So what happens to users that depend on these archs & drivers - get left behind on the last non-Rust kernel?
edit: ok so it sounds a bit mean if you interpret it as if I'm not considering UnkleJoe to be a real human being. What I mean is, I don't think UnkleJoe is trying to get new devices to talk to his prehistoric PPC embedded system just for fun. An organisation is maintaining these systems, and that organisation is mooching off of Linux's excellent backwards compatibility. What's the economical cost of supporting that organisation, versus the economical costs of not implementing ksmbd in Rust?
Security updates for one. People are always complaining about insecure IoT devices/routers that get conscripted into botnets. Dropping kernel support entirely compounds to manufacturer firmware negligence - meaning end users can't update to the latest Lineage/DD-WRT/OpenWRT firmware, even if they want to.
> Most software runs fine on 10 year old linux kernels.
If you had to wrangle with world of custom Android ROMs or open source router firmware - even as a user and not a developer - you know that software does not work fine on old kernels. Trying to support both old and new kernels is a royal pain, especially if you have add shims to support old kernels that lack new features.
I thank $DIETY for Linus' pragmatic approach, I trust him to balance user support against developer zealotry. I wonder how long it'll take for Linux to go to shit after he retires. Kernel notes for v42.3: dropped support for Risc-V. Remaining supported archs are x64 and ARMv21-24
I thought we were talking about new drivers being developed in Rust. Your IoT security upgrade probably doesn't include a hardware upgrade, so it will not require these new drivers.
Regardless, it is the responsibility of the device manufacturer to support their old devices. It is their problem, hold them accountable.
The discussion moved to what happens to drivers that are not ported to Rust, and GP was arguing they should be abandoned/dropped - as opposed to requiring that all driver's be switched over to Rust first.
> It is their problem, hold them accountable
How can they be held accountable, pragmatically? Also, owning a CVE-riddled, insecure device is actually the user's problem. Open firmware can at least ameliorate the situation right now: where manufacturers cant be bothered to publish/upstream/update their drivers to the latest kernel API in the same language. Asking them to rewrite drivers in Rust and upstream is obviously a tall order - barring the passing of new legislation.
I don't think that's realistic.
If we buy the idea that Rust will allow kernel developers to write more robust, safer, more secure drivers, I think that should override a desire to allow things like driver portability to arches that hardware manufacturers never expected or intended. I would judge that support to be a "nice to have", but it shouldn't block what I consider to be more important concerns (security and robustness).
Regardless, the Rust support here is still considered experimental. I would be surprised if any company adopted it for a Linux driver for at least a year or more. That's plenty of time for people to get esoteric platforms supported in LLVM/Rust if they want to. And it also gives more time for the Rust GCC backend to become more mature and usable, and I believe that will give Rust support for every arch that GCC supports (more or less), which should cover your objection.
(At any rate, LLVM+Rust does currently support PPC, so your specific argument here isn't relevant.)