[RFC][PATCH] (#7U4) file system auditing by location and name

Timothy R. Chavez tinytim at us.ibm.com
Mon May 9 14:57:35 UTC 2005


On Mon, 2005-05-09 at 10:45 -0400, Steve Grubb wrote:
> On Monday 09 May 2005 10:35, Timothy R. Chavez wrote:
> >Is that a fair analysis?
> 
> I guess. Is this good for you David?
> 
> >Admittedly, my approach was sloppier, and it appeared it wasn't 
> >working for you
> 
> Did it work for you?

Yes it did.  That's what weird to me... I briefly looked at the audit.31
kernel though and see the filesystem_init() function being called.  But,
regardless, I think I read somewhere people generally do not like those
types of "init" functions any way.

> 
> > Maybe we should be using unlikely() here?
> 
> Sure. The only other benefit that I could think of is that by deferring the 
> allocation, it only occurs IFF the filesystem auditing is used. Anyone doing 
> syscall only (or no auditing - just SE Linux avc denials) has a little bit of 
> memory saved.

Yeah, well the init function could have been macro'ed out.  I think its
water under the bridge now, unless there are in fact racey conditions.

> 
> -Steve
> 
> --
> Linux-audit mailing list
> Linux-audit at redhat.com
> http://www.redhat.com/mailman/listinfo/linux-audit
> 




More information about the Linux-audit mailing list