>  > For COMPAT_DARWIN, it seems I'll need a pseudo-device to be attached in
>  > the tree. Is it feasible?
> I can't see any reason for COMPAT_DARWIN to require this.

It finds devices by exploring the tree, and keyboard or mouse (don't
know yet) is handled though a pseudodevice tree called HIDIOSystem.

Maybe there is a way of handling it in a different way, but having the
pseudo-device attached to the tree would indeed be the simpliest one
from the compatibility layer point of view.

-- 
Emmanuel Dreyfus.
Avez vous entendu parler de Palladium?
http://www.lebars.org/sec/tcpa-faq.fr.html
manu@netbsd.org