[NET]: Fix neighbour destructor handling.

->neigh_destructor() is killed (not used), replaced with
->neigh_cleanup(), which is called when neighbor entry goes to dead
state. At this point everything is still valid: neigh->dev,
neigh->parms etc.

The device should guarantee that dead neighbor entries (neigh->dead !=
0) do not get private part initialized, otherwise nobody will cleanup
it.

I think this is enough for ipoib which is the only user of this thing.
Initialization private part of neighbor entries happens in ipib
start_xmit routine, which is not reached when device is down.  But it
would be better to add explicit test for neigh->dead in any case.

Signed-off-by: David S. Miller <davem@davemloft.net>

net/atm/clip.c

@@ -261,14 +261,6 @@ static void clip_pop(struct atm_vcc *vcc, struct sk_buff *skb)
 	spin_unlock_irqrestore(&PRIV(dev)->xoff_lock, flags);
 }
 
-static void clip_neigh_destroy(struct neighbour *neigh)
-{
-	DPRINTK("clip_neigh_destroy (neigh %p)\n", neigh);
-	if (NEIGH2ENTRY(neigh)->vccs)
-		printk(KERN_CRIT "clip_neigh_destroy: vccs != NULL !!!\n");
-	NEIGH2ENTRY(neigh)->vccs = (void *) NEIGHBOR_DEAD;
-}
-
 static void clip_neigh_solicit(struct neighbour *neigh, struct sk_buff *skb)
 {
 	DPRINTK("clip_neigh_solicit (neigh %p, skb %p)\n", neigh, skb);
@@ -342,7 +334,6 @@ static struct neigh_table clip_tbl = {
 	/* parameters are copied from ARP ... */
 	.parms = {
 		.tbl 			= &clip_tbl,
-		.neigh_destructor	= clip_neigh_destroy,
 		.base_reachable_time 	= 30 * HZ,
 		.retrans_time 		= 1 * HZ,
 		.gc_staletime 		= 60 * HZ,