NVIDIA on GENTOO - highly annoying

[Hayl]

as most of you know i am a big fan of debian. well --- i thought i'd give gentoo a try yesterday. so i emerged all day and have "some" apps now :) hehe

i am tryingto get my nvidia card to work properly and i am not sure what i am doing wrong.

i have tried to emerge it as well as tried compiling it from the tar files on nvidia's website.

i get the same error in my log each time i try to load it. i am sure it is modprobed. here is the output from lsmod
bash-2.05b# lsmod
Module Size Used by Tainted: P
snd-pcm-oss 39076 0 (autoclean)
snd-mixer-oss 13144 0 (autoclean) [snd-pcm-oss]
snd-emu10k1 71920 0
snd-pcm 64256 0 [snd-pcm-oss snd-emu10k1]
snd-timer 11688 0 [snd-pcm]
snd-rawmidi 14528 0 [snd-emu10k1]
snd-util-mem 1344 0 [snd-emu10k1]
snd-seq-device 4252 0 [snd-emu10k1 snd-rawmidi]
snd-hwdep 4288 0 [snd-emu10k1]
snd-ac97-codec 30724 0 [snd-emu10k1]
snd 29228 0 [snd-pcm-oss snd-mixer-oss snd-emu10k1 snd-pcm snd-timer snd-rawmidi snd-util-mem snd-seq-device snd-hwdep snd-ac97-codec]
soundcore 3524 6 [snd]
nvidia 1468768 0

here is the error in the XFree86.0.log

(**) NVIDIA(0): Depth 24, (--) framebuffer bpp 32
(==) NVIDIA(0): RGB weight 888
(==) NVIDIA(0): Default visual is TrueColor
(==) NVIDIA(0): Using gamma correction (1.0, 1.0, 1.0)
(--) NVIDIA(0): Linear framebuffer at 0xF4000000
(--) NVIDIA(0): MMIO registers at 0xF2000000
(WW) System lacks support for changing MTRRs
(EE) NVIDIA(0): Failed to initialize the NVIDIA kernel module!
(EE) NVIDIA(0): *** Aborting ***
(II) UnloadModule: "nvidia"
(II) UnloadModule: "vgahw"
(II) Unloading /usr/X11R6/lib/modules/libvgahw.a
(EE) Screen(s) found, but none have a usable configuration.

Fatal server error:
no screens found


does this have anything to do with the devfs system? what is MTRRs ??

[DarkJedi9]

Hmm... could you post your XF86Config, please? Also, it looks like you did get the nvidia drivers installed and they just won't work. Is that the case, or are you having problems with the installation as well?

[Hayl]

the install is fine and the module is loaded - X just wont start with the nvidia drivers. it's weird. ive done this many times with red hat and debian and it has always worked fine. i even had it working under gentoo a log time ago with gentoo 1.2. maybe im just not destined to use gentoo although id like to get this workign because everything else is working well (so far) :)

# /etc/X11/XF86Config
Section "Module"

Load "dbe" # Double buffer extension
SubSection "extmod"
Option "omit xfree86-dga" # don't initialise the DGA extension
EndSubSection
Load "type1"
Load "freetype"
# Load "glx"
EndSection

Section "Files"
RgbPath "/usr/X11R6/lib/X11/rgb"
FontPath "/usr/X11R6/lib/X11/fonts/local/"
FontPath "/usr/X11R6/lib/X11/fonts/misc/"
FontPath "/usr/X11R6/lib/X11/fonts/75dpi/:unscaled"
FontPath "/usr/X11R6/lib/X11/fonts/100dpi/:unscaled"
FontPath "/usr/X11R6/lib/X11/fonts/Type1/"
FontPath "/usr/X11R6/lib/X11/fonts/Speedo/"
FontPath "/usr/X11R6/lib/X11/fonts/75dpi/"
FontPath "/usr/X11R6/lib/X11/fonts/100dpi/"
# ModulePath "/usr/X11R6/lib/modules"
EndSection

Section "ServerFlags"
Option "NoTrapSignals"
# Option "DontZap"
# Option "Dont Zoom"
# Option "DisableVidModeExtension"
# Option "AllowNonLocalXvidtune"
# Option "DisableModInDev"
# Option "AllowNonLocalModInDev"
EndSection

Section "InputDevice"
Identifier "Keyboard1"
Driver "Keyboard"
Option "AutoRepeat" "500 30"
Option "XkbRules" "xfree86"
Option "XkbModel" "logiinetnav"
Option "XkbLayout" "us"
EndSection

Section "InputDevice"
Identifier "Mouse1"
Driver "mouse"
Option "Protocol" "IMPS/2"
Option "Device" "/dev/input/mouse0"
Option "ZAxisMapping" "4 5"
EndSection

Section "Monitor"
Identifier "NEC MultiSync FE950+"
HorizSync 30-96
VertRefresh 50-160
EndSection

Section "Device"
Identifier "NVIDIA GeForce"
# Driver "nvidia"
Driver "nv"
EndSection

Section "Screen"
Identifier "Screen 1"
Device "NVIDIA GeForce"
Monitor "NEC MultiSync FE950+"
DefaultDepth 24
Subsection "Display"
Depth 8
Modes "1600x1200" "1152x864" "1280x1024" "1024x768" "800x600" "640x480"
ViewPort 0 0
EndSubsection
Subsection "Display"
Depth 16
Modes "1600x1200" "1152x864" "1280x1024" "1024x768" "800x600" "640x480"
ViewPort 0 0
EndSubsection
Subsection "Display"
Depth 24
Modes "1600x1200" "1152x864" "1280x1024" "1024x768" "800x600" "640x480"
ViewPort 0 0
EndSubsection
EndSection

Section "ServerLayout"
Identifier "Simple Layout"
Screen "Screen 1"
InputDevice "Mouse1" "CorePointer"
InputDevice "Keyboard1" "CoreKeyboard"
EndSection

[Hayl]

bump.

[red_over_blue]

What module are you loading in /etc/modules.autoload?

It should be NVdriver in /etc/module.autoload and nvidia in XF86Config (which is what you have in there).

What happens if you type:


modprobe NVdriver

[dbfruth]

uncomment glx

[red_over_blue]

Try doing this (make sure you shutdown X first, just do it from a real terminal (tty?)) and leave your XF86Config unchanged.


emerge nvidia-kernel
emerge nvidia-glx
modprobe NVdriver
startx

[DarkJedi9]

First of all, uncomment the line reading Load "glx". See if that changes anything.

<Edit> Agh, beat to the punch!

[red_over_blue]

After seeing dbfruth's post, I noticed that you didn't have the nvidia driver in your XF86Config... you will need to comment nv and uncomment nvidia as well.

[dbfruth]

Originally posted by red_over_blue
What module are you loading in /etc/modules.autoload?

It should be NVdriver in /etc/module.autoload and nvidia in XF86Config (which is what you have in there).

What happens if you type:


modprobe NVdriver



Actually with the 4193 build the driver that needs to be loaded by modprobe has changed. it is no longer NVdriver but rather nvidia

[red_over_blue]

Are you showing the 4193 ebuild as being unmasked? The 3123 is the most current unmasked ebuild for the nvidia drivers that my portage tree is showing.

I started off trying to use the 4193, but had problems and decided to just use stable 3123 ebuilds.

If you already have some directories from the 4193 drivers, and you want to emerge the 3123 drivers, make sure you you get rid of everything the 4193 drivers installed.

I had a problem where there was a symbolic link that emergeing the 3123 drivers was not overwriting, so I still had problems because of the 4193 drivers until I did some digging around. I think it had to do with

/usr/lib/libGLcore.so.1

[Hayl]

Originally posted by red_over_blue
Are you showing the 4193 ebuild as being unmasked? The 3123 is the most current unmasked ebuild for the nvidia drivers that my portage tree is showing.

I started off trying to use the 4193, but had problems and decided to just use stable 3123 ebuilds.

If you already have some directories from the 4193 drivers, and you want to emerge the 3123 drivers, make sure you you get rid of everything the 4193 drivers installed.

I had a problem where there was a symbolic link that emergeing the 3123 drivers was not overwriting, so I still had problems because of the 4193 drivers until I did some digging around. I think it had to do with

/usr/lib/libGLcore.so.1

where would i find the 4193 stuff

oh - and for everyone else - i posted the XF86Config that works :) it doesnt work when i uncomment nvidia and glx and recomment nv :)

i am currently recompiling my kernel without the nvidia framebuffer support to see if it makes a difference.

[BFK4ever]

I am a relative noob, but if everything is installed correctly you should just uncomment 'nvidia' and either comment 'nv' or if your using Vi do an Esc then press 'dd' then enter to delete that entire line.

[Hayl]

resolved.

it was the nvidia framebuffer support taht was f**king it up

[red_over_blue]

You can emerge masked ebuilds in a few ways (assuming a masked 4193 ebuild exists)

1. There may be an entry in /usr/portage/profiles/package.mask that masks ebuilds. You can comment out the entry if it is there, then re-emerge the ebuild and it will accept the most current, stable or not.

2. Type the following:

export ACCEPT_KEYWORDS="~x86"
emerge nvidia-glx nvidia-kernel

However, if you do an emerge -u world, I think portage may try to downgrade to stable.

3. Type the full path to the ebuild (I am not at my gentoo machine right now, so the path may not be exactly correct as shown below).

emerge /usr/portage/media-video/nvidia-kernel-4193.ebuild

[dbfruth]

Originally posted by dbfruth
Actually with the 4193 build the driver that needs to be loaded by modprobe has changed. it is no longer NVdriver but rather nvidia

I think that I meant to say 4191

My bad

[budnoob]

not to be rude, but yeh it's 4191, i'm currently solving the same problem he is, but on RH 7.2 lol....all of your help is getting my X going again!

[GaMMa]

Cybr099 how did you get it working? I've been having the same problem, I have to add NVdriver to modules.autoload, and if I don't it spits out errors... It worked fine before without me adding it, but now it doesn't work without it.

[budnoob]

i still get the No Screens error in the exact way he does....

[GaMMa]

Well the way I fixed my problem was opened /etc/modules.autoload and put NVdriver in there. Rebooted the PC and boom X loads... I just want to know how he fixed it.

[budnoob]

lol oh my god, you won't believe what the fix was
it was a god damned number. One number!

alias char-major-196nvidia

the 196 was supposed to be 195...dont' even ask how I figured it out lol took me 3 hours

[GaMMa]

Did some searching on the Gentoo boards, and I found out how to fix the problem. I'm guessing you recently recompiled your kernel or updated it in some form. You probably forgot to run update-modules. I did that and rebooted and I'm running the nvidia drivers without putting a line in modules.autoload.