this post was submitted on 29 Aug 2024
510 points (98.7% liked)

Linux

48323 readers
632 users here now

From Wikipedia, the free encyclopedia

Linux is a family of open source Unix-like operating systems based on the Linux kernel, an operating system kernel first released on September 17, 1991 by Linus Torvalds. Linux is typically packaged in a Linux distribution (or distro for short).

Distributions include the Linux kernel and supporting system software and libraries, many of which are provided by the GNU Project. Many Linux distributions use the word "Linux" in their name, but the Free Software Foundation uses the name GNU/Linux to emphasize the importance of GNU software, causing some controversy.

Rules

Related Communities

Community icon by Alpár-Etele Méder, licensed under CC BY 3.0

founded 5 years ago
MODERATORS
 

Wedson Almeida Filho is a Microsoft engineer who has been prolific in his contributions to the Rust for the Linux kernel code over the past several years. Wedson has worked on many Rust Linux kernel features and even did a experimental EXT2 file-system driver port to Rust. But he's had enough and is now stepping away from the Rust for Linux efforts.

From Wedon's post on the kernel mailing list:

I am retiring from the project. After almost 4 years, I find myself lacking the energy and enthusiasm I once had to respond to some of the nontechnical nonsense, so it's best to leave it up to those who still have it in them.

...

I truly believe the future of kernels is with memory-safe languages. I am no visionary but if Linux doesn't internalize this, I'm afraid some other kernel will do to it what it did to Unix.

Lastly, I'll leave a small, 3min 30s, sample for context here: https://youtu.be/WiPp9YEBV0Q?t=1529 -- and to reiterate, no one is trying force anyone else to learn Rust nor prevent refactorings of C code."

you are viewing a single comment's thread
view the rest of the comments
[–] fartsparkles@sh.itjust.works 43 points 2 months ago* (last edited 2 months ago) (15 children)

No idea what you’re being downvoted. Just take a look at all the critical CVSS scored vulnerabilities in the Linux kernel over the past decade. They’re all overwhelmingly due to pitfalls of the C language - they’re rarely architectural issues but instead because some extra fluff wasn’t added to double check the size of an int or a struct etc resulting in memory corruption. Use after frees, out of bounds reads, etc.

These are pretty much wiped out entirely by Rust and caught at compile time (or at runtime with a panic).

The cognitive load of writing safe C, and the volume of extra code it requires, is the problem of C.

You can write safe C, if you know what you’re doing (but as shown by the volume of vulns, even the world’s best C programmers still make slip ups).

Rust forces safe(r) code without any of the cognitive load of C and without having to go out of your way to learn it and religiously implement it.

[–] qqq@lemmy.world 21 points 2 months ago* (last edited 2 months ago) (2 children)

They're being downvoted because it's a silly comment that is basically unrelated and also extremely unhelpful. Everyone can agree that C has footguns and isn't memory safe, but writing a kernel isn't memory safe. A kernel written in Rust will have tons of unsafe, just look at Redox: https://github.com/search?q=repo%3Aredox-os%2Fkernel%20unsafe&type=code That doesn't mean it isn't safer, even in kernel space, but the issues with introducing Rust into the kernel, which is already written in C and a massive project, are more nuanced than "C bad". The religious "C bad" and "C good" arguments are kinda exactly the issue on display in the OP.

I say this as someone who writes mostly Rust instead of C and is in favor of Rust in the kernel.

[–] witx@lemmy.sdf.org 8 points 2 months ago (1 children)

The difference is that now you have a scope of where the memory unsafe code might be(unsafe keyword) and you look there instead of all the C code.

[–] qqq@lemmy.world 2 points 2 months ago* (last edited 2 months ago)

I agree and think that should be helpful, but I hesitate to say how much easier that actually makes writing sound unsafe code. I'd think most experienced C developers also implicitly know when they're doing unsafe things, with or without an unsafe block in the language -- although I think the explicit unsafe should likely help code reviewers and tired developers.

It is possible to write highly unsafe code in Rust while each individual unsafe block appears sound. As a simple example: https://play.rust-lang.org/?version=stable&mode=debug&edition=2021&gist=6a1428d9cae5b9343b464709573648b4 [1] Run that on Debug and Release builds. Notice the output is different? Don't take that example as some sort of difficult case, you wouldn't write this code, but the concepts in it are a bit worrisome. That code is a silly example, but each individual unsafe block appears sound when trying to reason only within the block. There is unsafe behavior happening outside of the unsafe blocks (the do_some_things function should raise eyebrows), and the function we ultimately end up in has no idea something unsafe has happened.

Unsafe code in Rust is not easy, and to some extent it breaks abstractions (maybe pointers in general break abstractions to some extent?). noaliases in that playground code rightly assumes you can't have a &ref and &mut ref to the same thing, that's undefined behavior in Rust. Yet to understand the cause of that bug you have to look at all function calls on the way, just as you would have to in C, and one of the biggest issues in the code exists outside of an unsafe block.

[1]: If you don't want to click that link or it breaks, here is the code:

fn uhoh() {
    let val = 9;
    let val_ptr: *const usize = &val;
    do_some_things(val_ptr);
    println!("{}", val);
}

fn do_some_things(val: *const usize) {
    let valref = unsafe { val.as_ref().unwrap() };
    let mut_ptr: *mut usize = val as *mut usize;
    do_some_other_things(mut_ptr, valref);
}

fn do_some_other_things(val: *mut usize, normalref: &usize) {
    let mutref = unsafe { val.as_mut().unwrap() };
    noaliases(normalref, mutref);
}

fn noaliases(input: &usize, output: &mut usize) {
    if *input < 10 {
        *output = 15;
    }
    if *input > 10 {
        *output = 5;
    }
}

fn main() {
    uhoh();
}
load more comments (12 replies)