[PATCH] net: Fix skb_under_panic oops in neigh_resolve_output

From: Ramesh Nagappa <redacted>

The retry loop in neigh_resolve_output() and neigh_connected_output()
call dev_hard_header() with out reseting the skb to network_header.
This causes the retry to fail with skb_under_panic. The fix is to
reset the network_header within the retry loop.

Signed-off-by: Ramesh Nagappa <redacted>
Reviewed-by: Shawn Lu <redacted>
Reviewed-by: Robert Coulson <redacted>
Reviewed-by: Billie Alsup <redacted>
---
 net/core/neighbour.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/net/core/neighbour.c b/net/core/neighbour.c
index 117afaf..f50c542 100644
--- a/net/core/neighbour.c
+++ b/net/core/neighbour.c

@@ -1313,6 +1313,7 @@ int neigh_resolve_output(struct neighbour *neigh, struct sk_buff *skb)
 
 		do {
 			seq = read_seqbegin(&neigh->ha_lock);
+			__skb_pull(skb, skb_network_offset(skb));
 			err = dev_hard_header(skb, dev, ntohs(skb->protocol),
 					      neigh->ha, NULL, skb->len);
 		} while (read_seqretry(&neigh->ha_lock, seq));

@@ -1342,10 +1343,9 @@ int neigh_connected_output(struct neighbour *neigh, struct sk_buff *skb)
 	unsigned int seq;
 	int err;
 
-	__skb_pull(skb, skb_network_offset(skb));
-
 	do {
 		seq = read_seqbegin(&neigh->ha_lock);
+		__skb_pull(skb, skb_network_offset(skb));
 		err = dev_hard_header(skb, dev, ntohs(skb->protocol),
 				      neigh->ha, NULL, skb->len);
 	} while (read_seqretry(&neigh->ha_lock, seq));

-- 
1.7.11.4

On Thu, 2012-10-04 at 20:05 -0700, ramesh.nagappa@gmail.com wrote:

This __skb_pull() could be done before the read_seqbegin() to keep the
protected section short.

This __skb_pull() could be done before the read_seqbegin() to keep the
protected section short.

 		err = dev_hard_header(skb, dev, ntohs(skb->protocol),
 				      neigh->ha, NULL, skb->len);
 	} while (read_seqretry(&neigh->ha_lock, seq));

Thanks

Hi Eric,

Yes, that is a good optimization. neigh_resolve_output() also has the
__skb_pull() outside the loop, is that required ? The changes would be
like ...

neigh_resolve_output()
...
-        __skb_pull(skb, skb_network_offset(skb));


        if (!neigh_event_send(neigh, skb)) {
                int err;
                struct net_device *dev = neigh->dev;
                unsigned int seq;

                if (dev->header_ops->cache && !neigh->hh.hh_len)
                        neigh_hh_init(neigh, dst);

                do {                        
+			      __skb_pull(skb, skb_network_offset(skb));
                        seq = read_seqbegin(&neigh->ha_lock);
                        err = dev_hard_header(skb, dev, ntohs(skb->protocol),
                                              neigh->ha, NULL, skb->len);
                } while (read_seqretry(&neigh->ha_lock, seq));

Thanks,
Ramesh
 

-----Original Message-----
From: Eric Dumazet [mailto:eric.dumazet@gmail.com] 

quoted

 	do {
 		seq = read_seqbegin(&neigh->ha_lock);
+		__skb_pull(skb, skb_network_offset(skb));

This __skb_pull() could be done before the read_seqbegin() to 
keep the protected section short.

quoted

 		err = dev_hard_header(skb, dev, ntohs(skb->protocol),
 				      neigh->ha, NULL, skb->len);
 	} while (read_seqretry(&neigh->ha_lock, seq));

Thanks

On Fri, 2012-10-05 at 12:33 -0400, Ramesh Nagappa wrote:

Hi Eric,

Yes, that is a good optimization. neigh_resolve_output() also has the
__skb_pull() outside the loop, is that required ? The changes would be
like ...

neigh_resolve_output()
...
-        __skb_pull(skb, skb_network_offset(skb));


        if (!neigh_event_send(neigh, skb)) {
                int err;
                struct net_device *dev = neigh->dev;
                unsigned int seq;

                if (dev->header_ops->cache && !neigh->hh.hh_len)
                        neigh_hh_init(neigh, dst);

                do {                        
+			      __skb_pull(skb, skb_network_offset(skb));
                        seq = read_seqbegin(&neigh->ha_lock);
                        err = dev_hard_header(skb, dev, ntohs(skb->protocol),
                                              neigh->ha, NULL, skb->len);
                } while (read_seqretry(&neigh->ha_lock, seq));

All similar constructions should be audited.

For example in net/decnet/dn_neigh.c , dn_neigh_output_packet()