[people/ms/linux.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, const struct in6_addr *addr);

/*
 *	socket join an anycast group
 */

int ipv6_sock_ac_join(struct sock *sk, int ifindex, const 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;

	ASSERT_RTNL();

	if (!ns_capable(net->user_ns, 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)
		return -ENOMEM;
	pac->acl_next = NULL;
	pac->acl_addr = *addr;

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

		rt = rt6_lookup(net, addr, NULL, 0, 0);
		if (rt) {
			dev = rt->dst.dev;
			ip6_rt_put(rt);
		} else if (ishost) {
			err = -EADDRNOTAVAIL;
			goto error;
		} 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) {
		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(idev, addr);
	if (!err) {
		pac->acl_next = np->ipv6_ac_list;
		np->ipv6_ac_list = pac;
		pac = NULL;
	}

error:
	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, const 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);

	ASSERT_RTNL();

	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)
		return -ENOENT;
	if (prev_pac)
		prev_pac->acl_next = pac->acl_next;
	else
		np->ipv6_ac_list = pac->acl_next;

	dev = __dev_get_by_index(net, pac->acl_ifindex);
	if (dev)
		ipv6_dev_ac_dec(dev, &pac->acl_addr);

	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;

	if (!np->ipv6_ac_list)
		return;

	rtnl_lock();
	pac = np->ipv6_ac_list;
	np->ipv6_ac_list = NULL;

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

		if (pac->acl_ifindex != prev_index) {
			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;
	}
	rtnl_unlock();
}

static void aca_get(struct ifacaddr6 *aca)
{
	atomic_inc(&aca->aca_refcnt);
}

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);
	}
}

static struct ifacaddr6 *aca_alloc(struct rt6_info *rt,
				   const struct in6_addr *addr)
{
	struct inet6_dev *idev = rt->rt6i_idev;
	struct ifacaddr6 *aca;

	aca = kzalloc(sizeof(*aca), GFP_ATOMIC);
	if (!aca)
		return NULL;

	aca->aca_addr = *addr;
	in6_dev_hold(idev);
	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, 1);

	return aca;
}

/*
 *	device anycast group inc (add if not found)
 */
int __ipv6_dev_ac_inc(struct inet6_dev *idev, const struct in6_addr *addr)
{
	struct ifacaddr6 *aca;
	struct rt6_info *rt;
	int err;

	ASSERT_RTNL();

	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;
		}
	}

	rt = addrconf_dst_alloc(idev, addr, true);
	if (IS_ERR(rt)) {
		err = PTR_ERR(rt);
		goto out;
	}
	aca = aca_alloc(rt, addr);
	if (!aca) {
		ip6_rt_put(rt);
		err = -ENOMEM;
		goto out;
	}

	aca->aca_next = idev->ac_list;
	idev->ac_list = aca;

	/* Hold this for addrconf_join_solict() below before we unlock,
	 * it is already exposed via idev->ac_list.
	 */
	aca_get(aca);
	write_unlock_bh(&idev->lock);

	ip6_ins_rt(rt);

	addrconf_join_solict(idev->dev, &aca->aca_addr);

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

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

	ASSERT_RTNL();

	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 rtnl_lock() */
static int ipv6_dev_ac_dec(struct net_device *dev, const struct in6_addr *addr)
{
	struct inet6_dev *idev = __in6_dev_get(dev);

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

void ipv6_ac_destroy_dev(struct inet6_dev *idev)
{
	struct ifacaddr6 *aca;

	write_lock_bh(&idev->lock);
	while ((aca = idev->ac_list) != NULL) {
		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);

		write_lock_bh(&idev->lock);
	}
	write_unlock_bh(&idev->lock);
}

/*
 *	check if the interface has this anycast address
 *	called with rcu_read_lock()
 */
static bool ipv6_chk_acast_dev(struct net_device *dev, const 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 false;
}

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

	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 = true;
				break;
			}
	rcu_read_unlock();
	return found;
}

/*	check if this anycast address is link-local on given interface or
 *	is global
 */
bool ipv6_chk_acast_addr_src(struct net *net, struct net_device *dev,
			     const struct in6_addr *addr)
{
	return ipv6_chk_acast_addr(net,
				   (ipv6_addr_type(addr) & IPV6_ADDR_LINKLOCAL ?
				    dev : NULL),
				   addr);
}

#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_create("anycast6", S_IRUGO, net->proc_net, &ac6_seq_fops))
		return -ENOMEM;

	return 0;
}

void ac6_proc_exit(struct net *net)
{
	remove_proc_entry("anycast6", net->proc_net);
}
#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
b71d1d42	47	static int ipv6_dev_ac_dec(struct net_device dev, const struct in6_addr addr);
1da177e4	48
1da177e4 LT	49	/*
	50	* socket join an anycast group
	51	*/
	52
b71d1d42	53	int ipv6_sock_ac_join(struct sock sk, int ifindex, const struct in6_addr addr)
1da177e4 LT	54	{
	55	struct ipv6_pinfo *np = inet6_sk(sk);
	56	struct net_device *dev = NULL;
	57	struct inet6_dev *idev;
	58	struct ipv6_ac_socklist *pac;
6ab57e7e	59	struct net *net = sock_net(sk);
53b7997f	60	int ishost = !net->ipv6.devconf_all->forwarding;
1da177e4 LT	61	int err = 0;
1da177e4 LT	62
c4a6853d MRL	63	ASSERT_RTNL();
c4a6853d MRL	64
af31f412	65	if (!ns_capable(net->user_ns, CAP_NET_ADMIN))
1da177e4 LT	66	return -EPERM;
	67	if (ipv6_addr_is_multicast(addr))
	68	return -EINVAL;
6ab57e7e	69	if (ipv6_chk_addr(net, addr, NULL, 0))
1da177e4 LT	70	return -EINVAL;
	71
	72	pac = sock_kmalloc(sk, sizeof(struct ipv6_ac_socklist), GFP_KERNEL);
63159f29	73	if (!pac)
1da177e4 LT	74	return -ENOMEM;
1da177e4 LT	75	pac->acl_next = NULL;
4e3fd7a0	76	pac->acl_addr = *addr;
1da177e4 LT	77
	78	if (ifindex == 0) {
	79	struct rt6_info *rt;
	80
6ab57e7e	81	rt = rt6_lookup(net, addr, NULL, 0, 0);
1da177e4	82	if (rt) {
d1918542	83	dev = rt->dst.dev;
94e187c0	84	ip6_rt_put(rt);
1da177e4 LT	85	} else if (ishost) {
1da177e4 LT	86	err = -EADDRNOTAVAIL;
bb69ae04	87	goto error;
1da177e4 LT	88	} else {
1da177e4 LT	89	/* router, no matching interface: just pick one */
6c555490 WC	90	dev = __dev_get_by_flags(net, IFF_UP,
6c555490 WC	91	IFF_UP \| IFF_LOOPBACK);
1da177e4 LT	92	}
1da177e4 LT	93	} else
6c555490	94	dev = __dev_get_by_index(net, ifindex);
1da177e4	95
63159f29	96	if (!dev) {
1da177e4	97	err = -ENODEV;
bb69ae04	98	goto error;
1da177e4 LT	99	}
1da177e4 LT	100
bb69ae04	101	idev = __in6_dev_get(dev);
1da177e4 LT	102	if (!idev) {
	103	if (ifindex)
	104	err = -ENODEV;
	105	else
	106	err = -EADDRNOTAVAIL;
bb69ae04	107	goto error;
1da177e4 LT	108	}
	109	/* reset ishost, now that we have a specific device */
	110	ishost = !idev->cnf.forwarding;
1da177e4 LT	111
	112	pac->acl_ifindex = dev->ifindex;
	113
	114	/* XXX
	115	* For hosts, allow link-local or matching prefix anycasts.
	116	* This obviates the need for propagating anycast routes while
	117	* still allowing some non-router anycast participation.
	118	*/
52eeeb84	119	if (!ipv6_chk_prefix(addr, dev)) {
1da177e4 LT	120	if (ishost)
	121	err = -EADDRNOTAVAIL;
	122	if (err)
bb69ae04	123	goto error;
1da177e4 LT	124	}
1da177e4 LT	125
013b4d90	126	err = __ipv6_dev_ac_inc(idev, addr);
bb69ae04	127	if (!err) {
bb69ae04 ED	128	pac->acl_next = np->ipv6_ac_list;
bb69ae04 ED	129	np->ipv6_ac_list = pac;
bb69ae04 ED	130	pac = NULL;
bb69ae04 ED	131	}
1da177e4	132
bb69ae04	133	error:
bb69ae04 ED	134	if (pac)
bb69ae04 ED	135	sock_kfree_s(sk, pac, sizeof(*pac));
1da177e4 LT	136	return err;
	137	}
	138
	139	/*
	140	* socket leave an anycast group
	141	*/
b71d1d42	142	int ipv6_sock_ac_drop(struct sock sk, int ifindex, const struct in6_addr addr)
1da177e4 LT	143	{
	144	struct ipv6_pinfo *np = inet6_sk(sk);
	145	struct net_device *dev;
	146	struct ipv6_ac_socklist pac, prev_pac;
6ab57e7e	147	struct net *net = sock_net(sk);
1da177e4	148
c4a6853d MRL	149	ASSERT_RTNL();
c4a6853d MRL	150
1da177e4 LT	151	prev_pac = NULL;
	152	for (pac = np->ipv6_ac_list; pac; pac = pac->acl_next) {
	153	if ((ifindex == 0 \|\| pac->acl_ifindex == ifindex) &&
	154	ipv6_addr_equal(&pac->acl_addr, addr))
	155	break;
	156	prev_pac = pac;
	157	}
c4a6853d	158	if (!pac)
1da177e4	159	return -ENOENT;
1da177e4 LT	160	if (prev_pac)
	161	prev_pac->acl_next = pac->acl_next;
	162	else
	163	np->ipv6_ac_list = pac->acl_next;
	164
6c555490	165	dev = __dev_get_by_index(net, pac->acl_ifindex);
bb69ae04	166	if (dev)
1da177e4	167	ipv6_dev_ac_dec(dev, &pac->acl_addr);
bb69ae04	168
1da177e4 LT	169	sock_kfree_s(sk, pac, sizeof(*pac));
	170	return 0;
	171	}
	172
	173	void ipv6_sock_ac_close(struct sock *sk)
	174	{
	175	struct ipv6_pinfo *np = inet6_sk(sk);
	176	struct net_device *dev = NULL;
	177	struct ipv6_ac_socklist *pac;
6ab57e7e	178	struct net *net = sock_net(sk);
1da177e4 LT	179	int prev_index;
1da177e4 LT	180
0e1efe9d ED	181	if (!np->ipv6_ac_list)
	182	return;
	183
b03a9c04	184	rtnl_lock();
1da177e4 LT	185	pac = np->ipv6_ac_list;
1da177e4 LT	186	np->ipv6_ac_list = NULL;
1da177e4 LT	187
	188	prev_index = 0;
	189	while (pac) {
	190	struct ipv6_ac_socklist *next = pac->acl_next;
	191
	192	if (pac->acl_ifindex != prev_index) {
6c555490	193	dev = __dev_get_by_index(net, pac->acl_ifindex);
1da177e4 LT	194	prev_index = pac->acl_ifindex;
	195	}
	196	if (dev)
	197	ipv6_dev_ac_dec(dev, &pac->acl_addr);
	198	sock_kfree_s(sk, pac, sizeof(*pac));
	199	pac = next;
	200	}
a9ed4a29	201	rtnl_unlock();
1da177e4 LT	202	}
1da177e4 LT	203
83aa29ee WC	204	static void aca_get(struct ifacaddr6 *aca)
	205	{
	206	atomic_inc(&aca->aca_refcnt);
	207	}
	208
1da177e4 LT	209	static void aca_put(struct ifacaddr6 *ac)
	210	{
	211	if (atomic_dec_and_test(&ac->aca_refcnt)) {
	212	in6_dev_put(ac->aca_idev);
d8d1f30b	213	dst_release(&ac->aca_rt->dst);
1da177e4 LT	214	kfree(ac);
	215	}
	216	}
	217
83aa29ee WC	218	static struct ifacaddr6 aca_alloc(struct rt6_info rt,
	219	const struct in6_addr *addr)
	220	{
	221	struct inet6_dev *idev = rt->rt6i_idev;
	222	struct ifacaddr6 *aca;
	223
	224	aca = kzalloc(sizeof(*aca), GFP_ATOMIC);
63159f29	225	if (!aca)
83aa29ee WC	226	return NULL;
	227
	228	aca->aca_addr = *addr;
	229	in6_dev_hold(idev);
	230	aca->aca_idev = idev;
	231	aca->aca_rt = rt;
	232	aca->aca_users = 1;
	233	/* aca_tstamp should be updated upon changes */
	234	aca->aca_cstamp = aca->aca_tstamp = jiffies;
	235	atomic_set(&aca->aca_refcnt, 1);
83aa29ee WC	236
	237	return aca;
	238	}
	239
1da177e4 LT	240	/*
	241	* device anycast group inc (add if not found)
	242	*/
013b4d90	243	int __ipv6_dev_ac_inc(struct inet6_dev idev, const struct in6_addr addr)
1da177e4 LT	244	{
1da177e4 LT	245	struct ifacaddr6 *aca;
1da177e4 LT	246	struct rt6_info *rt;
	247	int err;
	248
a9ed4a29 SD	249	ASSERT_RTNL();
a9ed4a29 SD	250
1da177e4 LT	251	write_lock_bh(&idev->lock);
	252	if (idev->dead) {
	253	err = -ENODEV;
	254	goto out;
	255	}
	256
	257	for (aca = idev->ac_list; aca; aca = aca->aca_next) {
	258	if (ipv6_addr_equal(&aca->aca_addr, addr)) {
	259	aca->aca_users++;
	260	err = 0;
	261	goto out;
	262	}
	263	}
	264
8f031519	265	rt = addrconf_dst_alloc(idev, addr, true);
1da177e4	266	if (IS_ERR(rt)) {
1da177e4 LT	267	err = PTR_ERR(rt);
	268	goto out;
	269	}
83aa29ee	270	aca = aca_alloc(rt, addr);
63159f29	271	if (!aca) {
83aa29ee WC	272	ip6_rt_put(rt);
	273	err = -ENOMEM;
	274	goto out;
	275	}
1da177e4 LT	276
	277	aca->aca_next = idev->ac_list;
	278	idev->ac_list = aca;
83aa29ee WC	279
	280	/* Hold this for addrconf_join_solict() below before we unlock,
	281	* it is already exposed via idev->ac_list.
	282	*/
	283	aca_get(aca);
1da177e4 LT	284	write_unlock_bh(&idev->lock);
1da177e4 LT	285
84f59370	286	ip6_ins_rt(rt);
1da177e4	287
013b4d90	288	addrconf_join_solict(idev->dev, &aca->aca_addr);
1da177e4 LT	289
	290	aca_put(aca);
	291	return 0;
	292	out:
	293	write_unlock_bh(&idev->lock);
1da177e4 LT	294	return err;
	295	}
	296
	297	/*
	298	* device anycast group decrement
	299	*/
b71d1d42	300	int __ipv6_dev_ac_dec(struct inet6_dev idev, const struct in6_addr addr)
1da177e4 LT	301	{
	302	struct ifacaddr6 aca, prev_aca;
	303
a9ed4a29 SD	304	ASSERT_RTNL();
a9ed4a29 SD	305
1da177e4 LT	306	write_lock_bh(&idev->lock);
	307	prev_aca = NULL;
	308	for (aca = idev->ac_list; aca; aca = aca->aca_next) {
	309	if (ipv6_addr_equal(&aca->aca_addr, addr))
	310	break;
	311	prev_aca = aca;
	312	}
	313	if (!aca) {
	314	write_unlock_bh(&idev->lock);
	315	return -ENOENT;
	316	}
	317	if (--aca->aca_users > 0) {
	318	write_unlock_bh(&idev->lock);
	319	return 0;
	320	}
	321	if (prev_aca)
	322	prev_aca->aca_next = aca->aca_next;
	323	else
	324	idev->ac_list = aca->aca_next;
	325	write_unlock_bh(&idev->lock);
	326	addrconf_leave_solict(idev, &aca->aca_addr);
	327
d8d1f30b	328	dst_hold(&aca->aca_rt->dst);
84f59370	329	ip6_del_rt(aca->aca_rt);
1da177e4 LT	330
	331	aca_put(aca);
	332	return 0;
	333	}
	334
6c555490	335	/* called with rtnl_lock() */
b71d1d42	336	static int ipv6_dev_ac_dec(struct net_device dev, const struct in6_addr addr)
1da177e4	337	{
bb69ae04 ED	338	struct inet6_dev *idev = __in6_dev_get(dev);
bb69ae04 ED	339
63159f29	340	if (!idev)
1da177e4	341	return -ENODEV;
bb69ae04	342	return __ipv6_dev_ac_dec(idev, addr);
1da177e4	343	}
1ab1457c	344
381f4dca SD	345	void ipv6_ac_destroy_dev(struct inet6_dev *idev)
	346	{
	347	struct ifacaddr6 *aca;
	348
	349	write_lock_bh(&idev->lock);
	350	while ((aca = idev->ac_list) != NULL) {
	351	idev->ac_list = aca->aca_next;
	352	write_unlock_bh(&idev->lock);
	353
	354	addrconf_leave_solict(idev, &aca->aca_addr);
	355
	356	dst_hold(&aca->aca_rt->dst);
	357	ip6_del_rt(aca->aca_rt);
	358
	359	aca_put(aca);
	360
	361	write_lock_bh(&idev->lock);
	362	}
	363	write_unlock_bh(&idev->lock);
	364	}
	365
1da177e4 LT	366	/*
1da177e4 LT	367	* check if the interface has this anycast address
bb69ae04	368	* called with rcu_read_lock()
1da177e4	369	*/
a50feda5	370	static bool ipv6_chk_acast_dev(struct net_device dev, const struct in6_addr addr)
1da177e4 LT	371	{
	372	struct inet6_dev *idev;
	373	struct ifacaddr6 *aca;
	374
bb69ae04	375	idev = __in6_dev_get(dev);
1da177e4 LT	376	if (idev) {
	377	read_lock_bh(&idev->lock);
	378	for (aca = idev->ac_list; aca; aca = aca->aca_next)
	379	if (ipv6_addr_equal(&aca->aca_addr, addr))
	380	break;
	381	read_unlock_bh(&idev->lock);
cfcabdcc	382	return aca != NULL;
1da177e4	383	}
a50feda5	384	return false;
1da177e4 LT	385	}
	386
	387	/*
	388	* check if given interface (or any, if dev==0) has this anycast address
	389	*/
a50feda5 ED	390	bool ipv6_chk_acast_addr(struct net net, struct net_device dev,
a50feda5 ED	391	const struct in6_addr *addr)
1da177e4	392	{
a50feda5	393	bool found = false;
7562f876	394
c6d14c84	395	rcu_read_lock();
bb69ae04 ED	396	if (dev)
	397	found = ipv6_chk_acast_dev(dev, addr);
	398	else
	399	for_each_netdev_rcu(net, dev)
	400	if (ipv6_chk_acast_dev(dev, addr)) {
a50feda5	401	found = true;
bb69ae04 ED	402	break;
bb69ae04 ED	403	}
c6d14c84	404	rcu_read_unlock();
7562f876	405	return found;
1da177e4 LT	406	}
1da177e4 LT	407
7c90cc2d FLB	408	/* check if this anycast address is link-local on given interface or
	409	* is global
	410	*/
	411	bool ipv6_chk_acast_addr_src(struct net net, struct net_device dev,
	412	const struct in6_addr *addr)
	413	{
	414	return ipv6_chk_acast_addr(net,
	415	(ipv6_addr_type(addr) & IPV6_ADDR_LINKLOCAL ?
	416	dev : NULL),
	417	addr);
	418	}
1da177e4 LT	419
	420	#ifdef CONFIG_PROC_FS
	421	struct ac6_iter_state {
6ab57e7e	422	struct seq_net_private p;
1da177e4 LT	423	struct net_device *dev;
	424	struct inet6_dev *idev;
	425	};
	426
	427	#define ac6_seq_private(seq) ((struct ac6_iter_state *)(seq)->private)
	428
	429	static inline struct ifacaddr6 ac6_get_first(struct seq_file seq)
	430	{
	431	struct ifacaddr6 *im = NULL;
	432	struct ac6_iter_state *state = ac6_seq_private(seq);
6ab57e7e	433	struct net *net = seq_file_net(seq);
1da177e4	434
7562f876	435	state->idev = NULL;
ce81b76a	436	for_each_netdev_rcu(net, state->dev) {
1da177e4	437	struct inet6_dev *idev;
ce81b76a	438	idev = __in6_dev_get(state->dev);
1da177e4 LT	439	if (!idev)
	440	continue;
	441	read_lock_bh(&idev->lock);
	442	im = idev->ac_list;
	443	if (im) {
	444	state->idev = idev;
	445	break;
	446	}
	447	read_unlock_bh(&idev->lock);
	448	}
	449	return im;
	450	}
	451
	452	static struct ifacaddr6 ac6_get_next(struct seq_file seq, struct ifacaddr6 *im)
	453	{
	454	struct ac6_iter_state *state = ac6_seq_private(seq);
	455
	456	im = im->aca_next;
	457	while (!im) {
ce81b76a	458	if (likely(state->idev != NULL))
1da177e4	459	read_unlock_bh(&state->idev->lock);
ce81b76a ED	460
ce81b76a ED	461	state->dev = next_net_device_rcu(state->dev);
1da177e4 LT	462	if (!state->dev) {
	463	state->idev = NULL;
	464	break;
	465	}
ce81b76a	466	state->idev = __in6_dev_get(state->dev);
1da177e4 LT	467	if (!state->idev)
	468	continue;
	469	read_lock_bh(&state->idev->lock);
	470	im = state->idev->ac_list;
	471	}
	472	return im;
	473	}
	474
	475	static struct ifacaddr6 ac6_get_idx(struct seq_file seq, loff_t pos)
	476	{
	477	struct ifacaddr6 *im = ac6_get_first(seq);
	478	if (im)
	479	while (pos && (im = ac6_get_next(seq, im)) != NULL)
	480	--pos;
	481	return pos ? NULL : im;
	482	}
	483
	484	static void ac6_seq_start(struct seq_file seq, loff_t *pos)
ce81b76a	485	__acquires(RCU)
1da177e4	486	{
ce81b76a	487	rcu_read_lock();
1da177e4 LT	488	return ac6_get_idx(seq, *pos);
	489	}
	490
	491	static void ac6_seq_next(struct seq_file seq, void v, loff_t pos)
	492	{
ce81b76a ED	493	struct ifacaddr6 *im = ac6_get_next(seq, v);
ce81b76a ED	494
1da177e4 LT	495	++*pos;
	496	return im;
	497	}
	498
	499	static void ac6_seq_stop(struct seq_file seq, void v)
ce81b76a	500	__releases(RCU)
1da177e4 LT	501	{
1da177e4 LT	502	struct ac6_iter_state *state = ac6_seq_private(seq);
ce81b76a	503
1da177e4 LT	504	if (likely(state->idev != NULL)) {
1da177e4 LT	505	read_unlock_bh(&state->idev->lock);
ce81b76a	506	state->idev = NULL;
1da177e4	507	}
ce81b76a	508	rcu_read_unlock();
1da177e4 LT	509	}
	510
	511	static int ac6_seq_show(struct seq_file seq, void v)
	512	{
	513	struct ifacaddr6 im = (struct ifacaddr6 )v;
	514	struct ac6_iter_state *state = ac6_seq_private(seq);
	515
4b7a4274	516	seq_printf(seq, "%-4d %-15s %pi6 %5d\n",
1da177e4	517	state->dev->ifindex, state->dev->name,
b071195d	518	&im->aca_addr, im->aca_users);
1da177e4 LT	519	return 0;
	520	}
	521
56b3d975	522	static const struct seq_operations ac6_seq_ops = {
1da177e4 LT	523	.start = ac6_seq_start,
	524	.next = ac6_seq_next,
	525	.stop = ac6_seq_stop,
	526	.show = ac6_seq_show,
	527	};
	528
	529	static int ac6_seq_open(struct inode inode, struct file file)
	530	{
6ab57e7e DL	531	return seq_open_net(inode, file, &ac6_seq_ops,
6ab57e7e DL	532	sizeof(struct ac6_iter_state));
1da177e4 LT	533	}
1da177e4 LT	534
9a32144e	535	static const struct file_operations ac6_seq_fops = {
1da177e4 LT	536	.owner = THIS_MODULE,
	537	.open = ac6_seq_open,
	538	.read = seq_read,
	539	.llseek = seq_lseek,
6ab57e7e	540	.release = seq_release_net,
1da177e4 LT	541	};
1da177e4 LT	542
2c8c1e72	543	int __net_init ac6_proc_init(struct net *net)
1da177e4	544	{
d4beaa66	545	if (!proc_create("anycast6", S_IRUGO, net->proc_net, &ac6_seq_fops))
1da177e4 LT	546	return -ENOMEM;
	547
	548	return 0;
	549	}
	550
6ab57e7e	551	void ac6_proc_exit(struct net *net)
1da177e4	552	{
ece31ffd	553	remove_proc_entry("anycast6", net->proc_net);
1da177e4 LT	554	}
	555	#endif
	556