15.07.2018 12:58, Daniel Bauer пишет:

I start a new thread to not mix up things.

I have an Asus GL552V laptop, i7 6700, optimus graphics intel/nvidia running opensuse leap 42.3, KDE

The nvidia drivers are installed, but they are not loaded. I have no idea why.

Show dmesg output immediately after boot.

I completely removed everything nvidia, disabling the opensuse.nvidia repository. Then booted in init 3 and installed those drivers again using yast. I hoped this fresh istall would make them appear loaded, but no.

It is possible that due to the many things I tried to get this optimus stuff running again (it worked under 13.x using suse-prime) that there is some mess in the things installed or not installed. But I don't know what is needed and what not and where to begin to search....

By the way the is no /etc/X11/xorg.conf file on my system and yast said something about that when installing the nvidia drivers.

I have a xorg.conf.bak and a xorg.conf.install, furthermore a xorg.conf.nvidia-post and xorg.conf.nvidia-xconfig-original.

I copied xorg.conf.bak (and later xorg.conf.install) to xorg.conf, but with this I could not reach the graphical kde login anymore. So I deleted it again. An here I am.

So, if you think there is any hope that I can use the nvidia card on this optimus laptop, please help!

I guess the first thing is to have those nvidia drivers loaded. And then go further from there...

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

On 2018-07-15 12:30, Andrei Borzenkov wrote:

15.07.2018 13:14, jdd@dodin.org пишет:

...

Le 15/07/2018 à 11:58, Daniel Bauer a écrit :

...

I guess the first thing is to have those nvidia drivers loaded. And then go further from there...

no real idea. If nvidia is installed, nvidia-setings or nvidia-xconfig could be of some use, but on my own computer this write an xorg.conf file that prevent booting...

bumblebee config says also to disable nvidia

Of course it does. Bumblebee physically powers down dGPU when it is not used so you cannot have drivers loaded permanently. They are loaded when you start program using bumblebee.

Ah! Interesting. However, Daniel said on another post that he would be happy to use NVidia permanently. His Intel graphics do not work well, there is flicker and other issues. -- Cheers / Saludos, Carlos E. R. (from 42.3 x86_64 "Malachite" at Telcontar)

15.07.2018 14:47, Daniel Bauer пишет:

On 15.07.2018 12:28, Andrei Borzenkov wrote:

...

15.07.2018 12:58, Daniel Bauer пишет:

...

I start a new thread to not mix up things.

I have an Asus GL552V laptop, i7 6700, optimus graphics intel/nvidia running opensuse leap 42.3, KDE

The nvidia drivers are installed, but they are not loaded. I have no idea why.

Show dmesg output immediately after boot.

after boot, kde-login, open console, su to root: dmesg > dmsg2.txt

https://www.daniel-bauer.com/test/dmsg2.txt

Does "modprobe nvidia" work? What shows modinfo nvidia nvidia-drm -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org

15.07.2018 15:55, Daniel Bauer пишет:

On 15.07.2018 14:38, Andrei Borzenkov wrote:

...

15.07.2018 14:47, Daniel Bauer пишет:

...

On 15.07.2018 12:28, Andrei Borzenkov wrote:

...

15.07.2018 12:58, Daniel Bauer пишет:

...

I start a new thread to not mix up things.

I have an Asus GL552V laptop, i7 6700, optimus graphics intel/nvidia running opensuse leap 42.3, KDE

The nvidia drivers are installed, but they are not loaded. I have no idea why.

Show dmesg output immediately after boot.

after boot, kde-login, open console, su to root: dmesg > dmsg2.txt

https://www.daniel-bauer.com/test/dmsg2.txt

Does "modprobe nvidia" work? What shows

modinfo nvidia nvidia-drm

meitli:~ # modprobe nvidia modprobe: ERROR: could not find module by name='nvidia' modprobe: ERROR: could not insert 'nvidia': Function not implemented

meitli:~ # modinfo nvidia nvidia-drm modinfo: ERROR: Module nvidia not found.

Something is seriously wrong. What rpm -qa '*nvidia*' '*kernel*' says?

filename: /lib/modules/4.4.138-59-default/weak-updates/updates/nvidia-drm.ko> version:        390.67 supported:      external license:        MIT srcversion:     F3776C9F2655E8F760F5FD9 alias:          pci:v000010DEd00000E00sv*sd*bc04sc80i00* alias:          pci:v000010DEd*sv*sd*bc03sc02i00* alias:          pci:v000010DEd*sv*sd*bc03sc00i00* depends:        drm,drm_kms_helper,nvidia-modeset retpoline:      Y vermagic:       4.4.138-59-default SMP mod_unload modversions parm:           modeset:Enable atomic kernel modesetting (1 = enable, 0 = disable (default)) (bool)

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

On Sunday, 15 July 2018 23:59:33 ACST Daniel Bauer wrote:

On 15.07.2018 16:13, Andrei Borzenkov wrote:

...

15.07.2018 15:55, Daniel Bauer пишет:

...

On 15.07.2018 14:38, Andrei Borzenkov wrote:

...

15.07.2018 14:47, Daniel Bauer пишет:

...

On 15.07.2018 12:28, Andrei Borzenkov wrote:

...

15.07.2018 12:58, Daniel Bauer пишет: > I start a new thread to not mix up things. > > I have an Asus GL552V laptop, i7 6700, optimus graphics intel/nvidia > running opensuse leap 42.3, KDE > > The nvidia drivers are installed, but they are not loaded. I have no > idea why.

Show dmesg output immediately after boot.

after boot, kde-login, open console, su to root: dmesg > dmsg2.txt

https://www.daniel-bauer.com/test/dmsg2.txt

Does "modprobe nvidia" work? What shows

modinfo nvidia nvidia-drm

meitli:~ # modprobe nvidia modprobe: ERROR: could not find module by name='nvidia' modprobe: ERROR: could not insert 'nvidia': Function not implemented

meitli:~ # modinfo nvidia nvidia-drm modinfo: ERROR: Module nvidia not found.

Something is seriously wrong. What

rpm -qa '*nvidia*' '*kernel*'

says?

yes :-(

meitli:~ # rpm -qa '*nvidia*' '*kernel*' nvidia-glG04-390.67-8.1.x86_64 kernel-devel-4.4.138-59.1.noarch kernel-default-devel-4.4.138-59.1.x86_64 nvidia-computeG04-390.67-8.1.x86_64 kernel-firmware-20170530-20.1.noarch kernel-devel-4.4.132-53.1.noarch kernel-default-4.4.132-53.1.x86_64 kernel-macros-4.4.138-59.1.noarch kernel-default-devel-4.4.132-53.1.x86_64 kernel-default-4.4.138-59.1.x86_64 x11-video-nvidiaG04-390.67-8.1.x86_64 nvidia-gfxG04-kmp-default-390.67_k4.4.76_1-8.1.x86_64

The latest stable nvidia driver version is 396.24. Have you tried downloading and installing using the nvidia installer instead of the repo version? You'll need to have at least the kernel headers installed to do that. The drawback is that, unless you also setup dkms so that the nvidia drivers are automatically rebuilt when you reboot with an updated kernel, you'll need to manually rebuild and install the modules after a kernel update (and, if you build the drivers into initrd, you'll need to rebuild that too for the new kernel). The latest driver releases are announced here; https://devtalk.nvidia.com/default/topic/533434/linux/current-graphics-drive... You may also find information specific to bumblebee and the optimus chipset on the nvidia forums (although I haven't looked specifically for that). Is there a setting in the bios to choose which gpu to use by default? My guess is that the intel chipset is enabled by default at boot and the nvidia gpu is powered off, thus causing the nvidia modules to fail to load (since they're obviously not compatible with the intel embedded gpu). It may also be worth trying to include the nvidia drivers in initrd using dracut. This thread may be useful: https://forums.opensuse.org/showthread.php/529420-Problems-adding-nvidia-pro... Of course, the standard disclaimer applies: YMMV (your mileage may vary). None of the above is guaranteed to work, but it may be worth a try. -- ============================================================== Rodney Baker VK5ZTV rodney.baker@iinet.net.au CCNA #CSCO12880208 ============================================================== -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org

On 15.07.2018 20:02, Rodney Baker wrote:

On Sunday, 15 July 2018 23:59:33 ACST Daniel Bauer wrote:

...

On 15.07.2018 16:13, Andrei Borzenkov wrote:

...

15.07.2018 15:55, Daniel Bauer пишет:

...

On 15.07.2018 14:38, Andrei Borzenkov wrote:

...

15.07.2018 14:47, Daniel Bauer пишет:

...

On 15.07.2018 12:28, Andrei Borzenkov wrote: > 15.07.2018 12:58, Daniel Bauer пишет: >> I start a new thread to not mix up things. >> >> I have an Asus GL552V laptop, i7 6700, optimus graphics intel/nvidia >> running opensuse leap 42.3, KDE >> >> The nvidia drivers are installed, but they are not loaded. I have no >> idea why. > > Show dmesg output immediately after boot.

after boot, kde-login, open console, su to root: dmesg > dmsg2.txt

https://www.daniel-bauer.com/test/dmsg2.txt

Does "modprobe nvidia" work? What shows

modinfo nvidia nvidia-drm

meitli:~ # modprobe nvidia modprobe: ERROR: could not find module by name='nvidia' modprobe: ERROR: could not insert 'nvidia': Function not implemented

meitli:~ # modinfo nvidia nvidia-drm modinfo: ERROR: Module nvidia not found.

Something is seriously wrong. What

rpm -qa '*nvidia*' '*kernel*'

says?

yes :-(

meitli:~ # rpm -qa '*nvidia*' '*kernel*' nvidia-glG04-390.67-8.1.x86_64 kernel-devel-4.4.138-59.1.noarch kernel-default-devel-4.4.138-59.1.x86_64 nvidia-computeG04-390.67-8.1.x86_64 kernel-firmware-20170530-20.1.noarch kernel-devel-4.4.132-53.1.noarch kernel-default-4.4.132-53.1.x86_64 kernel-macros-4.4.138-59.1.noarch kernel-default-devel-4.4.132-53.1.x86_64 kernel-default-4.4.138-59.1.x86_64 x11-video-nvidiaG04-390.67-8.1.x86_64 nvidia-gfxG04-kmp-default-390.67_k4.4.76_1-8.1.x86_64

The latest stable nvidia driver version is 396.24. Have you tried downloading and installing using the nvidia installer instead of the repo version? You'll need to have at least the kernel headers installed to do that. The drawback is that, unless you also setup dkms so that the nvidia drivers are automatically rebuilt when you reboot with an updated kernel, you'll need to manually rebuild and install the modules after a kernel update (and, if you build the drivers into initrd, you'll need to rebuild that too for the new kernel).

The latest driver releases are announced here;

https://devtalk.nvidia.com/default/topic/533434/linux/current-graphics-drive...

Honestly, I am afraid of that. In old times (really long ago) I had the nvidia drivers directly from nvidia on a desktop machine, and, as you said, with every kernel update, I had to do a lot of install work again, with things that I followed line by line in my thick Linux-book without actually understanding it. I'm just a user... Nowadays, with so many kernel and other updates all the time and with so many changes that make online-instructions obsolete almost the moment they are published, I cannot risk that. Or I am not brace enough.

You may also find information specific to bumblebee and the optimus chipset on the nvidia forums (although I haven't looked specifically for that).

Is there a setting in the bios to choose which gpu to use by default? My guess is that the intel chipset is enabled by default at boot and the nvidia gpu is powered off, thus causing the nvidia modules to fail to load (since they're obviously not compatible with the intel embedded gpu).

Unfortunately there is nothing about that in my BIOS.

It may also be worth trying to include the nvidia drivers in initrd using dracut. This thread may be useful:

https://forums.opensuse.org/showthread.php/529420-Problems-adding-nvidia-pro...

I'll take a look.

Of course, the standard disclaimer applies: YMMV (your mileage may vary). None of the above is guaranteed to work, but it may be worth a try.

:-) -- Daniel Bauer photographer Basel Málaga https://www.patreon.com/danielbauer http://www.daniel-bauer.com -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org

On 15.07.2018 20:26, Andrei Borzenkov wrote:

15.07.2018 17:29, Daniel Bauer пишет:

...

On 15.07.2018 16:13, Andrei Borzenkov wrote:

...

15.07.2018 15:55, Daniel Bauer пишет:

...

On 15.07.2018 14:38, Andrei Borzenkov wrote:

...

15.07.2018 14:47, Daniel Bauer пишет:

...

On 15.07.2018 12:28, Andrei Borzenkov wrote: > 15.07.2018 12:58, Daniel Bauer пишет: >> I start a new thread to not mix up things. >> >> I have an Asus GL552V laptop, i7 6700, optimus graphics intel/nvidia >> running opensuse leap 42.3, KDE >> >> The nvidia drivers are installed, but they are not loaded. I have no >> idea why. >> > > Show dmesg output immediately after boot. >

after boot, kde-login, open console, su to root: dmesg > dmsg2.txt

https://www.daniel-bauer.com/test/dmsg2.txt

Does "modprobe nvidia" work? What shows

modinfo nvidia nvidia-drm

meitli:~ # modprobe nvidia modprobe: ERROR: could not find module by name='nvidia' modprobe: ERROR: could not insert 'nvidia': Function not implemented

meitli:~ # modinfo nvidia nvidia-drm modinfo: ERROR: Module nvidia not found.

Something is seriously wrong. What

rpm -qa '*nvidia*' '*kernel*'

says?

yes :-(

meitli:~ # rpm -qa '*nvidia*' '*kernel*' nvidia-glG04-390.67-8.1.x86_64 kernel-devel-4.4.138-59.1.noarch kernel-default-devel-4.4.138-59.1.x86_64 nvidia-computeG04-390.67-8.1.x86_64 kernel-firmware-20170530-20.1.noarch kernel-devel-4.4.132-53.1.noarch kernel-default-4.4.132-53.1.x86_64 kernel-macros-4.4.138-59.1.noarch kernel-default-devel-4.4.132-53.1.x86_64 kernel-default-4.4.138-59.1.x86_64 x11-video-nvidiaG04-390.67-8.1.x86_64 nvidia-gfxG04-kmp-default-390.67_k4.4.76_1-8.1.x86_64

This looks OK. And what

find /lib/modules -name 'nvidia*'

says?

meitli:~ # find /lib/modules -name 'nvidia*' /lib/modules/4.4.138-59-default/weak-updates/updates/nvidia-drm.ko /lib/modules/4.4.138-59-default/weak-updates/updates/nvidia-modeset.ko /lib/modules/4.4.138-59-default/kernel/drivers/net/ethernet/nvidia /lib/modules/4.4.76-1-default/updates/nvidia-uvm.ko /lib/modules/4.4.76-1-default/updates/nvidia.ko /lib/modules/4.4.76-1-default/updates/nvidia-drm.ko /lib/modules/4.4.76-1-default/updates/nvidia-modeset.ko /lib/modules/4.4.132-53-default/weak-updates/updates/nvidia-drm.ko /lib/modules/4.4.132-53-default/weak-updates/updates/nvidia-modeset.ko /lib/modules/4.4.132-53-default/kernel/drivers/net/ethernet/nvidia -- Daniel Bauer photographer Basel Málaga https://www.patreon.com/danielbauer http://www.daniel-bauer.com -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org

On 2018-07-15 20:33, Daniel Bauer wrote:

On 15.07.2018 20:26, Andrei Borzenkov wrote:

...

This looks OK. And what

find /lib/modules -name 'nvidia*'

says?

meitli:~ # find /lib/modules -name 'nvidia*' /lib/modules/4.4.138-59-default/weak-updates/updates/nvidia-drm.ko /lib/modules/4.4.138-59-default/weak-updates/updates/nvidia-modeset.ko /lib/modules/4.4.138-59-default/kernel/drivers/net/ethernet/nvidia /lib/modules/4.4.76-1-default/updates/nvidia-uvm.ko /lib/modules/4.4.76-1-default/updates/nvidia.ko /lib/modules/4.4.76-1-default/updates/nvidia-drm.ko /lib/modules/4.4.76-1-default/updates/nvidia-modeset.ko /lib/modules/4.4.132-53-default/weak-updates/updates/nvidia-drm.ko /lib/modules/4.4.132-53-default/weak-updates/updates/nvidia-modeset.ko /lib/modules/4.4.132-53-default/kernel/drivers/net/ethernet/nvidia

For comparison, this is mine: Telcontar:~ # find /lib/modules -name 'nvidia*' /lib/modules/4.4.136-56-default/kernel/drivers/net/ethernet/nvidia /lib/modules/4.4.136-56-default/weak-updates/updates/nvidia.ko /lib/modules/4.4.136-56-default/weak-updates/updates/nvidia-uvm.ko /lib/modules/4.4.138-59-default/kernel/drivers/net/ethernet/nvidia /lib/modules/4.4.138-59-default/weak-updates/updates/nvidia.ko /lib/modules/4.4.138-59-default/weak-updates/updates/nvidia-uvm.ko /lib/modules/scripts/nvidia /lib/modules/4.4.76-1-default/updates/nvidia.ko /lib/modules/4.4.76-1-default/updates/nvidia-uvm.ko Telcontar:~ # The one we are looking for is "/lib/modules/4.4.76-1-default/updates/nvidia.ko" Then on the current one, there is a symlink: Telcontar:~ # uname -a Linux Telcontar 4.4.138-59-default #1 SMP Mon Jun 18 13:48:42 UTC 2018 (f0b8f6b) x86_64 x86_64 x86_64 GNU/Linux Telcontar:~ # l /lib/modules/4.4.138-59-default/weak-updates/updates/nvidia.ko lrwxrwxrwx 1 root root 47 Jun 28 16:01 /lib/modules/4.4.138-59-default/weak-updates/updates/nvidia.ko -> /lib/modules/4.4.76-1-default/updates/nvidia.ko Telcontar:~ # And you have it. -- Cheers / Saludos, Carlos E. R. (from 42.3 x86_64 "Malachite" at Telcontar)

15.07.2018 21:33, Daniel Bauer пишет: ...

...

...

meitli:~ # rpm -qa '*nvidia*' '*kernel*' nvidia-glG04-390.67-8.1.x86_64 kernel-devel-4.4.138-59.1.noarch kernel-default-devel-4.4.138-59.1.x86_64 nvidia-computeG04-390.67-8.1.x86_64 kernel-firmware-20170530-20.1.noarch kernel-devel-4.4.132-53.1.noarch kernel-default-4.4.132-53.1.x86_64 kernel-macros-4.4.138-59.1.noarch kernel-default-devel-4.4.132-53.1.x86_64 kernel-default-4.4.138-59.1.x86_64 x11-video-nvidiaG04-390.67-8.1.x86_64 nvidia-gfxG04-kmp-default-390.67_k4.4.76_1-8.1.x86_64

This looks OK. And what

find /lib/modules -name 'nvidia*'

says?

meitli:~ # find /lib/modules -name 'nvidia*' /lib/modules/4.4.138-59-default/weak-updates/updates/nvidia-drm.ko /lib/modules/4.4.138-59-default/weak-updates/updates/nvidia-modeset.ko /lib/modules/4.4.138-59-default/kernel/drivers/net/ethernet/nvidia /lib/modules/4.4.76-1-default/updates/nvidia-uvm.ko /lib/modules/4.4.76-1-default/updates/nvidia.ko /lib/modules/4.4.76-1-default/updates/nvidia-drm.ko /lib/modules/4.4.76-1-default/updates/nvidia-modeset.ko /lib/modules/4.4.132-53-default/weak-updates/updates/nvidia-drm.ko /lib/modules/4.4.132-53-default/weak-updates/updates/nvidia-modeset.ko /lib/modules/4.4.132-53-default/kernel/drivers/net/ethernet/nvidia

Well, somehow it fails to add weak updates. Download package nvidia-gfxG04-kmp-default-390.67_k4.4.76_1-8.1.x86_64.rpm from repository on nVidia site and then do under script command to capture full output rpm -e --nodeps nvidia-gfxG04-kmp-default-390.67_k4.4.76_1-8.1.x86_64 find /lib/modules -name 'nvidia*' rpm -ivvh ./nvidia-gfxG04-kmp-default-390.67_k4.4.76_1-8.1.x86_64.rpm find /lib/modules -name 'nvidia*' (whatever path you have downloaded package in) The "rpm -ivv" will produce very verbose output. Upload file typescript created by script command somewhere. -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org

El 2018-07-15 a las 20:51 +0200, Daniel Bauer escribió:

...

I would reinstall the rpm "nvidia-glG03-340.106-11.1.x86_64" from the command line, (zypper) to see the exact output it prints. Maybe there is some error message that you do not see in YaST. Paste it here.

When I remove the G04 drivers, automatically the G03-drivers are installed - BUT: suse-prime is removed. When I install suse-prime automatically the G04 drivers are installed...

I don't know why suse-prime insists on G04, but I guess there is a reason ?

I don't understant that logic. Do what Andrei has written at 21 hours. You will have to remove packages using rpm directly, not zypper, because as you see zypper tries to resolve deps some other way. You have to manually download the rpms to a directory previously. I think you may run the commands to get the output to paste this way: rm outputlogfile program 2>&1 | tee -a outputlogfile echo >> outputlogfile echo "-----------------------" >> outputlogfile echo >> outputlogfile other_program 2>&1 | tee -a outputlogfile Finally: susepaste -n "Daniel" -t "rpm output" -e 40320 outputlogfile and paste here the link you obtain. Another method is in the terminal, select edit/select all, copy, then in an editor. But if the output is really verbose, the buffer may overflow and not capture everything. -- Cheers Carlos E. R. (from openSUSE 15.0 (Legolas))

* Carlos E. R. [07-15-18 17:49]:

El 2018-07-15 a las 20:51 +0200, Daniel Bauer escribió:

...

...

I would reinstall the rpm "nvidia-glG03-340.106-11.1.x86_64" from the command line, (zypper) to see the exact output it prints. Maybe there is some error message that you do not see in YaST. Paste it here.

When I remove the G04 drivers, automatically the G03-drivers are installed - BUT: suse-prime is removed. When I install suse-prime automatically the G04 drivers are installed...

I don't know why suse-prime insists on G04, but I guess there is a reason ?

I don't understant that logic.

Do what Andrei has written at 21 hours. You will have to remove packages using rpm directly, not zypper, because as you see zypper tries to resolve deps some other way.

no, not if you use "force" zypper -v rm -f <package> ps: that is incorrect, s/b zypper -v rm --force <package> -- (paka)Patrick Shanahan Plainfield, Indiana, USA @ptilopteri http://en.opensuse.org openSUSE Community Member facebook/ptilopteri Registered Linux User #207535 @ http://linuxcounter.net Photos: http://wahoo.no-ip.org/piwigo paka @ IRCnet freenode -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org

On 15.07.2018 21:00, Andrei Borzenkov wrote:

15.07.2018 21:33, Daniel Bauer пишет: ...

...

...

...

meitli:~ # rpm -qa '*nvidia*' '*kernel*' nvidia-glG04-390.67-8.1.x86_64 kernel-devel-4.4.138-59.1.noarch kernel-default-devel-4.4.138-59.1.x86_64 nvidia-computeG04-390.67-8.1.x86_64 kernel-firmware-20170530-20.1.noarch kernel-devel-4.4.132-53.1.noarch kernel-default-4.4.132-53.1.x86_64 kernel-macros-4.4.138-59.1.noarch kernel-default-devel-4.4.132-53.1.x86_64 kernel-default-4.4.138-59.1.x86_64 x11-video-nvidiaG04-390.67-8.1.x86_64 nvidia-gfxG04-kmp-default-390.67_k4.4.76_1-8.1.x86_64

This looks OK. And what

find /lib/modules -name 'nvidia*'

says?

meitli:~ # find /lib/modules -name 'nvidia*' /lib/modules/4.4.138-59-default/weak-updates/updates/nvidia-drm.ko /lib/modules/4.4.138-59-default/weak-updates/updates/nvidia-modeset.ko /lib/modules/4.4.138-59-default/kernel/drivers/net/ethernet/nvidia /lib/modules/4.4.76-1-default/updates/nvidia-uvm.ko /lib/modules/4.4.76-1-default/updates/nvidia.ko /lib/modules/4.4.76-1-default/updates/nvidia-drm.ko /lib/modules/4.4.76-1-default/updates/nvidia-modeset.ko /lib/modules/4.4.132-53-default/weak-updates/updates/nvidia-drm.ko /lib/modules/4.4.132-53-default/weak-updates/updates/nvidia-modeset.ko /lib/modules/4.4.132-53-default/kernel/drivers/net/ethernet/nvidia

Well, somehow it fails to add weak updates. Download package nvidia-gfxG04-kmp-default-390.67_k4.4.76_1-8.1.x86_64.rpm from repository on nVidia site and then do under script command to capture full output

rpm -e --nodeps nvidia-gfxG04-kmp-default-390.67_k4.4.76_1-8.1.x86_64 find /lib/modules -name 'nvidia*' rpm -ivvh ./nvidia-gfxG04-kmp-default-390.67_k4.4.76_1-8.1.x86_64.rpm find /lib/modules -name 'nvidia*'

(whatever path you have downloaded package in)

The "rpm -ivv" will produce very verbose output. Upload file typescript created by script command somewhere.

ok, I just did that. The output of the 4 commands is here: https://www.daniel-bauer.com/test/nvidia_rpm_output.txt -- Daniel Bauer photographer Basel Málaga https://www.patreon.com/danielbauer http://www.daniel-bauer.com -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org

On Mon, Jul 16, 2018 at 11:26 AM, Daniel Bauer wrote:

...

Well, somehow it fails to add weak updates. Download package nvidia-gfxG04-kmp-default-390.67_k4.4.76_1-8.1.x86_64.rpm from repository on nVidia site and then do under script command to capture full output

rpm -e --nodeps nvidia-gfxG04-kmp-default-390.67_k4.4.76_1-8.1.x86_64 find /lib/modules -name 'nvidia*' rpm -ivvh ./nvidia-gfxG04-kmp-default-390.67_k4.4.76_1-8.1.x86_64.rpm find /lib/modules -name 'nvidia*'

(whatever path you have downloaded package in)

The "rpm -ivv" will produce very verbose output. Upload file typescript created by script command somewhere.

ok, I just did that. The output of the 4 commands is here: https://www.daniel-bauer.com/test/nvidia_rpm_output.txt

rpm -e --nodeps nvidia-gfxG04-kmp-default-390.67_k4.4.76_1-8.1.x86_64

/usr/src/kernel-modules/nvidia-390.67-default / rm -f -r conftest make[1]: Entering directory '/usr/src/linux-4.4.138-59' make[2]: Entering directory '/usr/src/linux-4.4.138-59-obj/x86_64/default' CLEAN /usr/src/kernel-modules/nvidia-390.67-default CLEAN /usr/src/kernel-modules/nvidia-390.67-default/.tmp_versions CLEAN /usr/src/kernel-modules/nvidia-390.67-default/Module.symvers make[2]: Leaving directory '/usr/src/linux-4.4.138-59-obj/x86_64/default' make[1]: Leaving directory '/usr/src/linux-4.4.138-59' / update-alternatives: warning: alternative /usr/lib/nvidia/alternate-install-present-default (part of link group alternate-install-present) doesn't exist; removing from list of alternatives update-alternatives: warning: /etc/alternatives/alternate-install-present is dangling; it will be updated with best choice

-----------------------

find /lib/modules -name 'nvidia*'

/lib/modules/4.4.138-59-default/weak-updates/updates/nvidia-drm.ko /lib/modules/4.4.138-59-default/weak-updates/updates/nvidia-modeset.ko /lib/modules/4.4.138-59-default/kernel/drivers/net/ethernet/nvidia /lib/modules/4.4.132-53-default/weak-updates/updates/nvidia-drm.ko /lib/modules/4.4.132-53-default/weak-updates/updates/nvidia-modeset.ko /lib/modules/4.4.132-53-default/kernel/drivers/net/ethernet/nvidia

-----------------------

Well, that is the problem; these files must be symlinks to base modules which they are apparently not. Having output below would be interesting to confirm: ls -l /lib/modules/4.4.138-59-default/weak-updates/updates/nvidia-drm.ko /lib/modules/4.4.138-59-default/weak-updates/updates/nvidia-modeset.ko /lib/modules/4.4.132-53-default/weak-updates/updates/nvidia-drm.ko /lib/modules/4.4.132-53-default/weak-updates/updates/nvidia-modeset.ko

+ echo 'Warning: /lib/modules/4.4.138-59-default is inconsistent' Warning: /lib/modules/4.4.138-59-default is inconsistent + echo 'Warning: weak-updates symlinks might not be created' Warning: weak-updates symlinks might not be created

So try now 1. Remove package again 2. Delete files /lib/modules/4.4.138-59-default/weak-updates/updates/nvidia-drm.ko /lib/modules/4.4.138-59-default/weak-updates/updates/nvidia-modeset.ko /lib/modules/4.4.132-53-default/weak-updates/updates/nvidia-drm.ko /lib/modules/4.4.132-53-default/weak-updates/updates/nvidia-modeset.ko 3. Run (to be on safe side) depmod -a 4.4.132-53-default depmod -a 4.4.138-59-default 4. Reinstall package, verify that you have correct symlinks now. -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org

On 07/16/2018 03:26 AM, Daniel Bauer wrote:

Well, somehow it fails to add weak updates. Download package nvidia-gfxG04-kmp-default-390.67_k4.4.76_1-8.1.x86_64.rpm from repository on nVidia site and then do under script command to capture full output

I'm not sure what is going on, but from my earlier post looking at what was interesting in the dmesg output, I picked up on: [ 2.250135] nvidia_modeset: Unknown symbol nv_kthread_q_flush (err 0) [ 2.250168] nvidia_modeset: Unknown symbol nvidia_register_module (err 0) [ 2.250182] nvidia_modeset: Unknown symbol nv_kthread_q_schedule_q_item (err 0) [ 2.250204] nvidia_modeset: Unknown symbol nvidia_get_rm_ops (err 0) [ 2.250225] nvidia_modeset: Unknown symbol nv_kthread_q_item_init (err 0) [ 2.250235] nvidia_modeset: Unknown symbol nv_kthread_q_stop (err 0) [ 2.250243] nvidia_modeset: Unknown symbol nvidia_unregister_module (err 0) [ 2.250265] nvidia_modeset: Unknown symbol nv_kthread_q_init (err 0) Now from the current output Daniel posted, it indicates those will be provided by the driver: D: Requires: ksym(default:nv_kthread_q_flush) = a5a6ef41 YES (added provide) D: Requires: ksym(default:nv_kthread_q_init) = 1ad58eb YES (added provide) D: Requires: ksym(default:nv_kthread_q_item_init) = 519bbbeb YES (added provide) D: Requires: ksym(default:nv_kthread_q_run_self_test) = dc74e56f YES (added provide) D: Requires: ksym(default:nv_kthread_q_schedule_q_item) = 6f8d1d81 YES (added provide) D: Requires: ksym(default:nv_kthread_q_stop) = b3ad9cd7 YES (added provide) D: Requires: ksym(default:nvidia_get_rm_ops) = db5880e5 YES (added provide) D: Requires: ksym(default:nvidia_register_module) = e7da488 YES (added provide) D: Requires: ksym(default:nvidia_unregister_module) = 6f96aacb YES (added provide) But for some reason appear not to be found: ++ /sbin/depmod -b / -ae -E /var/tmp/weak-modules2.Wyd2zP/symvers-4.4.138-59-default 4.4.138-59-default + output='depmod: WARNING: //lib/modules/4.4.138-59-default/weak-updates/updates/nvidia-modeset.ko needs unknown symbol nv_kthread_q_flush depmod: WARNING: //lib/modules/4.4.138-59-default/weak-updates/updates/nvidia-modeset.ko needs unknown symbol nvidia_register_module depmod: WARNING: //lib/modules/4.4.138-59-default/weak-updates/updates/nvidia-modeset.ko needs unknown symbol nv_kthread_q_schedule_q_item depmod: WARNING: //lib/modules/4.4.138-59-default/weak-updates/updates/nvidia-modeset.ko needs unknown symbol nvidia_get_rm_ops depmod: WARNING: //lib/modules/4.4.138-59-default/weak-updates/updates/nvidia-modeset.ko needs unknown symbol nv_kthread_q_item_init depmod: WARNING: //lib/modules/4.4.138-59-default/weak-updates/updates/nvidia-modeset.ko needs unknown symbol nv_kthread_q_stop depmod: WARNING: //lib/modules/4.4.138-59-default/weak-updates/updates/nvidia-modeset.ko needs unknown symbol nvidia_unregister_module depmod: WARNING: //lib/modules/4.4.138-59-default/weak-updates/updates/nvidia-modeset.ko needs unknown symbol nv_kthread_q_init' ... This is just a guess looking at what parts of the dmesg and current output don't make sense. Hopefully this will all be corrected by the deletion suggested by Andrei. -- David C. Rankin, J.D.,P.E. -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org

On 16.07.2018 10:49, Andrei Borzenkov wrote

Well, that is the problem; these files must be symlinks to base modules which they are apparently not. Having output below would be interesting to confirm:

ls -l /lib/modules/4.4.138-59-default/weak-updates/updates/nvidia-drm.ko /lib/modules/4.4.138-59-default/weak-updates/updates/nvidia-modeset.ko /lib/modules/4.4.132-53-default/weak-updates/updates/nvidia-drm.ko /lib/modules/4.4.132-53-default/weak-updates/updates/nvidia-modeset.ko

...

+ echo 'Warning: /lib/modules/4.4.138-59-default is inconsistent' Warning: /lib/modules/4.4.138-59-default is inconsistent + echo 'Warning: weak-updates symlinks might not be created' Warning: weak-updates symlinks might not be created

So try now

1. Remove package again 2. Delete files

/lib/modules/4.4.138-59-default/weak-updates/updates/nvidia-drm.ko /lib/modules/4.4.138-59-default/weak-updates/updates/nvidia-modeset.ko /lib/modules/4.4.132-53-default/weak-updates/updates/nvidia-drm.ko /lib/modules/4.4.132-53-default/weak-updates/updates/nvidia-modeset.ko

3. Run (to be on safe side)

depmod -a 4.4.132-53-default depmod -a 4.4.138-59-default

4. Reinstall package, verify that you have correct symlinks now.

So I did: rpm -e --nodeps nvidia-gfxG04-kmp-default-390.67_k4.4.76_1-8.1.x86_64 rm /lib/modules/4.4.138-59-default/weak-updates/updates/nvidia-drm.ko rm /lib/modules/4.4.138-59-default/weak-updates/updates/nvidia-modeset.ko rm /lib/modules/4.4.132-53-default/weak-updates/updates/nvidia-drm.ko rm /lib/modules/4.4.132-53-default/weak-updates/updates/nvidia-modeset.ko depmod -a 4.4.132-53-default depmod -a 4.4.138-59-default rpm -ivvh ./nvidia-gfxG04-kmp-default-390.67_k4.4.76_1-8.1.x86_64.rpm the output of the last command is here: https://www.daniel-bauer.com/test/nvidia_rpm_output_2.txt how do I verify the symlinks? ls -l /lib/modules/4.4.138-59-default/weak-updates/updates/nvidia-drm.ko still gives: lrwxrwxrwx 1 root root 51 Jul 16 11:16 /lib/modules/4.4.138-59-default/weak-updates/updates/nvidia-drm.ko -> /lib/modules/4.4.76-1-default/updates/nvidia-drm.ko -- Daniel Bauer photographer Basel Málaga https://www.patreon.com/danielbauer http://www.daniel-bauer.com -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org

On 2018-07-16 11:29, Daniel Bauer wrote:

On 16.07.2018 10:49, Andrei Borzenkov wrote

...

rpm -ivvh ./nvidia-gfxG04-kmp-default-390.67_k4.4.76_1-8.1.x86_64.rpm

the output of the last command is here: https://www.daniel-bauer.com/test/nvidia_rpm_output_2.txt

how do I verify the symlinks?

I would reboot :-) -- Cheers / Saludos, Carlos E. R. (from 42.3 x86_64 "Malachite" at Telcontar)

On 16.07.2018 11:53, Andrei Borzenkov wrote:

On Mon, Jul 16, 2018 at 12:45 PM, Carlos E. R. wrote:

...

On 2018-07-16 11:29, Daniel Bauer wrote:

...

On 16.07.2018 10:49, Andrei Borzenkov wrote

...

...

rpm -ivvh ./nvidia-gfxG04-kmp-default-390.67_k4.4.76_1-8.1.x86_64.rpm

the output of the last command is here: https://www.daniel-bauer.com/test/nvidia_rpm_output_2.txt

how do I verify the symlinks?

I would reboot :-)

I did

modinfo nvidia

modinfo nvidia filename: /lib/modules/4.4.138-59-default/weak-updates/updates/nvidia.ko alias: char-major-195-* version: 390.67 supported: external license: NVIDIA srcversion: 209B1D0CB123DE466F700AD alias: pci:v000010DEd00000E00sv*sd*bc04sc80i00* alias: pci:v000010DEd*sv*sd*bc03sc02i00* alias: pci:v000010DEd*sv*sd*bc03sc00i00* depends: ipmi_msghandler retpoline: Y vermagic: 4.4.138-59-default SMP mod_unload modversions parm: NVreg_Mobile:int parm: NVreg_ResmanDebugLevel:int parm: NVreg_RmLogonRC:int parm: NVreg_ModifyDeviceFiles:int parm: NVreg_DeviceFileUID:int parm: NVreg_DeviceFileGID:int parm: NVreg_DeviceFileMode:int parm: NVreg_UpdateMemoryTypes:int parm: NVreg_InitializeSystemMemoryAllocations:int parm: NVreg_UsePageAttributeTable:int parm: NVreg_MapRegistersEarly:int parm: NVreg_RegisterForACPIEvents:int parm: NVreg_CheckPCIConfigSpace:int parm: NVreg_EnablePCIeGen3:int parm: NVreg_EnableMSI:int parm: NVreg_TCEBypassMode:int parm: NVreg_UseThreadedInterrupts:int parm: NVreg_EnableStreamMemOPs:int parm: NVreg_EnableBacklightHandler:int parm: NVreg_EnableUserNUMAManagement:int parm: NVreg_EnableIBMNPURelaxedOrderingMode:int parm: NVreg_MemoryPoolSize:int parm: NVreg_IgnoreMMIOCheck:int parm: NVreg_RegistryDwords:charp parm: NVreg_RegistryDwordsPerDevice:charp parm: NVreg_RmMsg:charp parm: NVreg_AssignGpus:charp So, I think the modules are here and loaded... Now how to continue with prime-select? If I run prime-select nvidia (after reboot) the script runs but then something (?) crashes - a message appears on screen but due to the flickering I could not read it. I cannot logout anymore, nor reboot with the kde-menu-button. I rebooted with shutdown from console... -- Daniel Bauer photographer Basel Málaga https://www.patreon.com/danielbauer http://www.daniel-bauer.com -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org

* Daniel Bauer [07-16-18 05:01]:

On 16.07.2018 10:49, Andrei Borzenkov wrote:

[...]

...

4. Reinstall package, verify that you have correct symlinks now.

rpm -e --nodeps nvidia-gfxG04-kmp-default-390.67_k4.4.76_1-8.1.x86_64 ?

zypper -v in nvidia-gfxG04-kmp-default-390.67_k4.4.76_1-8.1.x86_64 rpm -e deletes package, not install and --nodeps does not check that other packages may be affected -- (paka)Patrick Shanahan Plainfield, Indiana, USA @ptilopteri http://en.opensuse.org openSUSE Community Member facebook/ptilopteri Registered Linux User #207535 @ http://linuxcounter.net Photos: http://wahoo.no-ip.org/piwigo paka @ IRCnet freenode -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org

On 2018-07-16 13:34, Patrick Shanahan wrote:

* Daniel Bauer <> [07-16-18 05:01]:

...

On 16.07.2018 10:49, Andrei Borzenkov wrote:

[...]

...

...

4. Reinstall package, verify that you have correct symlinks now.

rpm -e --nodeps nvidia-gfxG04-kmp-default-390.67_k4.4.76_1-8.1.x86_64 ?

zypper -v in nvidia-gfxG04-kmp-default-390.67_k4.4.76_1-8.1.x86_64

rpm -e deletes package, not install and --nodeps does not check that other packages may be affected

Yes, exactly what was intended. First remove with no checks, then some manual file removes and checks, then another step that was not in the quote, install again with rpm command from local file, with certain options to get a very verbose output ("rpm -ivv"). No, zypper is not adequate for this particular set of operations, sorry. -- Cheers / Saludos, Carlos E. R. (from 42.3 x86_64 "Malachite" at Telcontar)