Re: auth handshake and rendevouz objects

[Neal H. Walfield]

> If (1) takes long time, the task deserves losing. (2) should be
> nothing more than a (hash-)table lookup, and it should be possible to
> arrange the arguments sent back and forth such that the lookup is
> *only* a bounds check, an array indexing, and a comparison of a few
> integers. At most a dozen machine instructions.

Despite what you project, it still behooves us to delegate the work.
Before (1), we would set the no preempt-bit and lipc to a worker
thread.  Then, if it is as fast as you say, we win.  If not, we do a
fixup in the kernel and the main thread gets scheduled again and can
do another receive.

Of course, this will not work on SMP.  I do not know how the L4 people
plan to handle this case.

> And (3) can perhaps be
> combined into the same l4 rpc operation as the receive.

It can.

> where the data pages can be made pageable again as soon as there are
> no outstanding handle transfers.

I am not convinced that this added hair is necessary.