Re: [9fans] why there is no srv->clunk in lib9p?

["Russ Cox" <rsc@swtch.com>]

> why there is no srv->clunk in lib9p?

>From 9p(2):

          Destroyfid
               When a Fid's reference count drops to zero (i.e., it
               has been clunked and there are no outstanding requests
               referring to it), destroyfid is called to allow the
               program to dispose of the fid->aux pointer.

There is no srv->clunk because srv->destroyfid
takes its place.  Srv->destroyfid is called once the
fid's reference count drops to zero, which usually
happens as a result of a Tclunk message, but not always.

In regular a 9P conversation, Tclunk will drop the
last reference to a fid, but you can come up with
various cases where this is not true.  For example,
consider a fid where Tread takes a long time (perhaps
a pipe or /dev/cons or network connection).  What
if a client sent Tread and then without waiting for
the Rread, sent Tclunk.  If there were a clunk handler,
then presumably it would free whatever data structures
were associated with the fid.  But that memory is still
being used by the Tread.  Until all the requests that
reference the fid are done, it is not safe to free the
fid.

Russ