DFP no longer works in 4349 & RH 9

[olddoc]

I had a working rh7.3 system with Nvidia 3123 drivers. I have an LCD that I only had a DVI connection to. It used to be fine -- start and see VGA, then startx and get 3D via the DVI input.
I upgraded to RH 9 and ran the 4349 install.
Now I can only get 3D if I connect an analog cable and not try to use DVI.
If I start with just the DVI cable now, I get the VGA boot up into run level 3 fine but a complete lock up when I startx -- it needs a hard reset.

I am using: Ge Force 3 ti 200 with Bios 03.20.00.18.19
Tyan S2426ung Thunder K7 Bios 2.12
RH 9 with 2.4.20-9smp (athlon) from RH update
BENQ FP2081 monitor
NVIDIA 4349 from the default .run script

Here is /etc/X11/XF86Config:
This works in analog but it looks SOOOO much better with a DVI Digital connector.
Why did it stop working?
If I tell it to use connected monitor DFP it locks!

# XFree86 4 configuration created by redhat-config-xfree86

Section "ServerLayout"
Identifier "XFree86 Configured"
Screen 0 "Screen0" 0 0
InputDevice "Mouse0" "CorePointer"
InputDevice "Keyboard0" "CoreKeyboard"
EndSection

Section "Files"
FontPath "unix/:7100"
EndSection

Section "Module"

# Load "record" # X event recorder
# You only need the following two modules if you do not use xfs.
# Load "freetype" # TrueType font handler
# Load "type1" # Adobe Type 1 font handler
Load "dbe" # Double-buffering
#Load "GLcore" # OpenGL support
#Load "dri" # Direct rendering infrastructure
Load "glx" # OpenGL X protocol interface
Load "extmod" # Misc. required extensions
Load "v4l" # Video4Linux
EndSection

Section "InputDevice"
Identifier "Keyboard0"
Driver "keyboard"
Option "XkbLayout" "us"
EndSection

Section "InputDevice"
Identifier "Mouse0"
Driver "mouse"
Option "Device" "/dev/mouse"
Option "Protocol" "IMPS/2"
Option "Emulate3Buttons" "off"
Option "ZAxisMapping" "4 5"
EndSection

Section "Monitor"
Identifier "Benq FP2081"
VendorName "Unknown"
ModelName "BENQ FP2081"
HorizSync 31.0 - 80.0
VertRefresh 60
Option "dpms"
#Option "ConnectedMonitor" "DFP"
Option "NoBandWidthTest" "1"
Option "NvAGP" "1"
Option "IgnoreDisplayDevices" "TV"
EndSection

Section "Device"
Identifier "nVidia Corporation|NV20 [GeForce3 Ti200]"
Driver "nvidia"
BoardName "NVIDIA GeForce 3 (generic)"
EndSection

Section "Device"
Identifier "Linux Frame Buffer"
Driver "fbdev"
BoardName "Unknown"
EndSection

Section "Screen"
Identifier "Screen0"
Device "nVidia Corporation|NV20 [GeForce3 Ti200]"
Monitor "Benq FP2081"
DefaultDepth 24
SubSection "Display"
Depth 24
Modes "1600x1200" "800x600"
EndSubSection
EndSection
___________________________

I also attached --verbose 5 log from when it starts ok using analog... not DFP

[bwkaz]

You probably want to see the driver errata (http://download.nvidia.com/XFree86/Linux-x86/1.0-4349/errata-1.0-4349.txt), which was available from the driver download page (it's linked off the bottom).

The solution is to either use a non-SMP kernel (do you actually have two processors? If not, SMP is doing you no good whatsoever anyway...), or use analog like you have been. Or go back to an earlier driver and wait for this one to figure itself out (probably in the next release; unfortunately no one at nVidia is saying when that'll be -- they probably just don't know or something). Oh, hang on -- you said you upgraded to RH 9. Well, you can't back off driver versions then -- 4349 is the only one that works with the new kernel and glibc that RH pulled into version 9.

[olddoc]

From the errata:

System hangs occur with smp systems driving digitial flatpanels.

The only solution, currently, is to revert to either the 1.0-3123
or 1.0-4191 drivers.
______________________________

OK, fooey! Yes, I actually have 2 processors.
There are a lot of inexpensive Athlon MP boards out now and the new Pentium 4 with Hyper Threading will need smp kernels.
NVIDIA will need to sort this one out soon!
Send me a beta, I'd be happy to help test it.

I have a Gentoo box on an Nforce2 mobo that emerged and ran 4349 perfectly well on a 2.4.20 kernel (no smp in the kernel) and also drives a DVI connected LCD.

Thanks for pointing the errata out.