> > > not using kernel_object for kernel memory that will never be pageable
> > > is a fine idea.  but could we think about reducing the number of interfaces
> > > and use more flags to get all the variations?
> > 
> > i prefer to have separate interfaces because they have
> > different arguments.
> > eg. to allocate wired memory, "preferred offset" isn't needed.
> 
> the "preferred offset" parameter doesn't appear to be used by
> any of the callers, so we could just get rid of it entirely.

some code actually uses "prefer".  eg. mips vmapbuf.
i think they're rare enough to be omitted from the gereric interface, though.
they can just use raw uvm_map().

> even if it were useful, I like have a unified interface even if
> it means that in some calls one of the arguments doesn't mean anything.

ok.

i've put the current diff on:
	ftp://ftp.netbsd.org/pub/NetBSD/misc/yamt/km.diff.gz
(i386 only)

if no problems, i'll start to convert the other ports.

YAMAMOTO Takashi