[RFC][PATCH] (#2) Prelim in-kernel file system auditing support

Serge Hallyn serue at us.ibm.com
Tue Jan 25 20:31:05 UTC 2005


I think that
On Tue, 2005-01-25 at 12:38 -0600, Timothy R. Chavez wrote:

> So let me better protect the list.  Should I simply include in the
> critical section a test to make sure that I've not already been added?

That's probably best.  We don't expect the list for one parent to get
very long (so checking for previous addition should be quick), and this
way we minimize the amount of time we're locked and excluding readers.
Otherwise we might end up sleeping on a kmalloc while holding the lock.

At least that's my take.  Does anyone feel the writelock should just be
held for all of audit_create_watch()?

-serge
-- 
Serge Hallyn <serue at us.ibm.com>




More information about the Linux-audit mailing list