Kernel-default

dagring · January 27, 2020, 4:45pm

Hi,

I had som problems with GPU drivers. I installed the pro versions of both nvidia and amdgpu. It’s a long story and will not bother you with that. I manged to get rid of the drivers, but i could not boot into the graphical interface. I tried to upgrade the newest kernel, but didn’t succeed. I got a message right before plasma should be loaded that kernel modules could not be loaded. My solution was to load an earlier kernel:

dagr@opensuse:~> uname -r
4.12.14-lp151.28.32-default

Now I can boot into plasma, and I’m using the nouvau driver. It seems to work fearly well, but it does not update well. The webpages can be a little scrobbly. My main concern is to update to the latest kernel, which in yast says it’s version 4.12.14-lp151.28.36.1. How can I do this and keep the setup which I have in kernel 28.32-1? When I now boot up the machine , I have to scroll down to the 28.32.1-kernel and chose this, I want to replace the old 4.12.14-lp151.28.36.1 with a kernel with the same version number, but works with my setup and load all kernels.

Dag R

nrickert · January 27, 2020, 5:01pm

Perhaps you should file a bug report:

https://en.opensuse.org/openSUSE:Submitting_bug_reports

The kernel maintainers are usually pretty good at handling this kind of problem. But you need to report the problem so that they know about it.

dagring · January 27, 2020, 5:17pm

Do you think it is a bug? I thought it was me messing it up withe several gpu drivers and changing gpus. Linux can be a little unstable if you mess around with different hardware and proprietary sotware. If you think it is a bug I will send a bug report.

Dag R

eng-int · January 27, 2020, 5:41pm

It may already have been fixed. You can check whether you still have problems with the latest stable kernel.
To do this ad the openSuse stable kernel:

 sudo zypper ar -f  http://ftp.gwdg.de/pub/opensuse/repositories/Kernel%3A/stable/standard/  Kernel-stable

Then use YaST Software Manager to add the kernel-default-5.4 and remove 4.12.14-lp151.28.36.1.
The reason for removing the “buggy” kernel is that by default the boot process only keeps the latest two kernels and purges the others.

dagring · January 27, 2020, 6:05pm

It occured an error when adding the repository to yast:

Det oppstod en feil under initalisering av pakkebrønner. [Kernel-stable|http://ftp.gwdg.de/pub/opensuse/repositories/Kernel%253A/stable/standard/] Valid metadata not found at specified URL Historie: - [Kernel-stable|http://ftp.gwdg.de/pub/opensuse/repositories/Kernel%253A/stable/standard/] Repository type can't be determined.

Here is the out put of the command:

sudo zypper ar -f  http://ftp.gwdg.de/pub/opensuse/repositories/Kernel%3A/stable/standard/  Kernel-stable
[sudo] passord for root: 
Legger til pakkebrønn 'Kernel-stable' ............................................................................................................................[fullført]
Pakkebrønnen 'Kernel-stable' ble lagt til

URI          : http://ftp.gwdg.de/pub/opensuse/repositories/Kernel%253A/stable/standard/
aktivert     : Ja                                                                       
GPG-kontroll : Ja                                                                       
Autooppdater : Ja                                                                       
Prioritet    : 99 (default priority)                                                    

Repository priorities in effect:                                                                                                            (See 'zypper lr -P' for details)
      90 (raised priority)  :  1 repository  
      99 (default priority) :  5 repositories

nrickert · January 27, 2020, 6:49pm

I’m not sure what went wrong there. But I see that “%3A” has changed to “%253A”. It looks as if it has percent-encoded the percent symbol.

I usually use:

http://download.opensuse.org/repositories/Kernel:/stable/standard/

for that repo (you seem to be using a mirror). It looks as if the percent-encoding confuses “zypper -ar”.

mrmazda · January 27, 2020, 7:07pm

Are you 100% unquestionably sure? NVidia driver installation messes with libs that must be restored to stock before FOSS drivers can work as expected.

Nouveau is the name of two different drivers, kernel nouveau module, and X nouveau (xf86-video-nouveau; DDX).

What GPUs do you have (marketing names + device IDs from lspci -nnk | grep -iA3 VGA)? Is this a laptop? Which X driver (DDX) causes “webpages can be a little scrobbly” (inxi -Gxx)?

eng-int · January 27, 2020, 7:46pm

My apologies. I did not notice that the “:” had transmuted to “%253A” while copy-pasting. You have already been given a working link.

The transformation happened during Plasma5 copy-paste. Intriguingly the version in the clipboard is the more common “%3A”, I think that it was the “25” that caused the problem.
And yes, I just happen to have a t tab open at the GWDG mirror at the moment.

dagring · January 27, 2020, 9:07pm

Here is the the output:

dagr@opensuse:~> sudo lspci -nnk | grep -iA3 VGA
[sudo] passord for root: 
01:00.0 VGA compatible controller [0300]: NVIDIA Corporation GM204 [GeForce GTX 970] [10de:13c2] (rev a1)
        Subsystem: Micro-Star International Co., Ltd. [MSI] Device [1462:3160]
        Kernel driver in use: nouveau
        Kernel modules: nouveau

I’m not sure if i have got rid of the prop drivers, but I tried my best to follow the procedure in the Opensuse wiki/documentation on the nvidia driver and I used the same procedure wwith the amdgpu pro.

Dag R

mrmazda · January 27, 2020, 10:49pm

Bugs reported about Nouveau trouble typically get resolved upstream with this boilerplate:

Thanks for your bug report!

Nouveau is an experimental driver under constant heavy development. This means that we cannot follow it closely, as we are not part of its team with  reverse engineered knowledge of NVIDIA cards.

In case you wish to stick with nouveau and to help us improve its support in openSUSE, you can try our latest kernel, Mesa, and xf86-video-nouveau packages:

  http://kernel.opensuse.org/
  https://build.opensuse.org/package/show/X11:XOrg/Mesa
  https://build.opensuse.org/package/show/X11:XOrg/xf86-video-nouveau

Testing the latest versions is a prerequisite in order to inform nouveau's upstream developers of any bugs you find:

  https://nouveau.freedesktop.org/wiki/Bugs/

Once you are aware of an upstream fix for your issue, please reopen the bug and  let us know. We will be happy to include it in your openSUSE distribution if it's technically feasible.

Alternatively, you can install NVIDIA's proprietary driver instead:

  https://en.opensuse.org/SDB:NVIDIA

That sudo lspci -nnk | grep -iA3 VGA output indicates a GeForce card 5-6 years old. It should work fine with a proper installation of NVidia’s drivers. However, they have a history of not playing nice with Plymouth. It’s probably worth trying again to install them but first either remove Plymouth altogether, or disable Plymouth via Grub cmdling option plymouth.enable=0. When (re-)installing, plymouth=0 is used instead.

dagring · January 27, 2020, 11:22pm

I did a update from a usb-stick because the change in kernels didn’t work out. Now it’s running in the 4.12.14xxxx which is the standard kernel for 15.1, but after the update I see the gpu are loading kernels from nvidia.

dagr@opensuse:~> sudo lspci -nnk | grep -iA3 VGA
[sudo] passord for root: 
01:00.0 VGA compatible controller [0300]: NVIDIA Corporation GM204 [GeForce GTX 970] [10de:13c2] (rev a1)
        Subsystem: Micro-Star International Co., Ltd. [MSI] Device [1462:3160]
        Kernel driver in use: nouveau
        Kernel modules: nouveau, nvidia_drm, nvidia

The initial qestion I asked concerned the kernel-default. It seems this is solved by doing the update and deleting the 5.4 kernel. The other question about the driver, it seems this is not fixed yet. As mrmazda said, the libs (?) must be restored to get the FOSS driver to work properly. I want first to get the noveau driver to work properly. I haven’t any use of 3D, so (as I have heard are the main difference between noveau and nvidia driver) the prop nvidia driver is not essential for me. I see some video and surf the net.

I would appreciate you guidance to solve this.

Dag R

mrmazda · January 28, 2020, 1:44am

To which nouveau do you here refer, kernel, or DDX? lspci is only showing the kernel involvement.

This is what any NVidia card using only FOSS should look approximately like:

# inxi -V | head -n1
inxi **3.0.37-00** (2019-11-19)
# inxi -GxxSM
System:    Host: p5bse Kernel: 4.12.14-lp151.28.36-default x86_64 bits: 64 compiler: gcc v: 7.4.1 Desktop: KDE 3 wm: kwin
           dm: N/A Distro: openSUSE Leap 15.1
Machine:   Type: Desktop Mobo: ASUSTeK model: P5B SE v: Rev 1.xx serial: MS6C79B32413550 BIOS: American Megatrends v: 1103
           date: 06/04/2009
Graphics:  Device-1: NVIDIA GF119 [NVS 310] vendor: Hewlett-Packard **driver: nouveau v: kernel** bus ID: 01:00.0
           chip ID: 10de:107d
           Display: server: **X.Org 1.20.3 driver: modesetting** unloaded: fbdev,vesa alternate: nouveau,nv,nvidia
           resolution: 1920x1200~60Hz
           OpenGL: renderer: llvmpipe (LLVM 7.0 128 bits) v: 3.3 Mesa 18.3.2 compat-v: 3.1 direct render: Yes
# lspci -nnk | grep -iA3 VGA
01:00.0 VGA compatible controller [0300]: NVIDIA Corporation GF119 [NVS 310] [10de:107d] (rev a1)
        Subsystem: Hewlett-Packard Company Device [103c:094e]
        Kernel driver in use: **nouveau**
        Kernel modules: **nouveau**
# lsmod | sort | grep vea
button                 16384  1 nouveau
drm                   491520  5 nouveau,ttm,drm_kms_helper
drm_kms_helper        204800  1 nouveau
i2c_algo_bit           16384  1 nouveau
mxm_wmi                16384  1 nouveau
nouveau              2150400  2
ttm                   126976  1 nouveau
video                  45056  1 nouveau
wmi                    28672  2 mxm_wmi,nouveau

Below is with xf86-video-nouveau installed, while above is without.

# inxi -Gxx
Graphics:  Device-1: NVIDIA GF119 [NVS 310] vendor: Hewlett-Packard **driver: nouveau v: kernel** bus ID: 01:00.0
           chip ID: 10de:107d
           Display: server: **X.Org 1.20.3 driver: nouveau** unloaded: fbdev,modesetting,vesa alternate: nv,nvidia
           resolution: 1920x1200~60Hz
           OpenGL: renderer: llvmpipe (LLVM 7.0 128 bits) v: 3.3 Mesa 18.3.2 compat-v: 3.1 direct render: Yes

Thus, it appears you still have some NVidia software eradication left to perform to have X utilize either the (reverse-engineered) Nouveau DDX or the (newer; upstream default) Modesetting DDX.

About eradicating NVidia all I know is the instructions for installation are supposed to provide instructions for removal that must be followed to restore FOSS DDX operation. What affected lib(s) this implies I have no idea.

dagring · January 28, 2020, 12:43pm

Ok. Thanks for all help on this matter. Since the thread involves several topics, and the main question is solved. I will make a new tread on the eradication of nvidia driver.

Dag R

hcvv · January 28, 2020, 12:56pm

Yes, a thread for each problem is much better. Reading the subject of this thread, already many people that are not realy Kernel adapts will not even open the thread.