In brief

> What about
> creating a new fs that works like debugfs but is for stable interfaces
> and can be mounted at root?
>
> /kernel?

Ugh. Now I have to look for kernel tuning knob is in /sys, /proc, or, now proposed, /kernel?

Just because /sys has an annoying *implementation* in the kernel doesn't mean we need a
whole new namespace for users.

It's way too late to undo the /proc vs /sys mistake. So: I have an idea. How about just putting
all the new stuff into sysfs. Fix the kernel-side API if it's too hard to use, don't make users
suffer. *mutter*