On 14/03/12, Steve Grubb wrote:
 On Wednesday, March 12, 2014 09:18:14 AM Eric Paris wrote:
 > On Wed, 2014-03-12 at 08:55 -0400, Steve Grubb wrote:
 > > On Wednesday, February 19, 2014 01:08:22 PM Richard Guy Briggs wrote:
 > > > Add a netlink multicast socket with one group to kaudit for
 > > > "best-effort"
 > > > delivery to read-only userspace clients such as systemd, in addition to
 > > > the
 > > > existing bidirectional unicast auditd userspace client.
 > > 
 > > One question...we do have to have the ability to separate of secadm_r and
 > > sysadm_r. By allowing this we will leak to a sysadmin that he is being
 > > audited by the security officer. In a lot of cases, they are one in the
 > > same person. But for others, they are not. I have a feeling this will
 > > cause problems for MLS systems. 
At first I had no idea what you were talking about but Eric's reply
helps to understand the context.
 > A good question.  But easily solved in policy.  Don't give
 > CAP_AUDIT_READ to sysadm_t if you don't want sysadm_t to be able to read
 > from the multicast socket. 
This seems like an easy one.
 That also means that we probably want an audit event for any
successful and 
 unsuccessful attempts to connect for _reading_ audit events. 
That could easily be added to the new custom netlink bind function.
 -Steve
 
 > As to what others who read from the journal I guess we can just make
 > sure it is a config option whether to collect or not.  Most everyone
 > would want to collect, but some configs might obviously not. 
This would be easy to add as a "feature", I'm guessing...
 > I'll roll around in the back of my head the ability for
auditctl to
 > disable the multicasting, but CAP_AUDIT_READ takes care of that a whole
 > lot more nicely... 
- RGB
--
Richard Guy Briggs <rbriggs(a)redhat.com>
Senior Software Engineer, Kernel Security, AMER ENG Base Operating Systems, Red Hat
Remote, Ottawa, Canada
Voice: +1.647.777.2635, Internal: (81) 32635, Alt: +1.613.693.0684x3545