bonding: do not ignore multicast in mode 4
authorEric Kinzie <ekinzie@brocade.com>
Tue, 1 Mar 2016 17:32:00 +0000 (09:32 -0800)
committerThomas Monjalon <thomas.monjalon@6wind.com>
Wed, 16 Mar 2016 18:05:47 +0000 (19:05 +0100)
The bonding PMD in mode 4 puts all enslaved interfaces into promiscuous
mode in order to receive LACPDUs and must filter unwanted packets
after the traffic has been "collected".  Allow broadcast and multicast
through so that ARP and IPv6 neighbor discovery continue to work.

Fixes: 46fb43683679 ("bond: add mode 4")

Signed-off-by: Eric Kinzie <ekinzie@brocade.com>
Signed-off-by: Stephen Hemminger <stephen@networkplumber.org>
Acked-by: Declan Doherty <declan.doherty@intel.com>
app/test/test_link_bonding_mode4.c
drivers/net/bonding/rte_eth_bond_pmd.c

index 713368d..31640cd 100644 (file)
@@ -747,8 +747,11 @@ test_mode4_rx(void)
        rte_eth_macaddr_get(test_params.bonded_port_id, &bonded_mac);
        ether_addr_copy(&bonded_mac, &dst_mac);
 
-       /* Assert that dst address is not bonding address */
-       dst_mac.addr_bytes[0]++;
+       /* Assert that dst address is not bonding address.  Do not set the
+        * least significant bit of the zero byte as this would create a
+        * multicast address.
+        */
+       dst_mac.addr_bytes[0] += 2;
 
        /* First try with promiscuous mode enabled.
         * Add 2 packets to each slave. First with bonding MAC address, second with
index b63c886..011150a 100644 (file)
@@ -171,6 +171,7 @@ bond_ethdev_rx_burst_8023ad(void *queue, struct rte_mbuf **bufs,
                         * mode and packet address does not match. */
                        if (unlikely(hdr->ether_type == ether_type_slow_be ||
                                !collecting || (!promisc &&
+                                       !is_multicast_ether_addr(&hdr->d_addr) &&
                                        !is_same_ether_addr(&bond_mac, &hdr->d_addr)))) {
 
                                if (hdr->ether_type == ether_type_slow_be) {