[openib-general] Re: ipoib: outstanding patches

[Roland Dreier]

    Roland> Crazy idea: can we just get away with never clearing
    Roland> ops->destructor?  ipoib_neigh_destructor() checks if the
    Roland> neighbour structure has an IPoIB structure attached and
    Roland> does nothing if it doesn't.  So does it hurt to leave
    Roland> ops->destructor set to ipoib_neigh_destructor() forever?

ie just do the following...

(It seems to me that if this approach has a problem, we're leaking
IPoIB neighbour structures already...)

--- infiniband/ulp/ipoib/ipoib_main.c	(revision 4929)
+++ infiniband/ulp/ipoib/ipoib_main.c	(working copy)
@@ -247,7 +247,6 @@ static void path_free(struct net_device 
 		if (neigh->ah)
 			ipoib_put_ah(neigh->ah);
 		*to_ipoib_neigh(neigh->neighbour) = NULL;
-		neigh->neighbour->ops->destructor = NULL;
 		kfree(neigh);
 	}
 
@@ -530,7 +529,6 @@ static void neigh_add_path(struct sk_buf
 err:
 	*to_ipoib_neigh(skb->dst->neighbour) = NULL;
 	list_del(&neigh->list);
-	neigh->neighbour->ops->destructor = NULL;
 	kfree(neigh);
 
 	++priv->stats.tx_dropped;
--- infiniband/ulp/ipoib/ipoib_multicast.c	(revision 4936)
+++ infiniband/ulp/ipoib/ipoib_multicast.c	(working copy)
@@ -114,7 +114,6 @@ static void ipoib_mcast_free(struct ipoi
 		if (neigh->ah)
 			ipoib_put_ah(neigh->ah);
 		*to_ipoib_neigh(neigh->neighbour) = NULL;
-		neigh->neighbour->ops->destructor = NULL;
 		kfree(neigh);
 	}