new nvidia problem

[VeeDubb]

I thought I had this fixed, but I guess not. I've even formatted my hard drive and re-installed, not helpfull.

 

I am using the latest nvidia linux driver version 4363. I installed it via the nvidia self installer running in console mode (X not running). After install, I can startx and everything is great, but every time I reboot the system, it locks up trying to start X and I have to re-install the nvidia driver every single time I boot my system. And frankly, it's getting a litle rediculous. I'm using a Soyo k7v Dragon motherboard with the via chipset and an nvidia GeForce4 mx440 AGP w/64MB video ram. 256MB system RAM. AGP is set to 4x and the apature is set to 128. Iv'e also tried a smaller apeture and 2x in all possible combinations, so I think tht is probably irrelivent.

 

Any clue what on earth could cause this kind of error???

 

I was using a different mobo and video card untill a few days ago, and ever since I rebuilt my sys, i've had this problem.

 

Any sugestions appreciated.

[Ixthusdan]

Couple of things.

 

Could you start in text mode by hitting esc (for lilo) and typing

linux3

 

This will generate an error log that you can see immediately.

When you say x locks up, what do you mean?

 

If you have reference to the nvidia driver in XFConfig-4, and the driver does not load, you will not be able to lauch x. I know all this is fundamental, but sometimes starting at the beginning will find better results.

[tyme]

post your XF86Config-4-i have some ideas but i need to see that before i suggest them. i know it's in your other post, but I'm going to guess it's changed since then :)

[VeeDubb]

# File generated by XFdrake.

 

# **********************************************************************

# Refer to the XF86Config man page for details about the format of

# this file.

# **********************************************************************

 

Section "Files"

# Multiple FontPath entries are allowed (they are concatenated together)

# By default, Mandrake 6.0 and later now use a font server independent of

# the X server to render fonts.

FontPath "unix/:-1"

EndSection

 

Section "ServerFlags"

#DontZap # disable <Crtl><Alt><BS> (server abort)

#DontZoom # disable <Crtl><Alt><KP_+>/<KP_-> (resolution switching)

AllowMouseOpenFail # allows the server to start up even if the mouse doesn't work

EndSection

 

Section "Module"

Load "dbe" # Double-Buffering Extension

Load "v4l" # Video for Linux

Load "extmod"

Load "type1"

Load "freetype"

Load "/usr/X11R6/lib/modules/extensions/libglx.so"

#Load "glx"

EndSection

 

Section "InputDevice"

Identifier "Keyboard1"

Driver "Keyboard"

Option "XkbModel" "pc105"

Option "XkbLayout" "us"

Option "XkbOptions" ""

EndSection

 

Section "InputDevice"

Identifier "Mouse1"

Driver "mouse"

Option "Protocol" "IMPS/2"

Option "Device" "/dev/psaux"

Option "ZAxisMapping" "4 5"

EndSection

 

Section "Monitor"

Identifier "monitor1"

VendorName "Gateway"

ModelName "Gateway EV700"

HorizSync 31-69

VertRefresh 50-110

 

# Sony Vaio C1(X,XS,VE,VN)?

# 1024x480 @ 85.6 Hz, 48 kHz hsync

ModeLine "1024x480" 65.00 1024 1032 1176 1344 480 488 494 563 -hsync -vsync

 

# TV fullscreen mode or DVD fullscreen output.

# 768x576 @ 79 Hz, 50 kHz hsync

ModeLine "768x576" 50.00 768 832 846 1000 576 590 595 630

 

# 768x576 @ 100 Hz, 61.6 kHz hsync

ModeLine "768x576" 63.07 768 800 960 1024 576 578 590 616

EndSection

 

Section "Device"

Identifier "device1"

VendorName "nVidia Corporation"

BoardName "NVIDIA GeForce2 DDR (generic)"

 

Driver "nvidia"

#Driver "nv"

Option "DPMS"

EndSection

 

Section "Screen"

Identifier "screen1"

Device "device1"

Monitor "monitor1"

DefaultColorDepth 24

 

Subsection "Display"

Depth 8

Modes "1024x768" "800x600" "640x480"

EndSubsection

 

Subsection "Display"

Depth 15

Modes "1024x768" "800x600" "640x480"

EndSubsection

 

Subsection "Display"

Depth 16

Modes "1024x768" "800x600" "640x480"

EndSubsection

 

Subsection "Display"

Depth 24

Modes "1024x768" "800x600" "640x480"

EndSubsection

EndSection

 

Section "ServerLayout"

Identifier "layout1"

InputDevice "Keyboard1" "CoreKeyboard"

InputDevice "Mouse1" "CorePointer"

Screen "screen1"

EndSection

[VeeDubb]

What I mean by locking up is actualy a couple things and it varies. Some times X just fails and it returns to a linux 3 type prompt. I'm happy whan that happens. Sometimes, X will try to load but the screen will fill with bizare symols and non-sense and then imediately become non-resonsive, to the point that all I can do is hit the reset button. ctrl-alt-dlt doesn't even work. ANd sometimes, It does the same thing, but the screen just goes black and stops.

 

 

 

As for starting linux 3, that's the only way I can boot into X. I start linux 3, login as root, re-instal the driver, logout, loging stephen, startX. When I try and startX before reinstalling the driver, it does just as I described before though I get back to a prompt more often, maybe half the tme. The error log has nothing useful. It either says the kernel module failed to load with no explanation of why, or it says there is an error on the last line that says EndSection and sometimes it tells me that there is no screen configured at all.

 

Also, you will notice that I have two lines refering to glx and one is commented out, and I also have a line for nvidia and one for nv but nv is commented out. This is just becaus ei have been switching back and forth so much trying to get it to work dependably, that I found it easier to just add or remove a "#" than to retype things. I have tried it without the commentd lines and it did not help, so they are not the problem.

[Guest fubar::chi]

have you tried going back to an earlier driver to see what happens?

also it may have something to do with your mobo settings.

Is the nvidia module actually loading at startup? try

modprobe nvidia

the next time you restart before reinstalling the driver. Also i think there should be /dev/nvidia*

[VeeDubb]

All the /dev/nvidia* fields are read-write all users. I tried modprobe nvidia and it givesan error about taining the kernel and fails to load. What on earth dos that mean?

 

Also, I have tried two different versions. The accelerated driver instaled by the mandrake installer (yes, it installed an nvidia driver, NOT an nv driver) and 4363. I will try one more, whatever was right before 4363.

[Ixthusdan]

Make sure you don't have parts of two different nvidia drivers trying to load. That will mess things up!

[Guest fubar::chi]

All the /dev/nvidia* fields are read-write all users.   I tried modprobe nvidia and it givesan error about taining the kernel and fails to load.  What on earth dos that mean?

you should get the warning about a tainted kernel driver but it should still load. something's not right. I also dont think you'd see the /dev/nvidia* files unload some kernel driver was loaded.

Also, I have tried two different versions.  The accelerated driver instaled by the mandrake installer (yes, it installed an nvidia driver, NOT an nv driver) and 4363.  I will try one more, whatever was right before 4363.

[tyme]

If the module fails to load it should give more than just a tainted kernel error message. Don't assume that because you receive the tainted kernel error message that it doesn't load. You see, the reason you get that message is because the NVIDIA drivers are not GPL. Using something not GPL "taints" the kernel-as it's being loaded into the kernel. This is not an actual software problem, more a differentiation of philosophies. It won't harm your system in any noticeable way.

 

If you are getting the message "module failed to load" when you do modprobe nvidia then obviously there are other issues. It's possible that after the original install of the NVIDIA drivers there were issues when you tried to install it again later-this could be due to the NVIDIA installer not removing previous installs correctly.

 

My suggestion? Go into /usr/src/linux/kernel/drivers/video (this could be off, I'm not in front of a linux machine write now, will check in about 2 hours unless someone can verify or correct) and delete anything beginning with nvidia, this can be done with: (# means you have to be root)

~# rm nvidia*

after doing that try re-installing the RPMs, and make sure you don't have X running (I'm sure you know this, of course :) )

 

addition: something I'm not sure of, but also you might wanna check dmesg and see if there are any nvidia errors in there (does Mandrake load nvidia during the bootup? i know i have to do this in gentoo...). i find the best way to check dmesg is to have it redirected to a file:

~# dmesg > dmesgfile

you can replace dmesgfile with any file name you choose, and the file shouldn't exist prior to your doing this. it will create a fille called dmesgfile (or whatever) that will have the dmesg output in it. look for anything mentioning nvidia and see if there are any errors.

[Guest fubar::chi]

deleting the nvidia files should be unnecessary as they are automatically removed everytime you install with nvidia's new script. if you used rpm versions before then it probably deleted those files.

[tyme]

but I've noticed when installing it sometimes says they may not have been removed correctly. unsure what this means, but better safe than sorry.

[Guest fubar::chi]

but I've noticed when installing it sometimes says they may not have been removed correctly.  unsure what this means, but better safe than sorry.

that was fast :D but yeah it does say that doesn't it

[VeeDubb]

Don't know what to say, it seems to be fixed. THe first time I had this problem, I realized that the mandrake nvidia RPM's hadn't been uninstalled correctly, so I removed them and it all worked. Some how, the mandrake nvidia RPM was back, I uninstalled it again and now everything works. Any idea how something like that could happen???

[Steve Scrimpshire]

I had the same problem kinda, VeeDubb. Dunno what to tell you. My *.rpm wasn't back but I tried removing the old *.rpm and use the nvidia installer and then reinstall the old *.rpm when the nvidia installer didn't work, even after telling me it was successful. I finally found out that I had to remove the old rpms and then go into the /lib/modules/2.4.20/video/blah/blah/blah and remove the nvidia.o file manually then run the nvidia installer.