> I said weird, because the whole notion of "remove a device, but have
> the kernel pretend some aspects of the device are still there," is
> just plain weird.  If a device is gone, it should be gone completely
> from the kernel. 

This is an interesting assertion to make; everything else you say
follows from it, but I'm not convinced that this should always be true
for every single subsystem. 

> I thought that much work had already been done towards that goal.
> It seems wasteful and ... less than optimal to throw that away and
> move forward on a plan that seems, at least to me, to provide
> fundamentally incorrect operation.

I'd love to be proven wrong, but as best as I can tell, the people who
did the earlier work have dropped the ball and nobody other than me
seems willing to pick it up.  I took a stab at merging their previous
work up to -current and the number and nature of the things touched by
the branch was such that it felt like the approach they took was not
entirely correct.

I'm exploring another alternative here.. I think regardless of what we
do here, it's going to be messy.

> Disclaimer: Not speaking for NetBSD, just expressing my own opinion.

ditto..

					- Bill