net: bridge: add support for IGMP/MLD stats and export them via netlink

This patch adds stats support for the currently used IGMP/MLD types by the
bridge. The stats are per-port (plus one stat per-bridge) and per-direction
(RX/TX). The stats are exported via netlink via the new linkxstats API
(RTM_GETSTATS). In order to minimize the performance impact, a new option
is used to enable/disable the stats - multicast_stats_enabled, similar to
the recent vlan stats. Also in order to avoid multiple IGMP/MLD type
lookups and checks, we make use of the current "igmp" member of the bridge
private skb->cb region to record the type on Rx (both host-generated and
external packets pass by multicast_rcv()). We can do that since the igmp
member was used as a boolean and all the valid IGMP/MLD types are positive
values. The normal bridge fast-path is not affected at all, the only
affected paths are the flooding ones and since we make use of the IGMP/MLD
type, we can quickly determine if the packet should be counted using
cache-hot data (cb's igmp member). We add counters for:
* IGMP Queries
* IGMP Leaves
* IGMP v1/v2/v3 reports

* MLD Queries
* MLD Leaves
* MLD v1/v2 reports

These are invaluable when monitoring or debugging complex multicast setups
with bridges.

Signed-off-by: Nikolay Aleksandrov <nikolay@cumulusnetworks.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

1 files changed, 203 insertions, 14 deletions

diff --git a/net/bridge/br_multicast.c b/net/bridge/br_multicast.c
index 43844144c9c4..e405eef0ae2e 100644
--- a/net/bridge/br_multicast.c
+++ b/net/bridge/br_multicast.c
@@ -361,7 +361,8 @@ out:
 }
 
 static struct sk_buff *br_ip4_multicast_alloc_query(struct net_bridge *br,
-						    __be32 group)
+						    __be32 group,
+						    u8 *igmp_type)
 {
 	struct sk_buff *skb;
 	struct igmphdr *ih;
@@ -411,6 +412,7 @@ static struct sk_buff *br_ip4_multicast_alloc_query(struct net_bridge *br,
 
 	skb_set_transport_header(skb, skb->len);
 	ih = igmp_hdr(skb);
+	*igmp_type = IGMP_HOST_MEMBERSHIP_QUERY;
 	ih->type = IGMP_HOST_MEMBERSHIP_QUERY;
 	ih->code = (group ? br->multicast_last_member_interval :
 			    br->multicast_query_response_interval) /
@@ -428,7 +430,8 @@ out:
 
 #if IS_ENABLED(CONFIG_IPV6)
 static struct sk_buff *br_ip6_multicast_alloc_query(struct net_bridge *br,
-						    const struct in6_addr *group)
+						    const struct in6_addr *grp,
+						    u8 *igmp_type)
 {
 	struct sk_buff *skb;
 	struct ipv6hdr *ip6h;
@@ -487,16 +490,17 @@ static struct sk_buff *br_ip6_multicast_alloc_query(struct net_bridge *br,
 	skb_set_transport_header(skb, skb->len);
 	mldq = (struct mld_msg *) icmp6_hdr(skb);
 
-	interval = ipv6_addr_any(group) ?
+	interval = ipv6_addr_any(grp) ?
 			br->multicast_query_response_interval :
 			br->multicast_last_member_interval;
 
+	*igmp_type = ICMPV6_MGM_QUERY;
 	mldq->mld_type = ICMPV6_MGM_QUERY;
 	mldq->mld_code = 0;
 	mldq->mld_cksum = 0;
 	mldq->mld_maxdelay = htons((u16)jiffies_to_msecs(interval));
 	mldq->mld_reserved = 0;
-	mldq->mld_mca = *group;
+	mldq->mld_mca = *grp;
 
 	/* checksum */
 	mldq->mld_cksum = csum_ipv6_magic(&ip6h->saddr, &ip6h->daddr,
@@ -513,14 +517,16 @@ out:
 #endif
 
 static struct sk_buff *br_multicast_alloc_query(struct net_bridge *br,
-						struct br_ip *addr)
+						struct br_ip *addr,
+						u8 *igmp_type)
 {
 	switch (addr->proto) {
 	case htons(ETH_P_IP):
-		return br_ip4_multicast_alloc_query(br, addr->u.ip4);
+		return br_ip4_multicast_alloc_query(br, addr->u.ip4, igmp_type);
 #if IS_ENABLED(CONFIG_IPV6)
 	case htons(ETH_P_IPV6):
-		return br_ip6_multicast_alloc_query(br, &addr->u.ip6);
+		return br_ip6_multicast_alloc_query(br, &addr->u.ip6,
+						    igmp_type);
 #endif
 	}
 	return NULL;
@@ -829,18 +835,23 @@ static void __br_multicast_send_query(struct net_bridge *br,
 				      struct br_ip *ip)
 {
 	struct sk_buff *skb;
+	u8 igmp_type;
 
-	skb = br_multicast_alloc_query(br, ip);
+	skb = br_multicast_alloc_query(br, ip, &igmp_type);
 	if (!skb)
 		return;
 
 	if (port) {
 		skb->dev = port->dev;
+		br_multicast_count(br, port, skb->protocol, igmp_type,
+				   BR_MCAST_DIR_TX);
 		NF_HOOK(NFPROTO_BRIDGE, NF_BR_LOCAL_OUT,
 			dev_net(port->dev), NULL, skb, NULL, skb->dev,
 			br_dev_queue_push_xmit);
 	} else {
 		br_multicast_select_own_querier(br, ip, skb);
+		br_multicast_count(br, port, skb->protocol, igmp_type,
+				   BR_MCAST_DIR_RX);
 		netif_rx(skb);
 	}
 }
@@ -918,7 +929,7 @@ static void br_ip6_multicast_port_query_expired(unsigned long data)
 }
 #endif
 
-void br_multicast_add_port(struct net_bridge_port *port)
+int br_multicast_add_port(struct net_bridge_port *port)
 {
 	port->multicast_router = MDB_RTR_TYPE_TEMP_QUERY;
 
@@ -930,6 +941,11 @@ void br_multicast_add_port(struct net_bridge_port *port)
 	setup_timer(&port->ip6_own_query.timer,
 		    br_ip6_multicast_port_query_expired, (unsigned long)port);
 #endif
+	port->mcast_stats = netdev_alloc_pcpu_stats(struct bridge_mcast_stats);
+	if (!port->mcast_stats)
+		return -ENOMEM;
+
+	return 0;
 }
 
 void br_multicast_del_port(struct net_bridge_port *port)
@@ -944,6 +960,7 @@ void br_multicast_del_port(struct net_bridge_port *port)
 		br_multicast_del_pg(br, pg);
 	spin_unlock_bh(&br->multicast_lock);
 	del_timer_sync(&port->multicast_router_timer);
+	free_percpu(port->mcast_stats);
 }
 
 static void br_multicast_enable(struct bridge_mcast_own_query *query)
@@ -1583,6 +1600,39 @@ static void br_ip6_multicast_leave_group(struct net_bridge *br,
 }
 #endif
 
+static void br_multicast_err_count(const struct net_bridge *br,
+				   const struct net_bridge_port *p,
+				   __be16 proto)
+{
+	struct bridge_mcast_stats __percpu *stats;
+	struct bridge_mcast_stats *pstats;
+
+	if (!br->multicast_stats_enabled)
+		return;
+
+	if (p)
+		stats = p->mcast_stats;
+	else
+		stats = br->mcast_stats;
+	if (WARN_ON(!stats))
+		return;
+
+	pstats = this_cpu_ptr(stats);
+
+	u64_stats_update_begin(&pstats->syncp);
+	switch (proto) {
+	case htons(ETH_P_IP):
+		pstats->mstats.igmp_parse_errors++;
+		break;
+#if IS_ENABLED(CONFIG_IPV6)
+	case htons(ETH_P_IPV6):
+		pstats->mstats.mld_parse_errors++;
+		break;
+#endif
+	}
+	u64_stats_update_end(&pstats->syncp);
+}
+
 static int br_multicast_ipv4_rcv(struct net_bridge *br,
 				 struct net_bridge_port *port,
 				 struct sk_buff *skb,
@@ -1599,11 +1649,12 @@ static int br_multicast_ipv4_rcv(struct net_bridge *br,
 			BR_INPUT_SKB_CB(skb)->mrouters_only = 1;
 		return 0;
 	} else if (err < 0) {
+		br_multicast_err_count(br, port, skb->protocol);
 		return err;
 	}
 
-	BR_INPUT_SKB_CB(skb)->igmp = 1;
 	ih = igmp_hdr(skb);
+	BR_INPUT_SKB_CB(skb)->igmp = ih->type;
 
 	switch (ih->type) {
 	case IGMP_HOST_MEMBERSHIP_REPORT:
@@ -1625,6 +1676,9 @@ static int br_multicast_ipv4_rcv(struct net_bridge *br,
 	if (skb_trimmed && skb_trimmed != skb)
 		kfree_skb(skb_trimmed);
 
+	br_multicast_count(br, port, skb->protocol, BR_INPUT_SKB_CB(skb)->igmp,
+			   BR_MCAST_DIR_RX);
+
 	return err;
 }
 
@@ -1645,11 +1699,12 @@ static int br_multicast_ipv6_rcv(struct net_bridge *br,
 			BR_INPUT_SKB_CB(skb)->mrouters_only = 1;
 		return 0;
 	} else if (err < 0) {
+		br_multicast_err_count(br, port, skb->protocol);
 		return err;
 	}
 
-	BR_INPUT_SKB_CB(skb)->igmp = 1;
 	mld = (struct mld_msg *)skb_transport_header(skb);
+	BR_INPUT_SKB_CB(skb)->igmp = mld->mld_type;
 
 	switch (mld->mld_type) {
 	case ICMPV6_MGM_REPORT:
@@ -1670,6 +1725,9 @@ static int br_multicast_ipv6_rcv(struct net_bridge *br,
 	if (skb_trimmed && skb_trimmed != skb)
 		kfree_skb(skb_trimmed);
 
+	br_multicast_count(br, port, skb->protocol, BR_INPUT_SKB_CB(skb)->igmp,
+			   BR_MCAST_DIR_RX);
+
 	return err;
 }
 #endif
@@ -1677,6 +1735,8 @@ static int br_multicast_ipv6_rcv(struct net_bridge *br,
 int br_multicast_rcv(struct net_bridge *br, struct net_bridge_port *port,
 		     struct sk_buff *skb, u16 vid)
 {
+	int ret = 0;
+
 	BR_INPUT_SKB_CB(skb)->igmp = 0;
 	BR_INPUT_SKB_CB(skb)->mrouters_only = 0;
 
@@ -1685,14 +1745,16 @@ int br_multicast_rcv(struct net_bridge *br, struct net_bridge_port *port,
 
 	switch (skb->protocol) {
 	case htons(ETH_P_IP):
-		return br_multicast_ipv4_rcv(br, port, skb, vid);
+		ret = br_multicast_ipv4_rcv(br, port, skb, vid);
+		break;
 #if IS_ENABLED(CONFIG_IPV6)
 	case htons(ETH_P_IPV6):
-		return br_multicast_ipv6_rcv(br, port, skb, vid);
+		ret = br_multicast_ipv6_rcv(br, port, skb, vid);
+		break;
 #endif
 	}
 
-	return 0;
+	return ret;
 }
 
 static void br_multicast_query_expired(struct net_bridge *br,
@@ -1831,6 +1893,8 @@ void br_multicast_dev_del(struct net_bridge *br)
 
 out:
 	spin_unlock_bh(&br->multicast_lock);
+
+	free_percpu(br->mcast_stats);
 }
 
 int br_multicast_set_router(struct net_bridge *br, unsigned long val)
@@ -2185,3 +2249,128 @@ unlock:
 	return ret;
 }
 EXPORT_SYMBOL_GPL(br_multicast_has_querier_adjacent);
+
+static void br_mcast_stats_add(struct bridge_mcast_stats __percpu *stats,
+			       __be16 proto, u8 type, u8 dir)
+{
+	struct bridge_mcast_stats *pstats = this_cpu_ptr(stats);
+
+	u64_stats_update_begin(&pstats->syncp);
+	switch (proto) {
+	case htons(ETH_P_IP):
+		switch (type) {
+		case IGMP_HOST_MEMBERSHIP_REPORT:
+			pstats->mstats.igmp_v1reports[dir]++;
+			break;
+		case IGMPV2_HOST_MEMBERSHIP_REPORT:
+			pstats->mstats.igmp_v2reports[dir]++;
+			break;
+		case IGMPV3_HOST_MEMBERSHIP_REPORT:
+			pstats->mstats.igmp_v3reports[dir]++;
+			break;
+		case IGMP_HOST_MEMBERSHIP_QUERY:
+			pstats->mstats.igmp_queries[dir]++;
+			break;
+		case IGMP_HOST_LEAVE_MESSAGE:
+			pstats->mstats.igmp_leaves[dir]++;
+			break;
+		}
+		break;
+#if IS_ENABLED(CONFIG_IPV6)
+	case htons(ETH_P_IPV6):
+		switch (type) {
+		case ICMPV6_MGM_REPORT:
+			pstats->mstats.mld_v1reports[dir]++;
+			break;
+		case ICMPV6_MLD2_REPORT:
+			pstats->mstats.mld_v2reports[dir]++;
+			break;
+		case ICMPV6_MGM_QUERY:
+			pstats->mstats.mld_queries[dir]++;
+			break;
+		case ICMPV6_MGM_REDUCTION:
+			pstats->mstats.mld_leaves[dir]++;
+			break;
+		}
+		break;
+#endif /* CONFIG_IPV6 */
+	}
+	u64_stats_update_end(&pstats->syncp);
+}
+
+void br_multicast_count(struct net_bridge *br, const struct net_bridge_port *p,
+			__be16 proto, u8 type, u8 dir)
+{
+	struct bridge_mcast_stats __percpu *stats;
+
+	/* if multicast_disabled is true then igmp type can't be set */
+	if (!type || !br->multicast_stats_enabled)
+		return;
+
+	if (p)
+		stats = p->mcast_stats;
+	else
+		stats = br->mcast_stats;
+	if (WARN_ON(!stats))
+		return;
+
+	br_mcast_stats_add(stats, proto, type, dir);
+}
+
+int br_multicast_init_stats(struct net_bridge *br)
+{
+	br->mcast_stats = netdev_alloc_pcpu_stats(struct bridge_mcast_stats);
+	if (!br->mcast_stats)
+		return -ENOMEM;
+
+	return 0;
+}
+
+static void mcast_stats_add_dir(u64 *dst, u64 *src)
+{
+	dst[BR_MCAST_DIR_RX] += src[BR_MCAST_DIR_RX];
+	dst[BR_MCAST_DIR_TX] += src[BR_MCAST_DIR_TX];
+}
+
+void br_multicast_get_stats(const struct net_bridge *br,
+			    const struct net_bridge_port *p,
+			    struct br_mcast_stats *dest)
+{
+	struct bridge_mcast_stats __percpu *stats;
+	struct br_mcast_stats tdst;
+	int i;
+
+	memset(dest, 0, sizeof(*dest));
+	if (p)
+		stats = p->mcast_stats;
+	else
+		stats = br->mcast_stats;
+	if (WARN_ON(!stats))
+		return;
+
+	memset(&tdst, 0, sizeof(tdst));
+	for_each_possible_cpu(i) {
+		struct bridge_mcast_stats *cpu_stats = per_cpu_ptr(stats, i);
+		struct br_mcast_stats temp;
+		unsigned int start;
+
+		do {
+			start = u64_stats_fetch_begin_irq(&cpu_stats->syncp);
+			memcpy(&temp, &cpu_stats->mstats, sizeof(temp));
+		} while (u64_stats_fetch_retry_irq(&cpu_stats->syncp, start));
+
+		mcast_stats_add_dir(tdst.igmp_queries, temp.igmp_queries);
+		mcast_stats_add_dir(tdst.igmp_leaves, temp.igmp_leaves);
+		mcast_stats_add_dir(tdst.igmp_v1reports, temp.igmp_v1reports);
+		mcast_stats_add_dir(tdst.igmp_v2reports, temp.igmp_v2reports);
+		mcast_stats_add_dir(tdst.igmp_v3reports, temp.igmp_v3reports);
+		tdst.igmp_parse_errors += temp.igmp_parse_errors;
+
+		mcast_stats_add_dir(tdst.mld_queries, temp.mld_queries);
+		mcast_stats_add_dir(tdst.mld_leaves, temp.mld_leaves);
+		mcast_stats_add_dir(tdst.mld_v1reports, temp.mld_v1reports);
+		mcast_stats_add_dir(tdst.mld_v2reports, temp.mld_v2reports);
+		tdst.mld_parse_errors += temp.mld_parse_errors;
+	}
+	memcpy(dest, &tdst, sizeof(*dest));
+}