Re: [RFC PATCH] ipv6: Split from and expires field in dst_entry out of union [net-next]

[Eric Dumazet <eric.dumazet@gmail.com>]

On Tue, 2013-02-19 at 16:49 -0500, Neil Horman wrote:
> On Tue, Feb 19, 2013 at 01:17:45PM -0800, Eric Dumazet wrote:
> > On Tue, 2013-02-19 at 15:28 -0500, Neil Horman wrote:
> > 
> > >  static inline void rt6_update_expires(struct rt6_info *rt, int timeout)
> > >  {
> > >  	if (!(rt->rt6i_flags & RTF_EXPIRES)) {
> > > -		if (rt->dst.from)
> > > -			dst_release(rt->dst.from);
> > > +		dst_release(rt->dst.from);
> > >  		/* dst_set_expires relies on expires == 0 
> > >  		 * if it has not been set previously.
> > >  		 */
> > >  		rt->dst.expires = 0;
> > > +		rt6->dst.from = NULL;
> > >  	}
> > >  
> > 
> > Sorry you didnt really address the problem, only reduce the race window.
> > 
> I kinda had a feeling you would say that, but the only other solution I see here
> is to either introduce some locking to protect the from pointer, or two revert
> the patch that introduced the from pointer alltogether, neither of which sounds
> appealing to me.  I suppose we could use an xchng to atomically update the from
> pointer, so there was only ever one context that was able to free it in
> rt6_update_path.  Does that seem reasonable to you?

I believe the setting of rt6->dst.from is safe :
It should be done at :
- dst creation time, when we are the only user.
- dst destry time, when we are the only user.

We only have to do the dst_release() at the right place, when we are the
last user of the dst.

So rt6_update_expires() should not mess with rt6->dst.from at all.