Really need some help with RTEMS networking semaphores

[gregory.menke at gsfc.nasa.gov]

As a followup for the file, I've continued experimenting with the mbuf &
mbuf cluster sizing.  If the mbuf cluster buffer is set to 16 * the mbuf
buffer, then as per rtems_bsdnet_show_mbuf_stats(), it looks like there
is slightly more than 1 cluster per mbuf.  The size of the mbuf buffer
is then set for whatever total footprint is desired, >= 100k seems to
provide fairly reliable behavior but the more the better.

I don't yet know to what extent mbufs are allocated without clusters,
but since at least the receive packets use mbufs with clusters, sizing
the cluster buffer so each mbuf can get one seems at least arguably
useful.

Gregm