Mailinglist Archive: opensuse (880 mails)

< Previous Next >
Re: [opensuse] Troubles with Nvidia
[07.08.2013 19:03] [Vojtěch Zeisek]:

OK, did you read
</usr/share/doc/packages/x11-video-nvidiaG03/html/optimus.html> then? Or
any other file in this directory?

Previously not. After reading it, still don't get the point...

The driver may be installed normally on Optimus systems, but the
NVIDIA X driver and the NVIDIA OpenGL driver may not be able to
display to the laptop's internal display panel unless a means to
connect the panel to the NVIDIA GPU (for example, a hardware
multiplexer, or "mux", often controllable by a BIOS setting) is
available.

You may need an external display to use the nvidia driver...


BTW, in supportedchips.html in this directory you can see that
x11-video-nvidiaG03 supports the GT630M chip so that you can uninstall
the G02 variants, just to avoid that they fight each other.

Why didn't I stayed with Intel graphics?

I do not know :-) But the make the fewest trouble on laptops.


I installed (for my G96 [GeForce 9400 GT]):
- nvidia-computeG03-319.32-15.1.x86_64
- nvidia-gfxG03-kmp-desktop-319.32_k3.7.10_1.1-15.1.x86_64
- x11-video-nvidiaG03-319.32-15.1.x86_64
all from http://download.nvidia.com/opensuse/12.3

Now I have only the 03 packages. When I run
# nvidia-debugdump --dumpall
FATAL: Module nvidia not found.
NVIDIA: failed to load the NVIDIA kernel module.
Error: nvmlInit(): Driver Not Loaded
# nvidia-modprobe
FATAL: Module nvidia not found.
# rpm -qa |grep nvidia



nvidia-settings-290.10-1.9.x86_64
nvidia-computeG03-319.32-15.1.x86_64
nvidia-gfxG03-kmp-desktop-319.32_k3.7.10_1.1-15.1.x86_64
x11-video-nvidiaG03-319.32-15.1.x86_64
# rpm -qa |grep bumblebee
bumblebee-2.4.1-4.1.x86_64
I feel lost. :-(

Don't :-). Better check if all the files are there. For example, with
"rpm -ql nvidia-gfxG03-kmp-desktop" you see the files that should be
there. Among these, ther is the module that you cannot load.

The following command should be entered in one line:

for DAT in $(rpm -ql nvidia-gfxG03-kmp-desktop) ; do if [ -e "$DAT" ];
then echo "exist: $DAT"; else echo "missing: $DAT"; fi; done

As a shell script, you can use:

#!/bin/bash
for MOD in nvidia-computeG03 nvidia-gfxG03-kmp-desktop \
x11-video-nvidiaG03; do
echo "checking: $MOD"
for DAT in $(rpm -ql "$MOD"); do
if [ -e "$DAT" ]; then
echo "exist: $DAT"
else
echo "missing: $DAT"
fi
done
done

If any file ist missing (according to "FATAL: Module nvidia not found."
this should be the case with
</lib/modules/3.7.10-1.1-desktop/updates/nvidia.ko>), re-install the
package with "zypper in -f "$packagename. For me, it happened sometimes,
that uninstalling one version of a kmp removed a file that belonged to
another kmp too. So maybe the uninstall of the G02 packages caused the
removal of the nvidia.ko file.


NVRM: loading NVIDIA UNIX x86_64 Kernel Module 319.32 Wed Jun 19
15:51:20
PDT 2013
2013-08-06T08:20:17.400659+02:00 veles kdm[948]: X server died during
startup 2013-08-06T08:20:17.401228+02:00 veles kdm[948]: X server for
display :0 cannot be started, session disabled

This might be because of the dual video chipset in your box. I only
found a russian text about it,
<http://ru.opensuse.org/SDB:NVidia_Optimus>, but no translations.

It has only Nvidia graphic card. BIOS has no setting related to graphic
card.
Pardon? Didn't I read

$ lspci | grep VGA
00:02.0 VGA compatible controller: Intel Corporation 3rd Gen Core
processor Graphics Controller (rev 09)
01:00.0 VGA compatible controller: NVIDIA Corporation GF108 [GeForce GT
630M] (rev a1)

in your first post? There are two graphics controllers, and bumblebee
should allow switching between them.

Hehe... HP's product page lists just the Nvidia as graphic card, so that I
automatically supposed it has only one card. Neither product specifications
mention 2 graphic controllers. Confusing.

Nvidia generates this xorg.conf:
I use nvidia drivers, and I have no xorg.conf. If you work on real
hardware, there are way too many "vmware" and "vbox" entries in this file.

It does't work. This is report of Xorg log:

X.Org X Server 1.13.2
Release Date: 2013-01-24
[ 71.426] X Protocol Version 11, Revision 0
[ 71.426] Build Operating System: openSUSE SUSE LINUX
[ 71.426] Current Operating System: Linux veles.site
3.7.10-1.16-desktop #1 SMP PREEMPT Fri May 31 20:21:23 UTC 2013 (97c14ba)
x86_64
[ 71.426] Kernel command line: BOOT_IMAGE=/vmlinuz-3.7.10-1.16-desktop
root=/dev/mapper/pocitac-koren ro elevator=cfq
[ 71.426] Build Date: 15 July 2013 12:22:00PM
[ 71.426]
[ 71.426] Current version of pixman: 0.28.2
[ 71.426] Before reporting problems, check http://wiki.x.org

to make sure that you have the latest version.

[ 71.427] (==) ModulePath set to
"/usr/lib64/xorg/modules/updates,/usr/lib64/xorg/modules"

[ 71.428] (II) LoadModule: "glamoregl"
[ 71.429] (II) Loading /usr/lib64/xorg/modules/libglamoregl.so
[ 71.434] (EE) Failed to load /usr/lib64/xorg/modules/libglamoregl.so:
/usr/lib64/xorg/modules/libglamoregl.so: undefined symbol:
_glapi_tls_Context
Hm. On my box, "rpm -qf /usr/lib64/xorg/modules/libglamoregl.so" gives
glamor-0.5-3.1.3.x86_64, directly from openSUSE distro. Do you have
another version? Where does it come from?

I use openSUSE 12.3 64 bit,
rpm -qf /usr/lib64/xorg/modules/libglamoregl.so
glamor-0.5-3.1.3.x86_64
from main OSS repository.

Ha. Thank you :-). I never noticed that my Xorg.0.log shows the same
error. At least we can say that it is not critical for the nvidia
driver, because it works for me...


[ 71.445] (EE) No devices detected.

Yes, and now we see that the driver does not find any device.

Yes, but why?

Because the driver does not like optimus. Oh, optimus and the driver are
from the same manufacturer? Never mind, in Windows everything is fine.
Huh, am I sarcastic?


I use VirtualBox (if that matters), but I have no idea why there is any
vmware section - I haven't installed it and I'm not going to install it.
I also lost boot splash screen although all packages should be installed.
Well, never mind at all. ;-)

Ahem... are you installing the driver inside a VirtualBox VM? If not,
you can delete all this, and all the VMWare things too. It is there to
give openSUSE some help when started inside a virtual machine, not on a
host that runs VirtualBox or VMWare Player. I suggest you rename
xorg.conf to something that xorg does not look for, like
xorg.conf.nonsense ;-). So xorg's autodetect feature will be startet,
and might become successful.

I installed openSUSE to real HW, HP ENVY dv6 notebook. I also use it as
VirtualBox host. You mean, that xorg.conf should be recreated during startup?
;-)

No, there is no xorg.conf any more, for several versions of openSUSE now
(since sax2 was abandoned). There is a magic feature in xorg that does a
lot autodetecting, and the directory /etc/X11/xorg.conf.d, where you
will find many snippets from the former xorg.conf. In 50-device conf you
can define your own constellation, but it should not be necessary. See
<http://en.opensuse.org/SDB:Configuring_graphics_cards> for a small
example (but replace "nv" with "nvidia"). Here you even are allowed to
create a file xorg.conf ;-)

Of course, autodetection - as any kind of magic - can go wrong...

I found
<http://forums.opensuse.org/english/get-technical-help-here/hardware/466596-nvidia-card-gt520m-not-useable-nvidia-driver-no-devices-detected-screen-blank-2.html>.
Ist this applicable for your case?

I still wish good luck :-)
Werner
--

--
To unsubscribe, e-mail: opensuse+unsubscribe@xxxxxxxxxxxx
To contact the owner, e-mail: opensuse+owner@xxxxxxxxxxxx

< Previous Next >
Follow Ups