doc arbiters

[Kevin Ryde]

A bit of a tweak to the arbiters node.  The old words suggested that
only the thread which locked an arbiter could release it, I propose to
relax that to reflect the way the code works, which is that anyone can
release.


5.17.1 Arbiters
---------------

Arbiters are synchronization objects, they can be used by threads to
control access to a shared resource.  An arbiter can be locked to
indicate a resource is in use, and unlocked when done.

   An arbiter is like a light-weight mutex (*note Low level thread
primitives::).  It uses a little less memory, but there's no means for
a thread to block waiting on an arbiter, it can only test and get the
status returned.

 -- Scheme Procedure: make-arbiter name
 -- C Function: scm_make_arbiter (name)
     Return an arbiter object, initially unlocked.  Currently NAME is
     only used for diagnostic output.

 -- Scheme Procedure: try-arbiter arb
 -- C Function: scm_try_arbiter (arb)
     If ARB is unlocked, then lock it and return `#t'.  If ARB is
     already locked, then do nothing and return `#f'.

 -- Scheme Procedure: release-arbiter arb
 -- C Function: scm_release_arbiter (arb)
     If ARB is locked, then unlock it and return `#t'.  If ARB is
     already unlocked, then do nothing and return `#f'.

     Typical usage is for the thread which locked an arbiter to later
     release it, but that's not required, any thread can release it.