audit-related slab memory leak in recent -mm kernels?

[Amy Griffis]

On Fri, Feb 24, 2006 at 03:11:44PM -0500, Valdis.Kletnieks at vt.edu wrote:
> On Fri, 24 Feb 2006 14:40:25 EST, Amy Griffis said:
> 
> > That kfree is misplaced.  It should be here.  Does this solve it on
> > your end?
> 
> Seems to, with a small tweak..
> 
> >  		if (context->names[i].name)
> >  			__putname(context->names[i].name);
> > +		kfree(context->names[i].ctx);
> 
> +		context->names[i].ctx = NULL;
> 
> > +	}
> 
> Can this thing legitimately get called twice for a given object?
> I needed the added line above to shut up a huge flood of 'double
> kfree'.  If this doesn't ring a bell, I'll go back and catch the
> dmesg so we can troubleshoot it.

Ah, yes.  Looks like audit_free_names is called on syscall exit and
task termination.  Your fix makes sense to me.

Amy