Linux 5.14 GPU Driver Updates Come In Heavy With ~300k New Lines Of Code

[NobodyXu]

This doesn't have anything to do with what you are talking about, you are just spurting random stuff here.



Not entirely true and stated in a misleading way. NVidia has a wrapper over the DKMS kernel driver which they do have to update somewhat frequently, so the ABI/API itself is not that stable (and its not even advertised as such)



Which again has nothing to do with the problems being talked about. This also is not a good thing because it means that bugs/fixes are blocked by a bureaucratic Linux process, which wouldn't be an issue if we had distributions running older stable Kernel versions.



Well no they don't, and thats the point. They are responsible for writing the driver, just as AMD/NVidia are on Windows. If there are bugs/issues, they work on it themselves and push the fixes/changes

That bureaucratic rules out many bugs and potential design problems.

Believe it or not, if you make a stable API/ABI of the linux and let’s every driver use that, distribution like debian/ubuntu/centos will reintroduce this bureaucratic process, but this time all their effort is separated and doesn share the same community anymore.

At that time where the community for kernel is divided, linux is going to slowly dies off as the community effort is no longer aligned.

That bureaucratic process is the proof that the linux community is working together to make linux better.

**It is not done to slow down innovation, but help rules out potential problems in them, and it is the biggest advantages of the open source ecosystem**

[mdedetrich]

Believe it or not, if you make a stable API/ABI of the linux and let’s every driver use that, distribution like debian/ubuntu/centos will reintroduce this bureaucratic process, but this time all their effort is separated and doesn share the same community anymore.

Thats actually better than what we have now, because right now we have to deal with both the bureaucracy of Linux AND the bureaucracy of the distrbution.

Also do note that the bureaucracy of the distribution can be completely bypassed by NVidia/AMD just providing the packages for the distro themselves (and they do this currently for a select few distros). If there was a stable graphics ABI in the kernel, NVidia/AMD can just provide their own package which users can install with apt-install or w/e.

Finally its also possible to distribute the driver using flatpak (which software is moving towards anyways to again bypass this complete mess on another level).

[NobodyXu]

Thats actually better than what we have now, because right now we have to deal with both the bureaucracy of Linux AND the bureaucracy of the distrbution.

Also do note that the bureaucracy of the distribution can be completely bypassed by NVidia/AMD just providing the packages for the distro themselves (and they do this currently for a select few distros). If there was a stable graphics ABI in the kernel, NVidia/AMD can just provide their own package which users can install with apt-install or w/e.

Finally its also possible to distribute the driver using flatpak (which software is moving towards anyways to again bypass this complete mess on another level).

I’d rather to keep this bureaucracy.

I want these linux kernel/distribution maintainers/reviewers to do the careful revision for me to make sure their code is good and rules out some of the bugs, and I also want them to keep these collaboration in check so that we know they are not playing some dirty tricks on their side for archiving false better performance in benchmark while no improvement in real life scenarios.

If you have payed attention to the news on here, you will know that not long ago, a research group in certain university try to submit deliberate buggy code as a patch, and if isn’t the contributors from linux who are keep it checked for us, god know wtf will they do.

What I trying to say is, do not trust these commercial companies more than the kernel contributors.

The later are actively maintaining a healthy environment for linux and open source while the former has no responsibility for that.

If you don’t let contributors from linux keep them in check, they sooner or later they are going to push more buggy code or code that are malicious.

Just because this approach is slower, doesn’t mean it’s worse.

Speed is nothing unless you can also guarantee its quality.