Rust Infrastructure Pull Request Submitted For Linux 6.1!

[phoronix]

Phoronix: Rust Infrastructure Pull Request Submitted For Linux 6.1!

It's happening, folks! Linus Torvalds already indicated recently he intends to pull the initial Rust programming language support into the Linux 6.1 kernel cycle and today that pull request was submitted to him. Linux 6.0 isn't out yet but should be on Sunday unless any last minute problems, which in turn will mark the start of the two week v6.1 merge window...

https://www.phoronix.com/news/Rust-For-Linux-6.1-Pull-Request

[timofonic]

Now GCC Rust effort needs to be taken more seriously, depending only in LLVM is a no go.

[oleid]

Now GCC Rust effort needs to be taken more seriously

More seriously compared to what? I mean, there are even two independent efforts. One independent implementation of rustc which will live in GCC tree (gccrs) and one as a backend to the official rust compiler.

depending only in LLVM is a no go.

There are currently three backends in rustc available: llvm, cranelift and gcc (via libgccjit, which also compiles ahead of time despite the name). The GCC backend is not far away from having no errors in rustc's the test suite. So it should be usable soon(ish).

[Azrael5]

What's the benefit to integrate Rust in the kernel?

[stormcrow]

What's the benefit to integrate Rust in the kernel?

It takes away the memory safety foot guns that C developers are so fond of using to blow their own feet off.

[Sin2x]

Yeah, the death of Linux is happening, hooray! *sarcasm*

[Jakobson]

Rustaceans like very much. But please give me some performance figures. If you don't have ay then I do not celebrate either.

[dragonn]

Rustaceans like very much. But please give me some performance figures. If you don't have ay then I do not celebrate either.

https://www.phoronix.com/news/LPC-2022-Rust-Linux here you go

[dlq84]

Rustaceans like very much. But please give me some performance figures. If you don't have ay then I do not celebrate either.

There is an experimental NVMe driver that has pretty much the same performance as the mainline one.

Read this: https://doc.rust-lang.org/beta/embed...tractions.html

[Mitch]

https://www.phoronix.com/news/LPC-2022-Rust-Linux here you go

I remember these tests. Someone correct me if I'm wrong, but I think it has been said and shown, ad nausium, that rust pays a higher overhead only during compile, not runtime. And whenever there's a runtime difference, it's within a margin of error. The compile-time overhead seems to help developers save time and focus their attention far less on memory and thread management so they can be more ambitious on things like performance and parallelism. I think Asahi Lina was talking about the pros and cons of Rust, but her overall impression while working on some M1 GPU stuff was very positive for Rust and her only criticism is that it could do with one or two additional features to close a parity gap (presumably with C).

I'm not a dev for this kind of programming, and I have no skin in this game. Just trying to recall what I remember.