In message <20040326202941.GB29369@bcd.geek.com.au>,
Daniel Carosone writes:

>This would be great, there's a perfect case for using something like
>this in cgd(4).

I'd say instead that cgd(4) should be using OCF (opencrypto(9)), so as to
take advantage of a hardware accelerator if one is present.

Strictly speaking, that's orthogonal to how OCF manages its continuations internally.