[linux-2.6-block.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);

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

		rt = rt6_lookup(net, addr, NULL, 0, 0);
		if (rt) {
			dev = rt->rt6i_dev;
			dev_hold(dev);
			dst_release(&rt->u.dst);
		} else if (ishost) {
			err = -EADDRNOTAVAIL;
			goto out_free_pac;
		} else {
			/* router, no matching interface: just pick one */

			dev = dev_get_by_flags(net, IFF_UP, IFF_UP|IFF_LOOPBACK);
		}
	} else
		dev = dev_get_by_index(net, ifindex);

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

	idev = in6_dev_get(dev);
	if (!idev) {
		if (ifindex)
			err = -ENODEV;
		else
			err = -EADDRNOTAVAIL;
		goto out_dev_put;
	}
	/* reset ishost, now that we have a specific device */
	ishost = !idev->cnf.forwarding;
	in6_dev_put(idev);

	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 out_dev_put;
	}

	err = ipv6_dev_ac_inc(dev, addr);
	if (err)
		goto out_dev_put;

	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);

	dev_put(dev);

	return 0;

out_dev_put:
	dev_put(dev);
out_free_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);

	dev = dev_get_by_index(net, pac->acl_ifindex);
	if (dev) {
		ipv6_dev_ac_dec(dev, &pac->acl_addr);
		dev_put(dev);
	}
	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;
	while (pac) {
		struct ipv6_ac_socklist *next = pac->acl_next;

		if (pac->acl_ifindex != prev_index) {
			if (dev)
				dev_put(dev);
			dev = dev_get_by_index(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;
	}
	if (dev)
		dev_put(dev);
}

#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->u.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->u.dst);
	ip6_del_rt(aca->aca_rt);

	aca_put(aca);
	return 0;
}

static int ipv6_dev_ac_dec(struct net_device *dev, struct in6_addr *addr)
{
	int ret;
	struct inet6_dev *idev = in6_dev_get(dev);
	if (idev == NULL)
		return -ENODEV;
	ret = __ipv6_dev_ac_dec(idev, addr);
	in6_dev_put(idev);
	return ret;
}

/*
 *	check if the interface has this anycast address
 */
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);
		in6_dev_put(idev);
		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;

	if (dev)
		return ipv6_chk_acast_dev(dev, addr);
	rcu_read_lock();
	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
	80	if (ifindex == 0) {
	81	struct rt6_info *rt;
	82
6ab57e7e	83	rt = rt6_lookup(net, addr, NULL, 0, 0);
1da177e4 LT	84	if (rt) {
	85	dev = rt->rt6i_dev;
	86	dev_hold(dev);
	87	dst_release(&rt->u.dst);
	88	} else if (ishost) {
	89	err = -EADDRNOTAVAIL;
	90	goto out_free_pac;
	91	} else {
	92	/* router, no matching interface: just pick one */
	93
6ab57e7e	94	dev = dev_get_by_flags(net, IFF_UP, IFF_UP\|IFF_LOOPBACK);
1da177e4 LT	95	}
1da177e4 LT	96	} else
6ab57e7e	97	dev = dev_get_by_index(net, ifindex);
1da177e4 LT	98
	99	if (dev == NULL) {
	100	err = -ENODEV;
	101	goto out_free_pac;
	102	}
	103
	104	idev = in6_dev_get(dev);
	105	if (!idev) {
	106	if (ifindex)
	107	err = -ENODEV;
	108	else
	109	err = -EADDRNOTAVAIL;
	110	goto out_dev_put;
	111	}
	112	/* reset ishost, now that we have a specific device */
	113	ishost = !idev->cnf.forwarding;
	114	in6_dev_put(idev);
	115
	116	pac->acl_ifindex = dev->ifindex;
	117
	118	/* XXX
	119	* For hosts, allow link-local or matching prefix anycasts.
	120	* This obviates the need for propagating anycast routes while
	121	* still allowing some non-router anycast participation.
	122	*/
52eeeb84	123	if (!ipv6_chk_prefix(addr, dev)) {
1da177e4 LT	124	if (ishost)
	125	err = -EADDRNOTAVAIL;
	126	if (err)
	127	goto out_dev_put;
	128	}
	129
	130	err = ipv6_dev_ac_inc(dev, addr);
	131	if (err)
	132	goto out_dev_put;
	133
	134	write_lock_bh(&ipv6_sk_ac_lock);
	135	pac->acl_next = np->ipv6_ac_list;
	136	np->ipv6_ac_list = pac;
	137	write_unlock_bh(&ipv6_sk_ac_lock);
	138
	139	dev_put(dev);
	140
	141	return 0;
	142
	143	out_dev_put:
	144	dev_put(dev);
	145	out_free_pac:
	146	sock_kfree_s(sk, pac, sizeof(*pac));
	147	return err;
	148	}
	149
	150	/*
	151	* socket leave an anycast group
	152	*/
	153	int ipv6_sock_ac_drop(struct sock sk, int ifindex, struct in6_addr addr)
	154	{
	155	struct ipv6_pinfo *np = inet6_sk(sk);
	156	struct net_device *dev;
	157	struct ipv6_ac_socklist pac, prev_pac;
6ab57e7e	158	struct net *net = sock_net(sk);
1da177e4 LT	159
	160	write_lock_bh(&ipv6_sk_ac_lock);
	161	prev_pac = NULL;
	162	for (pac = np->ipv6_ac_list; pac; pac = pac->acl_next) {
	163	if ((ifindex == 0 \|\| pac->acl_ifindex == ifindex) &&
	164	ipv6_addr_equal(&pac->acl_addr, addr))
	165	break;
	166	prev_pac = pac;
	167	}
	168	if (!pac) {
	169	write_unlock_bh(&ipv6_sk_ac_lock);
	170	return -ENOENT;
	171	}
	172	if (prev_pac)
	173	prev_pac->acl_next = pac->acl_next;
	174	else
	175	np->ipv6_ac_list = pac->acl_next;
	176
	177	write_unlock_bh(&ipv6_sk_ac_lock);
	178
6ab57e7e	179	dev = dev_get_by_index(net, pac->acl_ifindex);
1da177e4 LT	180	if (dev) {
	181	ipv6_dev_ac_dec(dev, &pac->acl_addr);
	182	dev_put(dev);
	183	}
	184	sock_kfree_s(sk, pac, sizeof(*pac));
	185	return 0;
	186	}
	187
	188	void ipv6_sock_ac_close(struct sock *sk)
	189	{
	190	struct ipv6_pinfo *np = inet6_sk(sk);
	191	struct net_device *dev = NULL;
	192	struct ipv6_ac_socklist *pac;
6ab57e7e	193	struct net *net = sock_net(sk);
1da177e4 LT	194	int prev_index;
	195
	196	write_lock_bh(&ipv6_sk_ac_lock);
	197	pac = np->ipv6_ac_list;
	198	np->ipv6_ac_list = NULL;
	199	write_unlock_bh(&ipv6_sk_ac_lock);
	200
	201	prev_index = 0;
	202	while (pac) {
	203	struct ipv6_ac_socklist *next = pac->acl_next;
	204
	205	if (pac->acl_ifindex != prev_index) {
	206	if (dev)
	207	dev_put(dev);
6ab57e7e	208	dev = dev_get_by_index(net, pac->acl_ifindex);
1da177e4 LT	209	prev_index = pac->acl_ifindex;
	210	}
	211	if (dev)
	212	ipv6_dev_ac_dec(dev, &pac->acl_addr);
	213	sock_kfree_s(sk, pac, sizeof(*pac));
	214	pac = next;
	215	}
	216	if (dev)
	217	dev_put(dev);
	218	}
	219
	220	#if 0
	221	/* The function is not used, which is funny. Apparently, author
	222	* supposed to use it to filter out datagrams inside udp/raw but forgot.
	223	*
	224	* It is OK, anycasts are not special comparing to delivery to unicasts.
	225	*/
	226
	227	int inet6_ac_check(struct sock sk, struct in6_addr addr, int ifindex)
	228	{
	229	struct ipv6_ac_socklist *pac;
	230	struct ipv6_pinfo *np = inet6_sk(sk);
	231	int found;
	232
	233	found = 0;
	234	read_lock(&ipv6_sk_ac_lock);
	235	for (pac=np->ipv6_ac_list; pac; pac=pac->acl_next) {
	236	if (ifindex && pac->acl_ifindex != ifindex)
	237	continue;
	238	found = ipv6_addr_equal(&pac->acl_addr, addr);
	239	if (found)
	240	break;
	241	}
	242	read_unlock(&ipv6_sk_ac_lock);
	243
	244	return found;
	245	}
	246
	247	#endif
	248
	249	static void aca_put(struct ifacaddr6 *ac)
	250	{
	251	if (atomic_dec_and_test(&ac->aca_refcnt)) {
	252	in6_dev_put(ac->aca_idev);
	253	dst_release(&ac->aca_rt->u.dst);
	254	kfree(ac);
	255	}
	256	}
	257
	258	/*
	259	* device anycast group inc (add if not found)
	260	*/
	261	int ipv6_dev_ac_inc(struct net_device dev, struct in6_addr addr)
	262	{
	263	struct ifacaddr6 *aca;
	264	struct inet6_dev *idev;
	265	struct rt6_info *rt;
	266	int err;
	267
	268	idev = in6_dev_get(dev);
	269
	270	if (idev == NULL)
	271	return -EINVAL;
	272
273	write_lock_bh(&idev->lock);
274	if (idev->dead) {
275	err = -ENODEV;
276	goto out;
277	}
278
279	for (aca = idev->ac_list; aca; aca = aca->aca_next) {
280	if (ipv6_addr_equal(&aca->aca_addr, addr)) {
281	aca->aca_users++;
282	err = 0;
283	goto out;
284	}
285	}
286
287	/*
288	* not found: create a new one.
289	*/
290
0c600eda	291	aca = kzalloc(sizeof(struct ifacaddr6), GFP_ATOMIC);
1da177e4 LT	292
	293	if (aca == NULL) {
	294	err = -ENOMEM;
	295	goto out;
	296	}
	297
	298	rt = addrconf_dst_alloc(idev, addr, 1);
	299	if (IS_ERR(rt)) {
	300	kfree(aca);
	301	err = PTR_ERR(rt);
	302	goto out;
	303	}
	304
1da177e4 LT	305	ipv6_addr_copy(&aca->aca_addr, addr);
	306	aca->aca_idev = idev;
	307	aca->aca_rt = rt;
	308	aca->aca_users = 1;
	309	/* aca_tstamp should be updated upon changes */
	310	aca->aca_cstamp = aca->aca_tstamp = jiffies;
	311	atomic_set(&aca->aca_refcnt, 2);
	312	spin_lock_init(&aca->aca_lock);
	313
	314	aca->aca_next = idev->ac_list;
	315	idev->ac_list = aca;
	316	write_unlock_bh(&idev->lock);
	317
84f59370	318	ip6_ins_rt(rt);
1da177e4 LT	319
	320	addrconf_join_solict(dev, &aca->aca_addr);
	321
	322	aca_put(aca);
	323	return 0;
	324	out:
	325	write_unlock_bh(&idev->lock);
	326	in6_dev_put(idev);
	327	return err;
	328	}
	329
	330	/*
	331	* device anycast group decrement
	332	*/
	333	int __ipv6_dev_ac_dec(struct inet6_dev idev, struct in6_addr addr)
	334	{
	335	struct ifacaddr6 aca, prev_aca;
	336
	337	write_lock_bh(&idev->lock);
	338	prev_aca = NULL;
	339	for (aca = idev->ac_list; aca; aca = aca->aca_next) {
	340	if (ipv6_addr_equal(&aca->aca_addr, addr))
	341	break;
	342	prev_aca = aca;
	343	}
	344	if (!aca) {
	345	write_unlock_bh(&idev->lock);
	346	return -ENOENT;
	347	}
	348	if (--aca->aca_users > 0) {
	349	write_unlock_bh(&idev->lock);
	350	return 0;
	351	}
	352	if (prev_aca)
	353	prev_aca->aca_next = aca->aca_next;
	354	else
	355	idev->ac_list = aca->aca_next;
	356	write_unlock_bh(&idev->lock);
	357	addrconf_leave_solict(idev, &aca->aca_addr);
	358
	359	dst_hold(&aca->aca_rt->u.dst);
84f59370	360	ip6_del_rt(aca->aca_rt);
1da177e4 LT	361
	362	aca_put(aca);
	363	return 0;
	364	}
	365
	366	static int ipv6_dev_ac_dec(struct net_device dev, struct in6_addr addr)
	367	{
	368	int ret;
	369	struct inet6_dev *idev = in6_dev_get(dev);
	370	if (idev == NULL)
	371	return -ENODEV;
	372	ret = __ipv6_dev_ac_dec(idev, addr);
	373	in6_dev_put(idev);
	374	return ret;
	375	}
1ab1457c	376
1da177e4 LT	377	/*
	378	* check if the interface has this anycast address
	379	*/
	380	static int ipv6_chk_acast_dev(struct net_device dev, struct in6_addr addr)
	381	{
	382	struct inet6_dev *idev;
	383	struct ifacaddr6 *aca;
	384
	385	idev = in6_dev_get(dev);
	386	if (idev) {
	387	read_lock_bh(&idev->lock);
	388	for (aca = idev->ac_list; aca; aca = aca->aca_next)
	389	if (ipv6_addr_equal(&aca->aca_addr, addr))
	390	break;
	391	read_unlock_bh(&idev->lock);
	392	in6_dev_put(idev);
cfcabdcc	393	return aca != NULL;
1da177e4 LT	394	}
	395	return 0;
	396	}
	397
	398	/*
	399	* check if given interface (or any, if dev==0) has this anycast address
	400	*/
6ab57e7e DL	401	int ipv6_chk_acast_addr(struct net net, struct net_device dev,
6ab57e7e DL	402	struct in6_addr *addr)
1da177e4	403	{
7562f876 PE	404	int found = 0;
7562f876 PE	405
1da177e4 LT	406	if (dev)
1da177e4 LT	407	return ipv6_chk_acast_dev(dev, addr);
c6d14c84 ED	408	rcu_read_lock();
c6d14c84 ED	409	for_each_netdev_rcu(net, dev)
7562f876 PE	410	if (ipv6_chk_acast_dev(dev, addr)) {
7562f876 PE	411	found = 1;
1da177e4	412	break;
7562f876	413	}
c6d14c84	414	rcu_read_unlock();
7562f876	415	return found;
1da177e4 LT	416	}
	417
	418
	419	#ifdef CONFIG_PROC_FS
	420	struct ac6_iter_state {
6ab57e7e	421	struct seq_net_private p;
1da177e4 LT	422	struct net_device *dev;
	423	struct inet6_dev *idev;
	424	};
	425
	426	#define ac6_seq_private(seq) ((struct ac6_iter_state *)(seq)->private)
	427
	428	static inline struct ifacaddr6 ac6_get_first(struct seq_file seq)
	429	{
	430	struct ifacaddr6 *im = NULL;
	431	struct ac6_iter_state *state = ac6_seq_private(seq);
6ab57e7e	432	struct net *net = seq_file_net(seq);
1da177e4	433
7562f876	434	state->idev = NULL;
ce81b76a	435	for_each_netdev_rcu(net, state->dev) {
1da177e4	436	struct inet6_dev *idev;
ce81b76a	437	idev = __in6_dev_get(state->dev);
1da177e4 LT	438	if (!idev)
	439	continue;
	440	read_lock_bh(&idev->lock);
	441	im = idev->ac_list;
	442	if (im) {
	443	state->idev = idev;
	444	break;
	445	}
	446	read_unlock_bh(&idev->lock);
	447	}
	448	return im;
	449	}
	450
	451	static struct ifacaddr6 ac6_get_next(struct seq_file seq, struct ifacaddr6 *im)
	452	{
	453	struct ac6_iter_state *state = ac6_seq_private(seq);
	454
	455	im = im->aca_next;
	456	while (!im) {
ce81b76a	457	if (likely(state->idev != NULL))
1da177e4	458	read_unlock_bh(&state->idev->lock);
ce81b76a ED	459
ce81b76a ED	460	state->dev = next_net_device_rcu(state->dev);
1da177e4 LT	461	if (!state->dev) {
	462	state->idev = NULL;
	463	break;
	464	}
ce81b76a	465	state->idev = __in6_dev_get(state->dev);
1da177e4 LT	466	if (!state->idev)
	467	continue;
	468	read_lock_bh(&state->idev->lock);
	469	im = state->idev->ac_list;
	470	}
	471	return im;
	472	}
	473
	474	static struct ifacaddr6 ac6_get_idx(struct seq_file seq, loff_t pos)
	475	{
	476	struct ifacaddr6 *im = ac6_get_first(seq);
	477	if (im)
	478	while (pos && (im = ac6_get_next(seq, im)) != NULL)
	479	--pos;
	480	return pos ? NULL : im;
	481	}
	482
	483	static void ac6_seq_start(struct seq_file seq, loff_t *pos)
ce81b76a	484	__acquires(RCU)
1da177e4	485	{
ce81b76a	486	rcu_read_lock();
1da177e4 LT	487	return ac6_get_idx(seq, *pos);
	488	}
	489
	490	static void ac6_seq_next(struct seq_file seq, void v, loff_t pos)
	491	{
ce81b76a ED	492	struct ifacaddr6 *im = ac6_get_next(seq, v);
ce81b76a ED	493
1da177e4 LT	494	++*pos;
	495	return im;
	496	}
	497
	498	static void ac6_seq_stop(struct seq_file seq, void v)
ce81b76a	499	__releases(RCU)
1da177e4 LT	500	{
1da177e4 LT	501	struct ac6_iter_state *state = ac6_seq_private(seq);
ce81b76a	502
1da177e4 LT	503	if (likely(state->idev != NULL)) {
1da177e4 LT	504	read_unlock_bh(&state->idev->lock);
ce81b76a	505	state->idev = NULL;
1da177e4	506	}
ce81b76a	507	rcu_read_unlock();
1da177e4 LT	508	}
	509
	510	static int ac6_seq_show(struct seq_file seq, void v)
	511	{
	512	struct ifacaddr6 im = (struct ifacaddr6 )v;
	513	struct ac6_iter_state *state = ac6_seq_private(seq);
	514
4b7a4274	515	seq_printf(seq, "%-4d %-15s %pi6 %5d\n",
1da177e4	516	state->dev->ifindex, state->dev->name,
b071195d	517	&im->aca_addr, im->aca_users);
1da177e4 LT	518	return 0;
	519	}
	520
56b3d975	521	static const struct seq_operations ac6_seq_ops = {
1da177e4 LT	522	.start = ac6_seq_start,
	523	.next = ac6_seq_next,
	524	.stop = ac6_seq_stop,
	525	.show = ac6_seq_show,
	526	};
	527
	528	static int ac6_seq_open(struct inode inode, struct file file)
	529	{
6ab57e7e DL	530	return seq_open_net(inode, file, &ac6_seq_ops,
6ab57e7e DL	531	sizeof(struct ac6_iter_state));
1da177e4 LT	532	}
1da177e4 LT	533
9a32144e	534	static const struct file_operations ac6_seq_fops = {
1da177e4 LT	535	.owner = THIS_MODULE,
	536	.open = ac6_seq_open,
	537	.read = seq_read,
	538	.llseek = seq_lseek,
6ab57e7e	539	.release = seq_release_net,
1da177e4 LT	540	};
1da177e4 LT	541
2c8c1e72	542	int __net_init ac6_proc_init(struct net *net)
1da177e4	543	{
6ab57e7e	544	if (!proc_net_fops_create(net, "anycast6", S_IRUGO, &ac6_seq_fops))
1da177e4 LT	545	return -ENOMEM;
	546
	547	return 0;
	548	}
	549
6ab57e7e	550	void ac6_proc_exit(struct net *net)
1da177e4	551	{
6ab57e7e	552	proc_net_remove(net, "anycast6");
1da177e4 LT	553	}
	554	#endif
	555