Linux Ethernet Bridging - Aug 2019 - [Bridge] [PATCH net-next v2 0/4] net: bridge: mdb: allow dump/add/del of host-joined entries

Hi,
This set makes the bridge dump host-joined mdb entries, they should be
treated as normal entries since they take a slot and are aging out.
We already have notifications for them but we couldn't dump them until
now so they remained hidden. We dump them similar to how they're
notified, in order to keep user-space compatibility with the dumped
objects (e.g. iproute2 dumps mdbs in a format which can be fed into
add/del commands) we allow host-joined groups also to be added/deleted via
mdb commands. That can later be used for L2 mcast MAC manipulation as
was recently discussed. Note that iproute2 changes are not necessary,
this set will work with the current user-space mdb code.

Patch 01 - a trivial comment move
Patch 02 - factors out the mdb filling code so it can be
           re-used for the host-joined entries
Patch 03 - dumps host-joined entries
Patch 04 - allows manipulation of host-joined entries via standard mdb
           calls

v2: change patch 04 to avoid double notification and improve host group
    manual removal if no ports are present in the group

Thanks,
 Nik

Nikolay Aleksandrov (4):
  net: bridge: mdb: move vlan comments
  net: bridge: mdb: factor out mdb filling
  net: bridge: mdb: dump host-joined entries as well
  net: bridge: mdb: allow add/delete for host-joined groups

 net/bridge/br_mdb.c       | 173 +++++++++++++++++++++++++-------------
 net/bridge/br_multicast.c |  30 +++++--
 net/bridge/br_private.h   |   2 +
 3 files changed, 141 insertions(+), 64 deletions(-)

-- 
2.21.0

Trivial patch to move the vlan comments in their proper places above the
vid 0 checks.

Signed-off-by: Nikolay Aleksandrov <nikolay at cumulusnetworks.com>
---
 net/bridge/br_mdb.c | 12 ++++++------
 1 file changed, 6 insertions(+), 6 deletions(-)

diff --git a/net/bridge/br_mdb.c b/net/bridge/br_mdb.c
index 428af1abf8cc..ee6208c6d946 100644
--- a/net/bridge/br_mdb.c
+++ b/net/bridge/br_mdb.c
@@ -653,9 +653,6 @@ static int br_mdb_add(struct sk_buff *skb, struct nlmsghdr
*nlh,
 
 	br = netdev_priv(dev);
 
-	/* If vlan filtering is enabled and VLAN is not specified
-	 * install mdb entry on all vlans configured on the port.
-	 */
 	pdev = __dev_get_by_index(net, entry->ifindex);
 	if (!pdev)
 		return -ENODEV;
@@ -665,6 +662,9 @@ static int br_mdb_add(struct sk_buff *skb, struct nlmsghdr
*nlh,
 		return -EINVAL;
 
 	vg = nbp_vlan_group(p);
+	/* If vlan filtering is enabled and VLAN is not specified
+	 * install mdb entry on all vlans configured on the port.
+	 */
 	if (br_vlan_enabled(br->dev) && vg && entry->vid == 0) {
 		list_for_each_entry(v, &vg->vlan_list, vlist) {
 			entry->vid = v->vid;
@@ -745,9 +745,6 @@ static int br_mdb_del(struct sk_buff *skb, struct nlmsghdr
*nlh,
 
 	br = netdev_priv(dev);
 
-	/* If vlan filtering is enabled and VLAN is not specified
-	 * delete mdb entry on all vlans configured on the port.
-	 */
 	pdev = __dev_get_by_index(net, entry->ifindex);
 	if (!pdev)
 		return -ENODEV;
@@ -757,6 +754,9 @@ static int br_mdb_del(struct sk_buff *skb, struct nlmsghdr
*nlh,
 		return -EINVAL;
 
 	vg = nbp_vlan_group(p);
+	/* If vlan filtering is enabled and VLAN is not specified
+	 * delete mdb entry on all vlans configured on the port.
+	 */
 	if (br_vlan_enabled(br->dev) && vg && entry->vid == 0) {
 		list_for_each_entry(v, &vg->vlan_list, vlist) {
 			entry->vid = v->vid;
-- 
2.21.0

We have to factor out the mdb fill portion in order to re-use it later for
the bridge mdb entries. No functional changes intended.

Signed-off-by: Nikolay Aleksandrov <nikolay at cumulusnetworks.com>
---
 net/bridge/br_mdb.c | 68 ++++++++++++++++++++++++---------------------
 1 file changed, 37 insertions(+), 31 deletions(-)

diff --git a/net/bridge/br_mdb.c b/net/bridge/br_mdb.c
index ee6208c6d946..77730983097e 100644
--- a/net/bridge/br_mdb.c
+++ b/net/bridge/br_mdb.c
@@ -77,6 +77,40 @@ static void __mdb_entry_to_br_ip(struct br_mdb_entry *entry,
struct br_ip *ip)
 #endif
 }
 
+static int __mdb_fill_info(struct sk_buff *skb,
+			   struct net_bridge_port_group *p)
+{
+	struct nlattr *nest_ent;
+	struct br_mdb_entry e;
+
+	memset(&e, 0, sizeof(e));
+	__mdb_entry_fill_flags(&e, p->flags);
+	e.ifindex = p->port->dev->ifindex;
+	e.vid = p->addr.vid;
+	if (p->addr.proto == htons(ETH_P_IP))
+		e.addr.u.ip4 = p->addr.u.ip4;
+#if IS_ENABLED(CONFIG_IPV6)
+	if (p->addr.proto == htons(ETH_P_IPV6))
+		e.addr.u.ip6 = p->addr.u.ip6;
+#endif
+	e.addr.proto = p->addr.proto;
+	nest_ent = nla_nest_start_noflag(skb,
+					 MDBA_MDB_ENTRY_INFO);
+	if (!nest_ent)
+		return -EMSGSIZE;
+
+	if (nla_put_nohdr(skb, sizeof(e), &e) ||
+	    nla_put_u32(skb,
+			MDBA_MDB_EATTR_TIMER,
+			br_timer_value(&p->timer))) {
+		nla_nest_cancel(skb, nest_ent);
+		return -EMSGSIZE;
+	}
+	nla_nest_end(skb, nest_ent);
+
+	return 0;
+}
+
 static int br_mdb_fill_info(struct sk_buff *skb, struct netlink_callback *cb,
 			    struct net_device *dev)
 {
@@ -95,7 +129,6 @@ static int br_mdb_fill_info(struct sk_buff *skb, struct
netlink_callback *cb,
 	hlist_for_each_entry_rcu(mp, &br->mdb_list, mdb_node) {
 		struct net_bridge_port_group *p;
 		struct net_bridge_port_group __rcu **pp;
-		struct net_bridge_port *port;
 
 		if (idx < s_idx)
 			goto skip;
@@ -108,41 +141,14 @@ static int br_mdb_fill_info(struct sk_buff *skb, struct
netlink_callback *cb,
 
 		for (pp = &mp->ports; (p = rcu_dereference(*pp)) != NULL;
 		      pp = &p->next) {
-			struct nlattr *nest_ent;
-			struct br_mdb_entry e;
-
-			port = p->port;
-			if (!port)
+			if (!p->port)
 				continue;
 
-			memset(&e, 0, sizeof(e));
-			e.ifindex = port->dev->ifindex;
-			e.vid = p->addr.vid;
-			__mdb_entry_fill_flags(&e, p->flags);
-			if (p->addr.proto == htons(ETH_P_IP))
-				e.addr.u.ip4 = p->addr.u.ip4;
-#if IS_ENABLED(CONFIG_IPV6)
-			if (p->addr.proto == htons(ETH_P_IPV6))
-				e.addr.u.ip6 = p->addr.u.ip6;
-#endif
-			e.addr.proto = p->addr.proto;
-			nest_ent = nla_nest_start_noflag(skb,
-							 MDBA_MDB_ENTRY_INFO);
-			if (!nest_ent) {
-				nla_nest_cancel(skb, nest2);
-				err = -EMSGSIZE;
-				goto out;
-			}
-			if (nla_put_nohdr(skb, sizeof(e), &e) ||
-			    nla_put_u32(skb,
-					MDBA_MDB_EATTR_TIMER,
-					br_timer_value(&p->timer))) {
-				nla_nest_cancel(skb, nest_ent);
+			err = __mdb_fill_info(skb, p);
+			if (err) {
 				nla_nest_cancel(skb, nest2);
-				err = -EMSGSIZE;
 				goto out;
 			}
-			nla_nest_end(skb, nest_ent);
 		}
 		nla_nest_end(skb, nest2);
 skip:
-- 
2.21.0

Currently we dump only the port mdb entries but we can have host-joined
entries on the bridge itself and they should be treated as normal temp
mdbs, they're already notified:
$ bridge monitor all
[MDB]dev br0 port br0 grp ff02::8 temp

The group will not be shown in the bridge mdb output, but it takes 1 slot
and it's timing out. If it's only host-joined then the mdb show output
can even be empty.

After this patch we show the host-joined groups:
$ bridge mdb show
dev br0 port br0 grp ff02::8 temp

Signed-off-by: Nikolay Aleksandrov <nikolay at cumulusnetworks.com>
---
 net/bridge/br_mdb.c | 41 +++++++++++++++++++++++++++++++----------
 1 file changed, 31 insertions(+), 10 deletions(-)

diff --git a/net/bridge/br_mdb.c b/net/bridge/br_mdb.c
index 77730983097e..985273425117 100644
--- a/net/bridge/br_mdb.c
+++ b/net/bridge/br_mdb.c
@@ -78,22 +78,35 @@ static void __mdb_entry_to_br_ip(struct br_mdb_entry *entry,
struct br_ip *ip)
 }
 
 static int __mdb_fill_info(struct sk_buff *skb,
+			   struct net_bridge_mdb_entry *mp,
 			   struct net_bridge_port_group *p)
 {
+	struct timer_list *mtimer;
 	struct nlattr *nest_ent;
 	struct br_mdb_entry e;
+	u8 flags = 0;
+	int ifindex;
 
 	memset(&e, 0, sizeof(e));
-	__mdb_entry_fill_flags(&e, p->flags);
-	e.ifindex = p->port->dev->ifindex;
-	e.vid = p->addr.vid;
-	if (p->addr.proto == htons(ETH_P_IP))
-		e.addr.u.ip4 = p->addr.u.ip4;
+	if (p) {
+		ifindex = p->port->dev->ifindex;
+		mtimer = &p->timer;
+		flags = p->flags;
+	} else {
+		ifindex = mp->br->dev->ifindex;
+		mtimer = &mp->timer;
+	}
+
+	__mdb_entry_fill_flags(&e, flags);
+	e.ifindex = ifindex;
+	e.vid = mp->addr.vid;
+	if (mp->addr.proto == htons(ETH_P_IP))
+		e.addr.u.ip4 = mp->addr.u.ip4;
 #if IS_ENABLED(CONFIG_IPV6)
-	if (p->addr.proto == htons(ETH_P_IPV6))
-		e.addr.u.ip6 = p->addr.u.ip6;
+	if (mp->addr.proto == htons(ETH_P_IPV6))
+		e.addr.u.ip6 = mp->addr.u.ip6;
 #endif
-	e.addr.proto = p->addr.proto;
+	e.addr.proto = mp->addr.proto;
 	nest_ent = nla_nest_start_noflag(skb,
 					 MDBA_MDB_ENTRY_INFO);
 	if (!nest_ent)
@@ -102,7 +115,7 @@ static int __mdb_fill_info(struct sk_buff *skb,
 	if (nla_put_nohdr(skb, sizeof(e), &e) ||
 	    nla_put_u32(skb,
 			MDBA_MDB_EATTR_TIMER,
-			br_timer_value(&p->timer))) {
+			br_timer_value(mtimer))) {
 		nla_nest_cancel(skb, nest_ent);
 		return -EMSGSIZE;
 	}
@@ -139,12 +152,20 @@ static int br_mdb_fill_info(struct sk_buff *skb, struct
netlink_callback *cb,
 			break;
 		}
 
+		if (mp->host_joined) {
+			err = __mdb_fill_info(skb, mp, NULL);
+			if (err) {
+				nla_nest_cancel(skb, nest2);
+				break;
+			}
+		}
+
 		for (pp = &mp->ports; (p = rcu_dereference(*pp)) != NULL;
 		      pp = &p->next) {
 			if (!p->port)
 				continue;
 
-			err = __mdb_fill_info(skb, p);
+			err = __mdb_fill_info(skb, mp, p);
 			if (err) {
 				nla_nest_cancel(skb, nest2);
 				goto out;
-- 
2.21.0

Currently this is needed only for user-space compatibility, so similar
object adds/deletes as the dumped ones would succeed. Later it can be
used for L2 mcast MAC add/delete.

v2: don't send a notification when used from user-space, arm the group
    timer if no ports are left after host entry del

Signed-off-by: Nikolay Aleksandrov <nikolay at cumulusnetworks.com>
---
 net/bridge/br_mdb.c       | 76 +++++++++++++++++++++++++++------------
 net/bridge/br_multicast.c | 30 ++++++++++++----
 net/bridge/br_private.h   |  2 ++
 3 files changed, 79 insertions(+), 29 deletions(-)

diff --git a/net/bridge/br_mdb.c b/net/bridge/br_mdb.c
index 985273425117..e0f789296920 100644
--- a/net/bridge/br_mdb.c
+++ b/net/bridge/br_mdb.c
@@ -616,6 +616,19 @@ static int br_mdb_add_group(struct net_bridge *br, struct
net_bridge_port *port,
 			return err;
 	}
 
+	/* host join */
+	if (!port) {
+		/* don't allow any flags for host-joined groups */
+		if (state)
+			return -EINVAL;
+		if (mp->host_joined)
+			return -EEXIST;
+
+		br_multicast_host_join(mp, false);
+
+		return 0;
+	}
+
 	for (pp = &mp->ports;
 	     (p = mlock_dereference(*pp, br)) != NULL;
 	     pp = &p->next) {
@@ -640,19 +653,21 @@ static int __br_mdb_add(struct net *net, struct net_bridge
*br,
 {
 	struct br_ip ip;
 	struct net_device *dev;
-	struct net_bridge_port *p;
+	struct net_bridge_port *p = NULL;
 	int ret;
 
 	if (!netif_running(br->dev) || !br_opt_get(br, BROPT_MULTICAST_ENABLED))
 		return -EINVAL;
 
-	dev = __dev_get_by_index(net, entry->ifindex);
-	if (!dev)
-		return -ENODEV;
+	if (entry->ifindex != br->dev->ifindex) {
+		dev = __dev_get_by_index(net, entry->ifindex);
+		if (!dev)
+			return -ENODEV;
 
-	p = br_port_get_rtnl(dev);
-	if (!p || p->br != br || p->state == BR_STATE_DISABLED)
-		return -EINVAL;
+		p = br_port_get_rtnl(dev);
+		if (!p || p->br != br || p->state == BR_STATE_DISABLED)
+			return -EINVAL;
+	}
 
 	__mdb_entry_to_br_ip(entry, &ip);
 
@@ -680,15 +695,19 @@ static int br_mdb_add(struct sk_buff *skb, struct nlmsghdr
*nlh,
 
 	br = netdev_priv(dev);
 
-	pdev = __dev_get_by_index(net, entry->ifindex);
-	if (!pdev)
-		return -ENODEV;
+	if (entry->ifindex != br->dev->ifindex) {
+		pdev = __dev_get_by_index(net, entry->ifindex);
+		if (!pdev)
+			return -ENODEV;
 
-	p = br_port_get_rtnl(pdev);
-	if (!p || p->br != br || p->state == BR_STATE_DISABLED)
-		return -EINVAL;
+		p = br_port_get_rtnl(pdev);
+		if (!p || p->br != br || p->state == BR_STATE_DISABLED)
+			return -EINVAL;
+		vg = nbp_vlan_group(p);
+	} else {
+		vg = br_vlan_group(br);
+	}
 
-	vg = nbp_vlan_group(p);
 	/* If vlan filtering is enabled and VLAN is not specified
 	 * install mdb entry on all vlans configured on the port.
 	 */
@@ -727,6 +746,15 @@ static int __br_mdb_del(struct net_bridge *br, struct
br_mdb_entry *entry)
 	if (!mp)
 		goto unlock;
 
+	/* host leave */
+	if (entry->ifindex == mp->br->dev->ifindex &&
mp->host_joined) {
+		br_multicast_host_leave(mp, false);
+		err = 0;
+		if (!mp->ports && netif_running(br->dev))
+			mod_timer(&mp->timer, jiffies);
+		goto unlock;
+	}
+
 	for (pp = &mp->ports;
 	     (p = mlock_dereference(*pp, br)) != NULL;
 	     pp = &p->next) {
@@ -759,9 +787,9 @@ static int br_mdb_del(struct sk_buff *skb, struct nlmsghdr
*nlh,
 {
 	struct net *net = sock_net(skb->sk);
 	struct net_bridge_vlan_group *vg;
+	struct net_bridge_port *p = NULL;
 	struct net_device *dev, *pdev;
 	struct br_mdb_entry *entry;
-	struct net_bridge_port *p;
 	struct net_bridge_vlan *v;
 	struct net_bridge *br;
 	int err;
@@ -772,15 +800,19 @@ static int br_mdb_del(struct sk_buff *skb, struct nlmsghdr
*nlh,
 
 	br = netdev_priv(dev);
 
-	pdev = __dev_get_by_index(net, entry->ifindex);
-	if (!pdev)
-		return -ENODEV;
+	if (entry->ifindex != br->dev->ifindex) {
+		pdev = __dev_get_by_index(net, entry->ifindex);
+		if (!pdev)
+			return -ENODEV;
 
-	p = br_port_get_rtnl(pdev);
-	if (!p || p->br != br || p->state == BR_STATE_DISABLED)
-		return -EINVAL;
+		p = br_port_get_rtnl(pdev);
+		if (!p || p->br != br || p->state == BR_STATE_DISABLED)
+			return -EINVAL;
+		vg = nbp_vlan_group(p);
+	} else {
+		vg = br_vlan_group(br);
+	}
 
-	vg = nbp_vlan_group(p);
 	/* If vlan filtering is enabled and VLAN is not specified
 	 * delete mdb entry on all vlans configured on the port.
 	 */
diff --git a/net/bridge/br_multicast.c b/net/bridge/br_multicast.c
index 9b379e110129..ad12fe3fca8c 100644
--- a/net/bridge/br_multicast.c
+++ b/net/bridge/br_multicast.c
@@ -148,8 +148,7 @@ static void br_multicast_group_expired(struct timer_list *t)
 	if (!netif_running(br->dev) || timer_pending(&mp->timer))
 		goto out;
 
-	mp->host_joined = false;
-	br_mdb_notify(br->dev, NULL, &mp->addr, RTM_DELMDB, 0);
+	br_multicast_host_leave(mp, true);
 
 	if (mp->ports)
 		goto out;
@@ -512,6 +511,27 @@ static bool br_port_group_equal(struct
net_bridge_port_group *p,
 	return ether_addr_equal(src, p->eth_addr);
 }
 
+void br_multicast_host_join(struct net_bridge_mdb_entry *mp, bool notify)
+{
+	if (!mp->host_joined) {
+		mp->host_joined = true;
+		if (notify)
+			br_mdb_notify(mp->br->dev, NULL, &mp->addr,
+				      RTM_NEWMDB, 0);
+	}
+	mod_timer(&mp->timer, jiffies +
mp->br->multicast_membership_interval);
+}
+
+void br_multicast_host_leave(struct net_bridge_mdb_entry *mp, bool notify)
+{
+	if (!mp->host_joined)
+		return;
+
+	mp->host_joined = false;
+	if (notify)
+		br_mdb_notify(mp->br->dev, NULL, &mp->addr, RTM_DELMDB, 0);
+}
+
 static int br_multicast_add_group(struct net_bridge *br,
 				  struct net_bridge_port *port,
 				  struct br_ip *group,
@@ -534,11 +554,7 @@ static int br_multicast_add_group(struct net_bridge *br,
 		goto err;
 
 	if (!port) {
-		if (!mp->host_joined) {
-			mp->host_joined = true;
-			br_mdb_notify(br->dev, NULL, &mp->addr, RTM_NEWMDB, 0);
-		}
-		mod_timer(&mp->timer, now + br->multicast_membership_interval);
+		br_multicast_host_join(mp, true);
 		goto out;
 	}
 
diff --git a/net/bridge/br_private.h b/net/bridge/br_private.h
index b7a4942ff1b3..ce2ab14ee605 100644
--- a/net/bridge/br_private.h
+++ b/net/bridge/br_private.h
@@ -702,6 +702,8 @@ void br_multicast_get_stats(const struct net_bridge *br,
 			    struct br_mcast_stats *dest);
 void br_mdb_init(void);
 void br_mdb_uninit(void);
+void br_multicast_host_join(struct net_bridge_mdb_entry *mp, bool notify);
+void br_multicast_host_leave(struct net_bridge_mdb_entry *mp, bool notify);
 
 #define mlock_dereference(X, br) \
 	rcu_dereference_protected(X, lockdep_is_held(&br->multicast_lock))
-- 
2.21.0

From: Nikolay Aleksandrov <nikolay at cumulusnetworks.com>
Date: Wed, 14 Aug 2019 20:04:57 +0300
> This set makes the bridge dump host-joined mdb entries, they should be
> treated as normal entries since they take a slot and are aging out.
 ...

Please respin with this warning fixed:

net/bridge/br_mdb.c: In function ?br_mdb_add?:
net/bridge/br_mdb.c:725:4: warning: ?p? may be used uninitialized in this
function [-Wmaybe-uninitialized]
    __br_mdb_notify(dev, p, entry, RTM_NEWMDB);
    ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

[davem at localhost net-next]$ gcc --version
gcc (GCC) 8.3.1 20190223 (Red Hat 8.3.1-2)
Copyright (C) 2018 Free Software Foundation, Inc.
This is free software; see the source for copying conditions.  There is NO
warranty; not even for MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.