LinuxQuestions.org
Help answer threads with 0 replies.
Go Back   LinuxQuestions.org > Forums > Linux Forums > Linux - General
User Name
Password
Linux - General This Linux forum is for general Linux questions and discussion.
If it is Linux Related and doesn't seem to fit in any other forum then this is the place.

Notices


Reply
  Search this Thread
Old 10-26-2002, 05:37 AM   #1
mikefi
LQ Newbie
 
Registered: Oct 2002
Posts: 8

Rep: Reputation: 0
Question NVIDIA Driver installation on RH8.0 problems.


Hi,

Hope someone can help. I've seen similar problems elsewhere on the forum (with installing nForce drivers) - but this one is slightly different.

I've downloaded the NVIDIA_GLX-1.0-3123.src.rpm, and NVIDIA_kernel-1.0-3123.src.rpm files. Having read the instructions (and installed once before on 7.3), I ran the rpmbuild --rebuild on the GLX file and it correctly generated the installable rpm in /usr/src/redhat.... yada yada yada. When I do the same with the kernel rpm, I get:

"Bad exit status from /var/tmp/rpm-tmp.65010 (%install)" - and no installable rpm is generated. I won't paste the whole error message here (there are pages of it!) but looking back through it all, the place where it all seems to go wrong is:

<paste error output excerpt>

++ echo 4.1
++ awk -F . '{print $3;}'
+ RPMPATCH=
+ '[' 4 -le 3 -a 1 -le 0 -a -lt 4 ']'
/var/tmp/rpm-tmp.85415: line 34: [: too many arguments
+ make NVdriver
echo \#define NV_COMPILER \"`cc -v 2>&1 | tail -1`\" > nv_compiler.h
cc -c -Wall -Wimplicit -Wreturn-type -Wswitch -Wformat -Wchar-subscripts -Wparentheses -Wpointer-arith -Wcast-qual -Wno-multichar -O -MD -D__KERNEL__ -DMODULE
-D_LOOSE_KERNEL_NAMES -DNTRM -D_GNU_SOURCE -DRM_HEAPMGR -D_LOOSE_KERNEL_NAMES -D__KERNEL__ -DMODULE -DNV_MAJOR_VERSION=1 -DNV_MINOR_VERSION=0 -DNV_PATCHLEVEL=3123 -DNV_UNIX -DNV_LINUX -DNVCPU_X86 -I. -I/lib/modules/2.4.18-17.8.0/build/include -Wno-cast-qual nv.c
In file included from nv-linux.h:27,
from nv.c:14:
/lib/modules/2.4.18-17.8.0/build/include/linux/kernel.h:60: invalid suffix on integer constant
/lib/modules/2.4.18-17.8.0/build/include/linux/kernel.h:60: parse error before numeric constant

</paste>

Can anyone make sense of this for me?

Thanks in advance,
Mike.
 
Old 10-26-2002, 06:08 AM   #2
jdc2048
Member
 
Registered: Jul 2002
Location: Ohio, USA
Distribution: Redhat, Gentoo, Solaris, HP-UX, etc...
Posts: 391

Rep: Reputation: 30
What distro (rh8, rh7.3)?, what version gcc?

I have succesfully compiled the NVIDIA_kernel-1.0-3123.src.rpm on RH8.0.

I downloaded the precompiled NVIDIA_GLX-1.0-3123.i386.rpm from nvidia.

It worked easily. The only problem I ran into was when I tried to compile 1.0-2880 on RH8, then it complained of a gcc version conflict.

Jeremiah
 
Old 10-26-2002, 10:31 AM   #3
mikefi
LQ Newbie
 
Registered: Oct 2002
Posts: 8

Original Poster
Rep: Reputation: 0
Hi Jeremiah,

The problem is on Redhat 8.0.

The machine is a clean install (not an upgrade from an earlier version), and will therefore be running whichever version of gcc that is distributed with 8.0.

The only slightly non-standard aspect of this PC, is that the "redhat network" client auto-installed a later kernel this morning. The kernel is now 2.4.18-17. I don't know if that has any bearing on the issue, because I never tried to compile these drivers before installing the later kernel.

Thanks for taking the time to reply,
Mike.
 
Old 10-27-2002, 08:09 AM   #4
jdc2048
Member
 
Registered: Jul 2002
Location: Ohio, USA
Distribution: Redhat, Gentoo, Solaris, HP-UX, etc...
Posts: 391

Rep: Reputation: 30
That does present an interesting twist. I am still running the stock 2.4.18-14 kernel on my install.

The gcc version that came stock is 3.2.7.

The error seems to state a file located in the 2.4.18.17 kernel.h file so I would have to download and mess with that kernel to figure out what was going on.

I will try to look at it sometime this week and get back to you. Until then maybe someone else could chip in.

Jeremiah
 
Old 10-27-2002, 08:17 AM   #5
mikefi
LQ Newbie
 
Registered: Oct 2002
Posts: 8

Original Poster
Rep: Reputation: 0
Hmmm - I still have the earlier kernel on the machine, so as soon as my currentl download finishes - I'll reboot using that one and then see if I can compile it.

Watch this space....
 
Old 10-27-2002, 09:31 AM   #6
mikefi
LQ Newbie
 
Registered: Oct 2002
Posts: 8

Original Poster
Rep: Reputation: 0
Thumbs down



Just tried it. Didn't work.



Just to confirm, "uname -r" reports "2.4.18-14"

Note in the following pasted output, that the rpmbuild appears to be using the correct directories when looking for the includes for this kernel. Also note the line I've singled out for attention with "^^^^^" characters.

Anyone?

I get the following when building the rpm:

rpmbuild --rebuild NVIDIA_kernel-1.0-3123.src.rpm

Installing NVIDIA_kernel-1.0-3123.src.rpm
warning: user buildmeister does not exist - using root
warning: group buildmeister does not exist - using root
warning: user buildmeister does not exist - using root
warning: group buildmeister does not exist - using root
Executing(%prep): /bin/sh -e /var/tmp/rpm-tmp.79853
+ umask 022
+ cd /usr/src/redhat/BUILD
+ LANG=C
+ export LANG
+ cd /usr/src/redhat/BUILD
+ rm -rf NVIDIA_kernel-1.0-3123
+ /usr/bin/gzip -dc /usr/src/redhat/SOURCES/NVIDIA_kernel-1.0-3123.tar.gz
+ tar -xf -
+ STATUS=0
+ '[' 0 -ne 0 ']'
+ cd NVIDIA_kernel-1.0-3123
++ /usr/bin/id -u
+ '[' 0 = 0 ']'
+ /bin/chown -Rhf root .
++ /usr/bin/id -u
+ '[' 0 = 0 ']'
+ /bin/chgrp -Rhf root .
+ /bin/chmod -Rf a+rX,g-w,o-w .
+ exit 0
Executing(%build): /bin/sh -e /var/tmp/rpm-tmp.79546
+ umask 022
+ cd /usr/src/redhat/BUILD
+ cd NVIDIA_kernel-1.0-3123
+ LANG=C
+ export LANG
+ set +e
++ rpm -q --queryformat '%{VERSION}' rpm
+ RPMVERSION=4.1
++ echo 4.1
++ awk -F . '{print $1;}'
+ RPMMAJOR=4
++ echo 4.1
++ awk -F . '{print $2;}'
+ RPMMINOR=1
++ echo 4.1
++ awk -F . '{print $3;}'
+ RPMPATCH=
+ '[' 4 -le 3 -a 1 -le 0 -a -lt 4 ']'

/var/tmp/rpm-tmp.79546: line 34: [: too many arguments
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
{does this look suspicious to you?}

+ make NVdriver
echo \#define NV_COMPILER \"`cc -v 2>&1 | tail -1`\" > nv_compiler.h
cc -c -Wall -Wimplicit -Wreturn-type -Wswitch -Wformat -Wchar-subscripts -Wparentheses -Wpointer-arith -Wcast-qual -Wno-multichar -O -MD -D__KERNEL__ -DMODULE
-D_LOOSE_KERNEL_NAMES -DNTRM -D_GNU_SOURCE -DRM_HEAPMGR -D_LOOSE_KERNEL_NAMES -D__KERNEL__ -DMODULE -DNV_MAJOR_VERSION=1 -DNV_MINOR_VERSION=0 -DNV_PATCHLEVEL=3123 -DNV_UNIX -DNV_LINUX -DNVCPU_X86 -I. -I/lib/modules/2.4.18-14/build/include -Wno-cast-qual nv.c
In file included from /usr/include/linux/tqueue.h:18,
from nv-linux.h:54,
from nv.c:14:
/usr/include/asm/bitops.h:333:2: warning: #warning This includefile is not available on all architectures.
/usr/include/asm/bitops.h:334:2: warning: #warning Using kernel headers in userspace.
In file included from nv-linux.h:54,
from nv.c:14:
/usr/include/linux/tqueue.h:94: parse error before "tqueue_lock"
In file included from nv.c:14:
nv-linux.h:57:75: linux/vmalloc.h: No such file or directory
In file included from /usr/include/linux/fs.h:23,
 
Old 10-27-2002, 10:25 AM   #7
Mara
Moderator
 
Registered: Feb 2002
Location: Grenoble
Distribution: Debian
Posts: 9,696

Rep: Reputation: 232Reputation: 232Reputation: 232
Do you have package kernel-headers installed? I mean the right version, the same as your kernel...
 
Old 10-27-2002, 11:43 AM   #8
mikefi
LQ Newbie
 
Registered: Oct 2002
Posts: 8

Original Poster
Rep: Reputation: 0
Funny you should ask that...

When I went through the excellent thread at http://www.nvnews.net/vbulletin/show...&threadid=2563 , I noticed that you have to have the kernel headers package installed.

So, I did the time-honoured rpm -qa and got this:

rpm -qa | grep head
glibc-kernheaders-2.4-7.20

Hmmm, thought I - that doesn't look right. So I tried to hunt down the file on redhats ftp site - only to draw a blank. The only version I could find on their site was the same as the one I already have.

I (incorrectly by the sounds of it) assumed that meant it couldn't be the problem and started looking elsewhere... (here).
 
Old 10-27-2002, 11:59 AM   #9
unSpawn
Moderator
 
Registered: May 2001
Posts: 29,415
Blog Entries: 55

Rep: Reputation: 3594Reputation: 3594Reputation: 3594Reputation: 3594Reputation: 3594Reputation: 3594Reputation: 3594Reputation: 3594Reputation: 3594Reputation: 3594Reputation: 3594
Most recent is kernel-headers-2.4.9-34 rpm, get it tru the RHN if you can't find it tru rpm.pbone.net or rpmfind.net
 
Old 10-30-2002, 01:37 PM   #10
jdc2048
Member
 
Registered: Jul 2002
Location: Ohio, USA
Distribution: Redhat, Gentoo, Solaris, HP-UX, etc...
Posts: 391

Rep: Reputation: 30
I would also check which version of kernel-source you have installed. Do a "rpm -qa | grep kernel" and post results.

Jeremiah
 
Old 10-31-2002, 06:02 AM   #11
Amerist
Member
 
Registered: Apr 2002
Location: Germany
Distribution: RedHat & SuSE
Posts: 106

Rep: Reputation: 15
Nvidia drivers work just fine on Redhat 8...

Here's a peice of advice. Get the .tar.gz source files for the NVIDIA_GLX And for the NVIDIA_kernel drivers and then install them according to the

NVlinuxR25notes_3123.pdf

If you follow the directions exactly, you shouldn't have any problems.

Additional info:

I just upgraded my kernel as probably many other people with redhat 8 have done, and of course when you update the kernel, you have to re-install this driver along with any other kernel drivers you have. It's a pain in the ass, but necessary nonetheless.

Download and install the kernel source from the redhat up2date applet.
Use a command prompt to type whereis kernel

the path will tell you where the kernel is installed, but you need the path to the includes, so if your path looks like this:

/usr/src/linux-2.4.18-17.8.0/kernel

then the include path is:

/usr/src/linux-2.4.18-17.8.0/include

Install the NVIDIA_GLX drivers according to the PDF

Then in a terminal, go to the directory where the source for the NVIDIA_kernel driver is extracted and type:

make -SYSINCLUDE=/usr/src/linux-2.4.18-17.8.0/include

Now you should be able to boot the new kernel and startx
 
Old 11-02-2002, 02:17 PM   #12
mikefi
LQ Newbie
 
Registered: Oct 2002
Posts: 8

Original Poster
Rep: Reputation: 0


Did as you suggested, still got major problems.

See output:

[root@mike NVIDIA_kernel-1.0-3123]# make SYSINCLUDE=/usr/src/linux-2.4.18-17.8.0/include
cc -c -Wall -Wimplicit -Wreturn-type -Wswitch -Wformat -Wchar-subscripts -Wparentheses -Wpointer-arith -Wcast-qual -Wno-multichar -O -MD -D__KERNEL__ -DMODULE
-D_LOOSE_KERNEL_NAMES -DNTRM -D_GNU_SOURCE -DRM_HEAPMGR -D_LOOSE_KERNEL_NAMES -D__KERNEL__ -DMODULE -DNV_MAJOR_VERSION=1 -DNV_MINOR_VERSION=0 -DNV_PATCHLEVEL=3123 -DNV_UNIX -DNV_LINUX -DNVCPU_X86 -I. -I/usr/src/linux-2.4.18-17.8.0/include -Wno-cast-qual nv.c
In file included from nv-linux.h:27,
from nv.c:14:
/usr/src/linux-2.4.18-17.8.0/include/linux/kernel.h:60: invalid suffix on integer constant
/usr/src/linux-2.4.18-17.8.0/include/linux/kernel.h:60: parse error before numeric constant
/usr/src/linux-2.4.18-17.8.0/include/linux/kernel.h:61: `panic_R_ver_str' declared as function returning a function
/usr/src/linux-2.4.18-17.8.0/include/linux/kernel.h:67: parse error before numeric constant
/usr/src/linux-2.4.18-17.8.0/include/linux/kernel.h:67: `simple_strtoul_R_ver_str' declared as function returning a function
/usr/src/linux-2.4.18-17.8.0/include/linux/kernel.h:68: invalid suffix on integer constant
/usr/src/linux-2.4.18-17.8.0/include/linux/kernel.h:68: parse error before numeric constant
/usr/src/linux-2.4.18-17.8.0/include/linux/kernel.h:68: `simple_strtol_R_ver_str' declared as function returning a function
/usr/src/linux-2.4.18-17.8.0/include/linux/kernel.h:69: invalid suffix on integer constant
/usr/src/linux-2.4.18-17.8.0/include/linux/kernel.h:69: parse error before numeric constant
/usr/src/linux-2.4.18-17.8.0/include/linux/kernel.h:69: `simple_strtoull_R_ver_str' declared as function returning a function
/usr/src/linux-2.4.18-17.8.0/include/linux/kernel.h:71: invalid suffix on integer constant
/usr/src/linux-2.4.18-17.8.0/include/linux/kernel.h:71: parse error before numeric constant
/usr/src/linux-2.4.18-17.8.0/include/linux/kernel.h:72: `sprintf_R_ver_str' declared as function returning a function
/usr/src/linux-2.4.18-17.8.0/include/linux/kernel.h:73: invalid suffix on integer constant
/usr/src/linux-2.4.18-17.8.0/include/linux/kernel.h:73: parse error before numeric constant
/usr/src/linux-2.4.18-17.8.0/include/linux/kernel.h:73: `vsprintf_R_ver_str' declared as function returning a function
/usr/src/linux-2.4.18-17.8.0/include/linux/kernel.h:74: invalid suffix on integer constant
/usr/src/linux-2.4.18-17.8.0/include/linux/kernel.h:74: parse error before numeric constant


And loads more of this until:

nv.c:2831: warning: implicit declaration of function `inter_module_get_Rf6a0ce24'
nv.c:2832: warning: implicit declaration of function `inter_module_put_R6b99f7d8'
make: *** [nv.o] Error 1

This looks no better to me...
 
Old 11-02-2002, 06:42 PM   #13
jdc2048
Member
 
Registered: Jul 2002
Location: Ohio, USA
Distribution: Redhat, Gentoo, Solaris, HP-UX, etc...
Posts: 391

Rep: Reputation: 30
DOH!

Just noticed in your post that you stated that

Quote:
Just to confirm, "uname -r" reports "2.4.18-14"
And also

Quote:
/usr/src/linux-2.4.18-17.8.0/include/linux/kernel.h:60: invalid suffix on integer constant
I think the RHN has a default that says that when you download a kernel update, that you don't install it.

It appears to me you are attempting to build against a kernel that is not running currently. Get your system running the 2.4.18-17 kernel first (and hint, change your /etc/inittab to default runlevel 3 instead of 5 before you reboot). Then when the uname -r tells you 2.4.18-17 try building agains the newer source dir (/usr/src/linux-2.4.18-17.8.0).

I just downloaded the kernel-2.4.18-17.8.0 i686.rpm and src.rpm from Redhat (ftp style) and installed the kernel from the i686.rpm, then installed the src.rpm (to /usr/src/linux-2.4.18-17.8.0) and then compiled the NVIDIA_kernel-1.0.3123.src.rpm using the rpmbuild method. I had no problems whatsoever.

HTH
Jeremiah
 
Old 11-03-2002, 12:51 AM   #14
mikefi
LQ Newbie
 
Registered: Oct 2002
Posts: 8

Original Poster
Rep: Reputation: 0
Quote:
HTH
Sorry Jeremiah - it doesn't.

You've extracted quotes from two different posts - missing one vital element.

If you look at my post previous to the one stating:

Quote:
Just to confirm, "uname -r" reports "2.4.18-14"
You'll see:

Quote:
Hmmm - I still have the earlier kernel on the machine, so as soon as my currentl download finishes - I'll reboot using that one and then see if I can compile it.
I was attempting to compile, whilst running the non-updated kernel to see if it was the updated one causing the problem.

Just to clear things up, other than that one attempt at compiling on the earlier kernel - all other attempts have been using the 2.4.18-17.8.0 version.

Given the number of people stating "I compiled it with no issues", and the fact that I seem to be getting the same sort of errors regardless of whether I use the .tar or the rpmbuild method - I think it points to some other unrelated problem with this install.

I'm beginning to consider winbloze methods, ie. fall back to reinstalling from scratch...

 
Old 11-03-2002, 12:58 AM   #15
mikefi
LQ Newbie
 
Registered: Oct 2002
Posts: 8

Original Poster
Rep: Reputation: 0
Doh!

When will I learn to proof read before clicking "POST"!?!?!

Quote:
some other unrelated problem
.... means of course unrelated to what we've been looking at so far, rather than unrelated to the reason why I can't compile the kernel drivers....
 
  


Reply


Thread Tools Search this Thread
Search this Thread:

Advanced Search

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off



Similar Threads
Thread Thread Starter Forum Replies Last Post
Display Driver(NVIDIA) crashes on boot up (RH8.0) sulabhv Linux - Laptop and Netbook 7 08-30-2004 01:53 PM
Trouble with new nVidia driver installation...no nvidia logo and KDE no start up mdb Linux - Software 8 08-18-2004 09:45 AM
Nvidia driver.. Installation problems deft Linux - Software 2 03-31-2004 05:47 PM
Nvidia RH8.0 driver problem on Dell 8100 nhasbek Linux - Hardware 1 12-13-2002 06:39 AM
Nvidia RH8 driver problem nhasbek Linux - General 1 12-13-2002 05:45 AM

LinuxQuestions.org > Forums > Linux Forums > Linux - General

All times are GMT -5. The time now is 01:30 AM.

Main Menu
Advertisement
My LQ
Write for LQ
LinuxQuestions.org is looking for people interested in writing Editorials, Articles, Reviews, and more. If you'd like to contribute content, let us know.
Main Menu
Syndicate
RSS1  Latest Threads
RSS1  LQ News
Twitter: @linuxquestions
Open Source Consulting | Domain Registration