Merge branch 'bridge'

Toshiaki Makita says:

====================
bridge: Fix corner case problems around local fdb entries

There are so many corner cases that are not handled properly around local
fdb entries.
- We might fail to delete the old entry and might delete an arbitrary local
  entry when changing mac address of a bridge port.
- We always fail to delete the old entry when changing mac address of the
  bridge device.
- We might incorrectly delete a necessary entry when detaching a bridge port.
- We might incorrectly delete a necessary entry when deleting a vlan.
and so on.

This is a patch series to fix these issues.

v3:
- Handle NTF_USE case in patch 1/9, commented by Vlad Yasevich.

- Tested port detach/attach and didn't find any problem with patch 5/9,
  suggested by Stephen Hemminger.

- Add comments about possible inconsistent state in current implementation
  into commit log of patch 5/9, found by the above test.

- Reword unintensive changelog of patch 7/9, commented by Vlad Yasevich.

v2:
- Change the way to find the old entry in br_fdb_changeaddr() from memorizing
  previous port address to introducing a new flag indicating whether a fdb
  entry is added by user or not, commented by Stephen Hemminger.

- Add a fix for the way to insert a new address in br_fdb_changeaddr().

- Prevent creating an entry such that its dst is NULL in br_add_if() to
  preserve old behavior, commented by Vlad Yasevich.

- Add more comments about slight behavior change, where the bridge device
  come to be able to receive traffic to an address it has during short
  window, to changelogs, commented by Vlad Yasevich.

- Add a fix for possible race in br_fdb_change_mac_address().
====================

Signed-off-by: David S. Miller <davem@davemloft.net>

7 files changed, 129 insertions, 63 deletions

diff --git a/net/bridge/br_device.c b/net/bridge/br_device.c
index d9a9b0f..63f0455 100644
--- a/net/bridge/br_device.c
+++ b/net/bridge/br_device.c
@@ -187,8 +187,7 @@ static int br_set_mac_address(struct net_device *dev, void *p)
 
 	spin_lock_bh(&br->lock);
 	if (!ether_addr_equal(dev->dev_addr, addr->sa_data)) {
-		memcpy(dev->dev_addr, addr->sa_data, ETH_ALEN);
-		br_fdb_change_mac_address(br, addr->sa_data);
+		/* Mac address will be changed in br_stp_change_bridge_id(). */
 		br_stp_change_bridge_id(br, addr->sa_data);
 	}
 	spin_unlock_bh(&br->lock);
diff --git a/net/bridge/br_fdb.c b/net/bridge/br_fdb.c
index c5f5a4a..9203d5a 100644
--- a/net/bridge/br_fdb.c
+++ b/net/bridge/br_fdb.c
@@ -27,6 +27,9 @@
 #include "br_private.h"
 
 static struct kmem_cache *br_fdb_cache __read_mostly;
+static struct net_bridge_fdb_entry *fdb_find(struct hlist_head *head,
+					     const unsigned char *addr,
+					     __u16 vid);
 static int fdb_insert(struct net_bridge *br, struct net_bridge_port *source,
 		      const unsigned char *addr, u16 vid);
 static void fdb_notify(struct net_bridge *br,
@@ -89,11 +92,57 @@ static void fdb_delete(struct net_bridge *br, struct net_bridge_fdb_entry *f)
 	call_rcu(&f->rcu, fdb_rcu_free);
 }
 
+/* Delete a local entry if no other port had the same address. */
+static void fdb_delete_local(struct net_bridge *br,
+			     const struct net_bridge_port *p,
+			     struct net_bridge_fdb_entry *f)
+{
+	const unsigned char *addr = f->addr.addr;
+	u16 vid = f->vlan_id;
+	struct net_bridge_port *op;
+
+	/* Maybe another port has same hw addr? */
+	list_for_each_entry(op, &br->port_list, list) {
+		if (op != p && ether_addr_equal(op->dev->dev_addr, addr) &&
+		    (!vid || nbp_vlan_find(op, vid))) {
+			f->dst = op;
+			f->added_by_user = 0;
+			return;
+		}
+	}
+
+	/* Maybe bridge device has same hw addr? */
+	if (p && ether_addr_equal(br->dev->dev_addr, addr) &&
+	    (!vid || br_vlan_find(br, vid))) {
+		f->dst = NULL;
+		f->added_by_user = 0;
+		return;
+	}
+
+	fdb_delete(br, f);
+}
+
+void br_fdb_find_delete_local(struct net_bridge *br,
+			      const struct net_bridge_port *p,
+			      const unsigned char *addr, u16 vid)
+{
+	struct hlist_head *head = &br->hash[br_mac_hash(addr, vid)];
+	struct net_bridge_fdb_entry *f;
+
+	spin_lock_bh(&br->hash_lock);
+	f = fdb_find(head, addr, vid);
+	if (f && f->is_local && !f->added_by_user && f->dst == p)
+		fdb_delete_local(br, p, f);
+	spin_unlock_bh(&br->hash_lock);
+}
+
 void br_fdb_changeaddr(struct net_bridge_port *p, const unsigned char *newaddr)
 {
 	struct net_bridge *br = p->br;
-	bool no_vlan = (nbp_get_vlan_info(p) == NULL) ? true : false;
+	struct net_port_vlans *pv = nbp_get_vlan_info(p);
+	bool no_vlan = !pv;
 	int i;
+	u16 vid;
 
 	spin_lock_bh(&br->hash_lock);
 
@@ -104,38 +153,34 @@ void br_fdb_changeaddr(struct net_bridge_port *p, const unsigned char *newaddr)
 			struct net_bridge_fdb_entry *f;
 
 			f = hlist_entry(h, struct net_bridge_fdb_entry, hlist);
-			if (f->dst == p && f->is_local) {
-				/* maybe another port has same hw addr? */
-				struct net_bridge_port *op;
-				u16 vid = f->vlan_id;
-				list_for_each_entry(op, &br->port_list, list) {
-					if (op != p &&
-					    ether_addr_equal(op->dev->dev_addr,
-							     f->addr.addr) &&
-					    nbp_vlan_find(op, vid)) {
-						f->dst = op;
-						goto insert;
-					}
-				}
-
+			if (f->dst == p && f->is_local && !f->added_by_user) {
 				/* delete old one */
-				fdb_delete(br, f);
-insert:
-				/* insert new address,  may fail if invalid
-				 * address or dup.
-				 */
-				fdb_insert(br, p, newaddr, vid);
+				fdb_delete_local(br, p, f);
 
 				/* if this port has no vlan information
 				 * configured, we can safely be done at
 				 * this point.
 				 */
 				if (no_vlan)
-					goto done;
+					goto insert;
 			}
 		}
 	}
 
+insert:
+	/* insert new address,  may fail if invalid address or dup. */
+	fdb_insert(br, p, newaddr, 0);
+
+	if (no_vlan)
+		goto done;
+
+	/* Now add entries for every VLAN configured on the port.
+	 * This function runs under RTNL so the bitmap will not change
+	 * from under us.
+	 */
+	for_each_set_bit(vid, pv->vlan_bitmap, VLAN_N_VID)
+		fdb_insert(br, p, newaddr, vid);
+
 done:
 	spin_unlock_bh(&br->hash_lock);
 }
@@ -146,10 +191,12 @@ void br_fdb_change_mac_address(struct net_bridge *br, const u8 *newaddr)
 	struct net_port_vlans *pv;
 	u16 vid = 0;
 
+	spin_lock_bh(&br->hash_lock);
+
 	/* If old entry was unassociated with any port, then delete it. */
 	f = __br_fdb_get(br, br->dev->dev_addr, 0);
 	if (f && f->is_local && !f->dst)
-		fdb_delete(br, f);
+		fdb_delete_local(br, NULL, f);
 
 	fdb_insert(br, NULL, newaddr, 0);
 
@@ -159,14 +206,16 @@ void br_fdb_change_mac_address(struct net_bridge *br, const u8 *newaddr)
 	 */
 	pv = br_get_vlan_info(br);
 	if (!pv)
-		return;
+		goto out;
 
 	for_each_set_bit_from(vid, pv->vlan_bitmap, VLAN_N_VID) {
 		f = __br_fdb_get(br, br->dev->dev_addr, vid);
 		if (f && f->is_local && !f->dst)
-			fdb_delete(br, f);
+			fdb_delete_local(br, NULL, f);
 		fdb_insert(br, NULL, newaddr, vid);
 	}
+out:
+	spin_unlock_bh(&br->hash_lock);
 }
 
 void br_fdb_cleanup(unsigned long _data)
@@ -235,25 +284,11 @@ void br_fdb_delete_by_port(struct net_bridge *br,
 
 			if (f->is_static && !do_all)
 				continue;
-			/*
-			 * if multiple ports all have the same device address
-			 * then when one port is deleted, assign
-			 * the local entry to other port
-			 */
-			if (f->is_local) {
-				struct net_bridge_port *op;
-				list_for_each_entry(op, &br->port_list, list) {
-					if (op != p &&
-					    ether_addr_equal(op->dev->dev_addr,
-							     f->addr.addr)) {
-						f->dst = op;
-						goto skip_delete;
-					}
-				}
-			}
 
-			fdb_delete(br, f);
-		skip_delete: ;
+			if (f->is_local)
+				fdb_delete_local(br, p, f);
+			else
+				fdb_delete(br, f);
 		}
 	}
 	spin_unlock_bh(&br->hash_lock);
@@ -397,6 +432,7 @@ static struct net_bridge_fdb_entry *fdb_create(struct hlist_head *head,
 		fdb->vlan_id = vid;
 		fdb->is_local = 0;
 		fdb->is_static = 0;
+		fdb->added_by_user = 0;
 		fdb->updated = fdb->used = jiffies;
 		hlist_add_head_rcu(&fdb->hlist, head);
 	}
@@ -447,7 +483,7 @@ int br_fdb_insert(struct net_bridge *br, struct net_bridge_port *source,
 }
 
 void br_fdb_update(struct net_bridge *br, struct net_bridge_port *source,
-		   const unsigned char *addr, u16 vid)
+		   const unsigned char *addr, u16 vid, bool added_by_user)
 {
 	struct hlist_head *head = &br->hash[br_mac_hash(addr, vid)];
 	struct net_bridge_fdb_entry *fdb;
@@ -473,13 +509,18 @@ void br_fdb_update(struct net_bridge *br, struct net_bridge_port *source,
 			/* fastpath: update of existing entry */
 			fdb->dst = source;
 			fdb->updated = jiffies;
+			if (unlikely(added_by_user))
+				fdb->added_by_user = 1;
 		}
 	} else {
 		spin_lock(&br->hash_lock);
 		if (likely(!fdb_find(head, addr, vid))) {
 			fdb = fdb_create(head, source, addr, vid);
-			if (fdb)
+			if (fdb) {
+				if (unlikely(added_by_user))
+					fdb->added_by_user = 1;
 				fdb_notify(br, fdb, RTM_NEWNEIGH);
+			}
 		}
 		/* else  we lose race and someone else inserts
 		 * it first, don't bother updating
@@ -647,6 +688,7 @@ static int fdb_add_entry(struct net_bridge_port *source, const __u8 *addr,
 
 		modified = true;
 	}
+	fdb->added_by_user = 1;
 
 	fdb->used = jiffies;
 	if (modified) {
@@ -664,7 +706,7 @@ static int __br_fdb_add(struct ndmsg *ndm, struct net_bridge_port *p,
 
 	if (ndm->ndm_flags & NTF_USE) {
 		rcu_read_lock();
-		br_fdb_update(p->br, p, addr, vid);
+		br_fdb_update(p->br, p, addr, vid, true);
 		rcu_read_unlock();
 	} else {
 		spin_lock_bh(&p->br->hash_lock);
@@ -749,8 +791,7 @@ out:
 	return err;
 }
 
-int fdb_delete_by_addr(struct net_bridge *br, const u8 *addr,
-		       u16 vlan)
+static int fdb_delete_by_addr(struct net_bridge *br, const u8 *addr, u16 vlan)
 {
 	struct hlist_head *head = &br->hash[br_mac_hash(addr, vlan)];
 	struct net_bridge_fdb_entry *fdb;
diff --git a/net/bridge/br_if.c b/net/bridge/br_if.c
index cffe1d6..54d207d 100644
--- a/net/bridge/br_if.c
+++ b/net/bridge/br_if.c
@@ -389,6 +389,9 @@ int br_add_if(struct net_bridge *br, struct net_device *dev)
 	if (br->dev->needed_headroom < dev->needed_headroom)
 		br->dev->needed_headroom = dev->needed_headroom;
 
+	if (br_fdb_insert(br, p, dev->dev_addr, 0))
+		netdev_err(dev, "failed insert local address bridge forwarding table\n");
+
 	spin_lock_bh(&br->lock);
 	changed_addr = br_stp_recalculate_bridge_id(br);
 
@@ -404,9 +407,6 @@ int br_add_if(struct net_bridge *br, struct net_device *dev)
 
 	dev_set_mtu(br->dev, br_min_mtu(br));
 
-	if (br_fdb_insert(br, p, dev->dev_addr, 0))
-		netdev_err(dev, "failed insert local address bridge forwarding table\n");
-
 	kobject_uevent(&p->kobj, KOBJ_ADD);
 
 	return 0;
diff --git a/net/bridge/br_input.c b/net/bridge/br_input.c
index bf8dc7d..28d5446 100644
--- a/net/bridge/br_input.c
+++ b/net/bridge/br_input.c
@@ -77,7 +77,7 @@ int br_handle_frame_finish(struct sk_buff *skb)
 	/* insert into forwarding database after filtering to avoid spoofing */
 	br = p->br;
 	if (p->flags & BR_LEARNING)
-		br_fdb_update(br, p, eth_hdr(skb)->h_source, vid);
+		br_fdb_update(br, p, eth_hdr(skb)->h_source, vid, false);
 
 	if (!is_broadcast_ether_addr(dest) && is_multicast_ether_addr(dest) &&
 	    br_multicast_rcv(br, p, skb, vid))
@@ -148,7 +148,7 @@ static int br_handle_local_finish(struct sk_buff *skb)
 
 	br_vlan_get_tag(skb, &vid);
 	if (p->flags & BR_LEARNING)
-		br_fdb_update(p->br, p, eth_hdr(skb)->h_source, vid);
+		br_fdb_update(p->br, p, eth_hdr(skb)->h_source, vid, false);
 	return 0;	 /* process further */
 }
 
diff --git a/net/bridge/br_private.h b/net/bridge/br_private.h
index fcd1233..3ba11bc 100644
--- a/net/bridge/br_private.h
+++ b/net/bridge/br_private.h
@@ -104,6 +104,7 @@ struct net_bridge_fdb_entry
 	mac_addr			addr;
 	unsigned char			is_local;
 	unsigned char			is_static;
+	unsigned char			added_by_user;
 	__u16				vlan_id;
 };
 
@@ -370,6 +371,9 @@ static inline void br_netpoll_disable(struct net_bridge_port *p)
 int br_fdb_init(void);
 void br_fdb_fini(void);
 void br_fdb_flush(struct net_bridge *br);
+void br_fdb_find_delete_local(struct net_bridge *br,
+			      const struct net_bridge_port *p,
+			      const unsigned char *addr, u16 vid);
 void br_fdb_changeaddr(struct net_bridge_port *p, const unsigned char *newaddr);
 void br_fdb_change_mac_address(struct net_bridge *br, const u8 *newaddr);
 void br_fdb_cleanup(unsigned long arg);
@@ -383,8 +387,7 @@ int br_fdb_fillbuf(struct net_bridge *br, void *buf, unsigned long count,
 int br_fdb_insert(struct net_bridge *br, struct net_bridge_port *source,
 		  const unsigned char *addr, u16 vid);
 void br_fdb_update(struct net_bridge *br, struct net_bridge_port *source,
-		   const unsigned char *addr, u16 vid);
-int fdb_delete_by_addr(struct net_bridge *br, const u8 *addr, u16 vid);
+		   const unsigned char *addr, u16 vid, bool added_by_user);
 
 int br_fdb_delete(struct ndmsg *ndm, struct nlattr *tb[],
 		  struct net_device *dev, const unsigned char *addr);
@@ -584,6 +587,7 @@ struct sk_buff *br_handle_vlan(struct net_bridge *br,
 int br_vlan_add(struct net_bridge *br, u16 vid, u16 flags);
 int br_vlan_delete(struct net_bridge *br, u16 vid);
 void br_vlan_flush(struct net_bridge *br);
+bool br_vlan_find(struct net_bridge *br, u16 vid);
 int br_vlan_filter_toggle(struct net_bridge *br, unsigned long val);
 int nbp_vlan_add(struct net_bridge_port *port, u16 vid, u16 flags);
 int nbp_vlan_delete(struct net_bridge_port *port, u16 vid);
@@ -665,6 +669,11 @@ static inline void br_vlan_flush(struct net_bridge *br)
 {
 }
 
+static inline bool br_vlan_find(struct net_bridge *br, u16 vid)
+{
+	return false;
+}
+
 static inline int nbp_vlan_add(struct net_bridge_port *port, u16 vid, u16 flags)
 {
 	return -EOPNOTSUPP;
diff --git a/net/bridge/br_stp_if.c b/net/bridge/br_stp_if.c
index 656a6f3..189ba1e 100644
--- a/net/bridge/br_stp_if.c
+++ b/net/bridge/br_stp_if.c
@@ -194,6 +194,8 @@ void br_stp_change_bridge_id(struct net_bridge *br, const unsigned char *addr)
 
 	wasroot = br_is_root_bridge(br);
 
+	br_fdb_change_mac_address(br, addr);
+
 	memcpy(oldaddr, br->bridge_id.addr, ETH_ALEN);
 	memcpy(br->bridge_id.addr, addr, ETH_ALEN);
 	memcpy(br->dev->dev_addr, addr, ETH_ALEN);
diff --git a/net/bridge/br_vlan.c b/net/bridge/br_vlan.c
index 4ca4d0a..8249ca76 100644
--- a/net/bridge/br_vlan.c
+++ b/net/bridge/br_vlan.c
@@ -275,9 +275,7 @@ int br_vlan_delete(struct net_bridge *br, u16 vid)
 	if (!pv)
 		return -EINVAL;
 
-	spin_lock_bh(&br->hash_lock);
-	fdb_delete_by_addr(br, br->dev->dev_addr, vid);
-	spin_unlock_bh(&br->hash_lock);
+	br_fdb_find_delete_local(br, NULL, br->dev->dev_addr, vid);
 
 	__vlan_del(pv, vid);
 	return 0;
@@ -295,6 +293,25 @@ void br_vlan_flush(struct net_bridge *br)
 	__vlan_flush(pv);
 }
 
+bool br_vlan_find(struct net_bridge *br, u16 vid)
+{
+	struct net_port_vlans *pv;
+	bool found = false;
+
+	rcu_read_lock();
+	pv = rcu_dereference(br->vlan_info);
+
+	if (!pv)
+		goto out;
+
+	if (test_bit(vid, pv->vlan_bitmap))
+		found = true;
+
+out:
+	rcu_read_unlock();
+	return found;
+}
+
 int br_vlan_filter_toggle(struct net_bridge *br, unsigned long val)
 {
 	if (!rtnl_trylock())
@@ -359,9 +376,7 @@ int nbp_vlan_delete(struct net_bridge_port *port, u16 vid)
 	if (!pv)
 		return -EINVAL;
 
-	spin_lock_bh(&port->br->hash_lock);
-	fdb_delete_by_addr(port->br, port->dev->dev_addr, vid);
-	spin_unlock_bh(&port->br->hash_lock);
+	br_fdb_find_delete_local(port->br, port, port->dev->dev_addr, vid);
 
 	return __vlan_del(pv, vid);
 }