kre, art and is - thanks for your replies.

kre - I hadn't counted on the buffer space and what you say now make
perfect sense. The only worry then would be, I need to decrypt anything
read off HD, or, anything I previously encrypted higher in the layer.
But am I decrypting (ever) too much, are there other times I would enter
that layer except from those two situations.

art - look interesting and almost exactly what I have done, except
they'd progressed further (finished) to pass the key along with the
ioctl().

is - I have no idea what you are meaning, sorry.


The question is now, do I take the OBSD vnd.c sources and try to massage
them into NetBSD, or is there someone official who will do that anyway?
They seem to have done some other larger changes as well, calling
VOP_READ/VOP_WRITE, but it doesn't look too bad to convert. (Just
missing .actf and .active in struct buf.)

Or do I just continue my work? I want to extend it a bit further, so you
can pick the cipher to use (with ioctl/from vnconfig) of all the
in-kernel available ciphers (as well as a few testers). As well as, once
it is done, bring it to on disk drivers (ccd or raid).

Comments?

Lundy


-- 
Jorgen "Lord" Lundman <lundman@lundman.net>
Technology Manager, Unix Administrator
Phone: +44 (0)20-86591860  Mobile: +44 (0)79-58642918
Pager: 07958642918@one2one.net
"Rare is the person who can weigh the faults of others 
 without putting his thumb on the scales": Byron J. Langenfeld