[RESEND] [PATCH] neigh: error out pending skbs if netlink invalidates incomplete neigh

[RESEND] [PATCH] neigh: error out pending skbs if netlink invalidates incomplete neigh

[Timo Teras]

The state transition code from incomplete to invalid via neigh_update()
is missing the proper clean up of skb queue. Separate the clean up
code from neigh_timer_handler() to a new function and make neigh_update()
is it also.

Signed-off-by: Timo Teras <timo.teras@iki.fi>
---

 net/core/neighbour.c |   46 ++++++++++++++++++++++++++++------------------
 1 files changed, 28 insertions(+), 18 deletions(-)

diff --git a/net/core/neighbour.c b/net/core/neighbour.c
index a1cbce7..a17cb9d 100644
--- a/net/core/neighbour.c
+++ b/net/core/neighbour.c
@@ -771,6 +771,28 @@ static __inline__ int neigh_max_probes(struct neighbour *n)
 		p->ucast_probes + p->app_probes + p->mcast_probes);
 }
 
+static void neigh_invalidate(struct neighbour *neigh)
+{
+	struct sk_buff *skb;
+
+	NEIGH_CACHE_STAT_INC(neigh->tbl, res_failed);
+	NEIGH_PRINTK2("neigh %p is failed.\n", neigh);
+	neigh->updated = jiffies;
+
+	/* It is very thin place. report_unreachable is very complicated
+	   routine. Particularly, it can hit the same neighbour entry!
+
+	   So that, we try to be accurate and avoid dead loop. --ANK
+	 */
+	while (neigh->nud_state == NUD_FAILED &&
+	       (skb = __skb_dequeue(&neigh->arp_queue)) != NULL) {
+		write_unlock(&neigh->lock);
+		neigh->ops->error_report(neigh, skb);
+		write_lock(&neigh->lock);
+	}
+	skb_queue_purge(&neigh->arp_queue);
+}
+
 /* Called when a timer expires for a neighbour entry. */
 
 static void neigh_timer_handler(unsigned long arg)
@@ -835,26 +857,9 @@ static void neigh_timer_handler(unsigned long arg)
 
 	if ((neigh->nud_state & (NUD_INCOMPLETE | NUD_PROBE)) &&
 	    atomic_read(&neigh->probes) >= neigh_max_probes(neigh)) {
-		struct sk_buff *skb;
-
 		neigh->nud_state = NUD_FAILED;
-		neigh->updated = jiffies;
 		notify = 1;
-		NEIGH_CACHE_STAT_INC(neigh->tbl, res_failed);
-		NEIGH_PRINTK2("neigh %p is failed.\n", neigh);
-
-		/* It is very thin place. report_unreachable is very complicated
-		   routine. Particularly, it can hit the same neighbour entry!
-
-		   So that, we try to be accurate and avoid dead loop. --ANK
-		 */
-		while (neigh->nud_state == NUD_FAILED &&
-		       (skb = __skb_dequeue(&neigh->arp_queue)) != NULL) {
-			write_unlock(&neigh->lock);
-			neigh->ops->error_report(neigh, skb);
-			write_lock(&neigh->lock);
-		}
-		skb_queue_purge(&neigh->arp_queue);
+		neigh_invalidate(neigh);
 	}
 
 	if (neigh->nud_state & NUD_IN_TIMER) {
@@ -1001,6 +1006,11 @@ int neigh_update(struct neighbour *neigh, const u8 *lladdr, u8 new,
 		neigh->nud_state = new;
 		err = 0;
 		notify = old & NUD_VALID;
+		if ((old & (NUD_INCOMPLETE | NUD_PROBE)) &&
+		    (new & NUD_FAILED)) {
+			neigh_invalidate(neigh);
+			notify = 1;
+		}
 		goto out;
 	}
 
-- 
1.6.0.4

Re: [RESEND] [PATCH] neigh: error out pending skbs if netlink invalidates incomplete neigh

[David Miller]

You don't need to resend, you're in the queue already:

http://patchwork.ozlabs.org/patch/28091/

Re: [RESEND] [PATCH] neigh: error out pending skbs if netlink invalidates incomplete neigh

[Timo Teräs]

David Miller wrote:
> You don't need to resend, you're in the queue already:
> 
> http://patchwork.ozlabs.org/patch/28091/

I looking at the queue, but somehow missed the netdev project
even though it's right there in the main page.

Sorry.

- Timo