From: Cong Wang <amwang@xxxxxxxxxx> This is a long-standing bug and reported several times: https://bugzilla.redhat.com/show_bug.cgi?id=880035 http://marc.info/?l=linux-netdev&m=136164389416341&w=2 This bug can be observed in virt environment, when a KVM guest communicates with the host via multicast. After some time (should be 260 sec, I didn't measure), the multicast traffic suddenly terminates. This is due to the mdb entry for bridge itself expires automatically, it should not expire as long as the bridge still generates multicast traffic. It should expire when the bridge leaves the multicast group, OR when there is no multicast traffic on this bridge. I fix this by adding another bool which is set when there is multicast traffic goes to the bridge, cleared in the expire timer and when IGMP leave is received. I ran omping for 15 minutes, everything looks good now. Cc: Herbert Xu <herbert@xxxxxxxxxxxxxxxxxxx> Cc: Stephen Hemminger <stephen@xxxxxxxxxxxxxxxxxx> Cc: "David S. Miller" <davem@xxxxxxxxxxxxx> Cc: Adam Baker <linux@xxxxxxxxxxxxxxxx> Signed-off-by: Cong Wang <amwang@xxxxxxxxxx> --- diff --git a/net/bridge/br_input.c b/net/bridge/br_input.c index 828e2bc..c706619 100644 --- a/net/bridge/br_input.c +++ b/net/bridge/br_input.c @@ -99,9 +99,11 @@ int br_handle_frame_finish(struct sk_buff *skb) else if (is_multicast_ether_addr(dest)) { mdst = br_mdb_get(br, skb, vid); if (mdst || BR_INPUT_SKB_CB_MROUTERS_ONLY(skb)) { - if ((mdst && mdst->mglist) || - br_multicast_is_router(br)) + bool to_br = mdst && mdst->mglist; + if (to_br || br_multicast_is_router(br)) skb2 = skb; + if (to_br) + mdst->busy = true; br_multicast_forward(mdst, skb, skb2); skb = NULL; if (!skb2) diff --git a/net/bridge/br_multicast.c b/net/bridge/br_multicast.c index 923fbea..4b12a0d 100644 --- a/net/bridge/br_multicast.c +++ b/net/bridge/br_multicast.c @@ -228,11 +228,16 @@ static void br_multicast_group_expired(unsigned long data) if (!netif_running(br->dev) || timer_pending(&mp->timer)) goto out; - mp->mglist = false; - if (mp->ports) goto out; + if (mp->mglist && mp->busy) { + mp->busy = false; + goto out; + } + + mp->mglist = false; + mdb = mlock_dereference(br->mdb, br); hlist_del_rcu(&mp->hlist[mdb->ver]); @@ -1280,6 +1285,7 @@ static void br_multicast_leave_group(struct net_bridge *br, mod_timer(&mp->timer, time); } + mp->busy = false; goto out; } diff --git a/net/bridge/br_private.h b/net/bridge/br_private.h index 3cbf5be..bb9a3e4 100644 --- a/net/bridge/br_private.h +++ b/net/bridge/br_private.h @@ -112,6 +112,7 @@ struct net_bridge_mdb_entry struct timer_list timer; struct br_ip addr; bool mglist; + bool busy; /* update this only when mglist == true */ }; struct net_bridge_mdb_htable
