And now we have FUSE. The good ideas do get taken up by the mainstream.

I feel like there's a difference between FUSE, an anomalous way to implement a filesystem, and having the user-space method be the primary mechanism to implement a filesystem. The latter ensures that the user-space thing doesn't have a quality gap with "real" FS drivers.

▲

marcosdumay 4 days ago | parent | prev [-]

We have entire userspace network protocols, ePBF, and to some extent even ePool pooling ideas from microkernels. But A single disgruntled kernel dev is enough to stop Rust device drivers from existing, so no, the idea is still not here.

	▲	steveklabnik 4 days ago \| parent \| next [-]
		That didn’t happen.
	▲	josefx 4 days ago \| parent \| prev [-]
		Even the Asahi Linux lead threatening Linus with a witchhunt against all kernel maintainers did not manage to finish off the ongoing Rust integration. People may not like it but it isn't going down easily.