[net-next-2.6.git] / net / ipv6 / anycast.c

/*
 *	Anycast support for IPv6
 *	Linux INET6 implementation
 *
 *	Authors:
 *	David L Stevens (dlstevens@us.ibm.com)
 *
 *	based heavily on net/ipv6/mcast.c
 *
 *	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/capability.h>
#include <linux/module.h>
#include <linux/errno.h>
#include <linux/types.h>
#include <linux/random.h>
#include <linux/string.h>
#include <linux/socket.h>
#include <linux/sockios.h>
#include <linux/net.h>
#include <linux/in6.h>
#include <linux/netdevice.h>
#include <linux/if_arp.h>
#include <linux/route.h>
#include <linux/init.h>
#include <linux/proc_fs.h>
#include <linux/seq_file.h>
#include <linux/slab.h>

#include <net/net_namespace.h>
#include <net/sock.h>
#include <net/snmp.h>

#include <net/ipv6.h>
#include <net/protocol.h>
#include <net/if_inet6.h>
#include <net/ndisc.h>
#include <net/addrconf.h>
#include <net/ip6_route.h>

#include <net/checksum.h>

static int ipv6_dev_ac_dec(struct net_device *dev, struct in6_addr *addr);

/* Big ac list lock for all the sockets */
static DEFINE_RWLOCK(ipv6_sk_ac_lock);


/*
 *	socket join an anycast group
 */

int ipv6_sock_ac_join(struct sock *sk, int ifindex, struct in6_addr *addr)
{
	struct ipv6_pinfo *np = inet6_sk(sk);
	struct net_device *dev = NULL;
	struct inet6_dev *idev;
	struct ipv6_ac_socklist *pac;
	struct net *net = sock_net(sk);
	int	ishost = !net->ipv6.devconf_all->forwarding;
	int	err = 0;

	if (!capable(CAP_NET_ADMIN))
		return -EPERM;
	if (ipv6_addr_is_multicast(addr))
		return -EINVAL;
	if (ipv6_chk_addr(net, addr, NULL, 0))
		return -EINVAL;

	pac = sock_kmalloc(sk, sizeof(struct ipv6_ac_socklist), GFP_KERNEL);
	if (pac == NULL)
		return -ENOMEM;
	pac->acl_next = NULL;
	ipv6_addr_copy(&pac->acl_addr, addr);

	rcu_read_lock();
	if (ifindex == 0) {
		struct rt6_info *rt;

		rt = rt6_lookup(net, addr, NULL, 0, 0);
		if (rt) {
			dev = rt->rt6i_dev;
			dst_release(&rt->dst);
		} else if (ishost) {
			err = -EADDRNOTAVAIL;
			goto error;
		} else {
			/* router, no matching interface: just pick one */
			dev = dev_get_by_flags_rcu(net, IFF_UP,
						   IFF_UP | IFF_LOOPBACK);
		}
	} else
		dev = dev_get_by_index_rcu(net, ifindex);

	if (dev == NULL) {
		err = -ENODEV;
		goto error;
	}

	idev = __in6_dev_get(dev);
	if (!idev) {
		if (ifindex)
			err = -ENODEV;
		else
			err = -EADDRNOTAVAIL;
		goto error;
	}
	/* reset ishost, now that we have a specific device */
	ishost = !idev->cnf.forwarding;

	pac->acl_ifindex = dev->ifindex;

	/* XXX
	 * For hosts, allow link-local or matching prefix anycasts.
	 * This obviates the need for propagating anycast routes while
	 * still allowing some non-router anycast participation.
	 */
	if (!ipv6_chk_prefix(addr, dev)) {
		if (ishost)
			err = -EADDRNOTAVAIL;
		if (err)
			goto error;
	}

	err = ipv6_dev_ac_inc(dev, addr);
	if (!err) {
		write_lock_bh(&ipv6_sk_ac_lock);
		pac->acl_next = np->ipv6_ac_list;
		np->ipv6_ac_list = pac;
		write_unlock_bh(&ipv6_sk_ac_lock);
		pac = NULL;
	}

error:
	rcu_read_unlock();
	if (pac)
		sock_kfree_s(sk, pac, sizeof(*pac));
	return err;
}

/*
 *	socket leave an anycast group
 */
int ipv6_sock_ac_drop(struct sock *sk, int ifindex, struct in6_addr *addr)
{
	struct ipv6_pinfo *np = inet6_sk(sk);
	struct net_device *dev;
	struct ipv6_ac_socklist *pac, *prev_pac;
	struct net *net = sock_net(sk);

	write_lock_bh(&ipv6_sk_ac_lock);
	prev_pac = NULL;
	for (pac = np->ipv6_ac_list; pac; pac = pac->acl_next) {
		if ((ifindex == 0 || pac->acl_ifindex == ifindex) &&
		     ipv6_addr_equal(&pac->acl_addr, addr))
			break;
		prev_pac = pac;
	}
	if (!pac) {
		write_unlock_bh(&ipv6_sk_ac_lock);
		return -ENOENT;
	}
	if (prev_pac)
		prev_pac->acl_next = pac->acl_next;
	else
		np->ipv6_ac_list = pac->acl_next;

	write_unlock_bh(&ipv6_sk_ac_lock);

	rcu_read_lock();
	dev = dev_get_by_index_rcu(net, pac->acl_ifindex);
	if (dev)
		ipv6_dev_ac_dec(dev, &pac->acl_addr);
	rcu_read_unlock();

	sock_kfree_s(sk, pac, sizeof(*pac));
	return 0;
}

void ipv6_sock_ac_close(struct sock *sk)
{
	struct ipv6_pinfo *np = inet6_sk(sk);
	struct net_device *dev = NULL;
	struct ipv6_ac_socklist *pac;
	struct net *net = sock_net(sk);
	int	prev_index;

	write_lock_bh(&ipv6_sk_ac_lock);
	pac = np->ipv6_ac_list;
	np->ipv6_ac_list = NULL;
	write_unlock_bh(&ipv6_sk_ac_lock);

	prev_index = 0;
	rcu_read_lock();
	while (pac) {
		struct ipv6_ac_socklist *next = pac->acl_next;

		if (pac->acl_ifindex != prev_index) {
			dev = dev_get_by_index_rcu(net, pac->acl_ifindex);
			prev_index = pac->acl_ifindex;
		}
		if (dev)
			ipv6_dev_ac_dec(dev, &pac->acl_addr);
		sock_kfree_s(sk, pac, sizeof(*pac));
		pac = next;
	}
	rcu_read_unlock();
}

#if 0
/* The function is not used, which is funny. Apparently, author
 * supposed to use it to filter out datagrams inside udp/raw but forgot.
 *
 * It is OK, anycasts are not special comparing to delivery to unicasts.
 */

int inet6_ac_check(struct sock *sk, struct in6_addr *addr, int ifindex)
{
	struct ipv6_ac_socklist *pac;
	struct ipv6_pinfo *np = inet6_sk(sk);
	int	found;

	found = 0;
	read_lock(&ipv6_sk_ac_lock);
	for (pac=np->ipv6_ac_list; pac; pac=pac->acl_next) {
		if (ifindex && pac->acl_ifindex != ifindex)
			continue;
		found = ipv6_addr_equal(&pac->acl_addr, addr);
		if (found)
			break;
	}
	read_unlock(&ipv6_sk_ac_lock);

	return found;
}

#endif

static void aca_put(struct ifacaddr6 *ac)
{
	if (atomic_dec_and_test(&ac->aca_refcnt)) {
		in6_dev_put(ac->aca_idev);
		dst_release(&ac->aca_rt->dst);
		kfree(ac);
	}
}

/*
 *	device anycast group inc (add if not found)
 */
int ipv6_dev_ac_inc(struct net_device *dev, struct in6_addr *addr)
{
	struct ifacaddr6 *aca;
	struct inet6_dev *idev;
	struct rt6_info *rt;
	int err;

	idev = in6_dev_get(dev);

	if (idev == NULL)
		return -EINVAL;

	write_lock_bh(&idev->lock);
	if (idev->dead) {
		err = -ENODEV;
		goto out;
	}

	for (aca = idev->ac_list; aca; aca = aca->aca_next) {
		if (ipv6_addr_equal(&aca->aca_addr, addr)) {
			aca->aca_users++;
			err = 0;
			goto out;
		}
	}

	/*
	 *	not found: create a new one.
	 */

	aca = kzalloc(sizeof(struct ifacaddr6), GFP_ATOMIC);

	if (aca == NULL) {
		err = -ENOMEM;
		goto out;
	}

	rt = addrconf_dst_alloc(idev, addr, 1);
	if (IS_ERR(rt)) {
		kfree(aca);
		err = PTR_ERR(rt);
		goto out;
	}

	ipv6_addr_copy(&aca->aca_addr, addr);
	aca->aca_idev = idev;
	aca->aca_rt = rt;
	aca->aca_users = 1;
	/* aca_tstamp should be updated upon changes */
	aca->aca_cstamp = aca->aca_tstamp = jiffies;
	atomic_set(&aca->aca_refcnt, 2);
	spin_lock_init(&aca->aca_lock);

	aca->aca_next = idev->ac_list;
	idev->ac_list = aca;
	write_unlock_bh(&idev->lock);

	ip6_ins_rt(rt);

	addrconf_join_solict(dev, &aca->aca_addr);

	aca_put(aca);
	return 0;
out:
	write_unlock_bh(&idev->lock);
	in6_dev_put(idev);
	return err;
}

/*
 *	device anycast group decrement
 */
int __ipv6_dev_ac_dec(struct inet6_dev *idev, struct in6_addr *addr)
{
	struct ifacaddr6 *aca, *prev_aca;

	write_lock_bh(&idev->lock);
	prev_aca = NULL;
	for (aca = idev->ac_list; aca; aca = aca->aca_next) {
		if (ipv6_addr_equal(&aca->aca_addr, addr))
			break;
		prev_aca = aca;
	}
	if (!aca) {
		write_unlock_bh(&idev->lock);
		return -ENOENT;
	}
	if (--aca->aca_users > 0) {
		write_unlock_bh(&idev->lock);
		return 0;
	}
	if (prev_aca)
		prev_aca->aca_next = aca->aca_next;
	else
		idev->ac_list = aca->aca_next;
	write_unlock_bh(&idev->lock);
	addrconf_leave_solict(idev, &aca->aca_addr);

	dst_hold(&aca->aca_rt->dst);
	ip6_del_rt(aca->aca_rt);

	aca_put(aca);
	return 0;
}

/* called with rcu_read_lock() */
static int ipv6_dev_ac_dec(struct net_device *dev, struct in6_addr *addr)
{
	struct inet6_dev *idev = __in6_dev_get(dev);

	if (idev == NULL)
		return -ENODEV;
	return __ipv6_dev_ac_dec(idev, addr);
}

/*
 *	check if the interface has this anycast address
 *	called with rcu_read_lock()
 */
static int ipv6_chk_acast_dev(struct net_device *dev, struct in6_addr *addr)
{
	struct inet6_dev *idev;
	struct ifacaddr6 *aca;

	idev = __in6_dev_get(dev);
	if (idev) {
		read_lock_bh(&idev->lock);
		for (aca = idev->ac_list; aca; aca = aca->aca_next)
			if (ipv6_addr_equal(&aca->aca_addr, addr))
				break;
		read_unlock_bh(&idev->lock);
		return aca != NULL;
	}
	return 0;
}

/*
 *	check if given interface (or any, if dev==0) has this anycast address
 */
int ipv6_chk_acast_addr(struct net *net, struct net_device *dev,
			struct in6_addr *addr)
{
	int found = 0;

	rcu_read_lock();
	if (dev)
		found = ipv6_chk_acast_dev(dev, addr);
	else
		for_each_netdev_rcu(net, dev)
			if (ipv6_chk_acast_dev(dev, addr)) {
				found = 1;
				break;
			}
	rcu_read_unlock();
	return found;
}


#ifdef CONFIG_PROC_FS
struct ac6_iter_state {
	struct seq_net_private p;
	struct net_device *dev;
	struct inet6_dev *idev;
};

#define ac6_seq_private(seq)	((struct ac6_iter_state *)(seq)->private)

static inline struct ifacaddr6 *ac6_get_first(struct seq_file *seq)
{
	struct ifacaddr6 *im = NULL;
	struct ac6_iter_state *state = ac6_seq_private(seq);
	struct net *net = seq_file_net(seq);

	state->idev = NULL;
	for_each_netdev_rcu(net, state->dev) {
		struct inet6_dev *idev;
		idev = __in6_dev_get(state->dev);
		if (!idev)
			continue;
		read_lock_bh(&idev->lock);
		im = idev->ac_list;
		if (im) {
			state->idev = idev;
			break;
		}
		read_unlock_bh(&idev->lock);
	}
	return im;
}

static struct ifacaddr6 *ac6_get_next(struct seq_file *seq, struct ifacaddr6 *im)
{
	struct ac6_iter_state *state = ac6_seq_private(seq);

	im = im->aca_next;
	while (!im) {
		if (likely(state->idev != NULL))
			read_unlock_bh(&state->idev->lock);

		state->dev = next_net_device_rcu(state->dev);
		if (!state->dev) {
			state->idev = NULL;
			break;
		}
		state->idev = __in6_dev_get(state->dev);
		if (!state->idev)
			continue;
		read_lock_bh(&state->idev->lock);
		im = state->idev->ac_list;
	}
	return im;
}

static struct ifacaddr6 *ac6_get_idx(struct seq_file *seq, loff_t pos)
{
	struct ifacaddr6 *im = ac6_get_first(seq);
	if (im)
		while (pos && (im = ac6_get_next(seq, im)) != NULL)
			--pos;
	return pos ? NULL : im;
}

static void *ac6_seq_start(struct seq_file *seq, loff_t *pos)
	__acquires(RCU)
{
	rcu_read_lock();
	return ac6_get_idx(seq, *pos);
}

static void *ac6_seq_next(struct seq_file *seq, void *v, loff_t *pos)
{
	struct ifacaddr6 *im = ac6_get_next(seq, v);

	++*pos;
	return im;
}

static void ac6_seq_stop(struct seq_file *seq, void *v)
	__releases(RCU)
{
	struct ac6_iter_state *state = ac6_seq_private(seq);

	if (likely(state->idev != NULL)) {
		read_unlock_bh(&state->idev->lock);
		state->idev = NULL;
	}
	rcu_read_unlock();
}

static int ac6_seq_show(struct seq_file *seq, void *v)
{
	struct ifacaddr6 *im = (struct ifacaddr6 *)v;
	struct ac6_iter_state *state = ac6_seq_private(seq);

	seq_printf(seq, "%-4d %-15s %pi6 %5d\n",
		   state->dev->ifindex, state->dev->name,
		   &im->aca_addr, im->aca_users);
	return 0;
}

static const struct seq_operations ac6_seq_ops = {
	.start	=	ac6_seq_start,
	.next	=	ac6_seq_next,
	.stop	=	ac6_seq_stop,
	.show	=	ac6_seq_show,
};

static int ac6_seq_open(struct inode *inode, struct file *file)
{
	return seq_open_net(inode, file, &ac6_seq_ops,
			    sizeof(struct ac6_iter_state));
}

static const struct file_operations ac6_seq_fops = {
	.owner		=	THIS_MODULE,
	.open		=	ac6_seq_open,
	.read		=	seq_read,
	.llseek		=	seq_lseek,
	.release	=	seq_release_net,
};

int __net_init ac6_proc_init(struct net *net)
{
	if (!proc_net_fops_create(net, "anycast6", S_IRUGO, &ac6_seq_fops))
		return -ENOMEM;

	return 0;
}

void ac6_proc_exit(struct net *net)
{
	proc_net_remove(net, "anycast6");
}
#endif
Commit	Line	Data
1da177e4 LT	1	/*
1da177e4 LT	2	* Anycast support for IPv6
1ab1457c	3	* Linux INET6 implementation
1da177e4 LT	4	*
	5	* Authors:
	6	* David L Stevens (dlstevens@us.ibm.com)
	7	*
	8	* based heavily on net/ipv6/mcast.c
	9	*
	10	* This program is free software; you can redistribute it and/or
	11	* modify it under the terms of the GNU General Public License
	12	* as published by the Free Software Foundation; either version
	13	* 2 of the License, or (at your option) any later version.
	14	*/
	15
4fc268d2	16	#include <linux/capability.h>
1da177e4 LT	17	#include <linux/module.h>
	18	#include <linux/errno.h>
	19	#include <linux/types.h>
	20	#include <linux/random.h>
	21	#include <linux/string.h>
	22	#include <linux/socket.h>
	23	#include <linux/sockios.h>
1da177e4 LT	24	#include <linux/net.h>
	25	#include <linux/in6.h>
	26	#include <linux/netdevice.h>
	27	#include <linux/if_arp.h>
	28	#include <linux/route.h>
	29	#include <linux/init.h>
	30	#include <linux/proc_fs.h>
	31	#include <linux/seq_file.h>
5a0e3ad6	32	#include <linux/slab.h>
1da177e4	33
457c4cbc	34	#include <net/net_namespace.h>
1da177e4 LT	35	#include <net/sock.h>
	36	#include <net/snmp.h>
	37
	38	#include <net/ipv6.h>
	39	#include <net/protocol.h>
	40	#include <net/if_inet6.h>
	41	#include <net/ndisc.h>
	42	#include <net/addrconf.h>
	43	#include <net/ip6_route.h>
	44
	45	#include <net/checksum.h>
	46
	47	static int ipv6_dev_ac_dec(struct net_device dev, struct in6_addr addr);
	48
	49	/* Big ac list lock for all the sockets */
	50	static DEFINE_RWLOCK(ipv6_sk_ac_lock);
	51
1da177e4 LT	52
	53	/*
	54	* socket join an anycast group
	55	*/
	56
	57	int ipv6_sock_ac_join(struct sock sk, int ifindex, struct in6_addr addr)
	58	{
	59	struct ipv6_pinfo *np = inet6_sk(sk);
	60	struct net_device *dev = NULL;
	61	struct inet6_dev *idev;
	62	struct ipv6_ac_socklist *pac;
6ab57e7e	63	struct net *net = sock_net(sk);
53b7997f	64	int ishost = !net->ipv6.devconf_all->forwarding;
1da177e4 LT	65	int err = 0;
	66
	67	if (!capable(CAP_NET_ADMIN))
	68	return -EPERM;
	69	if (ipv6_addr_is_multicast(addr))
	70	return -EINVAL;
6ab57e7e	71	if (ipv6_chk_addr(net, addr, NULL, 0))
1da177e4 LT	72	return -EINVAL;
	73
	74	pac = sock_kmalloc(sk, sizeof(struct ipv6_ac_socklist), GFP_KERNEL);
	75	if (pac == NULL)
	76	return -ENOMEM;
	77	pac->acl_next = NULL;
	78	ipv6_addr_copy(&pac->acl_addr, addr);
	79
bb69ae04	80	rcu_read_lock();
1da177e4 LT	81	if (ifindex == 0) {
	82	struct rt6_info *rt;
	83
6ab57e7e	84	rt = rt6_lookup(net, addr, NULL, 0, 0);
1da177e4 LT	85	if (rt) {
1da177e4 LT	86	dev = rt->rt6i_dev;
d8d1f30b	87	dst_release(&rt->dst);
1da177e4 LT	88	} else if (ishost) {
1da177e4 LT	89	err = -EADDRNOTAVAIL;
bb69ae04	90	goto error;
1da177e4 LT	91	} else {
1da177e4 LT	92	/* router, no matching interface: just pick one */
bb69ae04 ED	93	dev = dev_get_by_flags_rcu(net, IFF_UP,
bb69ae04 ED	94	IFF_UP \| IFF_LOOPBACK);
1da177e4 LT	95	}
1da177e4 LT	96	} else
bb69ae04	97	dev = dev_get_by_index_rcu(net, ifindex);
1da177e4 LT	98
	99	if (dev == NULL) {
	100	err = -ENODEV;
bb69ae04	101	goto error;
1da177e4 LT	102	}
1da177e4 LT	103
bb69ae04	104	idev = __in6_dev_get(dev);
1da177e4 LT	105	if (!idev) {
	106	if (ifindex)
	107	err = -ENODEV;
	108	else
	109	err = -EADDRNOTAVAIL;
bb69ae04	110	goto error;
1da177e4 LT	111	}
	112	/* reset ishost, now that we have a specific device */
	113	ishost = !idev->cnf.forwarding;
1da177e4 LT	114
	115	pac->acl_ifindex = dev->ifindex;
	116
	117	/* XXX
	118	* For hosts, allow link-local or matching prefix anycasts.
	119	* This obviates the need for propagating anycast routes while
	120	* still allowing some non-router anycast participation.
	121	*/
52eeeb84	122	if (!ipv6_chk_prefix(addr, dev)) {
1da177e4 LT	123	if (ishost)
	124	err = -EADDRNOTAVAIL;
	125	if (err)
bb69ae04	126	goto error;
1da177e4 LT	127	}
	128
	129	err = ipv6_dev_ac_inc(dev, addr);
bb69ae04 ED	130	if (!err) {
	131	write_lock_bh(&ipv6_sk_ac_lock);
	132	pac->acl_next = np->ipv6_ac_list;
	133	np->ipv6_ac_list = pac;
	134	write_unlock_bh(&ipv6_sk_ac_lock);
	135	pac = NULL;
	136	}
1da177e4	137
bb69ae04 ED	138	error:
	139	rcu_read_unlock();
	140	if (pac)
	141	sock_kfree_s(sk, pac, sizeof(*pac));
1da177e4 LT	142	return err;
	143	}
	144
	145	/*
	146	* socket leave an anycast group
	147	*/
	148	int ipv6_sock_ac_drop(struct sock sk, int ifindex, struct in6_addr addr)
	149	{
	150	struct ipv6_pinfo *np = inet6_sk(sk);
	151	struct net_device *dev;
	152	struct ipv6_ac_socklist pac, prev_pac;
6ab57e7e	153	struct net *net = sock_net(sk);
1da177e4 LT	154
	155	write_lock_bh(&ipv6_sk_ac_lock);
	156	prev_pac = NULL;
	157	for (pac = np->ipv6_ac_list; pac; pac = pac->acl_next) {
	158	if ((ifindex == 0 \|\| pac->acl_ifindex == ifindex) &&
	159	ipv6_addr_equal(&pac->acl_addr, addr))
	160	break;
	161	prev_pac = pac;
	162	}
	163	if (!pac) {
	164	write_unlock_bh(&ipv6_sk_ac_lock);
	165	return -ENOENT;
	166	}
	167	if (prev_pac)
	168	prev_pac->acl_next = pac->acl_next;
	169	else
	170	np->ipv6_ac_list = pac->acl_next;
	171
	172	write_unlock_bh(&ipv6_sk_ac_lock);
	173
bb69ae04 ED	174	rcu_read_lock();
	175	dev = dev_get_by_index_rcu(net, pac->acl_ifindex);
	176	if (dev)
1da177e4	177	ipv6_dev_ac_dec(dev, &pac->acl_addr);
bb69ae04 ED	178	rcu_read_unlock();
bb69ae04 ED	179
1da177e4 LT	180	sock_kfree_s(sk, pac, sizeof(*pac));
	181	return 0;
	182	}
	183
	184	void ipv6_sock_ac_close(struct sock *sk)
	185	{
	186	struct ipv6_pinfo *np = inet6_sk(sk);
	187	struct net_device *dev = NULL;
	188	struct ipv6_ac_socklist *pac;
6ab57e7e	189	struct net *net = sock_net(sk);
1da177e4 LT	190	int prev_index;
	191
	192	write_lock_bh(&ipv6_sk_ac_lock);
	193	pac = np->ipv6_ac_list;
	194	np->ipv6_ac_list = NULL;
	195	write_unlock_bh(&ipv6_sk_ac_lock);
	196
	197	prev_index = 0;
bb69ae04	198	rcu_read_lock();
1da177e4 LT	199	while (pac) {
	200	struct ipv6_ac_socklist *next = pac->acl_next;
	201
	202	if (pac->acl_ifindex != prev_index) {
bb69ae04	203	dev = dev_get_by_index_rcu(net, pac->acl_ifindex);
1da177e4 LT	204	prev_index = pac->acl_ifindex;
	205	}
	206	if (dev)
	207	ipv6_dev_ac_dec(dev, &pac->acl_addr);
	208	sock_kfree_s(sk, pac, sizeof(*pac));
	209	pac = next;
	210	}
bb69ae04	211	rcu_read_unlock();
1da177e4 LT	212	}
	213
	214	#if 0
	215	/* The function is not used, which is funny. Apparently, author
	216	* supposed to use it to filter out datagrams inside udp/raw but forgot.
	217	*
	218	* It is OK, anycasts are not special comparing to delivery to unicasts.
	219	*/
	220
	221	int inet6_ac_check(struct sock sk, struct in6_addr addr, int ifindex)
	222	{
	223	struct ipv6_ac_socklist *pac;
	224	struct ipv6_pinfo *np = inet6_sk(sk);
	225	int found;
	226
	227	found = 0;
	228	read_lock(&ipv6_sk_ac_lock);
	229	for (pac=np->ipv6_ac_list; pac; pac=pac->acl_next) {
	230	if (ifindex && pac->acl_ifindex != ifindex)
	231	continue;
	232	found = ipv6_addr_equal(&pac->acl_addr, addr);
	233	if (found)
	234	break;
	235	}
	236	read_unlock(&ipv6_sk_ac_lock);
	237
	238	return found;
	239	}
	240
	241	#endif
	242
	243	static void aca_put(struct ifacaddr6 *ac)
	244	{
	245	if (atomic_dec_and_test(&ac->aca_refcnt)) {
	246	in6_dev_put(ac->aca_idev);
d8d1f30b	247	dst_release(&ac->aca_rt->dst);
1da177e4 LT	248	kfree(ac);
	249	}
	250	}
	251
	252	/*
	253	* device anycast group inc (add if not found)
	254	*/
	255	int ipv6_dev_ac_inc(struct net_device dev, struct in6_addr addr)
	256	{
	257	struct ifacaddr6 *aca;
	258	struct inet6_dev *idev;
	259	struct rt6_info *rt;
	260	int err;
	261
	262	idev = in6_dev_get(dev);
	263
	264	if (idev == NULL)
	265	return -EINVAL;
	266
	267	write_lock_bh(&idev->lock);
	268	if (idev->dead) {
	269	err = -ENODEV;
	270	goto out;
	271	}
	272
	273	for (aca = idev->ac_list; aca; aca = aca->aca_next) {
	274	if (ipv6_addr_equal(&aca->aca_addr, addr)) {
	275	aca->aca_users++;
	276	err = 0;
	277	goto out;
	278	}
	279	}
	280
	281	/*
	282	* not found: create a new one.
	283	*/
	284
0c600eda	285	aca = kzalloc(sizeof(struct ifacaddr6), GFP_ATOMIC);
1da177e4 LT	286
	287	if (aca == NULL) {
	288	err = -ENOMEM;
	289	goto out;
	290	}
	291
	292	rt = addrconf_dst_alloc(idev, addr, 1);
	293	if (IS_ERR(rt)) {
	294	kfree(aca);
	295	err = PTR_ERR(rt);
	296	goto out;
	297	}
	298
1da177e4 LT	299	ipv6_addr_copy(&aca->aca_addr, addr);
	300	aca->aca_idev = idev;
	301	aca->aca_rt = rt;
	302	aca->aca_users = 1;
	303	/* aca_tstamp should be updated upon changes */
	304	aca->aca_cstamp = aca->aca_tstamp = jiffies;
	305	atomic_set(&aca->aca_refcnt, 2);
	306	spin_lock_init(&aca->aca_lock);
	307
	308	aca->aca_next = idev->ac_list;
	309	idev->ac_list = aca;
	310	write_unlock_bh(&idev->lock);
	311
84f59370	312	ip6_ins_rt(rt);
1da177e4 LT	313
	314	addrconf_join_solict(dev, &aca->aca_addr);
	315
	316	aca_put(aca);
	317	return 0;
	318	out:
	319	write_unlock_bh(&idev->lock);
	320	in6_dev_put(idev);
	321	return err;
	322	}
	323
	324	/*
	325	* device anycast group decrement
	326	*/
	327	int __ipv6_dev_ac_dec(struct inet6_dev idev, struct in6_addr addr)
	328	{
	329	struct ifacaddr6 aca, prev_aca;
	330
	331	write_lock_bh(&idev->lock);
	332	prev_aca = NULL;
	333	for (aca = idev->ac_list; aca; aca = aca->aca_next) {
	334	if (ipv6_addr_equal(&aca->aca_addr, addr))
	335	break;
	336	prev_aca = aca;
	337	}
	338	if (!aca) {
	339	write_unlock_bh(&idev->lock);
	340	return -ENOENT;
	341	}
	342	if (--aca->aca_users > 0) {
	343	write_unlock_bh(&idev->lock);
	344	return 0;
	345	}
	346	if (prev_aca)
	347	prev_aca->aca_next = aca->aca_next;
	348	else
	349	idev->ac_list = aca->aca_next;
	350	write_unlock_bh(&idev->lock);
	351	addrconf_leave_solict(idev, &aca->aca_addr);
	352
d8d1f30b	353	dst_hold(&aca->aca_rt->dst);
84f59370	354	ip6_del_rt(aca->aca_rt);
1da177e4 LT	355
	356	aca_put(aca);
	357	return 0;
	358	}
	359
bb69ae04	360	/* called with rcu_read_lock() */
1da177e4 LT	361	static int ipv6_dev_ac_dec(struct net_device dev, struct in6_addr addr)
1da177e4 LT	362	{
bb69ae04 ED	363	struct inet6_dev *idev = __in6_dev_get(dev);
bb69ae04 ED	364
1da177e4 LT	365	if (idev == NULL)
1da177e4 LT	366	return -ENODEV;
bb69ae04	367	return __ipv6_dev_ac_dec(idev, addr);
1da177e4	368	}
1ab1457c	369
1da177e4 LT	370	/*
1da177e4 LT	371	* check if the interface has this anycast address
bb69ae04	372	* called with rcu_read_lock()
1da177e4 LT	373	*/
	374	static int ipv6_chk_acast_dev(struct net_device dev, struct in6_addr addr)
	375	{
	376	struct inet6_dev *idev;
	377	struct ifacaddr6 *aca;
	378
bb69ae04	379	idev = __in6_dev_get(dev);
1da177e4 LT	380	if (idev) {
	381	read_lock_bh(&idev->lock);
	382	for (aca = idev->ac_list; aca; aca = aca->aca_next)
	383	if (ipv6_addr_equal(&aca->aca_addr, addr))
	384	break;
	385	read_unlock_bh(&idev->lock);
cfcabdcc	386	return aca != NULL;
1da177e4 LT	387	}
	388	return 0;
	389	}
	390
	391	/*
	392	* check if given interface (or any, if dev==0) has this anycast address
	393	*/
6ab57e7e DL	394	int ipv6_chk_acast_addr(struct net net, struct net_device dev,
6ab57e7e DL	395	struct in6_addr *addr)
1da177e4	396	{
7562f876 PE	397	int found = 0;
7562f876 PE	398
c6d14c84	399	rcu_read_lock();
bb69ae04 ED	400	if (dev)
	401	found = ipv6_chk_acast_dev(dev, addr);
	402	else
	403	for_each_netdev_rcu(net, dev)
	404	if (ipv6_chk_acast_dev(dev, addr)) {
	405	found = 1;
	406	break;
	407	}
c6d14c84	408	rcu_read_unlock();
7562f876	409	return found;
1da177e4 LT	410	}
	411
	412
	413	#ifdef CONFIG_PROC_FS
	414	struct ac6_iter_state {
6ab57e7e	415	struct seq_net_private p;
1da177e4 LT	416	struct net_device *dev;
	417	struct inet6_dev *idev;
	418	};
	419
	420	#define ac6_seq_private(seq) ((struct ac6_iter_state *)(seq)->private)
	421
	422	static inline struct ifacaddr6 ac6_get_first(struct seq_file seq)
	423	{
	424	struct ifacaddr6 *im = NULL;
	425	struct ac6_iter_state *state = ac6_seq_private(seq);
6ab57e7e	426	struct net *net = seq_file_net(seq);
1da177e4	427
7562f876	428	state->idev = NULL;
ce81b76a	429	for_each_netdev_rcu(net, state->dev) {
1da177e4	430	struct inet6_dev *idev;
ce81b76a	431	idev = __in6_dev_get(state->dev);
1da177e4 LT	432	if (!idev)
	433	continue;
	434	read_lock_bh(&idev->lock);
	435	im = idev->ac_list;
	436	if (im) {
	437	state->idev = idev;
	438	break;
	439	}
	440	read_unlock_bh(&idev->lock);
	441	}
	442	return im;
	443	}
	444
	445	static struct ifacaddr6 ac6_get_next(struct seq_file seq, struct ifacaddr6 *im)
	446	{
	447	struct ac6_iter_state *state = ac6_seq_private(seq);
	448
	449	im = im->aca_next;
	450	while (!im) {
ce81b76a	451	if (likely(state->idev != NULL))
1da177e4	452	read_unlock_bh(&state->idev->lock);
ce81b76a ED	453
ce81b76a ED	454	state->dev = next_net_device_rcu(state->dev);
1da177e4 LT	455	if (!state->dev) {
	456	state->idev = NULL;
	457	break;
	458	}
ce81b76a	459	state->idev = __in6_dev_get(state->dev);
1da177e4 LT	460	if (!state->idev)
	461	continue;
	462	read_lock_bh(&state->idev->lock);
	463	im = state->idev->ac_list;
	464	}
	465	return im;
	466	}
	467
	468	static struct ifacaddr6 ac6_get_idx(struct seq_file seq, loff_t pos)
	469	{
	470	struct ifacaddr6 *im = ac6_get_first(seq);
	471	if (im)
	472	while (pos && (im = ac6_get_next(seq, im)) != NULL)
	473	--pos;
	474	return pos ? NULL : im;
	475	}
	476
	477	static void ac6_seq_start(struct seq_file seq, loff_t *pos)
ce81b76a	478	__acquires(RCU)
1da177e4	479	{
ce81b76a	480	rcu_read_lock();
1da177e4 LT	481	return ac6_get_idx(seq, *pos);
	482	}
	483
	484	static void ac6_seq_next(struct seq_file seq, void v, loff_t pos)
	485	{
ce81b76a ED	486	struct ifacaddr6 *im = ac6_get_next(seq, v);
ce81b76a ED	487
1da177e4 LT	488	++*pos;
	489	return im;
	490	}
	491
	492	static void ac6_seq_stop(struct seq_file seq, void v)
ce81b76a	493	__releases(RCU)
1da177e4 LT	494	{
1da177e4 LT	495	struct ac6_iter_state *state = ac6_seq_private(seq);
ce81b76a	496
1da177e4 LT	497	if (likely(state->idev != NULL)) {
1da177e4 LT	498	read_unlock_bh(&state->idev->lock);
ce81b76a	499	state->idev = NULL;
1da177e4	500	}
ce81b76a	501	rcu_read_unlock();
1da177e4 LT	502	}
	503
	504	static int ac6_seq_show(struct seq_file seq, void v)
	505	{
	506	struct ifacaddr6 im = (struct ifacaddr6 )v;
	507	struct ac6_iter_state *state = ac6_seq_private(seq);
	508
4b7a4274	509	seq_printf(seq, "%-4d %-15s %pi6 %5d\n",
1da177e4	510	state->dev->ifindex, state->dev->name,
b071195d	511	&im->aca_addr, im->aca_users);
1da177e4 LT	512	return 0;
	513	}
	514
56b3d975	515	static const struct seq_operations ac6_seq_ops = {
1da177e4 LT	516	.start = ac6_seq_start,
	517	.next = ac6_seq_next,
	518	.stop = ac6_seq_stop,
	519	.show = ac6_seq_show,
	520	};
	521
	522	static int ac6_seq_open(struct inode inode, struct file file)
	523	{
6ab57e7e DL	524	return seq_open_net(inode, file, &ac6_seq_ops,
6ab57e7e DL	525	sizeof(struct ac6_iter_state));
1da177e4 LT	526	}
1da177e4 LT	527
9a32144e	528	static const struct file_operations ac6_seq_fops = {
1da177e4 LT	529	.owner = THIS_MODULE,
	530	.open = ac6_seq_open,
	531	.read = seq_read,
	532	.llseek = seq_lseek,
6ab57e7e	533	.release = seq_release_net,
1da177e4 LT	534	};
1da177e4 LT	535
2c8c1e72	536	int __net_init ac6_proc_init(struct net *net)
1da177e4	537	{
6ab57e7e	538	if (!proc_net_fops_create(net, "anycast6", S_IRUGO, &ac6_seq_fops))
1da177e4 LT	539	return -ENOMEM;
	540
	541	return 0;
	542	}
	543
6ab57e7e	544	void ac6_proc_exit(struct net *net)
1da177e4	545	{
6ab57e7e	546	proc_net_remove(net, "anycast6");
1da177e4 LT	547	}
	548	#endif
	549