Jump to content

Latest Nvidia driver won't work?


83mercedes
 Share

Recommended Posts

If you uninstalled 4363 then your output from 'file' says nvidias glx from 4363 is still installed, and should point to 4496 after 4496 is installed. Did you uninstall with --uninstall? (I guess so what else is there :roll: ). When you install 4496 does it ask if you want to install the glx files? Sorry, I don't know if it does 'normally, I've only used -ane with the installer

a=skip the license agreement (if I remember correctly)

n=no precompiled mods

e=expert (prompts for all options)

 

this way will ask everything. You some how need to get all the 4363 stuff out b4 installing 4496 (obviously) but I wonder why the installer isn't doing it?

 

I don't think it's the kernel.....someone else would have said something by now...I'd think?

Link to comment
Share on other sites

  • Replies 42
  • Created
  • Last Reply

Top Posters In This Topic

For clarification, ever since they made the installer, I have stopped forcing a compile and just allowed the installer to work. The latest one told me what it was doing every step. It found the old driver, removed it, looked for a pre-compile, didn't find one, and then asked if I wanted it to compile one on my machine. When I said yes, it compiled, installed, updated glx and modules.conf. It is a really nice installer, and I have had no nvidia related problems with my older video card.

Link to comment
Share on other sites

Actually, I have had no real problems (before) with the installer either.

bvc, sorry, I guess I misled you... the command you gave me, I did that with 4363 installed, not with the other, which would explain the simlink.

And yes I used:

nvidia-installer --uninstall

to remove 4363. Then I did

sh NVIDIA-Linux-x86-1.0-4496-pkg2.run

to install.

My config-4 file is good, OK.

I'll keep fooling with it, not cuz I really need it, but just out of sheer bull-headedness!:evil:

Thanks again.

Link to comment
Share on other sites

OK, this is getting weird!

Now, I installed 4496, and the ONLY way I can get into X is by booting in 'failsafe' mode, then choosing 'text mode & network', then logging in as a normal user (me), then doing 'startx'.

I re-ran lilo, still, it will not auto-login to kde on boot. (Using kde 3.1.3)

 

Any ideas at this point??

Thanks guys, for all your input. :)

Link to comment
Share on other sites

in your modules.conf, is there an "alias char-major-195 nvidia" ? if yes, just change the alias in /etc/modules.conf -- make it read alias /dev/nvidia* nvidia, then save/exit, and run /sbin/depmod -ae as root. It should autoload on the next boot.

hope this helps

Link to comment
Share on other sites

james!!!

Thanks, that did it! :cheeky:

What a hassle! Can you explain why the modules.conf would not have been properly setup with the new driver? I just don't get it, it has never been a problem with the older Nvidia installers?

And why just me? (sniffle...) I'm not a bad person...

:)

Thanks again for everyone's input.

Link to comment
Share on other sites

Awriigght!!! :)

 

Anyway, this seems to be happening since 1.0.4349, mostly on Mandrake 9.0 and 9.1. as for the reason, below is a cutnpaste from post of bwkaz on 04-02-2003 01:55 PM in nvidia forums.

 

" It probably changed alias /dev/nvidia* nvidia to alias char-major-195 nvidia, right? It'll use the char-major alias when you aren't running devfs -- how exactly did you boot the time you installed the drivers? Were you in like a rescue mode, or single-user mode, or failsafe mode, or something like that? Mandrake, IIRC, in failsafe mode, doesn't mount /dev. This will make the installer think that you don't use devfs (since it has no other way of knowing that I can think of -- in any case, this is what it checks), so it'll use the pre-devfs alias, which is char-major-195. If you have /dev mounted when the installer runs, then it'll find the /dev/.devfsd character device, and it'll use the devfs alias -- /dev/nvidia*"

 

another quote..

" The installer tries to add lines to certain config files to make the nvidia kernel module load automatically, either on boot, or when it's needed, but it doesn't always work -- and most often, that happens on Mandrake. Especially if you boot to failsafe mode to install the driver -- failsafe is not the same as "normal, just no X". In particular, failsafe doesn't mount the device filesystem (the in-memory one that keeps track of all devices), but uses the older way of using device files, where they're actually stored on disk. The method to get the nvidia module to autoload differs, depending on whether devfs is in use or not."

 

if the arguments are correct, then /sbin/init 3 is failsafe mode for Mandrake? or is this a bug in the installer? there had been a number of complaints on this that am puzzled this was not included in the nvidia readme up to now.

 

Anyway, glad to be part of the team effort! :)

Link to comment
Share on other sites

Im using the multimedia kernel so i wanted my own

NVIDIA RPMS i used Texstars spec files to start with

and altered them to fit my needs he has added

some code that makes all the needed modifications to

the Xfree config files in Mandrake.

Any way you can easily use theese ones to build rpms

for any kernel.

 

/thac

 

#

# Spec file for NVIDIA X server module and OpenGL libraries.

#

# /usr/X11R6 is relocatable.

#

# $Header: //sw/main/drivers/NVIDIA_GLX.spec#24 $

#

%define RELEASE 4496

%define rel %{?CUSTOM_RELEASE} %{!?CUSTOM_RELEASE:%RELEASE}

%define ver 1.0

%define nam NVIDIA_GLX

 

Summary: NVIDIA module for XFree86 4.0 X server and OpenGL libraries

Name: NVIDIA_multimedia_GLX

Version: %{ver}mdk

Release: %rel

Copyright: NVIDIA Corp. 2000

Group: User Interface/X Hardware Support

Source: ftp://ftp1.detonator.nvidia.com/pub/drive...RELEASE}.tar.gz

URL: http://www.nvidia.com

Vendor: NVIDIA Corp.

Packager: Buildmeister

Prefix: /usr/X11R6

Requires: XFree86 >= 4.0.1

Requires: NVIDIA_multimedia_kernel

Requires: kernel > 2.2.0

BuildRoot: %{_tmppath}/%{name}-%{PACKAGE_VERSION}

 

# disable build root strip policy

# XFree86 drivers may NOT be stripped...

%define __spec_install_post /usr/lib/rpm/brp-compress

 

%description

NVIDIA Architecture support for Common X and OpenGL for all Linux systems.

NVIDIA XFree86 4.0 server module and OpenGL 1.2 libraries for RIVA TNT/TNT2

and GeForce/Quadro based video cards. Older RIVA 128 based video cards are

supported by the server module shipping with XFree86 4.0, nv_drv.o. You

should install this package if you have one of the newer cards.

 

You must also install the NVIDIA_kernel module if you want to utilize these

drivers.

 

%prep

%setup -q -n %{nam}-%{ver}-${RPM_PACKAGE_RELEASE}

 

%build

 

%install

DSOMAJOR=1

DSOMINOR=0

DSOPATCH=4496

DSOREV=$DSOMAJOR.$DSOMINOR.$DSOPATCH

 

[ -n "$RPM_BUILD_ROOT" -a "$RPM_BUILD_ROOT" != / ] && rm -rf $RPM_BUILD_ROOT

mkdir -p $RPM_BUILD_ROOT/usr/X11R6/lib/modules/drivers

mkdir -p $RPM_BUILD_ROOT/usr/X11R6/lib/modules/extensions

mkdir -p $RPM_BUILD_ROOT/usr/lib

 

install usr/X11R6/lib/modules/drivers/nvidia_drv.o $RPM_BUILD_ROOT/usr/X11R6/lib/modules/drivers/nvidia_drv.o

install usr/X11R6/lib/modules/extensions/libglx.so.$DSOREV $RPM_BUILD_ROOT/usr/X11R6/lib/modules/extensions/libglx.so.$DSOREV

install usr/X11R6/lib/libXvMCNVIDIA.a $RPM_BUILD_ROOT/usr/X11R6/lib/libXvMCNVIDIA.a

install usr/X11R6/lib/libXvMCNVIDIA.so.$DSOREV $RPM_BUILD_ROOT/usr/X11R6/lib/libXvMCNVIDIA.so.$DSOREV

install usr/lib/libGLcore.so.$DSOREV $RPM_BUILD_ROOT/usr/lib/libGLcore.so.$DSOREV

install usr/lib/libGL.so.$DSOREV $RPM_BUILD_ROOT/usr/lib/libGL.so.$DSOREV

 

%pre

 

# if /usr/lib/libGL.so is pointing to libGL.so.1, we are upgrading.

# This is good. We just delete libGL.so. We'll set it up later.

# If not, leave it alone, we'll complain later

if [ -e /usr/lib/libGL.so ]; then

LIBGL=`ls -l /usr/lib/libGL.so | sed 's/.*->[ ]*(.*)$/1/'`

if [ X$LIBGL = X"libGL.so.1" ]; then

rm -f /usr/lib/libGL.so

fi

fi

 

problem=0

first_time=0

 

check_file() {

if [ -f $1 ]; then

# remove symlinks, and append .RPMSAVE to real files

if [ -h $1 ]; then

rm -f $1

else

problem=1

if [ $first_time -eq 0 ]; then

echo " --- WARNING!!"

first_time=1

fi

mv $1 xxx.$1.RPMSAVE

echo " $1"

fi

fi

}

 

final_check() {

if [ $problem -eq 1 ]; then

echo " --- The above file(s) possibly belong to a conflicting MESA rpm,"

echo " --- or to a previous stale NVIDIA distribution."

echo " --- They have been renamed to xxx.<originalFile>.RPMSAVE to"

echo " --- avoid conflicting with the files contained within this"

echo " --- package."

echo " --- Please see the FREQUENTLY ASKED QUESTIONS section of"

echo " --- /usr/share/doc/NVIDIA_GLX-1.0/README for more details."

fi

}

 

#

# Warn people that we have problems with those.

#

 

# Gaurd against the posibility that this directory doesn't exist

if [ -d /usr/X11R6/lib/modules/extensions ]; then

cd /usr/X11R6/lib/modules/extensions

check_file libglx.a

check_file libGLcore.a

fi

 

# shouldn't need to change directories here, but just in case..

# We do, RedHat's Mesa packages install to /usr/lib now

cd /usr/lib

check_file libGL.a

for file in libGL.so*

do

check_file $file

done

 

cd /usr/X11R6/lib

check_file libGL.a

for file in libGL.so*

do

check_file $file

done

 

for file in libGLcore.so*

do

check_file $file

done

 

final_check

 

%post

DSOMAJOR=1

DSOMINOR=0

DSOPATCH=4191

DSOREV=$DSOMAJOR.$DSOMINOR.$DSOPATCH

 

/sbin/ldconfig

/sbin/ldconfig -n /usr/X11R6/lib/modules/extensions

/sbin/ldconfig -n /usr/X11R6/lib

 

# create compile-time links

cd /usr/lib

ln -sf libGL.so.$DSOMAJOR libGL.so

 

%postun

 

first_time=0

 

restore_files () {

# Just get them all, hopefully noone else filled in any

for i in `ls xxx.*.RPMSAVE 2> /dev/null` ; do

# In earlier releases we could end up creating xxx.*.RPMSAVE files from

# our own files. There is no point in restoring those.

if [ "`grep -a -c -e "NVIDIA Corporation" $1`" -eq 0 ]; then

# Rename RPMSAVE files that are not our own

if [ $first_time -eq 0 ]; then

echo " --- WARNING!!"

first_time=1

fi

filename=`echo $i | sed -e "s/xxx.//" -e "s/.RPMSAVE//"`

echo " $i"

mv $i $filename

else

# remove our own

rm $i

fi

done

}

 

final_check() {

if [ $first_time -ne 0 ]; then

echo " --- The above files are apparently saved files from a"

echo " --- non-NVIDIA distribution (possibly MESA) that were"

echo " --- put aside while the NVIDIA rpm was installed."

echo " --- They are now being restored."

fi

}

 

 

# $1 contains the number of instances of this package that will be

# installed after the operation has completed. So [ $1 = 1 ] ==>

# upgrade and [ $1 = 0 ] ==> uninstall. In the uninstall case we

# need to clean up.

if [ $1 = 0 ]; then

# Restore the Mesa libraries that we saved away at install time

cd /usr/X11R6/lib/modules/extensions

restore_files

 

cd /usr/lib

restore_files

 

cd /usr/X11R6/lib

restore_files

 

final_check

 

# Remove any symlinks that would have been created for our shared libraries.

# Symlinkds required for Mesa will be recreated when ldconfig is run

# but symlinks specific to our drivers will be left dangling

# (/usr/lib/libGLcore.so.1 does this). These don't get removed by uninstall

# because we didn't install them, ldconfig created them

rm -f /usr/lib/libGLcore.so.1

rm -f /usr/lib/libGLcore.so

rm -f /usr/lib/libGL.so.1

rm -f /usr/lib/libGL.so

rm -f /usr/X11R6/lib/modules/extensions/libglx.so

fi

 

/sbin/ldconfig

 

%clean

[ -n "$RPM_BUILD_ROOT" -a "$RPM_BUILD_ROOT" != / ] && rm -rf $RPM_BUILD_ROOT

 

%files

%defattr (-, root, root)

%doc usr/share/doc/README

%doc usr/share/doc/README.DE

%doc usr/share/doc/license.txt

%doc usr/share/doc/NVIDIA_Changelog

%doc usr/share/doc/XF86Config.sample

# Install these headers in the doc dir so as not to destroy the installed

# headers but still give the user the option to use them

%doc usr/include

/usr/X11R6/lib/modules/drivers/nvidia_drv.o

/usr/X11R6/lib/modules/extensions/libglx.so.1.0.4496

%attr (0644, root, root) /usr/X11R6/lib/libXvMCNVIDIA.a

/usr/X11R6/lib/libXvMCNVIDIA.so.1.0.4496

/usr/lib/libGLcore.so.1.0.4496

/usr/lib/libGL.so.1.0.4496

 

%changelog

* Wed Jul 30 2003 Torbjorn Turpeinen <tobbe@nyvalls.se>

- Updated to latest version

 

* Mon Mar 17 2003 Texstar

- Build for Mandrake 9.1

 

* Mon May 13 2002 George Kyriazis <kyriazis@nvidia.com>

- Install libXvMCNVIDIA.so, too.

 

* Mon Mar 4 2002 Joseph Kain <jkain@nvidia.com>

- Install README.DE

 

* Tue Feb 26 2002 Joseph Kain <jkain@nvidia.com>

- libXvMCNVIDIA.so is now libXvMCNVIDIA.a

 

* Thu Feb 21 2002 George Kyriazis <kyriazis@nvidia.com>

- Added libXvMCNVIDIA.so

 

* Sat Feb 16 2002 Joseph Kain <jkain@nvidia.com>

- Protect against missing directories, specifically

/usr/X11R6/lib/modules/extensions

 

* Mon Nov 26 2001 Joseph Kain <jkain@nvidia.com>

- Install license file

 

* Mon Nov 26 2001 Joseph Kain <jkain@nvidia.com>

- Use %{_tmppath} instead of hardcoding /var/tmp

 

* Wed Mar 28 2001 Joseph Kain <jkain@nvidia.com>

- Stop renaming our own old libraries with xxx.*.RPMSAVE

- Detect uninstall case to restore saved Mesa libs and remove extra symlinks

- added gl.h and glx.h

 

* Mon May 1 2000 George Kyriazis <kyriazis@nvidia.com>

- Fixed installation of revisions of shared libraries.

 

* Thu Apr 27 2000 Gregory McLean <gregm@comstar.net>

- Added BuildRoot definition.

- Added some logic to deal with the two most common problems.

- Added the FAQ.

- Added logic for epoc builds if needed.

 

 

 

 

#

# Spec file for NVIDIA kernel module.

#

# $Header: //sw/main/drivers/resman/NVIDIA_kernel.spec#22 $

#

%define RELEASE 4496

%define CUSTOM_RELEASE 4496

%define rel %{?CUSTOM_RELEASE} %{!?CUSTOM_RELEASE:%RELEASE}

%define ver 1.0

%define nam NVIDIA_kernel

 

Summary: NVIDIA kernel module for NVIDIA Architecture support.

Name: NVIDIA_multimedia_kernel

Version: %{ver}mdk

Release: %rel

Copyright: NVIDIA Corp. 2000

Group: User Interface/X Hardware Support

Source: ftp://ftp1.detonator.nvidia.com/pub/drive...RELEASE}.tar.gz

URL: http://www.nvidia.com

Vendor: NVIDIA Corp.

BuildRoot: %{_tmppath}/%{name}-%{PACKAGE_VERSION}

Packager: Buildmeister

Requires: kernel >= 2.2.0

Requires: rpm >= 3.0.4

Requires: perl

Buildrequires: kernel-multimedia-source

 

%description

NVIDIA Architecture support for systems with updated or custom kernels.

 

%prep

%setup -q -n %{nam}-%{ver}-${RPM_PACKAGE_RELEASE}

 

%build

# don't exit immediately when something bad happens.

set +e

 

# building an rpm from an srpm does not check the requires field.

# make sure that rpm is good enough for us.

RPMVERSION=`rpm -q --queryformat "%{VERSION}" rpm`

RPMMAJOR=`echo $RPMVERSION | awk -F . '{print $1;}'`

RPMMINOR=`echo $RPMVERSION | awk -F . '{print $2;}'`

RPMPATCH=`echo $RPMVERSION | awk -F . '{print $3;}'`

if [ $RPMMAJOR -le 3 -a $RPMMINOR -le 0 -a $RPMPATCH -lt 4 ]; then

echo -e "nNeed rpm version 3.0.4 or greatern";

exit 1;

fi

 

make nvidia.o

 

# This will install the module itself onto the user's system, but not

# load it into the running system. This will allow our auto-builds to

# execute properly

%install

 

# create a file list for nvidia.o, to make sure the path is correct

# it really should be under $RPM_BUILD_ROOT, but the %files section

# does not honor it. Bummer.

rm -f /tmp/files.lst

 

# allow a build script to override the targetted kernel

# otherwise default to current running kernel

if [ -z "$TARGET_KERNEL" ]; then

export TARGET_KERNEL=`uname -r`

fi

 

if [ -d $RPM_ROOT/lib/modules/$TARGET_KERNEL/kernel ]; then

INSTALLPATH=/lib/modules/$TARGET_KERNEL/kernel/drivers/video

else

INSTALLPATH=/lib/modules/$TARGET_KERNEL/video

fi

mkdir -p $RPM_BUILD_ROOT/$INSTALLPATH

install -m 0444 nvidia.o $RPM_BUILD_ROOT/$INSTALLPATH

echo "%attr(444 root root) " $INSTALLPATH/nvidia.o > /tmp/files.lst

unset RPM_BUILD_ROOT # trick RPM 4.0 into not stripping us

 

%clean

[ -n "$RPM_BUILD_ROOT" -a "$RPM_BUILD_ROOT" != / ] && rm -rf $RPM_BUILD_ROOT

 

# Old versions of the tarball Makefile would install NVdriver in the

# wrong directory. Remove the old NVdriver, if that leaves it parent

# directory empty remove it as well. Note that we always install the

# nvidia.o module in the correct directory, so we don't need to worry

# about it here.

%pre

WRONGDIR=/lib/modules/`uname -r`/kernel/video

 

if [ -d $WRONGDIR ]; then

rm -f $WRONGDIR/NVdriver

rmdir --ignore-fail-on-non-empty $WRONGDIR

fi

 

 

# In %post we basically do "run-time" installation

# Note that our files are installed onto the system, and the packages

# are cleaned up and gone. There are no Makefiles or fancy environment

# variables to play with. I basically repeated the needed parts of the

# install.sh script here

%post

 

# unload it quietly if already loaded..

/sbin/rmmod NVdriver >& /dev/null

/sbin/rmmod nvidia >& /dev/null

 

rmmod_failed=0

/sbin/lsmod | grep nvidia >& /dev/null && rmmod_failed=1

/sbin/lsmod | grep NVdriver >& /dev/null && rmmod_failed=1

 

if [ $rmmod_failed -eq 1 ]; then

echo

echo "Unable to remove existing NVIDIA kernel module."

echo "Please be sure you have exited X before attempting"

echo "to install the NVIDIA_kernel rpm."

echo

exit 1

fi

 

# delete the old NVdriver

# XXX is there a better way to do this?

if [ -z "$TARGET_KERNEL" ]; then

export TARGET_KERNEL=`uname -r`

fi

if [ -d $RPM_ROOT/lib/modules/$TARGET_KERNEL/kernel ]; then

INSTALLPATH=/lib/modules/$TARGET_KERNEL/kernel/drivers/video

else

INSTALLPATH=/lib/modules/$TARGET_KERNEL/video

fi

rm -f $INSTALLPATH/NVdriver

 

function error {

echo "$(basename $0): fatal error, $1 failed"

exit 1

}

 

modconf=""

tmp=`mktemp -q /tmp/nvidia.XXXXXX` || error "mktemp"

 

paths[0]=/etc/conf.modules

paths[1]=/etc/modules.conf

paths[2]=/etc/modutils/aliases

 

for path in ${paths[@]}; do

test -f $path && modconf=$path

done

 

if [ ! -z $modconf ]; then

# Initialize the permissions on $tmp so that we don't change the

# permissions of $modconf when we mv it into place

cp -p $modconf $tmp

sed '/^alias.*(NVdriver|nvidia)/d' < $modconf > $tmp

if [ -c /dev/.devfsd ]; then

echo "alias /dev/nvidia* nvidia" >> $tmp

else

echo "alias char-major-195 nvidia" >> $tmp

fi

mv -f $tmp $modconf

test -x /sbin/update-modules && /sbin/update-modules

else

if [ -f /etc/rc.d/rc.modules ]; then

# Initialize the permissions on $tmp so that we don't change the

# permissions of /etc/rc.d/rc.modules when we mv it into place

cp -p /etc/rc.d/rc.modules $tmp

sed '/.*(NVdriver|nvidia).*/d' < /etc/rc.d/rc.modules > $tmp

echo "/sbin/modprobe nvidia" >> $tmp

mv -f $tmp /etc/rc.d/rc.modules

fi

fi

 

if [ ! -c /dev/.devfsd ]; then

for i in 0 1 2 3; do

node="/dev/nvidia$i"

rm -f $node

mknod $node c 195 $i || error "mknod "$node""

chmod 0666 $node || error "chmod "$node""

done

node="/dev/nvidiactl"

rm -f $node

mknod $node c 195 255 || error "mknod "$node""

chmod 0666 $node || error "chmod "$node""

fi

 

# Add it to the module dependency list.

/sbin/depmod -a && /sbin/modprobe nvidia

 

# Check to see if the module is actually loaded

if ! /sbin/lsmod | grep nvidia >& /dev/null ; then

echo

echo "Failed to install nvidia.o"

echo

exit 1

fi

 

# It worked

echo

echo nvidia.o installed successfully

 

perl -pi -e "s|Driver "nv"|Driver "nvidia"|" /etc/X11/XF86Config-4

 

echo XF86Config-4 file updated

 

echo

 

exit 0

 

%preun

 

perl -pi -e "s|Driver "nvidia"|Driver "nv"|" /etc/X11/XF86Config-4

 

echo XF86Config-4 file updated

 

if [ $1 = 0 ] ; then

#

# unload it quietly if already loaded..

#

# Only unload if unused, otherwise this scriptlet fails and

# RPM won't continue with uninstallation. So we also check

# if the usage count is zero.

#

if [ x`/sbin/lsmod | grep nvidia | tr -s " " | cut -f 3 -d " "` == "x0" ]; then

/sbin/rmmod nvidia >& /dev/null

fi

fi

 

 

 

%files -f /tmp/files.lst

 

%changelog

* Wed Jul 30 2003 Torbjorn Turpeinen <tobbe@nyvalls.se>

- Updated to latest version

 

* Sun Jan 05 2003 Joseph Kain <jkain@nvidia.com>

- Correctly preserve permissions on /etc/rc.d/rc.modules.

- Pass "-f" option to mv.

- Fix typos.

 

* Fri Jan 03 2003 Joseph Kain <jkain@nvidia.com>

- Preserve permissions on modules conf file.

 

* Thu Nov 07 2002 Joseph Kain <jkain@nvidia.com>

- Suppress lsmod output

 

* Fri Oct 11 2002 Andy Ritger <aritger@nvidia.com>

- Check that the module is loaded before reporting success.

 

* Fri Sep 06 2002 Gareth Hughes <gareth@nvidia.com>

- Change NVdriver to nvidia.o everywhere.

 

* Mon Nov 26 2001 Joseph Kain <jkain@nvidia.com>

- Use %{_tmppath} instead of hardcoding /var/tmp

 

* Tue Aug 24 2001 Joseph Kain <jkain@nvidia.com>

- in preun check usage count of NVdriver before trying to unload.

 

* Thu Aug 31 2000 Terence Ripperda <tripperda@nvidia.com>

- allow build scripts to specify a kernel version via TARGET_KERNEL variable

otherwise fallback to `uname -r`

- when searching for agpgart module, search in RPM_BUILD_ROOT, so we don't

pick up an errant symbol on a build machine.

 

* Tue Aug 15 2000 Mark Cooke <mpc@star.sr.bham.ac.uk>

- Change permissions on NVdriver to 444 to avoid symbol-stripping issues

with versions of RPM after 3.0.4

- Add a test to make sure the NVdriver registers correctly.

 

* Wed Aug 02 2000 Terence Ripperda <tripperda@nvidia.com>

- regress use of /proc/devices to get major number

 

* Tue May 02 2000 Terence Ripperda <tripperda@nvidia.com>

- regress use of "make install" in post, no Makefile present

 

* Fri Apr 27 2000 Nick Triantos <nick@nvidia.com>

- Use "NVdriver" and "install" targets on make

- Remove BuildReq item from header, was causing build error

 

* Thu Apr 27 2000 Gregory McLean <gregm@comstar.net>

- Cleaned out some redundant parts of the spec file.

- Set the buildroot.

- Made sure the preun gets called properly.

- Added a %clean section to keep building machine tidy.

- Set up logic for epoc builds if needed.

Link to comment
Share on other sites

Ok, weird stuff:

 

I have the same problem description that the topic starter has, but not the same solution...

 

What I did to install the new 4496 drivers (the drivers I was using are the last ones that came with rpm's, forgot the number) was just start with the

sh NVIDIA-Linux-x86-1.0-4496-pkg2.run

(after doing ctrl-alt-F1 to go to console from mdkkdm login manager, logging on as root and telinit 3 to go to text mode only)

It went fine, here's the log:

-> License accepted.

-> There appears to already be a driver installed on your system (version: 1.0-

  4496).  As part of installing this driver (version: 1.0-4496), the existing

  driver will be uninstalled.  Are you sure you want to continue? ('no' will a

  bort installation) (Answer: Yes)

-> A precompiled kernel interface for kernel 'Mandrake Linux 9.1 kernel

  2.4.21-0.13mdk i586' has been found.

  executing: 'cd ./usr/src/nv; /usr/bin/ld -r -o nvidia.o precompiled-nv-linux

  .o nv-kernel.o'...

-> Kernel module linked successfully.

-> Installing classic TLS OpenGL libraries.

-> Parsing log

-> done.

-> Validating previous installation:

-> done.

-> Uninstalling NVIDIA Accelerated Graphics Driver for Linux-x86 (1.0-4496):

-> done.

-> Uninstallation of existing driver: NVIDIA Accelerated Graphics Driver for

  Linux-x86 (1.0-4496) is complete.

-> Installing 'NVIDIA Accelerated Graphics Driver for Linux-x86' (1.0-4496):

  executing: './usr/src/nv/makedevices.sh'...

  executing: '/sbin/ldconfig'...

  executing: '/sbin/depmod -aq'...

-> done.

-> Driver file installation is complete.

-> Running post-install sanity check:

-> done.

-> Sanity check passed.

-> Installation of the NVIDIA Accelerated Graphics Driver for Linux-x86

  (version: 1.0-4496) is now complete.  Please update your XF86Config file as

  appropriate; see the file /usr/share/doc/NVIDIA_GLX-1.0/README for details.

 

and after deinstalling old stuff, installing the new, I could telinit 5 and it would start the login manager again.

 

BUT: after shutting down and rebooting (the next day), I had no X.

After fuzzing about a bit, I just thought, heck, I'll reinstall...

 

And that went fine, the same as before, so I could use X again.

 

Next day: again, no X. Reinstall, all is fine.

 

So for your entertainment, here's the output of lsmod; the Xlogfile (which I also have) just says no screens found...

modules loaded:



Module                  Size  Used by    Tainted: PF

lirc_i2c                4256   1

lirc_dev                9112   1  [lirc_i2c]

bttv                   78496   2

pwcx-gcc32             87552   0  (unused)

pwc                    45064   0  [pwcx-gcc32]

parport_pc             25096   1  (autoclean)

lp                      8096   0  (autoclean)

parport                34176   1  (autoclean) [parport_pc lp]

snd-seq-midi            5024   0  (autoclean) (unused)

snd-emu10k1-synth       5276   0  (autoclean) (unused)

snd-emux-synth         31644   0  (autoclean) [snd-emu10k1-synth]

snd-seq-midi-emul       6780   0  (autoclean) [snd-emux-synth]

snd-seq-virmidi         4572   0  (autoclean) [snd-emux-synth]

snd-seq-oss            31104   0  (unused)

snd-seq-midi-event      5640   0  [snd-seq-midi snd-seq-virmidi snd-seq-oss]

snd-seq                42608   2  [snd-seq-midi snd-emux-synth snd-seq-midi-emul snd-seq-virmidi snd-seq-oss snd-seq-midi-event]

snd-pcm-oss            43556   0  (unused)

snd-mixer-oss          14488   0  [snd-pcm-oss]

snd-emu10k1            69268   0  [snd-emu10k1-synth]

snd-rawmidi            17600   0  [snd-seq-midi snd-seq-virmidi snd-emu10k1]

snd-pcm                77536   0  [snd-pcm-oss snd-emu10k1]

snd-timer              18376   0  [snd-seq snd-pcm]

snd-page-alloc          7732   0  [snd-emu10k1 snd-pcm]

snd-util-mem            3008   0  [snd-emux-synth snd-emu10k1]

snd-ac97-codec         40160   0  [snd-emu10k1]

snd-seq-device          5832   0  [snd-seq-midi snd-emu10k1-synth snd-emux-synth snd-seq-oss snd-seq snd-emu10k1 snd-rawmidi]

snd-hwdep               6368   0  [snd-emu10k1]

snd                    40868   0  [snd-seq-midi snd-emux-synth snd-seq-virmidi snd-seq-oss snd-seq-midi-event snd-seq snd-pcm-oss snd-mixer-oss snd-emu10k1 snd-rawmidi snd-pcm snd-timer snd-util-mem snd-ac97-codec snd-seq-device snd-hwdep]

nfsd                   74256   8  (autoclean)

af_packet              14952   0  (autoclean)

w83781d                21872   0  (unused)

i2c-proc                9072   0  [w83781d]

i2c-viapro              4848   0  (unused)

sr_mod                 16920   0  (autoclean)

floppy                 55132   0

8139too                17160   1  (autoclean)

mii                     3832   0  (autoclean) [8139too]

nls_utf8                1312   3  (autoclean)

nls_cp437               5148   3  (autoclean)

vfat                   11820   3  (autoclean)

fat                    37944   0  (autoclean) [vfat]

supermount             15296   4  (autoclean)

ide-cd                 33856   0

cdrom                  31648   0  [sr_mod ide-cd]

tuner                  11744   1  (autoclean)

tvaudio                15068   0  (autoclean) (unused)

msp3400                17804   1  (autoclean)

i2c-algo-bit            9064   1  [bttv]

i2c-core               21192   0  [lirc_i2c bttv w83781d i2c-proc i2c-viapro tuner tvaudio msp3400 i2c-algo-bit]

soundcore               6276   0  [bttv snd]

ide-scsi               11280   0

scsi_mod              103284   2  [sr_mod ide-scsi]

NVdriver             1667200   0

usbmouse                2936   0  (unused)

keybdev                 2720   0  (unused)

mousedev                5268   0

hid                    20900   0  (unused)

input                   5664   0  [usbmouse keybdev mousedev hid]

videodev                7872   4  [bttv pwc]

printer                 8448   0  (unused)

ehci-hcd               18568   0  (unused)

usb-uhci               24652   0  (unused)

usbcore                72992   1  [pwc usbmouse hid printer ehci-hcd usb-uhci]

rtc                     8060   0  (autoclean)

ext3                   59916   8

jbd                    38972   8  [ext3]

 

I just also tried to reboot after doing a depmod -qa, didn't work...

Then did a modprobe nvidia, which failed (module used.. but not loaded..)

 

No clue on what else to try...

Link to comment
Share on other sites

I tried a bit more yesterday evening, like taking out the /dev/nvidia* part from /etc/security/console.parm in the dri section, chmod 644 /dev/nvidia* and chown root /dev/nvidia,... all to no avail.

 

Anyone any ideas?

 

I must just be missing something small, since reinstalling gets X up and running without any problems...

Link to comment
Share on other sites

@ aRTee,

Am still new in linux, here's my 2 cents nonetheless... :)

 

In your XF86Config-4 file, under the Subsection "Display", Depth xx, does it say Modes "default" ? if yes, please change to Modes "1024x768" "800x600" "640x480" (or any screen resolution you prefer).

 

In your /var/log/XFree86.0.log , is there an (EE) NVIDIA (0) failed to initialize the NVIDIA kernel module ?

Since your sanity checked has passed, the module is just not being loaded. Verify first that "alias /dev/nvidia* nvidia" (without quote) is in the /etc/modules.conf before doing "/sbin/modprobe nvidia" (without quote) then "/sbin/depmod -ae" (without quote). it should be in sequence, modprobe first to load the module then depmod to handle the dependencies. startx, logout, reboot, and hopefully X loads .....

 

that's what i can think of so far ....

 

(me goes back to installing kyro2 linux drivers (again) - MESA loads in user but powervr automaticaly loads in su, hmmm ...)

Link to comment
Share on other sites

bvc, a new user: haven't tried, but not likely to help.

Here's why: after installing the drivers (again), I can start mdkkdm as root, whereas before I can't. Since it's as root, a new user would not make a difference, mdkkdm should run as root.

Also, as root before I cannot start X by doing:

X

but after reinstalling I can...

There must just be a small setting somewhere..

 

 

James, thanks for thinking with me.

 

/etc/modules.conf does have

alias /dev/nvidia* nvidia

and I did do

modprobe nvidia

which gives me an error as mentioned above. So I get stuck there.

 

On the modes: I am using quite a few modelines; and no I am not using 'default' but specified resolutions. In the logfile it clearly shows that XFree never gets around to trying the modes, since it complains about no screens found.

 

I guess this may have to do with module nvidia not being able to load...

 

 

This last part is where it goes wrong; why would this module not be loaded but be 'in use' / not loadable, and how can I see which process is trying to do that? Is there any way I may try to force the loading?

Link to comment
Share on other sites

Join the conversation

You can post now and register later. If you have an account, sign in now to post with your account.
Note: Your post will require moderator approval before it will be visible.

Guest
Reply to this topic...

×   Pasted as rich text.   Paste as plain text instead

  Only 75 emoji are allowed.

×   Your link has been automatically embedded.   Display as a link instead

×   Your previous content has been restored.   Clear editor

×   You cannot paste images directly. Upload or insert images from URL.

Loading...
 Share


×
×
  • Create New...