Hello, the last 3 versions of Nvidia drivers are incompatible with Debian 11 and Kernel 5.19.x on my workstation !?
515.x series didn’t work with DKMS, latest driver 5.20.56.06 has problem with LIGHTDM manager ;-(
Could you please fix these bugs?? I haven’t been able to use any recent kernels for about 2 months now and have to settle for a 5.10 kernel which is only partially compatible with my hardware.
Thank you in advance because we pay high prices for graphics cards and we can’t use them ;-(
You should finally publish free drivers integrated into the kernel because by dint of not being able to use our hardware correctly we will end up turning to AMD…
For laptops with Intel and Nvidia cards, you can easily use the Intel card but it’s almost impossible to use the Nvidia card except on Ubuntu or Manjaro maybe, which are 2 OSes that I don’t like. Here too we have the feeling of not making our investment profitable because you are not doing anything for us ;-(
Please make an effort for your customers and stop sticking to ideas that don’t suit anyone except you, for a obscure reason.
I just noticed that when you deactivate and then reactivate the “Secure Boot” function in the bios of your machine (Asus ROG Crosshair VIII Dark Hero in my case), the Nvidia driver NVIDIA-Linux-x86_64-520.56.06 .run which initially worked perfectly is suddenly no longer able to boot Debian 11 with Kernel 6.0.8-1 (from Debian experimental section). It then displays an error message that it could not launch LIGHTDM manager ;-( If you deactivate “Secure Boot” again, the machine boots normally…
Apart from that the latest driver NVIDIA-Linux-x86_64-515.86.01 which just came out yesterday continues to give me problems with the 6.0.8 kernel whether secure boot is activated or not… My system cannot launch the xwindow manager lightdm ;-(
I experienced the same issue with Ubuntu 20.04 LTS. The GPUs were unable to load after upgrading to 22.04 and installing the most recent CUDA drivers and toolkit. It did not function even after downgrading to Ubuntu 20.04 LTS. After several hours of research, I discovered that your kernel module, clang, and any other packages necessary to install CUDA must match those listed in the CUDA Installation Documentation. The most current version of UBUNTU has a new Kernel Header, which may break some CUDA drivers or toolkits. So, before installing CUDA Drivers on Linux/Ubuntu, verify the CUDA Toolkit/Driver requirements. You can keep separate kernel modules with mainline to keep specific Linux Kernel Headers.
mainline --install 5.8.0
something like these to install kernel header with 5.8.0 version. You find more on google how to maintain different Kernel Headers in Linux.