Merge branch 'ipvlan-private-vepa'

Mahesh Bandewar says:

====================
add 'private' and 'vepa' attributes to ipvlan modes

IPvlan has always been operating in bridge-mode for its supported modes i.e.
if the packets are destined to the adjacent neighbor dev, then IPvlan driver
will switch the packet internally without needing the packets to hit the
wire or get routed. However, there are situations where this bridge-mode is
not needed. e.g. two private processes running inside two namespaces which
are having one IPvlan slave each for its namespace but sharing the master. These
processes should reach the outside world through the master device but at
the same time the bridge function should not work. Currently that's not
possible hence the private attribute for the selected mode comes in play.

VEPA or 802.1Qbg on the other hand has limited appeal with IPvlan since IPvlan
uses the mac-address of the lower device. So packets that are destined to
the adjacent neighbor slave-dev will have same src and dest mac. When these
packets reach the external switch/router, they will send you the redirect
message which the host will have to deal with. Having said that this attribute
will have appeal in debugging as IPvlan will not switch / short-circuit
packets internally. e.g. using VEPA mode with lower-device in loopback mode
will avoid some complicated set-ups that use non-local-bind with some route
jugglery.

This patch-set implements these attributes for the existing modes that
IPvlan has. Please see individual patches for their detailed implementation.
A subsequent ip-utils patch is needed and will be sent soon.
====================

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

5 files changed, 136 insertions, 12 deletions

diff --git a/Documentation/networking/ipvlan.txt b/Documentation/networking/ipvlan.txt
index 1fe42a874aae..812ef003e0a8 100644
--- a/Documentation/networking/ipvlan.txt
+++ b/Documentation/networking/ipvlan.txt
@@ -22,9 +22,21 @@ The driver can be built into the kernel (CONFIG_IPVLAN=y) or as a module
 	There are no module parameters for this driver and it can be configured
 using IProute2/ip utility.
 
-	ip link add link <master-dev> name <slave-dev> type ipvlan mode { l2 | l3 | l3s }
-
-	e.g. ip link add link eth0 name ipvl0 type ipvlan mode l2
+    ip link add link <master> name <slave> type ipvlan [ mode MODE ] [ FLAGS ]
+       where
+         MODE: l3 (default) | l3s | l2
+         FLAGS: bridge (default) | private | vepa
+
+    e.g.
+    (a) Following will create IPvlan link with eth0 as master in
+        L3 bridge mode
+          bash# ip link add link eth0 name ipvl0 type ipvlan
+    (b) This command will create IPvlan link in L2 bridge mode.
+          bash# ip link add link eth0 name ipvl0 type ipvlan mode l2 bridge
+    (c) This command will create an IPvlan device in L2 private mode.
+          bash# ip link add link eth0 name ipvlan type ipvlan mode l2 private
+    (d) This command will create an IPvlan device in L2 vepa mode.
+          bash# ip link add link eth0 name ipvlan type ipvlan mode l2 vepa
 
 
 4. Operating modes:
@@ -54,7 +66,29 @@ works in this mode and hence it is L3-symmetric (L3s). This will have slightly l
 performance but that shouldn't matter since you are choosing this mode over plain-L3
 mode to make conn-tracking work.
 
-5. What to choose (macvlan vs. ipvlan)?
+5. Mode flags:
+	At this time following mode flags are available
+
+5.1 bridge:
+	This is the default option. To configure the IPvlan port in this mode,
+user can choose to either add this option on the command-line or don't specify
+anything. This is the traditional mode where slaves can cross-talk among
+themseleves apart from talking through the master device.
+
+5.2 private:
+	If this option is added to the command-line, the port is set in private
+mode. i.e. port wont allow cross communication between slaves.
+
+5.3 vepa:
+	If this is added to the command-line, the port is set in VEPA mode.
+i.e. port will offload switching functionality to the external entity as
+described in 802.1Qbg
+Note: VEPA mode in IPvlan has limitations. IPvlan uses the mac-address of the
+master-device, so the packets which are emitted in this mode for the adjacent
+neighbor will have source and destination mac same. This will make the switch /
+router send the redirect message.
+
+6. What to choose (macvlan vs. ipvlan)?
 	These two devices are very similar in many regards and the specific use
 case could very well define which device to choose. if one of the following
 situations defines your use case then you can choose to use ipvlan -
diff --git a/drivers/net/ipvlan/ipvlan.h b/drivers/net/ipvlan/ipvlan.h
index ba8173a0b62e..5166575a164d 100644
--- a/drivers/net/ipvlan/ipvlan.h
+++ b/drivers/net/ipvlan/ipvlan.h
@@ -96,6 +96,7 @@ struct ipvl_port {
 	struct hlist_head	hlhead[IPVLAN_HASH_SIZE];
 	struct list_head	ipvlans;
 	u16			mode;
+	u16			flags;
 	u16			dev_id_start;
 	struct work_struct	wq;
 	struct sk_buff_head	backlog;
@@ -123,6 +124,36 @@ static inline struct ipvl_port *ipvlan_port_get_rtnl(const struct net_device *d)
 	return rtnl_dereference(d->rx_handler_data);
 }
 
+static inline bool ipvlan_is_private(const struct ipvl_port *port)
+{
+	return !!(port->flags & IPVLAN_F_PRIVATE);
+}
+
+static inline void ipvlan_mark_private(struct ipvl_port *port)
+{
+	port->flags |= IPVLAN_F_PRIVATE;
+}
+
+static inline void ipvlan_clear_private(struct ipvl_port *port)
+{
+	port->flags &= ~IPVLAN_F_PRIVATE;
+}
+
+static inline bool ipvlan_is_vepa(const struct ipvl_port *port)
+{
+	return !!(port->flags & IPVLAN_F_VEPA);
+}
+
+static inline void ipvlan_mark_vepa(struct ipvl_port *port)
+{
+	port->flags |= IPVLAN_F_VEPA;
+}
+
+static inline void ipvlan_clear_vepa(struct ipvl_port *port)
+{
+	port->flags &= ~IPVLAN_F_VEPA;
+}
+
 void ipvlan_init_secret(void);
 unsigned int ipvlan_mac_hash(const unsigned char *addr);
 rx_handler_result_t ipvlan_handle_frame(struct sk_buff **pskb);
diff --git a/drivers/net/ipvlan/ipvlan_core.c b/drivers/net/ipvlan/ipvlan_core.c
index 1f3295e274d0..034ae4c57196 100644
--- a/drivers/net/ipvlan/ipvlan_core.c
+++ b/drivers/net/ipvlan/ipvlan_core.c
@@ -514,10 +514,16 @@ static int ipvlan_xmit_mode_l3(struct sk_buff *skb, struct net_device *dev)
 	if (!lyr3h)
 		goto out;
 
-	addr = ipvlan_addr_lookup(ipvlan->port, lyr3h, addr_type, true);
-	if (addr)
-		return ipvlan_rcv_frame(addr, &skb, true);
-
+	if (!ipvlan_is_vepa(ipvlan->port)) {
+		addr = ipvlan_addr_lookup(ipvlan->port, lyr3h, addr_type, true);
+		if (addr) {
+			if (ipvlan_is_private(ipvlan->port)) {
+				consume_skb(skb);
+				return NET_XMIT_DROP;
+			}
+			return ipvlan_rcv_frame(addr, &skb, true);
+		}
+	}
 out:
 	ipvlan_skb_crossing_ns(skb, ipvlan->phy_dev);
 	return ipvlan_process_outbound(skb);
@@ -531,12 +537,18 @@ static int ipvlan_xmit_mode_l2(struct sk_buff *skb, struct net_device *dev)
 	void *lyr3h;
 	int addr_type;
 
-	if (ether_addr_equal(eth->h_dest, eth->h_source)) {
+	if (!ipvlan_is_vepa(ipvlan->port) &&
+	    ether_addr_equal(eth->h_dest, eth->h_source)) {
 		lyr3h = ipvlan_get_L3_hdr(skb, &addr_type);
 		if (lyr3h) {
 			addr = ipvlan_addr_lookup(ipvlan->port, lyr3h, addr_type, true);
-			if (addr)
+			if (addr) {
+				if (ipvlan_is_private(ipvlan->port)) {
+					consume_skb(skb);
+					return NET_XMIT_DROP;
+				}
 				return ipvlan_rcv_frame(addr, &skb, true);
+			}
 		}
 		skb = skb_share_check(skb, GFP_ATOMIC);
 		if (!skb)
diff --git a/drivers/net/ipvlan/ipvlan_main.c b/drivers/net/ipvlan/ipvlan_main.c
index f0ab55df57f1..a266aa435d4d 100644
--- a/drivers/net/ipvlan/ipvlan_main.c
+++ b/drivers/net/ipvlan/ipvlan_main.c
@@ -462,11 +462,29 @@ static int ipvlan_nl_changelink(struct net_device *dev,
 	struct ipvl_port *port = ipvlan_port_get_rtnl(ipvlan->phy_dev);
 	int err = 0;
 
-	if (data && data[IFLA_IPVLAN_MODE]) {
+	if (!data)
+		return 0;
+
+	if (data[IFLA_IPVLAN_MODE]) {
 		u16 nmode = nla_get_u16(data[IFLA_IPVLAN_MODE]);
 
 		err = ipvlan_set_port_mode(port, nmode);
 	}
+
+	if (!err && data[IFLA_IPVLAN_FLAGS]) {
+		u16 flags = nla_get_u16(data[IFLA_IPVLAN_FLAGS]);
+
+		if (flags & IPVLAN_F_PRIVATE)
+			ipvlan_mark_private(port);
+		else
+			ipvlan_clear_private(port);
+
+		if (flags & IPVLAN_F_VEPA)
+			ipvlan_mark_vepa(port);
+		else
+			ipvlan_clear_vepa(port);
+	}
+
 	return err;
 }
 
@@ -474,18 +492,34 @@ static size_t ipvlan_nl_getsize(const struct net_device *dev)
 {
 	return (0
 		+ nla_total_size(2) /* IFLA_IPVLAN_MODE */
+		+ nla_total_size(2) /* IFLA_IPVLAN_FLAGS */
 		);
 }
 
 static int ipvlan_nl_validate(struct nlattr *tb[], struct nlattr *data[],
 			      struct netlink_ext_ack *extack)
 {
-	if (data && data[IFLA_IPVLAN_MODE]) {
+	if (!data)
+		return 0;
+
+	if (data[IFLA_IPVLAN_MODE]) {
 		u16 mode = nla_get_u16(data[IFLA_IPVLAN_MODE]);
 
 		if (mode < IPVLAN_MODE_L2 || mode >= IPVLAN_MODE_MAX)
 			return -EINVAL;
 	}
+	if (data[IFLA_IPVLAN_FLAGS]) {
+		u16 flags = nla_get_u16(data[IFLA_IPVLAN_FLAGS]);
+
+		/* Only two bits are used at this moment. */
+		if (flags & ~(IPVLAN_F_PRIVATE | IPVLAN_F_VEPA))
+			return -EINVAL;
+		/* Also both flags can't be active at the same time. */
+		if ((flags & (IPVLAN_F_PRIVATE | IPVLAN_F_VEPA)) ==
+		    (IPVLAN_F_PRIVATE | IPVLAN_F_VEPA))
+			return -EINVAL;
+	}
+
 	return 0;
 }
 
@@ -502,6 +536,8 @@ static int ipvlan_nl_fillinfo(struct sk_buff *skb,
 	ret = -EMSGSIZE;
 	if (nla_put_u16(skb, IFLA_IPVLAN_MODE, port->mode))
 		goto err;
+	if (nla_put_u16(skb, IFLA_IPVLAN_FLAGS, port->flags))
+		goto err;
 
 	return 0;
 
@@ -549,6 +585,12 @@ int ipvlan_link_new(struct net *src_net, struct net_device *dev,
 	ipvlan_adjust_mtu(ipvlan, phy_dev);
 	INIT_LIST_HEAD(&ipvlan->addrs);
 
+	/* Flags are per port and latest update overrides. User has
+	 * to be consistent in setting it just like the mode attribute.
+	 */
+	if (data && data[IFLA_IPVLAN_FLAGS])
+		ipvlan->port->flags = nla_get_u16(data[IFLA_IPVLAN_FLAGS]);
+
 	/* If the port-id base is at the MAX value, then wrap it around and
 	 * begin from 0x1 again. This may be due to a busy system where lots
 	 * of slaves are getting created and deleted.
@@ -644,6 +686,7 @@ EXPORT_SYMBOL_GPL(ipvlan_link_setup);
 static const struct nla_policy ipvlan_nl_policy[IFLA_IPVLAN_MAX + 1] =
 {
 	[IFLA_IPVLAN_MODE] = { .type = NLA_U16 },
+	[IFLA_IPVLAN_FLAGS] = { .type = NLA_U16 },
 };
 
 static struct rtnl_link_ops ipvlan_link_ops = {
diff --git a/include/uapi/linux/if_link.h b/include/uapi/linux/if_link.h
index b037e0ab1975..81f26473d728 100644
--- a/include/uapi/linux/if_link.h
+++ b/include/uapi/linux/if_link.h
@@ -465,6 +465,7 @@ enum macsec_validation_type {
 enum {
 	IFLA_IPVLAN_UNSPEC,
 	IFLA_IPVLAN_MODE,
+	IFLA_IPVLAN_FLAGS,
 	__IFLA_IPVLAN_MAX
 };
 
@@ -477,6 +478,9 @@ enum ipvlan_mode {
 	IPVLAN_MODE_MAX
 };
 
+#define IPVLAN_F_PRIVATE	0x01
+#define IPVLAN_F_VEPA		0x02
+
 /* VXLAN section */
 enum {
 	IFLA_VXLAN_UNSPEC,