net 00/06: add GARP/GVRP support

net 00/06: add GARP/GVRP support

[Patrick McHardy]

These patches add an implementation of the GARP (Generic Attribute
Registration Protocol) applicant-only participant, which can be used
to dynamically register VLANs and multicast addresses with switches.
Applicant-only means it can only be used as a client, not on a bridge.
Also included is a GVRP (GARP VLAN Registration Protocol) application
on top.

I asked on opinions whether this should be merged a couple of weeks
ago, and Pekka Savola informed me that GARP was considered obsolete
by the IEEE. However, especially GVRP is supported by a huge number
of switches, even very cheap ones, so I think this is still useful
to support.

GARP uses the LLC_SAP_BSPAN LLC, which is also used by STP. Different
multicast addresses are used to distinguish the protocols. Since the
LLC code doesn't allow multiple registrations for the same SAP, a
small mac based demux layer is added that is used by both bridging
and GARP.

GVRP is not enabled by default yet. Since the applicant-only participant
should not be used on a bridge, we first need a reliable way to get
notified when a VLAN device becomes a bridge member. So currently,
the user needs to manually enable it using iproute. I plan to change
this soon.


 include/linux/if_vlan.h   |    1 +
 include/linux/netdevice.h |    2 +
 include/net/garp.h        |  128 +++++++++
 include/net/stp.h         |   14 +
 net/802/Kconfig           |    7 +
 net/802/Makefile          |    2 +
 net/802/garp.c            |  633 +++++++++++++++++++++++++++++++++++++++++++++
 net/802/stp.c             |  102 ++++++++
 net/8021q/Kconfig         |   10 +
 net/8021q/Makefile        |    9 +-
 net/8021q/vlan.c          |   33 ++-
 net/8021q/vlan.h          |   19 ++-
 net/8021q/vlan_dev.c      |   32 ++-
 net/8021q/vlan_gvrp.c     |   66 +++++
 net/8021q/vlan_netlink.c  |    7 +-
 net/Kconfig               |    1 +
 net/bridge/Kconfig        |    1 +
 net/bridge/br.c           |   16 +-
 net/bridge/br_private.h   |    5 +-
 net/bridge/br_stp_bpdu.c  |   12 +-
 20 files changed, 1051 insertions(+), 49 deletions(-)
 create mode 100644 include/net/garp.h
 create mode 100644 include/net/stp.h
 create mode 100644 net/802/Kconfig
 create mode 100644 net/802/garp.c
 create mode 100644 net/802/stp.c
 create mode 100644 net/8021q/vlan_gvrp.c

Patrick McHardy (6):
      net: Add STP demux layer
      bridge: Use STP demux
      net: Add GARP applicant-only participant
      vlan: Change vlan_dev_set_vlan_flag() to handle multiple flags at once
      vlan: Move device unregistration before lower dev cleanup
      vlan: Add GVRP support

net 01/06: Add STP demux layer

[Patrick McHardy]

net: Add STP demux layer

Add small STP demux layer for demuxing STP PDUs based on MAC address.
This is needed to run both GARP and STP in parallel (or even load the
modules) since both use LLC_SAP_BSPAN.

Signed-off-by: Patrick McHardy <kaber@trash.net>

---
commit 6c0be0e52a154f1991c221a6f60c6b9be71a10c8
tree 6dcdf3d929edeae8463c6957e688f1e91ca6cfe5
parent 44d28ab19c64d095314ac66f765d0c747519f4ed
author Patrick McHardy <kaber@trash.net> Thu, 03 Jul 2008 18:05:56 +0200
committer Patrick McHardy <kaber@trash.net> Thu, 03 Jul 2008 18:05:56 +0200

 include/net/stp.h |   14 +++++++
 net/802/Kconfig   |    3 ++
 net/802/Makefile  |    1 +
 net/802/stp.c     |  102 +++++++++++++++++++++++++++++++++++++++++++++++++++++
 net/Kconfig       |    1 +
 5 files changed, 121 insertions(+), 0 deletions(-)

diff --git a/include/net/stp.h b/include/net/stp.h
new file mode 100644
index 0000000..ad447f1
--- /dev/null
+++ b/include/net/stp.h
@@ -0,0 +1,14 @@
+#ifndef _NET_STP_H
+#define _NET_STP_H
+
+struct stp_proto {
+	unsigned char	group_address[ETH_ALEN];
+	void		(*rcv)(const struct stp_proto *, struct sk_buff *,
+			       struct net_device *);
+	void		*data;
+};
+
+extern int stp_proto_register(const struct stp_proto *proto);
+extern void stp_proto_unregister(const struct stp_proto *proto);
+
+#endif /* _NET_STP_H */
diff --git a/net/802/Kconfig b/net/802/Kconfig
new file mode 100644
index 0000000..01cb094
--- /dev/null
+++ b/net/802/Kconfig
@@ -0,0 +1,3 @@
+config STP
+	tristate
+	select LLC
diff --git a/net/802/Makefile b/net/802/Makefile
index 68569ff..c441d89 100644
--- a/net/802/Makefile
+++ b/net/802/Makefile
@@ -10,3 +10,4 @@ obj-$(CONFIG_FDDI)	+=                 fddi.o
 obj-$(CONFIG_HIPPI)	+=                 hippi.o
 obj-$(CONFIG_IPX)	+= p8022.o psnap.o p8023.o
 obj-$(CONFIG_ATALK)	+= p8022.o psnap.o
+obj-$(CONFIG_STP)	+= stp.o
diff --git a/net/802/stp.c b/net/802/stp.c
new file mode 100644
index 0000000..0b7a244
--- /dev/null
+++ b/net/802/stp.c
@@ -0,0 +1,102 @@
+/*
+ *	STP SAP demux
+ *
+ *	Copyright (c) 2008 Patrick McHardy <kaber@trash.net>
+ *
+ *	This program is free software; you can redistribute it and/or
+ *	modify it under the terms of the GNU General Public License
+ *	version 2 as published by the Free Software Foundation.
+ */
+#include <linux/mutex.h>
+#include <linux/skbuff.h>
+#include <linux/etherdevice.h>
+#include <linux/llc.h>
+#include <net/llc.h>
+#include <net/llc_pdu.h>
+#include <net/stp.h>
+
+/* 01:80:c2:00:00:20 - 01:80:c2:00:00:2F */
+#define GARP_ADDR_MIN	0x20
+#define GARP_ADDR_MAX	0x2F
+#define GARP_ADDR_RANGE	(GARP_ADDR_MAX - GARP_ADDR_MIN)
+
+static const struct stp_proto *garp_protos[GARP_ADDR_RANGE + 1] __read_mostly;
+static const struct stp_proto *stp_proto __read_mostly;
+
+static struct llc_sap *sap __read_mostly;
+static unsigned int sap_registered;
+static DEFINE_MUTEX(stp_proto_mutex);
+
+/* Called under rcu_read_lock from LLC */
+static int stp_pdu_rcv(struct sk_buff *skb, struct net_device *dev,
+		       struct packet_type *pt, struct net_device *orig_dev)
+{
+	const struct ethhdr *eh = eth_hdr(skb);
+	const struct llc_pdu_un *pdu = llc_pdu_un_hdr(skb);
+	const struct stp_proto *proto;
+
+	if (pdu->ssap != LLC_SAP_BSPAN ||
+	    pdu->dsap != LLC_SAP_BSPAN ||
+	    pdu->ctrl_1 != LLC_PDU_TYPE_U)
+		goto err;
+
+	if (eh->h_dest[5] >= GARP_ADDR_MIN && eh->h_dest[5] <= GARP_ADDR_MAX) {
+		proto = rcu_dereference(garp_protos[eh->h_dest[5] -
+						    GARP_ADDR_MIN]);
+		if (proto &&
+		    compare_ether_addr(eh->h_dest, proto->group_address))
+			goto err;
+	} else
+		proto = rcu_dereference(stp_proto);
+
+	if (!proto)
+		goto err;
+
+	proto->rcv(proto, skb, dev);
+	return 0;
+
+err:
+	kfree_skb(skb);
+	return 0;
+}
+
+int stp_proto_register(const struct stp_proto *proto)
+{
+	int err = 0;
+
+	mutex_lock(&stp_proto_mutex);
+	if (sap_registered++ == 0) {
+		sap = llc_sap_open(LLC_SAP_BSPAN, stp_pdu_rcv);
+		if (!sap) {
+			err = -ENOMEM;
+			goto out;
+		}
+	}
+	if (is_zero_ether_addr(proto->group_address))
+		rcu_assign_pointer(stp_proto, proto);
+	else
+		rcu_assign_pointer(garp_protos[proto->group_address[5] -
+					       GARP_ADDR_MIN], proto);
+out:
+	mutex_unlock(&stp_proto_mutex);
+	return err;
+}
+EXPORT_SYMBOL_GPL(stp_proto_register);
+
+void stp_proto_unregister(const struct stp_proto *proto)
+{
+	mutex_lock(&stp_proto_mutex);
+	if (is_zero_ether_addr(proto->group_address))
+		rcu_assign_pointer(stp_proto, NULL);
+	else
+		rcu_assign_pointer(garp_protos[proto->group_address[5] -
+					       GARP_ADDR_MIN], NULL);
+	synchronize_rcu();
+
+	if (--sap_registered == 0)
+		llc_sap_put(sap);
+	mutex_unlock(&stp_proto_mutex);
+}
+EXPORT_SYMBOL_GPL(stp_proto_unregister);
+
+MODULE_LICENSE("GPL");
diff --git a/net/Kconfig b/net/Kconfig
index acbf7c6..b986687 100644
--- a/net/Kconfig
+++ b/net/Kconfig
@@ -181,6 +181,7 @@ source "net/dccp/Kconfig"
 source "net/sctp/Kconfig"
 source "net/tipc/Kconfig"
 source "net/atm/Kconfig"
+source "net/802/Kconfig"
 source "net/bridge/Kconfig"
 source "net/8021q/Kconfig"
 source "net/decnet/Kconfig"

bridge 02/06: Use STP demux

[Patrick McHardy]

bridge: Use STP demux

Use the STP demux layer for receiving STP PDUs instead of directly
registering with LLC.

Signed-off-by: Patrick McHardy <kaber@trash.net>

---
commit 85ecc2934a05203a4f6372cd8f0b132c2d1f2f97
tree 331543d774aa6897e5902576b337515cfe70eb9c
parent 6c0be0e52a154f1991c221a6f60c6b9be71a10c8
author Patrick McHardy <kaber@trash.net> Thu, 03 Jul 2008 18:06:07 +0200
committer Patrick McHardy <kaber@trash.net> Thu, 03 Jul 2008 18:06:07 +0200

 net/bridge/Kconfig       |    1 +
 net/bridge/br.c          |   16 +++++++++-------
 net/bridge/br_private.h  |    5 +++--
 net/bridge/br_stp_bpdu.c |   12 +++---------
 4 files changed, 16 insertions(+), 18 deletions(-)

diff --git a/net/bridge/Kconfig b/net/bridge/Kconfig
index 12265af..e143ca6 100644
--- a/net/bridge/Kconfig
+++ b/net/bridge/Kconfig
@@ -5,6 +5,7 @@
 config BRIDGE
 	tristate "802.1d Ethernet Bridging"
 	select LLC
+	select STP
 	---help---
 	  If you say Y here, then your Linux box will be able to act as an
 	  Ethernet bridge, which means that the different Ethernet segments it
diff --git a/net/bridge/br.c b/net/bridge/br.c
index cede010..573acdf 100644
--- a/net/bridge/br.c
+++ b/net/bridge/br.c
@@ -18,21 +18,24 @@
 #include <linux/init.h>
 #include <linux/llc.h>
 #include <net/llc.h>
+#include <net/stp.h>
 
 #include "br_private.h"
 
 int (*br_should_route_hook)(struct sk_buff *skb);
 
-static struct llc_sap *br_stp_sap;
+static const struct stp_proto br_stp_proto = {
+	.rcv	= br_stp_rcv,
+};
 
 static int __init br_init(void)
 {
 	int err;
 
-	br_stp_sap = llc_sap_open(LLC_SAP_BSPAN, br_stp_rcv);
-	if (!br_stp_sap) {
+	err = stp_proto_register(&br_stp_proto);
+	if (err < 0) {
 		printk(KERN_ERR "bridge: can't register sap for STP\n");
-		return -EADDRINUSE;
+		return err;
 	}
 
 	err = br_fdb_init();
@@ -65,13 +68,13 @@ err_out2:
 err_out1:
 	br_fdb_fini();
 err_out:
-	llc_sap_put(br_stp_sap);
+	stp_proto_unregister(&br_stp_proto);
 	return err;
 }
 
 static void __exit br_deinit(void)
 {
-	rcu_assign_pointer(br_stp_sap->rcv_func, NULL);
+	stp_proto_unregister(&br_stp_proto);
 
 	br_netlink_fini();
 	unregister_netdevice_notifier(&br_device_notifier);
@@ -82,7 +85,6 @@ static void __exit br_deinit(void)
 	synchronize_net();
 
 	br_netfilter_fini();
-	llc_sap_put(br_stp_sap);
 	br_fdb_get_hook = NULL;
 	br_fdb_put_hook = NULL;
 
diff --git a/net/bridge/br_private.h b/net/bridge/br_private.h
index 8593c9f..815ed38 100644
--- a/net/bridge/br_private.h
+++ b/net/bridge/br_private.h
@@ -226,8 +226,9 @@ extern void br_stp_set_path_cost(struct net_bridge_port *p,
 extern ssize_t br_show_bridge_id(char *buf, const struct bridge_id *id);
 
 /* br_stp_bpdu.c */
-extern int br_stp_rcv(struct sk_buff *skb, struct net_device *dev,
-		      struct packet_type *pt, struct net_device *orig_dev);
+struct stp_proto;
+extern void br_stp_rcv(const struct stp_proto *proto, struct sk_buff *skb,
+		       struct net_device *dev);
 
 /* br_stp_timer.c */
 extern void br_stp_timer_init(struct net_bridge *br);
diff --git a/net/bridge/br_stp_bpdu.c b/net/bridge/br_stp_bpdu.c
index 9dc2de6..9964761 100644
--- a/net/bridge/br_stp_bpdu.c
+++ b/net/bridge/br_stp_bpdu.c
@@ -18,6 +18,7 @@
 #include <net/net_namespace.h>
 #include <net/llc.h>
 #include <net/llc_pdu.h>
+#include <net/stp.h>
 #include <asm/unaligned.h>
 
 #include "br_private.h"
@@ -131,10 +132,9 @@ void br_send_tcn_bpdu(struct net_bridge_port *p)
  *
  * NO locks, but rcu_read_lock (preempt_disabled)
  */
-int br_stp_rcv(struct sk_buff *skb, struct net_device *dev,
-	       struct packet_type *pt, struct net_device *orig_dev)
+void br_stp_rcv(const struct stp_proto *proto, struct sk_buff *skb,
+		struct net_device *dev)
 {
-	const struct llc_pdu_un *pdu = llc_pdu_un_hdr(skb);
 	const unsigned char *dest = eth_hdr(skb)->h_dest;
 	struct net_bridge_port *p = rcu_dereference(dev->br_port);
 	struct net_bridge *br;
@@ -146,11 +146,6 @@ int br_stp_rcv(struct sk_buff *skb, struct net_device *dev,
 	if (!p)
 		goto err;
 
-	if (pdu->ssap != LLC_SAP_BSPAN
-	    || pdu->dsap != LLC_SAP_BSPAN
-	    || pdu->ctrl_1 != LLC_PDU_TYPE_U)
-		goto err;
-
 	if (!pskb_may_pull(skb, 4))
 		goto err;
 
@@ -224,5 +219,4 @@ int br_stp_rcv(struct sk_buff *skb, struct net_device *dev,
 	spin_unlock(&br->lock);
  err:
 	kfree_skb(skb);
-	return 0;
 }

net 03/06: Add GARP applicant-only participant

[Patrick McHardy]

net: Add GARP applicant-only participant

Add an implementation of the GARP (Generic Attribute Registration Protocol)
applicant-only participant. This will be used by the following patch to
add GVRP support to the VLAN code.

Signed-off-by: Patrick McHardy <kaber@trash.net>

---
commit 4c8767ab0402e70217e9ec118d940e56255a8221
tree 5559fa726cb8ac236f69ccd2ead03adb26304e2b
parent 85ecc2934a05203a4f6372cd8f0b132c2d1f2f97
author Patrick McHardy <kaber@trash.net> Thu, 03 Jul 2008 18:11:18 +0200
committer Patrick McHardy <kaber@trash.net> Thu, 03 Jul 2008 18:11:18 +0200

 include/linux/netdevice.h |    2 
 include/net/garp.h        |  127 +++++++++
 net/802/Kconfig           |    4 
 net/802/Makefile          |    1 
 net/802/garp.c            |  633 +++++++++++++++++++++++++++++++++++++++++++++
 5 files changed, 767 insertions(+), 0 deletions(-)

diff --git a/include/linux/netdevice.h b/include/linux/netdevice.h
index 56dadb5..e009c6f 100644
--- a/include/linux/netdevice.h
+++ b/include/linux/netdevice.h
@@ -740,6 +740,8 @@ struct net_device
 	struct net_bridge_port	*br_port;
 	/* macvlan */
 	struct macvlan_port	*macvlan_port;
+	/* GARP */
+	struct garp_port	*garp_port;
 
 	/* class/net/name entry */
 	struct device		dev;
diff --git a/include/net/garp.h b/include/net/garp.h
new file mode 100644
index 0000000..73c7723
--- /dev/null
+++ b/include/net/garp.h
@@ -0,0 +1,127 @@
+#ifndef _NET_GARP_H
+#define _NET_GARP_H
+
+#include <net/stp.h>
+
+#define GARP_PROTOCOL_ID	0x1
+#define GARP_END_MARK		0x0
+
+struct garp_pdu_hdr {
+	__be16	protocol;
+};
+
+struct garp_msg_hdr {
+	u8	attrtype;
+};
+
+enum garp_attr_event {
+	GARP_LEAVE_ALL,
+	GARP_JOIN_EMPTY,
+	GARP_JOIN_IN,
+	GARP_LEAVE_EMPTY,
+	GARP_LEAVE_IN,
+	GARP_EMPTY,
+};
+
+struct garp_attr_hdr {
+	u8	len;
+	u8	event;
+	u8	data[];
+};
+
+struct garp_skb_cb {
+	u8	cur_type;
+};
+
+static inline struct garp_skb_cb *garp_cb(struct sk_buff *skb)
+{
+	BUILD_BUG_ON(sizeof(struct garp_skb_cb) >
+		     FIELD_SIZEOF(struct sk_buff, cb));
+	return (struct garp_skb_cb *)skb->cb;
+}
+
+enum garp_applicant_state {
+	GARP_APPLICANT_INVALID,
+	GARP_APPLICANT_VA,
+	GARP_APPLICANT_AA,
+	GARP_APPLICANT_QA,
+	GARP_APPLICANT_LA,
+	GARP_APPLICANT_VP,
+	GARP_APPLICANT_AP,
+	GARP_APPLICANT_QP,
+	GARP_APPLICANT_VO,
+	GARP_APPLICANT_AO,
+	GARP_APPLICANT_QO,
+	__GARP_APPLICANT_MAX
+};
+#define GARP_APPLICANT_MAX	(__GARP_APPLICANT_MAX - 1)
+
+enum garp_event {
+	GARP_EVENT_REQ_JOIN,
+	GARP_EVENT_REQ_LEAVE,
+	GARP_EVENT_R_JOIN_IN,
+	GARP_EVENT_R_JOIN_EMPTY,
+	GARP_EVENT_R_EMPTY,
+	GARP_EVENT_R_LEAVE_IN,
+	GARP_EVENT_R_LEAVE_EMPTY,
+	GARP_EVENT_TRANSMIT_PDU,
+	__GARP_EVENT_MAX
+};
+#define GARP_EVENT_MAX		(__GARP_EVENT_MAX - 1)
+
+enum garp_action {
+	GARP_ACTION_NONE,
+	GARP_ACTION_S_JOIN_IN,
+	GARP_ACTION_S_LEAVE_EMPTY,
+};
+
+struct garp_attr {
+	struct rb_node			node;
+	enum garp_applicant_state	state;
+	u8				type;
+	u8				dlen;
+	unsigned char			data[];
+};
+
+enum garp_applications {
+	__GARP_APPLICATION_MAX
+};
+#define GARP_APPLICATION_MAX	(__GARP_APPLICATION_MAX - 1)
+
+struct garp_application {
+	enum garp_applications	type;
+	unsigned int		maxattr;
+	struct stp_proto	proto;
+};
+
+struct garp_applicant {
+	struct garp_application	*app;
+	struct net_device	*dev;
+	struct timer_list	join_timer;
+
+	spinlock_t		lock;
+	struct sk_buff_head	queue;
+	struct sk_buff		*pdu;
+	struct rb_root		gid;
+};
+
+struct garp_port {
+	struct garp_applicant	*applicants[GARP_APPLICATION_MAX + 1];
+};
+
+extern int	garp_register_application(struct garp_application *app);
+extern void	garp_unregister_application(struct garp_application *app);
+
+extern int	garp_init_applicant(struct net_device *dev,
+				    struct garp_application *app);
+extern void	garp_uninit_applicant(struct net_device *dev,
+				      struct garp_application *app);
+
+extern int	garp_request_join(const struct net_device *dev,
+				  const struct garp_application *app,
+				  const void *data, u8 len, u8 type);
+extern void	garp_request_leave(const struct net_device *dev,
+				   const struct garp_application *app,
+				   const void *data, u8 len, u8 type);
+
+#endif /* _NET_GARP_H */
diff --git a/net/802/Kconfig b/net/802/Kconfig
index 01cb094..be33d27 100644
--- a/net/802/Kconfig
+++ b/net/802/Kconfig
@@ -1,3 +1,7 @@
 config STP
 	tristate
 	select LLC
+
+config GARP
+	tristate
+	select STP
diff --git a/net/802/Makefile b/net/802/Makefile
index c441d89..7893d67 100644
--- a/net/802/Makefile
+++ b/net/802/Makefile
@@ -11,3 +11,4 @@ obj-$(CONFIG_HIPPI)	+=                 hippi.o
 obj-$(CONFIG_IPX)	+= p8022.o psnap.o p8023.o
 obj-$(CONFIG_ATALK)	+= p8022.o psnap.o
 obj-$(CONFIG_STP)	+= stp.o
+obj-$(CONFIG_GARP)	+= garp.o
diff --git a/net/802/garp.c b/net/802/garp.c
new file mode 100644
index 0000000..3b78f7b
--- /dev/null
+++ b/net/802/garp.c
@@ -0,0 +1,633 @@
+/*
+ *	IEEE 802.1D Generic Attribute Registration Protocol (GARP)
+ *
+ *	Copyright (c) 2008 Patrick McHardy <kaber@trash.net>
+ *
+ *	This program is free software; you can redistribute it and/or
+ *	modify it under the terms of the GNU General Public License
+ *	version 2 as published by the Free Software Foundation.
+ */
+#include <linux/kernel.h>
+#include <linux/timer.h>
+#include <linux/skbuff.h>
+#include <linux/netdevice.h>
+#include <linux/etherdevice.h>
+#include <linux/rtnetlink.h>
+#include <linux/llc.h>
+#include <net/llc.h>
+#include <net/llc_pdu.h>
+#include <net/garp.h>
+#include <asm/unaligned.h>
+
+static unsigned int garp_join_time __read_mostly = 200;
+module_param(garp_join_time, uint, 0644);
+MODULE_PARM_DESC(garp_join_time, "Join time in ms (default 200ms)");
+MODULE_LICENSE("GPL");
+
+static const struct garp_state_trans {
+	u8	state;
+	u8	action;
+} garp_applicant_state_table[GARP_APPLICANT_MAX + 1][GARP_EVENT_MAX + 1] = {
+	[GARP_APPLICANT_VA] = {
+		[GARP_EVENT_TRANSMIT_PDU]	= { .state = GARP_APPLICANT_AA,
+						    .action = GARP_ACTION_S_JOIN_IN },
+		[GARP_EVENT_R_JOIN_IN]		= { .state = GARP_APPLICANT_AA },
+		[GARP_EVENT_R_JOIN_EMPTY]	= { .state = GARP_APPLICANT_VA },
+		[GARP_EVENT_R_EMPTY]		= { .state = GARP_APPLICANT_VA },
+		[GARP_EVENT_R_LEAVE_IN]		= { .state = GARP_APPLICANT_VA },
+		[GARP_EVENT_R_LEAVE_EMPTY]	= { .state = GARP_APPLICANT_VP },
+		[GARP_EVENT_REQ_JOIN]		= { .state = GARP_APPLICANT_INVALID },
+		[GARP_EVENT_REQ_LEAVE]		= { .state = GARP_APPLICANT_LA },
+	},
+	[GARP_APPLICANT_AA] = {
+		[GARP_EVENT_TRANSMIT_PDU]	= { .state = GARP_APPLICANT_QA,
+						    .action = GARP_ACTION_S_JOIN_IN },
+		[GARP_EVENT_R_JOIN_IN]		= { .state = GARP_APPLICANT_QA },
+		[GARP_EVENT_R_JOIN_EMPTY]	= { .state = GARP_APPLICANT_VA },
+		[GARP_EVENT_R_EMPTY]		= { .state = GARP_APPLICANT_VA },
+		[GARP_EVENT_R_LEAVE_IN]		= { .state = GARP_APPLICANT_VA },
+		[GARP_EVENT_R_LEAVE_EMPTY]	= { .state = GARP_APPLICANT_VP },
+		[GARP_EVENT_REQ_JOIN]		= { .state = GARP_APPLICANT_INVALID },
+		[GARP_EVENT_REQ_LEAVE]		= { .state = GARP_APPLICANT_LA },
+	},
+	[GARP_APPLICANT_QA] = {
+		[GARP_EVENT_TRANSMIT_PDU]	= { .state = GARP_APPLICANT_INVALID },
+		[GARP_EVENT_R_JOIN_IN]		= { .state = GARP_APPLICANT_QA },
+		[GARP_EVENT_R_JOIN_EMPTY]	= { .state = GARP_APPLICANT_VA },
+		[GARP_EVENT_R_EMPTY]		= { .state = GARP_APPLICANT_VA },
+		[GARP_EVENT_R_LEAVE_IN]		= { .state = GARP_APPLICANT_VP },
+		[GARP_EVENT_R_LEAVE_EMPTY]	= { .state = GARP_APPLICANT_VP },
+		[GARP_EVENT_REQ_JOIN]		= { .state = GARP_APPLICANT_INVALID },
+		[GARP_EVENT_REQ_LEAVE]		= { .state = GARP_APPLICANT_LA },
+	},
+	[GARP_APPLICANT_LA] = {
+		[GARP_EVENT_TRANSMIT_PDU]	= { .state = GARP_APPLICANT_VO,
+						    .action = GARP_ACTION_S_LEAVE_EMPTY },
+		[GARP_EVENT_R_JOIN_IN]		= { .state = GARP_APPLICANT_LA },
+		[GARP_EVENT_R_JOIN_EMPTY]	= { .state = GARP_APPLICANT_VO },
+		[GARP_EVENT_R_EMPTY]		= { .state = GARP_APPLICANT_LA },
+		[GARP_EVENT_R_LEAVE_IN]		= { .state = GARP_APPLICANT_LA },
+		[GARP_EVENT_R_LEAVE_EMPTY]	= { .state = GARP_APPLICANT_VO },
+		[GARP_EVENT_REQ_JOIN]		= { .state = GARP_APPLICANT_VA },
+		[GARP_EVENT_REQ_LEAVE]		= { .state = GARP_APPLICANT_INVALID },
+	},
+	[GARP_APPLICANT_VP] = {
+		[GARP_EVENT_TRANSMIT_PDU]	= { .state = GARP_APPLICANT_AA,
+						    .action = GARP_ACTION_S_JOIN_IN },
+		[GARP_EVENT_R_JOIN_IN]		= { .state = GARP_APPLICANT_AP },
+		[GARP_EVENT_R_JOIN_EMPTY]	= { .state = GARP_APPLICANT_VP },
+		[GARP_EVENT_R_EMPTY]		= { .state = GARP_APPLICANT_VP },
+		[GARP_EVENT_R_LEAVE_IN]		= { .state = GARP_APPLICANT_VP },
+		[GARP_EVENT_R_LEAVE_EMPTY]	= { .state = GARP_APPLICANT_VP },
+		[GARP_EVENT_REQ_JOIN]		= { .state = GARP_APPLICANT_INVALID },
+		[GARP_EVENT_REQ_LEAVE]		= { .state = GARP_APPLICANT_VO },
+	},
+	[GARP_APPLICANT_AP] = {
+		[GARP_EVENT_TRANSMIT_PDU]	= { .state = GARP_APPLICANT_QA,
+						    .action = GARP_ACTION_S_JOIN_IN },
+		[GARP_EVENT_R_JOIN_IN]		= { .state = GARP_APPLICANT_QP },
+		[GARP_EVENT_R_JOIN_EMPTY]	= { .state = GARP_APPLICANT_VP },
+		[GARP_EVENT_R_EMPTY]		= { .state = GARP_APPLICANT_VP },
+		[GARP_EVENT_R_LEAVE_IN]		= { .state = GARP_APPLICANT_VP },
+		[GARP_EVENT_R_LEAVE_EMPTY]	= { .state = GARP_APPLICANT_VP },
+		[GARP_EVENT_REQ_JOIN]		= { .state = GARP_APPLICANT_INVALID },
+		[GARP_EVENT_REQ_LEAVE]		= { .state = GARP_APPLICANT_AO },
+	},
+	[GARP_APPLICANT_QP] = {
+		[GARP_EVENT_TRANSMIT_PDU]	= { .state = GARP_APPLICANT_INVALID },
+		[GARP_EVENT_R_JOIN_IN]		= { .state = GARP_APPLICANT_QP },
+		[GARP_EVENT_R_JOIN_EMPTY]	= { .state = GARP_APPLICANT_VP },
+		[GARP_EVENT_R_EMPTY]		= { .state = GARP_APPLICANT_VP },
+		[GARP_EVENT_R_LEAVE_IN]		= { .state = GARP_APPLICANT_VP },
+		[GARP_EVENT_R_LEAVE_EMPTY]	= { .state = GARP_APPLICANT_VP },
+		[GARP_EVENT_REQ_JOIN]		= { .state = GARP_APPLICANT_INVALID },
+		[GARP_EVENT_REQ_LEAVE]		= { .state = GARP_APPLICANT_QO },
+	},
+	[GARP_APPLICANT_VO] = {
+		[GARP_EVENT_TRANSMIT_PDU]	= { .state = GARP_APPLICANT_INVALID },
+		[GARP_EVENT_R_JOIN_IN]		= { .state = GARP_APPLICANT_AO },
+		[GARP_EVENT_R_JOIN_EMPTY]	= { .state = GARP_APPLICANT_VO },
+		[GARP_EVENT_R_EMPTY]		= { .state = GARP_APPLICANT_VO },
+		[GARP_EVENT_R_LEAVE_IN]		= { .state = GARP_APPLICANT_VO },
+		[GARP_EVENT_R_LEAVE_EMPTY]	= { .state = GARP_APPLICANT_VO },
+		[GARP_EVENT_REQ_JOIN]		= { .state = GARP_APPLICANT_VP },
+		[GARP_EVENT_REQ_LEAVE]		= { .state = GARP_APPLICANT_INVALID },
+	},
+	[GARP_APPLICANT_AO] = {
+		[GARP_EVENT_TRANSMIT_PDU]	= { .state = GARP_APPLICANT_INVALID },
+		[GARP_EVENT_R_JOIN_IN]		= { .state = GARP_APPLICANT_QO },
+		[GARP_EVENT_R_JOIN_EMPTY]	= { .state = GARP_APPLICANT_VO },
+		[GARP_EVENT_R_EMPTY]		= { .state = GARP_APPLICANT_VO },
+		[GARP_EVENT_R_LEAVE_IN]		= { .state = GARP_APPLICANT_VO },
+		[GARP_EVENT_R_LEAVE_EMPTY]	= { .state = GARP_APPLICANT_VO },
+		[GARP_EVENT_REQ_JOIN]		= { .state = GARP_APPLICANT_AP },
+		[GARP_EVENT_REQ_LEAVE]		= { .state = GARP_APPLICANT_INVALID },
+	},
+	[GARP_APPLICANT_QO] = {
+		[GARP_EVENT_TRANSMIT_PDU]	= { .state = GARP_APPLICANT_INVALID },
+		[GARP_EVENT_R_JOIN_IN]		= { .state = GARP_APPLICANT_QO },
+		[GARP_EVENT_R_JOIN_EMPTY]	= { .state = GARP_APPLICANT_VO },
+		[GARP_EVENT_R_EMPTY]		= { .state = GARP_APPLICANT_VO },
+		[GARP_EVENT_R_LEAVE_IN]		= { .state = GARP_APPLICANT_VO },
+		[GARP_EVENT_R_LEAVE_EMPTY]	= { .state = GARP_APPLICANT_VO },
+		[GARP_EVENT_REQ_JOIN]		= { .state = GARP_APPLICANT_QP },
+		[GARP_EVENT_REQ_LEAVE]		= { .state = GARP_APPLICANT_INVALID },
+	},
+};
+
+static int garp_attr_cmp(const struct garp_attr *attr,
+			 const void *data, u8 len, u8 type)
+{
+	if (attr->type != type)
+		return attr->type - type;
+	if (attr->dlen != len)
+		return attr->dlen - len;
+	return memcmp(attr->data, data, len);
+}
+
+static struct garp_attr *garp_attr_lookup(const struct garp_applicant *app,
+					  const void *data, u8 len, u8 type)
+{
+	struct rb_node *parent = app->gid.rb_node;
+	struct garp_attr *attr;
+	int d;
+
+	while (parent) {
+		attr = rb_entry(parent, struct garp_attr, node);
+		d = garp_attr_cmp(attr, data, len, type);
+		if (d < 0)
+			parent = parent->rb_left;
+		else if (d > 0)
+			parent = parent->rb_right;
+		else
+			return attr;
+	}
+	return NULL;
+}
+
+static void garp_attr_insert(struct garp_applicant *app, struct garp_attr *new)
+{
+	struct rb_node *parent = NULL, **p = &app->gid.rb_node;
+	struct garp_attr *attr;
+	int d;
+
+	while (*p) {
+		parent = *p;
+		attr = rb_entry(parent, struct garp_attr, node);
+		d = garp_attr_cmp(attr, new->data, new->dlen, new->type);
+		if (d < 0)
+			p = &parent->rb_left;
+		else if (d > 0)
+			p = &parent->rb_right;
+	}
+	rb_link_node(&new->node, parent, p);
+	rb_insert_color(&new->node, &app->gid);
+}
+
+static struct garp_attr *garp_attr_create(struct garp_applicant *app,
+					  const void *data, u8 len, u8 type)
+{
+	struct garp_attr *attr;
+
+	attr = kmalloc(sizeof(*attr) + len, GFP_ATOMIC);
+	if (!attr)
+		return attr;
+	attr->state = GARP_APPLICANT_VO;
+	attr->type  = type;
+	attr->dlen  = len;
+	memcpy(attr->data, data, len);
+	garp_attr_insert(app, attr);
+	return attr;
+}
+
+static void garp_attr_destroy(struct garp_applicant *app, struct garp_attr *attr)
+{
+	rb_erase(&attr->node, &app->gid);
+	kfree(attr);
+}
+
+static int garp_pdu_init(struct garp_applicant *app)
+{
+	struct sk_buff *skb;
+	struct garp_pdu_hdr *gp;
+
+#define LLC_RESERVE	sizeof(struct llc_pdu_un)
+	skb = alloc_skb(app->dev->mtu + LL_RESERVED_SPACE(app->dev),
+			GFP_ATOMIC);
+	if (!skb)
+		return -ENOMEM;
+
+	skb->dev = app->dev;
+	skb->protocol = htons(ETH_P_802_2);
+	skb_reserve(skb, LL_RESERVED_SPACE(app->dev) + LLC_RESERVE);
+
+	gp = (struct garp_pdu_hdr *)__skb_put(skb, sizeof(*gp));
+	put_unaligned(htons(GARP_PROTOCOL_ID), &gp->protocol);
+
+	app->pdu = skb;
+	return 0;
+}
+
+static int garp_pdu_append_end_mark(struct garp_applicant *app)
+{
+	if (skb_tailroom(app->pdu) < sizeof(u8))
+		return -1;
+	*(u8 *)__skb_put(app->pdu, sizeof(u8)) = GARP_END_MARK;
+	return 0;
+}
+
+static void garp_pdu_queue(struct garp_applicant *app)
+{
+	if (!app->pdu)
+		return;
+
+	garp_pdu_append_end_mark(app);
+	garp_pdu_append_end_mark(app);
+
+	llc_pdu_header_init(app->pdu, LLC_PDU_TYPE_U, LLC_SAP_BSPAN,
+			    LLC_SAP_BSPAN, LLC_PDU_CMD);
+	llc_pdu_init_as_ui_cmd(app->pdu);
+	llc_mac_hdr_init(app->pdu, app->dev->dev_addr,
+			 app->app->proto.group_address);
+
+	skb_queue_tail(&app->queue, app->pdu);
+	app->pdu = NULL;
+}
+
+static void garp_queue_xmit(struct garp_applicant *app)
+{
+	struct sk_buff *skb;
+
+	while ((skb = skb_dequeue(&app->queue)))
+		dev_queue_xmit(skb);
+}
+
+static int garp_pdu_append_msg(struct garp_applicant *app, u8 attrtype)
+{
+	struct garp_msg_hdr *gm;
+
+	if (skb_tailroom(app->pdu) < sizeof(*gm))
+		return -1;
+	gm = (struct garp_msg_hdr *)__skb_put(app->pdu, sizeof(*gm));
+	gm->attrtype = attrtype;
+	garp_cb(app->pdu)->cur_type = attrtype;
+	return 0;
+}
+
+static int garp_pdu_append_attr(struct garp_applicant *app,
+				const struct garp_attr *attr,
+				enum garp_attr_event event)
+{
+	struct garp_attr_hdr *ga;
+	unsigned int len;
+	int err;
+again:
+	if (!app->pdu) {
+		err = garp_pdu_init(app);
+		if (err < 0)
+			return err;
+	}
+
+	if (garp_cb(app->pdu)->cur_type != attr->type) {
+		if (garp_cb(app->pdu)->cur_type &&
+		    garp_pdu_append_end_mark(app) < 0)
+			goto queue;
+		if (garp_pdu_append_msg(app, attr->type) < 0)
+			goto queue;
+	}
+
+	len = sizeof(*ga) + attr->dlen;
+	if (skb_tailroom(app->pdu) < len)
+		goto queue;
+	ga = (struct garp_attr_hdr *)__skb_put(app->pdu, len);
+	ga->len   = len;
+	ga->event = event;
+	memcpy(ga->data, attr->data, attr->dlen);
+	return 0;
+
+queue:
+	garp_pdu_queue(app);
+	goto again;
+}
+
+static void garp_attr_event(struct garp_applicant *app,
+			    struct garp_attr *attr, enum garp_event event)
+{
+	enum garp_applicant_state state;
+
+	state = garp_applicant_state_table[attr->state][event].state;
+	if (state == GARP_APPLICANT_INVALID)
+		return;
+
+	switch (garp_applicant_state_table[attr->state][event].action) {
+	case GARP_ACTION_NONE:
+		break;
+	case GARP_ACTION_S_JOIN_IN:
+		garp_pdu_append_attr(app, attr, GARP_JOIN_IN);
+		break;
+	case GARP_ACTION_S_LEAVE_EMPTY:
+		garp_pdu_append_attr(app, attr, GARP_LEAVE_EMPTY);
+		/* As a pure applicant, sending a leave message implies that
+		 * the attribute was unregistered and can be destroyed. */
+		garp_attr_destroy(app, attr);
+		return;
+	default:
+		WARN_ON(1);
+	}
+
+	attr->state = state;
+}
+
+int garp_request_join(const struct net_device *dev,
+		      const struct garp_application *appl,
+		      const void *data, u8 len, u8 type)
+{
+	struct garp_port *port = dev->garp_port;
+	struct garp_applicant *app = port->applicants[appl->type];
+	struct garp_attr *attr;
+
+	spin_lock_bh(&app->lock);
+	attr = garp_attr_create(app, data, len, type);
+	if (!attr) {
+		spin_unlock_bh(&app->lock);
+		return -ENOMEM;
+	}
+	garp_attr_event(app, attr, GARP_EVENT_REQ_JOIN);
+	spin_unlock_bh(&app->lock);
+	return 0;
+}
+EXPORT_SYMBOL_GPL(garp_request_join);
+
+void garp_request_leave(const struct net_device *dev,
+			const struct garp_application *appl,
+			const void *data, u8 len, u8 type)
+{
+	struct garp_port *port = dev->garp_port;
+	struct garp_applicant *app = port->applicants[appl->type];
+	struct garp_attr *attr;
+
+	spin_lock_bh(&app->lock);
+	attr = garp_attr_lookup(app, data, len, type);
+	if (!attr) {
+		spin_unlock_bh(&app->lock);
+		return;
+	}
+	garp_attr_event(app, attr, GARP_EVENT_REQ_LEAVE);
+	spin_unlock_bh(&app->lock);
+}
+EXPORT_SYMBOL_GPL(garp_request_leave);
+
+static void garp_gid_event(struct garp_applicant *app, enum garp_event event)
+{
+	struct rb_node *node, *next;
+	struct garp_attr *attr;
+
+	for (node = rb_first(&app->gid);
+	     next = node ? rb_next(node) : NULL, node != NULL;
+	     node = next) {
+		attr = rb_entry(node, struct garp_attr, node);
+		garp_attr_event(app, attr, event);
+	}
+}
+
+static void garp_join_timer_arm(struct garp_applicant *app)
+{
+	unsigned long delay;
+
+	delay = (u64)msecs_to_jiffies(garp_join_time) * net_random() >> 32;
+	mod_timer(&app->join_timer, jiffies + delay);
+}
+
+static void garp_join_timer(unsigned long data)
+{
+	struct garp_applicant *app = (struct garp_applicant *)data;
+
+	spin_lock(&app->lock);
+	garp_gid_event(app, GARP_EVENT_TRANSMIT_PDU);
+	garp_pdu_queue(app);
+	spin_unlock(&app->lock);
+
+	garp_queue_xmit(app);
+	garp_join_timer_arm(app);
+}
+
+static int garp_pdu_parse_end_mark(struct sk_buff *skb)
+{
+	if (!pskb_may_pull(skb, sizeof(u8)))
+		return -1;
+	if (*skb->data == GARP_END_MARK) {
+		skb_pull(skb, sizeof(u8));
+		return -1;
+	}
+	return 0;
+}
+
+static int garp_pdu_parse_attr(struct garp_applicant *app, struct sk_buff *skb,
+			       u8 attrtype)
+{
+	const struct garp_attr_hdr *ga;
+	struct garp_attr *attr;
+	enum garp_event event;
+	unsigned int dlen;
+
+	if (!pskb_may_pull(skb, sizeof(*ga)))
+		return -1;
+	ga = (struct garp_attr_hdr *)skb->data;
+	if (ga->len < sizeof(*ga))
+		return -1;
+
+	if (!pskb_may_pull(skb, ga->len))
+		return -1;
+	skb_pull(skb, ga->len);
+	dlen = sizeof(*ga) - ga->len;
+
+	if (attrtype > app->app->maxattr)
+		return 0;
+
+	switch (ga->event) {
+	case GARP_LEAVE_ALL:
+		if (dlen != 0)
+			return -1;
+		garp_gid_event(app, GARP_EVENT_R_LEAVE_EMPTY);
+		return 0;
+	case GARP_JOIN_EMPTY:
+		event = GARP_EVENT_R_JOIN_EMPTY;
+		break;
+	case GARP_JOIN_IN:
+		event = GARP_EVENT_R_JOIN_IN;
+		break;
+	case GARP_LEAVE_EMPTY:
+		event = GARP_EVENT_R_LEAVE_EMPTY;
+		break;
+	case GARP_EMPTY:
+		event = GARP_EVENT_R_EMPTY;
+		break;
+	default:
+		return 0;
+	}
+
+	if (dlen == 0)
+		return -1;
+	attr = garp_attr_lookup(app, ga->data, dlen, attrtype);
+	if (attr == NULL)
+		return 0;
+	garp_attr_event(app, attr, event);
+	return 0;
+}
+
+static int garp_pdu_parse_msg(struct garp_applicant *app, struct sk_buff *skb)
+{
+	const struct garp_msg_hdr *gm;
+
+	if (!pskb_may_pull(skb, sizeof(*gm)))
+		return -1;
+	gm = (struct garp_msg_hdr *)skb->data;
+	if (gm->attrtype == 0)
+		return -1;
+	skb_pull(skb, sizeof(*gm));
+
+	while (skb->len > 0) {
+		if (garp_pdu_parse_attr(app, skb, gm->attrtype) < 0)
+			return -1;
+		if (garp_pdu_parse_end_mark(skb) < 0)
+			break;
+	}
+	return 0;
+}
+
+static void garp_pdu_rcv(const struct stp_proto *proto, struct sk_buff *skb,
+			 struct net_device *dev)
+{
+	struct garp_application *appl = proto->data;
+	struct garp_port *port;
+	struct garp_applicant *app;
+	const struct garp_pdu_hdr *gp;
+
+	port = rcu_dereference(dev->garp_port);
+	if (!port)
+		goto err;
+	app = rcu_dereference(port->applicants[appl->type]);
+	if (!app)
+		goto err;
+
+	if (!pskb_may_pull(skb, sizeof(*gp)))
+		goto err;
+	gp = (struct garp_pdu_hdr *)skb->data;
+	if (get_unaligned(&gp->protocol) != htons(GARP_PROTOCOL_ID))
+		goto err;
+	skb_pull(skb, sizeof(*gp));
+
+	spin_lock(&app->lock);
+	while (skb->len > 0) {
+		if (garp_pdu_parse_msg(app, skb) < 0)
+			break;
+		if (garp_pdu_parse_end_mark(skb) < 0)
+			break;
+	}
+	spin_unlock(&app->lock);
+err:
+	kfree_skb(skb);
+}
+
+static int garp_init_port(struct net_device *dev)
+{
+	struct garp_port *port;
+
+	port = kzalloc(sizeof(*port), GFP_KERNEL);
+	if (!port)
+		return -ENOMEM;
+	rcu_assign_pointer(dev->garp_port, port);
+	return 0;
+}
+
+static void garp_release_port(struct net_device *dev)
+{
+	struct garp_port *port = dev->garp_port;
+	unsigned int i;
+
+	for (i = 0; i <= GARP_APPLICATION_MAX; i++) {
+		if (port->applicants[i])
+			return;
+	}
+	rcu_assign_pointer(dev->garp_port, NULL);
+	synchronize_rcu();
+	kfree(port);
+}
+
+int garp_init_applicant(struct net_device *dev, struct garp_application *appl)
+{
+	struct garp_applicant *app;
+	int err;
+
+	ASSERT_RTNL();
+
+	if (!dev->garp_port) {
+		err = garp_init_port(dev);
+		if (err < 0)
+			goto err1;
+	}
+
+	err = -ENOMEM;
+	app = kzalloc(sizeof(*app), GFP_KERNEL);
+	if (!app)
+		goto err2;
+
+	err = dev_mc_add(dev, appl->proto.group_address, ETH_ALEN, 0);
+	if (err < 0)
+		goto err3;
+
+	app->dev = dev;
+	app->app = appl;
+	app->gid = RB_ROOT;
+	spin_lock_init(&app->lock);
+	skb_queue_head_init(&app->queue);
+	rcu_assign_pointer(dev->garp_port->applicants[appl->type], app);
+	setup_timer(&app->join_timer, garp_join_timer, (unsigned long)app);
+	garp_join_timer_arm(app);
+	return 0;
+
+err3:
+	kfree(app);
+err2:
+	garp_release_port(dev);
+err1:
+	return err;
+}
+EXPORT_SYMBOL_GPL(garp_init_applicant);
+
+void garp_uninit_applicant(struct net_device *dev, struct garp_application *appl)
+{
+	struct garp_port *port = dev->garp_port;
+	struct garp_applicant *app = port->applicants[appl->type];
+
+	ASSERT_RTNL();
+
+	rcu_assign_pointer(port->applicants[appl->type], NULL);
+	synchronize_rcu();
+
+	/* Delete timer and generate a final TRANSMIT_PDU event to flush out
+	 * all pending messages before the applicant is gone. */
+	del_timer_sync(&app->join_timer);
+	garp_gid_event(app, GARP_EVENT_TRANSMIT_PDU);
+	garp_pdu_queue(app);
+	garp_queue_xmit(app);
+
+	dev_mc_delete(dev, appl->proto.group_address, ETH_ALEN, 0);
+	kfree(app);
+	garp_release_port(dev);
+}
+EXPORT_SYMBOL_GPL(garp_uninit_applicant);
+
+int garp_register_application(struct garp_application *appl)
+{
+	appl->proto.rcv = garp_pdu_rcv;
+	appl->proto.data = appl;
+	return stp_proto_register(&appl->proto);
+}
+EXPORT_SYMBOL_GPL(garp_register_application);
+
+void garp_unregister_application(struct garp_application *appl)
+{
+	stp_proto_unregister(&appl->proto);
+}
+EXPORT_SYMBOL_GPL(garp_unregister_application);

vlan 04/06: Change vlan_dev_set_vlan_flag() to handle multiple flags at once

[Patrick McHardy]

vlan: Change vlan_dev_set_vlan_flag() to handle multiple flags at once

Change vlan_dev_set_vlan_flag() to handle multiple flags at once and
rename to vlan_dev_change_flags(). This allows to to use it from the
netlink interface, which in turn allows to handle necessary adjustments
when changing flags centrally.

Signed-off-by: Patrick McHardy <kaber@trash.net>

---
commit 12ad21b911562368e08dc7166109370130c156c7
tree 1002d0cb3c6b1158a25fdd3494e81904c1931d17
parent 4c8767ab0402e70217e9ec118d940e56255a8221
author Patrick McHardy <kaber@trash.net> Thu, 03 Jul 2008 18:11:18 +0200
committer Patrick McHardy <kaber@trash.net> Thu, 03 Jul 2008 18:11:18 +0200

 net/8021q/vlan.c         |    6 +++---
 net/8021q/vlan.h         |    3 +--
 net/8021q/vlan_dev.c     |   20 +++++++++-----------
 net/8021q/vlan_netlink.c |    4 +---
 4 files changed, 14 insertions(+), 19 deletions(-)

diff --git a/net/8021q/vlan.c b/net/8021q/vlan.c
index ab2225d..b591bfc 100644
--- a/net/8021q/vlan.c
+++ b/net/8021q/vlan.c
@@ -591,9 +591,9 @@ static int vlan_ioctl_handler(struct net *net, void __user *arg)
 		err = -EPERM;
 		if (!capable(CAP_NET_ADMIN))
 			break;
-		err = vlan_dev_set_vlan_flag(dev,
-					     args.u.flag,
-					     args.vlan_qos);
+		err = vlan_dev_change_flags(dev,
+					    args.vlan_qos ? args.u.flag : 0,
+					    args.u.flag);
 		break;
 
 	case SET_VLAN_NAME_TYPE_CMD:
diff --git a/net/8021q/vlan.h b/net/8021q/vlan.h
index 5229a72..639e254 100644
--- a/net/8021q/vlan.h
+++ b/net/8021q/vlan.h
@@ -28,8 +28,7 @@ void vlan_dev_set_ingress_priority(const struct net_device *dev,
 				   u32 skb_prio, short vlan_prio);
 int vlan_dev_set_egress_priority(const struct net_device *dev,
 				 u32 skb_prio, short vlan_prio);
-int vlan_dev_set_vlan_flag(const struct net_device *dev,
-			   u32 flag, short flag_val);
+int vlan_dev_change_flags(const struct net_device *dev, u32 flag, u32 mask);
 void vlan_dev_get_realdev_name(const struct net_device *dev, char *result);
 void vlan_dev_get_vid(const struct net_device *dev, unsigned short *result);
 
diff --git a/net/8021q/vlan_dev.c b/net/8021q/vlan_dev.c
index 5d055c2..76c665c 100644
--- a/net/8021q/vlan_dev.c
+++ b/net/8021q/vlan_dev.c
@@ -507,18 +507,16 @@ int vlan_dev_set_egress_priority(const struct net_device *dev,
 }
 
 /* Flags are defined in the vlan_flags enum in include/linux/if_vlan.h file. */
-int vlan_dev_set_vlan_flag(const struct net_device *dev,
-			   u32 flag, short flag_val)
+int vlan_dev_change_flags(const struct net_device *dev, u32 flags, u32 mask)
 {
-	/* verify flag is supported */
-	if (flag == VLAN_FLAG_REORDER_HDR) {
-		if (flag_val)
-			vlan_dev_info(dev)->flags |= VLAN_FLAG_REORDER_HDR;
-		else
-			vlan_dev_info(dev)->flags &= ~VLAN_FLAG_REORDER_HDR;
-		return 0;
-	}
-	return -EINVAL;
+	struct vlan_dev_info *vlan = vlan_dev_info(dev);
+	u32 old_flags = vlan->flags;
+
+	if (mask & ~VLAN_FLAG_REORDER_HDR)
+		return -EINVAL;
+
+	vlan->flags = (old_flags & ~mask) | (flags & mask);
+	return 0;
 }
 
 void vlan_dev_get_realdev_name(const struct net_device *dev, char *result)
diff --git a/net/8021q/vlan_netlink.c b/net/8021q/vlan_netlink.c
index c93e69e..fd7cb19 100644
--- a/net/8021q/vlan_netlink.c
+++ b/net/8021q/vlan_netlink.c
@@ -75,7 +75,6 @@ static int vlan_validate(struct nlattr *tb[], struct nlattr *data[])
 static int vlan_changelink(struct net_device *dev,
 			   struct nlattr *tb[], struct nlattr *data[])
 {
-	struct vlan_dev_info *vlan = vlan_dev_info(dev);
 	struct ifla_vlan_flags *flags;
 	struct ifla_vlan_qos_mapping *m;
 	struct nlattr *attr;
@@ -83,8 +82,7 @@ static int vlan_changelink(struct net_device *dev,
 
 	if (data[IFLA_VLAN_FLAGS]) {
 		flags = nla_data(data[IFLA_VLAN_FLAGS]);
-		vlan->flags = (vlan->flags & ~flags->mask) |
-			      (flags->flags & flags->mask);
+		vlan_dev_change_flags(dev, flags->flags, flags->mask);
 	}
 	if (data[IFLA_VLAN_INGRESS_QOS]) {
 		nla_for_each_nested(attr, data[IFLA_VLAN_INGRESS_QOS], rem) {

vlan 05/06: Move device unregistration before lower dev cleanup

[Patrick McHardy]

vlan: Move device unregistration before lower dev cleanup

Move the unregister_netdevice() call for the VLAN device before cleanup
for the lower device. This is needed by GVRP so it can send a leave
message before the applicant on the lower device is cleaned up.

Signed-off-by: Patrick McHardy <kaber@trash.net>

---
commit 8f8f9dfb07f7c96492d58488d1aaf670329cc91a
tree 7748d32b10dc9342fe36e5f8259d259494c17641
parent 12ad21b911562368e08dc7166109370130c156c7
author Patrick McHardy <kaber@trash.net> Thu, 03 Jul 2008 18:11:19 +0200
committer Patrick McHardy <kaber@trash.net> Thu, 03 Jul 2008 18:11:19 +0200

 net/8021q/vlan.c |    4 ++--
 1 files changed, 2 insertions(+), 2 deletions(-)

diff --git a/net/8021q/vlan.c b/net/8021q/vlan.c
index b591bfc..8cae2da 100644
--- a/net/8021q/vlan.c
+++ b/net/8021q/vlan.c
@@ -165,6 +165,8 @@ void unregister_vlan_dev(struct net_device *dev)
 
 	synchronize_net();
 
+	unregister_netdevice(dev);
+
 	/* If the group is now empty, kill off the group. */
 	if (grp->nr_vlans == 0) {
 		if (real_dev->features & NETIF_F_HW_VLAN_RX)
@@ -178,8 +180,6 @@ void unregister_vlan_dev(struct net_device *dev)
 
 	/* Get rid of the vlan's reference to real_dev */
 	dev_put(real_dev);
-
-	unregister_netdevice(dev);
 }
 
 static void vlan_transfer_operstate(const struct net_device *dev,

vlan 06/06: Add GVRP support

[Patrick McHardy]

vlan: Add GVRP support

Add GVRP support for dynamically registering VLANs with switches.

By default GVRP is disabled because we only support the applicant-only
participant model, which means it should not be enabled on vlans that
are members of a bridge. Since there is currently no way to cleanly
determine that, the user is responsible for enabling it.

The code is pretty small and low impact, its wrapped in a config
option though because it depends on the GARP implementation and
the STP core.

Signed-off-by: Patrick McHardy <kaber@trash.net>

---
commit 48425f9bd27acfbc92688ac133dac7a536879082
tree f9794528a4352d8f48623495b0e53306f8cc33a7
parent 8f8f9dfb07f7c96492d58488d1aaf670329cc91a
author Patrick McHardy <kaber@trash.net> Thu, 03 Jul 2008 18:11:19 +0200
committer Patrick McHardy <kaber@trash.net> Thu, 03 Jul 2008 18:11:19 +0200

 include/linux/if_vlan.h  |    1 +
 include/net/garp.h       |    1 +
 net/8021q/Kconfig        |   10 +++++++
 net/8021q/Makefile       |    9 ++----
 net/8021q/vlan.c         |   23 +++++++++++++---
 net/8021q/vlan.h         |   16 +++++++++++
 net/8021q/vlan_dev.c     |   18 +++++++++++--
 net/8021q/vlan_gvrp.c    |   66 ++++++++++++++++++++++++++++++++++++++++++++++
 net/8021q/vlan_netlink.c |    3 +-
 9 files changed, 134 insertions(+), 13 deletions(-)

diff --git a/include/linux/if_vlan.h b/include/linux/if_vlan.h
index 15ace02..5190452 100644
--- a/include/linux/if_vlan.h
+++ b/include/linux/if_vlan.h
@@ -402,6 +402,7 @@ enum vlan_ioctl_cmds {
 
 enum vlan_flags {
 	VLAN_FLAG_REORDER_HDR	= 0x1,
+	VLAN_FLAG_GVRP		= 0x2,
 };
 
 enum vlan_name_types {
diff --git a/include/net/garp.h b/include/net/garp.h
index 73c7723..825f172 100644
--- a/include/net/garp.h
+++ b/include/net/garp.h
@@ -84,6 +84,7 @@ struct garp_attr {
 };
 
 enum garp_applications {
+	GARP_APPLICATION_GVRP,
 	__GARP_APPLICATION_MAX
 };
 #define GARP_APPLICATION_MAX	(__GARP_APPLICATION_MAX - 1)
diff --git a/net/8021q/Kconfig b/net/8021q/Kconfig
index c4a382e..fa073a5 100644
--- a/net/8021q/Kconfig
+++ b/net/8021q/Kconfig
@@ -17,3 +17,13 @@ config VLAN_8021Q
 	  will be called 8021q.
 
 	  If unsure, say N.
+
+config VLAN_8021Q_GVRP
+	bool "GVRP (GARP VLAN Registration Protocol) support"
+	depends on VLAN_8021Q
+	select GARP
+	help
+	  Select this to enable GVRP end-system support. GVRP is used for
+	  automatic propagation of registered VLANs to switches.
+
+	  If unsure, say N.
diff --git a/net/8021q/Makefile b/net/8021q/Makefile
index 10ca7f4..3006e9e 100644
--- a/net/8021q/Makefile
+++ b/net/8021q/Makefile
@@ -4,9 +4,6 @@
 
 obj-$(CONFIG_VLAN_8021Q) += 8021q.o
 
-8021q-objs := vlan.o vlan_dev.o vlan_netlink.o
-
-ifeq ($(CONFIG_PROC_FS),y)
-8021q-objs += vlanproc.o
-endif
-
+8021q-y				:= vlan.o vlan_dev.o vlan_netlink.o
+8021q-$(CONFIG_VLAN_8021Q_GVRP)	+= vlan_gvrp.o
+8021q-$(CONFIG_PROC_FS)		+= vlanproc.o
diff --git a/net/8021q/vlan.c b/net/8021q/vlan.c
index 8cae2da..b529110 100644
--- a/net/8021q/vlan.c
+++ b/net/8021q/vlan.c
@@ -169,6 +169,8 @@ void unregister_vlan_dev(struct net_device *dev)
 
 	/* If the group is now empty, kill off the group. */
 	if (grp->nr_vlans == 0) {
+		vlan_gvrp_uninit_applicant(real_dev);
+
 		if (real_dev->features & NETIF_F_HW_VLAN_RX)
 			real_dev->vlan_rx_register(real_dev, NULL);
 
@@ -249,15 +251,18 @@ int register_vlan_dev(struct net_device *dev)
 		ngrp = grp = vlan_group_alloc(real_dev);
 		if (!grp)
 			return -ENOBUFS;
+		err = vlan_gvrp_init_applicant(real_dev);
+		if (err < 0)
+			goto out_free_group;
 	}
 
 	err = vlan_group_prealloc_vid(grp, vlan_id);
 	if (err < 0)
-		goto out_free_group;
+		goto out_uninit_applicant;
 
 	err = register_netdevice(dev);
 	if (err < 0)
-		goto out_free_group;
+		goto out_uninit_applicant;
 
 	/* Account for reference in struct vlan_dev_info */
 	dev_hold(real_dev);
@@ -278,6 +283,9 @@ int register_vlan_dev(struct net_device *dev)
 
 	return 0;
 
+out_uninit_applicant:
+	if (ngrp)
+		vlan_gvrp_uninit_applicant(real_dev);
 out_free_group:
 	if (ngrp)
 		vlan_group_free(ngrp);
@@ -713,14 +721,20 @@ static int __init vlan_proto_init(void)
 	if (err < 0)
 		goto err2;
 
-	err = vlan_netlink_init();
+	err = vlan_gvrp_init();
 	if (err < 0)
 		goto err3;
 
+	err = vlan_netlink_init();
+	if (err < 0)
+		goto err4;
+
 	dev_add_pack(&vlan_packet_type);
 	vlan_ioctl_set(vlan_ioctl_handler);
 	return 0;
 
+err4:
+	vlan_gvrp_uninit();
 err3:
 	unregister_netdevice_notifier(&vlan_notifier_block);
 err2:
@@ -745,8 +759,9 @@ static void __exit vlan_cleanup_module(void)
 		BUG_ON(!hlist_empty(&vlan_group_hash[i]));
 
 	unregister_pernet_gen_device(vlan_net_id, &vlan_net_ops);
-
 	synchronize_net();
+
+	vlan_gvrp_uninit();
 }
 
 module_init(vlan_proto_init);
diff --git a/net/8021q/vlan.h b/net/8021q/vlan.h
index 639e254..097b2e0 100644
--- a/net/8021q/vlan.h
+++ b/net/8021q/vlan.h
@@ -37,6 +37,22 @@ void vlan_setup(struct net_device *dev);
 int register_vlan_dev(struct net_device *dev);
 void unregister_vlan_dev(struct net_device *dev);
 
+#ifdef CONFIG_VLAN_8021Q_GVRP
+extern int vlan_gvrp_request_join(const struct net_device *dev);
+extern void vlan_gvrp_request_leave(const struct net_device *dev);
+extern int vlan_gvrp_init_applicant(struct net_device *dev);
+extern void vlan_gvrp_uninit_applicant(struct net_device *dev);
+extern int vlan_gvrp_init(void);
+extern void vlan_gvrp_uninit(void);
+#else
+static inline int vlan_gvrp_request_join(const struct net_device *dev) { return 0; }
+static inline void vlan_gvrp_request_leave(const struct net_device *dev) {}
+static inline int vlan_gvrp_init_applicant(struct net_device *dev) { return 0; }
+static inline void vlan_gvrp_uninit_applicant(struct net_device *dev) {}
+static inline int vlan_gvrp_init(void) { return 0; }
+static inline void vlan_gvrp_uninit(void) {}
+#endif
+
 int vlan_netlink_init(void);
 void vlan_netlink_fini(void);
 
diff --git a/net/8021q/vlan_dev.c b/net/8021q/vlan_dev.c
index 76c665c..a0617bf 100644
--- a/net/8021q/vlan_dev.c
+++ b/net/8021q/vlan_dev.c
@@ -512,10 +512,17 @@ int vlan_dev_change_flags(const struct net_device *dev, u32 flags, u32 mask)
 	struct vlan_dev_info *vlan = vlan_dev_info(dev);
 	u32 old_flags = vlan->flags;
 
-	if (mask & ~VLAN_FLAG_REORDER_HDR)
+	if (mask & ~(VLAN_FLAG_REORDER_HDR | VLAN_FLAG_GVRP))
 		return -EINVAL;
 
 	vlan->flags = (old_flags & ~mask) | (flags & mask);
+
+	if (netif_running(dev) && (vlan->flags ^ old_flags) & VLAN_FLAG_GVRP) {
+		if (vlan->flags & VLAN_FLAG_GVRP)
+			vlan_gvrp_request_join(dev);
+		else
+			vlan_gvrp_request_leave(dev);
+	}
 	return 0;
 }
 
@@ -550,12 +557,19 @@ static int vlan_dev_open(struct net_device *dev)
 	if (dev->flags & IFF_PROMISC)
 		dev_set_promiscuity(real_dev, 1);
 
+	if (vlan->flags & VLAN_FLAG_GVRP)
+		vlan_gvrp_request_join(dev);
+
 	return 0;
 }
 
 static int vlan_dev_stop(struct net_device *dev)
 {
-	struct net_device *real_dev = vlan_dev_info(dev)->real_dev;
+	struct vlan_dev_info *vlan = vlan_dev_info(dev);
+	struct net_device *real_dev = vlan->real_dev;
+
+	if (vlan->flags & VLAN_FLAG_GVRP)
+		vlan_gvrp_request_leave(dev);
 
 	dev_mc_unsync(real_dev, dev);
 	dev_unicast_unsync(real_dev, dev);
diff --git a/net/8021q/vlan_gvrp.c b/net/8021q/vlan_gvrp.c
new file mode 100644
index 0000000..db97816
--- /dev/null
+++ b/net/8021q/vlan_gvrp.c
@@ -0,0 +1,66 @@
+/*
+ * 	IEEE 802.1Q GARP VLAN Registration Protocol (GVRP)
+ *
+ * 	Copyright (c) 2008 Patrick McHardy <kaber@trash.net>
+ *
+ *	This program is free software; you can redistribute it and/or
+ *	modify it under the terms of the GNU General Public License
+ *	version 2 as published by the Free Software Foundation.
+ */
+#include <linux/types.h>
+#include <linux/if_vlan.h>
+#include <net/garp.h>
+#include "vlan.h"
+
+#define GARP_GVRP_ADDRESS	{ 0x01, 0x80, 0xc2, 0x00, 0x00, 0x21 }
+
+enum gvrp_attributes {
+	GVRP_ATTR_INVALID,
+	GVRP_ATTR_VID,
+	__GVRP_ATTR_MAX
+};
+#define GVRP_ATTR_MAX	(__GVRP_ATTR_MAX - 1)
+
+static struct garp_application vlan_gvrp_app __read_mostly = {
+	.proto.group_address	= GARP_GVRP_ADDRESS,
+	.maxattr		= GVRP_ATTR_MAX,
+	.type			= GARP_APPLICATION_GVRP,
+};
+
+int vlan_gvrp_request_join(const struct net_device *dev)
+{
+	const struct vlan_dev_info *vlan = vlan_dev_info(dev);
+	__be16 vid = htons(vlan->vlan_id);
+
+	return garp_request_join(vlan->real_dev, &vlan_gvrp_app,
+				 &vid, sizeof(vid), GVRP_ATTR_VID);
+}
+
+void vlan_gvrp_request_leave(const struct net_device *dev)
+{
+	const struct vlan_dev_info *vlan = vlan_dev_info(dev);
+	__be16 vid = htons(vlan->vlan_id);
+
+	garp_request_leave(vlan->real_dev, &vlan_gvrp_app,
+			   &vid, sizeof(vid), GVRP_ATTR_VID);
+}
+
+int vlan_gvrp_init_applicant(struct net_device *dev)
+{
+	return garp_init_applicant(dev, &vlan_gvrp_app);
+}
+
+void vlan_gvrp_uninit_applicant(struct net_device *dev)
+{
+	garp_uninit_applicant(dev, &vlan_gvrp_app);
+}
+
+int __init vlan_gvrp_init(void)
+{
+	return garp_register_application(&vlan_gvrp_app);
+}
+
+void vlan_gvrp_uninit(void)
+{
+	garp_unregister_application(&vlan_gvrp_app);
+}
diff --git a/net/8021q/vlan_netlink.c b/net/8021q/vlan_netlink.c
index fd7cb19..e9c91dc 100644
--- a/net/8021q/vlan_netlink.c
+++ b/net/8021q/vlan_netlink.c
@@ -59,7 +59,8 @@ static int vlan_validate(struct nlattr *tb[], struct nlattr *data[])
 	}
 	if (data[IFLA_VLAN_FLAGS]) {
 		flags = nla_data(data[IFLA_VLAN_FLAGS]);
-		if ((flags->flags & flags->mask) & ~VLAN_FLAG_REORDER_HDR)
+		if ((flags->flags & flags->mask) &
+		    ~(VLAN_FLAG_REORDER_HDR | VLAN_FLAG_GVRP))
 			return -EINVAL;
 	}
 

Re: net 01/06: Add STP demux layer

[Stephen Hemminger]

On Thu,  3 Jul 2008 19:04:42 +0200 (MEST)
Patrick McHardy <kaber@trash.net> wrote:

> net: Add STP demux layer
> 
> Add small STP demux layer for demuxing STP PDUs based on MAC address.
> This is needed to run both GARP and STP in parallel (or even load the
> modules) since both use LLC_SAP_BSPAN.
> 
> Signed-off-by: Patrick McHardy <kaber@trash.net>
> 
> ---
> commit 6c0be0e52a154f1991c221a6f60c6b9be71a10c8
> tree 6dcdf3d929edeae8463c6957e688f1e91ca6cfe5
> parent 44d28ab19c64d095314ac66f765d0c747519f4ed
> author Patrick McHardy <kaber@trash.net> Thu, 03 Jul 2008 18:05:56 +0200
> committer Patrick McHardy <kaber@trash.net> Thu, 03 Jul 2008 18:05:56 +0200
> 
>  include/net/stp.h |   14 +++++++
>  net/802/Kconfig   |    3 ++
>  net/802/Makefile  |    1 +
>  net/802/stp.c     |  102 +++++++++++++++++++++++++++++++++++++++++++++++++++++
>  net/Kconfig       |    1 +
>  5 files changed, 121 insertions(+), 0 deletions(-)
> 
> diff --git a/include/net/stp.h b/include/net/stp.h
> new file mode 100644
> index 0000000..ad447f1
> --- /dev/null
> +++ b/include/net/stp.h
> @@ -0,0 +1,14 @@
> +#ifndef _NET_STP_H
> +#define _NET_STP_H
> +
> +struct stp_proto {
> +	unsigned char	group_address[ETH_ALEN];
> +	void		(*rcv)(const struct stp_proto *, struct sk_buff *,
> +			       struct net_device *);
> +	void		*data;
> +};
> +
> +extern int stp_proto_register(const struct stp_proto *proto);
> +extern void stp_proto_unregister(const struct stp_proto *proto);
> +
> +#endif /* _NET_STP_H */
> diff --git a/net/802/Kconfig b/net/802/Kconfig
> new file mode 100644
> index 0000000..01cb094
> --- /dev/null
> +++ b/net/802/Kconfig
> @@ -0,0 +1,3 @@
> +config STP
> +	tristate
> +	select LLC
> diff --git a/net/802/Makefile b/net/802/Makefile
> index 68569ff..c441d89 100644
> --- a/net/802/Makefile
> +++ b/net/802/Makefile
> @@ -10,3 +10,4 @@ obj-$(CONFIG_FDDI)	+=                 fddi.o
>  obj-$(CONFIG_HIPPI)	+=                 hippi.o
>  obj-$(CONFIG_IPX)	+= p8022.o psnap.o p8023.o
>  obj-$(CONFIG_ATALK)	+= p8022.o psnap.o
> +obj-$(CONFIG_STP)	+= stp.o
> diff --git a/net/802/stp.c b/net/802/stp.c
> new file mode 100644
> index 0000000..0b7a244
> --- /dev/null
> +++ b/net/802/stp.c
> @@ -0,0 +1,102 @@
> +/*
> + *	STP SAP demux
> + *
> + *	Copyright (c) 2008 Patrick McHardy <kaber@trash.net>
> + *
> + *	This program is free software; you can redistribute it and/or
> + *	modify it under the terms of the GNU General Public License
> + *	version 2 as published by the Free Software Foundation.
> + */
> +#include <linux/mutex.h>
> +#include <linux/skbuff.h>
> +#include <linux/etherdevice.h>
> +#include <linux/llc.h>
> +#include <net/llc.h>
> +#include <net/llc_pdu.h>
> +#include <net/stp.h>
> +
> +/* 01:80:c2:00:00:20 - 01:80:c2:00:00:2F */
> +#define GARP_ADDR_MIN	0x20
> +#define GARP_ADDR_MAX	0x2F
> +#define GARP_ADDR_RANGE	(GARP_ADDR_MAX - GARP_ADDR_MIN)
> +
> +static const struct stp_proto *garp_protos[GARP_ADDR_RANGE + 1] __read_mostly;
> +static const struct stp_proto *stp_proto __read_mostly;
> +
> +static struct llc_sap *sap __read_mostly;
> +static unsigned int sap_registered;
> +static DEFINE_MUTEX(stp_proto_mutex);
> +
> +/* Called under rcu_read_lock from LLC */
> +static int stp_pdu_rcv(struct sk_buff *skb, struct net_device *dev,
> +		       struct packet_type *pt, struct net_device *orig_dev)
> +{
> +	const struct ethhdr *eh = eth_hdr(skb);
> +	const struct llc_pdu_un *pdu = llc_pdu_un_hdr(skb);
> +	const struct stp_proto *proto;
> +
> +	if (pdu->ssap != LLC_SAP_BSPAN ||
> +	    pdu->dsap != LLC_SAP_BSPAN ||
> +	    pdu->ctrl_1 != LLC_PDU_TYPE_U)
> +		goto err;
> +
> +	if (eh->h_dest[5] >= GARP_ADDR_MIN && eh->h_dest[5] <= GARP_ADDR_MAX) {
> +		proto = rcu_dereference(garp_protos[eh->h_dest[5] -
> +						    GARP_ADDR_MIN]);
> +		if (proto &&
> +		    compare_ether_addr(eh->h_dest, proto->group_address))
> +			goto err;
> +	} else
> +		proto = rcu_dereference(stp_proto);
> +
> +	if (!proto)
> +		goto err;
> +
> +	proto->rcv(proto, skb, dev);
> +	return 0;
> +
> +err:
> +	kfree_skb(skb);
> +	return 0;
> +}
> +
> +int stp_proto_register(const struct stp_proto *proto)
> +{
> +	int err = 0;
> +
> +	mutex_lock(&stp_proto_mutex);
> +	if (sap_registered++ == 0) {
> +		sap = llc_sap_open(LLC_SAP_BSPAN, stp_pdu_rcv);
> +		if (!sap) {
> +			err = -ENOMEM;
> +			goto out;
> +		}
> +	}
> +	if (is_zero_ether_addr(proto->group_address))
> +		rcu_assign_pointer(stp_proto, proto);
> +	else
> +		rcu_assign_pointer(garp_protos[proto->group_address[5] -
> +					       GARP_ADDR_MIN], proto);
> +out:
> +	mutex_unlock(&stp_proto_mutex);
> +	return err;
> +}
> +EXPORT_SYMBOL_GPL(stp_proto_register);
> +
> +void stp_proto_unregister(const struct stp_proto *proto)
> +{
> +	mutex_lock(&stp_proto_mutex);
> +	if (is_zero_ether_addr(proto->group_address))
> +		rcu_assign_pointer(stp_proto, NULL);
> +	else
> +		rcu_assign_pointer(garp_protos[proto->group_address[5] -
> +					       GARP_ADDR_MIN], NULL);
> +	synchronize_rcu();
> +
> +	if (--sap_registered == 0)
> +		llc_sap_put(sap);
> +	mutex_unlock(&stp_proto_mutex);
> +}
> +EXPORT_SYMBOL_GPL(stp_proto_unregister);
> +
> +MODULE_LICENSE("GPL");
> diff --git a/net/Kconfig b/net/Kconfig
> index acbf7c6..b986687 100644
> --- a/net/Kconfig
> +++ b/net/Kconfig
> @@ -181,6 +181,7 @@ source "net/dccp/Kconfig"
>  source "net/sctp/Kconfig"
>  source "net/tipc/Kconfig"
>  source "net/atm/Kconfig"
> +source "net/802/Kconfig"
>  source "net/bridge/Kconfig"
>  source "net/8021q/Kconfig"
>  source "net/decnet/Kconfig"

Why not just have LLC handle multiple registrations for same SAP?

Re: net 01/06: Add STP demux layer

[Patrick McHardy]

Stephen Hemminger wrote:
> On Thu,  3 Jul 2008 19:04:42 +0200 (MEST)
> Patrick McHardy <kaber@trash.net> wrote:
> 
>> net: Add STP demux layer
>>
>> Add small STP demux layer for demuxing STP PDUs based on MAC address.
>> This is needed to run both GARP and STP in parallel (or even load the
>> modules) since both use LLC_SAP_BSPAN.
>>
>>
> Why not just have LLC handle multiple registrations for same SAP?


I considered that, but it looks like a lot more work without a
clear advantage - nothing else seems to need it.

Re: net 00/06: add GARP/GVRP support

[David Miller]

From: Patrick McHardy <kaber@trash.net>
Date: Thu,  3 Jul 2008 19:04:41 +0200 (MEST)

> These patches add an implementation of the GARP (Generic Attribute
> Registration Protocol) applicant-only participant, which can be used
> to dynamically register VLANs and multicast addresses with switches.
> Applicant-only means it can only be used as a client, not on a bridge.
> Also included is a GVRP (GARP VLAN Registration Protocol) application
> on top.

These look good, applied and pushed out to net-next-2.6, thanks!