> I have at least one concern about the wisdom of this entire approach.

>     o   Because this hooks into the system at the kernel descriptor
>         level, i.e., the outermost possible layer, it needs to
>         implement every little semantic someone might expect of
>         pipes, as well as do its own SMP locks, as it adds an entire
>         third I/O type to the previous choice of merely: vnode or
>         socket.

On the other hand, this could be looked on as good, because it forces
you to make that layer properly orthogonal.  It was having multiple
ports that forced NetBSD into doing the MI/MD split right; it was
having multiple devices that forced the development of proper device
driver interfaces; it was having multiple filesystems that forced
proper development of the vnode layer.  I say the right thing is to
either clean up this layer, which using it for pipes could help with,
or toss it and subsume everything into (say) vnodes.

Of course, this is orthogonal to other considerations.

					der Mouse

			       mouse@rodents.montreal.qc.ca
		     7D C8 61 52 5D E7 2D 39  4E F1 31 3E E8 B3 27 4B