/*
 *	Forwarding decision
 *	Linux ethernet bridge
 *
 *	Authors:
 *	Lennert Buytenhek		<buytenh@gnu.org>
 *
 *	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/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"

#include <avm/pa/avm_pa.h>

static int deliver_clone(const struct net_bridge_port *prev,
			 struct sk_buff *skb,
			 void (*__packet_hook)(const struct net_bridge_port *p,
					       struct sk_buff *skb));

/* 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) &&
		br_allowed_egress(vg, skb) && p->state == BR_STATE_FORWARDING;
}

#ifdef CONFIG_AVM_BRIDGE_MULTICAST_TO_UNICAST
static inline int br_forward_as_unicast(struct sk_buff *skb)
{
	struct net_bridge_port            *port = br_port_get_rcu(skb->dev);
	struct net_bridge_port_group      *pg = BR_OUTPUT_SKB_CB(skb)->pg;
	struct net_bridge_group_recipient *p;

	/* Nothing to do? */
	if (likely((port->flags & BR_MULTICAST_TO_UNICAST) == 0) || !pg)
		return 0;

	/* Should be BUG_ON() but I don't dare reboots now */
	if (WARN(pg->port != port, "Ports do not match for group %s vs %s\n",
			 pg->port->dev->name, port->dev->name)) {
		return 0;
	}

	/* Explicit MDB entries added by user space do not have snooped
	 * recipients, proceed as multicast. Likewise if there are too many recipients */
	if (pg->num_recipients == 0 || pg->num_recipients > port->multicast_to_unicast_threshold)
		return 0;

	/* CPA/TMA 2016/04/27: skb is cloned if transmitted to more than one port or
	 * if passed to the local system as well. If not cloned we try to re-use
	 * the original skb.
	 * We have to unclone skb here as the packet must not go through anyone
	 * else, because we manipulate the actual data. skb_cow() does
	 * exactly that but clone must not be shared (this is the case here because
	 * we come from deliver_clone()).
	 */
	if (skb_cow(skb, 0) != 0 || !pskb_may_pull(skb, ETH_ALEN)) {
		/* proceed as multicast if out-of-memory */
		return 0;
	}

	/* CPA/TMA 2016/04/27: skb is cloned if transmitted to more than one port or
	 * if passed to the local system as well. If not cloned we try to re-use
	 * the original skb.
	 * We have to unclone skb here as the packet must not go through anyone
	 * else, because we manipulate the actual data. skb_cow() does
	 * exactly that but clone must not be shared (this is the case here because
	 * we come from deliver_clone()).
	 */
	if (skb_cow(skb, 0) != 0 || !pskb_may_pull(skb, ETH_ALEN)) {
		/* proceed as multicast if out-of-memory */
		return 0;
	}

	list_for_each_entry_rcu(p, &pg->recipients, list) {
		struct net_bridge_fdb_entry *fdb = p->recipient;
		struct sk_buff *ucast_skb;
		struct ethhdr  *ucast_eth;

		/* Re-use the original skb if possible. The skb must be copied
		 * for every recipient except the last one where we can
		 * use the (possibly uncloned) original skb */
		if (list_is_last(&p->list, &pg->recipients))
			ucast_skb = skb;
		else {
			ucast_skb = pskb_copy(skb, GFP_ATOMIC);
			if (unlikely(!ucast_skb)) {
				/* Abort if out-of-memory. Here we prefer to drop
				 * instead of giving recipients duplicated packets. */
				kfree_skb(skb);
				break;
			}
		}

		/* modify and send */
		ucast_eth = eth_hdr(ucast_skb);
		memcpy(ucast_eth->h_dest, &fdb->addr.addr, ETH_ALEN);
		dev_queue_xmit(ucast_skb);
	}

	return 1;
}

static inline int is_udp(struct sk_buff *skb)
{
	/*
	 * only change UDP pakets to unicast
	 *   2014-12-12, calle
	 */
	switch (skb->protocol) {
		case htons(ETH_P_IP):
			return ip_hdr(skb)->protocol == IPPROTO_UDP;
#if defined(CONFIG_IPV6) || defined(CONFIG_IPV6_MODULE)
		case htons(ETH_P_IPV6):
			return ipv6_hdr(skb)->nexthdr == IPPROTO_UDP;
#endif
		default:
			return 0;
	}
}

static inline int should_convert_to_unicast(struct sk_buff *skb, const u8 *dest)
{
	return is_multicast_ether_addr(dest) && !is_broadcast_ether_addr(dest) && is_udp(skb);
}

#else
static inline int br_forward_as_unicast(struct sk_buff *skb)
{
	return 0;
}
static inline int should_convert_to_unicast(struct sk_buff *skb, const u8 *dest)
{
	return 0;
}
#endif

int br_dev_queue_push_xmit(struct net *net, struct sock *sk, struct sk_buff *skb)
{
	u8 *dest;

	dest = skb_push(skb, ETH_HLEN);

	if (!is_skb_forwardable(skb->dev, skb))
		goto drop;

	br_drop_fake_rtable(skb);
	skb_sender_cpu_clear(skb);

	if (skb->ip_summed == CHECKSUM_PARTIAL &&
	    (skb->protocol == htons(ETH_P_8021Q) ||
	     skb->protocol == htons(ETH_P_8021AD))) {
		int depth;

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

		skb_set_network_header(skb, depth);
	}

	if (!should_convert_to_unicast(skb, dest) || !br_forward_as_unicast(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)
{
	return BR_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_deliver(const struct net_bridge_port *to, struct sk_buff *skb)
{
	struct net_bridge_vlan_group *vg;

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

	skb->dev = to->dev;

	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(NFPROTO_BRIDGE, NF_BR_LOCAL_OUT,
		dev_net(skb->dev), NULL, skb,NULL, skb->dev,
		br_forward_finish);
}

static void __br_forward(const struct net_bridge_port *to, struct sk_buff *skb)
{
	struct net_bridge_vlan_group *vg;
	struct net_device *indev;

	if (skb_warn_if_lro(skb)) {
		kfree_skb(skb);
		return;
	}

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

	indev = skb->dev;
	skb->dev = to->dev;
	skb_forward_csum(skb);

	BR_HOOK(NFPROTO_BRIDGE, NF_BR_FORWARD,
		dev_net(indev), NULL, skb, indev, skb->dev,
		br_forward_finish);
}

/* called with rcu_read_lock */
void br_deliver(const struct net_bridge_port *to, struct sk_buff *skb)
{
	if (to && should_deliver(to, skb)) {
		__br_deliver(to, skb);
		return;
	}

	kfree_skb(skb);
}

/* called with rcu_read_lock */
void br_forward(const struct net_bridge_port *to, struct sk_buff *skb, struct sk_buff *skb0)
{
	if (to && should_deliver(to, skb) && !(to->flags & BR_ISOLATE_MODE)) {
		if (skb0)
			deliver_clone(to, skb, __br_forward);
		else
			__br_forward(to, skb);
		return;
	}

	if (!skb0)
		kfree_skb(skb);
}

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

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

	__packet_hook(prev, skb);
	return 0;
}

static struct net_bridge_port *maybe_deliver(
	struct net_bridge_port *prev, struct net_bridge_port *p,
	struct sk_buff *skb,
	void (*__packet_hook)(const struct net_bridge_port *p,
			      struct sk_buff *skb))
{
	int err;

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

	if (!prev)
		goto out;

	err = deliver_clone(prev, skb, __packet_hook);
	if (err)
		return ERR_PTR(err);

out:
	return p;
}

/* called under bridge lock */
static void br_flood(struct net_bridge *br, struct sk_buff *skb,
		     struct sk_buff *skb0,
		     void (*__packet_hook)(const struct net_bridge_port *p,
					   struct sk_buff *skb),
		     				bool unicast, bool forward)
{
	struct net_bridge_port *p;
	struct net_bridge_port *prev;
	enum br_pkt_type pkt_type = BR_PKT_UNICAST;
	const unsigned char *dest;

	dest = eth_hdr(skb)->h_dest;
	if (is_broadcast_ether_addr(dest))
		pkt_type = BR_PKT_BROADCAST;
	else if (is_multicast_ether_addr(dest))
		pkt_type = BR_PKT_MULTICAST;

	prev = NULL;
	/* AVM/TMA 20170315 JZ-33183
	 * There are some cases where we come here for multicast traffic, for example
	 * if all mdb entries are gone due to down'ing the bridge (multid may do this).
	 * In the past, br_forward_as_unicast() accessed an uninitialized pointer.
	 */
	BR_OUTPUT_SKB_CB(skb)->pg = NULL;

	list_for_each_entry_rcu(p, &br->port_list, list) {
		if (forward && (p->flags & BR_ISOLATE_MODE))
			continue;
		/* 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_INPUT_SKB_CB(skb)->proxyarp_replied)
			continue;

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

	if (!prev)
		goto out;

	if (skb0)
		deliver_clone(prev, skb, __packet_hook);
	else
		__packet_hook(prev, skb);
	return;

out:
	if (!skb0)
		kfree_skb(skb);
}


/* called with rcu_read_lock */
void br_flood_deliver(struct net_bridge *br, struct sk_buff *skb, bool unicast)
{
	/* FIXME: Try to accelerate */
	avm_pa_do_not_accelerate(skb);
	br_flood(br, skb, NULL, __br_deliver, unicast, false);
}
EXPORT_SYMBOL_GPL(br_deliver);

/* called under bridge lock */
void br_flood_forward(struct net_bridge *br, struct sk_buff *skb,
		      struct sk_buff *skb2, bool unicast)
{
	/* FIXME: Try to accelerate */
	avm_pa_do_not_accelerate(skb);
	if (br_flood_rl(br, skb, skb->dev))
		br_flood(br, skb, skb2, __br_forward, unicast, true);
	else if (!skb2)
		kfree_skb(skb);
}
EXPORT_SYMBOL_GPL(br_forward);

#ifdef CONFIG_BRIDGE_IGMP_SNOOPING
/* called with rcu_read_lock */
static void br_multicast_flood(struct net_bridge_mdb_entry *mdst,
			       struct sk_buff *skb, struct sk_buff *skb0,
			       void (*__packet_hook)(
					const struct net_bridge_port *p,
					struct sk_buff *skb))
{
	struct net_device *dev = BR_INPUT_SKB_CB(skb)->brdev;
	struct net_bridge *br = netdev_priv(dev);
	struct net_bridge_port *prev = NULL;
	struct net_bridge_port_group *p;
	struct hlist_node *rp;

	rp = rcu_dereference(hlist_first_rcu(&br->router_list));
	p = mdst ? rcu_dereference(mdst->ports) : NULL;
	BR_OUTPUT_SKB_CB(skb)->pg = NULL;

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

		lport = p ? p->port : NULL;
		rport = rp ? hlist_entry(rp, struct net_bridge_port, rlist) :
			     NULL;

		port = (unsigned long)lport > (unsigned long)rport ?
		       lport : rport;

		prev = maybe_deliver(prev, port, skb, __packet_hook);
		if (IS_ERR(prev))
			goto out;

		/* store port_group which is needed by br_forward_as_unicast() to avoid
		 * costly lookup in hot path. BR_OUTPUT_SKB_CB(skb)->pg must correspond to the
		 * prev port, because the skb will be delivered to that one next. Therefore a)
		 * assign BR_OUTPUT_SKB_CB(skb)->pg before p is updated and b) only change it when
		 * maybe_deliver() was not a no-op */
		if (prev == lport)
			BR_OUTPUT_SKB_CB(skb)->pg = p;
		else if (prev == rport)
			BR_OUTPUT_SKB_CB(skb)->pg = NULL;
		/* else: prev == (old) prev; i.e. should_deliver() was no-op and port was skipped */

		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 (skb0)
		deliver_clone(prev, skb, __packet_hook);
	else
		__packet_hook(prev, skb);
	return;

out:
	if (!skb0)
		kfree_skb(skb);
}

/* called with rcu_read_lock */
void br_multicast_deliver(struct net_bridge_mdb_entry *mdst,
			  struct sk_buff *skb)
{
	br_multicast_flood(mdst, skb, NULL, __br_deliver);
}

/* called with rcu_read_lock */
void br_multicast_forward(struct net_bridge_mdb_entry *mdst,
			  struct sk_buff *skb, struct sk_buff *skb2)
{
	br_multicast_flood(mdst, skb, skb2, __br_forward);
}
#endif