Nvidia drivers

[Larsson]

I have a Geforce 2 GTS that i am trying to install.

This card is classfied by Nvidia as legacy card, from Nvida:

 

These GPUs will continue to be maintained through the special legacy NVIDIA GPU driver releases.

 

I have not been able to locate any special driver packs for those so i took which i think is the latest driver pack that support my card -> 1.0-6629

 

However when i try to install them after following the instrucion as good as i could, installing the kernels - > http://www.mandrake.tips.4.free.fr/nvidiainst.html

 

I got a message telling me somewhat close to this (taken from log cause i am to lazy to get the exact words but the meaning is the same)

 

The rivafb driver conflicts with the NVIDIA driver; the

NVIDIA kernel module will still be built and installed,

but be aware that the NVIDIA driver will not be able to

function properly if the rivafb module is loaded!

 

After the process of making a kernel interface i got errors that it failed and that i should check the log.

 

*** Failed rivafb module sanity check, but continuing! ***

 

...

 

[make[4]: *** [/root/tmp/selfgz1860/NVIDIA-Linux-x86-1.0-6629-pkg1/usr/src/nv

/nv.o] Fel 1

make[3]: *** [_module_/root/tmp/selfgz1860/NVIDIA-Linux-x86-1.0-6629-pkg1/us

r/src/nv] Fel 2

make[2]: *** [modules] Fel 2

NVIDIA: left KBUILD.

nvidia.ko failed to build!

make[1]: *** [module] Fel 1

make: *** [module] Fel 2

Error.

 

Please help me, and remember than i am a newbie so make detailed description of what i should do. Also i could mention that i got this card working for Suse 9.3 by using Yast so it is not impossible but then i might got the wrong drivers that i try to install.

 

One last thing, could someone tell me what rivafb is and what it does.

[tyme]

rivafb is a module for a different video card that is a part of the mandrake kernel install, as long as you don't load it (and you won't with an nvidia card) it won't cause any problems, so it's safe to ignore that error.

 

please try to post abit more of the error messages you got, what you've posted doesn't appear to point the the problem. it may be further up in the output.

[Larsson]

My log, tryied to add it as an attachment to save room but it was apparently not allowed to add .log.

 

nvidia-installer log file '/var/log/nvidia-installer.log'

creation time: Fri Aug 12 23:15:01 2005

 

option status:

  license pre-accepted    : false

  update                  : false

  force update            : false

  expert                  : false

  uninstall              : false

  driver info            : false

  no precompiled interface: false

  no ncurses color        : false

  query latest driver ver : false

  OpenGL header files    : false

  no questions            : false

  silent                  : false

  no backup              : false

  kernel module only      : false

  sanity                  : false

  add this kernel        : false

  no runlevel check      : false

  no network              : false

  no ABI note            : false

  no RPMs                : false

  force tls              : (not specified)

  force compat32 tls      : (not specified)

  X install prefix        : /usr/X11R6

  OpenGL install prefix  : /usr

  compat32 install prefix : (not specified)

  installer install prefix: /usr

  utility install prefix  : /usr

  kernel name            : (not specified)

  kernel include path    : (not specified)

  kernel source path      : (not specified)

  kernel output path      : (not specified)

  kernel install path    : (not specified)

  proc mount point        : /proc

  ui                      : (not specified)

  tmpdir                  : /root/tmp

  ftp mirror              : ftp://download.nvidia.com

  RPM file list          : (not specified)

 

Using: nvidia-installer ncurses user interface

-> License accepted.

-> No precompiled kernel interface was found to match your kernel; would you li

  ke the installer to attempt to download a kernel interface for your kernel f

  rom the NVIDIA ftp site (ftp://download.nvidia.com)? (Answer: No)

-> No precompiled kernel interface was found to match your kernel; this means

  that the installer will need to compile a new kernel interface.

-> Kernel source path: '/usr/src/linux'

-> Performing CC test with CC="cc".

-> Performing rivafb check.

-> Performing rivafb module check.

WARNING: Your kernel was configured to include rivafb support as

        a loadable kernel module.

       

        The rivafb driver conflicts with the NVIDIA driver; the

        NVIDIA kernel module will still be built and installed,

        but be aware that the NVIDIA driver will not be able to

        function properly if the rivafb module is loaded!

-> Cleaning kernel module build directory.

  executing: 'cd ./usr/src/nv; make clean'...

  rm -f -f nv.o nv-vm.o os-agp.o os-interface.o os-registry.o nv.o nv-vm.o os-

  agp.o os-interface.o os-registry.o nvidia.mod.o

  rm -f -f build-in.o nv-linux.o *.d .*.{cmd,flags}

  rm -f -f nvidia.{o,ko,mod.{o,c}} nv_compiler.h *~

  rm -f -rf .tmp_versions

-> Building kernel module:

  executing: 'cd ./usr/src/nv; make module SYSSRC=/usr/src/linux SYSOUT=/usr/s

  rc/linux'...

  Your kernel was configured to include rivafb support as

  a loadable kernel module.

 

  The rivafb driver conflicts with the NVIDIA driver; the

  NVIDIA kernel module will still be built and installed,

  but be aware that the NVIDIA driver will not be able to

  function properly if the rivafb module is loaded!

 

  *** Failed rivafb module sanity check, but continuing! ***

 

 

  NVIDIA: calling KBUILD...

  make CC=cc  KBUILD_VERBOSE=1 -C /usr/src/linux SUBDIRS=/root/tmp/selfgz7292/

  NVIDIA-Linux-x86-1.0-6629-pkg1/usr/src/nv modules

  mkdir -p /root/tmp/selfgz7292/NVIDIA-Linux-x86-1.0-6629-pkg1/usr/src/nv/.tmp

  _versions

  make -f scripts/Makefile.build obj=/root/tmp/selfgz7292/NVIDIA-Linux-x86-1.0

  -6629-pkg1/usr/src/nv

  echo \#define NV_COMPILER \"`cc -v 2>&1 | tail -n 1`\" > /root/tmp/selfgz729

  2/NVIDIA-Linux-x86-1.0-6629-pkg1/usr/src/nv/nv_compiler.h

    cc -Wp,-MD,/root/tmp/selfgz7292/NVIDIA-Linux-x86-1.0-6629-pkg1/usr/src/nv/

  .nv.o.d -nostdinc -isystem /usr/lib/gcc/i586-mandrake-linux-gnu/3.4.3/includ

  e -D__KERNEL__ -Iinclude  -Wall -Wstrict-prototypes -Wno-trigraphs -fno-stri

  ct-aliasing -fno-common -ffreestanding -O2    -fno-omit-frame-pointer -pipe

    -mpreferred-stack-boundary=2 -fno-unit-at-a-time -march=i686 -Iinclude/asm-

  i386/mach-default -Wdeclaration-after-statement  -I/root/tmp/selfgz7292/NVID

  IA-Linux-x86-1.0-6629-pkg1/usr/src/nv -Wall -Wimplicit -Wreturn-type -Wswitc

  h -Wformat -Wchar-subscripts -Wparentheses -Wpointer-arith  -Wno-multichar 

  -Werror -O -fno-common -MD  -Wno-cast-qual -Wno-error -D_LOOSE_KERNEL_NAMES

  -D__KERNEL__ -DMODULE  -DNTRM -D_GNU_SOURCE -D_LOOSE_KERNEL_NAMES -D__KERNEL

  __ -DMODULE  -DNV_MAJOR_VERSION=1 -DNV_MINOR_VERSION=0 -DNV_PATCHLEVEL=6629

  -DNV_UNIX  -DNV_LINUX  -DNV_INT64_OK  -DNVCPU_X86      -UDEBUG -U_DEBUG -

  DNDEBUG -DNV_REMAP_PFN_RANGE_PRESENT -DNV_CHANGE_PAGE_ATTR_PRESENT -DNV_PCI_

  DISABLE_DEVICE_PRESENT -DNV_CLASS_SIMPLE_CREATE_PRESENT -DNV_PCI_GET_CLASS_P

  RESENT  -DMODULE -DKBUILD_BASENAME=nv -DKBUILD_MODNAME=nvidia -c -o /root/tm

  p/selfgz7292/NVIDIA-Linux-x86-1.0-6629-pkg1/usr/src/nv/nv.o /root/tmp/s

  elfgz7292/NVIDIA-Linux-x86-1.0-6629-pkg1/usr/src/nv/nv.c

  In file included from include/linux/dmapool.h:14,

                    from include/linux/pci.h:863,

                    from /root/tmp/selfgz7292/NVIDIA-Linux-x86-1.0-6629-pkg1/us

  r/src/nv/nv-linux.h:75,

                    from /root/tmp/selfgz7292/NVIDIA-Linux-x86-1.0-6629-pkg1/us

  r/src/nv/nv.c:14:

  include/asm/io.h: In function `check_signature':

  include/asm/io.h:242: warning: wrong type argument to increment

  /root/tmp/selfgz7292/NVIDIA-Linux-x86-1.0-6629-pkg1/usr/src/nv/nv.c: In func

  tion `nvidia_init_module':

  /root/tmp/selfgz7292/NVIDIA-Linux-x86-1.0-6629-pkg1/usr/src/nv/nv.c:930: war

  ning: `pm_register' is deprecated (declared at include/linux/pm.h:106)

  /root/tmp/selfgz7292/NVIDIA-Linux-x86-1.0-6629-pkg1/usr/src/nv/nv.c: In func

  tion `nvidia_exit_module':

  /root/tmp/selfgz7292/NVIDIA-Linux-x86-1.0-6629-pkg1/usr/src/nv/nv.c:1051: wa

  rning: `pm_unregister' is deprecated (declared at include/linux/pm.h:111)

  /root/tmp/selfgz7292/NVIDIA-Linux-x86-1.0-6629-pkg1/usr/src/nv/nv.c: In func

  tion `_get_phys_address':

  /root/tmp/selfgz7292/NVIDIA-Linux-x86-1.0-6629-pkg1/usr/src/nv/nv.c:2509: er

  ror: request for member `pgd' in something not a structure or union

  /root/tmp/selfgz7292/NVIDIA-Linux-x86-1.0-6629-pkg1/usr/src/nv/nv.c: In func

  tion `nv_agp_init':

  /root/tmp/selfgz7292/NVIDIA-Linux-x86-1.0-6629-pkg1/usr/src/nv/nv.c:2992: wa

  rning: `inter_module_put' is deprecated (declared at include/linux/module.h:

  578)

  make[3]: *** [/root/tmp/selfgz7292/NVIDIA-Linux-x86-1.0-6629-pkg1/usr/src/nv

  /nv.o] Fel 1  ( = ERROR 1)

  make[2]: *** [_module_/root/tmp/selfgz7292/NVIDIA-Linux-x86-1.0-6629-pkg1/us

  r/src/nv] Fel 2  ( = ERROR 2)

  NVIDIA: left KBUILD.

  nvidia.ko failed to build!

  make[1]: *** [module] Fel 1  ( = ERROR 1)

  make: *** [module] Fel 2  ( = 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.

 

I dont really know what to make out of the errors, fel 1 is = error 1 as i have added into the log.

Thanks for all the help i might get and i as i said before, i am a newbie so please give a straightforware answer of what to do, if possible.

[tyme]

Follow the directions in this post.

 

Specifically, you'll want to download the fill attached the the very first post, which is here.

 

Put both files in the same directory.

 

first, we want to get exit the X environment and login as root. so:

init 3

should close out the graphical session. You'll be asked to login, login in as root.

 

Now, taking the instructions on that page, go into the directory you downloaded the files to (cd /path/to/files) and then do these steps:

# sh ./NVIDIA-Linux-x86-1.0-6629-pkg1.run --extract-only
# cd NVIDIA-Linux-x86-1.0-6629-pkg1
# patch -p0 < NVIDIA_kernel-1.0-6629-02.28.2005.diff.txt
# modprobe -q agpgart
# ./nvidia-installer -n

 

feel free to ask questions if i'm not making sense and i'll try to clarify :)

[Larsson]

New error message:

 

ERROR: Unable to load the kernel module 'nvidia.ko'.  This is most likely

      because the kernel module was built using the wrong kernel source files.

      Please make sure you have installed the kernel source files for your

      kernel; on Red Hat Linux systems, for example, be sure you have the

      'kernel-source' rpm installed.  If you know the correct kernel source

      files are installed, you may specify the kernel source path with the

      '--kernel-source-path' commandline option.

-> Kernel module load error: insmod: error inserting './usr/src/nv/nvidia.ko':

  -1 Invalid module format

-> Kernel messages:

  eth0: Setting Rx mode to 4 addresses.

  IPv6 over IPv4 tunneling driver

  Disabled Privacy Extensions on device d2c26c00(sit0)

  eth0: no IPv6 routers present

  nvidia: version magic '2.6.11-12mdk 686 gcc-3.4' should be '2.6.11-6mdk 686

  gcc-3.4'

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.

 

# modprobe -q agpgart <-- on Linux 2.6, only!

 

isnt mandriva 10.2 3.XXX on kernel or what does the above mean?

 

Want it installed first then i come back to what i have done =)

[tyme]

you get that message because you have the wrong kernel sources for the kernel you have installed. You have two choices, you can either upgrade the kernel:

urpmi kernel

and this should grab the new one (which is what you got the sources for)

or you can take the easier (read: more likely to work without a hassle) way and just get the sources for the kernel you have:

urpmi kernel-source-2.6.11-6mdk

you may need to remove the kernel sources you already installed, IIRC you do that with:

urpme kernel-source-2.6.11-12mdk

[Larsson]

Installed them when i follewed a installation tutorial:

urpmi kernel or urpmi kernel-source

 

Everything is aldready installed

Had no success with the following:

urpmi kernel-source-2.6.11-6mdk

 

there is no package with the kernel name kernel-source-2.6.11-6mdk

urpme kernel-source-2.6.11-12mdk

 

unknown package: kernel-source-2.6.11-12mdk

 

nvidia: version magic '2.6.11-12mdk 686 gcc-3.4' should be '2.6.11-6mdk 686

gcc-3.4'

is that really the kernel isnt that some of Nvidias alternativly some other reproduktion that i might have installed in panic attack.

[Qchem]

If you run MCC and go to software management -> install -> search

 

and look for "kernel" (without the quotes) you should be able to find the package that matches your kernel source - or find the kernel source that matches your running kernel.