[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_and_depth(skb, skb->protocol, &depth))
			goto drop;

		skb_set_network_header(skb, depth);
	}

	br_switchdev_frame_set_offload_fwd_mark(skb);

	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_clear_tstamp(skb);
	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;

	/* Mark the skb for forwarding offload early so that br_handle_vlan()
	 * can know whether to pop the VLAN header on egress or keep it.
	 */
	nbp_switchdev_frame_mark_tx_fwd_offload(to, skb);

	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;

	nbp_switchdev_frame_mark_tx_fwd_to_hwdom(p, skb);

	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,
	      u16 vid)
{
	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 (BR_INPUT_SKB_CB(skb)->proxyarp_replied &&
		    ((p->flags & BR_PROXYARP_WIFI) ||
		     br_is_neigh_suppress_enabled(p, vid)))
			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
	1	// SPDX-License-Identifier: GPL-2.0-or-later
	2	/*
	3	* Forwarding decision
	4	* Linux ethernet bridge
	5	*
	6	* Authors:
	7	* Lennert Buytenhek <buytenh@gnu.org>
	8	*/
	9
	10	#include <linux/err.h>
	11	#include <linux/slab.h>
	12	#include <linux/kernel.h>
	13	#include <linux/netdevice.h>
	14	#include <linux/netpoll.h>
	15	#include <linux/skbuff.h>
	16	#include <linux/if_vlan.h>
	17	#include <linux/netfilter_bridge.h>
	18	#include "br_private.h"
	19
	20	/* Don't forward packets to originating port or forwarding disabled */
	21	static inline int should_deliver(const struct net_bridge_port *p,
	22	const struct sk_buff *skb)
	23	{
	24	struct net_bridge_vlan_group *vg;
	25
	26	vg = nbp_vlan_group_rcu(p);
	27	return ((p->flags & BR_HAIRPIN_MODE) \|\| skb->dev != p->dev) &&
	28	p->state == BR_STATE_FORWARDING && br_allowed_egress(vg, skb) &&
	29	nbp_switchdev_allowed_egress(p, skb) &&
	30	!br_skb_isolated(p, skb);
	31	}
	32
	33	int br_dev_queue_push_xmit(struct net net, struct sock sk, struct sk_buff *skb)
	34	{
	35	skb_push(skb, ETH_HLEN);
	36	if (!is_skb_forwardable(skb->dev, skb))
	37	goto drop;
	38
	39	br_drop_fake_rtable(skb);
	40
	41	if (skb->ip_summed == CHECKSUM_PARTIAL &&
	42	eth_type_vlan(skb->protocol)) {
	43	int depth;
	44
	45	if (!vlan_get_protocol_and_depth(skb, skb->protocol, &depth))
	46	goto drop;
	47
	48	skb_set_network_header(skb, depth);
	49	}
	50
	51	br_switchdev_frame_set_offload_fwd_mark(skb);
	52
	53	dev_queue_xmit(skb);
	54
	55	return 0;
	56
	57	drop:
	58	kfree_skb(skb);
	59	return 0;
	60	}
	61	EXPORT_SYMBOL_GPL(br_dev_queue_push_xmit);
	62
	63	int br_forward_finish(struct net net, struct sock sk, struct sk_buff *skb)
	64	{
	65	skb_clear_tstamp(skb);
	66	return NF_HOOK(NFPROTO_BRIDGE, NF_BR_POST_ROUTING,
	67	net, sk, skb, NULL, skb->dev,
	68	br_dev_queue_push_xmit);
	69
	70	}
	71	EXPORT_SYMBOL_GPL(br_forward_finish);
	72
	73	static void __br_forward(const struct net_bridge_port *to,
	74	struct sk_buff *skb, bool local_orig)
	75	{
	76	struct net_bridge_vlan_group *vg;
	77	struct net_device *indev;
	78	struct net *net;
	79	int br_hook;
	80
	81	/* Mark the skb for forwarding offload early so that br_handle_vlan()
	82	* can know whether to pop the VLAN header on egress or keep it.
	83	*/
	84	nbp_switchdev_frame_mark_tx_fwd_offload(to, skb);
	85
	86	vg = nbp_vlan_group_rcu(to);
	87	skb = br_handle_vlan(to->br, to, vg, skb);
	88	if (!skb)
	89	return;
	90
	91	indev = skb->dev;
	92	skb->dev = to->dev;
	93	if (!local_orig) {
	94	if (skb_warn_if_lro(skb)) {
	95	kfree_skb(skb);
	96	return;
	97	}
	98	br_hook = NF_BR_FORWARD;
	99	skb_forward_csum(skb);
	100	net = dev_net(indev);
	101	} else {
	102	if (unlikely(netpoll_tx_running(to->br->dev))) {
	103	skb_push(skb, ETH_HLEN);
	104	if (!is_skb_forwardable(skb->dev, skb))
	105	kfree_skb(skb);
	106	else
	107	br_netpoll_send_skb(to, skb);
	108	return;
	109	}
	110	br_hook = NF_BR_LOCAL_OUT;
	111	net = dev_net(skb->dev);
	112	indev = NULL;
	113	}
	114
	115	NF_HOOK(NFPROTO_BRIDGE, br_hook,
	116	net, NULL, skb, indev, skb->dev,
	117	br_forward_finish);
	118	}
	119
	120	static int deliver_clone(const struct net_bridge_port *prev,
	121	struct sk_buff *skb, bool local_orig)
	122	{
	123	struct net_device *dev = BR_INPUT_SKB_CB(skb)->brdev;
	124
	125	skb = skb_clone(skb, GFP_ATOMIC);
	126	if (!skb) {
	127	dev->stats.tx_dropped++;
	128	return -ENOMEM;
	129	}
	130
	131	__br_forward(prev, skb, local_orig);
	132	return 0;
	133	}
	134
	135	/**
	136	* br_forward - forward a packet to a specific port
	137	* @to: destination port
	138	* @skb: packet being forwarded
	139	* @local_rcv: packet will be received locally after forwarding
	140	* @local_orig: packet is locally originated
	141	*
	142	* Should be called with rcu_read_lock.
	143	*/
	144	void br_forward(const struct net_bridge_port *to,
	145	struct sk_buff *skb, bool local_rcv, bool local_orig)
	146	{
	147	if (unlikely(!to))
	148	goto out;
	149
	150	/* redirect to backup link if the destination port is down */
	151	if (rcu_access_pointer(to->backup_port) && !netif_carrier_ok(to->dev)) {
	152	struct net_bridge_port *backup_port;
	153
	154	backup_port = rcu_dereference(to->backup_port);
	155	if (unlikely(!backup_port))
	156	goto out;
	157	to = backup_port;
	158	}
	159
	160	if (should_deliver(to, skb)) {
	161	if (local_rcv)
	162	deliver_clone(to, skb, local_orig);
	163	else
	164	__br_forward(to, skb, local_orig);
	165	return;
	166	}
	167
	168	out:
	169	if (!local_rcv)
	170	kfree_skb(skb);
	171	}
	172	EXPORT_SYMBOL_GPL(br_forward);
	173
	174	static struct net_bridge_port *maybe_deliver(
	175	struct net_bridge_port prev, struct net_bridge_port p,
	176	struct sk_buff *skb, bool local_orig)
	177	{
	178	u8 igmp_type = br_multicast_igmp_type(skb);
	179	int err;
	180
	181	if (!should_deliver(p, skb))
	182	return prev;
	183
	184	nbp_switchdev_frame_mark_tx_fwd_to_hwdom(p, skb);
	185
	186	if (!prev)
	187	goto out;
	188
	189	err = deliver_clone(prev, skb, local_orig);
	190	if (err)
	191	return ERR_PTR(err);
	192	out:
	193	br_multicast_count(p->br, p, skb, igmp_type, BR_MCAST_DIR_TX);
	194
	195	return p;
	196	}
	197
	198	/* called under rcu_read_lock */
	199	void br_flood(struct net_bridge br, struct sk_buff skb,
	200	enum br_pkt_type pkt_type, bool local_rcv, bool local_orig,
	201	u16 vid)
	202	{
	203	struct net_bridge_port *prev = NULL;
	204	struct net_bridge_port *p;
	205
	206	list_for_each_entry_rcu(p, &br->port_list, list) {
	207	/* Do not flood unicast traffic to ports that turn it off, nor
	208	* other traffic if flood off, except for traffic we originate
	209	*/
	210	switch (pkt_type) {
	211	case BR_PKT_UNICAST:
	212	if (!(p->flags & BR_FLOOD))
	213	continue;
	214	break;
	215	case BR_PKT_MULTICAST:
	216	if (!(p->flags & BR_MCAST_FLOOD) && skb->dev != br->dev)
	217	continue;
	218	break;
	219	case BR_PKT_BROADCAST:
	220	if (!(p->flags & BR_BCAST_FLOOD) && skb->dev != br->dev)
	221	continue;
	222	break;
	223	}
	224
	225	/* Do not flood to ports that enable proxy ARP */
	226	if (p->flags & BR_PROXYARP)
	227	continue;
	228	if (BR_INPUT_SKB_CB(skb)->proxyarp_replied &&
	229	((p->flags & BR_PROXYARP_WIFI) \|\|
	230	br_is_neigh_suppress_enabled(p, vid)))
	231	continue;
	232
	233	prev = maybe_deliver(prev, p, skb, local_orig);
	234	if (IS_ERR(prev))
	235	goto out;
	236	}
	237
	238	if (!prev)
	239	goto out;
	240
	241	if (local_rcv)
	242	deliver_clone(prev, skb, local_orig);
	243	else
	244	__br_forward(prev, skb, local_orig);
	245	return;
	246
	247	out:
	248	if (!local_rcv)
	249	kfree_skb(skb);
	250	}
	251
	252	#ifdef CONFIG_BRIDGE_IGMP_SNOOPING
	253	static void maybe_deliver_addr(struct net_bridge_port p, struct sk_buff skb,
	254	const unsigned char *addr, bool local_orig)
	255	{
	256	struct net_device *dev = BR_INPUT_SKB_CB(skb)->brdev;
	257	const unsigned char *src = eth_hdr(skb)->h_source;
	258
	259	if (!should_deliver(p, skb))
	260	return;
	261
	262	/* Even with hairpin, no soliloquies - prevent breaking IPv6 DAD */
	263	if (skb->dev == p->dev && ether_addr_equal(src, addr))
	264	return;
	265
	266	skb = skb_copy(skb, GFP_ATOMIC);
	267	if (!skb) {
	268	dev->stats.tx_dropped++;
	269	return;
	270	}
	271
	272	if (!is_broadcast_ether_addr(addr))
	273	memcpy(eth_hdr(skb)->h_dest, addr, ETH_ALEN);
	274
	275	__br_forward(p, skb, local_orig);
	276	}
	277
	278	/* called with rcu_read_lock */
	279	void br_multicast_flood(struct net_bridge_mdb_entry *mdst,
	280	struct sk_buff *skb,
	281	struct net_bridge_mcast *brmctx,
	282	bool local_rcv, bool local_orig)
	283	{
	284	struct net_bridge_port *prev = NULL;
	285	struct net_bridge_port_group *p;
	286	bool allow_mode_include = true;
	287	struct hlist_node *rp;
	288
	289	rp = br_multicast_get_first_rport_node(brmctx, skb);
	290
	291	if (mdst) {
	292	p = rcu_dereference(mdst->ports);
	293	if (br_multicast_should_handle_mode(brmctx, mdst->addr.proto) &&
	294	br_multicast_is_star_g(&mdst->addr))
	295	allow_mode_include = false;
	296	} else {
	297	p = NULL;
	298	}
	299
	300	while (p \|\| rp) {
	301	struct net_bridge_port port, lport, *rport;
	302
	303	lport = p ? p->key.port : NULL;
	304	rport = br_multicast_rport_from_node_skb(rp, skb);
	305
	306	if ((unsigned long)lport > (unsigned long)rport) {
	307	port = lport;
	308
	309	if (port->flags & BR_MULTICAST_TO_UNICAST) {
	310	maybe_deliver_addr(lport, skb, p->eth_addr,
	311	local_orig);
	312	goto delivered;
	313	}
	314	if ((!allow_mode_include &&
	315	p->filter_mode == MCAST_INCLUDE) \|\|
	316	(p->flags & MDB_PG_FLAGS_BLOCKED))
	317	goto delivered;
	318	} else {
	319	port = rport;
	320	}
	321
	322	prev = maybe_deliver(prev, port, skb, local_orig);
	323	if (IS_ERR(prev))
	324	goto out;
	325	delivered:
	326	if ((unsigned long)lport >= (unsigned long)port)
	327	p = rcu_dereference(p->next);
	328	if ((unsigned long)rport >= (unsigned long)port)
	329	rp = rcu_dereference(hlist_next_rcu(rp));
	330	}
	331
	332	if (!prev)
	333	goto out;
	334
	335	if (local_rcv)
	336	deliver_clone(prev, skb, local_orig);
	337	else
	338	__br_forward(prev, skb, local_orig);
	339	return;
	340
	341	out:
	342	if (!local_rcv)
	343	kfree_skb(skb);
	344	}
	345	#endif