[SOLVED] Nvidia Slackbuild and Slackware 13.37 to 13.1

CFet · 04-01-2011, 08:47 PM

Newbie Alert

So I gleefully installed Slackware 13.37 but alas, it seems I cannot install nvidia drivers. I think this is due to the slackbuilds.org packages (13.1) not matching my kernel (13.37). Restarting the X server lands me quite a few errors.

Am I right with this assumption (mismatched build and local kernels) and can this be fixed by me learning to build my own Slackbuild script?

Secondly, as an aside, is it possible to downgrade to Slackware 13.1? If so, where would I begin?

Kind regards,
Chris

willysr · 04-01-2011, 08:52 PM

What NVidia version did you use?
I used 260.19.44 and it's working on my -Current machines

please post the error messages as well. It helps us to debug your problem

CFet · 04-01-2011, 09:00 PM

Quote:

Originally Posted by willysr

What NVidia version did you use?
I used 260.19.44 and it's working on my -Current machines

Thanks for the quick reply! I used 260.19.29 from http://slackbuilds.org/result/?search=nvidia&sv=13.1 to build the packages and then install them.

I'll repost the errors in a few~

willysr · 04-01-2011, 09:13 PM

Ah.. it hasn't supported your kernel

2.6.37.x requires 260.19.36 and newer (.44 is the latest one)

try editing the VERSION line and get the latest binary driver from NVidia website http://www.nvidia.com/object/unix.html

CFet · 04-01-2011, 09:16 PM

Here's the errors upon killing and restarting X:

Code:

FATAL: Error inserting nvidia(/lib/modules/2.6.37.5/kernel/drivers/video/nvidia.ko): No such device
(EE)NVIDIA: Failed to load the NVIDIA kernel module. Please
(EE)NVIDIA: check your systems's kernel log for additional error messages.
(EE)NVIDIA: No drivers available

Fatal server error:
no screens found

xinit: giving up
xinit: unable to connect to X server: Connection refused
xinit: server error

---------- Post added 04-01-11 at 09:17 PM ----------

Quote:

Originally Posted by willysr

Ah.. it hasn't supported your kernel

2.6.37.x requires 260.19.36 and newer (.44 is the latest one)

try editing the VERSION line and get the latest binary driver from NVidia website http://www.nvidia.com/object/unix.html

OK! Will let you know how I make out.

CFet · 04-01-2011, 10:03 PM

Well seems I have no luck with rebuilt packages from the latest and greatest Nvidia binaries. Getting the same error.

I have to hit ctrl-alt-backspace to kill the x server, I can't seem to find an easier or "gentler" way to do it at the moment. Upon being kicked to command line, I run "startx" and the wall of errors ensues

.

Is there a way to log what happens when I am in cli and hit startx? I'm having to re-type what I see, writing it down on paper in the process :S
Chris

Mol_Bolom · 04-01-2011, 11:09 PM

That's odd. It looks similar to my setup, but I don't have any problems with it.

Run "find /lib -name nvidia.ko" and check to make sure it was compiled, or at least added to the kernel you boot with. Mine was installed to 2.6.37.5-smp and not 2.6.37.5.

Code:

# find /var/log/packages \( -name mesa* -o -name nvid* -o -name kernel*smp* \)
/var/log/packages/kernel-generic-smp-2.6.37.5_smp-i686-1
/var/log/packages/kernel-headers-2.6.37.5_smp-x86-1
/var/log/packages/kernel-huge-smp-2.6.37.5_smp-i686-1
/var/log/packages/kernel-modules-smp-2.6.37.5_smp-i686-1
/var/log/packages/kernel-source-2.6.37.5_smp-noarch-1
/var/log/packages/mesa-7.9.2-i486-1
/var/log/packages/nvidia-driver-260.19.29-x86-2_SBo
/var/log/packages/nvidia-kernel-260.19.29_2.6.37.4_smp-x86-1_SBo
/var/log/packages/nvidia-kernel-260.19.29_2.6.37.5_smp-x86-2_SBo


# find /usr/lib -name *.260.*
/usr/lib/libGL.so.260.19.29
/usr/lib/libGL.so.260.19.29-nvidia
/usr/lib/libXvMCNVIDIA.so.260.19.29
/usr/lib/libcuda.so.260.19.29
/usr/lib/libnvidia-cfg.so.260.19.29
/usr/lib/libnvidia-compiler.so.260.19.29
/usr/lib/libnvidia-glcore.so.260.19.29
/usr/lib/libnvidia-tls.so.260.19.29
/usr/lib/tls/libnvidia-tls.so.260.19.29
/usr/lib/vdpau/libvdpau_nvidia.so.260.19.29
/usr/lib/xorg/modules/extensions/libglx.so.260.19.29
/usr/lib/xorg/modules/extensions/libglx.so.260.19.29-nvidia
/usr/lib/xorg/modules/libnvidia-wfb.so.260.19.29

CFet · 04-01-2011, 11:22 PM

Thank you for the help Mol_Bolom

Code:

find /lib -name nvidia.ko

yields:

Code:

/lib/modules/2.6.37.5/kernel/drivers/video/nvidia.ko

question: If I run glxgears, does that indicate the presence of 3D rendering? I ran the NVIDIA_longfilename.run directly and followed the prompts until the install "failed'. Interestingly, my computer did fire up into X after a reboot and /etc/X11/xorg.conf does indicate "nivida" as the driver in the Device section...

Mol_Bolom · 04-01-2011, 11:38 PM

Quote:

Originally Posted by CFet

Thank you for the help Mol_Bolom

Code:

find /lib -name nvidia.ko

yields:

Code:

/lib/modules/2.6.37.5/kernel/drivers/video/nvidia.ko

Ok, so it was installed, and considering the slackbuild builds it for the kernel currently loaded it should be the one that you booted up with. Still is kind of odd that it's not working correctly.

Quote:

question: If I run glxgears, does that indicate the presence of 3D rendering? I ran the NVIDIA_longfilename.run directly and followed the prompts until the install "failed'. Interestingly, my computer did fire up into X after a reboot and /etc/X11/xorg.conf does indicate "nivida" as the driver in the Device section...

It wouldn't necessarily mean that it's working, but "glxinfo | grep render" should yield "direct rendering: yes" if it is working correctly.

Also, check the output of the following command to check if there is any errors when the nvidia.ko driver loads.
# dmesg | grep nvidia

<Edit>

Quote:

Interestingly, my computer did fire up into X after a reboot and /etc/X11/xorg.conf does indicate "nivida" as the driver in the Device section...

Just noticed this. (Sometimes my brain happens to "think" it's read something when it hasn't). It might be working now.

CFet · 04-01-2011, 11:48 PM

Code:

bash-4.1$ glxinfo | grep render
direct rendering: Yes
OpenGL renderer string: GeForce GTS 250/PCI/SSE2
    GL_NV_conditional_render, GL_NV_copy_depth_to_color, GL_NV_copy_image, 
    GL_NVX_conditional_render, GL_NVX_gpu_memory_info,

Code:

bash-4.1$ dmesg | grep nvidia
klogctl: Operation not permitted
bash-4.1$ sudo dmesg | grep nvidia
Password:
[    4.505200] nvidia: module license 'NVIDIA' taints kernel.
[    5.080498] nvidia 0000:01:00.0: PCI INT A -> GSI 16 (level, low) -> IRQ 16
[    5.080597] nvidia 0000:01:00.0: setting latency timer to 64

I have no idea how to interpret the last section, but going by the direct rendering: Yes, seems safe to say that it is working? I loathe these mystery fixes. I almost want to nuke all the nvidia packages and try a direct ./NVIDIA_longfilename.run and see if that does the trick.

Thanks so much for your help! I'll hold off on the [Solved] tag for the moment.

Mol_Bolom · 04-02-2011, 12:01 AM

Quote:

Originally Posted by CFet

Code:

bash-4.1$ glxinfo | grep render
direct rendering: Yes
OpenGL renderer string: GeForce GTS 250/PCI/SSE2
    GL_NV_conditional_render, GL_NV_copy_depth_to_color, GL_NV_copy_image, 
    GL_NVX_conditional_render, GL_NVX_gpu_memory_info,

Code:

bash-4.1$ dmesg | grep nvidia
klogctl: Operation not permitted
bash-4.1$ sudo dmesg | grep nvidia
Password:
[    4.505200] nvidia: module license 'NVIDIA' taints kernel.
[    5.080498] nvidia 0000:01:00.0: PCI INT A -> GSI 16 (level, low) -> IRQ 16
[    5.080597] nvidia 0000:01:00.0: setting latency timer to 64

I have no idea how to interpret the last section, but going by the direct rendering: Yes, seems safe to say that it is working? I loathe these mystery fixes. I almost want to nuke all the nvidia packages and try a direct ./NVIDIA_longfilename.run and see if that does the trick.

Thanks so much for your help! I'll hold off on the [Solved] tag for the moment.

I believe it is solved. This is the exact same thing I see when I boot. The reason you see "nvidia: module license 'NVIDIA' taints kernel." is because it is a closed source, or some other term that I don't know as of yet, which isn't a normal kernel module. In simple terms, it's something the kernel writes to screen to give the user/admin information that a module was loaded that could be a virus. However, I am horrible with explaining that, but if you did a search on google you should be able to find a better explanation.

willysr · 04-02-2011, 03:18 AM

have you make sure that your /etc/X11/xorg.conf has loaded "nvidia" ??

wadsworth · 04-02-2011, 05:38 AM

Quote:

Use one of the provided generic kernels for daily use. Do not report
bugs until/unless you have reproduced them using one of the stock
generic kernels. You will need to create an initrd in order to boot
the generic kernels - see /boot/README.initrd for instructions.
The huge kernels are primarily intended as "installer" and "emergency"
kernels in case you forget to make an initrd. For most systems, you
should use the generic SMP kernel if it will run, even if your system is
not SMP-capable. Some newer hardware needs the local APIC enabled in the
SMP kernel, and theoretically there should not be a performance penalty
with using the SMP-capable kernel on a uniprocessor machine, as the SMP
kernel tests for this and makes necessary adjustments. Furthermore, the
kernel sources shipped with Slackware are configured for SMP usage, so you
won't have to modify those to build external modules (such as NVidia or
ATI proprietary drivers) if you use the SMP kernel.

If you decide to use one of the non-SMP kernels, you will need to follow the
instructions in /extra/linux-2.6.37.6-nosmp-sdk/README.TXT to modify your
kernel sources for non-SMP usage. Note that this only applies if you are
using the Slackware-provided non-SMP kernel - if you build a custom kernel,
the symlinks at /lib/modules/$(uname -r)/{build,source} will point to the
correct kernel source so long as you don't (re)move it.

http://slackware.osuosl.org/slackwar..._AND_HINTS.TXT

CFet · 04-02-2011, 12:28 PM

Thanks all for the help, seems to be up and running!
On to multilib stuff for wine/skype.

Cheers,
Chris