ipvlan: fix multicast processing

In an IPvlan setup when master is set in loopback mode e.g.

  ethtool -K eth0 set loopback on

  where eth0 is master device for IPvlan setup.

The failure is caused by the faulty logic that determines if the
packet is from TX-path vs. RX-path by just looking at the mac-
addresses on the packet while processing multicast packets.

In the loopback-mode where this crash was happening, the packets
that are sent out are reflected by the NIC and are processed on
the RX path, but mac-address check tricks into thinking this
packet is from TX path and falsely uses dev_forward_skb() to pass
packets to the slave (virtual) devices.

This patch records the path while queueing packets and eliminates
logic of looking at mac-addresses for the same decision.

------------[ cut here ]------------
kernel BUG at include/linux/skbuff.h:1737!
Call Trace:
 [<ffffffff921fbbc2>] dev_forward_skb+0x92/0xd0
 [<ffffffffc031ac65>] ipvlan_process_multicast+0x395/0x4c0 [ipvlan]
 [<ffffffffc031a9a7>] ? ipvlan_process_multicast+0xd7/0x4c0 [ipvlan]
 [<ffffffff91cdfea7>] ? process_one_work+0x147/0x660
 [<ffffffff91cdff09>] process_one_work+0x1a9/0x660
 [<ffffffff91cdfea7>] ? process_one_work+0x147/0x660
 [<ffffffff91ce086d>] worker_thread+0x11d/0x360
 [<ffffffff91ce0750>] ? rescuer_thread+0x350/0x350
 [<ffffffff91ce960b>] kthread+0xdb/0xe0
 [<ffffffff91c05c70>] ? _raw_spin_unlock_irq+0x30/0x50
 [<ffffffff91ce9530>] ? flush_kthread_worker+0xc0/0xc0
 [<ffffffff92348b7a>] ret_from_fork+0x9a/0xd0
 [<ffffffff91ce9530>] ? flush_kthread_worker+0xc0/0xc0

Fixes: ba35f858 ("ipvlan: Defer multicast / broadcast processing to a work-queue")
Signed-off-by: Mahesh Bandewar <maheshb@google.com>
CC: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

drivers/net/ipvlan/ipvlan.h

@@ -99,6 +99,11 @@ struct ipvl_port {
 	int			count;
 };
 
+struct ipvl_skb_cb {
+	bool tx_pkt;
+};
+#define IPVL_SKB_CB(_skb) ((struct ipvl_skb_cb *)&((_skb)->cb[0]))
+
 static inline struct ipvl_port *ipvlan_port_get_rcu(const struct net_device *d)
 {
 	return rcu_dereference(d->rx_handler_data);

drivers/net/ipvlan/ipvlan_core.c

@@ -198,7 +198,7 @@ void ipvlan_process_multicast(struct work_struct *work)
 	unsigned int mac_hash;
 	int ret;
 	u8 pkt_type;
-	bool hlocal, dlocal;
+	bool tx_pkt;
 
 	__skb_queue_head_init(&list);
 
@@ -211,7 +211,7 @@ void ipvlan_process_multicast(struct work_struct *work)
 		bool consumed = false;
 
 		ethh = eth_hdr(skb);
-		hlocal = ether_addr_equal(ethh->h_source, port->dev->dev_addr);
+		tx_pkt = IPVL_SKB_CB(skb)->tx_pkt;
 		mac_hash = ipvlan_mac_hash(ethh->h_dest);
 
 		if (ether_addr_equal(ethh->h_dest, port->dev->broadcast))
@@ -219,13 +219,10 @@ void ipvlan_process_multicast(struct work_struct *work)
 		else
 			pkt_type = PACKET_MULTICAST;
 
-		dlocal = false;
 		rcu_read_lock();
 		list_for_each_entry_rcu(ipvlan, &port->ipvlans, pnode) {
-			if (hlocal && (ipvlan->dev == dev)) {
-				dlocal = true;
+			if (tx_pkt && (ipvlan->dev == skb->dev))
 				continue;
-			}
 			if (!test_bit(mac_hash, ipvlan->mac_filters))
 				continue;
 			if (!(ipvlan->dev->flags & IFF_UP))
@@ -238,7 +235,7 @@ void ipvlan_process_multicast(struct work_struct *work)
 				consumed = true;
 				nskb->pkt_type = pkt_type;
 				nskb->dev = ipvlan->dev;
-				if (hlocal)
+				if (tx_pkt)
 					ret = dev_forward_skb(ipvlan->dev, nskb);
 				else
 					ret = netif_rx(nskb);
@@ -248,7 +245,7 @@ void ipvlan_process_multicast(struct work_struct *work)
 		}
 		rcu_read_unlock();
 
-		if (dlocal) {
+		if (tx_pkt) {
 			/* If the packet originated here, send it out. */
 			skb->dev = port->dev;
 			skb->pkt_type = pkt_type;
@@ -480,13 +477,20 @@ static int ipvlan_process_outbound(struct sk_buff *skb)
 }
 
 static void ipvlan_multicast_enqueue(struct ipvl_port *port,
-				     struct sk_buff *skb)
+				     struct sk_buff *skb, bool tx_pkt)
 {
 	if (skb->protocol == htons(ETH_P_PAUSE)) {
 		kfree_skb(skb);
 		return;
 	}
 
+	/* Record that the deferred packet is from TX or RX path. By
+	 * looking at mac-addresses on packet will lead to erronus decisions.
+	 * (This would be true for a loopback-mode on master device or a
+	 * hair-pin mode of the switch.)
+	 */
+	IPVL_SKB_CB(skb)->tx_pkt = tx_pkt;
+
 	spin_lock(&port->backlog.lock);
 	if (skb_queue_len(&port->backlog) < IPVLAN_QBACKLOG_LIMIT) {
 		if (skb->dev)
@@ -549,7 +553,7 @@ static int ipvlan_xmit_mode_l2(struct sk_buff *skb, struct net_device *dev)
 
 	} else if (is_multicast_ether_addr(eth->h_dest)) {
 		ipvlan_skb_crossing_ns(skb, NULL);
-		ipvlan_multicast_enqueue(ipvlan->port, skb);
+		ipvlan_multicast_enqueue(ipvlan->port, skb, true);
 		return NET_XMIT_SUCCESS;
 	}
 
@@ -646,7 +650,7 @@ static rx_handler_result_t ipvlan_handle_mode_l2(struct sk_buff **pskb,
 			 */
 			if (nskb) {
 				ipvlan_skb_crossing_ns(nskb, NULL);
-				ipvlan_multicast_enqueue(port, nskb);
+				ipvlan_multicast_enqueue(port, nskb, false);
 			}
 		}
 	} else {