moving audit_free() up into do_exit()

[Alexander Viro]

On Wed, Mar 29, 2006 at 05:21:34PM -0500, Steve Grubb wrote:
> On Wednesday 29 March 2006 17:03, Alexander Viro wrote:
> > NOTE: it does change the moment when record is generated and if
> > something in userland depends on having it postponed until the time when
> > zombie gets reaped, we will have to change such userland code. ?I doubt
> > we have any such place, but that definitely needs an ACK from userland
> > side of things.
> 
> User land - e.g. auditd - does not have any dependency. It is a change that 
> people doing EAL documentation may want to be aware of.

OK, preliminary patches attached; the first one is minimal "take care of
deadlocks", hopefully suitable for 2.6.16.2.  The second is incremental
to that (actually, a combination of several smaller steps from git tree)
doing the audit_free() move and cleanups.  Warning: both are absolutely
untested.

Question: do we have agreed upon variants of "Inotify kernel API" and
"filesystem location based auditing" patches against the current mainline
(or FC, no difference in that respect) tree?