[PATCH] Audit: save audit_backlog_limit audit messages in case auditd comes back

[Eric Paris]

On Thu, 2008-03-27 at 17:50 -0400, Steve Grubb wrote:
> On Thursday 27 March 2008 17:37:44 Eric Paris wrote:
> > This is useful to collect audit messages during bootup and even when auditd
> > is stopped.  This is NOT a reliable mechanism, it does not ever call
> > audit_panic, nor should it. 
> 
> Thanks Eric for working on this. We've needed this for quite a while so that 
> we can see some of the avcs that happen during boot.
> 
> 
> > If auditd never starts the kernel will hold by default up to 64 messages
> > in memory forever.
> 
> I have an idea. Maybe this behavior could be enabled if audit=1 is passed as a 
> boot parameter. In this way, you would know that the user intended for the 
> audit daemon to start at some point. You could then call audit panic or 
> whatever else is normal. If no audit=1 is passed, you could just do the 
> printk like usual and not waste memory. Would this be helpful?

I could probably do that.  I also could conditionalize it on auditd ever
having run.  I can't imagine it is normal for auditd to be running and
then stopped forever....

Anyone else see value in that situation?  Only do it on boot if audit=1
is passed?  Does anyone actually use that command line option?

-Eric