nfp: ensure skb network header is set for packet redirect

Packets received at the NFP driver may be redirected to egress of another
netdev (e.g. in the case of OvS internal ports). On the egress path, some
processes, like TC egress hooks, may expect the network header offset
field in the skb to be correctly set. If this is not the case there is
potential for abnormal behaviour and even the triggering of BUG() calls.

Set the skb network header field before the mac header pull when doing a
packet redirect.

Fixes: 27f54b582567 ("nfp: allow fallback packets from non-reprs")
Signed-off-by: John Hurley <john.hurley@netronome.com>
Reviewed-by: Jakub Kicinski <jakub.kicinski@netronome.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

diff --git a/drivers/net/ethernet/netronome/nfp/nfp_net_common.c b/drivers/net/ethernet/netronome/nfp/nfp_net_common.c
index b82b684..36a3bd3 100644 (file)
--- a/drivers/net/ethernet/netronome/nfp/nfp_net_common.c
+++ b/drivers/net/ethernet/netronome/nfp/nfp_net_common.c
@@ -1867,6 +1867,7 @@ static int nfp_net_rx(struct nfp_net_rx_ring *rx_ring, int budget)
                        napi_gro_receive(&rx_ring->r_vec->napi, skb);
                } else {
                        skb->dev = netdev;
+                       skb_reset_network_header(skb);
                        __skb_push(skb, ETH_HLEN);
                        dev_queue_xmit(skb);
                }