>> I've rewritten much of the tun driver, and among other things made
>> it capable of handling multiple address families.  Only a little of
>> this has made it back into the tree, probably largely because I did
>> a special device with a text interface rather than a slew of new
>> ioctl()s as the interface to frobbing options.

> Hmm, why not just add this slew of ioctl()s and get this stuff into
> the tree?

The usual reason: I have no !@*(&%#@# time.  Anybody wanna hire me to
work on NetBSD?  In Montreal?  Don't everybody speak up at once.

> Or just convince the others of the beauty of your text interface and
> get your version into the tree :-)

Yeah, I wish.  Core seems to have followed Berkeley in wandering even
farther from the "everything is a file" paradigm; there's little chance
I alone can drag them all back, even had I the energy to undertake the
task.  Look at sysctl().  Look at the resistance evidenced any time
anyone suggests actually using kernfs or procfs for anything.  Look at
this case. :-/

					der Mouse

			       mouse@rodents.montreal.qc.ca
		     7D C8 61 52 5D E7 2D 39  4E F1 31 3E E8 B3 27 4B