Linux Developers May Discuss Allowing Rust Code Within The Kernel

[phoronix]

Phoronix: Linux Developers May Discuss Allowing Rust Code Within The Kernel

A Google engineer is looking to discuss at this year's Linux Plumbers Conference the possibility of allowing in-tree Rust language support...

http://www.phoronix.com/scan.php?page=news_item&px=Linux-Plumbers-2020-Rust

[bug77]

Niiice. I mean, small steps and everything, we wouldn't want to mess anything up, but adding Rust to the mix is a win imho.

[oleid]

There was recently a discussion on allowing other languages than C for drivers. Something like that could make sense maintenance wise.

[maarten]

I hope this will motivate gcc (and others) to start a rust frontend.

[rabcor]

Well I mean it's just a programming language, not a big deal, isn't it basically just a more verbose version of C++ anyways? I don't see the harm.

[kpedersen]

Well I mean it's just a programming language, not a big deal, isn't it basically just a more verbose version of C++ anyways? I don't see the harm.

Well C++ isn't in the kernel for a number of different reasons. Those same reasons are pretty much going to apply to Rust too.

[wswartzendruber]

Well C++ isn't in the kernel for a number of different reasons. Those same reasons are pretty much going to apply to Rust too.

It would be a lot easier to discuss this if you told us what those reasons are.

[shmerl]

Great idea.

[shmerl]

Well C++ isn't in the kernel for a number of different reasons. Those same reasons are pretty much going to apply to Rust too.

From what I gather, those reasons (like various implicit behavior) don't apply to Rust, which aims to be more explicit.

[jrch2k8]

Well, they can discuss it but the answer will be very skewed toward nope, i simply can't see Linus accepting LLVM as a dependency because honestly LLVM is a can of worms on its own and very far still from "stable" for something that low level.

Also this will require some very ugly hacks for debugging a kernel at runtime? , kexec will work if the runtime of rust had ABI changes? we would need to pull rust runtimes or deps for initcpio images? how rust runtimes will handle security injector(or readers dunno the right terminology) like grsecurity, hardening, lkrg, selinux? is a bug actually in rust or in the kernel? which one will need recompile? is rust actually safe(thread, allocations) at this level(remember the stuff like pthreads and the likes are several layers up the actual kernel code)?

Dunno, this seems like a damn lot of work for getting a lot more work without any benefit outside of "but it use the new language, so ..." and before any Rust evangelist comes crying thread/allocation safe(which is not, at least 100%) remember that is when dealing with user space, at kernel level that handling is infinitely more complex, for, rust this will be like go from typescript to deal with vectored ASM.