how to upgrade debian 10 kernel from the backports without r

[L_V]

Where did you see you have an Intel graphic chipset ?
+

Code: Select all

systemctl list-unit-files | grep nvidia

dpkg -l | grep nouveau

[stevepusser]

More explicit output might be helpful...
run:

Code: Select all

nvidia-smi

and check "Persistence-M" parameter.
+

Code: Select all

systemctl status nvidia-persistenced-manual.service

I don't have nvidia graphics to help.
If nvidia-driver buster driver is really not compatible with kernel 5.20, I wonder why there is no apt compatibility issue.

+ did you install the linux-headers ? If no, try to see if it helps.

Enable unlimited scrollback on your terminal, or use a terminal that supports that, then run with sudo or as root in that terminal:

Code: Select all

dpkg-reconfigure nvidia-kernel-dkms

Edit: sorry, had the wrong command, now fixed.

At the end, you can scroll back up to where the error in building the driver occurred. I'm sure it's there, because the stock 418.74 driver won't build on a 5.2 kernel. Nothing of the other suggestions from users here will change that cold, hard fact.

[marietto]

lspci | grep Display
00:02.0 Display controller : intel corporation device

dpkg -l | grep nouveau
libdrm-m-nouveau2:amd64 2.4.97
libdrm-m-nouveau2:i386 2.4.97
xserver-xorg-video -nouveau 1:1.0.16

systemctl list-unit-files |grep nvidia | grep less
nvidia-persistenced.service enabled

dpkg-reconfigure nvidia-driver
nvidia-driver not installed

apt install nvidia-driver
nvidia-kernel-dkms
nvidia-driver
sub-process /usr/bin/dkpg returned an error

[stevepusser]

I'm setting up an OBS Buster repo with backported Debian 5.3.2 kernel, the latest firmware (which is also in buster-backports), nvidia-driver 435.21, and nvidia-settings 435.21. The same sort of backports are working for MX 19 buster-based users as well as for MX 18 stretch-based users, though I don't have any data (yet) if Mesa needs to be backported for your Intel card. You could try updating from my repo the non-kernel packages now--the kernel builds take many hours, and are touchy, sometimes failing for no good reason. You can check the "linux-source-5.3" package to see if it finished successfully.

https://build.opensuse.org/project/show ... ort-buster

[marietto]

write me here the commands that I need to give,thanks. In the while I've installed ubuntu 19.04 and I've configured it to make work the GPU pass,following this tutorial :

https://blog.zerosector.io/2018/07/28/k ... ssthrough/

I think that I've been able to do that,according to this output :

ziomario@ziomario-Z390-AORUS-PRO:~$ dmesg | grep -i vfio
[ 1.709915] VFIO - User Level meta-driver version: 0.3
[ 8.280538] vfio_pci: add [10de:1e04[ffffffff:ffffffff]] class 0x000000/00000000
[ 8.297402] vfio_pci: add [10de:10f7[ffffffff:ffffffff]] class 0x000000/00000000
[ 8.297404] vfio_pci: add [10de:1ad6[ffffffff:ffffffff]] class 0x000000/00000000
[ 8.297406] vfio_pci: add [10de:1ad7[ffffffff:ffffffff]] class 0x000000/00000000

but anyway,I prefer more debian than ubuntu. So I don't want to give up here. On ubuntu are working good the nvidia card on the monitor A and the intel chipset driver on monitor B.

[stevepusser]

Sorry, I messed up the command to tell the system to try and rebuild the nvidia kernel drivers. I fixed it in my original post, and the correct one is (as root or with sudo):

Code: Select all

dpkg-reconfigure nvidia-kernel-dkms

The 5.3.2 kernel built successfully in my OBS repo. Can you install inxi on any Linux you have running, as it should be in the standard repos, and give us the results of

Code: Select all

inxi -Fxxz

so I can check if you need any other updated drivers for a 5.3.2 kernel? I know if you want to run Virtual Box on a 5.3 kernel, 6.0.12 will work, but Buster's won't, and there's no version in buster-backports at all...so I will probably add a backport of 6.0.12 in that OBS repo as well.

The instructions to add that repo and its key to a Buster install are:

As root:

Code: Select all

echo 'deb http://download.opensuse.org/repositories/home:/stevenpusser:/new-hardware-support-buster/Debian_10/ /' > /etc/apt/sources.list.d/home:stevenpusser:new-hardware-support-buster.list

rm -rf Release.key && wget -nv https://download.opensuse.org/repositories/home:stevenpusser:new-hardware-support-buster/Debian_10/Release.key -O Release.key && apt-key add - < Release.key && rm -rf Release.key

apt-get update

Now any Nvidia driver packages, nvidia-settings, and firmware packages you have installed from Buster should appear as upgradeable. Newer kernel versions will not appear as automatic upgrades--they get installed alongside the stock Debian kernel. You can search in Synaptic for 5.3.0-1 to show my 5.3 kernel's packages; you want the linux-image and linux-headers packages for amd64.

I don't still know if your Intel GPU requires a backport of Mesa, but I expect we'll find out.

[marietto]

while I was doing : dpkg-reconfigure nvidia-kernel-dkms an error happened :

job for systemd-modules-load.service failed because the control process exited with error code
see systemctl status systemd-modules-load.service. and journalctl -xe for details

now the monitor A attached to the nvidia card is working. the monitor B,attached to the intel chipset is turned off. And debian is not able to recognize it on the control panel.

mario@DESKTOP-N9UN2H3:/home/mariozio# inxi -Fxxz
System:
Host: DESKTOP Kernel: 5.3.2 x86_64 bits: 64 compiler: gcc v: 8.3.0
Desktop: MATE 1.20.4 wm: gnome-shell dm: GDM3
Distro: Debian GNU/Linux 10 (buster)
Machine:
Type: Desktop System: Gigabyte product: Z390 AORUS PRO v: N/A serial: N/A
Mobo: Gigabyte model: Z390 AORUS PRO-CF v: x.x serial: N/A
UEFI [Legacy]: American Megatrends v: F10 date: 06/05/2019
CPU:
Topology: 8-Core model: Intel Core i9-9900K bits: 64 type: MT MCP
arch: Kaby Lake rev: D L1 cache: 512 KiB L2 cache: 16.0 MiB
L3 cache: 16.0 MiB
flags: avx avx2 lm nx pae sse sse2 sse3 sse4_1 sse4_2 ssse3 vmx
bogomips: 115200
Speed: 800 MHz min/max: 800/5000 MHz Core speeds (MHz): 1: 800 2: 800
3: 800 4: 800 5: 801 6: 801 7: 801 8: 801 9: 800 10: 800 11: 800 12: 800
13: 800 14: 800 15: 800 16: 800
Graphics:
Device-1: Intel vendor: Gigabyte driver: i915 v: kernel bus ID: 00:02.0
chip ID: 8086:3e98
Device-2: NVIDIA TU102 [GeForce RTX 2080 Ti] vendor: ZOTAC driver: nvidia
v: 435.21 bus ID: 01:00.0 chip ID: 10de:1e04
Display: x11 server: X.Org 1.20.4 driver: modesetting,nvidia
unloaded: fbdev,nouveau,vesa alternate: nv compositor: gnome-shell
resolution: 1920x1080~60Hz
OpenGL: renderer: GeForce RTX 2080 Ti/PCIe/SSE2 v: 4.6.0 NVIDIA 435.21
direct render: Yes
Audio:
Device-1: Intel Cannon Lake PCH cAVS vendor: Gigabyte
driver: snd_hda_intel v: kernel bus ID: 00:1f.3 chip ID: 8086:a348
Device-2: NVIDIA TU102 High Definition Audio vendor: ZOTAC
driver: snd_hda_intel v: kernel bus ID: 01:00.1 chip ID: 10de:10f7
Sound Server: ALSA v: k5.3.2

[marietto]

mariozio@DESKTOP-N9UN2H3:~$ systemctl status systemd-modules-load

systemd-modules-load.service - Load Kernel Modules
Loaded: loaded (/lib/systemd/system/systemd-modules-load.service; static; vendor preset: enabled)
Active: failed (Result: exit-code) since Wed 2019-10-09 21:27:27 CEST; 2h 42min ago
Docs: man:systemd-modules-load.service(8)
man:modules-load.d(5)
Process: 848 ExecStart=/lib/systemd/systemd-modules-load (code=exited, status=1/FAILURE)
Main PID: 848 (code=exited, status=1/FAILURE)

root@DESKTOP-N9UN2H3:/home/mariozio# journalctl _PID=848
-- Logs begin at Wed 2019-10-09 21:27:21 CEST, end at Thu 2019-10-10 00:14:29 CEST. --
ott 09 21:27:27 DESKTOP-N9UN2H3 systemd-modules-load[848]: Failed to find module '“vfio-pci”'
ott 09 21:27:27 DESKTOP-N9UN2H3 systemd-modules-load[848]: Failed to find module '“vfio”'
ott 09 21:27:27 DESKTOP-N9UN2H3 systemd-modules-load[848]: Failed to find module '“vfio_iommu_type1”'

[marietto]

I've removed the vfio modules. and now :

systemd-modules-load.service - Load Kernel Modules
Loaded: loaded (/lib/systemd/system/systemd-modules-load.service; static; vendor preset: enabled)
Active: active (exited) since Wed 2019-10-09 22:29:47 CEST; 2h 1min ago
Docs: man:systemd-modules-load.service(8)
man:modules-load.d(5)
Process: 381 ExecStart=/lib/systemd/systemd-modules-load (code=exited, status=0/SUCCESS)
Main PID: 381 (code=exited, status=0/SUCCESS)

ott 09 22:29:46 DESKTOP-N9UN2H3 systemd-modules-load[381]: Inserted module 'lp'
ott 09 22:29:46 DESKTOP-N9UN2H3 systemd-modules-load[381]: Inserted module 'ppdev'
ott 09 22:29:46 DESKTOP-N9UN2H3 systemd-modules-load[381]: Inserted module 'parport_pc'
ott 09 22:29:47 DESKTOP-N9UN2H3 systemd-modules-load[381]: Inserted module 'nvidia_drm'
ott 09 22:29:47 DESKTOP-N9UN2H3 systemd[1]: Started Load Kernel Modules.
Warning: Journal has been rotated since unit was started. Log output is incomplete or unavailable.

THE Monitor attached to the intel chipset is OFF.

[marietto]

I did a mistake inside the file /etc/modprobe.d/vfio.conf. I wrote :

options vfio_pci ids=10de:1b80,10de:10f0

instead of :

options vfio-pci ids=10de:1b80,10de:10f0

now it seems that VFIO works great :

root@DESKTOP-N9UN2H3:/home/mariozio# dmesg | grep -i vfio
[ 12.705059] VFIO - User Level meta-driver version: 0.3
[ 12.740758] vfio_pci: add [10de:1e04[ffffffff:ffffffff]] class 0x000000/00000000
[ 12.758429] vfio_pci: add [10de:10f7[ffffffff:ffffffff]] class 0x000000/00000000

BUT the monitor attached to the intel graphic cpu is turned off and debian does not recognized it.