvell(R).
ACPI: PCI interrupt 0000:03:00.0[A] -> GSI 16 (level, low) -> IRQ 16
PCI: Setting latency timer of device 0000:03:00.0 to 64
PCI: Unable to reserve mem region #1:ffffffffbe700000@140000000 for
device 0000:03:00.0
SKGE: PCI setup failed: 2

All other things seem to work (at first glance), even my
self-compiled Xvnc runs ok. Why is it always only the non-GPL
drivers that have problems? :->>

Dr.techn. Alexander K. Seewald

Austrian Research Institute     alexsee@oefai.at
for Artificial Intelligence         www.oefai.at
         Information wants to be free;
Information also wants to be expensive (S.Brant)
--------------- alex.seewald.at ----------------

Reply to: