[linux-block.git] / net / bridge / br_forward.c

// SPDX-License-Identifier: GPL-2.0-or-later
/*
 *	Forwarding decision
 *	Linux ethernet bridge
 *
 *	Authors:
 *	Lennert Buytenhek		<buytenh@gnu.org>
 */

#include <linux/err.h>
#include <linux/slab.h>
#include <linux/kernel.h>
#include <linux/netdevice.h>
#include <linux/netpoll.h>
#include <linux/skbuff.h>
#include <linux/if_vlan.h>
#include <linux/netfilter_bridge.h>
#include "br_private.h"

/* Don't forward packets to originating port or forwarding disabled */
static inline int should_deliver(const struct net_bridge_port *p,
				 const struct sk_buff *skb)
{
	struct net_bridge_vlan_group *vg;

	vg = nbp_vlan_group_rcu(p);
	return ((p->flags & BR_HAIRPIN_MODE) || skb->dev != p->dev) &&
		p->state == BR_STATE_FORWARDING && br_allowed_egress(vg, skb) &&
		nbp_switchdev_allowed_egress(p, skb) &&
		!br_skb_isolated(p, skb);
}

int br_dev_queue_push_xmit(struct net *net, struct sock *sk, struct sk_buff *skb)
{
	skb_push(skb, ETH_HLEN);
	if (!is_skb_forwardable(skb->dev, skb))
		goto drop;

	br_drop_fake_rtable(skb);

	if (skb->ip_summed == CHECKSUM_PARTIAL &&
	    eth_type_vlan(skb->protocol)) {
		int depth;

		if (!__vlan_get_protocol(skb, skb->protocol, &depth))
			goto drop;

		skb_set_network_header(skb, depth);
	}

	dev_queue_xmit(skb);

	return 0;

drop:
	kfree_skb(skb);
	return 0;
}
EXPORT_SYMBOL_GPL(br_dev_queue_push_xmit);

int br_forward_finish(struct net *net, struct sock *sk, struct sk_buff *skb)
{
	skb->tstamp = 0;
	return NF_HOOK(NFPROTO_BRIDGE, NF_BR_POST_ROUTING,
		       net, sk, skb, NULL, skb->dev,
		       br_dev_queue_push_xmit);

}
EXPORT_SYMBOL_GPL(br_forward_finish);

static void __br_forward(const struct net_bridge_port *to,
			 struct sk_buff *skb, bool local_orig)
{
	struct net_bridge_vlan_group *vg;
	struct net_device *indev;
	struct net *net;
	int br_hook;

	vg = nbp_vlan_group_rcu(to);
	skb = br_handle_vlan(to->br, to, vg, skb);
	if (!skb)
		return;

	indev = skb->dev;
	skb->dev = to->dev;
	if (!local_orig) {
		if (skb_warn_if_lro(skb)) {
			kfree_skb(skb);
			return;
		}
		br_hook = NF_BR_FORWARD;
		skb_forward_csum(skb);
		net = dev_net(indev);
	} else {
		if (unlikely(netpoll_tx_running(to->br->dev))) {
			skb_push(skb, ETH_HLEN);
			if (!is_skb_forwardable(skb->dev, skb))
				kfree_skb(skb);
			else
				br_netpoll_send_skb(to, skb);
			return;
		}
		br_hook = NF_BR_LOCAL_OUT;
		net = dev_net(skb->dev);
		indev = NULL;
	}

	NF_HOOK(NFPROTO_BRIDGE, br_hook,
		net, NULL, skb, indev, skb->dev,
		br_forward_finish);
}

static int deliver_clone(const struct net_bridge_port *prev,
			 struct sk_buff *skb, bool local_orig)
{
	struct net_device *dev = BR_INPUT_SKB_CB(skb)->brdev;

	skb = skb_clone(skb, GFP_ATOMIC);
	if (!skb) {
		dev->stats.tx_dropped++;
		return -ENOMEM;
	}

	__br_forward(prev, skb, local_orig);
	return 0;
}

/**
 * br_forward - forward a packet to a specific port
 * @to: destination port
 * @skb: packet being forwarded
 * @local_rcv: packet will be received locally after forwarding
 * @local_orig: packet is locally originated
 *
 * Should be called with rcu_read_lock.
 */
void br_forward(const struct net_bridge_port *to,
		struct sk_buff *skb, bool local_rcv, bool local_orig)
{
	if (unlikely(!to))
		goto out;

	/* redirect to backup link if the destination port is down */
	if (rcu_access_pointer(to->backup_port) && !netif_carrier_ok(to->dev)) {
		struct net_bridge_port *backup_port;

		backup_port = rcu_dereference(to->backup_port);
		if (unlikely(!backup_port))
			goto out;
		to = backup_port;
	}

	if (should_deliver(to, skb)) {
		if (local_rcv)
			deliver_clone(to, skb, local_orig);
		else
			__br_forward(to, skb, local_orig);
		return;
	}

out:
	if (!local_rcv)
		kfree_skb(skb);
}
EXPORT_SYMBOL_GPL(br_forward);

static struct net_bridge_port *maybe_deliver(
	struct net_bridge_port *prev, struct net_bridge_port *p,
	struct sk_buff *skb, bool local_orig)
{
	u8 igmp_type = br_multicast_igmp_type(skb);
	int err;

	if (!should_deliver(p, skb))
		return prev;

	if (!prev)
		goto out;

	err = deliver_clone(prev, skb, local_orig);
	if (err)
		return ERR_PTR(err);
out:
	br_multicast_count(p->br, p, skb, igmp_type, BR_MCAST_DIR_TX);

	return p;
}

/* called under rcu_read_lock */
void br_flood(struct net_bridge *br, struct sk_buff *skb,
	      enum br_pkt_type pkt_type, bool local_rcv, bool local_orig)
{
	struct net_bridge_port *prev = NULL;
	struct net_bridge_port *p;

	list_for_each_entry_rcu(p, &br->port_list, list) {
		/* Do not flood unicast traffic to ports that turn it off, nor
		 * other traffic if flood off, except for traffic we originate
		 */
		switch (pkt_type) {
		case BR_PKT_UNICAST:
			if (!(p->flags & BR_FLOOD))
				continue;
			break;
		case BR_PKT_MULTICAST:
			if (!(p->flags & BR_MCAST_FLOOD) && skb->dev != br->dev)
				continue;
			break;
		case BR_PKT_BROADCAST:
			if (!(p->flags & BR_BCAST_FLOOD) && skb->dev != br->dev)
				continue;
			break;
		}

		/* Do not flood to ports that enable proxy ARP */
		if (p->flags & BR_PROXYARP)
			continue;
		if ((p->flags & (BR_PROXYARP_WIFI | BR_NEIGH_SUPPRESS)) &&
		    BR_INPUT_SKB_CB(skb)->proxyarp_replied)
			continue;

		prev = maybe_deliver(prev, p, skb, local_orig);
		if (IS_ERR(prev))
			goto out;
	}

	if (!prev)
		goto out;

	if (local_rcv)
		deliver_clone(prev, skb, local_orig);
	else
		__br_forward(prev, skb, local_orig);
	return;

out:
	if (!local_rcv)
		kfree_skb(skb);
}

#ifdef CONFIG_BRIDGE_IGMP_SNOOPING
static void maybe_deliver_addr(struct net_bridge_port *p, struct sk_buff *skb,
			       const unsigned char *addr, bool local_orig)
{
	struct net_device *dev = BR_INPUT_SKB_CB(skb)->brdev;
	const unsigned char *src = eth_hdr(skb)->h_source;

	if (!should_deliver(p, skb))
		return;

	/* Even with hairpin, no soliloquies - prevent breaking IPv6 DAD */
	if (skb->dev == p->dev && ether_addr_equal(src, addr))
		return;

	skb = skb_copy(skb, GFP_ATOMIC);
	if (!skb) {
		dev->stats.tx_dropped++;
		return;
	}

	if (!is_broadcast_ether_addr(addr))
		memcpy(eth_hdr(skb)->h_dest, addr, ETH_ALEN);

	__br_forward(p, skb, local_orig);
}

/* called with rcu_read_lock */
void br_multicast_flood(struct net_bridge_mdb_entry *mdst,
			struct sk_buff *skb,
			struct net_bridge_mcast *brmctx,
			bool local_rcv, bool local_orig)
{
	struct net_bridge_port *prev = NULL;
	struct net_bridge_port_group *p;
	bool allow_mode_include = true;
	struct hlist_node *rp;

	rp = br_multicast_get_first_rport_node(brmctx, skb);

	if (mdst) {
		p = rcu_dereference(mdst->ports);
		if (br_multicast_should_handle_mode(brmctx, mdst->addr.proto) &&
		    br_multicast_is_star_g(&mdst->addr))
			allow_mode_include = false;
	} else {
		p = NULL;
	}

	while (p || rp) {
		struct net_bridge_port *port, *lport, *rport;

		lport = p ? p->key.port : NULL;
		rport = br_multicast_rport_from_node_skb(rp, skb);

		if ((unsigned long)lport > (unsigned long)rport) {
			port = lport;

			if (port->flags & BR_MULTICAST_TO_UNICAST) {
				maybe_deliver_addr(lport, skb, p->eth_addr,
						   local_orig);
				goto delivered;
			}
			if ((!allow_mode_include &&
			     p->filter_mode == MCAST_INCLUDE) ||
			    (p->flags & MDB_PG_FLAGS_BLOCKED))
				goto delivered;
		} else {
			port = rport;
		}

		prev = maybe_deliver(prev, port, skb, local_orig);
		if (IS_ERR(prev))
			goto out;
delivered:
		if ((unsigned long)lport >= (unsigned long)port)
			p = rcu_dereference(p->next);
		if ((unsigned long)rport >= (unsigned long)port)
			rp = rcu_dereference(hlist_next_rcu(rp));
	}

	if (!prev)
		goto out;

	if (local_rcv)
		deliver_clone(prev, skb, local_orig);
	else
		__br_forward(prev, skb, local_orig);
	return;

out:
	if (!local_rcv)
		kfree_skb(skb);
}
#endif
Commit	Line	Data
2874c5fd	1	// SPDX-License-Identifier: GPL-2.0-or-later
1da177e4 LT	2	/*
	3	* Forwarding decision
	4	* Linux ethernet bridge
	5	*
	6	* Authors:
	7	* Lennert Buytenhek <buytenh@gnu.org>
1da177e4 LT	8	*/
1da177e4 LT	9
025d89c2	10	#include <linux/err.h>
5a0e3ad6	11	#include <linux/slab.h>
1da177e4 LT	12	#include <linux/kernel.h>
1da177e4 LT	13	#include <linux/netdevice.h>
c06ee961	14	#include <linux/netpoll.h>
1da177e4	15	#include <linux/skbuff.h>
85ca719e	16	#include <linux/if_vlan.h>
1da177e4 LT	17	#include <linux/netfilter_bridge.h>
	18	#include "br_private.h"
	19
1a81a2e0	20	/* Don't forward packets to originating port or forwarding disabled */
9d6f229f	21	static inline int should_deliver(const struct net_bridge_port *p,
1da177e4 LT	22	const struct sk_buff *skb)
1da177e4 LT	23	{
2594e906 NA	24	struct net_bridge_vlan_group *vg;
2594e906 NA	25
907b1e6e	26	vg = nbp_vlan_group_rcu(p);
a97bfc1d	27	return ((p->flags & BR_HAIRPIN_MODE) \|\| skb->dev != p->dev) &&
ac0e932d	28	p->state == BR_STATE_FORWARDING && br_allowed_egress(vg, skb) &&
7d850abd NA	29	nbp_switchdev_allowed_egress(p, skb) &&
7d850abd NA	30	!br_skb_isolated(p, skb);
1da177e4 LT	31	}
1da177e4 LT	32
0c4b51f0	33	int br_dev_queue_push_xmit(struct net net, struct sock sk, struct sk_buff *skb)
1da177e4	34	{
28c1382f	35	skb_push(skb, ETH_HLEN);
df356d5e TM	36	if (!is_skb_forwardable(skb->dev, skb))
	37	goto drop;
	38
df356d5e	39	br_drop_fake_rtable(skb);
df356d5e TM	40
df356d5e TM	41	if (skb->ip_summed == CHECKSUM_PARTIAL &&
a98c0c47	42	eth_type_vlan(skb->protocol)) {
df356d5e TM	43	int depth;
	44
	45	if (!__vlan_get_protocol(skb, skb->protocol, &depth))
	46	goto drop;
	47
	48	skb_set_network_header(skb, depth);
1da177e4 LT	49	}
1da177e4 LT	50
df356d5e TM	51	dev_queue_xmit(skb);
	52
	53	return 0;
	54
	55	drop:
	56	kfree_skb(skb);
1da177e4 LT	57	return 0;
1da177e4 LT	58	}
34666d46	59	EXPORT_SYMBOL_GPL(br_dev_queue_push_xmit);
1da177e4	60
0c4b51f0	61	int br_forward_finish(struct net net, struct sock sk, struct sk_buff *skb)
1da177e4	62	{
41d1c883	63	skb->tstamp = 0;
29a26a56 EB	64	return NF_HOOK(NFPROTO_BRIDGE, NF_BR_POST_ROUTING,
29a26a56 EB	65	net, sk, skb, NULL, skb->dev,
9ef513be	66	br_dev_queue_push_xmit);
1da177e4	67
1da177e4	68	}
34666d46	69	EXPORT_SYMBOL_GPL(br_forward_finish);
1da177e4	70
37b090e6 NA	71	static void __br_forward(const struct net_bridge_port *to,
37b090e6 NA	72	struct sk_buff *skb, bool local_orig)
1da177e4	73	{
2594e906	74	struct net_bridge_vlan_group *vg;
37b090e6 NA	75	struct net_device *indev;
	76	struct net *net;
	77	int br_hook;
2594e906	78
907b1e6e	79	vg = nbp_vlan_group_rcu(to);
11538d03	80	skb = br_handle_vlan(to->br, to, vg, skb);
78851988 VY	81	if (!skb)
	82	return;
	83
37b090e6	84	indev = skb->dev;
1da177e4	85	skb->dev = to->dev;
37b090e6 NA	86	if (!local_orig) {
37b090e6 NA	87	if (skb_warn_if_lro(skb)) {
91d2c34a	88	kfree_skb(skb);
37b090e6	89	return;
91d2c34a	90	}
37b090e6 NA	91	br_hook = NF_BR_FORWARD;
	92	skb_forward_csum(skb);
	93	net = dev_net(indev);
	94	} else {
	95	if (unlikely(netpoll_tx_running(to->br->dev))) {
28c1382f YW	96	skb_push(skb, ETH_HLEN);
28c1382f YW	97	if (!is_skb_forwardable(skb->dev, skb))
37b090e6	98	kfree_skb(skb);
28c1382f	99	else
37b090e6	100	br_netpoll_send_skb(to, skb);
37b090e6 NA	101	return;
	102	}
	103	br_hook = NF_BR_LOCAL_OUT;
	104	net = dev_net(skb->dev);
	105	indev = NULL;
91d2c34a HX	106	}
91d2c34a HX	107
37b090e6 NA	108	NF_HOOK(NFPROTO_BRIDGE, br_hook,
37b090e6 NA	109	net, NULL, skb, indev, skb->dev,
713aefa3	110	br_forward_finish);
1da177e4 LT	111	}
1da177e4 LT	112
37b090e6 NA	113	static int deliver_clone(const struct net_bridge_port *prev,
37b090e6 NA	114	struct sk_buff *skb, bool local_orig)
1da177e4	115	{
37b090e6	116	struct net_device *dev = BR_INPUT_SKB_CB(skb)->brdev;
1da177e4	117
37b090e6 NA	118	skb = skb_clone(skb, GFP_ATOMIC);
	119	if (!skb) {
	120	dev->stats.tx_dropped++;
	121	return -ENOMEM;
1da177e4 LT	122	}
1da177e4 LT	123
37b090e6 NA	124	__br_forward(prev, skb, local_orig);
37b090e6 NA	125	return 0;
1da177e4 LT	126	}
1da177e4 LT	127
37b090e6 NA	128	/**
	129	* br_forward - forward a packet to a specific port
	130	* @to: destination port
	131	* @skb: packet being forwarded
	132	* @local_rcv: packet will be received locally after forwarding
	133	* @local_orig: packet is locally originated
	134	*
	135	* Should be called with rcu_read_lock.
	136	*/
	137	void br_forward(const struct net_bridge_port *to,
	138	struct sk_buff *skb, bool local_rcv, bool local_orig)
1da177e4	139	{
2756f68c NA	140	if (unlikely(!to))
	141	goto out;
	142
	143	/* redirect to backup link if the destination port is down */
	144	if (rcu_access_pointer(to->backup_port) && !netif_carrier_ok(to->dev)) {
	145	struct net_bridge_port *backup_port;
	146
	147	backup_port = rcu_dereference(to->backup_port);
	148	if (unlikely(!backup_port))
	149	goto out;
	150	to = backup_port;
	151	}
	152
	153	if (should_deliver(to, skb)) {
b35c5f63	154	if (local_rcv)
37b090e6	155	deliver_clone(to, skb, local_orig);
7f7708f0	156	else
37b090e6	157	__br_forward(to, skb, local_orig);
1da177e4 LT	158	return;
	159	}
	160
2756f68c	161	out:
b35c5f63	162	if (!local_rcv)
7f7708f0	163	kfree_skb(skb);
1da177e4	164	}
37b090e6	165	EXPORT_SYMBOL_GPL(br_forward);
025d89c2 HX	166
	167	static struct net_bridge_port *maybe_deliver(
	168	struct net_bridge_port prev, struct net_bridge_port p,
37b090e6	169	struct sk_buff *skb, bool local_orig)
025d89c2	170	{
847d44ef	171	u8 igmp_type = br_multicast_igmp_type(skb);
025d89c2 HX	172	int err;
	173
	174	if (!should_deliver(p, skb))
	175	return prev;
	176
	177	if (!prev)
	178	goto out;
	179
37b090e6	180	err = deliver_clone(prev, skb, local_orig);
025d89c2 HX	181	if (err)
025d89c2 HX	182	return ERR_PTR(err);
025d89c2	183	out:
847d44ef PNA	184	br_multicast_count(p->br, p, skb, igmp_type, BR_MCAST_DIR_TX);
847d44ef PNA	185
025d89c2 HX	186	return p;
	187	}
	188
37b090e6 NA	189	/* called under rcu_read_lock */
37b090e6 NA	190	void br_flood(struct net_bridge br, struct sk_buff skb,
8addd5e7	191	enum br_pkt_type pkt_type, bool local_rcv, bool local_orig)
1da177e4	192	{
37b090e6	193	struct net_bridge_port *prev = NULL;
1080ab95	194	struct net_bridge_port *p;
1da177e4	195
1da177e4	196	list_for_each_entry_rcu(p, &br->port_list, list) {
99f906e9 MM	197	/* Do not flood unicast traffic to ports that turn it off, nor
	198	* other traffic if flood off, except for traffic we originate
	199	*/
	200	switch (pkt_type) {
	201	case BR_PKT_UNICAST:
	202	if (!(p->flags & BR_FLOOD))
	203	continue;
	204	break;
	205	case BR_PKT_MULTICAST:
	206	if (!(p->flags & BR_MCAST_FLOOD) && skb->dev != br->dev)
	207	continue;
	208	break;
	209	case BR_PKT_BROADCAST:
	210	if (!(p->flags & BR_BCAST_FLOOD) && skb->dev != br->dev)
	211	continue;
	212	break;
	213	}
95850116 KP	214
	215	/* Do not flood to ports that enable proxy ARP */
	216	if (p->flags & BR_PROXYARP)
	217	continue;
821f1b21	218	if ((p->flags & (BR_PROXYARP_WIFI \| BR_NEIGH_SUPPRESS)) &&
842a9ae0 JM	219	BR_INPUT_SKB_CB(skb)->proxyarp_replied)
842a9ae0 JM	220	continue;
95850116	221
37b090e6	222	prev = maybe_deliver(prev, p, skb, local_orig);
025d89c2 HX	223	if (IS_ERR(prev))
025d89c2 HX	224	goto out;
1da177e4 LT	225	}
1da177e4 LT	226
b33084be HX	227	if (!prev)
	228	goto out;
	229
b35c5f63	230	if (local_rcv)
37b090e6	231	deliver_clone(prev, skb, local_orig);
025d89c2	232	else
37b090e6	233	__br_forward(prev, skb, local_orig);
b33084be	234	return;
1da177e4	235
b33084be	236	out:
b35c5f63	237	if (!local_rcv)
b33084be	238	kfree_skb(skb);
1da177e4 LT	239	}
1da177e4 LT	240
5cb5e947	241	#ifdef CONFIG_BRIDGE_IGMP_SNOOPING
5b9d6b15 AB	242	static void maybe_deliver_addr(struct net_bridge_port p, struct sk_buff skb,
	243	const unsigned char *addr, bool local_orig)
	244	{
	245	struct net_device *dev = BR_INPUT_SKB_CB(skb)->brdev;
	246	const unsigned char *src = eth_hdr(skb)->h_source;
	247
	248	if (!should_deliver(p, skb))
	249	return;
	250
	251	/* Even with hairpin, no soliloquies - prevent breaking IPv6 DAD */
	252	if (skb->dev == p->dev && ether_addr_equal(src, addr))
	253	return;
	254
	255	skb = skb_copy(skb, GFP_ATOMIC);
	256	if (!skb) {
	257	dev->stats.tx_dropped++;
	258	return;
	259	}
	260
	261	if (!is_broadcast_ether_addr(addr))
	262	memcpy(eth_hdr(skb)->h_dest, addr, ETH_ALEN);
	263
	264	__br_forward(p, skb, local_orig);
	265	}
	266
5cb5e947	267	/* called with rcu_read_lock */
37b090e6 NA	268	void br_multicast_flood(struct net_bridge_mdb_entry *mdst,
37b090e6 NA	269	struct sk_buff *skb,
adc47037	270	struct net_bridge_mcast *brmctx,
37b090e6	271	bool local_rcv, bool local_orig)
5cb5e947	272	{
afe0159d	273	struct net_bridge_port *prev = NULL;
5cb5e947	274	struct net_bridge_port_group *p;
36cfec73	275	bool allow_mode_include = true;
5cb5e947 HX	276	struct hlist_node *rp;
5cb5e947 HX	277
adc47037	278	rp = br_multicast_get_first_rport_node(brmctx, skb);
44ebb081	279
36cfec73 NA	280	if (mdst) {
36cfec73 NA	281	p = rcu_dereference(mdst->ports);
adc47037	282	if (br_multicast_should_handle_mode(brmctx, mdst->addr.proto) &&
36cfec73 NA	283	br_multicast_is_star_g(&mdst->addr))
	284	allow_mode_include = false;
	285	} else {
	286	p = NULL;
	287	}
	288
5cb5e947	289	while (p \|\| rp) {
afe0159d	290	struct net_bridge_port port, lport, *rport;
afe0159d	291
085b53c8	292	lport = p ? p->key.port : NULL;
44ebb081	293	rport = br_multicast_rport_from_node_skb(rp, skb);
5cb5e947	294
6db6f0ea FF	295	if ((unsigned long)lport > (unsigned long)rport) {
	296	port = lport;
	297
	298	if (port->flags & BR_MULTICAST_TO_UNICAST) {
	299	maybe_deliver_addr(lport, skb, p->eth_addr,
	300	local_orig);
	301	goto delivered;
	302	}
36cfec73 NA	303	if ((!allow_mode_include &&
	304	p->filter_mode == MCAST_INCLUDE) \|\|
	305	(p->flags & MDB_PG_FLAGS_BLOCKED))
	306	goto delivered;
6db6f0ea FF	307	} else {
	308	port = rport;
	309	}
5cb5e947	310
37b090e6	311	prev = maybe_deliver(prev, port, skb, local_orig);
5cb5e947 HX	312	if (IS_ERR(prev))
5cb5e947 HX	313	goto out;
847d44ef	314	delivered:
5cb5e947	315	if ((unsigned long)lport >= (unsigned long)port)
83f6a740	316	p = rcu_dereference(p->next);
5cb5e947	317	if ((unsigned long)rport >= (unsigned long)port)
e8051688	318	rp = rcu_dereference(hlist_next_rcu(rp));
5cb5e947 HX	319	}
	320
	321	if (!prev)
	322	goto out;
	323
b35c5f63	324	if (local_rcv)
37b090e6	325	deliver_clone(prev, skb, local_orig);
5cb5e947	326	else
37b090e6	327	__br_forward(prev, skb, local_orig);
5cb5e947 HX	328	return;
	329
	330	out:
b35c5f63	331	if (!local_rcv)
5cb5e947 HX	332	kfree_skb(skb);
5cb5e947 HX	333	}
5cb5e947	334	#endif