Fedora Core 3

[lovswr]

Ok, I just loaded the nvdia driver & it told me that the module rivafb was configured in my current (default) kernel. Of course it did not work.

Can someone tell me how to make sure that rivafb does not load, or better yet, remove it completely? /sbin/modprobe nvivia will load the nvidia driver but when I try to startx I just get a "grey screen of death"

TIA.

[Void Main]

Ok, I just loaded the nvdia driver & it told me that the module rivafb was configured in my current (default) kernel. Of course it did not work.

Can someone tell me how to make sure that rivafb does not load, or better yet, remove it completely? /sbin/modprobe nvivia will load the nvidia driver but when I try to startx I just get a "grey screen of death" :D

TIA.

I think the rivafb thing is just a warning that it shouldn't be loaded. I got the same warning but it works fine. Did you follow the instructions in the README about modifying your /etc/X11/xorg.conf file? There are like two changes you have to make to it, one of the changes is change the driver name "nv" to "nvidia". Read the README for anything else that needs to be changed. Now did you read the other posts in this thread about "udev" and nVidia drivers? If you haven't updated since doing the base install then you'll need to copy your /dev/nvidia* device files to /etc/udev/devices and set permissions on them (after installing your nVidia driver but before any reboot):

# cp -a /dev/nvidia* /etc/udev/devices
# chown root:root /etc/udev/devices/nvidia*
# chmod 666 /etc/udev/devices/nvida*

Now, if you *have* updated your your system with the latest errata (including udev) you will also have an entry in your "/etc/udev/permissions.d/50-udev.permissions" file in the "dri devices" section. Change the 0660 on that line to 0666.

[lovswr]

Yes Void I did "try" the first suggeston from the readme. But I don't have that directory!

cp: cannot stat `/dev/nvidia*': No such file or directory
[root@matrix1 dev]#

I also looked at your second suggestion, & it is indeed set to 0660, but I figured that I better leave that be until I work out this directory thing.

p.s. Yes I modified the /etc/x11/xorg.conf file too.

[Void Main]

Yes Void I did "try" the first suggeston from the readme. But I don't have that directory! :(

cp: cannot stat `/dev/nvidia*': No such file or directory
[root@matrix1 dev]#

It sounds like you installed the driver and have rebooted since in which case the /dev/nvidia* special files would not exist, which is why your nVidia driver isn't working. You have to successfully install that driver (go ahead and inistall it again with the *.run file) and before rebooting do the copy commands again. I would also upgrade the udev package before doing anything if you haven't already done so (apt-get update; apt-get dist-upgrade if you have apt set up).

I also looked at your second suggestion, & it is indeed set to 0660, but I figured that I better leave that be until I work out this directory thing.

It won't hurt anything to make this change any time.

[Tux]

If you have rebooted and the /dev/nvidia files are nomlonger there, you can just change a line in rc.sysinit.

Change other="", to other="nvidia"
Bingo!

[Void Main]

That is another way but if you upgrade your initscripts that will be lost. I also don't believe either is necessary if you have updated to the last udev package. The only thing that would be required would be to change the permissions file in /etc/udev/permissions.d. In fact that isn't even required if you are running in runlevel 5 because then X is spawned from a root level process.

[lovswr]

Well I have done everything that everybody has suggested. No luck, just grey screen & complete lock up. I can not get to another terminal. I had to resort to booting with the 1st CD of my FC2 set to get to a point where I could edit /etc/X11/xorg.conf with pico just to get back.

I have been using the command init3 to stop X at the command line, but I know there is a way to change something in a file so that I boot straight to the command prompt. Could somebody help me with that?

Thanks for all the other suggestionss. I'll probably try again after Turkey Day to get this nvidia driver to work.

[Tux]

I have been using the command init3 to stop X at the command line, but I know there is a way to change something in a file so that I boot straight to the command prompt. Could somebody help me with that?

Change initdefault to 3 in /etc/inittab

[Void Main]

After setting your initdefault to 3 as Tux mentioned and after successfully installing the nVidia driver verify that you have the proper /dev/nvidia* files:

Code: Select all

$ ls -l /dev/nvidia*
crw-rw-rw-  1 root root 195,   0 Nov 12 19:53 /dev/nvidia0
crw-rw-rw-  1 root root 195, 255 Nov 12 19:53 /dev/nvidiactl


and verify that the module is loaded:

Code: Select all

$ /sbin/lsmod | grep nvid
nvidia               3473436  12


If it is loaded and everything is installed correctly and the devices are there with appropriate permissions (crw-rw-rw- root root) then if it still doesn't work I would like to see a copy of your /etc/xorg.conf and a copy of your Xorg.0.log. You can upload them to my ftp server:

$ ftp voidmain.is-a-geek.net

log in as "anonymous" with any password and then:

ftp> cd uploads
ftp> mkdir lovswr
ftp> cd lovswr
ftp> lcd /etc/X11
ftp> put xorg.conf
ftp> lcd /var/log
ftp> put Xorg.0.log

I can look over them and see if something looks obviously wrong.

Or even easier you can download a script I just wrote that will check all of the above and upload the above files. I would actually prefer if you do this:

Code: Select all

$ wget http://voidmain.is-a-geek.net/files/scripts/ftp-xorg-to-void.sh
$ sh ftp-xorg-to-void.sh


When the above script is run it will create a /tmp/void.txt that contains the output of all of the commands mentioned above and it will then show you a list of files it found (up to 5 files) that it is going to upload (void.txt, xorg.conf, XF86Config, Xorg.0.log, XFree86.0.log) and then prompt for your username on this site. The only thing the username is for is it will create a directory on my ftp server with whatever name you enter there and upload the files to that directory. You can only successfully run the script once with any one particular username because it will not overwrite any files that have already been uploaded. If you want to run it a second time you will have to add a number or something to your username (lovswr2) so it can create a new directory.

Also, once the files are uploaded to my FTP server you will not be able to see them on the server if you do a directory listing in ftp but they will be there.

[lovswr]

Void. The files are on your ftp server under "lovswr".

[insomnia]

Did you do the "modprobe nvidia" in runlevel 3 ?

[Void Main]

The first thing I notice is that your "nvidia" module is not loaded. Try loading it and starting X:

# /sbin/modprobe nvidia

I also notice your udev is not the latest, you might want to update it:

http://ayo.freshrpms.net/fedora/linux/3 ... 2.i386.rpm

The "nvidia*" line in your /etc/udev/permissions.d/50-udev.permissions file is still 0660. I would set it to 0666 not that it should be effecting anything right now.

Your xorg.conf looks good although I have a GeForce card, I see you have a Riva, don't know if there is anything special for your card (yet). The log looks good which makes me think the kernel module had to be loaded. If it still doesn't work after verifying the kernel module is loaded then I will be stumped. You might make a backup copy of your /etc/X11/xorg.conf and then run "system-config-display" and setting a config and see if that makes it work (it will probably reset your video driver to "nv" though). If none of this works then it would be off to the nvidia forums I would go:

http://www.nvnews.net/vbulletin/forumdi ... forumid=14

EDIT: I have seen indication that there may be problems with the TNT2 in the latest nVidia driver:

http://www.nvnews.net/vbulletin/showthr ... ge=3&pp=15

Although the person with the problem suggests that it does work, just not certain options. Some posts indicate that the new driver works just fine. One thing I am wondering is if maybe if your riva kernel module *is* installed since you actually do have a Riva card. See what the output of this is:

$ /sbin/lsmod | grep riva

If you have a "riva" driver installed and not the "nvidia" then it surely will not work. "rmmod" your riva driver and "modprobe nvidia" and verify it is loaded then try X again. If it works we can fix you up permanently.

EDIT2: The README has a section that addresses a specific problem with the TNT cards that you may be running in to:

Code: Select all

__________________________________________________________________________

(app-h) APPENDIX H: TNT SPECIFIC ISSUES
__________________________________________________________________________

Most issues pertaining to SGRAM/SDRAM TNT cards should be resolved.
There is the rare chance, however, that your video card has the wrong
BIOS installed, and that this driver will continue to fail for you.

If this driver fails for you, do the following:

  o watch your monitor as the system boots. The very first, brief screen
    will identify the type of video memory your card has. This will be
    either SGRAM or SDRAM.

  o edit the file "os-registry.c" from the kernel module sources.  Look
    for the variable "NVreg_VideoMemoryTypeOverride".  Set the value of
    the variable to the type of memory you have (numerically, see the
    line just above it).

  o since we do not normally use this variable, change the "#if 0" that is
    about 10 lines above the variable to "#if 1".

  o rebuild and reinstall the new driver ("make")


[lovswr]

insomnia, yes I did do that. & void, no the module was not loaded until I did /.sbin/modprobe nvidia. I 'm at work right now, but I have some documentation at home that specificall warned against the module not loading. I will respond in greater detail later today.

[Void Main]

insomnia, yes I did do that. & void, no the module was not loaded until I did /.sbin/modprobe nvidia. I 'm at work right now, but I have some documentation at home that specificall warned against the module not loading. I will respond in greater detail later today.

The module not loading automatically is a common problem which we have already addressed. By your quote above it sounds like the module loaded with no errors after you did an "/sbin/modprobe nvidia". If this is true (verify with an "lsmod") then did everything work at that time when you did a startx?

[lovswr]

insomnia, yes I did do that. & void, no the module was not loaded until I did /.sbin/modprobe nvidia. I 'm at work right now, but I have some documentation at home that specificall warned against the module not loading. I will respond in greater detail later today.

The module not loading automatically is a common problem which we have already addressed. By your quote above it sounds like the module loaded with no errors after you did an "/sbin/modprobe nvidia". If this is true (verify with an "lsmod") then did everything work at that time when you did a startx?

Yes void, it did load, but I still get the 'grey screen of death".