> 
> Some things are already available. For instance, we have zero-copy TCP 
> write using the vm system. And the vm system method is set up so that 
> writing to the data while it's piped-out for transmit will do the right 
> thing (trigger a COW).
>
That's what I assumed.  Is TCP receive also 0-copy?
 
> 
> My main suggestion is to step back and talk about what you really want to 
> do. There may well be easier ways to hook everything together to get there 
> both more quickly and in a more-reusable manner.
> 
Hmm, what I'm actually doing would take a pretty long time to explain, and most
of it is doable in the user space.  Yes, I agree, what I've described reminds a
lot of aio.  What I want to do is truly 0-copy disk/network aio to/from a wired
shared memory region.  

Thanks and best regards,
  Zeljko.