Nvidia 177.82 howto for F10

[leigh123linux]

1. Install the release key


su
rpm -Uvh http://download1.rpmfusion.org/free/fedora/development/i386/os/rpmfusion-free-release-10-1.noarch.rpm http://download1.rpmfusion.org/nonfree/fedora/development/i386/os/rpmfusion-nonfree-release-10-1.noarch.rpm



2. Install the driver

su
yum install kmod-nvidia xorg-x11-drv-nvidia-libs xorg-x11-drv-nvidia



Or if you like playing with unreleased Koji kernels

su
yum install akmod-nvidia xorg-x11-drv-nvidia-libs xorg-x11-drv-nvidia



64 bit users will also need to install the 32 bit libs for apps like Wine.

su
yum install xorg-x11-drv-nvidia-libs.i386


3, Reboot.

[Dies]

Nice! Thanks.

nv works just fine now so I wasn't too concerned but was getting a little tired of the lifeless windows...

[Hlingler]

Because of the <insert any number> kernel the nvidia driver needs patching !So, what else is new? :rolleyes:

Really sorry, I just couldn't resist.... :D

V

[cloneu2]

Thanks for the info.. Now I can make f10 work for me... had that crummy kernel-devel error even though it was on the machine... fought it this am and here you have a fix the same night....

[leigh123linux]

Thanks for the info.. Now I can make f10 work for me... had that crummy kernel-devel error even though it was on the machine... fought it this am and here you have a fix the same night....


I haven't managed to get the driver to build on any kernel higher than kernel-2.6.27-0.208.rc1 at the moment .

[vallimar]

Make sure you are re-extracting the sources when trying to compile for a new
kernel. I was running into problems after awhile by not doing that. Since I use
rawhide kernels, and they change constantly, I wrote a little build script to auto-
mate the tediousness. I've attached it below.

Download script to director with patch and driver.
Strip off the .txt extension (optional)
Change to init 3 (script checks this too)
sh nvidia-build.sh --uninstall
sh nvidia-build.sh [any extra args, like -a]
init 5

The script passes -Nn to the nvidia-installer, pass any others on the command line.
You can skip the runlevel check by passing --no-x-check, but the installer will still
do its checks for a running module and such, so this is only useful when you aren't
currenly using the nvidia driver.

Feel free to modify, update, pass it around to friends and pets, whatever.

I'm currently running the nvidia driver on kernel-2.6.27-0.237.rc2.fc10.x86_64.

[leigh123linux]

Make sure you are re-extracting the sources when trying to compile for a new
kernel. I was running into problems after awhile by not doing that. Since I use
rawhide kernels, and they change constantly, I wrote a little build script to auto-
mate the tediousness. I've attached it below.

Download script to director with patch and driver.
Strip off the .txt extension (optional)
Change to init 3 (script checks this too)
sh nvidia-build.sh --uninstall
sh nvidia-build.sh [any extra args, like -a]
init 5

The script passes -Nn to the nvidia-installer, pass any others on the command line.
You can skip the runlevel check by passing --no-x-check, but the installer will still
do its checks for a running module and such, so this is only useful when you aren't
currenly using the nvidia driver.

Feel free to modify, update, pass it around to friends and pets, whatever.

I'm currently running the nvidia driver on kernel-2.6.27-0.237.rc2.fc10.x86_64.


Thank you for posting the patch on the Nv news forum :cool:

http://nvnews.net/vbulletin/showpost.php?p=1731112&postcount=1

[vallimar]

No problem - Thankfuly others had done some nicer stuff already so I decided to use
their work instead of the primitive hack-job I started with on my own, then just added back
the needed bits for the newer kernels.

Let me know if you still have problems with newer kernels on a fresh extraction.
I think one time I had to do a manual module build first.. To do that, just extract
the driver, enter the usr/src/nv dir, and do "make module", provided that works,
you should be able to install okay.

[leigh123linux]

I still get the same error on the newer 2.6.27 kernels !


Using: nvidia-installer ncurses user interface
-> License accepted.
-> Installing NVIDIA driver version 177.13.
-> Not probing for precompiled kernel interfaces.
-> Performing CC sanity check with CC="cc".
-> Performing CC version check with CC="cc".
-> Kernel source path: '/lib/modules/2.6.27-0.237.rc2.fc10.x86_64/source'
-> Kernel output path: '/lib/modules/2.6.27-0.237.rc2.fc10.x86_64/build'
ERROR: If you are using a Linux 2.4 kernel, please make sure
you either have configured kernel sources matching your
kernel or the correct set of kernel headers installed
on your system.

If you are using a Linux 2.6 kernel, please make sure
you have configured kernel sources matching your kernel
installed on your system. If you specified a separate
output directory using either the "KBUILD_OUTPUT" or
the "O" KBUILD parameter, make sure to specify this
directory with the SYSOUT environment variable or with
the equivalent nvidia-installer command line option.

Depending on where and how the kernel sources (or the
kernel headers) were installed, you may need to specify
their location with the SYSSRC environment variable or
the equivalent nvidia-installer command line option.
ERROR: Installation has failed. Please see the file
'/var/log/nvidia-installer.log' for details. You may find suggestions
on fixing installation problems in the README available on the Linux
driver download page at www.nvidia.com.

[vallimar]

Strange, I just removed and re-installed the driver with no problems.
What errors do you get if you do a manual kmod compile from the
extracted sources?

[leigh123linux]

Strange, I just removed and re-installed the driver with no problems.
What errors do you get if you do a manual kmod compile from the
extracted sources?

Here the output !


[root@localhost ~]# cd /home/leigh/NVIDIA-Linux-x86_64-177.13-pkg2/usr/src/nv/
[root@localhost nv]# make module
If you are using a Linux 2.4 kernel, please make sure
you either have configured kernel sources matching your
kernel or the correct set of kernel headers installed
on your system.

If you are using a Linux 2.6 kernel, please make sure
you have configured kernel sources matching your kernel
installed on your system. If you specified a separate
output directory using either the "KBUILD_OUTPUT" or
the "O" KBUILD parameter, make sure to specify this
directory with the SYSOUT environment variable or with
the equivalent nvidia-installer command line option.

Depending on where and how the kernel sources (or the
kernel headers) were installed, you may need to specify
their location with the SYSSRC environment variable or
the equivalent nvidia-installer command line option.

*** Unable to determine the target kernel version. ***

make: *** [select_makefile] Error 1
[root@localhost nv]# rpm -qa |grep kernel
kernel-devel-2.6.27-0.237.rc2.fc10.x86_64
kernel-2.6.27-0.208.rc1.git2.fc10.x86_64
kerneloops-0.11-1.fc10.x86_64
kernel-devel-2.6.27-0.226.rc1.git5.fc10.x86_64
kernel-headers-2.6.27-0.237.rc2.fc10.x86_64
kernel-2.6.27-0.226.rc1.git5.fc10.x86_64
kernel-2.6.27-0.237.rc2.fc10.x86_64
kernel-devel-2.6.27-0.208.rc1.git2.fc10.x86_64
kernel-firmware-2.6.27-0.237.rc2.fc10.noarch
[root@localhost nv]# uname -r
2.6.27-0.237.rc2.fc10.x86_64
[root@localhost nv]#


Please could you upload your nvidia.ko for kernel-2.6.27-0.237 as your also running 64bit :cool: ( I can manually copy it to it's correct location )


Thanks

Leigh

[vallimar]

I'm pretty sure I found the problem, I forgot I had done this before, but I just updated to the 238 kernel.. It seems the packager is stripping out more stuff in the kernel-devel package. You have to enter the /usr/src/kernels/<kernel> for your current kernel and do "make prepare". This will recreate the include/asm symlink and a couple other things and then error out. Don't worry about the error. After you've done this, you should be able to compile the module. Kinda dumb, but there you go. If still not working, let me know and I'll upload the kmod. Check the nvnews thread as well, I made a slightly updated patch to remove usage of a deprecated symbol. Should be good, but I'm not a kernel-dev, so....

[cloneu2]

Thanks guys.. did a rawhide (alpha) reinstall and then updated the kernel. Applied the patch and then did the other trick and IT COMPILED.. Got Nvidia driver working on a dual head and now I can play and see how long it takes to completely trash the system!

[leigh123linux]

I'm pretty sure I found the problem, I forgot I had done this before, but I just updated to the 238 kernel.. It seems the packager is stripping out more stuff in the kernel-devel package. You have to enter the /usr/src/kernels/<kernel> for your current kernel and do "make prepare". This will recreate the include/asm symlink and a couple other things and then error out. Don't worry about the error. After you've done this, you should be able to compile the module. Kinda dumb, but there you go. If still not working, let me know and I'll upload the kmod. Check the nvnews thread as well, I made a slightly updated patch to remove usage of a deprecated symbol. Should be good, but I'm not a kernel-dev, so....

Thanks :cool:


Running ' make prepare ' has fixed the error ! :)

[bseltzer]

Seems the kernel-2.6.27-0.244 kernel update this morning breaks something. After installing this morning's update, the procedure described above, extracting and patching the installer and the doing a "make prepare" in /usr/src/kernels/`uname -r`, then running the installer, produces the same errors Leigh originally reported.

Regards,
Bert

[leigh123linux]

Seems the kernel-2.6.27-0.244 kernel update this morning breaks something. After installing this morning's update, the procedure described above, extracting and patching the installer and the doing a "make prepare" in /usr/src/kernels/`uname -r`, then running the installer, produces the same errors Leigh originally reported.

Regards,
Bert

I have the same here !

Changelog

* Fri Aug 08 2008 Kristian Høgsberg <krh@redhat.com>
- Export module device table for radeon DRM driver.
- Comment out firewire patch 2300 for now, numbering conflict.

* Fri Aug 08 2008 Adam Jackson <ajax@redhat.com>
- Silence the PIIX3 PCI quirk message in quiet boot.

* Fri Aug 08 2008 Dave Airlie <airlied@redhat.com>
- attempt to fix oops in drm_open

* Thu Aug 07 2008 Dave Jones <davej@redhat.com>
- include/asm needs to be called such, not asm-x86.

[bseltzer]

Leigh,

Thanks for the hint... Seems "make prepare" doesn't do what it's supposed to in 2.6.27-0.244. What I found was /usr/src/`uname -r`/include had the "asm" directory and an empty "asm-x86" directory" with no symbolic link. Soooo... I removed the "asm-x86" directory, moved "asm" to "asm-x86", and created a symbolic link called "asm" pointing to "asm-x86" all manually.

Then I went back to the nvidia installer directory, ran the installer script as before, and all is well again.

Aaah, the joys of alpha-ness :D

Regards,
Bert

[vallimar]

Yeah, ran into that as well. For some reason they are messing with the includes, for no reason, and just breaking stuff. They apparently don't understand you need the files in the proper directory and the symlink pointing to the directory for your architecture. The architecture files point to their specific directory and not the generic asm, that is for external inclusion only. Hopefully they'll get hit by a cluebus soon and stop messing up the headers.

[bseltzer]

Hopefully they'll get hit by a cluebus soon and stop messing up the headers.

Not to put to fine a point on this, but I'd rather they got on the cluebus rather than be hit by it. :D

Regards,
Bert

[leigh123linux]

Leigh,

Thanks for the hint... Seems "make prepare" doesn't do what it's supposed to in 2.6.27-0.244. What I found was /usr/src/`uname -r`/include had the "asm" directory and an empty "asm-x86" directory" with no symbolic link. Soooo... I removed the "asm-x86" directory, moved "asm" to "asm-x86", and created a symbolic link called "asm" pointing to "asm-x86" all manually.

Then I went back to the nvidia installer directory, ran the installer script as before, and all is well again.

Aaah, the joys of alpha-ness :D

Regards,
Bert

I fixed it with this command!


ln -sf /usr/src/kernels/2.6.27-0.244.rc2.git1.fc10.x86_64/include/asm /usr/src/kernels/2.6.27-0.244.rc2.git1.fc10.x86_64/include/asm-x86

Then I built the module with this command whilst still running the old kernel ( 2.6.27-0.238.rc2.git1.fc10.x86_64 )

./nvidia-installer --kernel-name=2.6.27-0.244.rc2.git1.fc10.x86_64 --no-runlevel-check --kernel-module-only

[hceylan]

For those like me have no luck with nvidia on 2.6.27 kernel:

You don't have to use 2.6.27 kernel to use F 10 Alpha.

If you are more after F 10 rather then kernel 2.6.27, then just download the 2.6.26.XX kernels from fedora 9 repository and install them.

This way you do not have to mess patches and everything....

Hasan Ceylan

[vallimar]

The new nvidia driver has several fixes rendering a lot of the old patches unneeded.
Based on their work, I've massively reduced the complexity of the patch I am using.

[leigh123linux]

The new nvidia driver has several fixes rendering a lot of the old patches unneeded.
Based on their work, I've massively reduced the complexity of the patch I am using.

Your patch is much better then my attempt :cool: , so I deleted my previous posts .
Your new patch is also i386 & x86_64 compatible :cool:



Thank you for posting your patch .


Leigh

[leigh123linux]

updated post #1

[Debroyston]

Because of the 2.6.27 kernel the nvidia driver needs patching !

32 bit

su -
yum install wget
wget ftp://download.nvidia.com/XFree86/Linux-x86/177.68/NVIDIA-Linux-x86-177.68-pkg1.run
wget http://www.linux-ati-drivers.homecall.co.uk/patch-no-xen.patch
sh NVIDIA-Linux-x86-177.68-pkg1.run --apply-patch ./patch-no-xen.patch
ln -sf /usr/src/kernels/`uname -r`/include/asm /usr/src/kernels/`uname -r`/include/asm-x86
init 3
cd /home/*
sh NVIDIA-Linux-x86-177.68-pkg1-custom.run -an


32 bit module update (run these commands after any new kernel updates)

su -
ln -sf /usr/src/kernels/`uname -r`/include/asm /usr/src/kernels/`uname -r`/include/asm-x86
init 3
cd /home/*
sh NVIDIA-Linux-x86-177.68-pkg1-custom.run -an





64 bit

su -
yum install wget
wget ftp://download.nvidia.com/XFree86/Linux-x86_64/177.68/NVIDIA-Linux-x86_64-177.68-pkg2.run
wget http://www.linux-ati-drivers.homecall.co.uk/patch-no-xen.patch
sh NVIDIA-Linux-x86_64-177.68-pkg2.run --apply-patch ./patch-no-xen.patch
ln -sf /usr/src/kernels/`uname -r`/include/asm /usr/src/kernels/`uname -r`/include/asm-x86
init 3
cd /home/*
sh NVIDIA-Linux-x86_64-177.68-pkg2-custom.run -an




64 bit module update (run these commands after any new kernel updates)

su -
ln -sf /usr/src/kernels/`uname -r`/include/asm /usr/src/kernels/`uname -r`/include/asm-x86
init 3
cd /home/*
NVIDIA-Linux-x86_64-177.68-pkg2-custom.run -an
I am using a BenQ FP767 monitor with a GT8500 graphics card with Fedora 10 alpha x86_64 OS.
Unfortunately this did not work for me.
Previous Fedora OSs used Livna which has not covered this release.
Any further advice would be appreciated.

[vallimar]

You've not exactly given us anything to go on...
How did it not work for you?

[vallimar]

Just an aside, the newer koji kernels finally reverted the include/asm back to a symlink.
* Wed Aug 20 18:00:00 2008 Dave Jones <davej@redhat.com>
- include/asm needs to be a symlink, not a dir.Hopefully these won't take much longer to get pushed out.
Though I've noticed rawhide still has yet to be updated.

[Debroyston]

You've not exactly given us anything to go on...
How did it not work for you?
Everything went OK until the final line:-
sh NVIDIA-Linux-x86__64-177.68-pkg2-custom.run -an.
The cd /home/* pointed me to the Lost & Found folder. Is this correct?

[ryptyde]

Everything went OK until the final line:-
sh NVIDIA-Linux-x86__64-177.68-pkg2-custom.run -an.
The cd /home/* pointed me to the Lost & Found folder. Is this correct?

Running cd /home/* put me in my home directory where the nvidia package was d'loaded to. :)

[vallimar]

Everything went OK until the final line:-
sh NVIDIA-Linux-x86__64-177.68-pkg2-custom.run -an.
The cd /home/* pointed me to the Lost & Found folder. Is this correct?That line is more a demonstrative example pointing out that you need
change directories to where you downloaded the driver. Once you
are in the same directory, execute that last line on it the driver file.
Or you can run it from anywhere doing: "sh </path/to/driver.run> -an"

[neighborlee]

The new nvidia driver has several fixes rendering a lot of the old patches unneeded.
Based on their work, I've massively reduced the complexity of the patch I am using.


I got it compiled fine after inserting what livecd left out (grins), but glxinfo wont run, showing:
glxinfo: error while loading shared libraries: libGL.so.1: failed to map segment from shared object: Permission denied.

I noticed the latest /include/asm comment, but today when I installed fedora and then got the latest updates, I the rebooted and chose 244 kernel, which has asm as NOT a symlink, so I am finding this thread rather confusing ;)

thx anyone
nl

[leigh123linux]

I got it compiled fine after inserting what livecd left out (grins), but glxinfo wont run, showing:
glxinfo: error while loading shared libraries: libGL.so.1: failed to map segment from shared object: Permission denied.

I noticed the latest /include/asm comment, but today when I installed fedora and then got the latest updates, I the rebooted and chose 244 kernel, which has asm as NOT a symlink, so I am finding this thread rather confusing ;)

thx anyone
nl


Selinux blocks glxinfo

Try

su -
chcon -t unconfined_execmem_exec_t '/usr/bin/nvidia-settings'
chcon -t unconfined_execmem_exec_t '/usr/bin/glxinfo'



I the rebooted and chose 244 kernel, which has asm as NOT a symlink, so I am finding this thread rather confusing ;)

You need to re-run the installer !

su -
ln -sf /usr/src/kernels/`uname -r`/include/asm /usr/src/kernels/`uname -r`/include/asm-x86
init 3
cd /home/*
sh NVIDIA-Linux-x86-177.68-pkg1-custom.run -an

[neighborlee]

Selinux blocks glxinfo

Try

su -
chcon -t unconfined_execmem_exec_t '/usr/bin/nvidia-settings'
chcon -t unconfined_execmem_exec_t '/usr/bin/glxinfo'





You need to re-run the installer !

su -
ln -sf /usr/src/kernels/`uname -r`/include/asm /usr/src/kernels/`uname -r`/include/asm-x86
init 3
cd /home/*
sh NVIDIA-Linux-x86-177.68-pkg1-custom.run -an


ok thx on that..it was just my perception I guess that wasn't working, because atm my system is running very slowly for a 8800gtx dual core setup, and looking at system monitor its clear that Ive got alot of cpu activity with it sometimes going over 100% between both cores, which I know is wrong due to similar viewing in xp.

I suspect this is due to 244 and alpha status combined, assuming everyone is running 244 kernel, as that seems to be the latest one , and yes I did update successfully yesterday.

thx ;)
lee

[leigh123linux]

Rpm method added to post #1



su
wget http://www.linux-ati-drivers.homecall.co.uk/nvidia.repo -O /etc/yum.repos.d/nvidia.repo
yum install nvidia-x11-drv*
/usr/sbin/nvidia-xconfig
reboot

[chepioq]

thank leigh...
That's work!!!

[vallimar]

Not that I need it personally, by why no 177.70 version?

http://www.nvnews.net/vbulletin/showthread.php?t=118602

[leigh123linux]

Not that I need it personally, by why no 177.70 version?

http://www.nvnews.net/vbulletin/showthread.php?t=118602


Thanks for the link :cool:

I have a 7600 .

# Fixed a text rendering performance regression that affected
GeForce 6 and 7 series GPUs.
# Fixed a regression that caused text to be missing or
corrupt on GeForce 6 and 7 series GPUs.


I am building it now ! :)

Edit : built and uploaded to the server !

http://www.dfm.uninsubria.it/compiz/fusion-testing/9/nvidia/x86_64/

[leigh123linux]

Guide updated to 177.76 .

[Perpetual]

Would this issue with Nvidia be the reason for me not being able to boot the live cd? I can boot F9 just fine but not the F10 Alpha. FYI, same issue when I tried Ubuntu. Can boot the current version but not the Alphas.

Thanks.

[crainey69]

Is this working for everyone?

I'm running kernel 2.6.27-0.352.rc7.git1.fc10.i686 and followed the steps (that I believe current in post 1) in the easy section and it fails. Only error that comes to mind is, "no screens found".

Thanks

Cory

3d isn't too important as it will be months before I make the real change after the final comes out but, I'm beginning on the work for the move now.

[Demz]

another Newer Linux Beta Driver released 177.78 http://www.nvnews.net/vbulletin/showthread.php?t=120052

[cloneu2]

the new beta driver works just fine here. Had to do a little trick for the twin view to get 1680x1050 but that is ok.

--------------Xorg.conf------------------------------------
Section "ServerLayout"
Identifier "Layout0"
Screen 0 "Screen0" 0 0
InputDevice "Keyboard0" "CoreKeyboard"
InputDevice "Mouse0" "CorePointer"
EndSection

Section "Files"
EndSection

Section "Module"
Load "dbe"
Load "extmod"
Load "type1"
Load "freetype"
Load "glx"
EndSection

Section "ServerFlags"
Option "Xinerama" "0"
EndSection

Section "InputDevice"

# generated from default
Identifier "Mouse0"
Driver "mouse"
Option "Protocol" "auto"
Option "Device" "/dev/input/mice"
Option "Emulate3Buttons" "no"
Option "ZAxisMapping" "4 5"
EndSection

Section "InputDevice"

# generated from data in "/etc/sysconfig/keyboard"
Identifier "Keyboard0"
Driver "kbd"
Option "XkbLayout" "us"
Option "XkbModel" "pc105"
EndSection

Section "Monitor"
Identifier "Monitor0"
VendorName "Unknown"
ModelName "CRT-0"
HorizSync 28.0 - 55.0
VertRefresh 43.0 - 72.0
Option "DPMS"
EndSection

Section "Device"
Identifier "Device0"
Driver "nvidia"
VendorName "NVIDIA Corporation"
BoardName "GeForce 7100 GS"
Option "IgnoreEDID" "true"
EndSection

Section "Screen"
Identifier "Screen0"
Device "Device0"
Monitor "Monitor0"
DefaultDepth 24
Option "AddARGBGLXVisuals" "True"
Option "TripleBuffer" "True"
Option "TwinView" "1"
Option "TwinViewXineramaInfoOrder" "DFP-0"
Option "CustomEDID" "CRT:/etc/X11/edid.bin"
Option "metamodes" "CRT: 1680x1050_60 +1680+0, DFP: nvidia-auto-select +0+0"
SubSection "Display"
Depth 24
EndSubSection
EndSection

Section "Extensions"
Option "Composite" "Enable"
EndSection


The custom EDID fixed the second screen from a LOUSY resolution!

[crainey69]

Just an update on my end.

I haven't messed around much with the nvidia driver in F10 much since my last post. A few days ago, my boot began failing after starting cups (turns out to be where it was starting X). The nvidia packages that I was using to try to get it working were the cause. Took them all back off, renamed XF86Config and renamed xorg.conf and at least I can boot to X again.

The XF86Config file showed the driver as nv though instead of nvidia. Don't remember changing that I did change the xorg.conf to nv though after nvidia failed.

[leigh123linux]

Repo & guide updated to 177.80

[leigh123linux]

MAKEDEV-3.23-7 breaks the Nvidia driver so be careful when updating today ( 10/10/08 )

/dev/nvidia0 is omitted on reboot


https://bugzilla.redhat.com/show_bug.cgi?id=466385

[ninja600j]

Hi,

Can you point me in the right direction to fixing this please? I already updated today and killed it.

Thanks.

[leigh123linux]

Hi,

Can you point me in the right direction to fixing this please? I already updated today and killed it.

Thanks.

Download the old MAKEDEV package ( grab the right arch )

http://koji.fedoraproject.org/koji/buildinfo?buildID=63659


and replace with this command

rpm -Uvh --replacefiles --replacepkgs --oldpackage MAKEDEV-3.23-6*rpm



also edit /etc/yum.conf

and add this line

exclude=MAKEDEV-3.23-7

i.e

[main]
cachedir=/var/cache/yum
keepcache=0
debuglevel=2
logfile=/var/log/yum.log
exactarch=1
exclude=MAKEDEV-3.23-7
obsoletes=1
gpgcheck=1
plugins=1
installonly_limit=3

# This is the default, if you make this bigger yum won't see if the metadata
# is newer on the remote and so you'll "gain" the bandwidth of not having to
# download the new metadata and "pay" for it by yum not having correct
# information.
# It is esp. important, to have correct metadata, for distributions like
# Fedora which don't keep old packages around. If you don't like this checking
# interupting your command line usage, it's much better to have something
# manually check the metadata once an hour (yum-updatesd will do this).
# metadata_expire=90m

# PUT YOUR REPOS HERE OR IN separate files named file.repo
# in /etc/yum.repos.d

[ninja600j]

Worked a treat. Thank you.

[daf1983]

thanks leigh123@linux!

[neighborlee]

MAKEDEV-3.23-7 breaks the Nvidia driver so be careful when updating today ( 10/10/08 )

/dev/nvidia0 is omitted on reboot


https://bugzilla.redhat.com/show_bug.cgi?id=466385

Does this still apply, even though I just updated this morning..I suspect so as after update and having reinstalled nvidia and ran nvidia-xconfig I started having trouble with xserver, but I wanted to be sure its not something else.

cheer
nl

[leigh123linux]

Does this still apply, even though I just updated this morning..I suspect so as after update and having reinstalled nvidia and ran nvidia-xconfig I started having trouble with xserver, but I wanted to be sure its not something else.

cheer
nl
It still applies .

[Captn]

IF i do this on my freshly installed & updated FC9 64bit


yum install kernel-devel kernel-headers


1.akmod-nvidia

2.kmod-nvidia

3.Reboot my system



What would i have to do to install the 177.80 drivers at nvidias website when everything is fixed?

or should i just wait ?

[quantumburnz]

Really wish I would have seen this post earlier... I let my livna-nvidia drivers get updated and I can no longer boot into FC10beta any more. It begins to load like normal, gets to "Starting cpus: [OK]", the screen starts to flash some and then it just hangs. I tried "/usr/sbin/nvidia-config-display disable" but that didn't seem to do much good. Any way to get myself out of this jam?

[daf1983]

Ctrl+Alt+F2, become root, go to xorg.conf and change your driver to 'nv', startx and then you can downgrade your MAKEDEV as described in leigh's post.

you can also do it from the console but it depends on you.

[quantumburnz]

Is there any way to remove the livna nvidia drivers rather than going back to the old MAKEDEV?

[daf1983]

well, probably there is but i found downgrading MAKEDEV much easier. you can try to uninstall akmods and kmods but i did it first without success as the driver kept loading after each reboot so i finally went back to MAKEDEV-3.23-6 and it works like a charm with kmod-nvidia.

[Captn]

Is the MAKEDEV busted in FC9 also, or is this a FC10 only problem? Whats the command so i can see what version of MAKEDEV im running? Sorry 1 more thing hehe will leigh123@linux post to edit out also edit /etc/yum.conf

and add this line

exclude=MAKEDEV-3.23-7


work for FC9 64bit? Thank you in advance.

[ogetbilo]

Is the MAKEDEV busted in FC9 also, or is this a FC10 only problem? Whats the command so i can see what version of MAKEDEV im running? Sorry 1 more thing hehe will leigh123@linux post to edit out also edit /etc/yum.conf

and add this line

exclude=MAKEDEV-3.23-7


work for FC9 64bit? Thank you in advance.

As the title says the problem is specific to F10. If you have not enabled rawhide repo you are fine.

[Captn]

As the title says the problem is specific to F10. If you have not enabled rawhide repo you are fine.


Ok, i was asking because after i install the 177.80 from nvidias site and when i go to startx i cannot. This is right after a fresh install & updates are done.

[leigh123linux]

Ok, i was asking because after i install the 177.80 from nvidias site and when i go to startx i cannot. This is right after a fresh install & updates are done.

If you have followed post #1 I am not surprised as this guide is specific to F10 only .

[chepioq]

Hello leigh
This tuto for F10 work fine, but for F9?
I use actually the nvidia drivers from livna, but there is only the 173.14 version of nvidia drivers with livna. (and for F10 beta there 177.80...)
Do you think that your tuto work with F9?

[leigh123linux]

Hello leigh
This tuto for F10 work fine, but for F9?
I use actually the nvidia drivers from livna, but there is only the 173.14 version of nvidia drivers with livna. (and for F10 beta there 177.80...)
Do you think that your tuto work with F9?


For F9 64 bit

su
yum groupinstall "Development Tools"
yum install kernel-devel kernel-headers wget
wget http://us.download.nvidia.com/XFree86/Linux-x86_64/177.80/NVIDIA-Linux-x86_64-177.80-pkg2.run
su -
init 3
cd /home/*
sh NVIDIA-Linux-x86_64-177.80-pkg2.run -an


Or rebuild the SRPM from this post

http://forums.fedoraforum.org/showthread.php?t=201292

su
yum groupinstall "Development Tools"
yum install rpmdevtools wget
wget http://www.dfm.uninsubria.it/compiz/fusion-testing/9/nvidia/SRPM/nvidia-x11-drv-177.80-1.fc9.src.rpm
yum-buildep nvidia-x11-drv-177.80-1.fc9.src.rpm
exit
rpmdev-setuptree
rpmbuild --rebuild nvidia-x11-drv-177.80-1.fc9.src.rpm

[chepioq]

sorry that dont work:
# yum-buildeps nvidia-x11-drv-177.80-1.fc9.src.rpm
-bash: yum-buildeps: command not found


Development Tools, rpmdevtools and wget are installed.

[leigh123linux]

sorry that dont work:
# yum-buildeps nvidia-x11-drv-177.80-1.fc9.src.rpm
-bash: yum-buildeps: command not found


Development Tools, rpmdevtools and wget are installed.

sorry it's a typo


try


yum-builddep nvidia-x11-drv-177.80-1.fc9.src.rpm

[Waggoneer]

I was able to get up and running on Nvidia with the aforementioned walk-through. I do get one error when I try and run glxinfo.

[Jason@linbox ~]$ glxinfo
name of display: :0.0
Error: glXCreateContext failed

This is running on the EVGA Nvidia GeForce 8600 GT.

[chepioq]

@leigh
There is another typo:
yum-buildep nvidia-x11-drv-177.80-1.fc9.src.rpm
don't work but
yum-builddep nvidia-x11-drv-177.80-1.fc9.src.rpm
work

[chepioq]

Well... i have installed the 177.80 version and that work fine...
But now if I want to go to the livna version I have a problem:
I remove the package of 177.80 and install livna one but when I reboot, X don't launch.
The errors is that there is a 177.80 version of nvidia driver (I don't note the error's messages).
What file I do remove?

[leigh123linux]

Well... i have installed the 177.80 version and that work fine...
But now if I want to go to the livna version I have a problem:
I remove the package of 177.80 and install livna one but when I reboot, X don't launch.
The errors is that there is a 177.80 version of nvidia driver (I don't note the error's messages).
What file I do remove?

Post


rpm -qa |grep nvidia

[chepioq]

Sorry leigh, but I have re-install the 177.80 version and remove it, and install the livna one and now that work...
I don't know what I do...

[yunhui]

I'm use Dell D630 nvs 135M
give error follow, how to fix it?
(EE) NVIDIA(0): Failed to initialize the NVIDIA graphics device PCI:1:0:0.
(EE) NVIDIA(0): Please see the COMMON PROBLEMS section in the README for
(EE) NVIDIA(0): additional information.
(EE) NVIDIA(0): Failed to initialize the NVIDIA graphics device!

[leigh123linux]

I'm use Dell D630 nvs 135M
give error follow, how to fix it?
(EE) NVIDIA(0): Failed to initialize the NVIDIA graphics device PCI:1:0:0.
(EE) NVIDIA(0): Please see the COMMON PROBLEMS section in the README for
(EE) NVIDIA(0): additional information.
(EE) NVIDIA(0): Failed to initialize the NVIDIA graphics device!


Try downgrading MAKEDEV ( see posts 46 & 48 )

[Boricua]

For some reason, my module build keeps failing :(. leigh' s "automatic" method worked originally for me while using my nvidia 8600 GT card. However, the card was damaged and I switched to a nvidia 7200. Now, the system starts trying to build the new module for several seconds and then fails :eek:. This is my 64 bit Fedora 10 system, up-to-date, so any help is appreciated :).

[kosh]

I've got nvidia module installed without any problems.
#init 3
#./NVIDIA-Linux-x86-177.80-pkg1.run
Then alter '/etc/X11xorg.conf'

[vallimar]

Try downgrading MAKEDEV ( see posts 46 & 48 )
Has anyone tried doing MAKEDEV nvidia -m 3 to create some nvidia devices instead of
downgrading the entire package? Or even manually adding rules to udev?

If nobody else has messed with it, I may have to fire up the nvidia driver and see what
I can work out. I've been lazy and just using the nv driver now that it supports AIGLX
and been using cheap compositing in metacity for transparency.

[vallimar]

http://koji.fedoraproject.org/koji/buildinfo?buildID=67080
* Wed Oct 22 2008 Chris Lumens <clumens@redhat.com> 3.24-1
- Fix speed up patch (jakub, #466485).
Might fix the issue since it was the speedup patch that caused the errors.

[leigh123linux]

http://koji.fedoraproject.org/koji/buildinfo?buildID=67080

Might fix the issue since it was the speedup patch that caused the errors.


I can confirm that the new MAKEDEV package fixes the issue :cool:

[leigh@localhost ~]$ uname -r
2.6.27.3-30.rc1.fc10.x86_64
[leigh@localhost ~]$ ls -al /dev |grep nvidia
crw------- 1 leigh root 195, 0 2008-10-22 16:48 nvidia0
crw------- 1 leigh root 195, 1 2008-10-22 16:48 nvidia1
crw------- 1 leigh root 195, 2 2008-10-22 16:48 nvidia2
crw------- 1 leigh root 195, 3 2008-10-22 16:48 nvidia3
crw------- 1 leigh root 195, 255 2008-10-22 16:48 nvidiactl
[leigh@localhost ~]$ ls -al /dev |grep loop
brw-rw---- 1 root disk 7, 0 2008-10-22 16:48 loop0
brw-rw---- 1 root disk 7, 1 2008-10-22 16:48 loop1
brw-rw---- 1 root disk 7, 2 2008-10-22 16:48 loop2
brw-rw---- 1 root disk 7, 3 2008-10-22 16:48 loop3
brw-rw---- 1 root disk 7, 4 2008-10-22 16:48 loop4
brw-rw---- 1 root disk 7, 5 2008-10-22 16:48 loop5
brw-rw---- 1 root disk 7, 6 2008-10-22 16:48 loop6
brw-rw---- 1 root disk 7, 7 2008-10-22 16:48 loop7
[leigh@localhost ~]$ ls -al /dev |grep lp
crw-rw---- 1 root lp 6, 0 2008-10-22 16:48 lp0
crw-rw---- 1 root lp 6, 1 2008-10-22 16:48 lp1
crw-rw---- 1 root lp 6, 2 2008-10-22 16:48 lp2
crw-rw---- 1 root lp 6, 3 2008-10-22 16:48 lp3
[leigh@localhost ~]$ rpm -q MAKEDEV
MAKEDEV-3.24-1.x86_64
[leigh@localhost ~]$

[jrummy27]

Should this work with the legacy drivers as well? I tried your procedure from post #1 which led to this error upon load: (WW) NVIDIA(0): The NVIDIA Quadro FX 330 GPU installed in this system is
(WW) NVIDIA(0): supported through the NVIDIA 173.14.xx Legacy drivers.
(WW) NVIDIA(0): Please visit http://www.nvidia.com/object/unix.html for
(WW) NVIDIA(0): more information. The 177.80 NVIDIA driver will ignore
(WW) NVIDIA(0): this GPU. Continuing probe...
(EE) No devices detected. I have a nVidia Corporation NV37GL [Quadro PCI-E Series] so I assume that's right. So I went and downloaded the legacy version of the driver and followed your procedure with it, however I get this error:
-> Building kernel module:
executing: 'cd ./usr/src/nv; make module SYSSRC=/lib/modules/2.6.27.3-39.fc1
0.x86_64/source SYSOUT=/lib/modules/2.6.27.3-39.fc10.x86_64/build'...
NVIDIA: calling KBUILD...
make CC=cc KBUILD_OUTPUT=/lib/modules/2.6.27.3-39.fc10.x86_64/build KBUILD_V
ERBOSE=1 -C /lib/modules/2.6.27.3-39.fc10.x86_64/source SUBDIRS=/tmp/selfgz6
594/NVIDIA-Linux-x86_64-173.14.12-pkg2-custom/usr/src/nv modules
make -C /lib/modules/2.6.27.3-39.fc10.x86_64/build \
KBUILD_SRC=/usr/src/kernels/2.6.27.3-39.fc10.x86_64 \
KBUILD_EXTMOD="/tmp/selfgz6594/NVIDIA-Linux-x86_64-173.14.12-pkg2-custom/us
r/src/nv" -f /usr/src/kernels/2.6.27.3-39.fc10.x86_64/Makefile \
modules
test -e include/linux/autoconf.h -a -e include/config/auto.conf || ( \
echo; \
echo " ERROR: Kernel configuration is invalid."; \
echo " include/linux/autoconf.h or include/config/auto.conf are mis
sing."; \
echo " Run 'make oldconfig && make prepare' on kernel src to fix it
."; \
echo; \
/bin/false)
mkdir -p /tmp/selfgz6594/NVIDIA-Linux-x86_64-173.14.12-pkg2-custom/usr/src/n
v/.tmp_versions ; rm -f /tmp/selfgz6594/NVIDIA-Linux-x86_64-173.14.12-pkg2-c
ustom/usr/src/nv/.tmp_versions/*
make -f /usr/src/kernels/2.6.27.3-39.fc10.x86_64/scripts/Makefile.build obj=
/tmp/selfgz6594/NVIDIA-Linux-x86_64-173.14.12-pkg2-custom/usr/src/nv
cc -Wp,-MD,/tmp/selfgz6594/NVIDIA-Linux-x86_64-173.14.12-pkg2-custom/usr/s
rc/nv/.nv.o.d -nostdinc -isystem /usr/lib/gcc/x86_64-redhat-linux/4.3.2/inc
lude -D__KERNEL__ -Iinclude -Iinclude2 -I/usr/src/kernels/2.6.27.3-39.fc10.x
86_64/include -I/usr/src/kernels/2.6.27.3-39.fc10.x86_64/arch/x86/include -i
nclude include/linux/autoconf.h -I/tmp/selfgz6594/NVIDIA-Linux-x86_64-173.1
4.12-pkg2-custom/usr/src/nv -Wall -Wundef -Wstrict-prototypes -Wno-trigraphs
-fno-strict-aliasing -fno-common -Werror-implicit-function-declaration -Os -
m64 -mtune=generic -mno-red-zone -mcmodel=kernel -funit-at-a-time -maccumula
te-outgoing-args -DCONFIG_AS_CFI=1 -DCONFIG_AS_CFI_SIGNAL_FRAME=1 -pipe -Wno
-sign-compare -fno-asynchronous-unwind-tables -mno-sse -mno-mmx -mno-sse2 -m
no-3dnow -I/usr/src/kernels/2.6.27.3-39.fc10.x86_64/include/asm-x86/mach-def
ault -Iinclude/asm-x86/mach-default -fno
-stack-protector -fno-omit-frame-pointer -fno-optimize-sibling-calls -g -Wde
claration-after-statement -Wno-pointer-sign -I/tmp/selfgz6594/NVIDIA-Linux-
x86_64-173.14.12-pkg2-custom/usr/src/nv -Wall -Wimplicit -Wreturn-type -Wswi
tch -Wformat -Wchar-subscripts -Wparentheses -Wpointer-arith -Wno-multichar
-Werror -mcmodel=kernel -mno-red-zone -MD -Wsign-compare -Wno-cast-qual -Wno
-error -D__KERNEL__ -DMODULE -DNVRM -DNV_VERSION_STRING=\"173.14.12\" -UDEBU
G -U_DEBUG -DNDEBUG -DMODULE -D"KBUILD_STR(s)=#s" -D"KBUILD_BASENAME=KBUILD_
STR(nv)" -D"KBUILD_MODNAME=KBUILD_STR(nvidia)" -c -o /tmp/selfgz6594/NVIDIA
-Linux-x86_64-173.14.12-pkg2-custom/usr/src/nv/nv.o /tmp/selfgz6594/NVIDIA-L
inux-x86_64-173.14.12-pkg2-custom/usr/src/nv/nv.c
In file included from include/linux/bitops.h:17,
from include/linux/kernel.h:15,
from include/linux/sched.h:52,
from include/linux/utsname.h:35,
from /tmp/selfgz6594/NVIDIA-Linux-x86_64-173.14.12-pkg2-cus
tom/usr/src/nv/nv-linux.h:19,
from /tmp/selfgz6594/NVIDIA-Linux-x86_64-173.14.12-pkg2-cus
tom/usr/src/nv/nv.c:14:
include/asm/bitops.h: In function ‘set_bit’:
include/asm/bitops.h:60: warning: pointer of type ‘void *’ used in arith
metic
include/asm/bitops.h: In function ‘clear_bit’:
include/asm/bitops.h:97: warning: pointer of type ‘void *’ used in arith
metic
In file included from include/linux/utsname.h:35,
from /tmp/selfgz6594/NVIDIA-Linux-x86_64-173.14.12-pkg2-cus
tom/usr/src/nv/nv-linux.h:19,
from /tmp/selfgz6594/NVIDIA-Linux-x86_64-173.14.12-pkg2-cus
tom/usr/src/nv/nv.c:14:
include/linux/sched.h: In function ‘object_is_on_stack’:
include/linux/sched.h:1983: warning: pointer of type ‘void *’ used in ar
ithmetic
In file included from include/asm/dma-mapping.h:9,
from include/linux/dma-mapping.h:52,
from include/asm-generic/pci-dma-compat.h:7,
from include/asm/pci.h:94,
from include/linux/pci.h:983,
from /tmp/selfgz6594/NVIDIA-Linux-x86_64-173.14.12-pkg2-cus
tom/usr/src/nv/nv-linux.h:86,
from /tmp/selfgz6594/NVIDIA-Linux-x86_64-173.14.12-pkg2-cus
tom/usr/src/nv/nv.c:14:
include/linux/scatterlist.h: In function ‘sg_virt’:
include/linux/scatterlist.h:199: warning: pointer of type ‘void *’ used
in arithmetic
In file included from /tmp/selfgz6594/NVIDIA-Linux-x86_64-173.14.12-pkg2-cus
tom/usr/src/nv/nv.c:14:
/tmp/selfgz6594/NVIDIA-Linux-x86_64-173.14.12-pkg2-custom/usr/src/nv/nv-linu
x.h:107:27: error: asm/semaphore.h: No such file or directory
In file included from /tmp/selfgz6594/NVIDIA-Linux-x86_64-173.14.12-pkg2-cus
tom/usr/src/nv/nv-linux.h:109,
from /tmp/selfgz6594/NVIDIA-Linux-x86_64-173.14.12-pkg2-cus
tom/usr/src/nv/nv.c:14:
include/linux/highmem.h: In function ‘zero_user_segments’:
include/linux/highmem.h:134: warning: pointer of type ‘void *’ used in a
rithmetic
include/linux/highmem.h:137: warning: pointer of type ‘void *’ used in a
rithmetic
In file included from include/linux/compat.h:14,
from include/asm/mtrr.h:141,
from /tmp/selfgz6594/NVIDIA-Linux-x86_64-173.14.12-pkg2-cus
tom/usr/src/nv/nv-linux.h:121,
from /tmp/selfgz6594/NVIDIA-Linux-x86_64-173.14.12-pkg2-cus
tom/usr/src/nv/nv.c:14:
include/asm/compat.h: In function ‘compat_alloc_user_space’:
include/asm/compat.h:210: warning: pointer of type ‘void *’ used in arit
hmetic
In file included from /tmp/selfgz6594/NVIDIA-Linux-x86_64-173.14.12-pkg2-cus
tom/usr/src/nv/nv.c:14:
/tmp/selfgz6594/NVIDIA-Linux-x86_64-173.14.12-pkg2-custom/usr/src/nv/nv-linu
x.h: In function ‘nv_execute_on_all_cpus’:
/tmp/selfgz6594/NVIDIA-Linux-x86_64-173.14.12-pkg2-custom/usr/src/nv/nv-linu
x.h:674: error: too many arguments to function ‘on_each_cpu’
/tmp/selfgz6594/NVIDIA-Linux-x86_64-173.14.12-pkg2-custom/usr/src/nv/nv.c: I
n function ‘__nv_setup_pat_entries’:
/tmp/selfgz6594/NVIDIA-Linux-x86_64-173.14.12-pkg2-custom/usr/src/nv/nv.c:94
4: warning: comparison between signed and unsigned
/tmp/selfgz6594/NVIDIA-Linux-x86_64-173.14.12-pkg2-custom/usr/src/nv/nv.c: I
n function ‘__nv_restore_pat_entries’:
/tmp/selfgz6594/NVIDIA-Linux-x86_64-173.14.12-pkg2-custom/usr/src/nv/nv.c:97
0: warning: comparison between signed and unsigned
/tmp/selfgz6594/NVIDIA-Linux-x86_64-173.14.12-pkg2-custom/usr/src/nv/nv.c: I
n function ‘nv_kern_cpu_callback’:
/tmp/selfgz6594/NVIDIA-Linux-x86_64-173.14.12-pkg2-custom/usr/src/nv/nv.c:12
96: warning: comparison between signed and unsigned
/tmp/selfgz6594/NVIDIA-Linux-x86_64-173.14.12-pkg2-custom/usr/src/nv/nv.c:12
99: error: too many arguments to function ‘smp_call_function’
/tmp/selfgz6594/NVIDIA-Linux-x86_64-173.14.12-pkg2-custom/usr/src/nv/nv.c:13
03: warning: comparison between signed and unsigned
/tmp/selfgz6594/NVIDIA-Linux-x86_64-173.14.12-pkg2-custom/usr/src/nv/nv.c:13
06: error: too many arguments to function ‘smp_call_function’
make[4]: *** [/tmp/selfgz6594/NVIDIA-Linux-x86_64-173.14.12-pkg2-custom/usr/
src/nv/nv.o] Error 1
make[3]: *** [_module_/tmp/selfgz6594/NVIDIA-Linux-x86_64-173.14.12-pkg2-cus
tom/usr/src/nv] Error 2
make[2]: *** [sub-make] Error 2
NVIDIA: left KBUILD.
nvidia.ko failed to build!
make[1]: *** [module] Error 1
make: *** [module] Error 2
-> Error.
ERROR: Unable to build the NVIDIA kernel module.
ERROR: Installation has failed. Please see the file
'/var/log/nvidia-installer.log' for details. You may find suggestions
on fixing installation problems in the README available on the Linux
driver download page at www.nvidia.com.
Any ideas?

[leigh123linux]

Should this work with the legacy drivers as well? I tried your procedure from post #1 which led to this error upon load: (WW) NVIDIA(0): The NVIDIA Quadro FX 330 GPU installed in this system is
(WW) NVIDIA(0): supported through the NVIDIA 173.14.xx Legacy drivers.
(WW) NVIDIA(0): Please visit http://www.nvidia.com/object/unix.html for
(WW) NVIDIA(0): more information. The 177.80 NVIDIA driver will ignore
(WW) NVIDIA(0): this GPU. Continuing probe...
(EE) No devices detected. I have a nVidia Corporation NV37GL [Quadro PCI-E Series] so I assume that's right. So I went and downloaded the legacy version of the driver and followed your procedure with it, however I get this error:
-> Building kernel module:
executing: 'cd ./usr/src/nv; make module SYSSRC=/lib/modules/2.6.27.3-39.fc1
0.x86_64/source SYSOUT=/lib/modules/2.6.27.3-39.fc10.x86_64/build'...
NVIDIA: calling KBUILD...
make CC=cc KBUILD_OUTPUT=/lib/modules/2.6.27.3-39.fc10.x86_64/build KBUILD_V
ERBOSE=1 -C /lib/modules/2.6.27.3-39.fc10.x86_64/source SUBDIRS=/tmp/selfgz6
594/NVIDIA-Linux-x86_64-173.14.12-pkg2-custom/usr/src/nv modules
make -C /lib/modules/2.6.27.3-39.fc10.x86_64/build \
KBUILD_SRC=/usr/src/kernels/2.6.27.3-39.fc10.x86_64 \
KBUILD_EXTMOD="/tmp/selfgz6594/NVIDIA-Linux-x86_64-173.14.12-pkg2-custom/us
r/src/nv" -f /usr/src/kernels/2.6.27.3-39.fc10.x86_64/Makefile \
modules
test -e include/linux/autoconf.h -a -e include/config/auto.conf || ( \
echo; \
echo " ERROR: Kernel configuration is invalid."; \
echo " include/linux/autoconf.h or include/config/auto.conf are mis
sing."; \
echo " Run 'make oldconfig && make prepare' on kernel src to fix it
."; \
echo; \
/bin/false)
mkdir -p /tmp/selfgz6594/NVIDIA-Linux-x86_64-173.14.12-pkg2-custom/usr/src/n
v/.tmp_versions ; rm -f /tmp/selfgz6594/NVIDIA-Linux-x86_64-173.14.12-pkg2-c
ustom/usr/src/nv/.tmp_versions/*
make -f /usr/src/kernels/2.6.27.3-39.fc10.x86_64/scripts/Makefile.build obj=
/tmp/selfgz6594/NVIDIA-Linux-x86_64-173.14.12-pkg2-custom/usr/src/nv
cc -Wp,-MD,/tmp/selfgz6594/NVIDIA-Linux-x86_64-173.14.12-pkg2-custom/usr/s
rc/nv/.nv.o.d -nostdinc -isystem /usr/lib/gcc/x86_64-redhat-linux/4.3.2/inc
lude -D__KERNEL__ -Iinclude -Iinclude2 -I/usr/src/kernels/2.6.27.3-39.fc10.x
86_64/include -I/usr/src/kernels/2.6.27.3-39.fc10.x86_64/arch/x86/include -i
nclude include/linux/autoconf.h -I/tmp/selfgz6594/NVIDIA-Linux-x86_64-173.1
4.12-pkg2-custom/usr/src/nv -Wall -Wundef -Wstrict-prototypes -Wno-trigraphs
-fno-strict-aliasing -fno-common -Werror-implicit-function-declaration -Os -
m64 -mtune=generic -mno-red-zone -mcmodel=kernel -funit-at-a-time -maccumula
te-outgoing-args -DCONFIG_AS_CFI=1 -DCONFIG_AS_CFI_SIGNAL_FRAME=1 -pipe -Wno
-sign-compare -fno-asynchronous-unwind-tables -mno-sse -mno-mmx -mno-sse2 -m
no-3dnow -I/usr/src/kernels/2.6.27.3-39.fc10.x86_64/include/asm-x86/mach-def
ault -Iinclude/asm-x86/mach-default -fno
-stack-protector -fno-omit-frame-pointer -fno-optimize-sibling-calls -g -Wde
claration-after-statement -Wno-pointer-sign -I/tmp/selfgz6594/NVIDIA-Linux-
x86_64-173.14.12-pkg2-custom/usr/src/nv -Wall -Wimplicit -Wreturn-type -Wswi
tch -Wformat -Wchar-subscripts -Wparentheses -Wpointer-arith -Wno-multichar
-Werror -mcmodel=kernel -mno-red-zone -MD -Wsign-compare -Wno-cast-qual -Wno
-error -D__KERNEL__ -DMODULE -DNVRM -DNV_VERSION_STRING=\"173.14.12\" -UDEBU
G -U_DEBUG -DNDEBUG -DMODULE -D"KBUILD_STR(s)=#s" -D"KBUILD_BASENAME=KBUILD_
STR(nv)" -D"KBUILD_MODNAME=KBUILD_STR(nvidia)" -c -o /tmp/selfgz6594/NVIDIA
-Linux-x86_64-173.14.12-pkg2-custom/usr/src/nv/nv.o /tmp/selfgz6594/NVIDIA-L
inux-x86_64-173.14.12-pkg2-custom/usr/src/nv/nv.c
In file included from include/linux/bitops.h:17,
from include/linux/kernel.h:15,
from include/linux/sched.h:52,
from include/linux/utsname.h:35,
from /tmp/selfgz6594/NVIDIA-Linux-x86_64-173.14.12-pkg2-cus
tom/usr/src/nv/nv-linux.h:19,
from /tmp/selfgz6594/NVIDIA-Linux-x86_64-173.14.12-pkg2-cus
tom/usr/src/nv/nv.c:14:
include/asm/bitops.h: In function ‘set_bit’:
include/asm/bitops.h:60: warning: pointer of type ‘void *’ used in arith
metic
include/asm/bitops.h: In function ‘clear_bit’:
include/asm/bitops.h:97: warning: pointer of type ‘void *’ used in arith
metic
In file included from include/linux/utsname.h:35,
from /tmp/selfgz6594/NVIDIA-Linux-x86_64-173.14.12-pkg2-cus
tom/usr/src/nv/nv-linux.h:19,
from /tmp/selfgz6594/NVIDIA-Linux-x86_64-173.14.12-pkg2-cus
tom/usr/src/nv/nv.c:14:
include/linux/sched.h: In function ‘object_is_on_stack’:
include/linux/sched.h:1983: warning: pointer of type ‘void *’ used in ar
ithmetic
In file included from include/asm/dma-mapping.h:9,
from include/linux/dma-mapping.h:52,
from include/asm-generic/pci-dma-compat.h:7,
from include/asm/pci.h:94,
from include/linux/pci.h:983,
from /tmp/selfgz6594/NVIDIA-Linux-x86_64-173.14.12-pkg2-cus
tom/usr/src/nv/nv-linux.h:86,
from /tmp/selfgz6594/NVIDIA-Linux-x86_64-173.14.12-pkg2-cus
tom/usr/src/nv/nv.c:14:
include/linux/scatterlist.h: In function ‘sg_virt’:
include/linux/scatterlist.h:199: warning: pointer of type ‘void *’ used
in arithmetic
In file included from /tmp/selfgz6594/NVIDIA-Linux-x86_64-173.14.12-pkg2-cus
tom/usr/src/nv/nv.c:14:
/tmp/selfgz6594/NVIDIA-Linux-x86_64-173.14.12-pkg2-custom/usr/src/nv/nv-linu
x.h:107:27: error: asm/semaphore.h: No such file or directory
In file included from /tmp/selfgz6594/NVIDIA-Linux-x86_64-173.14.12-pkg2-cus
tom/usr/src/nv/nv-linux.h:109,
from /tmp/selfgz6594/NVIDIA-Linux-x86_64-173.14.12-pkg2-cus
tom/usr/src/nv/nv.c:14:
include/linux/highmem.h: In function ‘zero_user_segments’:
include/linux/highmem.h:134: warning: pointer of type ‘void *’ used in a
rithmetic
include/linux/highmem.h:137: warning: pointer of type ‘void *’ used in a
rithmetic
In file included from include/linux/compat.h:14,
from include/asm/mtrr.h:141,
from /tmp/selfgz6594/NVIDIA-Linux-x86_64-173.14.12-pkg2-cus
tom/usr/src/nv/nv-linux.h:121,
from /tmp/selfgz6594/NVIDIA-Linux-x86_64-173.14.12-pkg2-cus
tom/usr/src/nv/nv.c:14:
include/asm/compat.h: In function ‘compat_alloc_user_space’:
include/asm/compat.h:210: warning: pointer of type ‘void *’ used in arit
hmetic
In file included from /tmp/selfgz6594/NVIDIA-Linux-x86_64-173.14.12-pkg2-cus
tom/usr/src/nv/nv.c:14:
/tmp/selfgz6594/NVIDIA-Linux-x86_64-173.14.12-pkg2-custom/usr/src/nv/nv-linu
x.h: In function ‘nv_execute_on_all_cpus’:
/tmp/selfgz6594/NVIDIA-Linux-x86_64-173.14.12-pkg2-custom/usr/src/nv/nv-linu
x.h:674: error: too many arguments to function ‘on_each_cpu’
/tmp/selfgz6594/NVIDIA-Linux-x86_64-173.14.12-pkg2-custom/usr/src/nv/nv.c: I
n function ‘__nv_setup_pat_entries’:
/tmp/selfgz6594/NVIDIA-Linux-x86_64-173.14.12-pkg2-custom/usr/src/nv/nv.c:94
4: warning: comparison between signed and unsigned
/tmp/selfgz6594/NVIDIA-Linux-x86_64-173.14.12-pkg2-custom/usr/src/nv/nv.c: I
n function ‘__nv_restore_pat_entries’:
/tmp/selfgz6594/NVIDIA-Linux-x86_64-173.14.12-pkg2-custom/usr/src/nv/nv.c:97
0: warning: comparison between signed and unsigned
/tmp/selfgz6594/NVIDIA-Linux-x86_64-173.14.12-pkg2-custom/usr/src/nv/nv.c: I
n function ‘nv_kern_cpu_callback’:
/tmp/selfgz6594/NVIDIA-Linux-x86_64-173.14.12-pkg2-custom/usr/src/nv/nv.c:12
96: warning: comparison between signed and unsigned
/tmp/selfgz6594/NVIDIA-Linux-x86_64-173.14.12-pkg2-custom/usr/src/nv/nv.c:12
99: error: too many arguments to function ‘smp_call_function’
/tmp/selfgz6594/NVIDIA-Linux-x86_64-173.14.12-pkg2-custom/usr/src/nv/nv.c:13
03: warning: comparison between signed and unsigned
/tmp/selfgz6594/NVIDIA-Linux-x86_64-173.14.12-pkg2-custom/usr/src/nv/nv.c:13
06: error: too many arguments to function ‘smp_call_function’
make[4]: *** [/tmp/selfgz6594/NVIDIA-Linux-x86_64-173.14.12-pkg2-custom/usr/
src/nv/nv.o] Error 1
make[3]: *** [_module_/tmp/selfgz6594/NVIDIA-Linux-x86_64-173.14.12-pkg2-cus
tom/usr/src/nv] Error 2
make[2]: *** [sub-make] Error 2
NVIDIA: left KBUILD.
nvidia.ko failed to build!
make[1]: *** [module] Error 1
make: *** [module] Error 2
-> Error.
ERROR: Unable to build the NVIDIA kernel module.
ERROR: Installation has failed. Please see the file
'/var/log/nvidia-installer.log' for details. You may find suggestions
on fixing installation problems in the README available on the Linux
driver download page at www.nvidia.com.
Any ideas?

The patch only works on the 177 drivers , I haven't seen any patches for the 173 driver.

[jrummy27]

And just trying the legacy driver from livna fails when akmod tries to build it 23 Oct 17:03:06 akmodsbuild: The kernel you are installing for is a Xen kernel!
23 Oct 17:03:06 akmodsbuild:
23 Oct 17:03:06 akmodsbuild: The NVIDIA driver does not currently work on Xen kernels. If
23 Oct 17:03:06 akmodsbuild: you are using a stock distribution kernel, please install
23 Oct 17:03:06 akmodsbuild: a variant of this kernel without Xen support; if this is a
23 Oct 17:03:06 akmodsbuild: custom kernel, please install a standard Linux kernel. Then
23 Oct 17:03:06 akmodsbuild: try installing the NVIDIA kernel module again.
23 Oct 17:03:06 akmodsbuild:
23 Oct 17:03:06 akmodsbuild: *** Failed Xen sanity check. Bailing out! ***
23 Oct 17:03:06 akmodsbuild:
23 Oct 17:03:06 akmodsbuild: make[1]: *** [xen-sanity-check] Error 1
23 Oct 17:03:06 akmodsbuild: make[1]: *** Waiting for unfinished jobs....
23 Oct 17:03:06 akmodsbuild: make: *** [module] Error 2
23 Oct 17:03:06 akmodsbuild: error: Bad exit status from /var/tmp/rpm-tmp.LpIjZe (%build)
I am guessing this is because it was not built with the no-xen patch.

[jrummy27]

Woohoo! Got it working. Here's the procedure to get the 173.14.12 legacy driver going on F10: # su
# yum install wget
# wget http://us.download.nvidia.com/XFree86/Linux-x86_64/173.14.12/NVIDIA-Linux-x86_64-173.14.12-pkg2.run
# wget http://www.linux-ati-drivers.homecall.co.uk/patch-no-xen.patch
# wget http://kanotix.com/files/173.14.12/NVIDIA_173.14.12_2.6.27.patch
# sh NVIDIA-Linux-x86_64-173.14.12-pkg2.run --apply-patch patch-no-xen.patch
# sh NVIDIA-Linux-x86_64-173.14.12-pkg2-custom.run --apply-patch NVIDIA_173.14.12_2.6.27.patch
# su -
# init 3
# cd /home/*
# sh NVIDIA-Linux-x86_64-177.80-pkg2-custom.run -an The second patch came from the last post here: http://www.nvnews.net/vbulletin/showthread.php?t=118595

Whoops. Forgot to add the wget for the second patch

[Moncho]

For F9 64 bit

su
yum groupinstall "Development Tools"
yum install kernel-devel kernel-headers wget
wget http://us.download.nvidia.com/XFree86/Linux-x86_64/177.80/NVIDIA-Linux-x86_64-177.80-pkg2.run
su -
init 3
cd /home/*
sh NVIDIA-Linux-x86_64-177.80-pkg2.run -an


Or rebuild the SRPM from this post

http://forums.fedoraforum.org/showthread.php?t=201292

su
yum groupinstall "Development Tools"
yum install rpmdevtools wget
wget http://www.dfm.uninsubria.it/compiz/fusion-testing/9/nvidia/SRPM/nvidia-x11-drv-177.80-1.fc9.src.rpm
yum-buildep nvidia-x11-drv-177.80-1.fc9.src.rpm
exit
rpmdev-setuptree
rpmbuild --rebuild nvidia-x11-drv-177.80-1.fc9.src.rpm

Although I don't have Fedora 10 Beta, I've been having problems installing 177.80 on my Fedora 9 64 bit OS (GPU 8800 GTX). I tried both options above without success. The following is the output of /var/log/nvidia log:

Using: nvidia-installer ncurses user interface
-> Tagging shared libraries with chcon -t textrel_shlib_t.
-> License accepted by command line option.
-> Installing NVIDIA driver version 177.80.
-> Not probing for precompiled kernel interfaces.
-> Performing CC sanity check with CC="cc".
-> Performing CC version check with CC="cc".
ERROR: Unable to find the kernel source tree for the currently running kernel.
Please make sure you have installed the kernel source files for your
kernel and that they are properly configured; on Red Hat Linux systems,
for example, be sure you have the 'kernel-source' or 'kernel-devel' RPM
installed. If you know the correct kernel source files are installed,
you may specify the kernel source path with the '--kernel-source-path'
command line option.
ERROR: Installation has failed. Please see the file
'/var/log/nvidia-installer.log' for details. You may find suggestions
on fixing installation problems in the README available on the Linux
driver download page at www.nvidia.com.
[root@localhost log]#

Any help would be greatly appreciated.

Cheers...

[leigh123linux]

Although I don't have Fedora 10 Beta, I've been having problems installing 177.80 on my Fedora 9 64 bit OS (GPU 8800 GTX). I tried both options above without success. The following is the output of /var/log/nvidia log:

Using: nvidia-installer ncurses user interface
-> Tagging shared libraries with chcon -t textrel_shlib_t.
-> License accepted by command line option.
-> Installing NVIDIA driver version 177.80.
-> Not probing for precompiled kernel interfaces.
-> Performing CC sanity check with CC="cc".
-> Performing CC version check with CC="cc".
ERROR: Unable to find the kernel source tree for the currently running kernel.
Please make sure you have installed the kernel source files for your
kernel and that they are properly configured; on Red Hat Linux systems,
for example, be sure you have the 'kernel-source' or 'kernel-devel' RPM
installed. If you know the correct kernel source files are installed,
you may specify the kernel source path with the '--kernel-source-path'
command line option.
ERROR: Installation has failed. Please see the file
'/var/log/nvidia-installer.log' for details. You may find suggestions
on fixing installation problems in the README available on the Linux
driver download page at www.nvidia.com.
[root@localhost log]#

Any help would be greatly appreciated.

Cheers...

Post


rpm -qa |grep kernel

[Moncho]

Hi Leigh,

I went ahead and installed slightly older drivers (173.14), so I'm ok. However the following is the output to your request:

[root@localhost Backup]# rpm -qa |grep kernel
kerneloops-0.10-11.fc9.x86_64
kernel-headers-2.6.26.6-79.fc9.x86_64
kernel-devel-2.6.26.6-79.fc9.x86_64
kernel-2.6.26.6-79.fc9.x86_64
kernel-2.6.25-14.fc9.x86_64
[root@localhost Backup]#

Cheers...

[leigh123linux]

Post #1 edited .

[net_boy]

Total crash for me after following this from a completely fresh install of the preview..

My machine boots to a blank screen and no response. It's obv a major crash as there is nothing on the keyboard, even the caps light won't work (no interrupts ?)

Any help before I re install again...

Cheers

Steve

[popatopalous]

1. Install the release key


su
rpm -Uvh http://download1.rpmfusion.org/nonfree/fedora/development/i386/os/rpmfusion-nonfree-release-9.90-5.noarch.rpm



2. Install the driver

su
yum install kmod-nvidia xorg-x11-drv-nvidia-libs xorg-x11-drv-nvidia



Or if you like playing with unreleased Koji kernels

su
yum install akmod-nvidia xorg-x11-drv-nvidia-libs xorg-x11-drv-nvidia



64 bit users will also need to install the 32 bit libs for apps like Wine.

su
yum install xorg-x11-drv-nvidia-libs.i386


3, Reboot.

For x86_64 this is installing for kernel 2.6.27.4-68.fc10.x86_64. However the latest kernel is 2.6.27.4-79.fc10.x86_64. What's up?

yum install kmod-nvidia-2.6.27.4-79.fc10.x86_64
Loaded plugins: refresh-packagekit
Setting up Install Process
Parsing package install arguments
No package kmod-nvidia-2.6.27.4-79.fc10.x86_64 available.
Nothing to do


?

[leigh123linux]

For x86_64 this is installing for kernel 2.6.27.4-68.fc10.x86_64. However the latest kernel is 2.6.27.4-79.fc10.x86_64. What's up?

yum install kmod-nvidia-2.6.27.4-79.fc10.x86_64
Loaded plugins: refresh-packagekit
Setting up Install Process
Parsing package install arguments
No package kmod-nvidia-2.6.27.4-79.fc10.x86_64 available.
Nothing to do


?

Try using the akmod-nvidia driver.

su
yum install akmod-nvidia xorg-x11-drv-nvidia-libs xorg-x11-drv-nvidia

[leigh123linux]

Total crash for me after following this from a completely fresh install of the preview..

My machine boots to a blank screen and no response. It's obv a major crash as there is nothing on the keyboard, even the caps light won't work (no interrupts ?)

Any help before I re install again...

Cheers

Steve


Did you update first ?

[net_boy]

Hi Leigh

I'm really new to Linux, so apologies for the newbie'ness but no I didn't update.

I take it I need to yum update before following your first post ?

Many thanks for the help..

Best

Steve

[leigh123linux]

Hi Leigh

I'm really new to Linux, so apologies for the newbie'ness but no I didn't update.

I take it I need to yum update before following your first post ?

Many thanks for the help..

Best

Steve

It's best to update first.


Try using this command to install the driver

su
yum install akmod-nvidia xorg-x11-drv-nvidia-libs xorg-x11-drv-nvidia

[net_boy]

Thanks Leigh, much appreciated.

Steve

[popatopalous]

Try using the akmod-nvidia driver.

su
yum install akmod-nvidia xorg-x11-drv-nvidia-libs xorg-x11-drv-nvidia

Thanks leigh123. That worked. In F-10 x86_64 with latest kernel nvidia is working with akmod-nvidia.

[net_boy]

Hi Leigh,

I seem to bet getting somewhere now, however when Fedora now boots into X and I try to access the nvidia driver i get this :

http://img89.imageshack.us/img89/3826/nvidiaissuegp7.th.jpg (http://img89.imageshack.us/my.php?image=nvidiaissuegp7.jpg)http://img89.imageshack.us/images/thpix.gif (http://g.imageshack.us/thpix.php)

the output of the xorg.conf file is this :

# Xorg configuration created by livna-config-display

Section "Files"
ModulePath "/usr/lib64/xorg/modules"
EndSection

Section "ServerFlags"
Option "AIGLX" "on"
EndSection

Section "Device"
Identifier "Videocard0"
Driver "vesa"
EndSection

Section "Extensions"
Option "Composite" "Enable"
EndSection


now obv the card doesn't seem right in the device section, however when I run xorg-conf as root in a terminal window it says it completes ok, after a reboot the file reverts to this...

Any ideas ?

Many thanks

Steve

[net_boy]

Help....:(

Anybody.....

[leigh123linux]

Help....:(

Anybody.....


Post


rpm -qa |grep nvidia

[a_small_cake]

You can manually change "vesa" to "nvidia".

[net_boy]

Post


rpm -qa |grep nvidia

Hi Leigh, here it is...

[nelsons@homer ~]$ rpm -qa |grep nvidia
kmod-nvidia-2.6.27.4-68.fc10.x86_64-177.80-1.fc10.3.x86_64
kmod-nvidia-177.80-1.fc10.3.x86_64
xorg-x11-drv-nvidia-libs-177.80-1.fc10.x86_64
xorg-x11-drv-nvidia-libs-177.80-1.fc10.i386
xorg-x11-drv-nvidia-177.80-1.fc10.x86_64
[nelsons@homer ~]$

[leigh123linux]

Hi Leigh, here it is...

[nelsons@homer ~]$ rpm -qa |grep nvidia
kmod-nvidia-2.6.27.4-68.fc10.x86_64-177.80-1.fc10.3.x86_64
kmod-nvidia-177.80-1.fc10.3.x86_64
xorg-x11-drv-nvidia-libs-177.80-1.fc10.x86_64
xorg-x11-drv-nvidia-libs-177.80-1.fc10.i386
xorg-x11-drv-nvidia-177.80-1.fc10.x86_64
[nelsons@homer ~]$


Run system-config-display ( select the correct monitor and card settings ) to create a xorg.conf and nvidia-xconfig to finish it

su
yum install system-config-display
exit
system-config-display
nvidia-xconfig

[chepioq]

@net_boy
what is your current kernel?
uname -r

[net_boy]

Run system-config-display ( select the correct monitor and card settings ) to create a xorg.conf and nvidia-xconfig to finish it

su
yum install system-config-display
exit
system-config-display
nvidia-xconfig

Hi Leigh,

Nope no joy, the system config display works but the only reference I can see to the card is the "nv" driver. The card is a 9800 GT

When I ran the commands and restarted x, I was greeted by a black screen and X would not start.

I did a reboot and the box came back and started X but the xorg.conf file had been changed to use the vesa driver.

I think I am close, but just not there, do you think it's possible that there is no support my card in this release of drivers ??

Ta

Steve

[hiberphoptik]

don't forget, it's very possible you will need to add your PCI busID to your xorg.conf if you are getting the fast flicker and black screen problem after installing the drivers.

a bit off topic but...so far F10 preview has been good to me aside from the GIANT login box in GDM when running multiple xinerama screen setups...

instead of having a fixed size the login box is a % of the screensize so it is HUUUUUUUUUUGE with 4 monitors

[net_boy]

@net_boy
what is your current kernel?
uname -r

chepioq.

Here it is :

[nelsons@homer ~]$ uname -r
2.6.27.4-79.fc10.x86_64
[nelsons@homer ~]$

[leigh123linux]

Hi Leigh,

Nope no joy, the system config display works but the only reference I can see to the card is the "nv" driver. The card is a 9800 GT

When I ran the commands and restarted x, I was greeted by a black screen and X would not start.

I did a reboot and the box came back and started X but the xorg.conf file had been changed to use the vesa driver.

I think I am close, but just not there, do you think it's possible that there is no support my card in this release of drivers ??

Ta

Steve

Have a look a posts #29 & 30

http://forums.fedoraforum.org/showthread.php?t=203553&page=2


su -
nvidia-xconfig --enable-all-gpus

[leigh123linux]

Hi Leigh, here it is...

[nelsons@homer ~]$ rpm -qa |grep nvidia
kmod-nvidia-2.6.27.4-68.fc10.x86_64-177.80-1.fc10.3.x86_64
kmod-nvidia-177.80-1.fc10.3.x86_64
xorg-x11-drv-nvidia-libs-177.80-1.fc10.x86_64
xorg-x11-drv-nvidia-libs-177.80-1.fc10.i386
xorg-x11-drv-nvidia-177.80-1.fc10.x86_64
[nelsons@homer ~]$


You module doesn't match your kernel

Try

su
yum erase *nvidia*
yum install akmod-nvidia xorg-x11-drv-nvidia-libs xorg-x11-drv-nvidia

Reboot

[net_boy]

Thanks Leigh,

I'll give it a whirl..

[net_boy]

OK after the last install I rebooted and got the black screen again, no command line, nothing. No caps lock light, another total crash...

So I reinstalled again, went through the process, applied a yum update and followed the post number 1 again, except this time I used Leigh's recommendation of :

yum install akmod-nvidia xorg-x11-drv-nvidia-libs xorg-x11-drv-nvidia

and then reboot....

Again the black screen and the box will not boot at all, I am gutted, looks like I am staring down the barrel of another rebuild....

I know I am new to this, but surely this can't be THAT difficult. I work in telecoms for a living and hold a CCIE so I am used to technology, but this.... Argh!!!!!

Help me !!

Thanks

Anyone help out there...?

I am thinking it may be wise to not install the nvidia card drivers until after the final release of FC10, then again I have a nice all singing all dancing video card that is just waiting for compiz..

Cheers

Steve

[sj200449]

kmod-nvidia-2.6.27.4-79.fc10.x86_64.x86_64 (note the "79") is available now , it wasnt yesterday and my kernel "79" wasnt supported so reverted back to 68 kernel.

Works fine. RPM Fusion stuff.

gts880 btw

[popatopalous]

Hi Leigh,

I seem to bet getting somewhere now, however when Fedora now boots into X and I try to access the nvidia driver i get this :

http://img89.imageshack.us/img89/3826/nvidiaissuegp7.th.jpg (http://img89.imageshack.us/my.php?image=nvidiaissuegp7.jpg)http://img89.imageshack.us/images/thpix.gif (http://g.imageshack.us/thpix.php)

the output of the xorg.conf file is this :

# Xorg configuration created by livna-config-display

Section "Files"
ModulePath "/usr/lib64/xorg/modules"
EndSection

Section "ServerFlags"
Option "AIGLX" "on"
EndSection

Section "Device"
Identifier "Videocard0"
Driver "vesa"
EndSection

Section "Extensions"
Option "Composite" "Enable"
EndSection


now obv the card doesn't seem right in the device section, however when I run xorg-conf as root in a terminal window it says it completes ok, after a reboot the file reverts to this...

Any ideas ?

Many thanks

Steve

Be sure that kernel you are booting in to matches your kmod-nvidia or akmod nvidia kernel. To determine which kernel you are booted into from Konsole [in KDE] or terminial [Gnome]:

# uname -r
2.6.27.4-79.fc10.x86_64

Then you could simply open Yumex and search whichever you have installed kmod or akmod and see if you have [this is in my case]:

kmod-nvidia-2.6.27.4-79.fc10.x86_64

which does match and is what I need. If not then you need to reboot into correct kernel for your nvidia install. If you are already in the correct kernel try running in Konsole or terminial as root [su -]:

# nvidia-xconfig

Using X configuration file: "/etc/X11/xorg.conf".
Backed up file '/etc/X11/xorg.conf' as '/etc/X11/xorg.conf.backup'
New X configuration file written to '/etc/X11/xorg.conf'
# modprobe nvidia


Then either logout and log back in or reboot and see if nvidia driver is working. Hope this helps.:D

[net_boy]

kmod-nvidia-2.6.27.4-79.fc10.x86_64.x86_64 (note the "79") is available now , it wasnt yesterday and my kernel "79" wasnt supported so reverted back to 68 kernel.

Works fine. RPM Fusion stuff.

gts880 btw

All present and correct, still no luck.... :(
Well, I tried again but still no luck.

I am using this kernel 2.6.27.4-79.fc10.x86_64 after the last yum update.

I again followed the steps in post 1 but checked at each stage that that the kmod matched the kernel, which it did.

Bonus I thought!

So I went for a reboot, still nothing, again a major kernel issue with black screen and no response at all from the box...

Still I learned that I didn't need a rebuild and that I could access a boot menu by pressing escape just as the boot process was starting, so at least no rebuild was necessary...

So I am now booting into 2.6.27.4-68.fc10.x86_64 and I have the same issue as I posted in post 93

I also tried the nvidia-xconfig --enable-all-gpus after I did a nvidia-xconfig

Nothing....

I am at my wits end here, does anyone think there is any mileage in me calling nvidia ?

HELP !! PLEASE...

[net_boy]

Just some more info I thought may be of use.

The card is in a PCI-X 16bit slot

Here is the output showing the correct mods, note the incorrect kernel as I have had to revert due to the system hanging when booting from 2.6.27.4-79.fc10.x86_64

http://img526.imageshack.us/img526/5854/nvidiaco6.th.jpg (http://img526.imageshack.us/my.php?image=nvidiaco6.jpg)

[net_boy]

# nvidia-xconfig

Using X configuration file: "/etc/X11/xorg.conf".
Backed up file '/etc/X11/xorg.conf' as '/etc/X11/xorg.conf.backup'
New X configuration file written to '/etc/X11/xorg.conf'
# modprobe nvidia


Then either logout and log back in or reboot and see if nvidia driver is working. Hope this helps.:D[/QUOTE]

Another update.

I am not sure what modprobe does, but whatever it is it does not look healthy !!

[root@homer nelsons]# nvidia-xconfig

Using X configuration file: "/etc/X11/xorg.conf".
Backed up file '/etc/X11/xorg.conf' as '/etc/X11/xorg.conf.backup'
New X configuration file written to '/etc/X11/xorg.conf'

[root@homer nelsons]# modprobe nvidia
FATAL: Module nvidia not found.
[root@homer nelsons]#
[root@homer nelsons]#

[net_boy]

No replies, I am guessing that this either has everyone stumped or I have some faulty hardware !!

[leigh123linux]

Post

rpm -qa |grep nvidia

and

uname -r

[net_boy]

Unfortunately not Leigh.

I removed the Nvidia drivers using :

yum erase *nvidia*

I did this because if I didn't press escape at bootup and select 2.6.27.4-68.fc10.x86_64 as the kernel then the machine would try and boot into 2.6.27.4-79.fc10.x86_64 as that was installed as part of of the yum update.

Seeing as I didn't want to sit and press escape in order to get my box to boot I uninstalled the nvidia drivers, it now boots straight into 2.6.27.4-79.fc10.x86_64 with no issues except it is using the vesa driver.

I am willing to give it another shot as I realise now that I dont have to rebuild after every kernel crash..

Do you have any pearls of wisdom as to what may be the next steps ?

Thanks

Steve

[hephasteus]

HI net_boy. If you want to use the nvidia just reload it. Go to system -> administration-> bootloader and set it to a kernel that you have the nvidia driver for then keep a close eye on your updates. When an nivdia driver pops up that supports the newer kernel you can then change your bootloader to the new kernel and remove the old nvidia package so that if you decide to remove the old kernel it won't give a dependency problem. I had to use the kernel before 79 for about 4 days before nvidia upped to a 79 kernel driver. Now I have the new 94 kernel but i'm bootloading on 79 until nvidia updates driver which shouldn't take too long. I bet theres a 94 nvidia driver update before weekend.

[net_boy]

Hi hephasteus

The problem is that even with the correct drivers, mod etc and a matching kernel my box still spectacularly crashes, at least it does with 2.6.27.4-79.fc10.x86_64

So I actually can't get it to work at all !!

Thanks for the tips though, it makes sense to me what you have said and how it works

All the best

Steve

[DrWilken]

HI net_boy. If you want to use the nvidia just reload it. Go to system -> administration-> bootloader and set it to a kernel that you have the nvidia driver for then keep a close eye on your updates. When an nivdia driver pops up that supports the newer kernel you can then change your bootloader to the new kernel and remove the old nvidia package so that if you decide to remove the old kernel it won't give a dependency problem. I had to use the kernel before 79 for about 4 days before nvidia upped to a 79 kernel driver. Now I have the new 94 kernel but i'm bootloading on 79 until nvidia updates driver which shouldn't take too long. I bet theres a 94 nvidia driver update before weekend.

Just did a yum update:
[snip]
nvidia (177.80-1.fc10): Installing module.
.............
......

Installed:
kernel.x86_64 0:2.6.27.5-94.fc10 kernel-devel.x86_64 0:2.6.27.5-94.fc10
[snip]

[net_boy]

Just did a yum update:
[snip]
nvidia (177.80-1.fc10): Installing module.
.............
......

Installed:
kernel.x86_64 0:2.6.27.5-94.fc10 kernel-devel.x86_64 0:2.6.27.5-94.fc10
[snip]

hmmmm, I must be missing something here, I again installed the nvidia drivers and thought I would be able to yum update to the 27.5-94 kernel as you have above but......

[root@homer nelsons]# rpm -qa |grep nvidia
kmod-nvidia-2.6.27.4-79.fc10.x86_64-177.80-1.fc10.4.x86_64
xorg-x11-drv-nvidia-177.80-1.fc10.x86_64
xorg-x11-drv-nvidia-libs-177.80-1.fc10.x86_64
kmod-nvidia-177.80-1.fc10.4.x86_64
[root@homer nelsons]#
[root@homer nelsons]#
[root@homer nelsons]# yum update
Loaded plugins: refresh-packagekit
Setting up Update Process
No Packages marked for Update
[root@homer nelsons]#

[DrWilken]

hmmmm, I must be missing something here, I again installed the nvidia drivers and thought I would be able to yum update to the 27.5-94 kernel as you have above but......

[root@homer nelsons]# rpm -qa |grep nvidia
kmod-nvidia-2.6.27.4-79.fc10.x86_64-177.80-1.fc10.4.x86_64
xorg-x11-drv-nvidia-177.80-1.fc10.x86_64
xorg-x11-drv-nvidia-libs-177.80-1.fc10.x86_64
kmod-nvidia-177.80-1.fc10.4.x86_64
[root@homer nelsons]#
[root@homer nelsons]#
[root@homer nelsons]# yum update
Loaded plugins: refresh-packagekit
Setting up Update Process
No Packages marked for Update
[root@homer nelsons]#

Hmmm... Weird...

I still can't enable compiz-fusion through "Desktop Effects" though...

[DrWilken]

Do You have the new(?) nvidia repo:
[root@xpc ~]# yum repolist
Loaded plugins: refresh-packagekit
repo id repo name status
fedora Fedora 10 - x86_64 enabled : 14,304
nvidia nvidia repository enabled : 2
updates Fedora 10 - x86_64 - Updates enabled : 14,304
repolist: 28,610

[net_boy]

DrWilken

It would appear not, how do I get it ?

[root@homer nelsons]# yum repolist
Loaded plugins: refresh-packagekit
repo id repo name status
adobe-linux-i386 Adobe Systems Incorporated enabled : 17
fedora Fedora 10 - x86_64 enabled : 14,304
rpmfusion-free-rawhi RPM Fusion for Fedora Rawhide - Free enabled : 417
rpmfusion-nonfree-ra RPM Fusion for Fedora Rawhide - Nonfree enabled : 154
updates Fedora 10 - x86_64 - Updates enabled : 14,304
repolist: 29,196
[root@homer nelsons]#



Ta

Steve

[DrWilken]

I noticed that I had created a xorg.conf so I tried removing it and restarting X (ctrl-alt-backspace)...

When I try to enable compiz-fusion through "Desktop Effect" I get a white screen and I have to restart X again...

I will have a look at /var/log/Xorg.0.log and see what's going on...

[DrWilken]

DrWilken

It would appear not, how do I get it ?

[root@homer nelsons]# yum repolist
Loaded plugins: refresh-packagekit
repo id repo name status
adobe-linux-i386 Adobe Systems Incorporated enabled : 17
fedora Fedora 10 - x86_64 enabled : 14,304
rpmfusion-free-rawhi RPM Fusion for Fedora Rawhide - Free enabled : 417
rpmfusion-nonfree-ra RPM Fusion for Fedora Rawhide - Nonfree enabled : 154
updates Fedora 10 - x86_64 - Updates enabled : 14,304
repolist: 29,196
[root@homer nelsons]#



Ta

Steve

I'm new to fedora (and yum) so I really don't know how to do it via the yum command but You can try selecting it in Gnome "System"->"Administration"->"Add/Remove Software"...

Under "System"->"Software Sources"...

You might have to use the "Refresh Package Lists" first... (?) ;)

EDIT:
maybe this does the trick (?):
yum --enablerepo=nvidia

[DrWilken]

After looking at /var/log/Xorg.0.log I can see that I'm NOT using the nvidia module but rather the nv module so apparently the autoconfiguration of X doesn't detect the nvidia module... I think I'll have to create an xorg.conf file again...

[hephasteus]

hmmmm, I must be missing something here, I again installed the nvidia drivers and thought I would be able to yum update to the 27.5-94 kernel as you have above but......

[root@homer nelsons]# rpm -qa |grep nvidia
kmod-nvidia-2.6.27.4-79.fc10.x86_64-177.80-1.fc10.4.x86_64
xorg-x11-drv-nvidia-177.80-1.fc10.x86_64
xorg-x11-drv-nvidia-libs-177.80-1.fc10.x86_64
kmod-nvidia-177.80-1.fc10.4.x86_64
[root@homer nelsons]#
[root@homer nelsons]#
[root@homer nelsons]# yum update
Loaded plugins: refresh-packagekit
Setting up Update Process
No Packages marked for Update
[root@homer nelsons]#

See that kmod-nvidia-2.6.27.4-79 means that the video driver is custom to the 79 kernel and will only work on the 79 kernel. The latest kernel is a 94 kernel so you have to set the system to boot the 79 kernel if you keep that one. When you get an update that says something like this..
kmod-nvidia-2.6.27.4-94.fc10x86_64-177.80-1.fc10.4x86_64
Then you can switch the system over to boot the 94 kernel. Everything before .fc10 on a kmod-nvidia driver means the kernel it's designed to work wtih. Everything after x86_64- is the driver version. You're using 177.80 driver version.
That supports everything from geforce 6 series and they've added support for some new ones.
# Added support for the following new GPUs:

* GeForce GTX 260
* GeForce GTX 280
* GeForce 9800 GTX+
* GeForce 9800 GT
* GeForce 9700M GTS
* GeForce 9500 GT
* GeForce 8100P
* nForce 780a SLI
* nForce 750a SLI
* Quadro FX 770M
* Quadro NVS 160M
* Quadro NVS 150M

So unless you're video card is really old or something stolen out the nvidia R&D lab i can't see why it would black screen. Using an old driver on a new kernel will usually black screen or give you some sort of untenable x window desktop. (wrong screen resolution and can't be changed, windows overwriting each other, mouse cursors ghosting, etc.) There's usually not that many kernels with each release but I think all the wifi and sound card action in fedora lately is making new kernels more common until all that gets sorted out. There will probably one more at release and then one more a month or two down the line from release so don't forget how to set it till ya get new video drivers.

[net_boy]

Good try but nope.... :)

[root@homer nelsons]# yum --enablerepo=nvidia
Loaded plugins: refresh-packagekit


Error getting repository data for nvidia, repository not found
[root@homer nelsons]#

[net_boy]

See that kmod-nvidia-2.6.27.4-79 means that the video driver is custom to the 79 kernel and will only work on the 79 kernel. The latest kernel is a 94 kernel so you have to set the system to boot the 79 kernel if you keep that one. When you get an update that says something like this..
kmod-nvidia-2.6.27.4-94.fc10x86_64-177.80-1.fc10.4x86_64
Then you can switch the system over to boot the 94 kernel. Everything before .fc10 on a kmod-nvidia driver means the kernel it's designed to work wtih. Everything after x86_64- is the driver version. You're using 177.80 driver version.
That supports everything from geforce 6 series and they've added support for some new ones.
# Added support for the following new GPUs:

* GeForce GTX 260
* GeForce GTX 280
* GeForce 9800 GTX+
* GeForce 9800 GT
* GeForce 9700M GTS
* GeForce 9500 GT
* GeForce 8100P
* nForce 780a SLI
* nForce 750a SLI
* Quadro FX 770M
* Quadro NVS 160M
* Quadro NVS 150M

So unless you're video card is really old or something stolen out the nvidia R&D lab i can't see why it would black screen. Using an old driver on a new kernel will usually black screen or give you some sort of untenable x window desktop. (wrong screen resolution and can't be changed, windows overwriting each other, mouse cursors ghosting, etc.) There's usually not that many kernels with each release but I think all the wifi and sound card action in fedora lately is making new kernels more common until all that gets sorted out. There will probably one more at release and then one more a month or two down the line from release so don't forget how to set it till ya get new video drivers.

I understand all that, thanks.

However I have a brand new GeForce 9800 GT which was bought about 10 days ago, it's an official nvidia card too not one of the OEM ones.

I am not joking when I say that even when all the mods etc match with the kernel then I get a black screen and the whole machine dies, infact when I rebooted into the previous kernel I got a message in Gnome at the bottom right of the sreen saying my box had suffered a kernel failure and did I want to send it to the developers, which of course I said yes.

That said I am still here, no card driver works for me, no 3d acceleration. It sucks.....

cheers

Steve

[IBBoard]

What kind of performance are people getting out of the 177.80 drivers on the 2.6.27.5-94 kernel?

I've installed them from the RPM Fusion repos, they've been enabled, direct rendering reports okay, but comparing Compiz benchmarks my 256MB 7950GT at home seriously beats the 512MB 9800GTX+ at work at ~350fps versus ~55fps (both unlimited). It's not too slow with Compiz to be particularly noticeable, but it doesn't quite seem right that the newer, faster, higher memory card should be doing so much worse. The 7950 is on F9 rather than F10 and is still on the .26 kernel.

[net_boy]

Can someone tell how I can get nvidia repos installed please, I would like to see if this driver also causes my machine to crash with a black screen.

I can't get the update with the standard yum update as the repos are not there :

[root@homer nelsons]# yum repolist
Loaded plugins: refresh-packagekit
repo id repo name status
adobe-linux-i386 Adobe Systems Incorporated enabled : 17
fedora Fedora 10 - x86_64 enabled : 14,304
rpmfusion-free-rawhi RPM Fusion for Fedora Rawhide - Free enabled : 417
rpmfusion-nonfree-ra RPM Fusion for Fedora Rawhide - Nonfree enabled : 154
updates Fedora 10 - x86_64 - Updates enabled : 14,304
repolist: 29,196
[root@homer nelsons]#

Thanks

Steve

[DrWilken]

After looking at /var/log/Xorg.0.log I can see that I'm NOT using the nvidia module but rather the nv module so apparently the autoconfiguration of X doesn't detect the nvidia module... I think I'll have to create an xorg.conf file again...

I got it working with the following setup:
[drwilken@xpc ~]$ rpm -qa | grep nvidia
kmod-nvidia-2.6.27.4-79.fc10.x86_64-177.80-1.fc10.4.x86_64
kmod-nvidia-177.80-1.fc10.4.x86_64
xorg-x11-drv-nvidia-libs-177.80-1.fc10.x86_64
xorg-x11-drv-nvidia-177.80-1.fc10.x86_64

I removed the nvidia packages (via the graphical installer :eek:) and installed the kmod-nvidia metapackage with dependencies (the libs, driver etc.)...

Booting with the 79 kernel works and I can now enable compiz-fusion... :D

I have to use the 79 kernel until there's a kmod-nvidia package for the 94 kernel as stated by hephasteus...

[net_boy]

I was asked if I had the nvidia repo in my list as per below :


Do You have the new(?) nvidia repo:
[root@xpc ~]# yum repolist
Loaded plugins: refresh-packagekit
repo id repo name status
fedora Fedora 10 - x86_64 enabled : 14,304
nvidia nvidia repository enabled : 2
updates Fedora 10 - x86_64 - Updates enabled : 14,304
repolist: 28,610

I don't....

Can someone tell me how to add it please ?

Thanks

Steve

[leigh123linux]

What kind of performance are people getting out of the 177.80 drivers on the 2.6.27.5-94 kernel?

I've installed them from the RPM Fusion repos, they've been enabled, direct rendering reports okay, but comparing Compiz benchmarks my 256MB 7950GT at home seriously beats the 512MB 9800GTX+ at work at ~350fps versus ~55fps (both unlimited). It's not too slow with Compiz to be particularly noticeable, but it doesn't quite seem right that the newer, faster, higher memory card should be doing so much worse. The 7950 is on F9 rather than F10 and is still on the .26 kernel.


This might explain the poor performance .

http://forums.fedoraforum.org/showpost.php?p=1108292&postcount=103

[IBBoard]

net_boy: I'm using the RPM Fusion repositories that Leigh mentions on the front page. From reading bits of this thread a couple of days ago I think the nvidia repository was one Leigh set up for a while, but his instructions have changed now to the RPM Fusion.

DrWilken: I'm using 2.6.27.5-94 without a problem from RPM Fusion. There's been a 64-bit KMod since at least this morning and the AKMod should recompile itself anyway if you have everything you need installed.

[Edit] Leigh, I'll take a look at that. As I said, the card works fine overall (so I wouldn't have thought of looking in those threads/posts), it's just a bit slow. I'll report back after I've tried it out.

[net_boy]

IBBoard

Thanks for that, now i have the nvidia repos installed...

However can I just say :mad::mad:ARRRRRGHHHHH!!!!!!!!! :mad::mad:

See below

[root@homer nelsons]# yum repolist
Loaded plugins: refresh-packagekit
repo id repo name status
adobe-linux-i386 Adobe Systems Incorporated enabled : 17
fedora Fedora 10 - x86_64 enabled : 14,304
nvidia nvidia repository enabled : 2
rpmfusion-free-rawhi RPM Fusion for Fedora Rawhide - Free enabled : 417
rpmfusion-nonfree-ra RPM Fusion for Fedora Rawhide - Nonfree enabled : 154
updates Fedora 10 - x86_64 - Updates enabled : 14,304
repolist: 29,198
[root@homer nelsons]#
[root@homer nelsons]# yum update
Loaded plugins: refresh-packagekit
Setting up Update Process
No Packages marked for Update
[root@homer nelsons]#

no updates !!!!!!!!!!!!!

So where did I put that length of rope.......:(

Best

Steve

[IBBoard]

If you've installed from the nVidia distributed file then you won't see any updates as the RPM Fusion is built from it but uses a different format with some extra features (such as re-building as an RPM). Try running "yum install akmod-nvidia" and it should pull in the dependencies.

Leigh: So far it isn't looking promising. The enable didn't like my XOrg.conf (apparently I was missing a monitor, but it has been working fine up until now) and now GLX is dead and Compiz won't work! I'll see what I can work out.

[edit] Oh, and it appears there isn't a kmod build for the latest kernel yet (although I'm sure I got an nVidia update this morning) but the akmod works with the newest kernel as long as you've got all of the dependencies.

[edit2] Okay, GLX problem solved (I ran "nvidia-config-display disable" and "nvidia-config-display enable" and it added in some module paths), but frame rates are still ~60fps.

[ruien]

You can just download 177.80 from nvidia.com and install kernel-devel, gcc ++ . Then kill x or boot into textmode and run the Nvidia installer. Works both on 2.6.27.5-94.i686 and PAE kernel wich I run atm.

[net_boy]

Nope, no good... can't compile the kernel....

[root@homer nelsons]# yum install kernel-devel
Loaded plugins: refresh-packagekit
Setting up Install Process
Parsing package install arguments
Package kernel-devel-2.6.27.5-94.fc10.x86_64 already installed and latest version
Nothing to do
[root@homer nelsons]# yum install gcc
Loaded plugins: refresh-packagekit
Setting up Install Process
Parsing package install arguments
Package gcc-4.3.2-7.x86_64 already installed and latest version
Nothing to do
[root@homer nelsons]#


Back to the drawing board

[cloneu2]

You also need kernel-headers to compile the kernel for the nvidia driver. It works just fine here with
[root@incognito Documents]# uname -r
2.6.27.5-101.fc10.x86_64
[root@incognito Documents]#

[DrWilken]

DrWilken: I'm using 2.6.27.5-94 without a problem from RPM Fusion. There's been a 64-bit KMod since at least this morning and the AKMod should recompile itself anyway if you have everything you need installed.


OK, I'll try it out... ;) Thx for the info...

[DrWilken]

[edit] Oh, and it appears there isn't a kmod build for the latest kernel yet (although I'm sure I got an nVidia update this morning) but the akmod works with the newest kernel as long as you've got all of the dependencies.

Works here with akmod on the 2.6.27.5-94.fc10.x86_64 kernel... ;)

Just did:

yum install akmod-nvidia

nvidia-display-config enable

reboot :D

[net_boy]

Same result here for every method, black screen, crash and reboot into old kernel / rebuild in some cases necessary.

I am beginning to think it's a hardware issue, I searched online and someone had the same issues as me with the same manufacturer of motherboard (not the same model) and a bios update fixed it for him

So I'll go and search for a bios update for the board and see how that goes.... BTW its a gigabyte board, the latest one with the 4 LAN ports on and a good few SATA / USB etc etc

Wish me luck ?

Steve

[ruien]

You are booting the .94 kernel same as kernel-dev? What error do you get trying to run the nvidia installer when it fails? If you installed gcc you should have the kernel headers to. This should be pretty easy to do if you have installed kernel-dev and compilers for your running kernel.

[net_boy]

You are booting the .94 kernel same as kernel-dev? What error do you get trying to run the nvidia installer when it fails? If you installed gcc you should have the kernel headers to. This should be pretty easy to do if you have installed kernel-dev and compilers for your running kernel.

Yes it was very easy, I had forgotten to install make after I rebuilt my box, that said once the Nvidia app said all was OK, I rebooted and the box had the black screen no response crash. When this happens I have to powercycle to get the box back in either A. Another kernel or B. Do a complete rebuild as the older kernel on the box also give the same effect.

It's very very frustrating.

[hephasteus]

Yes it was very easy, I had forgotten to install make after I rebuilt my box, that said once the Nvidia app said all was OK, I rebooted and the box had the black screen no response crash. When this happens I have to powercycle to get the box back in either A. Another kernel or B. Do a complete rebuild as the older kernel on the box also give the same effect.

It's very very frustrating.

Are you running this on Linux only or do you have windows? I'd get it working under windows to make sure the hardware is working right before trying linux. I also just want to say that it might be a good idea take it back and pick up a much much cheaper 8000 or 7000 card. There's new technology to keep chips from leaking electrons all over the place. To put it bluntly you bought an expensive dinosaur.

http://www.intel.com/technology/architecture-silicon/45nm-core2/?cid=cim:ggl|penryn_us_brand|kEB59|s

If you want to learn a bit more about your card look here.
http://www.overclock.net/hardware-news/312289-it168-9800gt-review.html

[JoshuaK]

With nVidia from RPM Fusion it will not work with 94 kernel... only the (?) 70 i think.:mad:

[net_boy]

Are you running this on Linux only or do you have windows? I'd get it working under windows to make sure the hardware is working right before trying linux. I also just want to say that it might be a good idea take it back and pick up a much much cheaper 8000 or 7000 card. There's new technology to keep chips from leaking electrons all over the place. To put it bluntly you bought an expensive dinosaur.

http://www.intel.com/technology/architecture-silicon/45nm-core2/?cid=cim:ggl|penryn_us_brand|kEB59|s

If you want to learn a bit more about your card look here.
http://www.overclock.net/hardware-news/312289-it168-9800gt-review.html

Thanks for that, to be honest I don't really need all the whistles and bells as I am only running Linux for the performance of Dynagen, but this has become a sort of nemesis and I won't let it drop until its done.... Yeah Yeah I must be mad !! I think I am !!:D

As for your question, I have only tried on Linux at the moment but tomorrow I'll give it a whirl on Windows, hopefully the hardware is OK.

I also have the latest bios for the motherboard, I have found out I am on Rev 5.0 and the latest is v12.0 so I am going to go with that too...

I'll report back, if you don't hear from me assume I have given up and gone to live as a monk in the hills somewhere !!

Best

Steve

[Demz]

leigh

are you gonna bother building the 177.82 driver or leave it for RPMFusion now? http://www.nvidia.com/object/unix.html

[leigh123linux]

With nVidia from RPM Fusion it will not work with 94 kernel... only the (?) 70 i think.:mad:


Balderdash :p


[leigh@localhost ~]$ rpm -qa |grep nvidia
xorg-x11-drv-nvidia-libs-177.80-1.fc10.i386
akmod-nvidia-177.80-1.fc10.i686
kmod-nvidia-2.6.27.5-94.fc10.i686-177.80-1.fc10.i686
xorg-x11-drv-nvidia-177.80-1.fc10.i386
[leigh@localhost ~]$ uname -r
2.6.27.5-94.fc10.i686
[leigh@localhost ~]$ glxinfo
name of display: :0.0
display: :0 screen: 0
direct rendering: Yes
server glx vendor string: NVIDIA Corporation
server glx version string: 1.4
server glx extensions:
GLX_EXT_visual_info, GLX_EXT_visual_rating, GLX_SGIX_fbconfig,
GLX_SGIX_pbuffer, GLX_SGI_video_sync, GLX_SGI_swap_control,
GLX_EXT_texture_from_pixmap, GLX_ARB_multisample, GLX_NV_float_buffer,
GLX_ARB_fbconfig_float
client glx vendor string: NVIDIA Corporation
client glx version string: 1.4
client glx extensions:
GLX_ARB_get_proc_address, GLX_ARB_multisample, GLX_EXT_visual_info,
GLX_EXT_visual_rating, GLX_EXT_import_context, GLX_SGI_video_sync,
GLX_NV_swap_group, GLX_NV_video_out, GLX_SGIX_fbconfig, GLX_SGIX_pbuffer,
GLX_SGI_swap_control, GLX_NV_float_buffer, GLX_ARB_fbconfig_float,
GLX_EXT_fbconfig_packed_float, GLX_EXT_texture_from_pixmap,
GLX_EXT_framebuffer_sRGB, GLX_NV_present_video,
GLX_NV_multisample_coverage
GLX version: 1.3
GLX extensions:
GLX_EXT_visual_info, GLX_EXT_visual_rating, GLX_SGIX_fbconfig,
GLX_SGIX_pbuffer, GLX_SGI_video_sync, GLX_SGI_swap_control,
GLX_EXT_texture_from_pixmap, GLX_ARB_multisample, GLX_NV_float_buffer,
GLX_ARB_fbconfig_float, GLX_ARB_get_proc_address
OpenGL vendor string: NVIDIA Corporation
OpenGL renderer string: GeForce 7600 GT/PCI/SSE2/3DNOW!
OpenGL version string: 2.1.2 NVIDIA 177.80

[leigh123linux]

leigh

are you gonna bother building the 177.82 driver or leave it for RPMFusion now? http://www.nvidia.com/object/unix.html


No I not , I am busy with other projects at the moment ( qbittorrent & compiz-fusion )


Besides it doesn't really offer anything new .

Version: 177.82
Operating System: Linux x64 (AMD64/EM64T)
Release Date: November 12, 2008

Release Highlights

* Added support for the following new GPUs:
o Quadro NVS 450
o Quadro FX 370 LP
o Quadro FX 5800
o Quadro FX 4800
o Quadro FX 470
o Quadro CX
* Fixed a problem on recent mobile GPUs that caused a power management resume from S3 to take 30+ seconds.
* Fixed a problem with hotkey switching on some recent mobile GPUs.
* Fixed an image corruption issue seen in FireFox 3.

[Demz]

No I not , I am busy with other projects at the moment ( qbittorrent & compiz-fusion )

are you not gonna submit qbittorrent as a RFE in bugzilla for Fedora11?

[leigh123linux]

are you not gonna submit qbittorrent as a RFE in bugzilla for Fedora11?


I was hoping to get a Fedora Dev to adopt qbittorrent :cool: as this would be the quickest way to get qbittorrent into Fedora .

[Demz]

I was hoping to get a Fedora Dev to adopt qbittorrent :cool: as this would be the quickest way to get qbittorrent into Fedora .

it would.... i dont know if Rex Dieter has done anything about it as he thought the package looked pretty good ( he mentioned it in another one of your threads ) as far as i knew he only mentioned to make it as a RFE in bugzilla

[hephasteus]

Thanks for that, to be honest I don't really need all the whistles and bells as I am only running Linux for the performance of Dynagen, but this has become a sort of nemesis and I won't let it drop until its done.... Yeah Yeah I must be mad !! I think I am !!:D

As for your question, I have only tried on Linux at the moment but tomorrow I'll give it a whirl on Windows, hopefully the hardware is OK.

I also have the latest bios for the motherboard, I have found out I am on Rev 5.0 and the latest is v12.0 so I am going to go with that too...

I'll report back, if you don't hear from me assume I have given up and gone to live as a monk in the hills somewhere !!

Best

Steve

There's a volcano with about 120,000 Newest nvidia video card virgins bones in it. Glad to see they can still get people to jump into it. :D I'll stick with my 7600GS. :p

[net_boy]

Whoaaaaaaaaaaaaaaaa


Bios update on gigabyte board to v12 and akmod-nvidia kernel 101

sorted.

http://img356.imageshack.us/img356/4927/whoaaaaol9.th.jpg (http://img356.imageshack.us/my.php?image=whoaaaaol9.jpg)http://img356.imageshack.us/images/thpix.gif (http://g.imageshack.us/thpix.php)

[sj200449]

Got the compiled version working this evening using : NVIDIA-Linux-x86_64-177.82-pkg2.run

Here's the context...

uname -r:
2.6.27.5-101.fc10.x86_64

this kernel was updated today for me when I ran "yum update"

Initially the nvidia.ko bit of the boot process complain there was no module, so the yum update had not been provided with a kmod-nvidia... for this kernel (the -101 release)

The gfx btw is an nvidia GTS 8800.

So,

I did a yum install for

kernel-devel
kernel-headers
gcc
gcc-++
xorg-x11-drv-nvidia-libs.i386 (this wasnt really necessary but I did it anyway to allow 32-bit stuff such as Wine to work)


(kernel-headers were there already)

Went to Nvidia site and downloaded the NVIDIA-Linux-x86_64-177.82-pkg2.run

Attempted to "sh <nvidia download>" while booted into the newer kernel (-101 release) , this complained as I was running an X server at the time.

killed the X process - which restarted , so as I couldnt remember how to stop it doing that , I just editted /etc/inittab and changed the run level to 3 instead of 5

Redid the "sh <nvidia download>" and it began to work then moaned about "make" not being available.

As, without X running my wireless didnt pick up correctly I reverted to X (run level 5 again) and yum installed "make", then went back to runlevel 3 (no X) and tried 3rd attempt at "sh <nvidia stuff>"

The nvidia.sh script worked this time. It asked me if I wanted to connect to nvidia.com to see if a precompiled module for the nvidia interface for my kernel was available and I selected No - I want to compile it (the whole point of the exercise), it went away and tried to compile it and this was successful. It then asked if I wanted to run the nvidia-xconfig tool to update my Xorg.conf which I selected Yes for and it just did this - no output other than .. done ok.

To test I reverted to the runlevel 5 again and rebooted my machine.

Machine booted into X login manager with no fails and it was a nice resolution again rather than the 800x600 guff I get if it doesnt work.

All Good.


-Stephen.


NB : Although it worked, I noticed that if I watch the boot messages there is still the failure to load module nvidia.ko ... not loading until one is available - Now this is normally resolved if I just download an available kmod-nvidia for the running kernel from rpmfusion (with yum update) and I had expected it to be gone. However even though the warning is still there , it does work and Im mostly happy.

Hope this is of use.

[jerboyd]

just installed f10 32 bit preview on my laptop (core 2 duo with nvidia 9600m gt), i believe i have the PAE kernel installed. i updated all my packages and then installed the nividia driver with
yum install -y xorg-x11-drv-nvidia

after a reboot i was surprised that it worked. no tinkering required. yesterday i did a yum update and after the newest kernel installed and i rebooted, the nvidia driver doesn't load. also, i had compiz running and after logging in i got a white screen after compiz took over. i booted up at runlevel 3 and uninstalled compiz so i could get back in to my desktop (i'm not so good at fixing things in the shell).

i tried reinstalling the driver but it still doesn't load, i'm assuming it is because the kmod-nvidia package hasn't been recompiled for the newest kernel. my question is this, am i going to have this problem every time i update my kernel? is there any way to get around this other than excluding the kernel from yum updates? also, if i just wait a couple days will my nvidia driver update itself and set everything aright?

thanks
jeremy boyd

[leigh123linux]

just installed f10 32 bit preview on my laptop (core 2 duo with nvidia 9600m gt), i believe i have the PAE kernel installed. i updated all my packages and then installed the nividia driver with
yum install -y xorg-x11-drv-nvidia

after a reboot i was surprised that it worked. no tinkering required. yesterday i did a yum update and after the newest kernel installed and i rebooted, the nvidia driver doesn't load. also, i had compiz running and after logging in i got a white screen after compiz took over. i booted up at runlevel 3 and uninstalled compiz so i could get back in to my desktop (i'm not so good at fixing things in the shell).

i tried reinstalling the driver but it still doesn't load, i'm assuming it is because the kmod-nvidia package hasn't been recompiled for the newest kernel. my question is this, am i going to have this problem every time i update my kernel? is there any way to get around this other than excluding the kernel from yum updates? also, if i just wait a couple days will my nvidia driver update itself and set everything aright?

thanks
jeremy boyd


Or if you like playing with unreleased Koji kernels

Code:

su
yum install akmod-nvidia xorg-x11-drv-nvidia-libs xorg-x11-drv-nvidia


Use the akmod instead.

[net_boy]

I am getting this error after what seems to be a bit of a botched install *doh!

root@homer nelsons]# yum update
Loaded plugins: refresh-packagekit
Setting up Update Process
Resolving Dependencies
--> Running transaction check
---> Package xorg-x11-drv-nvidia-libs.x86_64 0:177.82-1.fc10 set to be updated
---> Package xorg-x11-drv-nvidia.x86_64 0:177.82-1.fc10 set to be updated
--> Processing Dependency: nvidia-kmod >= 177.82 for package: xorg-x11-drv-nvidia
--> Finished Dependency Resolution
xorg-x11-drv-nvidia-177.82-1.fc10.x86_64 from rpmfusion-nonfree-rawhide has depsolving problems
--> Missing Dependency: nvidia-kmod >= 177.82 is needed by package xorg-x11-drv-nvidia-177.82-1.fc10.x86_64 (rpmfusion-nonfree-rawhide)
Error: Missing Dependency: nvidia-kmod >= 177.82 is needed by package xorg-x11-drv-nvidia-177.82-1.fc10.x86_64 (rpmfusion-nonfree-rawhide)
[root@homer nelsons]#
[root@homer nelsons]#
[root@homer nelsons]#
[root@homer nelsons]#
[root@homer nelsons]#
[root@homer nelsons]# rpm -qa |grep nvidia
akmod-nvidia-177.80-1.fc10.x86_64
kmod-nvidia-2.6.27.5-101.fc10.x86_64-177.80-1.fc10.x86_64
xorg-x11-drv-nvidia-libs-177.80-1.fc10.x86_64
xorg-x11-drv-nvidia-177.80-1.fc10.x86_64
[root@homer nelsons]#

That said the drivers are all working fine and I assume that is because I installed the akmod-nvidia module.

Any tips on how to straighten it out, or opinions on whether I need to at all ?

Thanks

Steve

[spartas]

net_boy,

I am getting the same thing. Seems that rpmfusion published new xorg drivers for the new 177.82 nvidia release, but haven't released the nvidia-kmod package for it yet. Wait a while and you should be able to update.

[Demz]

I am getting this error after what seems to be a bit of a botched install *doh!

root@homer nelsons]# yum update
Loaded plugins: refresh-packagekit
Setting up Update Process
Resolving Dependencies
--> Running transaction check
---> Package xorg-x11-drv-nvidia-libs.x86_64 0:177.82-1.fc10 set to be updated
---> Package xorg-x11-drv-nvidia.x86_64 0:177.82-1.fc10 set to be updated
--> Processing Dependency: nvidia-kmod >= 177.82 for package: xorg-x11-drv-nvidia
--> Finished Dependency Resolution
xorg-x11-drv-nvidia-177.82-1.fc10.x86_64 from rpmfusion-nonfree-rawhide has depsolving problems
--> Missing Dependency: nvidia-kmod >= 177.82 is needed by package xorg-x11-drv-nvidia-177.82-1.fc10.x86_64 (rpmfusion-nonfree-rawhide)
Error: Missing Dependency: nvidia-kmod >= 177.82 is needed by package xorg-x11-drv-nvidia-177.82-1.fc10.x86_64 (rpmfusion-nonfree-rawhide)
[root@homer nelsons]#
[root@homer nelsons]#
[root@homer nelsons]#
[root@homer nelsons]#
[root@homer nelsons]#
[root@homer nelsons]# rpm -qa |grep nvidia
akmod-nvidia-177.80-1.fc10.x86_64
kmod-nvidia-2.6.27.5-101.fc10.x86_64-177.80-1.fc10.x86_64
xorg-x11-drv-nvidia-libs-177.80-1.fc10.x86_64
xorg-x11-drv-nvidia-177.80-1.fc10.x86_64
[root@homer nelsons]#

That said the drivers are all working fine and I assume that is because I installed the akmod-nvidia module.

Any tips on how to straighten it out, or opinions on whether I need to at all ?

Thanks

Steve

wait till your Mirrors are synced ..but if you use the akmod-nvidia you wont need to depend on waiting for the kmod all the time to be released

[jdeslip]

Is this going to be updated to 180.06 driver for F10?

[leigh123linux]

Is this going to be updated to 180.06 driver for F10?


What version? :confused:

@rpmfusion

akmod-nvidia-177.82-1.fc10


or for openGL 3 developers ( needs GF8 + to work )

akmod-nvidia-beta-177.61.02-2.fc10

[jdeslip]

180.06 was released today by nvidia with some nice fixes and features. I am wondering if this is going to come rpmfusion repos or if they are going to stick with the 178 for F10 like they stuck with 173 for F9.

[ogetbilo]

jdeslip,
We already have so many branches to deal with. Now this 180.06 is a second beta branch in addition to 177.61.02 . I will see what we can do at rpmfusion.

[Demz]

Is this going to be updated to 180.06 driver for F10?

i dont think that driver will be out for awhile,,as it has a lot of new features that include OpenGL3.0 support an a few others

[jdeslip]

Ya, I understand. As a suggestion, I'd recommending building as part of the main branch (what is currently 177.82), but keeping it in the testing repo until the stable 180 version comes out. This makes sense because the 180 driver will be the successor of 177. Who knows when the opengl 3.0 stuff will get merged in...

[jdeslip]

i dont think that driver will be out for awhile,,as it has a lot of new features that include OpenGL3.0 support an a few others

It was released today! (there are lots of new features, but OpenGL3.0 was left out)

[Demz]

thanks.. i just found it ..

[hephasteus]

I am getting this error after what seems to be a bit of a botched install *doh!

root@homer nelsons]# yum update
Loaded plugins: refresh-packagekit
Setting up Update Process
Resolving Dependencies
--> Running transaction check
---> Package xorg-x11-drv-nvidia-libs.x86_64 0:177.82-1.fc10 set to be updated
---> Package xorg-x11-drv-nvidia.x86_64 0:177.82-1.fc10 set to be updated
--> Processing Dependency: nvidia-kmod >= 177.82 for package: xorg-x11-drv-nvidia
--> Finished Dependency Resolution
xorg-x11-drv-nvidia-177.82-1.fc10.x86_64 from rpmfusion-nonfree-rawhide has depsolving problems
--> Missing Dependency: nvidia-kmod >= 177.82 is needed by package xorg-x11-drv-nvidia-177.82-1.fc10.x86_64 (rpmfusion-nonfree-rawhide)
Error: Missing Dependency: nvidia-kmod >= 177.82 is needed by package xorg-x11-drv-nvidia-177.82-1.fc10.x86_64 (rpmfusion-nonfree-rawhide)
[root@homer nelsons]#
[root@homer nelsons]#
[root@homer nelsons]#
[root@homer nelsons]#
[root@homer nelsons]#
[root@homer nelsons]# rpm -qa |grep nvidia
akmod-nvidia-177.80-1.fc10.x86_64
kmod-nvidia-2.6.27.5-101.fc10.x86_64-177.80-1.fc10.x86_64
xorg-x11-drv-nvidia-libs-177.80-1.fc10.x86_64
xorg-x11-drv-nvidia-177.80-1.fc10.x86_64
[root@homer nelsons]#

That said the drivers are all working fine and I assume that is because I installed the akmod-nvidia module.

Any tips on how to straighten it out, or opinions on whether I need to at all ?

Thanks

Steve

They updated to kernel 109. Go to system -> software -> sources. And set RPM fusion to fedora 10 free, RPM fusion for fedora 10-free updates and the same for the non free pair. The 109 nvidia driver is in those. Try it and if it works alright stick with it or go back and set your bootloader back to the 101 you just got working. It's working pretty good for me but im getting some kernel hiccups on my ethernet.

[ruien]

I have a 8600M GT card on this laptop and with theese new drivers (177.82 atm) my desktop hangs all the time.Running kernel 2.6.27.5-109.fc10.x86_64. Gnome and KDE are sluggish and unusable. Its reported in nvnews her : http://www.nvnews.net/vbulletin/showthread.php?t=123003
and here: http://www.nvnews.net/vbulletin/showthread.php?t=115724

Is it possible to get older nvidia drivers known to work fire on 8600 series cards? like 169.12 or even older? Never had any problems before.

[leigh123linux]

For nvidia performance issues try

http://www.nvnews.net/vbulletin/showthread.php?t=118088

It implements some of the new beta ( 180.06 ) driver changes

[ruien]

For nvidia performance issues try

http://www.nvnews.net/vbulletin/showthread.php?t=118088

It implements some of the new beta ( 180.06 ) driver changes


Thanks. That helped alot but still its not optimal. Installed 180.06 drivers and did this:
nvidia-settings -a InitialPixmapPlacement=2 -a GlyphCache=1 (should add this to ~xinitrc)
and added this to xorg.conf:
Option "AllowSHMPixmaps" "0"
Option "PixmapCacheSize" "2000000"

However. I installed akmod-nvidia before i downloaded the 180.06 driver from nvidia server and even though i uninstalled akmod-nvidia via yum the akmod scripts still stayed in the system. Wich means that the scripts starts with boot and try to build the nvidia driver again. That will naturally fail (no nvidia.ko module found) and akmod scripts reset the working xorg.conf from nvidia driver to nv. It took me some time to figure out that i needed to remove the akmod init and build scripts in order to get the new driver to load properly. Is this a preview problem or doesnt theese akmod scripts get removed when you uninstall akmod's regardless FC version? Anyway. I did a locate and deleted all akmod init and build scripts. Edited xorg.conf back to nvidia and rebooted into a working 180.06 desktop. This should be sticky somewhere if akmod's doesnt remove everything duiring uninstall. I couldn't find anything about this here. I just read somewhere that akmod-nvidia auto biuld nvidia drivers duiring boot. That made me look for the scripts.

My desktop is usable and the response is much better now so I'm satisfied for now. Haven't tried out Unreal or Urban Terror yet tho. I'll see how it goes when it's gametime:D

[Karmaflute]

Use this to install nvidia drivers for F10:

http://download1.rpmfusion.org/free/fedora/development/i386/os/rpmfusion-free-release-10-1.noarch.rpm

http://download1.rpmfusion.org/nonfree/fedora/development/i386/os/rpmfusion-nonfree-release-10-1.noarch.rpm

yum install akmod-nvidia xorg-x11-drv-nvidia-libs xorg-x11-drv-nvidia

[jostd]

Any idea when the 180 driver will be available on rpmfusion-nonfree?
I wonder whether to wait or try the nvidia installer.

[ruien]

Any idea when the 180 driver will be available on rpmfusion-nonfree?
I wonder whether to wait or try the nvidia installer.

I have no idea. It's still a beta driver but it works fine for me. FC10 isnt released yet either :) If you don't have a performance problem or some other issues with the rpmfusion driver then stick to it for now.