Nvidia 0.95 on RH 7.0, agpgart problem

Nvidia 0.95 on RH 7.0, agpgart problem

Post by Anton Degue » Sat, 11 Nov 2000 04:00:00

> Hello Anton,

> When searching on the web for info with respect to the problems
> mentioned in the subject header I came across your posting and the
> replies.

> I also have a Dell Dimension 4100 PIII with the Geforce 2 GTS and
> use RH 7.0. So I came across the same problems.

> First you write:
> > Xfree 4.0.1 is installed by default and the card is identified as an
> > "NVIDIA GeForce 2 (generic)".  It works fine without accelerated 3D.

> Do you mean using the standard `nv' driver included in RH 7.0?
> I haven't tried it because I assumed it would not work or would
> not work well, maybe giving hardware problems.
> If `nv' works I can live without 3D.

I had no problem with the 'nv' driver that came with RH 7.0, for 2D
only.  I have read somewhere (I can't remember where) that the XFree
used by RedHat is a cvs snapshot of XFree and therefore has a few
features that are not in the default Xfree 4.0.1, including a driver for
GeForce 2 cards.  Once again, I have 3 workstations here with the
default 'nv' drivers and it seems to be stable.

Quote:> Second you write:
> > about the messages you get at loading with respect to agpgart.

> I assume you mean when using the nVidia make file to compile
> the stuff in the NVIDIA_Kernel source package which uses `insmod'
> itself at the end. I get exactly the same messages.

> Have you gotten it to work yet and what did you do eventually.
> The replies I saw where not especially helpfull. I don't feel like
> recompiling the kernel without AGP just to get the Geforce 2 running.

I had problems with Nvidia rpms.  I have read many times that the rpms
were "buggy" and I thought it could not be true since it is easy to
write some good rpms.  The truth is that YOU MUST AVOID RPMS
INSTALLATION, at least with RedHat 7.0.  Just download the tgz files,
follow the instructions in the Nvidia FAQ and be sure that you don't
have a single old file remaining (like libGLcore, libglx, etc ...).  I
use "find" to be very systematic.  I did that on two computers:

- install kernel sources, do make xconfig, just save, don't recompile
the kernel.
- install from tar files, check this link:

The first computer is a DELL PIII 866, chipset i815, with the GeForce 2
and I have a user working with OpenGL + TclTk
applications/compilations.  It is used about 8 hours a day, intensive
OpenGL, up for 2 weeks (since installation). It is stable.

The second computer is a DELL PIII 600, chipset i820, with a TNT2 M64.
Same kind of use, very stable but much slower, as expected.

Finally, we found that Nvidia's OpenGL is not very robust (at least less
that SGI, DEC, Mesa) and tends to crash easily if the OpenGL programm is
poorly written.  If you have a crash, it is sometime a good occasion to
debug your own code :-).


-- Computer Integrated Surgery Systems Technology --
------------- Johns Hopkins University -------------

-- Office NEB 26 ------------------- 410 516 5261 --


1. Nvidia 0.95 on Redhat 7.0, agpgart problem.


I am trying to install nvidia drivers on a Dell Dimension 4100 PIII with
a GeForce2 card and RedHat 7.0.

Xfree 4.0.1 is installed by default and the card is identified as an
"NVIDIA GeForce 2 (generic)".  It works fine without accelerated 3D.

I downloaded the NVIDIA_kernel source rpm and NVIDIA_GLX both 0.95.
After installing kernel sources (2.2.16-22) I have been able to compile
the NVIDIA kernel rpm itself.  I then tried to install it.

I have a major problem with agpgart.  I added in /etc/modules.conf the
alias char-major-10-175 agpgart

When I try to manually load this module, I get:
Using /lib/modules/2.2.16-22/misc/agpgart.o
agpgart: agpgart: Detected an Intel i815, but could not find the
secondary device.
/lib/modules/2.2.16-22/misc/agpgart.o: init_module: Device or resource
Hint: insmod errors can be caused by incorrect module parameters,
including invalid IO or IRQ parameters

After that, loading NVdriver is impossible, I logically get agp symbols
I checked with cat /proc/pci and the IRQ used by the card (11) does not
seem to be shared.

Anyone has any clue?


-- Computer Integrated Surgery Systems Technology --
-- Anton Deguet ------------- anton<at>cs.jhu.edu --
-- Office NEB 26 ------------------- 410 516 5261 --

2. dd and floppies for bit copies, device busy

3. nvidia 0.95 + mdk 7.1 problems

4. Q: HSSI/X.21 card for PC w/ Linux

5. nVidia 0.95 under 2.4


7. Nvidia 0.95 + Xfree86 v4.0.1

8. (reposting) how to get DMA'able memory within 4GB on 64-bit machi ne

9. How do I compile nvidia 0.95 on an SMP system under Suse 7?

10. xf86 4.0, NvidiaGLX 0.95 - problems with monitor synchronisation

11. NVDriver, Nvidia M64, agpgart and MPV3 chipset problems

12. nVidia Geforce 2 MX 400 and RH 7.0

13. nVidia GeForce2 and RH 7.0