> > complete answer.  It is, at least, more correct than just "stick 0xF1000000
> > in a pointer variable and off you go".  (Off into the weeds, more like it.)
> You're right that I left all that out (mainly as I don't know how to do
> it). But I DID mention that I assumed that the memory block was already
> mapped, and that it'd be better to base things off of the address returned
> by the mapping and not an absolute address.

Ironically enough, Theo D. sent me mail saying the questioner evidently isn't
using NetBSD, but is just spamming several technical lists hoping to come
up with the right answer.  If he really *is* doing direct embedded programming,
then my answer was exactly wrong, since then he *would* be talking to the bare
metal and *should* just "stick 0xF1000000 in a pointer variable"...  (Well,
assuming his C compiler implements the straightforward method of casting
integers into pointers, but I'm way out of the Language Lawyer habit, so
I won't go into that...)

(Actually, if it IS a 68K embedded controller, the shared memory MIGHT be
addressed via one of the alternate "spaces" available, but it would take a
REALLY, REALLY sadistic hardware designer to do that.  Of course, as anyone
who has done software for low level hardware knows all to well, *ALL* hardware
designers are REALLY, REALLY sadistic...  "but it saved a gate!")