On Tue, Aug 11, 2020 at 9:05 AM Al Viro [off-list ref] wrote:

Except that you suddenly see non-directory dentries get children.
And a lot of dcache-related logics needs to be changed if that
becomes possible.

Yeah, I think you'd basically need to associate a (dynamic)
mount-point to that path when you start doing O_ALT. Or something.

And it might not be reasonably implementable. I just think that as
_interface_ it's unambiguous and fairly clean, and if Miklos can
implement something like that, I think it would be maintainable.

No?

                Linus