Linux-for-Rust or Rust-for-Linux

Sorry, the example of using it there wasn't really my point.

But like you said, using the wifi example again, if someone were to add rust bindings to mac80211 (which in your view would be a good thing to have), the point is that I'm still going to have to maintain them once they're in the tree, but I get nothing out of it because I can't even use them. So I have to maintain something I can't even _use_ myself.

And I think that's similar for the filesystems: as long as you can't actually rely on rust in general, most stuff will have to be maintained in C.

And I think for me at least, the point is that for those of us working on sufficiently core things rust therefore just adds work, no value. And I can't even really learn rust in the context of the code I'm maintaining, because I can't actually use it there myself! I'd totally be at the mercy of whoever adds the rust bindings, and then try to figure out what that does, but I don't think there'd be much of a chance if I'm not going to use it.

Do I think it's possible to learn rust? Sure. But I don't think it'll happen overnight (nor within a couple of weeks), but more importantly I don't think I'm going to learn rust and how to apply it to mac80211 without even being able to use it there.

(Given enough time I could of course write a toy driver for it or something, but in reality I don't see where to find the time for that. Most maintainers are already overworked. There's probably some promise in rust actually making that better, but that only applies when you can actually use rust for the work you need to do, which is nowhere near true yet for most of us.)