rt_finalize WTFs?

[dsimcha]

== Quote from Martin Nowak (dawg at dawgfoto.de)'s article
> > More promising is to put pool addresses ranges in a trie.
> >
> > addr[7]          [...      .     ...]
> >                       /     |    \
> > addr[6]     [...   .   ...]    [...   .   ...]
> >                 /   |    \         /   |   \
> > addr[5]     pool:8             [...   .  ...]
> >                                /   |   \
> > addr[4]                  pool:8 [....] pool:5
> >
> Actually 64-bit should use a hashtable for the upper 32-bit and then
> the the 32-bit trie for lower.

Why do you expect this to be faster than a binary search?  I'm not saying it won't
be, just that it's not a home run that deserves a high priority as an
optimization.  You still have a whole bunch of indirections, probably more than
you would ever have for binary search.