On Tue, 2005-04-05 at 23:30 +0100, David Woodhouse wrote:
 I believe that there is already an implementation of such a hash
table
 floating around, because it was once suggested for the i_security field.
 Serge, do you have one? 
I do have still have a patch laying around from october, but I notice
that it was entirely protected by a spinlock.  If we in fact decide to
switch to using a hash table, it would seem worthwhile to switch to an
rcu based hash table, as there should be very few writes.
-serge
-- 
Serge Hallyn <serue(a)us.ibm.com>