nvidia & 2.6.6 & x11

Message

njlg · Post by **njlg** » Wed Jun 02, 2004 3:44 pm

(I am posting these because I have noticed some people having this problem on this forum, but no one has posted a solution. I hope one exists.)

I recency ran into problems with 2.6.6 & nvidia-kernel (who hasn't really). So after I tried to get them working together for acouple weeks, I decided to go back to 2.6.5 where I had nvidia-kernel (quasi) working.

When my system comes up with 2.6.5 the nvidia module is auto probed.

Code: Select all

% startx

hostname: Unknown host

Release Date: 18 December 2003
X Protocol Version 11, Revision 0, Release 6.7
Build Operating System: Linux 2.6.5-gentoo-r1 i686 [ELF] 
Current Operating System: Linux cN*2 2.6.5-gentoo-r1 #2 SMP Thu Apr 22 10:33:34 PDT 2004 i686
Build Date: 23 April 2004
	Before reporting problems, check http://wiki.X.Org
	to make sure that you have the latest version.
Module Loader present
Markers: (--) probed, (**) from config file, (==) default setting,
	(++) from command line, (!!) notice, (II) informational,
	(WW) warning, (EE) error, (NI) not implemented, (??) unknown.
(==) Log file: "/var/log/Xorg.0.log", Time: Tue Jun  1 13:08:00 2004
(==) Using config file: "/etc/X11/XF86Config"
Using vt 7

(EE) NVIDIA(0): Failed to initialize the NVIDIA kernel module!
(EE) NVIDIA(0):  *** Aborting ***
(EE) Screen(s) found, but none have a usable configuration.

Fatal server error:
no screens found

Please consult the The X.Org Foundation support 
	 at http://wiki.X.Org
 for help. 
Please also check the log file at "/var/log/Xorg.0.log" for additional information.

XIO:  fatal IO error 104 (Connection reset by peer) on X server ":0.0"

      after 0 requests (0 known processed) with 0 events remaining.

Then I uninstall all versions of nvidia-kernel & nvidia-glx from my system. Double check that /usr/src/linux is pointing to 2.6.5-gentoo-r1. Then I emerge nvidia-kernel-1.0.5336-r2 &
nvidia-glx-1.0.5336-r2. Then I try to run x11 and get exactly the same message as before. Then I try the following

Code: Select all

% sh /usr/portage/distfiles/NVIDIA-Linux-x86-1.0-5341-pkg1.run -a -s

chown: `0.0' syntax is obsolete; use `:' since this will be removed in the future
make: *** [select_makefile] Error 1

ERROR: Unable to determine the NVIDIA kernel module filename.


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.

Code: Select all

% cat /var/log/nvidia-installer.log

nvidia-installer log file '/var/log/nvidia-installer.log'
creation time: Tue Jun  1 13:08:53 2004

option status:
  license pre-accepted    : true
  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            : true
  silent                  : true
  XFree86 install prefix  : /usr/X11R6
  OpenGL install prefix   : /usr
  Installer install prefix: /usr
  kernel source path      : (not specified)
  kernel install path     : (not specified)
  proc mount point        : /proc
  ui                      : none
  tmpdir                  : /tmp
  ftp site                : ftp://download.nvidia.com

Using built-in stream user interface
-> License accepted by command line option.
-> There appears to already be a driver installed on your system (version: 1.0-
   5341).  As part of installing this driver (version: 1.0-5341), the existing 
   driver will be uninstalled.  Are you sure you want to continue? ('no' will a
   bort installation) (Answer: Yes)
-> 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: Yes)
-> No matching precompiled kernel interface was found on the NVIDIA ftp site;
   this means that the installer will need to compile a kernel interface for
   your kernel.
-> Kernel source path: '/usr/src/linux'
-> Performing cc_version_check with CC="gcc".
ERROR: Unable to determine the NVIDIA kernel module filename.
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.

Next I reboot into 2.6.6-rc1. The nvidia module is auto probed this time too. When the system comes up I try

Code: Select all

% startx
hostname: Unknown host

Release Date: 18 December 2003
X Protocol Version 11, Revision 0, Release 6.7
Build Operating System: Linux 2.6.5-gentoo-r1 i686 [ELF] 
Current Operating System: Linux cN*2 2.6.6-rc1 #2 SMP Tue May 18 10:45:37 PDT 2004 i686
Build Date: 23 April 2004
	Before reporting problems, check http://wiki.X.Org
	to make sure that you have the latest version.
Module Loader present
Markers: (--) probed, (**) from config file, (==) default setting,
	(++) from command line, (!!) notice, (II) informational,
	(WW) warning, (EE) error, (NI) not implemented, (??) unknown.
(==) Log file: "/var/log/Xorg.0.log", Time: Tue Jun  1 13:13:16 2004
(==) Using config file: "/etc/X11/XF86Config"
Using vt 7
(EE) NVIDIA(0): Failed to initialize the NVIDIA kernel module!
(EE) NVIDIA(0):  *** Aborting ***
(EE) Screen(s) found, but none have a usable configuration.

Fatal server error:
no screens found

Please consult the The X.Org Foundation support 
	 at http://wiki.X.Org
 for help. 
Please also check the log file at "/var/log/Xorg.0.log" for additional information.

XIO:  fatal IO error 104 (Connection reset by peer) on X server ":0.0"

      after 0 requests (0 known processed) with 0 events remaining.

Code: Select all

% lsmod

Module                  Size  Used by
nvidia               2079592  0

The only way I know how to get nvidia-kernel working with x11 is to manually run the installer from Nvidia.com.

Code: Select all

% sh /usr/portage/distfiles/NVIDIA-Linux-x86-1.0-5341-pkg1.run -a -s

Verifying archive integrity... OK
Uncompressing NVIDIA Accelerated Graphics Driver for Linux-x86 1.0-5341........................................................................................
.........................................................................................

chown: `0.0' syntax is obsolete; use `:' since this will be removed in the future

WARNING: Unable to restore symbolic link /usr/X11R6/lib/libGL.so.1 ->
         libMesaGL.so (File exists).

But keep in mind I still have these emerged

Code: Select all

%qpkg -I -v nvidia
nvidia-kernel-1.0.5336-r3
nvidia-glx-1.0.5336-r2

And x11 starts just fine. The only problem is I have to do this every time I restart my computer (which happens a lot now, because as far as I can tell the nvidia driver locks up my system sporadically).

When I restart the system in 2.6.6-rc1 this is the error message

Code: Select all

% startx
hostname: Unknown host

Release Date: 18 December 2003

X Protocol Version 11, Revision 0, Release 6.7
Build Operating System: Linux 2.6.5-gentoo-r1 i686 [ELF] 
Current Operating System: Linux cN*2 2.6.6-rc1 #2 SMP Tue May 18 10:45:37 PDT 2004 i686
Build Date: 23 April 2004
	Before reporting problems, check http://wiki.X.Org
	to make sure that you have the latest version.
Module Loader present
Markers: (--) probed, (**) from config file, (==) default setting,
	(++) from command line, (!!) notice, (II) informational,
	(WW) warning, (EE) error, (NI) not implemented, (??) unknown.
(==) Log file: "/var/log/Xorg.0.log", Time: Tue Jun  1 17:58:51 2004
(==) Using config file: "/etc/X11/XF86Config"
Using vt 7
dlopen: /usr/lib/opengl/nvidia/lib/libGLcore.so.1: undefined symbol: _nv000024gl
(EE) Failed to load /usr/X11R6/lib/modules/extensions/libglx.so
(EE) Failed to load module "glx" (loader failed, 7)
(EE) NVIDIA(0): Failed to initialize the NVIDIA kernel module!
(EE) NVIDIA(0):  *** Aborting ***
(EE) Screen(s) found, but none have a usable configuration.

Fatal server error:
no screens found

Please consult the The X.Org Foundation support 
	 at http://wiki.X.Org
 for help. 
Please also check the log file at "/var/log/Xorg.0.log" for additional information.

XIO:  fatal IO error 104 (Connection reset by peer) on X server ":0.0"

      after 0 requests (0 known processed) with 0 events remaining.

I think I will try love-sources later and see if that helps anything. I also need to experiment with 2.6.6-rc1 kernel settings (ie. some people get nvidia to work without fb).

ikaro · Post by **ikaro** » Wed Jun 02, 2004 5:44 pm

I got a good working solution...

emerge -C nvidia-kernel nvidia-glx

then edit the xorg.conf and change:

"nvidia" to "nv"

now you can startx

Aurisor · Post by **Aurisor** » Wed Jun 02, 2004 6:14 pm

ikaro wrote:I got a good working solution...

emerge -C nvidia-kernel nvidia-glx

then edit the xorg.conf and change:

"nvidia" to "nv"

now you can startx

Just in case you couldn't tell, he's being a wiseass.

Cintra · Post by **Cintra** » Wed Jun 02, 2004 6:32 pm

ikaro wrote:I got a good working solution...

emerge -C nvidia-kernel nvidia-glx

then edit the xorg.conf and change:

"nvidia" to "nv"

now you can startx

Been there..

mvh

ikaro · Post by **ikaro** » Wed Jun 02, 2004 7:49 pm

ishan wrote:
ikaro wrote:I got a good working solution...

emerge -C nvidia-kernel nvidia-glx

then edit the xorg.conf and change:

"nvidia" to "nv"

now you can startx
Just in case you couldn't tell, he's being a wiseass.

It is known that the nvidia drivers are not working properlly with the kernels 2.6.x
Im not using the nvidia drivers, and not going to waste more time trying to, until nvidia releases decent drivers.

You can take my advice and so the same, or not

Whatever you might chose to do, peace! and good luck.

Drooling Iguana · Post by **Drooling Iguana** » Wed Jun 02, 2004 8:46 pm

They work fine with 2.6.4, but not on anything above that.

MaDxRaY · Post by **MaDxRaY** » Wed Jun 02, 2004 9:07 pm

Drooling Iguana wrote:They work fine with 2.6.4, but not on anything above that.

I wouldn't say it this way...gentoo-dev-sources (2.6.5-r3) and nvidia/nvidia-glx(newest, 1.0.5336-r3/r2) are working fine here.

@co je Jidda Sha Ti'l:
Your problem sound's like missing module...are you shure it's up(lsmod)?
In my eyes it could also work with kernel 2.6.6 and above, cause mainly libs, headers and some links where changed, and their path should be the same...what about mesa, does it stuck with new kernel?

IvanHoe · Post by **IvanHoe** » Thu Jun 03, 2004 12:12 am

Do you have CONFIG_REGPARMS or CONFIG_4KSTACKS enabled? If you do then disable them and rebuild your kerrnel. If you don't see those options then check arch/i386/Kconfig in your linux source folder for "config REGPARMS" or "config 4KSTACKS". You'll see something like this

Code: Select all

config REGPARMS
    def_bool=y

Just set it to "n".

njlg · Post by **njlg** » Thu Jun 03, 2004 7:02 pm

I am using nv right now. Wow! I never really used this driver longer than a few moments. Everything is all jerky. This is great! =]

I am not surprised by nvidia flat out not working with 2.6.6-rc1. Though I am midly surprised that it does work, but I just have to install it every time I reboot. And even more so that I cannot get nvidia working with 2.6.5, even though I had it working perfectly (sort of) before.

I believe the only difference between the first time I had nvidia working with 2.6.5 and now, is that I emerged the arch modutils and module-init-tools.

Thanks for everyones responces. I guess I will just have to play around with backing things off of arch and see if that helps.

Deathwing00 · Post by **Deathwing00** » Thu Jun 03, 2004 7:46 pm

Moved from Desktop Environments.
Check bug: http://bugs.gentoo.org/show_bug.cgi?id=47692

teutzz · Post by **teutzz** » Thu Jun 03, 2004 9:21 pm

try the latest driver available in arch="~x86"
i've been using it without any problem in conjuction with 2.6.6-love4

njlg · Post by **njlg** » Thu Jun 03, 2004 9:38 pm

Deathwing00,
Thanks for moving to the correct location (wherever that may be); but the bug is not related to my problem.

teutzz (and all),
I have tried this with 1.0.5??? to 1.0.5341-r1 with no luck. Also, I just tried 2.6.7-rc2-love2 with no luck either.

teutzz · Post by **teutzz** » Thu Jun 03, 2004 9:49 pm

are you using 4k stack in your kernel config? because nvidia don't cope with that

danone · Post by **danone** » Thu Jun 03, 2004 10:47 pm

pure question do you use udev? are sysfstools,hotplug ..i ran 2 time in this problem and made everything new until I checked /dev for nvidia0 & nvidiactl..hehe solution was NVmakedevices..but this is for udev..

njlg · Post by **njlg** » Fri Jun 04, 2004 3:30 am

danon: I call NVmakedevices on startup.

teutzz: this is not related to 4k stack problem, because the module loads and worksine (after I install it). =]

John5788 · Post by **John5788** » Fri Jun 04, 2004 3:46 am

m@dxray wrote:
Drooling Iguana wrote:They work fine with 2.6.4, but not on anything above that.
I wouldn't say it this way...gentoo-dev-sources (2.6.5-r3) and nvidia/nvidia-glx(newest, 1.0.5336-r3/r2) are working fine here.

@co je Jidda Sha Ti'l:
Your problem sound's like missing module...are you shure it's up(lsmod)?
In my eyes it could also work with kernel 2.6.6 and above, cause mainly libs, headers and some links where changed, and their path should be the same...what about mesa, does it stuck with new kernel?

same here

teutzz · Post by **teutzz** » Fri Jun 04, 2004 9:42 am

man, there must be something really wrong with your box; I never had any problems with nvidia drivers not in conjunction with 2.4 or 2.6 kernels
One thing comes to my mind (stupid question by the way, so please don't flame me for asking it): are you sure you are propelly installing your newly configured kernel?
(had to ask it

)

Bonkie · Post by **Bonkie** » Fri Jun 04, 2004 6:03 pm

danone wrote:pure question do you use udev? are sysfstools,hotplug ..i ran 2 time in this problem and made everything new until I checked /dev for nvidia0 & nvidiactl..hehe solution was NVmakedevices..but this is for udev..

Well you did solve my issues

Gonna try 2.6.7 now that 2.6.4 works

dan2003 · Post by **dan2003** » Fri Jun 04, 2004 8:39 pm

Bah, i just ran into this too.. silly thing is my box has been fine.

I booted and logged in fine.. I was trying to build summin in kdevelop 3.0.1 and ran into some weird build issue with things that were on an nfs mount.

I started an update of kdevelop but didn't pay attention to what else it was updating apart from noticed kdelibs 3.2.2 was becoming 3.2.2-r1 there were 8 things to do in total, but i cancelled it (ctrl+c) on the seventh cos i wanted to umount my nfs mount and something was keeping it busy within kde so i wanted to log out and back in. I logged out and went to log back in but X wouldnt start (i use qingy 0.4 by the way).

After much poking around i noticed that nvidia0 and nvidiactl are not in /dev anymore. I added them manaully but i stilll cant get X to start I have remerged baselayout, devfsd, nvidia-kernel, nvidia-glx and now emerged xorg instead and remerged nvidia again but i still cant get X back..

.

So i'm stumped at the mo.

[Edit} By the way im using kernel 2.6.4-ck2

njlg · Post by **njlg** » Fri Jun 04, 2004 10:46 pm

danone wrote:pure question do you use udev? are sysfstools,hotplug ..i ran 2 time in this problem and made everything new until I checked /dev for nvidia0 & nvidiactl..hehe solution was NVmakedevices..but this is for udev..

Damnit! You were right. I did have NVmakedevices.sh on startup, but I guess I forgot to save my local.start when I upgraded. I feel so stupid right now.

So this total fixed my problem with all of the kernels I have and a new one =]

Code: Select all

Linux cN*2 2.6.7-rc2-love2 #1 SMP Thu Jun 3 12:31:34 PDT 2004 i686 AMD Athlon(tm) XP 2400+ AuthenticAMD GNU/Linux

Thanks for everyone's help and opinions!

thechris · Post by **thechris** » Tue Jun 08, 2004 10:06 pm

ok, firstly i'm trying to fix sound without losing video. alsa sucks with aureal cards. i'm hoping that by ugrading to 2.6.7 sources that i'll fix the sound.

but nvidia won't work.

4k stacks and reg-arguments aren't in the kernel. 8k stacks.
i've modified /etc/conf.d/local.start to have mknod for nvidia0-7 and ctl
re-emerged the 5336 kernel driver and glx packages
tried the patch script that fixed "invalid module format" in the past
tried modprobe nvidia

what log gives detailed specs. pretty much it just says failed to load nvidia -- no screens with valid conf found.
this is from the log listed in the errror.

tried the NVmakedevices.sh script as well.

so far nothing has worked. and i don't get a detailed error message.

the_sphynx · Post by **the_sphynx** » Thu Jun 17, 2004 3:31 am

Hey All,

I just wanted to let you know that the 2.6.7 Kernel came out today as a stable release according to Kernel.org. I downloaded it and complied and I am getting the "invalid module format" error as well after makeing sure that /usr/src/linux is pointing to linux-2.6.7. I hope that NVidia gets off their buts and gets a driver to work for anything higher than my 2.6.5 kernel (that is as high up as I can go).

HydroSan · Post by **HydroSan** » Thu Jun 17, 2004 3:49 am

the_sphynx wrote:Hey All,

I just wanted to let you know that the 2.6.7 Kernel came out today as a stable release according to Kernel.org. I downloaded it and complied and I am getting the "invalid module format" error as well after makeing sure that /usr/src/linux is pointing to linux-2.6.7. I hope that NVidia gets off their buts and gets a driver to work for anything higher than my 2.6.5 kernel (that is as high up as I can go).

I got nVidia working with 2.6.7-rc3. I don't think there would be THAT much of a difference between 2.6.6 and 2.6.7.

the_sphynx · Post by **the_sphynx** » Thu Jun 17, 2004 6:13 am

the_sphynx wrote:
Hey All,

I just wanted to let you know that the 2.6.7 Kernel came out today as a stable release according to Kernel.org. I downloaded it and complied and I am getting the "invalid module format" error as well after makeing sure that /usr/src/linux is pointing to linux-2.6.7. I hope that NVidia gets off their buts and gets a driver to work for anything higher than my 2.6.5 kernel (that is as high up as I can go).
.
I got nVidia working with 2.6.7-rc3. I don't think there would be THAT much of a difference between 2.6.6 and 2.6.7.

What card are you using? I am using a GForce2 Ti with 32M/RAM and I can't get anything but the 2.6.5 Kernel to work. Did you compile something different in the 2.6.7-rc3 kernel than the previous version?[/quote]

HydroSan · Post by **HydroSan** » Thu Jun 17, 2004 6:19 am

HydroSan wrote:
Hey All,

I just wanted to let you know that the 2.6.7 Kernel came out today as a stable release according to Kernel.org. I downloaded it and complied and I am getting the "invalid module format" error as well after makeing sure that /usr/src/linux is pointing to linux-2.6.7. I hope that NVidia gets off their buts and gets a driver to work for anything higher than my 2.6.5 kernel (that is as high up as I can go).
I got nVidia working with 2.6.7-rc3. I don't think there would be THAT much of a difference between 2.6.6 and 2.6.7.
What card are you using? I am using a GForce2 Ti with 32M/RAM and I can't get anything but the 2.6.5 Kernel to work. Did you compile something different in the 2.6.7-rc3 kernel than the previous version?

GeForceFX 5600 256MB. Works like a charm with 8k stacks.