Re:Next steps for Rust in the kernel
by menaerus on Thursday September 22, 2022 @08:28AM
(#32937919)
> That doesn't excuse downplaying the security issues so common when using languages that do not offer memory safety.
Sure thing but can Rust live up to its memory-safety promise? Or even data-race freedom promise?
I must say that, coming from systems programming background, I was always skeptical about those promises because IMO these problems fundamentally cannot be solved in bare-metal programming language. They're implied by the complexity of HW not loop-holes in programming languages.
That said I see nothing wrong with Rust but I think that people are vastly overselling its core features while at the same time a quick glance over https://cve.mitre.org/cgi-bin/cvekey.cgi?keyword=rust will give you exactly a list of bugs which Rust seems to promise not to run into.
Sure thing but can Rust live up to its memory-safety promise? Or even data-race freedom promise?
I must say that, coming from systems programming background, I was always skeptical about those promises because IMO these problems fundamentally cannot be solved in bare-metal programming language. They're implied by the complexity of HW not loop-holes in programming languages.
That said I see nothing wrong with Rust but I think that people are vastly overselling its core features while at the same time a quick glance over https://cve.mitre.org/cgi-bin/cvekey.cgi?keyword=rust will give you exactly a list of bugs which Rust seems to promise not to run into.
Another very recent example, and a very good blog post about memory corruption taking place in Rust: https://www.graplsecurity.com/post/attacking-firecracker