> On 7/14/19 2:45 AM, Lorenzo Bianconi wrote:
> > @@ -1124,7 +1125,9 @@ int __neigh_event_send(struct neighbour *neigh, 
> > struct sk_buff *skb)
> >  
> >                     atomic_set(&neigh->probes,
> >                                NEIGH_VAR(neigh->parms, UCAST_PROBES));
> > -                   neigh->nud_state     = NUD_INCOMPLETE;
> > +                   if (check_timer)
> > +                           neigh_del_timer(neigh);
> 
> Why not just always call neigh_del_timer and avoid the check_timer flag?
> Let the NUD_IN_TIMER flag handle whether anything needs to be done.

ack, I have been too paranoid here. I will post a v3 fixing it.

Regards,
Lorenzo

> 
> > +                   neigh->nud_state = NUD_INCOMPLETE;
> >                     neigh->updated = now;
> >                     next = now + max(NEIGH_VAR(neigh->parms, RETRANS_TIME),
> >                                      HZ/2);
> > @@ -1140,6 +1143,8 @@ int __neigh_event_send(struct neighbour *neigh, 
> > struct sk_buff *skb)
> >             }
> >     } else if (neigh->nud_state & NUD_STALE) {
> >             neigh_dbg(2, "neigh %p is delayed\n", neigh);
> > +           if (check_timer)
> > +                   neigh_del_timer(neigh);
> >             neigh->nud_state = NUD_DELAY;
> >             neigh->updated = jiffies;
> >             neigh_add_timer(neigh, jiffies +
> 

Attachment: signature.asc
Description: PGP signature

Reply via email to