Is anyone interested in working on this?  It's beyond what I can do
without some difficulty, but I can probably get it working eventually.
;)

I'm going to write 4front and ask if there was any particular reason
they neglected to release their NetBSD-specific code.  Granted, they
released 4.0 and not 3.9, but I somehow doubt it is so different as to
render the NetBSD code useless.  Why port the FreeBSD OS layer, etc.
if we don't have to?

My main motives for doing this:

-- "Virtual" mixer device (and perhaps using this with code from the
"Ethernet Speaker System" which  I think would be especially useful in
certain embedded applications, see google for more information)
-- mmap() access to DMA space on more devices, e.g. azalia
-- Better OSS compatability (obviously)

One thing I'm especially interested in is how this should behave with
regard to the existing audio subsystem.  For example, what if someone
wanted to use an OSS-supported device on a platform which also has a
built-in device supported only by NetBSD?  Ideally, they'd be able to
use both devices with the OSS in-kernel mixer.

Regards,

--Blair