Nvidia Linux drivers causing random hard crashes and now a major security risk still not fixed after 5+ months
The recent fiasco with Nvidia trying to block Hardware Unboxed from future GPU review samples for the content of their review is one example of how they choose to play this game. This hatred is not only shared by reviewers, but also developers and especially Linux users.
The infamous Torvalds videos still traverse the web today as Nvidia conjures up another evil plan to suck up more of your money and market share. This is not just one off shoot case; oh how much I wish it was. I just want my computer to work.
If anyone has used Sway-WM with an Nvidia GPU I’m sure they would remember the –my-next-gpu-wont-be-nvidia
option.
These are a few examples of many.
The Nvidia Linux drivers have never been good but whatever has been happening at Nvidia for the past decade has to stop today. The topic in question today is this bug: [https://forums.developer.nvidia.com/t/bug-report-455-23-04-kernel-panic-due-to-null-pointer-dereference]
This bug causes hard irrecoverable crashes from driver 440+. This issue is still happening 5+ months later with no end in sight. At first users could work around this by using an older DKMS driver along with a LTS kernel. However today this is no longer possible. Many distributions of Linux are now dropping the old kernels. DKMS cannot build. The users are now FORCED with this “choice”:
{Use an older driver and risk security implications} or {“use” the new drivers that cause random irrecoverable crashes.}
This issue is only going to get more and more prevalent as the kernel is a core dependency by definition. This is just another example of the implications of an unsafe older kernel causing issue for users: https://archlinux.org/news/moving-to-zstandard-images-by-default-on-mkinitcpio/
If you use Linux or care about the implications of a GPU monopoly, consider AMD. Nvidia is already rearing its ugly head and AMD is actually putting up a fight this year.