nbd: Rewrite thread passing to use semaphore rather than pipe

nbd: Rewrite thread passing to use semaphore rather than pipe

I ran some local testing on my Fedora 29 system via:
$ ./nbdkit memory 1m
$ for i in `seq 10`; do
    ./nbdkit -U - --filter=stats nbd \
      statsappend=true statsfile=$file hostname=localhost port=10809 \
      --run '~/libnbd/examples/threaded-reads-and-writes $unixsocket'
  done

Pre-patch, the runs averaged  1.266s, 1.30E+08 bits/s
Post-patch, the runs averaged 1.154s, 1.42E+08 bits/s

This is roughly 9% performance gain, all because semaphores are
lighter weight than pipes for inter-thread locking.  The refactoring
will also make it a bit easier to incorporate the use of libnbd.

Signed-off-by: Eric Blake <eblake@redhat.com>