Of course. I've followed the ATI install directions exactly. I've run and rerun the install a dozen times. I've rebooted after every attempt.

Anything in the log files, /var/log/XF*

I get some interesting info when I run dmesg, but I don't know what it means:

Linux agpgart interface v0.99 (c) Jeff Hartmann
agpgart: Maximum main memory to use for agp memory: 439M
agpgart: Detected NVIDIA nForce2 chipset
agpgart: AGP aperture is 64M @ 0xe0000000
[fglrx] Maximum main memory to use for locked dma buffers: 428 MBytes.
[fglrx] module loaded - fglrx 3.11.1 [Aug 4 2004] on minor 0
Fire GL built-in AGP-support
Based on agpgart interface v0.99 (c) Jeff Hartmann
agpgart: Maximum main memory to use for agp memory: 439M
agpgart: Detected nVidia nForce2 chipset
agpgart: Bridge device is in AGP v1/v2 mode
agpgart: AGP aperture is 64M @ 0xe0000000
Power management callback for AGP chipset installed
[fglrx] AGP detected, AgpState = 0x00000000 (hardware caps of chipset)
[fglrx:firegl_unlock] *ERROR* Process 6341 using kernel context 0

what is runnning on PID 6341?

ps -aux |grep 6341

I think you should probably not run an external agpgart. There should be on build into the driver.

Well, the PID changes every time, but I rebooted and then immediately checked dmesg and then ran the ps command like you said. I came up with this:

/usr/X11R6/bin/X:0 -auth /var/gdm/:0.Xauth vt7

But where is the external AGPGart being loaded from, then (if that's the problem)? The kernel?

jonmills' XF86Config-4 clearly says
.

But checking out The ATI Radeon How-to suggests that in fact this setting (UseInternalAGPGART) should be set to "no" for an nForce mobo, and the kernel AGPGART should be used.

Sounds like it's worth a try, anyway. Can't really get much worse than what it is now .

Yes, I did read the ATI HowTo, and what it says is that on an nForce mobo, it's best to use the external agpgart, but that it's only supported on kernel 2.4.22 and later. Red Hat Enterprise Linux uses 2.4.21. So, I can't use the external agpgart. It also says that the internal agp is supported on nForce mobos in ATI driver versions 3.7.6 and later, and I'm using 3.11. Ergo, I must use the internal agp.

Driver 3.11 is later than version 3.7.6.

11 comes after 7.

Trust me, I've been using ATI binary drivers since version 3.2.8, and I do know their order

3.2.8=>3.7.0=>3.7.6=>3.11.1

The 3.11 are the current version, the 3.7.6 are the previous version directly before.

I would try the external apgart anyway, since the kernel difference is 0.0.01 revisions.

You really can't be much worse off than you are now in terms of the drivers working.

Oh for pete's sake, I'm not a freakin newbie. YES, I know 3.11 is after 3.7.6. That's what I said! Did you read my post? I'm using internal agp because it's the only option I have. I can't use external agp because the 2.4.21 kernel doesn't support it on the nForce mobo. As in, it will crash my system. I've tried it.

You're right, I misread your post.

My apologies.

I take it you don't have the opportunity to upgrade your kernel a couple of tenths of a revision?

It seems strange, but it looks like the program you have returned with the ps command is the display manager. I seem to remember a version of RedHat or maybe it was Fedora that I had a problem starting a second X server on with my system.

Can you try this..

Start up the system in runlevel 3, or maybe telinit into runlevel three. Then run startx, or `xinit -e xterm` would be better. Then see how things work. Maybe the display manager is part of the problem. It may be that the problem will only allow you to run X that way.