The exception is allowing the subsystem to break Rust code temporarily. If you accept a patch in C that breaks Rust code, and the Rust for Linux team doesn't fix it quickly enough, you either need to fix the Rust code yourself, remove it, or re-write it in C. All of this would take time and energy from all the non-R4L kernel devs.
This is why people are reluctant to accept too much mixing of the C and Rust codebases, because even the Rust for Linux team isn't promising to fix breakages in Rust for Linux code.
Have I gotten that right?
And then you're presenting this situation as "the Rust for Linux team isn't promising to fix breakages in Rust for Linux code". Somewhat disingenuous.
> The Rust maintainers have committed to handling all maintenance of Rust code, and handling all breakage of their code by changes on the C side. The only "burden" the C maintainers have to carry is to CC a couple of extra people on commits when APIs change.
But this is not actually true it seems. Even the Rust for Linux policy doesn't say this. But because of the incorrect statement that keeps getting repeated, people are calling Kernel devs unreasonable for being reluctant to Rust patches.
Well, firstly, "randos" aren't getting their patches easily accepted anyway.
And secondly, what's the problem with this? You want one of the following options:
1. Everyone who wants to submit a patch also be proficient in Rust,
Or
2. You want the reviewer to also be proficient in Rust
You don't think that's an unnecessary burden for the existing maintainers?
The burden should be on the people who want to introduce the second language.
> However, exceptionally, for Rust, a subsystem may allow to temporarily break Rust code. The intention is to facilitate friendly adoption of Rust in a subsystem without introducing a burden to existing maintainers who may be working on urgent fixes for the C side. The breakage should nevertheless be fixed as soon as possible, ideally before the breakage reaches Linus.