Hello, List:
I am hoping someone can help me install these Nvidia
drivers. I am trying to install
NVIDIA_kernel-0.9-6.tar.gz and
NVIDIA_GLX-0.9-6.tar.gz.
My system is the following:
Pentium III 866MHz
384 Megs PC-133 RAM
Hercules Prophet II Geforce 2 GTS w/ 64megs of DDR
Asus CUSL2 Motherboard (Intel 815E chipset)
Soundblaster Live! Platinum
I download the drivers from the nvidia.com website to
my /tmp directory and decompress them with tar -xvzf.
I then make a change to my os-registry.c as per
http://www.evil3d.net/articles/linux/nv95v96_2/ This
is a perfermance bug fix and should not affect my
installation.
Before install I removed the Mesa rpm using rpm -e
because Mesa is not compatible with the nvidia
drivers.
I have tried installing these nvidia drivers probably
30 times. Several times I have tried from a clean
install of Mandrake 7.2. Currently I am trying to
install these drivers on Mandrake 7.2 with all of the
updates and KDE 2.1 beta 1.
I have followed the instructions on linuxnewbie.org
and mandrakeuser.org (both short and long) as best as
I can. I have also followed the Nvidia
instructions/FAQ thoroughly. However, all of these
(including nvidia) seem to be written for the old 9-5
drivers.
After decompressing and applying the performance fix
as mentioned above, I first go to the
NVIDIA_kernel-0.9-6 directory. I had to make
/usr/X11R6/lib/modules/extensions and
/usr/X11R6/lib/modules/drivers directories because
they were needed for the driver install, but did not
previously exist. As root I have tried compiling with
make, make install, and make
SYSINCLUDE=/usr/src/linux/include. Sometimes after
the compile I try running sh ./makedevices.sh. These
all SEEM to end with the same results.
Following is the output from make (please notice the
warnings):
[root@localhost NVIDIA_kernel-0.9-6]# make
rm -f nv.o os-interface.o os-registry.o Module-linux
NVdriver
cc -c -Wall -Wunknown-pragmas -Wno-multichar -O
-D__KERNEL__ -DMODULE -D_LOOSE_KERNEL_NAMES -DUNIX
-DLINUX -DNV4_HW -DNTRM -DRM20 -D_X86_=1 -Di386=1
-D_GNU_SOURCE -DRM_HEAPMGR -D_LOOSE_KERNEL_NAMES
-I. -I/usr/src/linux/include nv.c
/tmp/ccvvfRE1.s: Assembler messages:
/tmp/ccvvfRE1.s:9: Warning: Ignoring changed section
attributes for .modinfo
cc -c -Wall -Wunknown-pragmas -Wno-multichar -O
-D__KERNEL__ -DMODULE -D_LOOSE_KERNEL_NAMES -DUNIX
-DLINUX -DNV4_HW -DNTRM -DRM20 -D_X86_=1 -Di386=1
-D_GNU_SOURCE -DRM_HEAPMGR -D_LOOSE_KERNEL_NAMES
-I. -I/usr/src/linux/include os-interface.c
cc -c -Wall -Wunknown-pragmas -Wno-multichar -O
-D__KERNEL__ -DMODULE -D_LOOSE_KERNEL_NAMES -DUNIX
-DLINUX -DNV4_HW -DNTRM -DRM20 -D_X86_=1 -Di386=1
-D_GNU_SOURCE -DRM_HEAPMGR -D_LOOSE_KERNEL_NAMES
-I. -I/usr/src/linux/include os-registry.c
/tmp/ccgZ6UM6.s: Assembler messages:
/tmp/ccgZ6UM6.s:16: Warning: Ignoring changed section
attributes for .modinfo
ld -r -o Module-linux nv.o os-interface.o
os-registry.o
ld -r -o NVdriver Module-linux Module-nvkernel
size NVdriver
text data bss dec hex filename
444879 28160 19528 492567 78417 NVdriver
NVdriver installed successfully.
[root@localhost NVIDIA_kernel-0.9-6]#
I then cd ../NVIDIA_GLX-0.9-6 where I do another make.
This results in the following with a new error
(earlier I did not have this error, but still had the
same situation [meaning GLX was not loaded]):
[root@localhost NVIDIA_GLX-0.9-6]# make
Removing all old and conflicting files!
rm -f //usr/lib/libGL.*
rm -f //usr/lib/libGLcore.*
rm -f
//usr/X11R6/lib/modules/extensions/libGLcore.*
rm -f
//usr/X11R6/lib/modules/extensions/libglx.*
Installing new drivers
install libGL.so.1.0.6 //usr/lib
install libGLcore.so.1.0.6 //usr/lib
install nvidia_drv.o
//usr/X11R6/lib/modules/drivers
install libglx.so.1.0.6
//usr/X11R6/lib/modules/extensions
/sbin/ldconfig
/sbin/ldconfig: warning: can't open
/usr/X11R6/lib/libGL.so.1 (No such file or directory),
skipping
cd //usr/lib/; ln -fs libGL.so.1.0.6 libGL.so
cd //usr/X11R6/lib/modules/extensions; \
ln -fs libglx.so.1.0.6 libglx.so
[root@localhost NVIDIA_GLX-0.9-6]#
At this point (after the completion of the GLX make)
running ldconfig seems to work okay.
After a reboot, I do not have the NVdriver module
loaded. Loading this manually with /sbin/insmod
NVdriver seems to work, but lsmod gives the following
peculiar results (notice the unused):
[root@localhost NVIDIA_GLX-0.9-6]# lsmod
Module Size Used by
NVdriver 608976 0 (unused)
tulip 33392 1 (autoclean)
pci-scan 2624 0 (autoclean) [tulip]
emu10k1 42480 0
soundcore 2800 4 [emu10k1]
nls_cp437 3952 2 (autoclean)
vfat 9408 1 (autoclean)
fat 30432 1 (autoclean) [vfat]
supermount 14224 3 (autoclean)
ide-scsi 7664 0
[root@localhost NVIDIA_GLX-0.9-6]#
The nvidia FAQ suggests viewing the results of
xdpyinfo and looking for glx. An xpdyinfo > test
reveals no glx or GLX in my test file.
Nvidia suggests that failing to find glx may mean that
my XF86Config (I think they meant XF86Config-4) needs
the Load "glx". However, it has that. The following
is a portion of my /etc/X11/XF86Config-4:
Section "Module"
# This loads the GLX module
Load "glx"
# This loads the DBE extension module.
Load "dbe"
# This loads the miscellaneous extensions module, and
disables
# initialisation of the XFree86-DGA extension within
that module.
SubSection "extmod"
#Option "omit xfree86-dga"
EndSubSection
# This loads the Type1 and FreeType font modules
Load "type1"
Load "freetype"
EndSection
I have tried commenting out the subsection as I had
seen in some help on the web. This made no difference
that I am aware of. Incidentally, I did change the
"nv" to "nvidia" in the driver portion.
The nvidia FAQ also suggests checking
/var/log/XFree86.0.log for information about why the
GLX extension did not load. However, this log file
does not seem to exist on my machine. :(
Finally, when trying to run Unreal Tournament
(installed with ut-install-436.run from
http://www.lokigames.com) I receive the following
error:
Input system initialized for SDLViewport0
Opening SDL viewport.
Bound to SDLGLDrv.so
Loaded render device class.
Initializing SDLGLDrv...
binding libGL.so.1
Resizing SDL viewport. X: 640 Y: 480
OpenGL
Xlib: extension "GLX" missing on display ":0.0".
Xlib: extension "GLX" missing on display ":0.0".
appError called:
Couldn't set video mode: Couldn't find matching GLX
visual
Executing UObject::StaticShutdownAfterError
Executing USDLClient::ShutdownAfterError
Executing USDLViewport::ShutdownAfterError
USDLGLRenderDevice::ShutdownAfterError
Signal: SIGIOT [iot trap]
Aborting.
Exiting.
Name subsystem shut down
[root@localhost NVIDIA_GLX-0.9-6]#
I have been working on this for quite some time and
have been trying to research all avenues I could find.
I did try building rpms from the src.rpm at
nvidia.com. However, this did not work either (and
does not allow me to execute the performance fix). I
also tried installing the 7.1 binary .rpms which did
not work. Help here would be greatly appreciated.
Thank you for your efforts.
__________________________________________________
Get personalized email addresses from Yahoo! Mail - only $35
a year! http://personal.mail.yahoo.com/
