[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 <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);

static int
ip6_onlink(struct in6_addr *addr, struct net_device *dev)
{
	struct inet6_dev	*idev;
	struct inet6_ifaddr	*ifa;
	int	onlink;

	onlink = 0;
	rcu_read_lock();
	idev = __in6_dev_get(dev);
	if (idev) {
		read_lock_bh(&idev->lock);
		for (ifa=idev->addr_list; ifa; ifa=ifa->if_next) {
			onlink = ipv6_prefix_equal(addr, &ifa->addr,
						   ifa->prefix_len);
			if (onlink)
				break;
		}
		read_unlock_bh(&idev->lock);
	}
	rcu_read_unlock();
	return onlink;
}

/*
 *	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;
	int	ishost = !ipv6_devconf.forwarding;
	int	err = 0;

	if (!capable(CAP_NET_ADMIN))
		return -EPERM;
	if (ipv6_addr_is_multicast(addr))
		return -EINVAL;
	if (ipv6_chk_addr(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(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(&init_net, IFF_UP, IFF_UP|IFF_LOOPBACK);
		}
	} else
		dev = dev_get_by_index(&init_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 (!ip6_onlink(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;

	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(&init_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;
	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(&init_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);

	dst_hold(&rt->u.dst);
	if (ip6_ins_rt(rt))
		dst_release(&rt->u.dst);

	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);
	if (ip6_del_rt(aca->aca_rt))
		dst_free(&aca->aca_rt->u.dst);
	else
		dst_release(&aca->aca_rt->u.dst);

	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 != 0;
	}
	return 0;
}

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

	if (dev)
		return ipv6_chk_acast_dev(dev, addr);
	read_lock(&dev_base_lock);
	for_each_netdev(&init_net, dev)
		if (ipv6_chk_acast_dev(dev, addr)) {
			found = 1;
			break;
		}
	read_unlock(&dev_base_lock);
	return found;
}


#ifdef CONFIG_PROC_FS
struct ac6_iter_state {
	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);

	state->idev = NULL;
	for_each_netdev(&init_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);
		in6_dev_put(idev);
	}
	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);
			in6_dev_put(state->idev);
		}
		state->dev = next_net_device(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)
{
	read_lock(&dev_base_lock);
	return ac6_get_idx(seq, *pos);
}

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

static void ac6_seq_stop(struct seq_file *seq, void *v)
{
	struct ac6_iter_state *state = ac6_seq_private(seq);
	if (likely(state->idev != NULL)) {
		read_unlock_bh(&state->idev->lock);
		in6_dev_put(state->idev);
	}
	read_unlock(&dev_base_lock);
}

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 " NIP6_SEQFMT " %5d\n",
		   state->dev->ifindex, state->dev->name,
		   NIP6(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)
{
	struct seq_file *seq;
	int rc = -ENOMEM;
	struct ac6_iter_state *s = kzalloc(sizeof(*s), GFP_KERNEL);

	if (!s)
		goto out;

	rc = seq_open(file, &ac6_seq_ops);
	if (rc)
		goto out_kfree;

	seq = file->private_data;
	seq->private = s;
out:
	return rc;
out_kfree:
	kfree(s);
	goto out;
}

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

int __init ac6_proc_init(void)
{
	if (!proc_net_fops_create(&init_net, "anycast6", S_IRUGO, &ac6_seq_fops))
		return -ENOMEM;

	return 0;
}

void ac6_proc_exit(void)
{
	proc_net_remove(&init_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>
	32
457c4cbc	33	#include <net/net_namespace.h>
1da177e4 LT	34	#include <net/sock.h>
	35	#include <net/snmp.h>
	36
	37	#include <net/ipv6.h>
	38	#include <net/protocol.h>
	39	#include <net/if_inet6.h>
	40	#include <net/ndisc.h>
	41	#include <net/addrconf.h>
	42	#include <net/ip6_route.h>
	43
	44	#include <net/checksum.h>
	45
	46	static int ipv6_dev_ac_dec(struct net_device dev, struct in6_addr addr);
	47
	48	/* Big ac list lock for all the sockets */
	49	static DEFINE_RWLOCK(ipv6_sk_ac_lock);
	50
	51	static int
	52	ip6_onlink(struct in6_addr addr, struct net_device dev)
	53	{
	54	struct inet6_dev *idev;
	55	struct inet6_ifaddr *ifa;
	56	int onlink;
	57
	58	onlink = 0;
8814c4b5	59	rcu_read_lock();
1da177e4 LT	60	idev = __in6_dev_get(dev);
	61	if (idev) {
	62	read_lock_bh(&idev->lock);
	63	for (ifa=idev->addr_list; ifa; ifa=ifa->if_next) {
	64	onlink = ipv6_prefix_equal(addr, &ifa->addr,
	65	ifa->prefix_len);
	66	if (onlink)
	67	break;
	68	}
	69	read_unlock_bh(&idev->lock);
	70	}
8814c4b5	71	rcu_read_unlock();
1da177e4 LT	72	return onlink;
	73	}
	74
	75	/*
	76	* socket join an anycast group
	77	*/
	78
	79	int ipv6_sock_ac_join(struct sock sk, int ifindex, struct in6_addr addr)
	80	{
	81	struct ipv6_pinfo *np = inet6_sk(sk);
	82	struct net_device *dev = NULL;
	83	struct inet6_dev *idev;
	84	struct ipv6_ac_socklist *pac;
	85	int ishost = !ipv6_devconf.forwarding;
	86	int err = 0;
	87
	88	if (!capable(CAP_NET_ADMIN))
	89	return -EPERM;
	90	if (ipv6_addr_is_multicast(addr))
	91	return -EINVAL;
	92	if (ipv6_chk_addr(addr, NULL, 0))
	93	return -EINVAL;
	94
	95	pac = sock_kmalloc(sk, sizeof(struct ipv6_ac_socklist), GFP_KERNEL);
	96	if (pac == NULL)
	97	return -ENOMEM;
	98	pac->acl_next = NULL;
	99	ipv6_addr_copy(&pac->acl_addr, addr);
	100
	101	if (ifindex == 0) {
	102	struct rt6_info *rt;
	103
	104	rt = rt6_lookup(addr, NULL, 0, 0);
	105	if (rt) {
	106	dev = rt->rt6i_dev;
	107	dev_hold(dev);
	108	dst_release(&rt->u.dst);
	109	} else if (ishost) {
	110	err = -EADDRNOTAVAIL;
	111	goto out_free_pac;
	112	} else {
	113	/* router, no matching interface: just pick one */
	114
881d966b	115	dev = dev_get_by_flags(&init_net, IFF_UP, IFF_UP\|IFF_LOOPBACK);
1da177e4 LT	116	}
1da177e4 LT	117	} else
881d966b	118	dev = dev_get_by_index(&init_net, ifindex);
1da177e4 LT	119
	120	if (dev == NULL) {
	121	err = -ENODEV;
	122	goto out_free_pac;
	123	}
	124
	125	idev = in6_dev_get(dev);
	126	if (!idev) {
	127	if (ifindex)
	128	err = -ENODEV;
	129	else
	130	err = -EADDRNOTAVAIL;
	131	goto out_dev_put;
	132	}
	133	/* reset ishost, now that we have a specific device */
	134	ishost = !idev->cnf.forwarding;
	135	in6_dev_put(idev);
	136
	137	pac->acl_ifindex = dev->ifindex;
	138
	139	/* XXX
	140	* For hosts, allow link-local or matching prefix anycasts.
	141	* This obviates the need for propagating anycast routes while
	142	* still allowing some non-router anycast participation.
	143	*/
	144	if (!ip6_onlink(addr, dev)) {
	145	if (ishost)
	146	err = -EADDRNOTAVAIL;
	147	if (err)
	148	goto out_dev_put;
	149	}
	150
	151	err = ipv6_dev_ac_inc(dev, addr);
	152	if (err)
	153	goto out_dev_put;
	154
	155	write_lock_bh(&ipv6_sk_ac_lock);
	156	pac->acl_next = np->ipv6_ac_list;
	157	np->ipv6_ac_list = pac;
	158	write_unlock_bh(&ipv6_sk_ac_lock);
	159
	160	dev_put(dev);
	161
	162	return 0;
	163
	164	out_dev_put:
	165	dev_put(dev);
	166	out_free_pac:
	167	sock_kfree_s(sk, pac, sizeof(*pac));
	168	return err;
	169	}
	170
	171	/*
	172	* socket leave an anycast group
	173	*/
	174	int ipv6_sock_ac_drop(struct sock sk, int ifindex, struct in6_addr addr)
	175	{
	176	struct ipv6_pinfo *np = inet6_sk(sk);
	177	struct net_device *dev;
	178	struct ipv6_ac_socklist pac, prev_pac;
	179
	180	write_lock_bh(&ipv6_sk_ac_lock);
	181	prev_pac = NULL;
	182	for (pac = np->ipv6_ac_list; pac; pac = pac->acl_next) {
183	if ((ifindex == 0 \|\| pac->acl_ifindex == ifindex) &&
184	ipv6_addr_equal(&pac->acl_addr, addr))
185	break;
186	prev_pac = pac;
187	}
188	if (!pac) {
189	write_unlock_bh(&ipv6_sk_ac_lock);
190	return -ENOENT;
191	}
192	if (prev_pac)
193	prev_pac->acl_next = pac->acl_next;
194	else
195	np->ipv6_ac_list = pac->acl_next;
196
197	write_unlock_bh(&ipv6_sk_ac_lock);
198
881d966b	199	dev = dev_get_by_index(&init_net, pac->acl_ifindex);
1da177e4 LT	200	if (dev) {
	201	ipv6_dev_ac_dec(dev, &pac->acl_addr);
	202	dev_put(dev);
	203	}
	204	sock_kfree_s(sk, pac, sizeof(*pac));
	205	return 0;
	206	}
	207
	208	void ipv6_sock_ac_close(struct sock *sk)
	209	{
	210	struct ipv6_pinfo *np = inet6_sk(sk);
	211	struct net_device *dev = NULL;
	212	struct ipv6_ac_socklist *pac;
	213	int prev_index;
	214
	215	write_lock_bh(&ipv6_sk_ac_lock);
	216	pac = np->ipv6_ac_list;
	217	np->ipv6_ac_list = NULL;
	218	write_unlock_bh(&ipv6_sk_ac_lock);
	219
	220	prev_index = 0;
	221	while (pac) {
	222	struct ipv6_ac_socklist *next = pac->acl_next;
	223
	224	if (pac->acl_ifindex != prev_index) {
	225	if (dev)
	226	dev_put(dev);
881d966b	227	dev = dev_get_by_index(&init_net, pac->acl_ifindex);
1da177e4 LT	228	prev_index = pac->acl_ifindex;
	229	}
	230	if (dev)
	231	ipv6_dev_ac_dec(dev, &pac->acl_addr);
	232	sock_kfree_s(sk, pac, sizeof(*pac));
	233	pac = next;
	234	}
	235	if (dev)
	236	dev_put(dev);
	237	}
	238
	239	#if 0
	240	/* The function is not used, which is funny. Apparently, author
	241	* supposed to use it to filter out datagrams inside udp/raw but forgot.
	242	*
	243	* It is OK, anycasts are not special comparing to delivery to unicasts.
	244	*/
	245
	246	int inet6_ac_check(struct sock sk, struct in6_addr addr, int ifindex)
	247	{
	248	struct ipv6_ac_socklist *pac;
	249	struct ipv6_pinfo *np = inet6_sk(sk);
	250	int found;
	251
	252	found = 0;
	253	read_lock(&ipv6_sk_ac_lock);
	254	for (pac=np->ipv6_ac_list; pac; pac=pac->acl_next) {
	255	if (ifindex && pac->acl_ifindex != ifindex)
	256	continue;
	257	found = ipv6_addr_equal(&pac->acl_addr, addr);
	258	if (found)
	259	break;
	260	}
	261	read_unlock(&ipv6_sk_ac_lock);
	262
	263	return found;
	264	}
	265
	266	#endif
	267
	268	static void aca_put(struct ifacaddr6 *ac)
	269	{
	270	if (atomic_dec_and_test(&ac->aca_refcnt)) {
	271	in6_dev_put(ac->aca_idev);
	272	dst_release(&ac->aca_rt->u.dst);
	273	kfree(ac);
	274	}
	275	}
	276
	277	/*
	278	* device anycast group inc (add if not found)
	279	*/
	280	int ipv6_dev_ac_inc(struct net_device dev, struct in6_addr addr)
	281	{
	282	struct ifacaddr6 *aca;
	283	struct inet6_dev *idev;
	284	struct rt6_info *rt;
	285	int err;
	286
	287	idev = in6_dev_get(dev);
	288
	289	if (idev == NULL)
	290	return -EINVAL;
	291
292	write_lock_bh(&idev->lock);
293	if (idev->dead) {
294	err = -ENODEV;
295	goto out;
296	}
297
298	for (aca = idev->ac_list; aca; aca = aca->aca_next) {
299	if (ipv6_addr_equal(&aca->aca_addr, addr)) {
300	aca->aca_users++;
301	err = 0;
302	goto out;
303	}
304	}
305
306	/*
307	* not found: create a new one.
308	*/
309
0c600eda	310	aca = kzalloc(sizeof(struct ifacaddr6), GFP_ATOMIC);
1da177e4 LT	311
	312	if (aca == NULL) {
	313	err = -ENOMEM;
	314	goto out;
	315	}
	316
	317	rt = addrconf_dst_alloc(idev, addr, 1);
	318	if (IS_ERR(rt)) {
	319	kfree(aca);
	320	err = PTR_ERR(rt);
	321	goto out;
	322	}
	323
1da177e4 LT	324	ipv6_addr_copy(&aca->aca_addr, addr);
	325	aca->aca_idev = idev;
	326	aca->aca_rt = rt;
	327	aca->aca_users = 1;
	328	/* aca_tstamp should be updated upon changes */
	329	aca->aca_cstamp = aca->aca_tstamp = jiffies;
	330	atomic_set(&aca->aca_refcnt, 2);
	331	spin_lock_init(&aca->aca_lock);
	332
	333	aca->aca_next = idev->ac_list;
	334	idev->ac_list = aca;
	335	write_unlock_bh(&idev->lock);
	336
	337	dst_hold(&rt->u.dst);
40e22e8f	338	if (ip6_ins_rt(rt))
1da177e4 LT	339	dst_release(&rt->u.dst);
	340
	341	addrconf_join_solict(dev, &aca->aca_addr);
	342
	343	aca_put(aca);
	344	return 0;
	345	out:
	346	write_unlock_bh(&idev->lock);
	347	in6_dev_put(idev);
	348	return err;
	349	}
	350
	351	/*
	352	* device anycast group decrement
	353	*/
	354	int __ipv6_dev_ac_dec(struct inet6_dev idev, struct in6_addr addr)
	355	{
	356	struct ifacaddr6 aca, prev_aca;
	357
	358	write_lock_bh(&idev->lock);
	359	prev_aca = NULL;
	360	for (aca = idev->ac_list; aca; aca = aca->aca_next) {
	361	if (ipv6_addr_equal(&aca->aca_addr, addr))
	362	break;
	363	prev_aca = aca;
	364	}
	365	if (!aca) {
	366	write_unlock_bh(&idev->lock);
	367	return -ENOENT;
	368	}
	369	if (--aca->aca_users > 0) {
	370	write_unlock_bh(&idev->lock);
	371	return 0;
	372	}
	373	if (prev_aca)
	374	prev_aca->aca_next = aca->aca_next;
	375	else
	376	idev->ac_list = aca->aca_next;
	377	write_unlock_bh(&idev->lock);
	378	addrconf_leave_solict(idev, &aca->aca_addr);
	379
	380	dst_hold(&aca->aca_rt->u.dst);
72d3b2c9	381	if (ip6_del_rt(aca->aca_rt))
1da177e4 LT	382	dst_free(&aca->aca_rt->u.dst);
	383	else
	384	dst_release(&aca->aca_rt->u.dst);
	385
	386	aca_put(aca);
	387	return 0;
	388	}
	389
	390	static int ipv6_dev_ac_dec(struct net_device dev, struct in6_addr addr)
	391	{
	392	int ret;
	393	struct inet6_dev *idev = in6_dev_get(dev);
	394	if (idev == NULL)
	395	return -ENODEV;
	396	ret = __ipv6_dev_ac_dec(idev, addr);
	397	in6_dev_put(idev);
	398	return ret;
	399	}
1ab1457c	400
1da177e4 LT	401	/*
	402	* check if the interface has this anycast address
	403	*/
	404	static int ipv6_chk_acast_dev(struct net_device dev, struct in6_addr addr)
	405	{
	406	struct inet6_dev *idev;
	407	struct ifacaddr6 *aca;
	408
	409	idev = in6_dev_get(dev);
	410	if (idev) {
	411	read_lock_bh(&idev->lock);
	412	for (aca = idev->ac_list; aca; aca = aca->aca_next)
	413	if (ipv6_addr_equal(&aca->aca_addr, addr))
	414	break;
	415	read_unlock_bh(&idev->lock);
	416	in6_dev_put(idev);
	417	return aca != 0;
	418	}
	419	return 0;
	420	}
	421
	422	/*
	423	* check if given interface (or any, if dev==0) has this anycast address
	424	*/
	425	int ipv6_chk_acast_addr(struct net_device dev, struct in6_addr addr)
	426	{
7562f876 PE	427	int found = 0;
7562f876 PE	428
1da177e4 LT	429	if (dev)
	430	return ipv6_chk_acast_dev(dev, addr);
	431	read_lock(&dev_base_lock);
881d966b	432	for_each_netdev(&init_net, dev)
7562f876 PE	433	if (ipv6_chk_acast_dev(dev, addr)) {
7562f876 PE	434	found = 1;
1da177e4	435	break;
7562f876	436	}
1da177e4	437	read_unlock(&dev_base_lock);
7562f876	438	return found;
1da177e4 LT	439	}
	440
	441
	442	#ifdef CONFIG_PROC_FS
	443	struct ac6_iter_state {
	444	struct net_device *dev;
	445	struct inet6_dev *idev;
	446	};
	447
	448	#define ac6_seq_private(seq) ((struct ac6_iter_state *)(seq)->private)
	449
	450	static inline struct ifacaddr6 ac6_get_first(struct seq_file seq)
	451	{
	452	struct ifacaddr6 *im = NULL;
	453	struct ac6_iter_state *state = ac6_seq_private(seq);
	454
7562f876	455	state->idev = NULL;
881d966b	456	for_each_netdev(&init_net, state->dev) {
1da177e4 LT	457	struct inet6_dev *idev;
	458	idev = in6_dev_get(state->dev);
	459	if (!idev)
	460	continue;
	461	read_lock_bh(&idev->lock);
	462	im = idev->ac_list;
	463	if (im) {
	464	state->idev = idev;
	465	break;
	466	}
	467	read_unlock_bh(&idev->lock);
aa6e4a96	468	in6_dev_put(idev);
1da177e4 LT	469	}
	470	return im;
	471	}
	472
	473	static struct ifacaddr6 ac6_get_next(struct seq_file seq, struct ifacaddr6 *im)
	474	{
	475	struct ac6_iter_state *state = ac6_seq_private(seq);
	476
	477	im = im->aca_next;
	478	while (!im) {
	479	if (likely(state->idev != NULL)) {
	480	read_unlock_bh(&state->idev->lock);
	481	in6_dev_put(state->idev);
	482	}
7562f876	483	state->dev = next_net_device(state->dev);
1da177e4 LT	484	if (!state->dev) {
	485	state->idev = NULL;
	486	break;
	487	}
	488	state->idev = in6_dev_get(state->dev);
	489	if (!state->idev)
	490	continue;
	491	read_lock_bh(&state->idev->lock);
	492	im = state->idev->ac_list;
	493	}
	494	return im;
	495	}
	496
	497	static struct ifacaddr6 ac6_get_idx(struct seq_file seq, loff_t pos)
	498	{
	499	struct ifacaddr6 *im = ac6_get_first(seq);
	500	if (im)
	501	while (pos && (im = ac6_get_next(seq, im)) != NULL)
	502	--pos;
	503	return pos ? NULL : im;
	504	}
	505
	506	static void ac6_seq_start(struct seq_file seq, loff_t *pos)
	507	{
	508	read_lock(&dev_base_lock);
	509	return ac6_get_idx(seq, *pos);
	510	}
	511
	512	static void ac6_seq_next(struct seq_file seq, void v, loff_t pos)
	513	{
	514	struct ifacaddr6 *im;
	515	im = ac6_get_next(seq, v);
	516	++*pos;
	517	return im;
	518	}
	519
	520	static void ac6_seq_stop(struct seq_file seq, void v)
	521	{
	522	struct ac6_iter_state *state = ac6_seq_private(seq);
	523	if (likely(state->idev != NULL)) {
	524	read_unlock_bh(&state->idev->lock);
	525	in6_dev_put(state->idev);
	526	}
	527	read_unlock(&dev_base_lock);
	528	}
	529
	530	static int ac6_seq_show(struct seq_file seq, void v)
	531	{
	532	struct ifacaddr6 im = (struct ifacaddr6 )v;
	533	struct ac6_iter_state *state = ac6_seq_private(seq);
	534
	535	seq_printf(seq,
9343e79a	536	"%-4d %-15s " NIP6_SEQFMT " %5d\n",
1da177e4 LT	537	state->dev->ifindex, state->dev->name,
	538	NIP6(im->aca_addr),
	539	im->aca_users);
	540	return 0;
	541	}
	542
56b3d975	543	static const struct seq_operations ac6_seq_ops = {
1da177e4 LT	544	.start = ac6_seq_start,
	545	.next = ac6_seq_next,
	546	.stop = ac6_seq_stop,
	547	.show = ac6_seq_show,
	548	};
	549
	550	static int ac6_seq_open(struct inode inode, struct file file)
	551	{
	552	struct seq_file *seq;
	553	int rc = -ENOMEM;
0c600eda	554	struct ac6_iter_state s = kzalloc(sizeof(s), GFP_KERNEL);
1da177e4 LT	555
	556	if (!s)
	557	goto out;
	558
	559	rc = seq_open(file, &ac6_seq_ops);
	560	if (rc)
	561	goto out_kfree;
	562
	563	seq = file->private_data;
	564	seq->private = s;
1da177e4 LT	565	out:
	566	return rc;
	567	out_kfree:
	568	kfree(s);
	569	goto out;
	570	}
	571
9a32144e	572	static const struct file_operations ac6_seq_fops = {
1da177e4 LT	573	.owner = THIS_MODULE,
	574	.open = ac6_seq_open,
	575	.read = seq_read,
	576	.llseek = seq_lseek,
	577	.release = seq_release_private,
	578	};
	579
	580	int __init ac6_proc_init(void)
	581	{
457c4cbc	582	if (!proc_net_fops_create(&init_net, "anycast6", S_IRUGO, &ac6_seq_fops))
1da177e4 LT	583	return -ENOMEM;
	584
	585	return 0;
	586	}
	587
	588	void ac6_proc_exit(void)
	589	{
457c4cbc	590	proc_net_remove(&init_net, "anycast6");
1da177e4 LT	591	}
	592	#endif
	593