[AFS3-std] Standardization of GetCapabilties RPCs forAFS3 client and services

[Buhrmaster, Gary]

> I'm not sure I like the implication that every new RPC should have an=20
> associated capability bit, either.  I see a potential here=20
> for a lot of bloat with relatively little benefit. =20

I would hope one could avoid that.  (new/changed) RPC's seem
likely to be rare, and introduced/supported at "version" boundries.
If it is possible to group them (say "byte range locking
capability V1", or "Vx.y extensions"), that would be better
(I am, perhaps, demonstrating my age, but VM/XA (and beyond)
used a DIAG 00 call to indicate capabilities of the base OS,
and one needed only to test the bit for the level of the OS
of when the feature was introduced to know whether you could
use it - of course IBM is big into backward compatibility,
so once introduced, the feature was (never?) removed, nor
was backward compatibility impacted (at least in theory)).

Gary