watch dir problem

David Woodhouse dwmw2 at infradead.org
Thu May 26 21:20:15 UTC 2005


On Thu, 2005-05-26 at 15:18 -0500, Timothy R. Chavez wrote:
> Are you using a .49 or .50 kernel?  I wasn't quite sure if David said "OK" to just
> using the hash table from now on or to use the i_audit field until we're in the
> clear.

Yes, let's just use the hash table from now on.

> David has incorporated most of the cleanups I had and then he made some 
> more which I've also added to my tree.. and the last major thing I can think of
> is to correctly handle audit_notify_watch() in fs/namei.c

OK, I'll leave that one to you while I play with trying to allocate
inode audit data for only those inodes which need it -- which will make
the decision to use a hash table sane for upstream too.

If we're going to be poking at the same piece of code we should probably
have a communication channel which has lower latency than email. I've
created a #audit channel on irc.freenode.net.

Not that you're going to get _much_ more out of me tonight, mind you :)
> 
-- 
dwmw2





More information about the Linux-audit mailing list