[AFS3-std] updated extended callback draft

[Chaskiel M Grundman]

>
> That's supposed to be handled by implementing a write barrier on lock
> release.  Tom had indicated this in conversation, too.
That already happens. releasing a write lock does write back the contents
before releasing the lock. how is client 1 supposed to implement the
corresponding read barrier if callback delivery is not synchronous? Do you
want it to do an extra fetchstatus after the lock succeeds?
>