Skip to contentSkip to conversation

RFC: [PATCH] bridge vlan integration

Bridge David Kimdon 5 participants 10 messages Sep 11, 2006

Subsystems (3)
Networking Network drivers Ethernet bridge
lore
Share & tools
lore

David Kimdon

author
#0 lore raw

Hi,

The attached patches enables the bridge to filter and forward packets according to their IEEE 802.1q headers. The goals behind this change include :

- Enable running STP on 802.1q tagged networks. STP packets must be untagged. It isn't obvious how else to enable STP with the current bridge and vlan code.
- Add native support for an untagged vlan. Currently an untagged vlan can be implimented using ebtables or similar.
- On devices bridging a large number of interfaces across various vlans this significantly simplifies configuration and, depending on configuration, can improve performance.

Comments appreciated,

David

From: Simon Barber <[email protected]>

Signed-off-by Simon Barber <[email protected]>
Signed-off-by David Kimdon <[email protected]>
include/linux/if_bridge.h +10 -0
Index: wireless-dev/include/linux/if_bridge.h
===================================================================
@@ -44,6 +44,10 @@
 #define BRCTL_SET_PORT_PRIORITY 16
 #define BRCTL_SET_PATH_COST 17
 #define BRCTL_GET_FDB_ENTRIES 18
+#define BRCTL_SET_PORT_UNTAGGED_VLAN 19
+#define BRCTL_ADD_PORT_VLAN 20
+#define BRCTL_DEL_PORT_VLAN 21
+#define BRCTL_GET_PORT_VLAN_INFO 22
 
 #define BR_STATE_DISABLED 0
 #define BR_STATE_LISTENING 1

@@ -91,6 +95,12 @@ struct __port_info
 	__u32 hold_timer_value;
 };
 
+struct __vlan_info
+{
+	__u32 untagged;
+	__u8 filter[4096/8];
+};
+
 struct __fdb_entry
 {
 	__u8 mac_addr[6];
include/linux/skbuff.h +3 -0
Index: wireless-dev/include/linux/skbuff.h
===================================================================
@@ -296,6 +296,9 @@ struct sk_buff {
 #endif
 	__u32			nfmark;
 #endif /* CONFIG_NETFILTER */
+#ifdef CONFIG_BRIDGE_VLAN
+	unsigned int 		vlan;
+#endif
 #ifdef CONFIG_NET_SCHED
 	__u16			tc_index;	/* traffic control index */
 #ifdef CONFIG_NET_CLS_ACT
net/bridge/br_forward.c +14 -1
Index: wireless-dev/net/bridge/br_forward.c
===================================================================
@@ -24,7 +24,16 @@
 static inline int should_deliver(const struct net_bridge_port *p, 
 				 const struct sk_buff *skb)
 {
-	return (skb->dev != p->dev && p->state == BR_STATE_FORWARDING);
+	if (skb->dev == p->dev ||
+	    p->state != BR_STATE_FORWARDING)
+		return 0;
+
+#ifdef CONFIG_BRIDGE_VLAN
+	if (skb->vlan && br_vlan_filter(skb, &p->vlan))
+		return 0;
+#endif
+
+	return 1;
 }
 
 static inline unsigned packet_length(const struct sk_buff *skb)

@@ -47,6 +56,10 @@ int br_dev_queue_push_xmit(struct sk_buf
 		{
 			skb_push(skb, ETH_HLEN);
 
+	 		if (br_vlan_output_frame(&skb,
+ 						 skb->dev->br_port->vlan.untagged))
+ 				return 0;
+
 			dev_queue_xmit(skb);
 		}
 	}
net/bridge/br_if.c +2 -0
Index: wireless-dev/net/bridge/br_if.c
===================================================================
@@ -227,6 +227,7 @@ static struct net_device *new_bridge_dev
 	INIT_LIST_HEAD(&br->age_list);
 
 	br_stp_timer_init(br);
+	br_vlan_init(&br->vlan);
 
 	return dev;
 }

@@ -278,6 +279,7 @@ static struct net_bridge_port *new_nbp(s
 	p->state = BR_STATE_DISABLED;
 	INIT_WORK(&p->carrier_check, port_carrier_check, dev);
 	br_stp_port_timer_init(p);
+	br_vlan_init(&p->vlan);
 
 	kobject_init(&p->kobj);
 	kobject_set_name(&p->kobj, SYSFS_BRIDGE_PORT_ATTR);
net/bridge/br_input.c +12 -0
Index: wireless-dev/net/bridge/br_input.c
===================================================================
@@ -26,12 +26,20 @@ static void br_pass_frame_up(struct net_
 {
 	struct net_device *indev;
 
+	if (br_vlan_filter(skb, &br->vlan)) {
+		kfree_skb(skb);
+		return;
+	}
+
 	br->statistics.rx_packets++;
 	br->statistics.rx_bytes += skb->len;
 
 	indev = skb->dev;
 	skb->dev = br->dev;
 
+	if (br_vlan_output_frame(&skb, br->vlan.untagged))
+		return;
+
 	NF_HOOK(PF_BRIDGE, NF_BR_LOCAL_IN, skb, indev, NULL,
 		netif_receive_skb);
 }

@@ -136,6 +144,10 @@ int br_handle_frame(struct net_bridge_po
 	}
 
 	if (p->state == BR_STATE_FORWARDING || p->state == BR_STATE_LEARNING) {
+		if (br_vlan_input_frame(skb, &p->vlan)) {
+			return 1;
+		}
+
 		if (br_should_route_hook) {
 			if (br_should_route_hook(pskb)) 
 				return 0;
net/bridge/br_ioctl.c +12 -0
Index: wireless-dev/net/bridge/br_ioctl.c
===================================================================
@@ -302,6 +302,18 @@ static int old_dev_ioctl(struct net_devi
 	case BRCTL_GET_FDB_ENTRIES:
 		return get_fdb_entries(br, (void __user *)args[1], 
 				       args[2], args[3]);
+
+#ifdef CONFIG_BRIDGE_VLAN
+	case BRCTL_SET_PORT_UNTAGGED_VLAN:
+		return br_vlan_set_untagged(br, args[1], args[2]);
+
+	case BRCTL_ADD_PORT_VLAN:
+	case BRCTL_DEL_PORT_VLAN:
+		return br_vlan_set_filter(br, args[0], args[1], args[2]);
+
+	case BRCTL_GET_PORT_VLAN_INFO:
+		return br_vlan_get_info(br, (void *)args[1], args[2]);
+#endif
 	}
 
 	return -EOPNOTSUPP;
net/bridge/br_private.h +42 -0

64 lines not on this page · raw message

net/bridge/br_vlan.c new file +203 -0

203 lines not on this page · raw message

net/bridge/Makefile +2 -0
Index: wireless-dev/net/bridge/Makefile
===================================================================
@@ -12,4 +12,6 @@ bridge-$(CONFIG_SYSFS) += br_sysfs_if.o 
 
 bridge-$(CONFIG_BRIDGE_NETFILTER) += br_netfilter.o
 
+bridge-$(CONFIG_BRIDGE_VLAN) += br_vlan.o
+
 obj-$(CONFIG_BRIDGE_NF_EBTABLES) += netfilter/
net/core/skbuff.c +6 -0
Index: wireless-dev/net/core/skbuff.c
===================================================================
@@ -486,6 +486,9 @@ struct sk_buff *skb_clone(struct sk_buff
 	nf_bridge_get(skb->nf_bridge);
 #endif
 #endif /*CONFIG_NETFILTER*/
+#ifdef CONFIG_BRIDGE_VLAN
+	C(vlan);
+#endif
 #ifdef CONFIG_NET_SCHED
 	C(tc_index);
 #ifdef CONFIG_NET_CLS_ACT

@@ -550,6 +553,9 @@ static void copy_skb_header(struct sk_bu
 	nf_bridge_get(old->nf_bridge);
 #endif
 #endif
+#ifdef CONFIG_BRIDGE_VLAN
+	new->vlan	= old->vlan;
+#endif
 #ifdef CONFIG_NET_SCHED
 #ifdef CONFIG_NET_CLS_ACT
 	new->tc_verd = old->tc_verd;
net/bridge/Kconfig +10 -0
Index: wireless-dev/net/bridge/Kconfig
===================================================================
@@ -30,3 +30,13 @@ config BRIDGE
 	  will be called bridge.
 
 	  If unsure, say N.
+
+config BRIDGE_VLAN
+	bool "802.1Q bridge support"
+	depends on BRIDGE
+	---help---
+	  If you say Y here, then your bridge will be able to filter and
+	  forward packets according to their IEEE 802.1Q headers.
+
+	  If unsure, say N.
+
net/bridge/br_device.c +3 -0
Index: wireless-dev/net/bridge/br_device.c
===================================================================
@@ -34,6 +34,9 @@ int br_dev_xmit(struct sk_buff *skb, str
 	const unsigned char *dest = skb->data;
 	struct net_bridge_fdb_entry *dst;
 
+	if (br_vlan_input_frame(skb, &br->vlan))
+		return 0;
+
 	br->statistics.tx_packets++;
 	br->statistics.tx_bytes += skb->len;
author
#1 lore raw

Hi,

Here is the corresponding patch to brctl.

-David

brctl/brctl_cmd.c +107 -0
Index: bridge-utils-1.1/brctl/brctl_cmd.c
===================================================================
@@ -390,6 +390,105 @@ static int br_cmd_showmacs(int argc, cha
 	return 0;
 }
 
+static int br_cmd_setuntagged(int argc, char*const* argv)
+{
+	int vlan_id, err;
+
+	if (!strcmp(argv[3], "none"))
+		vlan_id = 0;
+	else	if (sscanf(argv[3], "%i", &vlan_id) != 1 ||
+			vlan_id < 1 || vlan_id > 4094) {
+		fprintf(stderr, "vlan id must be 'none' or from 1 to 4094\n");
+		return 1;
+	}
+
+	if (!strcmp(argv[2], "local"))
+		err = br_set_port_untagged_vlan(argv[1], 0, vlan_id);
+	else
+		err = br_set_port_untagged_vlan(argv[1], argv[2], vlan_id);
+
+	if (err)
+		fprintf(stderr, "set untagged vlan failed: %s\n",
+			strerror(err));
+	return err != 0;
+}
+
+static int br_cmd_vlanfilter(char*const* argv, int (*func)(const char *, const char *, unsigned int))
+{
+	int vlan_id, err;
+
+	if (!strcmp(argv[3], "all"))
+		vlan_id = 0;
+	else	if (sscanf(argv[3], "%i", &vlan_id) != 1 ||
+			vlan_id < 1 || vlan_id > 4094) {
+		fprintf(stderr, "vlan id must be 'all' or from 1 to 4094\n");
+		return 1;
+	}
+
+	err = func(argv[1], strcmp(argv[2], "local")?argv[2]:0, vlan_id);
+	if (err)
+		fprintf(stderr, "set vlan filter failed: %s\n",
+			strerror(err));
+	return err != 0;
+}
+
+static int br_cmd_addvlan(int argc, char*const* argv)
+{
+	return br_cmd_vlanfilter(argv, br_add_vlan);
+}
+
+static int br_cmd_delvlan(int argc, char*const* argv)
+{
+	return br_cmd_vlanfilter(argv, br_del_vlan);
+}
+
+int show_port_vlan(const char *brname, const char *port,
+					   void *arg )
+{
+	int err, count = 0, i;
+	struct vlan_info v;
+
+	err = br_get_port_vlan_info(brname, port, &v);
+	if (err)
+		fprintf(stderr, "get single vlan failed: %s\n",
+			strerror(err));
+
+	printf("%s\t", port?port:brname);
+
+	if (v.untagged)
+		printf("%d\t\t", v.untagged);
+	else
+		printf("none\t\t");
+
+	for(i = 1; i < 4095; i++)
+		if (v.filter[i / 8] & (1 << (i & 7)))
+			count++;
+
+	if (count == 4094)
+		printf("all\n");
+	else if (count == 0)
+		printf("none\n");
+	else {
+		int first = 1;
+
+		for(i = 1; i < 4095; i++)
+			if (v.filter[i / 8] & (1 << (i & 7))) {
+				printf("%s%d", first?"":", ", i);
+				first = 0;
+			}
+		printf("\n");
+	}
+
+	return err != 0;
+}
+
+static int br_cmd_showvlans(int argc, char*const* argv)
+{
+	printf("Port\tUntagged vlan\tEnabled Vlans\n");
+	show_port_vlan(argv[1], 0, 0);
+	return br_foreach_port(argv[1], show_port_vlan, 0);
+}
+
 static const struct command commands[] = {
 	{ 1, "addbr", br_cmd_addbr, "<bridge>\t\tadd bridge" },
 	{ 1, "delbr", br_cmd_delbr, "<bridge>\t\tdelete bridge" },

@@ -418,6 +517,14 @@ static const struct command commands[] =
 	  "<bridge>\t\tshow bridge stp info"},
 	{ 2, "stp", br_cmd_stp,
 	  "<bridge> {on|off}\tturn stp on/off" },
+ 	{ 3, "setuntagged", br_cmd_setuntagged,
+ 	  "<bridge> <port>|local <vlan_id>|none\tset untagged vlan id" },
+ 	{ 3, "addvlan", br_cmd_addvlan,
+ 	  "<bridge> <port>|local <vlan_id>|all\tenable specific or all vlans" },
+ 	{ 3, "delvlan", br_cmd_delvlan,
+ 	  "<bridge> <port>|local <vlan_id>|all\tdisable specific or all vlans" },
+ 	{ 1, "showvlans", br_cmd_showvlans,
+ 	  "<bridge>\t\tshow vlan info"},
 };
 
 const struct command *command_lookup(const char *cmd)
libbridge/libbridge.h +15 -0
Index: bridge-utils-1.1/libbridge/libbridge.h
===================================================================
@@ -76,6 +76,12 @@ struct port_info
 	struct timeval hold_timer_value;
 };
 
+struct vlan_info
+{
+	unsigned untagged;
+	unsigned char filter[4096/8];
+};
+
 extern int br_init(void);
 extern int br_refresh(void);
 extern void br_shutdown(void);

@@ -107,4 +113,13 @@ extern int br_set_path_cost(const char *
 			    int path_cost);
 extern int br_read_fdb(const char *br, struct fdb_entry *fdbs, 
 		       unsigned long skip, int num);
+
+int br_set_port_untagged_vlan(const char *brname, const char *port,
+		     unsigned int vlan_id);
+int br_add_vlan(const char *brname, const char *port,
+		     unsigned int vlan_id);
+int br_del_vlan(const char *brname, const char *port,
+		     unsigned int vlan_id);
+int br_get_port_vlan_info(const char *brname, const char *port,
+		     struct vlan_info *info);
 #endif
libbridge/libbridge_devif.c +126 -0

129 lines not on this page · raw message

#2 lore raw
In reply to David Kimdon
Quoted 15 lines · David Kimdon

Hi,

The attached patches enables the bridge to filter and forward packets according to their IEEE 802.1q headers. The goals behind this change include :

- Enable running STP on 802.1q tagged networks. STP packets must be untagged. It isn't obvious how else to enable STP with the current bridge and vlan code.
- Add native support for an untagged vlan. Currently an untagged vlan can be implimented using ebtables or similar.
- On devices bridging a large number of interfaces across various vlans this significantly simplifies configuration and, depending on configuration, can improve performance.

Quoted 3 lines

Comments appreciated,

David

David,

It looks like this code specifically ignores (which is OK for now) and clears (which is not OK) the frame's 802.1p priority. Have you tested priority-tagged frames to see if they are passed correctly? It seems like you should consider adding another field to the sk_buff structure so priority of the incoming frame can be tracked as well as add logic to br_vlan_output_frame and br_vlan_input_frame so the tag is saved. Something like this would probably be fine:

Signed-off-by Andy Gospodarek <[email protected]>
include/linux/skbuff.h 2006-09-11 14:43:27.000000000 -0400 renamed from include/linux/skbuff.h.gospo 2006-09-11 14:41:54.000000000 -0400 +0 -0
@@ -297,6 +297,7 @@ struct sk_buff {
       __u32                   nfmark;

#endif /* CONFIG_NETFILTER */
#ifdef CONFIG_BRIDGE_VLAN

+       unsigned int            vlan_priority;
       unsigned int            vlan;

#endif
#ifdef CONFIG_NET_SCHED

include/linux/if_vlan.h 2006-09-11 14:53:59.000000000 -0400 renamed from include/linux/if_vlan.h.gospo 2006-09-11 14:53:24.000000000 -0400 +0 -0
@@ -60,6 +60,7 @@ struct vlan_hdr {

};

#define VLAN_VID_MASK 0xfff
+#define VLAN_PRI_MASK 0xf000

/* found in socket.c */
extern void vlan_ioctl_set(int (*hook)(void __user *));

net/bridge/br_vlan.c 2006-09-11 14:52:06.000000000 -0400 renamed from net/bridge/br_vlan.c.gospo 2006-09-11 14:42:02.000000000 -0400 +3 -1
@@ -28,8 +28,10 @@ int br_vlan_input_frame(struct sk_buff *
               struct vlan_ethhdr *vhdr = (struct vlan_ethhdr *)(skb->mac.raw);
               unsigned short vlan_TCI = ntohs(vhdr->h_vlan_TCI);
               unsigned short vid = (vlan_TCI & VLAN_VID_MASK);
+               unsigned short pri = (vlan_TCI & VLAN_PRI_MASK);
 
               skb->vlan = vid ? vid : vlan->untagged;
+              skb->vlan_priority = pri ? pri : 0;
       }
 
       if (skb->vlan == 0)

@@ -100,7 +102,7 @@ int br_vlan_output_frame(struct sk_buff
                               ETH_ALEN * 2);
                       vhdr = (struct vlan_ethhdr *)skb->mac.raw;
                       vhdr->h_vlan_proto = htons(ETH_P_8021Q);
-                       vhdr->h_vlan_TCI = htons(skb->vlan);
+                       vhdr->h_vlan_TCI = htons(skb->vlan |

(skb->vlan_priority << 12));
               } else {
                       /* ensure VID is correct */
                       struct vlan_ethhdr *vhdr =
Quoted 512 lines

From: Simon Barber <[email protected]>

Signed-off-by Simon Barber <[email protected]>
Signed-off-by David Kimdon <[email protected]>

Index: wireless-dev/include/linux/if_bridge.h
===================================================================

include/linux/if_bridge.h +10 -0
@@ -44,6 +44,10 @@
 #define BRCTL_SET_PORT_PRIORITY 16
 #define BRCTL_SET_PATH_COST 17
 #define BRCTL_GET_FDB_ENTRIES 18
+#define BRCTL_SET_PORT_UNTAGGED_VLAN 19
+#define BRCTL_ADD_PORT_VLAN 20
+#define BRCTL_DEL_PORT_VLAN 21
+#define BRCTL_GET_PORT_VLAN_INFO 22
 
 #define BR_STATE_DISABLED 0
 #define BR_STATE_LISTENING 1

@@ -91,6 +95,12 @@ struct __port_info
        __u32 hold_timer_value;
 };
 
+struct __vlan_info
+{
+       __u32 untagged;
+       __u8 filter[4096/8];
+};
+
 struct __fdb_entry
 {
        __u8 mac_addr[6];

Index: wireless-dev/include/linux/skbuff.h
===================================================================

include/linux/skbuff.h +3 -0
@@ -296,6 +296,9 @@ struct sk_buff {
 #endif
        __u32                   nfmark;
 #endif /* CONFIG_NETFILTER */
+#ifdef CONFIG_BRIDGE_VLAN
+       unsigned int            vlan;
+#endif
 #ifdef CONFIG_NET_SCHED
        __u16                   tc_index;       /* traffic control index */
 #ifdef CONFIG_NET_CLS_ACT

Index: wireless-dev/net/bridge/br_forward.c
===================================================================

net/bridge/br_forward.c +14 -1
@@ -24,7 +24,16 @@
 static inline int should_deliver(const struct net_bridge_port *p,
                                 const struct sk_buff *skb)
 {
-       return (skb->dev != p->dev && p->state == BR_STATE_FORWARDING);
+       if (skb->dev == p->dev ||
+           p->state != BR_STATE_FORWARDING)
+               return 0;
+
+#ifdef CONFIG_BRIDGE_VLAN
+       if (skb->vlan && br_vlan_filter(skb, &p->vlan))
+               return 0;
+#endif
+
+       return 1;
 }
 
 static inline unsigned packet_length(const struct sk_buff *skb)

@@ -47,6 +56,10 @@ int br_dev_queue_push_xmit(struct sk_buf
                {
                        skb_push(skb, ETH_HLEN);
 
+                       if (br_vlan_output_frame(&skb,
+                                                skb->dev->br_port->vlan.untagged))
+                               return 0;
+
                        dev_queue_xmit(skb);
                }
        }

Index: wireless-dev/net/bridge/br_if.c
===================================================================

net/bridge/br_if.c +2 -0
@@ -227,6 +227,7 @@ static struct net_device *new_bridge_dev
        INIT_LIST_HEAD(&br->age_list);
 
        br_stp_timer_init(br);
+       br_vlan_init(&br->vlan);
 
        return dev;
 }

@@ -278,6 +279,7 @@ static struct net_bridge_port *new_nbp(s
        p->state = BR_STATE_DISABLED;
        INIT_WORK(&p->carrier_check, port_carrier_check, dev);
        br_stp_port_timer_init(p);
+       br_vlan_init(&p->vlan);
 
        kobject_init(&p->kobj);
        kobject_set_name(&p->kobj, SYSFS_BRIDGE_PORT_ATTR);

Index: wireless-dev/net/bridge/br_input.c
===================================================================

net/bridge/br_input.c +12 -0
@@ -26,12 +26,20 @@ static void br_pass_frame_up(struct net_
 {
        struct net_device *indev;
 
+       if (br_vlan_filter(skb, &br->vlan)) {
+               kfree_skb(skb);
+               return;
+       }
+
        br->statistics.rx_packets++;
        br->statistics.rx_bytes += skb->len;
 
        indev = skb->dev;
        skb->dev = br->dev;
 
+       if (br_vlan_output_frame(&skb, br->vlan.untagged))
+               return;
+
        NF_HOOK(PF_BRIDGE, NF_BR_LOCAL_IN, skb, indev, NULL,
                netif_receive_skb);
 }

@@ -136,6 +144,10 @@ int br_handle_frame(struct net_bridge_po
        }
 
        if (p->state == BR_STATE_FORWARDING || p->state == BR_STATE_LEARNING) {
+               if (br_vlan_input_frame(skb, &p->vlan)) {
+                       return 1;
+               }
+
                if (br_should_route_hook) {
                        if (br_should_route_hook(pskb))
                                return 0;

Index: wireless-dev/net/bridge/br_ioctl.c
===================================================================

net/bridge/br_ioctl.c +12 -0
@@ -302,6 +302,18 @@ static int old_dev_ioctl(struct net_devi
        case BRCTL_GET_FDB_ENTRIES:
                return get_fdb_entries(br, (void __user *)args[1],
                                       args[2], args[3]);
+
+#ifdef CONFIG_BRIDGE_VLAN
+       case BRCTL_SET_PORT_UNTAGGED_VLAN:
+               return br_vlan_set_untagged(br, args[1], args[2]);
+
+       case BRCTL_ADD_PORT_VLAN:
+       case BRCTL_DEL_PORT_VLAN:
+               return br_vlan_set_filter(br, args[0], args[1], args[2]);
+
+       case BRCTL_GET_PORT_VLAN_INFO:
+               return br_vlan_get_info(br, (void *)args[1], args[2]);
+#endif
        }
 
        return -EOPNOTSUPP;

Index: wireless-dev/net/bridge/br_private.h
===================================================================

net/bridge/br_private.h +42 -0
@@ -59,6 +59,14 @@ struct net_bridge_fdb_entry
        unsigned char                   is_static;
 };
 
+#ifdef CONFIG_BRIDGE_VLAN
+struct net_bridge_port_vlan
+{
+       int                             untagged;
+       u8                              filter[4096/8];
+};
+#endif
+
 struct net_bridge_port
 {
        struct net_bridge               *br;

@@ -84,6 +92,9 @@ struct net_bridge_port
        struct kobject                  kobj;
        struct work_struct              carrier_check;
        struct rcu_head                 rcu;
+#ifdef CONFIG_BRIDGE_VLAN
+       struct net_bridge_port_vlan     vlan;
+#endif
 };
 
 struct net_bridge

@@ -96,6 +107,9 @@ struct net_bridge
        struct hlist_head               hash[BR_HASH_SIZE];
        struct list_head                age_list;
        unsigned long                   feature_mask;
+#ifdef CONFIG_BRIDGE_VLAN
+       struct net_bridge_port_vlan     vlan;
+#endif
 
        /* STP */
        bridge_id                       designated_root;

@@ -258,4 +272,32 @@ extern void br_sysfs_delbr(struct net_de
 #define br_sysfs_delbr(dev)    do { } while(0)
 #endif /* CONFIG_SYSFS */
 
+#ifdef CONFIG_BRIDGE_VLAN
+#include <linux/if_vlan.h>
+
+static inline int br_vlan_filter(const struct sk_buff *skb,
+                                const struct net_bridge_port_vlan *vlan)
+{
+       return !(vlan->filter[skb->vlan / 8] & (1 << (skb->vlan & 7)));
+}
+
+/* br_vlan.c */
+extern int br_vlan_input_frame(struct sk_buff *skb,
+                              struct net_bridge_port_vlan *vlan);
+extern int br_vlan_output_frame(struct sk_buff **pskb, unsigned int untagged);
+extern void br_vlan_init(struct net_bridge_port_vlan *vlan);
+extern int br_vlan_set_untagged(struct net_bridge *br,
+                               unsigned int port, unsigned int vid);
+extern int br_vlan_set_filter(struct net_bridge *br,
+                             unsigned int cmd,
+                             unsigned int port, unsigned int vid);
+extern int br_vlan_get_info(struct net_bridge *br,
+                           void *user_mem, unsigned long port);
+#else
+
+#define br_vlan_filter(skb, vlan)              (0)
+#define br_vlan_input_frame(skb, vlan)         (0)
+#define br_vlan_output_frame(pskb, untagged)   (0)
+#define br_vlan_init(vlan)                     do { } while(0)
+#endif /* CONFIG_BRIDGE_VLAN */
 #endif

Index: wireless-dev/net/bridge/br_vlan.c
===================================================================

net/bridge/br_vlan.c new file +203 -0
@@ -0,0 +1,203 @@
+/*
+ *     VLAN support
+ *     Linux ethernet bridge
+ *
+ *     Authors:
+ *     Simon Barber                    <[email protected]>
+ *
+ *     This program is free software; you can redistribute it and/or
+ *     modify it under the terms of the GNU General Public License
+ *     as published by the Free Software Foundation; either version
+ *     2 of the License, or (at your option) any later version.
+ */
+
+#include <linux/kernel.h>
+#include <linux/netdevice.h>
+#include <linux/inetdevice.h>
+#include <linux/skbuff.h>
+#include <linux/if_bridge.h>
+#include <linux/netfilter_bridge.h>
+#include <asm/uaccess.h>
+#include "br_private.h"
+
+int br_vlan_input_frame(struct sk_buff *skb, struct net_bridge_port_vlan *vlan)
+{
+       if (skb->protocol != htons(ETH_P_8021Q)) {
+               skb->vlan = vlan->untagged;
+       } else {
+               struct vlan_ethhdr *vhdr = (struct vlan_ethhdr *)(skb->mac.raw);
+               unsigned short vlan_TCI = ntohs(vhdr->h_vlan_TCI);
+               unsigned short vid = (vlan_TCI & VLAN_VID_MASK);
+
+               skb->vlan = vid ? vid : vlan->untagged;
+       }
+
+       if (skb->vlan == 0)
+               goto err;
+
+       if (br_vlan_filter(skb, vlan))
+               goto err;
+
+       return 0;
+
+err:
+       kfree_skb(skb);
+       return 1;
+}
+
+int br_vlan_output_frame(struct sk_buff **pskb, unsigned int untagged)
+{
+       struct sk_buff *skb = *pskb;
+
+       if (skb->vlan == 0)     /* don't touch the frame */
+               return 0;
+
+       if (skb->vlan == untagged) {
+               /* frame should be untagged */
+               if (eth_hdr(skb)->h_proto == htons(ETH_P_8021Q)) {
+                       /* remove VLAN tag */
+                       if (skb_cloned(skb) || skb_shared(skb)) {
+                               struct sk_buff *new_skb;
+
+                               new_skb = skb_copy(skb, GFP_ATOMIC);
+                               kfree_skb(skb);
+                               if (!new_skb)
+                                       return 1;
+                               *pskb = skb = new_skb;
+                       }
+
+                       skb->mac.raw += VLAN_HLEN;
+                       memmove(skb->mac.raw, skb->data, ETH_ALEN * 2);
+                       skb_pull(skb, VLAN_HLEN);
+               }
+       } else {
+               /* frame should be tagged */
+               if (eth_hdr(skb)->h_proto != htons(ETH_P_8021Q)) {
+                       /* add VLAN tag */
+                       struct vlan_ethhdr *vhdr;
+                       if (skb_cloned(skb) || skb_shared(skb)) {
+                               struct sk_buff *new_skb;
+
+                               new_skb = skb_copy(skb, GFP_ATOMIC);
+                               kfree_skb(skb);
+                               if (!new_skb)
+                                       return 1;
+                               *pskb = skb = new_skb;
+                       }
+
+                       if (skb_headroom(skb) < VLAN_HLEN) {
+                               if (pskb_expand_head(skb, VLAN_HLEN, 0,
+                                                    GFP_ATOMIC)) {
+                                       kfree_skb(skb);
+                                       return 1;
+                               }
+                       }
+
+                       skb_push(skb, VLAN_HLEN);
+
+                       skb->mac.raw -= VLAN_HLEN;
+                       memmove(skb->mac.raw, skb->mac.raw + VLAN_HLEN,
+                               ETH_ALEN * 2);
+                       vhdr = (struct vlan_ethhdr *)skb->mac.raw;
+                       vhdr->h_vlan_proto = htons(ETH_P_8021Q);
+                       vhdr->h_vlan_TCI = htons(skb->vlan);
+               } else {
+                       /* ensure VID is correct */
+                       struct vlan_ethhdr *vhdr =
+                               (struct vlan_ethhdr *)skb->mac.raw;
+                       vhdr->h_vlan_TCI = (vhdr->h_vlan_TCI & htons(~VLAN_VID_MASK)) |
+                                           htons(skb->vlan);
+               }
+               // TODO: set priority in tag correctly
+       }
+
+       return 0;
+}
+
+void br_vlan_init(struct net_bridge_port_vlan *vlan)
+{
+       vlan->untagged = 1;
+       vlan->filter[0] = 1 << 1;
+}
+
+/* ioctl functions */
+int br_vlan_set_untagged(struct net_bridge *br,
+                        unsigned int port, unsigned int vid)
+{
+       struct net_bridge_port_vlan *vlan;
+
+       if (vid > 4094)
+               return -EINVAL;
+
+       if (port) {
+               struct net_bridge_port *p = br_get_port(br, port);
+
+               if (p == NULL)
+                       return -EINVAL;
+               vlan = &p->vlan;
+       } else {
+               vlan = &br->vlan;
+       }
+
+       vlan->untagged = vid;
+
+       return 0;
+}
+
+int br_vlan_set_filter(struct net_bridge *br,
+                      unsigned int cmd, unsigned int port, unsigned int vid)
+{
+       struct net_bridge_port_vlan *vlan;
+       int add = (cmd == BRCTL_ADD_PORT_VLAN);
+
+       if (vid > 4094)
+               return -EINVAL;
+
+       if (port) {
+               struct net_bridge_port *p = br_get_port(br, port);
+
+               if (p == NULL)
+                       return -EINVAL;
+               vlan = &p->vlan;
+       } else {
+               vlan = &br->vlan;
+       }
+
+       if (vid == 0) {
+               /* special case - add/del for all vlans */
+               memset(vlan->filter, add ? 255 : 0, 4096 / 8);
+               if (add) {
+                       vlan->filter[4095 / 8] &= ~(1 << (4095 & 7));
+               }
+       } else if (add)
+               vlan->filter[vid / 8] |= 1 << (vid & 7);
+       else
+               vlan->filter[vid / 8] &= ~(1 << (vid & 7));
+
+       return 0;
+}
+
+int br_vlan_get_info(struct net_bridge *br, void *user_mem, unsigned long port)
+{
+       struct net_bridge_port_vlan *vlan;
+       struct __vlan_info v;
+
+       if (port) {
+               struct net_bridge_port *p = br_get_port(br, port);
+
+               if (p == NULL)
+                       return -EINVAL;
+               vlan = &p->vlan;
+       } else {
+               vlan = &br->vlan;
+       }
+
+       memset(&v, 0, sizeof(v));
+       v.untagged = vlan->untagged;
+       memcpy(v.filter, vlan->filter, 4096 / 8);
+
+       if (copy_to_user((void __user *)user_mem, &v, sizeof(v)))
+               return -EFAULT;
+
+       return 0;
+}

Index: wireless-dev/net/bridge/Makefile
===================================================================

net/bridge/Makefile +2 -0
@@ -12,4 +12,6 @@ bridge-$(CONFIG_SYSFS) += br_sysfs_if.o
 
 bridge-$(CONFIG_BRIDGE_NETFILTER) += br_netfilter.o
 
+bridge-$(CONFIG_BRIDGE_VLAN) += br_vlan.o
+
 obj-$(CONFIG_BRIDGE_NF_EBTABLES) += netfilter/

Index: wireless-dev/net/core/skbuff.c
===================================================================

net/core/skbuff.c +6 -0
@@ -486,6 +486,9 @@ struct sk_buff *skb_clone(struct sk_buff
        nf_bridge_get(skb->nf_bridge);
 #endif
 #endif /*CONFIG_NETFILTER*/
+#ifdef CONFIG_BRIDGE_VLAN
+       C(vlan);
+#endif
 #ifdef CONFIG_NET_SCHED
        C(tc_index);
 #ifdef CONFIG_NET_CLS_ACT

@@ -550,6 +553,9 @@ static void copy_skb_header(struct sk_bu
        nf_bridge_get(old->nf_bridge);
 #endif
 #endif
+#ifdef CONFIG_BRIDGE_VLAN
+       new->vlan       = old->vlan;
+#endif
 #ifdef CONFIG_NET_SCHED
 #ifdef CONFIG_NET_CLS_ACT
        new->tc_verd = old->tc_verd;

Index: wireless-dev/net/bridge/Kconfig
===================================================================

net/bridge/Kconfig +10 -0
@@ -30,3 +30,13 @@ config BRIDGE
          will be called bridge.
 
          If unsure, say N.
+
+config BRIDGE_VLAN
+       bool "802.1Q bridge support"
+       depends on BRIDGE
+       ---help---
+         If you say Y here, then your bridge will be able to filter and
+         forward packets according to their IEEE 802.1Q headers.
+
+         If unsure, say N.
+

Index: wireless-dev/net/bridge/br_device.c
===================================================================

net/bridge/br_device.c +3 -0
@@ -34,6 +34,9 @@ int br_dev_xmit(struct sk_buff *skb, str
        const unsigned char *dest = skb->data;
        struct net_bridge_fdb_entry *dst;
 
+       if (br_vlan_input_frame(skb, &br->vlan))
+               return 0;
+
        br->statistics.tx_packets++;
        br->statistics.tx_bytes += skb->len;

_______________________________________________
Bridge mailing list
[email protected]
https://lists.osdl.org/mailman/listinfo/bridge

#7 lore raw
In reply to David Kimdon

Have you tested this in a real world scenario?

I attempted to use a linux bridge to bridge a few vlans and found that it didn't work as expected. At first I thought it was a performance problem relating to the linux bridge code because it was dropping packets, but I believe the problem is inherent in using a bridge in a vlan environment.

The problem lies in the fact that the switch connected to the bridge (which has the vlans tagged) gets confused as to which port that particular mac address is out of. Take the following scenario: (based on my understanding of how things should work)

1. Computer a on vlan 6 wants to send out a packet to computer b on vlan 7.
2. Computer a sends out an arp request for computer b. (core switch records computer A as being on port A14)
3. The arp request reaches the linux bridge, which doesn't know computer b yet, so it send out an arp request on all other vlans its bridging. (core switch records computer A as being on port A1 because it saw a packet from that mac on port A1)
4. Computer b responds to the arp request. (core switch records computer b as being on port A10)
5. bridge forwards the arp response out A10 to computer a (core switch records computer b as being on A10)

As you can see, the core switch (and I've seen this happen on our HP 5300 switch) gets confused as to which port to send out unicast packets -- hence sometimes when it _thinks_ it sent out a packet to the computer, the computer didn't actually receive it and the packet needs to be resent multiple times until the mac cache times out on the core switch. Under other circumstances, the switch doesn't seem to get confused, so it appears to be flaky, not completely broken.

Is there something I could have done differently which would have avoided that problem? Does this patch somehow address that problem? Is my HP switch just inept and do most switches keep a mac-port table per vlan? (I doubt this, because the SNMP mib doesn't allow that to be represented.)

Ethan Sommer
UNIX Systems Administrator
Gustavus Adolphus College

Quoted 532 lines · David Kimdon

Hi,

The attached patches enables the bridge to filter and forward packets according to their IEEE 802.1q headers. The goals behind this change include :

- Enable running STP on 802.1q tagged networks. STP packets must be untagged. It isn't obvious how else to enable STP with the current bridge and vlan code.
- Add native support for an untagged vlan. Currently an untagged vlan can be implimented using ebtables or similar.
- On devices bridging a large number of interfaces across various vlans this significantly simplifies configuration and, depending on configuration, can improve performance.

Comments appreciated,

David

From: Simon Barber <[email protected]>

Signed-off-by Simon Barber <[email protected]>
Signed-off-by David Kimdon <[email protected]>

Index: wireless-dev/include/linux/if_bridge.h
===================================================================

include/linux/if_bridge.h +10 -0
@@ -44,6 +44,10 @@
 #define BRCTL_SET_PORT_PRIORITY 16
 #define BRCTL_SET_PATH_COST 17
 #define BRCTL_GET_FDB_ENTRIES 18
+#define BRCTL_SET_PORT_UNTAGGED_VLAN 19
+#define BRCTL_ADD_PORT_VLAN 20
+#define BRCTL_DEL_PORT_VLAN 21
+#define BRCTL_GET_PORT_VLAN_INFO 22
 
 #define BR_STATE_DISABLED 0
 #define BR_STATE_LISTENING 1

@@ -91,6 +95,12 @@ struct __port_info
 	__u32 hold_timer_value;
 };
 
+struct __vlan_info
+{
+	__u32 untagged;
+	__u8 filter[4096/8];
+};
+
 struct __fdb_entry
 {
 	__u8 mac_addr[6];

Index: wireless-dev/include/linux/skbuff.h
===================================================================

include/linux/skbuff.h +3 -0
@@ -296,6 +296,9 @@ struct sk_buff {
 #endif
 	__u32			nfmark;
 #endif /* CONFIG_NETFILTER */
+#ifdef CONFIG_BRIDGE_VLAN
+	unsigned int 		vlan;
+#endif
 #ifdef CONFIG_NET_SCHED
 	__u16			tc_index;	/* traffic control index */
 #ifdef CONFIG_NET_CLS_ACT

Index: wireless-dev/net/bridge/br_forward.c
===================================================================

net/bridge/br_forward.c +14 -1
@@ -24,7 +24,16 @@
 static inline int should_deliver(const struct net_bridge_port *p, 
 				 const struct sk_buff *skb)
 {
-	return (skb->dev != p->dev && p->state == BR_STATE_FORWARDING);
+	if (skb->dev == p->dev ||
+	    p->state != BR_STATE_FORWARDING)
+		return 0;
+
+#ifdef CONFIG_BRIDGE_VLAN
+	if (skb->vlan && br_vlan_filter(skb, &p->vlan))
+		return 0;
+#endif
+
+	return 1;
 }
 
 static inline unsigned packet_length(const struct sk_buff *skb)

@@ -47,6 +56,10 @@ int br_dev_queue_push_xmit(struct sk_buf
 		{
 			skb_push(skb, ETH_HLEN);
 
+	 		if (br_vlan_output_frame(&skb,
+ 						 skb->dev->br_port->vlan.untagged))
+ 				return 0;
+
 			dev_queue_xmit(skb);
 		}
 	}

Index: wireless-dev/net/bridge/br_if.c
===================================================================

net/bridge/br_if.c +2 -0
@@ -227,6 +227,7 @@ static struct net_device *new_bridge_dev
 	INIT_LIST_HEAD(&br->age_list);
 
 	br_stp_timer_init(br);
+	br_vlan_init(&br->vlan);
 
 	return dev;
 }

@@ -278,6 +279,7 @@ static struct net_bridge_port *new_nbp(s
 	p->state = BR_STATE_DISABLED;
 	INIT_WORK(&p->carrier_check, port_carrier_check, dev);
 	br_stp_port_timer_init(p);
+	br_vlan_init(&p->vlan);
 
 	kobject_init(&p->kobj);
 	kobject_set_name(&p->kobj, SYSFS_BRIDGE_PORT_ATTR);

Index: wireless-dev/net/bridge/br_input.c
===================================================================

net/bridge/br_input.c +12 -0
@@ -26,12 +26,20 @@ static void br_pass_frame_up(struct net_
 {
 	struct net_device *indev;
 
+	if (br_vlan_filter(skb, &br->vlan)) {
+		kfree_skb(skb);
+		return;
+	}
+
 	br->statistics.rx_packets++;
 	br->statistics.rx_bytes += skb->len;
 
 	indev = skb->dev;
 	skb->dev = br->dev;
 
+	if (br_vlan_output_frame(&skb, br->vlan.untagged))
+		return;
+
 	NF_HOOK(PF_BRIDGE, NF_BR_LOCAL_IN, skb, indev, NULL,
 		netif_receive_skb);
 }

@@ -136,6 +144,10 @@ int br_handle_frame(struct net_bridge_po
 	}
 
 	if (p->state == BR_STATE_FORWARDING || p->state == BR_STATE_LEARNING) {
+		if (br_vlan_input_frame(skb, &p->vlan)) {
+			return 1;
+		}
+
 		if (br_should_route_hook) {
 			if (br_should_route_hook(pskb)) 
 				return 0;

Index: wireless-dev/net/bridge/br_ioctl.c
===================================================================

net/bridge/br_ioctl.c +12 -0
@@ -302,6 +302,18 @@ static int old_dev_ioctl(struct net_devi
 	case BRCTL_GET_FDB_ENTRIES:
 		return get_fdb_entries(br, (void __user *)args[1], 
 				       args[2], args[3]);
+
+#ifdef CONFIG_BRIDGE_VLAN
+	case BRCTL_SET_PORT_UNTAGGED_VLAN:
+		return br_vlan_set_untagged(br, args[1], args[2]);
+
+	case BRCTL_ADD_PORT_VLAN:
+	case BRCTL_DEL_PORT_VLAN:
+		return br_vlan_set_filter(br, args[0], args[1], args[2]);
+
+	case BRCTL_GET_PORT_VLAN_INFO:
+		return br_vlan_get_info(br, (void *)args[1], args[2]);
+#endif
 	}
 
 	return -EOPNOTSUPP;

Index: wireless-dev/net/bridge/br_private.h
===================================================================

net/bridge/br_private.h +42 -0
@@ -59,6 +59,14 @@ struct net_bridge_fdb_entry
 	unsigned char			is_static;
 };
 
+#ifdef CONFIG_BRIDGE_VLAN
+struct net_bridge_port_vlan
+{
+	int				untagged;
+	u8				filter[4096/8];
+};
+#endif
+
 struct net_bridge_port
 {
 	struct net_bridge		*br;

@@ -84,6 +92,9 @@ struct net_bridge_port
 	struct kobject			kobj;
 	struct work_struct		carrier_check;
 	struct rcu_head			rcu;
+#ifdef CONFIG_BRIDGE_VLAN
+	struct net_bridge_port_vlan 	vlan;
+#endif
 };
 
 struct net_bridge

@@ -96,6 +107,9 @@ struct net_bridge
 	struct hlist_head		hash[BR_HASH_SIZE];
 	struct list_head		age_list;
 	unsigned long			feature_mask;
+#ifdef CONFIG_BRIDGE_VLAN
+	struct net_bridge_port_vlan 	vlan;
+#endif
 
 	/* STP */
 	bridge_id			designated_root;

@@ -258,4 +272,32 @@ extern void br_sysfs_delbr(struct net_de
 #define br_sysfs_delbr(dev)	do { } while(0)
 #endif /* CONFIG_SYSFS */
 
+#ifdef CONFIG_BRIDGE_VLAN
+#include <linux/if_vlan.h>
+
+static inline int br_vlan_filter(const struct sk_buff *skb,
+				 const struct net_bridge_port_vlan *vlan)
+{
+	return !(vlan->filter[skb->vlan / 8] & (1 << (skb->vlan & 7)));
+}
+
+/* br_vlan.c */
+extern int br_vlan_input_frame(struct sk_buff *skb,
+			       struct net_bridge_port_vlan *vlan);
+extern int br_vlan_output_frame(struct sk_buff **pskb, unsigned int untagged);
+extern void br_vlan_init(struct net_bridge_port_vlan *vlan);
+extern int br_vlan_set_untagged(struct net_bridge *br,
+				unsigned int port, unsigned int vid);
+extern int br_vlan_set_filter(struct net_bridge *br,
+			      unsigned int cmd,
+			      unsigned int port, unsigned int vid);
+extern int br_vlan_get_info(struct net_bridge *br,
+			    void *user_mem, unsigned long port);
+#else
+
+#define br_vlan_filter(skb, vlan)		(0)
+#define br_vlan_input_frame(skb, vlan)		(0)
+#define br_vlan_output_frame(pskb, untagged)	(0)
+#define br_vlan_init(vlan)			do { } while(0)
+#endif /* CONFIG_BRIDGE_VLAN */
 #endif

Index: wireless-dev/net/bridge/br_vlan.c
===================================================================

net/bridge/br_vlan.c new file +203 -0
@@ -0,0 +1,203 @@
+/*
+ *	VLAN support
+ *	Linux ethernet bridge
+ *
+ *	Authors:
+ *	Simon Barber			<[email protected]>
+ *
+ *	This program is free software; you can redistribute it and/or
+ *	modify it under the terms of the GNU General Public License
+ *	as published by the Free Software Foundation; either version
+ *	2 of the License, or (at your option) any later version.
+ */
+
+#include <linux/kernel.h>
+#include <linux/netdevice.h>
+#include <linux/inetdevice.h>
+#include <linux/skbuff.h>
+#include <linux/if_bridge.h>
+#include <linux/netfilter_bridge.h>
+#include <asm/uaccess.h>
+#include "br_private.h"
+
+int br_vlan_input_frame(struct sk_buff *skb, struct net_bridge_port_vlan *vlan)
+{
+	if (skb->protocol != htons(ETH_P_8021Q)) {
+		skb->vlan = vlan->untagged;
+	} else {
+		struct vlan_ethhdr *vhdr = (struct vlan_ethhdr *)(skb->mac.raw);
+		unsigned short vlan_TCI = ntohs(vhdr->h_vlan_TCI);
+		unsigned short vid = (vlan_TCI & VLAN_VID_MASK);
+
+		skb->vlan = vid ? vid : vlan->untagged;
+	}
+
+	if (skb->vlan == 0)
+		goto err;
+
+	if (br_vlan_filter(skb, vlan))
+		goto err;
+
+	return 0;
+
+err:
+	kfree_skb(skb);
+	return 1;
+}
+
+int br_vlan_output_frame(struct sk_buff **pskb, unsigned int untagged)
+{
+	struct sk_buff *skb = *pskb;
+
+	if (skb->vlan == 0)	/* don't touch the frame */
+		return 0;
+
+	if (skb->vlan == untagged) {
+		/* frame should be untagged */
+		if (eth_hdr(skb)->h_proto == htons(ETH_P_8021Q)) {
+			/* remove VLAN tag */
+			if (skb_cloned(skb) || skb_shared(skb)) {
+				struct sk_buff *new_skb;
+
+				new_skb = skb_copy(skb, GFP_ATOMIC);
+				kfree_skb(skb);
+				if (!new_skb)
+					return 1;
+				*pskb = skb = new_skb;
+			}
+
+			skb->mac.raw += VLAN_HLEN;
+			memmove(skb->mac.raw, skb->data, ETH_ALEN * 2);
+			skb_pull(skb, VLAN_HLEN);
+		}
+	} else {
+		/* frame should be tagged */
+		if (eth_hdr(skb)->h_proto != htons(ETH_P_8021Q)) {
+			/* add VLAN tag */
+			struct vlan_ethhdr *vhdr;
+			if (skb_cloned(skb) || skb_shared(skb)) {
+				struct sk_buff *new_skb;
+
+				new_skb = skb_copy(skb, GFP_ATOMIC);
+				kfree_skb(skb);
+				if (!new_skb)
+					return 1;
+				*pskb = skb = new_skb;
+			}
+
+			if (skb_headroom(skb) < VLAN_HLEN) {
+				if (pskb_expand_head(skb, VLAN_HLEN, 0,
+						     GFP_ATOMIC)) {
+					kfree_skb(skb);
+					return 1;
+				}
+			}
+
+			skb_push(skb, VLAN_HLEN);
+
+			skb->mac.raw -= VLAN_HLEN;
+			memmove(skb->mac.raw, skb->mac.raw + VLAN_HLEN,
+				ETH_ALEN * 2);
+			vhdr = (struct vlan_ethhdr *)skb->mac.raw;
+			vhdr->h_vlan_proto = htons(ETH_P_8021Q);
+			vhdr->h_vlan_TCI = htons(skb->vlan);
+		} else {
+			/* ensure VID is correct */
+			struct vlan_ethhdr *vhdr =
+				(struct vlan_ethhdr *)skb->mac.raw;
+			vhdr->h_vlan_TCI = (vhdr->h_vlan_TCI & htons(~VLAN_VID_MASK)) |
+					    htons(skb->vlan);
+		}
+		// TODO: set priority in tag correctly
+	}
+
+	return 0;
+}
+
+void br_vlan_init(struct net_bridge_port_vlan *vlan)
+{
+	vlan->untagged = 1;
+	vlan->filter[0] = 1 << 1;
+}
+
+/* ioctl functions */
+int br_vlan_set_untagged(struct net_bridge *br,
+			 unsigned int port, unsigned int vid)
+{
+	struct net_bridge_port_vlan *vlan;
+
+	if (vid > 4094)
+		return -EINVAL;
+
+	if (port) {
+		struct net_bridge_port *p = br_get_port(br, port);
+
+		if (p == NULL)
+			return -EINVAL;
+		vlan = &p->vlan;
+	} else {
+		vlan = &br->vlan;
+	}
+
+	vlan->untagged = vid;
+
+	return 0;
+}
+
+int br_vlan_set_filter(struct net_bridge *br,
+		       unsigned int cmd, unsigned int port, unsigned int vid)
+{
+	struct net_bridge_port_vlan *vlan;
+	int add = (cmd == BRCTL_ADD_PORT_VLAN);
+
+	if (vid > 4094)
+		return -EINVAL;
+
+	if (port) {
+		struct net_bridge_port *p = br_get_port(br, port);
+
+		if (p == NULL)
+			return -EINVAL;
+		vlan = &p->vlan;
+	} else {
+		vlan = &br->vlan;
+	}
+
+	if (vid == 0) {
+		/* special case - add/del for all vlans */
+		memset(vlan->filter, add ? 255 : 0, 4096 / 8);
+		if (add) {
+			vlan->filter[4095 / 8] &= ~(1 << (4095 & 7));
+		}
+	} else if (add)
+		vlan->filter[vid / 8] |= 1 << (vid & 7);
+	else
+		vlan->filter[vid / 8] &= ~(1 << (vid & 7));
+
+	return 0;
+}
+
+int br_vlan_get_info(struct net_bridge *br, void *user_mem, unsigned long port)
+{
+	struct net_bridge_port_vlan *vlan;
+	struct __vlan_info v;
+
+	if (port) {
+		struct net_bridge_port *p = br_get_port(br, port);
+
+		if (p == NULL)
+			return -EINVAL;
+		vlan = &p->vlan;
+	} else {
+		vlan = &br->vlan;
+	}
+
+	memset(&v, 0, sizeof(v));
+	v.untagged = vlan->untagged;
+	memcpy(v.filter, vlan->filter, 4096 / 8);
+
+	if (copy_to_user((void __user *)user_mem, &v, sizeof(v)))
+		return -EFAULT;
+
+	return 0;
+}

Index: wireless-dev/net/bridge/Makefile
===================================================================

net/bridge/Makefile +2 -0
@@ -12,4 +12,6 @@ bridge-$(CONFIG_SYSFS) += br_sysfs_if.o 
 
 bridge-$(CONFIG_BRIDGE_NETFILTER) += br_netfilter.o
 
+bridge-$(CONFIG_BRIDGE_VLAN) += br_vlan.o
+
 obj-$(CONFIG_BRIDGE_NF_EBTABLES) += netfilter/

Index: wireless-dev/net/core/skbuff.c
===================================================================

net/core/skbuff.c +6 -0
@@ -486,6 +486,9 @@ struct sk_buff *skb_clone(struct sk_buff
 	nf_bridge_get(skb->nf_bridge);
 #endif
 #endif /*CONFIG_NETFILTER*/
+#ifdef CONFIG_BRIDGE_VLAN
+	C(vlan);
+#endif
 #ifdef CONFIG_NET_SCHED
 	C(tc_index);
 #ifdef CONFIG_NET_CLS_ACT

@@ -550,6 +553,9 @@ static void copy_skb_header(struct sk_bu
 	nf_bridge_get(old->nf_bridge);
 #endif
 #endif
+#ifdef CONFIG_BRIDGE_VLAN
+	new->vlan	= old->vlan;
+#endif
 #ifdef CONFIG_NET_SCHED
 #ifdef CONFIG_NET_CLS_ACT
 	new->tc_verd = old->tc_verd;

Index: wireless-dev/net/bridge/Kconfig
===================================================================

net/bridge/Kconfig +10 -0
@@ -30,3 +30,13 @@ config BRIDGE
 	  will be called bridge.
 
 	  If unsure, say N.
+
+config BRIDGE_VLAN
+	bool "802.1Q bridge support"
+	depends on BRIDGE
+	---help---
+	  If you say Y here, then your bridge will be able to filter and
+	  forward packets according to their IEEE 802.1Q headers.
+
+	  If unsure, say N.
+

Index: wireless-dev/net/bridge/br_device.c
===================================================================

net/bridge/br_device.c +3 -0
@@ -34,6 +34,9 @@ int br_dev_xmit(struct sk_buff *skb, str
 	const unsigned char *dest = skb->data;
 	struct net_bridge_fdb_entry *dst;
 
+	if (br_vlan_input_frame(skb, &br->vlan))
+		return 0;
+
 	br->statistics.tx_packets++;
 	br->statistics.tx_bytes += skb->len;

_______________________________________________
Bridge mailing list
[email protected]
https://lists.osdl.org/mailman/listinfo/bridge

Signature · 5 lines
Ethan Sommer
UNIX Systems Administrator
Gustavus Adolphus College
507-933-7042
[email protected]
#3 lore raw
In reply to Andy Gospodarek
Quoted 36 lines · Andy Gospodarek
Quoted 15 lines · David Kimdon

Hi,

The attached patches enables the bridge to filter and forward packets according to their IEEE 802.1q headers. The goals behind this change include :

- Enable running STP on 802.1q tagged networks. STP packets must be untagged. It isn't obvious how else to enable STP with the current bridge and vlan code.
- Add native support for an untagged vlan. Currently an untagged vlan can be implimented using ebtables or similar.
- On devices bridging a large number of interfaces across various vlans this significantly simplifies configuration and, depending on configuration, can improve performance.

Quoted 3 lines

Comments appreciated,

David

David,

It looks like this code specifically ignores (which is OK for now) and clears (which is not OK) the frame's 802.1p priority. Have you tested priority-tagged frames to see if they are passed correctly? It seems like you should consider adding another field to the sk_buff structure so priority of the incoming frame can be tracked as well as add logic to br_vlan_output_frame and br_vlan_input_frame so the tag is saved. Something like this would probably be fine:

Signed-off-by Andy Gospodarek <[email protected]>

include/linux/skbuff.h 2006-09-11 14:43:27.000000000 -0400 struct sk_buff { · line 302

Quoted 6 lines
include/linux/skbuff.h 2006-09-11 14:43:27.000000000 -0400 renamed from include/linux/skbuff.h.gospo 2006-09-11 14:41:54.000000000 -0400 +1 -0
@@ -297,6 +297,7 @@ struct sk_buff {
        __u32                   nfmark;
 #endif /* CONFIG_NETFILTER */
 #ifdef CONFIG_BRIDGE_VLAN
+       unsigned int            vlan_priority;
        unsigned int            vlan;
 #endif

SKB bloat is not good for performance. You can store the priority and the VID in a single 32-bit number with room to spare...

The skb->priority is also a possible place to store the VLAN priority.

The actual VLAN code has a mapping from vlan-priority -> skb priority on ingress, and another mapping of skb->priority -> vlan-priority on egress. You probably don't need to worry about this mapping for a bridge, however.

Ben

Signature · 2 lines
Ben Greear <[email protected]>
Candela Technologies Inc  http://www.candelatech.com
#8 lore raw
In reply to Ethan Sommer
Quoted 5 lines · Ethan Sommer

Have you tested this in a real world scenario?

I attempted to use a linux bridge to bridge a few vlans and found that it didn't work as expected. At first I thought it was a performance problem relating to the linux bridge code because it was dropping packets, but I believe the problem is inherent in using a bridge in a vlan environment.

The problem lies in the fact that the switch connected to the bridge (which has the vlans tagged) gets confused as to which port that particular mac address is out of. Take the following scenario: (based on my understanding of how things should work)

Cisco switches have this bug, though it can be worked around on some switches. The Netgear switches I've used do not have any problem with the same MAC on different vlans.

Either way, you can change the MAC of the Linux VLAN device using the standard ifconfig or ip tool commands.

Ben

Signature · 2 lines
Ben Greear <[email protected]>
Candela Technologies Inc  http://www.candelatech.com
#4 lore raw
In reply to Ben Greear
Quoted 47 lines · Ben Greear
Quoted 36 lines · Andy Gospodarek
Quoted 15 lines · David Kimdon

Hi,

The attached patches enables the bridge to filter and forward packets according to their IEEE 802.1q headers. The goals behind this change include :

- Enable running STP on 802.1q tagged networks. STP packets must be untagged. It isn't obvious how else to enable STP with the current bridge and vlan code.
- Add native support for an untagged vlan. Currently an untagged vlan can be implimented using ebtables or similar.
- On devices bridging a large number of interfaces across various vlans this significantly simplifies configuration and, depending on configuration, can improve performance.

Quoted 3 lines

Comments appreciated,

David

David,

It looks like this code specifically ignores (which is OK for now) and clears (which is not OK) the frame's 802.1p priority. Have you tested priority-tagged frames to see if they are passed correctly? It seems like you should consider adding another field to the sk_buff structure so priority of the incoming frame can be tracked as well as add logic to br_vlan_output_frame and br_vlan_input_frame so the tag is saved. Something like this would probably be fine:

Signed-off-by Andy Gospodarek <[email protected]>

include/linux/skbuff.h 2006-09-11 14:43:27.000000000 -0400 struct sk_buff { · line 302

Quoted 6 lines
include/linux/skbuff.h 2006-09-11 14:43:27.000000000 -0400 renamed from include/linux/skbuff.h.gospo 2006-09-11 14:41:54.000000000 -0400 +1 -0
@@ -297,6 +297,7 @@ struct sk_buff {
        __u32                   nfmark;
 #endif /* CONFIG_NETFILTER */
 #ifdef CONFIG_BRIDGE_VLAN
+       unsigned int            vlan_priority;
        unsigned int            vlan;
 #endif

SKB bloat is not good for performance. You can store the priority and the VID in a single 32-bit number with room to spare...

Quoted 1 line

The skb->priority is also a possible place to store the VLAN priority.

Good. I didn't like the idea of increasing the bloat factor of an skb.

Quoted 4 lines

The actual VLAN code has a mapping from vlan-priority -> skb priority on ingress, and another mapping of skb->priority -> vlan-priority on egress. You probably don't need to worry about this mapping for a bridge, however.

If the bridge destroys your existing priority then that's bad however. Am I seeing this wrong or will this not reset bridged frames to priority zero in the case where untagged frames need to be given a tag as they are bridged?

Quoted 8 lines

Ben

--
Ben Greear <[email protected]>
Candela Technologies Inc http://www.candelatech.com

author
#6 lore raw
In reply to Andy Gospodarek
Quoted 5 lines · Andy Gospodarek

On 9/11/06, David Kimdon <[email protected]> wrote:
David,

It looks like this code specifically ignores (which is OK for now) and clears (which is not OK) the frame's 802.1p priority.i

ah, yes, good point.

Quoted 2 lines

Have you tested
priority-tagged frames to see if they are passed correctly?

well, they get mapped to the correct vlan (untagged), but, yes, we lose the priority.

Quoted 1 line

Something like this would probably be fine:

yes, thanks.

-David

author
#5 lore raw
In reply to Ben Greear
Quoted 7 lines
Quoted 7 lines
include/linux/skbuff.h 2006-09-11 14:43:27.000000000 -0400 renamed from include/linux/skbuff.h.gospo 2006-09-11 14:41:54.000000000 -0400 +1 -0
@@ -297,6 +297,7 @@ struct sk_buff {
       __u32                   nfmark;
 #endif /* CONFIG_NETFILTER */
 #ifdef CONFIG_BRIDGE_VLAN
+       unsigned int            vlan_priority;
Quoted 8 lines

include/linux/skbuff.h 2006-09-11 14:43:27.000000000 -0400 struct sk_buff { · line 302

Quoted 2 lines
include/linux/skbuff.h 2006-09-11 14:43:27.000000000 -0400 renamed from include/linux/skbuff.h.gospo 2006-09-11 14:41:54.000000000 -0400 +0 -0
       unsigned int            vlan;
 #endif

SKB bloat is not good for performance. You can store the priority and the VID in
a single 32-bit number with room to spare...

I wonder if we can use skb->cb?

-David

#9 lore raw
In reply to David Kimdon
Quoted 1 line
Quoted 1 line
Quoted 1 line
Quoted 1 line
Quoted 1 line

"DK" == David Kimdon <[email protected]> writes:

DK> Hi, The attached patches enables the bridge to filter and forward DK> packets according to their IEEE 802.1q headers.

How does this work in the presence of hardware VLAN acceleration (as almost all important cards/drivers offer)? In that case it seems that the tag is stripped and forgotten about. Does this code somehow avoid the problem?

/Benny