addrconf.c 144.5 KB
Newer Older
L
Linus Torvalds 已提交
1 2 3 4 5
/*
 *	IPv6 Address [auto]configuration
 *	Linux INET6 implementation
 *
 *	Authors:
6
 *	Pedro Roque		<roque@di.fc.ul.pt>
L
Linus Torvalds 已提交
7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35
 *	Alexey Kuznetsov	<kuznet@ms2.inr.ac.ru>
 *
 *	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.
 */

/*
 *	Changes:
 *
 *	Janos Farkas			:	delete timer on ifdown
 *	<chexum@bankinf.banki.hu>
 *	Andi Kleen			:	kill double kfree on module
 *						unload.
 *	Maciej W. Rozycki		:	FDDI support
 *	sekiya@USAGI			:	Don't send too many RS
 *						packets.
 *	yoshfuji@USAGI			:       Fixed interval between DAD
 *						packets.
 *	YOSHIFUJI Hideaki @USAGI	:	improved accuracy of
 *						address validation timer.
 *	YOSHIFUJI Hideaki @USAGI	:	Privacy Extensions (RFC3041)
 *						support.
 *	Yuji SEKIYA @USAGI		:	Don't assign a same IPv6
 *						address on a same interface.
 *	YOSHIFUJI Hideaki @USAGI	:	ARCnet support
 *	YOSHIFUJI Hideaki @USAGI	:	convert /proc/net/if_inet6 to
 *						seq_file.
36 37 38
 *	YOSHIFUJI Hideaki @USAGI	:	improved source address
 *						selection; consider scope,
 *						status etc.
L
Linus Torvalds 已提交
39 40
 */

41 42
#define pr_fmt(fmt) "IPv6: " fmt

L
Linus Torvalds 已提交
43 44
#include <linux/errno.h>
#include <linux/types.h>
45
#include <linux/kernel.h>
L
Linus Torvalds 已提交
46 47 48
#include <linux/socket.h>
#include <linux/sockios.h>
#include <linux/net.h>
49
#include <linux/inet.h>
L
Linus Torvalds 已提交
50 51
#include <linux/in6.h>
#include <linux/netdevice.h>
52
#include <linux/if_addr.h>
L
Linus Torvalds 已提交
53 54 55 56 57 58
#include <linux/if_arp.h>
#include <linux/if_arcnet.h>
#include <linux/if_infiniband.h>
#include <linux/route.h>
#include <linux/inetdevice.h>
#include <linux/init.h>
59
#include <linux/slab.h>
L
Linus Torvalds 已提交
60 61 62
#ifdef CONFIG_SYSCTL
#include <linux/sysctl.h>
#endif
63
#include <linux/capability.h>
L
Linus Torvalds 已提交
64 65
#include <linux/delay.h>
#include <linux/notifier.h>
66
#include <linux/string.h>
E
Eric Dumazet 已提交
67
#include <linux/hash.h>
L
Linus Torvalds 已提交
68

69
#include <net/net_namespace.h>
L
Linus Torvalds 已提交
70 71 72
#include <net/sock.h>
#include <net/snmp.h>

73
#include <net/af_ieee802154.h>
74
#include <net/firewire.h>
L
Linus Torvalds 已提交
75 76 77 78 79 80 81
#include <net/ipv6.h>
#include <net/protocol.h>
#include <net/ndisc.h>
#include <net/ip6_route.h>
#include <net/addrconf.h>
#include <net/tcp.h>
#include <net/ip.h>
82
#include <net/netlink.h>
83
#include <net/pkt_sched.h>
L
Linus Torvalds 已提交
84 85
#include <linux/if_tunnel.h>
#include <linux/rtnetlink.h>
86
#include <linux/netconf.h>
L
Linus Torvalds 已提交
87
#include <linux/random.h>
S
Stephen Hemminger 已提交
88
#include <linux/uaccess.h>
89
#include <asm/unaligned.h>
L
Linus Torvalds 已提交
90 91 92

#include <linux/proc_fs.h>
#include <linux/seq_file.h>
93
#include <linux/export.h>
L
Linus Torvalds 已提交
94 95 96 97 98

/* Set to 3 to get tracing... */
#define ACONF_DEBUG 2

#if ACONF_DEBUG >= 3
99
#define ADBG(fmt, ...) printk(fmt, ##__VA_ARGS__)
L
Linus Torvalds 已提交
100
#else
101
#define ADBG(fmt, ...) do { if (0) printk(fmt, ##__VA_ARGS__); } while (0)
L
Linus Torvalds 已提交
102 103 104
#endif

#define	INFINITY_LIFE_TIME	0xFFFFFFFF
105

106 107 108
#define IPV6_MAX_STRLEN \
	sizeof("ffff:ffff:ffff:ffff:ffff:ffff:255.255.255.255")

109 110 111 112
static inline u32 cstamp_delta(unsigned long cstamp)
{
	return (cstamp - INITIAL_JIFFIES) * 100UL / HZ;
}
L
Linus Torvalds 已提交
113 114

#ifdef CONFIG_SYSCTL
115
static int addrconf_sysctl_register(struct inet6_dev *idev);
116 117
static void addrconf_sysctl_unregister(struct inet6_dev *idev);
#else
118
static inline int addrconf_sysctl_register(struct inet6_dev *idev)
119
{
120
	return 0;
121 122 123 124 125
}

static inline void addrconf_sysctl_unregister(struct inet6_dev *idev)
{
}
L
Linus Torvalds 已提交
126 127
#endif

S
Sorin Dumitru 已提交
128 129
static void __ipv6_regen_rndid(struct inet6_dev *idev);
static void __ipv6_try_regen_rndid(struct inet6_dev *idev, struct in6_addr *tmpaddr);
L
Linus Torvalds 已提交
130 131
static void ipv6_regen_rndid(unsigned long data);

132
static int ipv6_generate_eui64(u8 *eui, struct net_device *dev);
L
Linus Torvalds 已提交
133
static int ipv6_count_addresses(struct inet6_dev *idev);
134 135 136
static int ipv6_generate_stable_address(struct in6_addr *addr,
					u8 dad_count,
					const struct inet6_dev *idev);
L
Linus Torvalds 已提交
137 138 139 140

/*
 *	Configured unicast address hash table
 */
141
static struct hlist_head inet6_addr_lst[IN6_ADDR_HSIZE];
142
static DEFINE_SPINLOCK(addrconf_hash_lock);
L
Linus Torvalds 已提交
143

144 145 146
static void addrconf_verify(void);
static void addrconf_verify_rtnl(void);
static void addrconf_verify_work(struct work_struct *);
L
Linus Torvalds 已提交
147

148 149
static struct workqueue_struct *addrconf_wq;
static DECLARE_DELAYED_WORK(addr_chk_work, addrconf_verify_work);
L
Linus Torvalds 已提交
150 151 152 153

static void addrconf_join_anycast(struct inet6_ifaddr *ifp);
static void addrconf_leave_anycast(struct inet6_ifaddr *ifp);

154 155
static void addrconf_type_change(struct net_device *dev,
				 unsigned long event);
L
Linus Torvalds 已提交
156 157
static int addrconf_ifdown(struct net_device *dev, int how);

158 159 160 161 162
static struct rt6_info *addrconf_get_prefix_route(const struct in6_addr *pfx,
						  int plen,
						  const struct net_device *dev,
						  u32 flags, u32 noflags);

163
static void addrconf_dad_start(struct inet6_ifaddr *ifp);
164
static void addrconf_dad_work(struct work_struct *w);
L
Linus Torvalds 已提交
165
static void addrconf_dad_completed(struct inet6_ifaddr *ifp);
166
static void addrconf_dad_run(struct inet6_dev *idev);
L
Linus Torvalds 已提交
167 168 169 170
static void addrconf_rs_timer(unsigned long data);
static void __ipv6_ifa_notify(int event, struct inet6_ifaddr *ifa);
static void ipv6_ifa_notify(int event, struct inet6_ifaddr *ifa);

171
static void inet6_prefix_notify(int event, struct inet6_dev *idev,
L
Linus Torvalds 已提交
172
				struct prefix_info *pinfo);
173 174
static bool ipv6_chk_same_addr(struct net *net, const struct in6_addr *addr,
			       struct net_device *dev);
L
Linus Torvalds 已提交
175

176
static struct ipv6_devconf ipv6_devconf __read_mostly = {
L
Linus Torvalds 已提交
177 178 179 180 181 182 183
	.forwarding		= 0,
	.hop_limit		= IPV6_DEFAULT_HOPLIMIT,
	.mtu6			= IPV6_MIN_MTU,
	.accept_ra		= 1,
	.accept_redirects	= 1,
	.autoconf		= 1,
	.force_mld_version	= 0,
184 185
	.mldv1_unsolicited_report_interval = 10 * HZ,
	.mldv2_unsolicited_report_interval = HZ,
L
Linus Torvalds 已提交
186 187 188 189
	.dad_transmits		= 1,
	.rtr_solicits		= MAX_RTR_SOLICITATIONS,
	.rtr_solicit_interval	= RTR_SOLICITATION_INTERVAL,
	.rtr_solicit_delay	= MAX_RTR_SOLICITATION_DELAY,
190
	.use_tempaddr		= 0,
L
Linus Torvalds 已提交
191 192 193 194 195
	.temp_valid_lft		= TEMP_VALID_LIFETIME,
	.temp_prefered_lft	= TEMP_PREFERRED_LIFETIME,
	.regen_max_retry	= REGEN_MAX_RETRY,
	.max_desync_factor	= MAX_DESYNC_FACTOR,
	.max_addresses		= IPV6_MAX_ADDRESSES,
196
	.accept_ra_defrtr	= 1,
197
	.accept_ra_from_local	= 0,
198
	.accept_ra_min_hop_limit= 1,
199
	.accept_ra_pinfo	= 1,
200 201
#ifdef CONFIG_IPV6_ROUTER_PREF
	.accept_ra_rtr_pref	= 1,
202
	.rtr_probe_interval	= 60 * HZ,
203 204 205
#ifdef CONFIG_IPV6_ROUTE_INFO
	.accept_ra_rt_info_max_plen = 0,
#endif
206
#endif
207
	.proxy_ndp		= 0,
208
	.accept_source_route	= 0,	/* we do not accept RH0 by default. */
209
	.disable_ipv6		= 0,
210
	.accept_dad		= 1,
211
	.suppress_frag_ndisc	= 1,
212
	.accept_ra_mtu		= 1,
213 214
	.stable_secret		= {
		.initialized = false,
215 216
	},
	.use_oif_addrs_only	= 0,
217
	.ignore_routes_with_linkdown = 0,
L
Linus Torvalds 已提交
218 219
};

220
static struct ipv6_devconf ipv6_devconf_dflt __read_mostly = {
L
Linus Torvalds 已提交
221 222 223 224 225 226
	.forwarding		= 0,
	.hop_limit		= IPV6_DEFAULT_HOPLIMIT,
	.mtu6			= IPV6_MIN_MTU,
	.accept_ra		= 1,
	.accept_redirects	= 1,
	.autoconf		= 1,
227 228 229
	.force_mld_version	= 0,
	.mldv1_unsolicited_report_interval = 10 * HZ,
	.mldv2_unsolicited_report_interval = HZ,
L
Linus Torvalds 已提交
230 231 232 233 234 235 236 237 238 239
	.dad_transmits		= 1,
	.rtr_solicits		= MAX_RTR_SOLICITATIONS,
	.rtr_solicit_interval	= RTR_SOLICITATION_INTERVAL,
	.rtr_solicit_delay	= MAX_RTR_SOLICITATION_DELAY,
	.use_tempaddr		= 0,
	.temp_valid_lft		= TEMP_VALID_LIFETIME,
	.temp_prefered_lft	= TEMP_PREFERRED_LIFETIME,
	.regen_max_retry	= REGEN_MAX_RETRY,
	.max_desync_factor	= MAX_DESYNC_FACTOR,
	.max_addresses		= IPV6_MAX_ADDRESSES,
240
	.accept_ra_defrtr	= 1,
241
	.accept_ra_from_local	= 0,
242
	.accept_ra_min_hop_limit= 1,
243
	.accept_ra_pinfo	= 1,
244 245
#ifdef CONFIG_IPV6_ROUTER_PREF
	.accept_ra_rtr_pref	= 1,
246
	.rtr_probe_interval	= 60 * HZ,
247 248 249
#ifdef CONFIG_IPV6_ROUTE_INFO
	.accept_ra_rt_info_max_plen = 0,
#endif
250
#endif
251
	.proxy_ndp		= 0,
252
	.accept_source_route	= 0,	/* we do not accept RH0 by default. */
253
	.disable_ipv6		= 0,
254
	.accept_dad		= 1,
255
	.suppress_frag_ndisc	= 1,
256
	.accept_ra_mtu		= 1,
257 258 259
	.stable_secret		= {
		.initialized = false,
	},
260
	.use_oif_addrs_only	= 0,
261
	.ignore_routes_with_linkdown = 0,
L
Linus Torvalds 已提交
262 263
};

264
/* Check if a valid qdisc is available */
265
static inline bool addrconf_qdisc_ok(const struct net_device *dev)
266
{
267
	return !qdisc_tx_is_noop(dev);
268 269
}

270
static void addrconf_del_rs_timer(struct inet6_dev *idev)
L
Linus Torvalds 已提交
271
{
272 273 274 275
	if (del_timer(&idev->rs_timer))
		__in6_dev_put(idev);
}

276
static void addrconf_del_dad_work(struct inet6_ifaddr *ifp)
277
{
278
	if (cancel_delayed_work(&ifp->dad_work))
L
Linus Torvalds 已提交
279 280 281
		__in6_ifa_put(ifp);
}

282 283 284 285 286 287 288
static void addrconf_mod_rs_timer(struct inet6_dev *idev,
				  unsigned long when)
{
	if (!timer_pending(&idev->rs_timer))
		in6_dev_hold(idev);
	mod_timer(&idev->rs_timer, jiffies + when);
}
L
Linus Torvalds 已提交
289

290 291
static void addrconf_mod_dad_work(struct inet6_ifaddr *ifp,
				   unsigned long delay)
L
Linus Torvalds 已提交
292
{
293
	if (!delayed_work_pending(&ifp->dad_work))
L
Linus Torvalds 已提交
294
		in6_ifa_hold(ifp);
295
	mod_delayed_work(addrconf_wq, &ifp->dad_work, delay);
L
Linus Torvalds 已提交
296 297
}

298 299
static int snmp6_alloc_dev(struct inet6_dev *idev)
{
300 301
	int i;

W
WANG Cong 已提交
302 303
	idev->stats.ipv6 = alloc_percpu(struct ipstats_mib);
	if (!idev->stats.ipv6)
304
		goto err_ip;
305 306 307

	for_each_possible_cpu(i) {
		struct ipstats_mib *addrconf_stats;
W
WANG Cong 已提交
308
		addrconf_stats = per_cpu_ptr(idev->stats.ipv6, i);
309 310 311 312
		u64_stats_init(&addrconf_stats->syncp);
	}


313 314 315
	idev->stats.icmpv6dev = kzalloc(sizeof(struct icmpv6_mib_device),
					GFP_KERNEL);
	if (!idev->stats.icmpv6dev)
316
		goto err_icmp;
317 318 319
	idev->stats.icmpv6msgdev = kzalloc(sizeof(struct icmpv6msg_mib_device),
					   GFP_KERNEL);
	if (!idev->stats.icmpv6msgdev)
320
		goto err_icmpmsg;
321 322 323

	return 0;

324
err_icmpmsg:
325
	kfree(idev->stats.icmpv6dev);
326
err_icmp:
W
WANG Cong 已提交
327
	free_percpu(idev->stats.ipv6);
328
err_ip:
329
	return -ENOMEM;
330 331
}

332
static struct inet6_dev *ipv6_add_dev(struct net_device *dev)
L
Linus Torvalds 已提交
333 334
{
	struct inet6_dev *ndev;
335
	int err = -ENOMEM;
L
Linus Torvalds 已提交
336 337 338 339

	ASSERT_RTNL();

	if (dev->mtu < IPV6_MIN_MTU)
340
		return ERR_PTR(-EINVAL);
L
Linus Torvalds 已提交
341

342
	ndev = kzalloc(sizeof(struct inet6_dev), GFP_KERNEL);
343
	if (!ndev)
344
		return ERR_PTR(err);
345 346 347

	rwlock_init(&ndev->lock);
	ndev->dev = dev;
348
	INIT_LIST_HEAD(&ndev->addr_list);
349 350
	setup_timer(&ndev->rs_timer, addrconf_rs_timer,
		    (unsigned long)ndev);
351
	memcpy(&ndev->cnf, dev_net(dev)->ipv6.devconf_dflt, sizeof(ndev->cnf));
352 353 354
	ndev->cnf.mtu6 = dev->mtu;
	ndev->cnf.sysctl = NULL;
	ndev->nd_parms = neigh_parms_alloc(dev, &nd_tbl);
355
	if (!ndev->nd_parms) {
356
		kfree(ndev);
357
		return ERR_PTR(err);
358
	}
359 360
	if (ndev->cnf.forwarding)
		dev_disable_lro(dev);
361 362
	/* We refer to the device */
	dev_hold(dev);
L
Linus Torvalds 已提交
363

364
	if (snmp6_alloc_dev(ndev) < 0) {
365
		ADBG(KERN_WARNING
366
			"%s: cannot allocate memory for statistics; dev=%s.\n",
367
			__func__, dev->name);
368
		neigh_parms_release(&nd_tbl, ndev->nd_parms);
R
Roy Li 已提交
369 370
		dev_put(dev);
		kfree(ndev);
371
		return ERR_PTR(err);
372
	}
L
Linus Torvalds 已提交
373

374
	if (snmp6_register_dev(ndev) < 0) {
375
		ADBG(KERN_WARNING
376
			"%s: cannot create /proc/net/dev_snmp6/%s\n",
377
			__func__, dev->name);
378
		goto err_release;
379 380 381 382 383 384
	}

	/* One reference from device.  We must do this before
	 * we invoke __ipv6_regen_rndid().
	 */
	in6_dev_hold(ndev);
L
Linus Torvalds 已提交
385

386 387 388
	if (dev->flags & (IFF_NOARP | IFF_LOOPBACK))
		ndev->cnf.accept_dad = -1;

A
Amerigo Wang 已提交
389
#if IS_ENABLED(CONFIG_IPV6_SIT)
390
	if (dev->type == ARPHRD_SIT && (dev->priv_flags & IFF_ISATAP)) {
391
		pr_info("%s: Disabled Multicast RS\n", dev->name);
392 393 394 395
		ndev->cnf.rtr_solicits = 0;
	}
#endif

396
	INIT_LIST_HEAD(&ndev->tempaddr_list);
397
	setup_timer(&ndev->regen_timer, ipv6_regen_rndid, (unsigned long)ndev);
398 399
	if ((dev->flags&IFF_LOOPBACK) ||
	    dev->type == ARPHRD_TUNNEL ||
400
	    dev->type == ARPHRD_TUNNEL6 ||
401 402
	    dev->type == ARPHRD_SIT ||
	    dev->type == ARPHRD_NONE) {
403 404 405 406 407
		ndev->cnf.use_tempaddr = -1;
	} else {
		in6_dev_hold(ndev);
		ipv6_regen_rndid((unsigned long) ndev);
	}
408

409
	ndev->token = in6addr_any;
L
Linus Torvalds 已提交
410

411
	if (netif_running(dev) && addrconf_qdisc_ok(dev))
412 413
		ndev->if_flags |= IF_READY;

414 415
	ipv6_mc_init_dev(ndev);
	ndev->tstamp = jiffies;
416 417 418 419 420 421
	err = addrconf_sysctl_register(ndev);
	if (err) {
		ipv6_mc_destroy_dev(ndev);
		del_timer(&ndev->regen_timer);
		goto err_release;
	}
422
	/* protected by rtnl_lock */
423
	rcu_assign_pointer(dev->ip6_ptr, ndev);
424

425 426 427
	/* Join interface-local all-node multicast group */
	ipv6_dev_mc_inc(dev, &in6addr_interfacelocal_allnodes);

428
	/* Join all-node multicast group */
429
	ipv6_dev_mc_inc(dev, &in6addr_linklocal_allnodes);
430

431
	/* Join all-router multicast group if forwarding is set */
L
Li Wei 已提交
432
	if (ndev->cnf.forwarding && (dev->flags & IFF_MULTICAST))
433 434
		ipv6_dev_mc_inc(dev, &in6addr_linklocal_allrouters);

L
Linus Torvalds 已提交
435
	return ndev;
436 437 438 439 440 441

err_release:
	neigh_parms_release(&nd_tbl, ndev->nd_parms);
	ndev->dead = 1;
	in6_dev_finish_destroy(ndev);
	return ERR_PTR(err);
L
Linus Torvalds 已提交
442 443
}

444
static struct inet6_dev *ipv6_find_idev(struct net_device *dev)
L
Linus Torvalds 已提交
445 446 447 448 449
{
	struct inet6_dev *idev;

	ASSERT_RTNL();

S
Stephen Hemminger 已提交
450 451 452
	idev = __in6_dev_get(dev);
	if (!idev) {
		idev = ipv6_add_dev(dev);
453
		if (IS_ERR(idev))
L
Linus Torvalds 已提交
454 455
			return NULL;
	}
456

L
Linus Torvalds 已提交
457 458 459 460 461
	if (dev->flags&IFF_UP)
		ipv6_mc_up(idev);
	return idev;
}

462 463 464 465 466
static int inet6_netconf_msgsize_devconf(int type)
{
	int size =  NLMSG_ALIGN(sizeof(struct netconfmsg))
		    + nla_total_size(4);	/* NETCONFA_IFINDEX */

467 468
	/* type -1 is used for ALL */
	if (type == -1 || type == NETCONFA_FORWARDING)
469
		size += nla_total_size(4);
470
#ifdef CONFIG_IPV6_MROUTE
471 472
	if (type == -1 || type == NETCONFA_MC_FORWARDING)
		size += nla_total_size(4);
473
#endif
474 475
	if (type == -1 || type == NETCONFA_PROXY_NEIGH)
		size += nla_total_size(4);
476

477 478 479
	if (type == -1 || type == NETCONFA_IGNORE_ROUTES_WITH_LINKDOWN)
		size += nla_total_size(4);

480 481 482 483 484 485 486 487 488 489 490 491 492
	return size;
}

static int inet6_netconf_fill_devconf(struct sk_buff *skb, int ifindex,
				      struct ipv6_devconf *devconf, u32 portid,
				      u32 seq, int event, unsigned int flags,
				      int type)
{
	struct nlmsghdr  *nlh;
	struct netconfmsg *ncm;

	nlh = nlmsg_put(skb, portid, seq, event, sizeof(struct netconfmsg),
			flags);
493
	if (!nlh)
494 495 496 497 498 499 500 501
		return -EMSGSIZE;

	ncm = nlmsg_data(nlh);
	ncm->ncm_family = AF_INET6;

	if (nla_put_s32(skb, NETCONFA_IFINDEX, ifindex) < 0)
		goto nla_put_failure;

502 503
	/* type -1 is used for ALL */
	if ((type == -1 || type == NETCONFA_FORWARDING) &&
504 505
	    nla_put_s32(skb, NETCONFA_FORWARDING, devconf->forwarding) < 0)
		goto nla_put_failure;
506
#ifdef CONFIG_IPV6_MROUTE
507 508 509 510
	if ((type == -1 || type == NETCONFA_MC_FORWARDING) &&
	    nla_put_s32(skb, NETCONFA_MC_FORWARDING,
			devconf->mc_forwarding) < 0)
		goto nla_put_failure;
511
#endif
512 513 514 515
	if ((type == -1 || type == NETCONFA_PROXY_NEIGH) &&
	    nla_put_s32(skb, NETCONFA_PROXY_NEIGH, devconf->proxy_ndp) < 0)
		goto nla_put_failure;

516 517 518 519 520
	if ((type == -1 || type == NETCONFA_IGNORE_ROUTES_WITH_LINKDOWN) &&
	    nla_put_s32(skb, NETCONFA_IGNORE_ROUTES_WITH_LINKDOWN,
			devconf->ignore_routes_with_linkdown) < 0)
		goto nla_put_failure;

521 522
	nlmsg_end(skb, nlh);
	return 0;
523 524 525 526 527 528

nla_put_failure:
	nlmsg_cancel(skb, nlh);
	return -EMSGSIZE;
}

529 530
void inet6_netconf_notify_devconf(struct net *net, int type, int ifindex,
				  struct ipv6_devconf *devconf)
531 532 533 534 535
{
	struct sk_buff *skb;
	int err = -ENOBUFS;

	skb = nlmsg_new(inet6_netconf_msgsize_devconf(type), GFP_ATOMIC);
536
	if (!skb)
537 538 539 540 541 542 543 544 545 546 547 548 549
		goto errout;

	err = inet6_netconf_fill_devconf(skb, ifindex, devconf, 0, 0,
					 RTM_NEWNETCONF, 0, type);
	if (err < 0) {
		/* -EMSGSIZE implies BUG in inet6_netconf_msgsize_devconf() */
		WARN_ON(err == -EMSGSIZE);
		kfree_skb(skb);
		goto errout;
	}
	rtnl_notify(skb, net, 0, RTNLGRP_IPV6_NETCONF, NULL, GFP_ATOMIC);
	return;
errout:
550
	rtnl_set_sk_err(net, RTNLGRP_IPV6_NETCONF, err);
551 552
}

553 554 555
static const struct nla_policy devconf_ipv6_policy[NETCONFA_MAX+1] = {
	[NETCONFA_IFINDEX]	= { .len = sizeof(int) },
	[NETCONFA_FORWARDING]	= { .len = sizeof(int) },
556
	[NETCONFA_PROXY_NEIGH]	= { .len = sizeof(int) },
557
	[NETCONFA_IGNORE_ROUTES_WITH_LINKDOWN]	= { .len = sizeof(int) },
558 559 560
};

static int inet6_netconf_get_devconf(struct sk_buff *in_skb,
561
				     struct nlmsghdr *nlh)
562 563 564 565 566 567 568 569 570 571 572 573 574 575 576 577 578 579 580 581 582 583 584 585 586 587 588 589 590 591
{
	struct net *net = sock_net(in_skb->sk);
	struct nlattr *tb[NETCONFA_MAX+1];
	struct netconfmsg *ncm;
	struct sk_buff *skb;
	struct ipv6_devconf *devconf;
	struct inet6_dev *in6_dev;
	struct net_device *dev;
	int ifindex;
	int err;

	err = nlmsg_parse(nlh, sizeof(*ncm), tb, NETCONFA_MAX,
			  devconf_ipv6_policy);
	if (err < 0)
		goto errout;

	err = EINVAL;
	if (!tb[NETCONFA_IFINDEX])
		goto errout;

	ifindex = nla_get_s32(tb[NETCONFA_IFINDEX]);
	switch (ifindex) {
	case NETCONFA_IFINDEX_ALL:
		devconf = net->ipv6.devconf_all;
		break;
	case NETCONFA_IFINDEX_DEFAULT:
		devconf = net->ipv6.devconf_dflt;
		break;
	default:
		dev = __dev_get_by_index(net, ifindex);
592
		if (!dev)
593 594
			goto errout;
		in6_dev = __in6_dev_get(dev);
595
		if (!in6_dev)
596 597 598 599 600 601 602
			goto errout;
		devconf = &in6_dev->cnf;
		break;
	}

	err = -ENOBUFS;
	skb = nlmsg_new(inet6_netconf_msgsize_devconf(-1), GFP_ATOMIC);
603
	if (!skb)
604 605 606 607 608 609 610 611 612 613 614 615 616 617 618 619 620
		goto errout;

	err = inet6_netconf_fill_devconf(skb, ifindex, devconf,
					 NETLINK_CB(in_skb).portid,
					 nlh->nlmsg_seq, RTM_NEWNETCONF, 0,
					 -1);
	if (err < 0) {
		/* -EMSGSIZE implies BUG in inet6_netconf_msgsize_devconf() */
		WARN_ON(err == -EMSGSIZE);
		kfree_skb(skb);
		goto errout;
	}
	err = rtnl_unicast(skb, net, NETLINK_CB(in_skb).portid);
errout:
	return err;
}

621 622 623 624 625 626 627 628 629 630 631 632 633 634 635 636 637
static int inet6_netconf_dump_devconf(struct sk_buff *skb,
				      struct netlink_callback *cb)
{
	struct net *net = sock_net(skb->sk);
	int h, s_h;
	int idx, s_idx;
	struct net_device *dev;
	struct inet6_dev *idev;
	struct hlist_head *head;

	s_h = cb->args[0];
	s_idx = idx = cb->args[1];

	for (h = s_h; h < NETDEV_HASHENTRIES; h++, s_idx = 0) {
		idx = 0;
		head = &net->dev_index_head[h];
		rcu_read_lock();
638 639
		cb->seq = atomic_read(&net->ipv6.dev_addr_genid) ^
			  net->dev_base_seq;
640 641 642 643 644 645 646 647 648 649 650 651 652
		hlist_for_each_entry_rcu(dev, head, index_hlist) {
			if (idx < s_idx)
				goto cont;
			idev = __in6_dev_get(dev);
			if (!idev)
				goto cont;

			if (inet6_netconf_fill_devconf(skb, dev->ifindex,
						       &idev->cnf,
						       NETLINK_CB(cb->skb).portid,
						       cb->nlh->nlmsg_seq,
						       RTM_NEWNETCONF,
						       NLM_F_MULTI,
653
						       -1) < 0) {
654 655 656
				rcu_read_unlock();
				goto done;
			}
657
			nl_dump_check_consistent(cb, nlmsg_hdr(skb));
658 659 660 661 662 663 664 665 666 667 668
cont:
			idx++;
		}
		rcu_read_unlock();
	}
	if (h == NETDEV_HASHENTRIES) {
		if (inet6_netconf_fill_devconf(skb, NETCONFA_IFINDEX_ALL,
					       net->ipv6.devconf_all,
					       NETLINK_CB(cb->skb).portid,
					       cb->nlh->nlmsg_seq,
					       RTM_NEWNETCONF, NLM_F_MULTI,
669
					       -1) < 0)
670 671 672 673 674 675 676 677 678 679
			goto done;
		else
			h++;
	}
	if (h == NETDEV_HASHENTRIES + 1) {
		if (inet6_netconf_fill_devconf(skb, NETCONFA_IFINDEX_DEFAULT,
					       net->ipv6.devconf_dflt,
					       NETLINK_CB(cb->skb).portid,
					       cb->nlh->nlmsg_seq,
					       RTM_NEWNETCONF, NLM_F_MULTI,
680
					       -1) < 0)
681 682 683 684 685 686 687 688 689 690 691
			goto done;
		else
			h++;
	}
done:
	cb->args[0] = h;
	cb->args[1] = idx;

	return skb->len;
}

L
Linus Torvalds 已提交
692 693 694 695 696 697 698 699 700
#ifdef CONFIG_SYSCTL
static void dev_forward_change(struct inet6_dev *idev)
{
	struct net_device *dev;
	struct inet6_ifaddr *ifa;

	if (!idev)
		return;
	dev = idev->dev;
701 702
	if (idev->cnf.forwarding)
		dev_disable_lro(dev);
A
Amerigo Wang 已提交
703
	if (dev->flags & IFF_MULTICAST) {
704
		if (idev->cnf.forwarding) {
705
			ipv6_dev_mc_inc(dev, &in6addr_linklocal_allrouters);
706 707 708
			ipv6_dev_mc_inc(dev, &in6addr_interfacelocal_allrouters);
			ipv6_dev_mc_inc(dev, &in6addr_sitelocal_allrouters);
		} else {
709
			ipv6_dev_mc_dec(dev, &in6addr_linklocal_allrouters);
710 711 712
			ipv6_dev_mc_dec(dev, &in6addr_interfacelocal_allrouters);
			ipv6_dev_mc_dec(dev, &in6addr_sitelocal_allrouters);
		}
L
Linus Torvalds 已提交
713
	}
714 715

	list_for_each_entry(ifa, &idev->addr_list, if_list) {
M
Michal Wrobel 已提交
716 717
		if (ifa->flags&IFA_F_TENTATIVE)
			continue;
L
Linus Torvalds 已提交
718 719 720 721 722
		if (idev->cnf.forwarding)
			addrconf_join_anycast(ifa);
		else
			addrconf_leave_anycast(ifa);
	}
723 724
	inet6_netconf_notify_devconf(dev_net(dev), NETCONFA_FORWARDING,
				     dev->ifindex, &idev->cnf);
L
Linus Torvalds 已提交
725 726 727
}


728
static void addrconf_forward_change(struct net *net, __s32 newf)
L
Linus Torvalds 已提交
729 730 731 732
{
	struct net_device *dev;
	struct inet6_dev *idev;

733
	for_each_netdev(net, dev) {
L
Linus Torvalds 已提交
734 735
		idev = __in6_dev_get(dev);
		if (idev) {
736 737
			int changed = (!idev->cnf.forwarding) ^ (!newf);
			idev->cnf.forwarding = newf;
L
Linus Torvalds 已提交
738 739 740 741 742
			if (changed)
				dev_forward_change(idev);
		}
	}
}
743

744
static int addrconf_fixup_forwarding(struct ctl_table *table, int *p, int newf)
745
{
746
	struct net *net;
747 748 749 750
	int old;

	if (!rtnl_trylock())
		return restart_syscall();
751 752

	net = (struct net *)table->extra2;
753 754
	old = *p;
	*p = newf;
755

756
	if (p == &net->ipv6.devconf_dflt->forwarding) {
757 758 759 760
		if ((!newf) ^ (!old))
			inet6_netconf_notify_devconf(net, NETCONFA_FORWARDING,
						     NETCONFA_IFINDEX_DEFAULT,
						     net->ipv6.devconf_dflt);
761 762
		rtnl_unlock();
		return 0;
E
Eric W. Biederman 已提交
763
	}
764

765 766 767
	if (p == &net->ipv6.devconf_all->forwarding) {
		net->ipv6.devconf_dflt->forwarding = newf;
		addrconf_forward_change(net, newf);
768 769 770 771
		if ((!newf) ^ (!old))
			inet6_netconf_notify_devconf(net, NETCONFA_FORWARDING,
						     NETCONFA_IFINDEX_ALL,
						     net->ipv6.devconf_all);
772
	} else if ((!newf) ^ (!old))
773
		dev_forward_change((struct inet6_dev *)table->extra1);
774
	rtnl_unlock();
775

776
	if (newf)
777
		rt6_purge_dflt_routers(net);
778
	return 1;
779
}
780 781 782 783 784 785 786 787 788 789 790 791 792 793 794 795 796 797 798 799 800 801 802 803 804 805 806 807 808 809 810 811 812 813 814 815 816 817 818 819 820 821 822 823 824 825 826 827 828 829 830 831 832 833 834 835 836

static void addrconf_linkdown_change(struct net *net, __s32 newf)
{
	struct net_device *dev;
	struct inet6_dev *idev;

	for_each_netdev(net, dev) {
		idev = __in6_dev_get(dev);
		if (idev) {
			int changed = (!idev->cnf.ignore_routes_with_linkdown) ^ (!newf);

			idev->cnf.ignore_routes_with_linkdown = newf;
			if (changed)
				inet6_netconf_notify_devconf(dev_net(dev),
							     NETCONFA_IGNORE_ROUTES_WITH_LINKDOWN,
							     dev->ifindex,
							     &idev->cnf);
		}
	}
}

static int addrconf_fixup_linkdown(struct ctl_table *table, int *p, int newf)
{
	struct net *net;
	int old;

	if (!rtnl_trylock())
		return restart_syscall();

	net = (struct net *)table->extra2;
	old = *p;
	*p = newf;

	if (p == &net->ipv6.devconf_dflt->ignore_routes_with_linkdown) {
		if ((!newf) ^ (!old))
			inet6_netconf_notify_devconf(net,
						     NETCONFA_IGNORE_ROUTES_WITH_LINKDOWN,
						     NETCONFA_IFINDEX_DEFAULT,
						     net->ipv6.devconf_dflt);
		rtnl_unlock();
		return 0;
	}

	if (p == &net->ipv6.devconf_all->ignore_routes_with_linkdown) {
		net->ipv6.devconf_dflt->ignore_routes_with_linkdown = newf;
		addrconf_linkdown_change(net, newf);
		if ((!newf) ^ (!old))
			inet6_netconf_notify_devconf(net,
						     NETCONFA_IGNORE_ROUTES_WITH_LINKDOWN,
						     NETCONFA_IFINDEX_ALL,
						     net->ipv6.devconf_all);
	}
	rtnl_unlock();

	return 1;
}

L
Linus Torvalds 已提交
837 838
#endif

839
/* Nobody refers to this ifaddr, destroy it */
L
Linus Torvalds 已提交
840 841
void inet6_ifa_finish_destroy(struct inet6_ifaddr *ifp)
{
842
	WARN_ON(!hlist_unhashed(&ifp->addr_lst));
843

L
Linus Torvalds 已提交
844
#ifdef NET_REFCNT_DEBUG
845
	pr_debug("%s\n", __func__);
L
Linus Torvalds 已提交
846 847 848 849
#endif

	in6_dev_put(ifp->idev);

850 851 852
	if (cancel_delayed_work(&ifp->dad_work))
		pr_notice("delayed DAD work was pending while freeing ifa=%p\n",
			  ifp);
L
Linus Torvalds 已提交
853

854
	if (ifp->state != INET6_IFADDR_STATE_DEAD) {
855
		pr_warn("Freeing alive inet6 address %p\n", ifp);
L
Linus Torvalds 已提交
856 857
		return;
	}
A
Amerigo Wang 已提交
858
	ip6_rt_put(ifp->rt);
L
Linus Torvalds 已提交
859

860
	kfree_rcu(ifp, rcu);
L
Linus Torvalds 已提交
861 862
}

B
Brian Haley 已提交
863 864 865
static void
ipv6_link_dev_addr(struct inet6_dev *idev, struct inet6_ifaddr *ifp)
{
866
	struct list_head *p;
867
	int ifp_scope = ipv6_addr_src_scope(&ifp->addr);
B
Brian Haley 已提交
868 869 870 871 872

	/*
	 * Each device address list is sorted in order of scope -
	 * global before linklocal.
	 */
873 874 875
	list_for_each(p, &idev->addr_list) {
		struct inet6_ifaddr *ifa
			= list_entry(p, struct inet6_ifaddr, if_list);
876
		if (ifp_scope >= ipv6_addr_src_scope(&ifa->addr))
B
Brian Haley 已提交
877 878 879
			break;
	}

880
	list_add_tail(&ifp->if_list, p);
B
Brian Haley 已提交
881 882
}

E
Eric Dumazet 已提交
883
static u32 inet6_addr_hash(const struct in6_addr *addr)
884
{
E
Eric Dumazet 已提交
885
	return hash_32(ipv6_addr_hash(addr), IN6_ADDR_HSIZE_SHIFT);
886 887
}

L
Linus Torvalds 已提交
888 889 890
/* On success it returns ifp with increased reference count */

static struct inet6_ifaddr *
891 892
ipv6_add_addr(struct inet6_dev *idev, const struct in6_addr *addr,
	      const struct in6_addr *peer_addr, int pfxlen,
893
	      int scope, u32 flags, u32 valid_lft, u32 prefered_lft)
L
Linus Torvalds 已提交
894 895 896
{
	struct inet6_ifaddr *ifa = NULL;
	struct rt6_info *rt;
897
	unsigned int hash;
L
Linus Torvalds 已提交
898
	int err = 0;
899 900 901 902 903 904 905
	int addr_type = ipv6_addr_type(addr);

	if (addr_type == IPV6_ADDR_ANY ||
	    addr_type & IPV6_ADDR_MULTICAST ||
	    (!(idev->dev->flags & IFF_LOOPBACK) &&
	     addr_type & IPV6_ADDR_LOOPBACK))
		return ERR_PTR(-EADDRNOTAVAIL);
L
Linus Torvalds 已提交
906

907
	rcu_read_lock_bh();
L
Linus Torvalds 已提交
908 909 910 911 912
	if (idev->dead) {
		err = -ENODEV;			/*XXX*/
		goto out2;
	}

913
	if (idev->cnf.disable_ipv6) {
914 915 916 917
		err = -EACCES;
		goto out2;
	}

918
	spin_lock(&addrconf_hash_lock);
L
Linus Torvalds 已提交
919 920

	/* Ignore adding duplicate addresses on an interface */
921
	if (ipv6_chk_same_addr(dev_net(idev->dev), addr, idev->dev)) {
922
		ADBG("ipv6_add_addr: already assigned\n");
L
Linus Torvalds 已提交
923 924 925 926
		err = -EEXIST;
		goto out;
	}

927
	ifa = kzalloc(sizeof(struct inet6_ifaddr), GFP_ATOMIC);
L
Linus Torvalds 已提交
928

929
	if (!ifa) {
930
		ADBG("ipv6_add_addr: malloc failed\n");
L
Linus Torvalds 已提交
931 932 933 934
		err = -ENOBUFS;
		goto out;
	}

935
	rt = addrconf_dst_alloc(idev, addr, false);
L
Linus Torvalds 已提交
936 937 938 939 940
	if (IS_ERR(rt)) {
		err = PTR_ERR(rt);
		goto out;
	}

941 942
	neigh_parms_data_state_setall(idev->nd_parms);

A
Alexey Dobriyan 已提交
943
	ifa->addr = *addr;
944 945
	if (peer_addr)
		ifa->peer_addr = *peer_addr;
L
Linus Torvalds 已提交
946 947

	spin_lock_init(&ifa->lock);
948
	INIT_DELAYED_WORK(&ifa->dad_work, addrconf_dad_work);
949
	INIT_HLIST_NODE(&ifa->addr_lst);
L
Linus Torvalds 已提交
950 951 952
	ifa->scope = scope;
	ifa->prefix_len = pfxlen;
	ifa->flags = flags | IFA_F_TENTATIVE;
953 954
	ifa->valid_lft = valid_lft;
	ifa->prefered_lft = prefered_lft;
L
Linus Torvalds 已提交
955
	ifa->cstamp = ifa->tstamp = jiffies;
956
	ifa->tokenized = false;
L
Linus Torvalds 已提交
957

958 959
	ifa->rt = rt;

L
Linus Torvalds 已提交
960 961 962 963 964 965
	ifa->idev = idev;
	in6_dev_hold(idev);
	/* For caller */
	in6_ifa_hold(ifa);

	/* Add to big hash table */
E
Eric Dumazet 已提交
966
	hash = inet6_addr_hash(addr);
L
Linus Torvalds 已提交
967

968 969
	hlist_add_head_rcu(&ifa->addr_lst, &inet6_addr_lst[hash]);
	spin_unlock(&addrconf_hash_lock);
L
Linus Torvalds 已提交
970 971 972

	write_lock(&idev->lock);
	/* Add to inet6_dev unicast addr list. */
B
Brian Haley 已提交
973
	ipv6_link_dev_addr(idev, ifa);
L
Linus Torvalds 已提交
974 975

	if (ifa->flags&IFA_F_TEMPORARY) {
976
		list_add(&ifa->tmp_list, &idev->tempaddr_list);
L
Linus Torvalds 已提交
977 978 979 980 981 982
		in6_ifa_hold(ifa);
	}

	in6_ifa_hold(ifa);
	write_unlock(&idev->lock);
out2:
983
	rcu_read_unlock_bh();
L
Linus Torvalds 已提交
984

985
	if (likely(err == 0))
986
		inet6addr_notifier_call_chain(NETDEV_UP, ifa);
L
Linus Torvalds 已提交
987 988 989 990 991 992 993
	else {
		kfree(ifa);
		ifa = ERR_PTR(err);
	}

	return ifa;
out:
994
	spin_unlock(&addrconf_hash_lock);
L
Linus Torvalds 已提交
995 996 997
	goto out2;
}

998 999 1000 1001 1002 1003 1004 1005 1006 1007 1008 1009 1010 1011 1012 1013 1014 1015 1016 1017 1018 1019 1020 1021 1022 1023 1024 1025 1026 1027 1028 1029 1030 1031 1032 1033 1034 1035 1036 1037 1038 1039 1040 1041 1042 1043 1044 1045 1046 1047 1048 1049 1050 1051 1052 1053 1054 1055 1056 1057 1058 1059 1060 1061 1062 1063 1064 1065 1066 1067 1068 1069 1070 1071 1072 1073 1074 1075 1076 1077 1078 1079
enum cleanup_prefix_rt_t {
	CLEANUP_PREFIX_RT_NOP,    /* no cleanup action for prefix route */
	CLEANUP_PREFIX_RT_DEL,    /* delete the prefix route */
	CLEANUP_PREFIX_RT_EXPIRE, /* update the lifetime of the prefix route */
};

/*
 * Check, whether the prefix for ifp would still need a prefix route
 * after deleting ifp. The function returns one of the CLEANUP_PREFIX_RT_*
 * constants.
 *
 * 1) we don't purge prefix if address was not permanent.
 *    prefix is managed by its own lifetime.
 * 2) we also don't purge, if the address was IFA_F_NOPREFIXROUTE.
 * 3) if there are no addresses, delete prefix.
 * 4) if there are still other permanent address(es),
 *    corresponding prefix is still permanent.
 * 5) if there are still other addresses with IFA_F_NOPREFIXROUTE,
 *    don't purge the prefix, assume user space is managing it.
 * 6) otherwise, update prefix lifetime to the
 *    longest valid lifetime among the corresponding
 *    addresses on the device.
 *    Note: subsequent RA will update lifetime.
 **/
static enum cleanup_prefix_rt_t
check_cleanup_prefix_route(struct inet6_ifaddr *ifp, unsigned long *expires)
{
	struct inet6_ifaddr *ifa;
	struct inet6_dev *idev = ifp->idev;
	unsigned long lifetime;
	enum cleanup_prefix_rt_t action = CLEANUP_PREFIX_RT_DEL;

	*expires = jiffies;

	list_for_each_entry(ifa, &idev->addr_list, if_list) {
		if (ifa == ifp)
			continue;
		if (!ipv6_prefix_equal(&ifa->addr, &ifp->addr,
				       ifp->prefix_len))
			continue;
		if (ifa->flags & (IFA_F_PERMANENT | IFA_F_NOPREFIXROUTE))
			return CLEANUP_PREFIX_RT_NOP;

		action = CLEANUP_PREFIX_RT_EXPIRE;

		spin_lock(&ifa->lock);

		lifetime = addrconf_timeout_fixup(ifa->valid_lft, HZ);
		/*
		 * Note: Because this address is
		 * not permanent, lifetime <
		 * LONG_MAX / HZ here.
		 */
		if (time_before(*expires, ifa->tstamp + lifetime * HZ))
			*expires = ifa->tstamp + lifetime * HZ;
		spin_unlock(&ifa->lock);
	}

	return action;
}

static void
cleanup_prefix_route(struct inet6_ifaddr *ifp, unsigned long expires, bool del_rt)
{
	struct rt6_info *rt;

	rt = addrconf_get_prefix_route(&ifp->addr,
				       ifp->prefix_len,
				       ifp->idev->dev,
				       0, RTF_GATEWAY | RTF_DEFAULT);
	if (rt) {
		if (del_rt)
			ip6_del_rt(rt);
		else {
			if (!(rt->rt6i_flags & RTF_EXPIRES))
				rt6_set_expires(rt, expires);
			ip6_rt_put(rt);
		}
	}
}


L
Linus Torvalds 已提交
1080 1081 1082 1083
/* This function wants to get referenced ifp and releases it before return */

static void ipv6_del_addr(struct inet6_ifaddr *ifp)
{
1084
	int state;
1085 1086
	enum cleanup_prefix_rt_t action = CLEANUP_PREFIX_RT_NOP;
	unsigned long expires;
L
Linus Torvalds 已提交
1087

1088 1089
	ASSERT_RTNL();

1090
	spin_lock_bh(&ifp->lock);
1091
	state = ifp->state;
1092
	ifp->state = INET6_IFADDR_STATE_DEAD;
1093
	spin_unlock_bh(&ifp->lock);
1094 1095 1096

	if (state == INET6_IFADDR_STATE_DEAD)
		goto out;
L
Linus Torvalds 已提交
1097

1098 1099 1100
	spin_lock_bh(&addrconf_hash_lock);
	hlist_del_init_rcu(&ifp->addr_lst);
	spin_unlock_bh(&addrconf_hash_lock);
L
Linus Torvalds 已提交
1101

1102
	write_lock_bh(&ifp->idev->lock);
1103

L
Linus Torvalds 已提交
1104
	if (ifp->flags&IFA_F_TEMPORARY) {
1105 1106 1107 1108
		list_del(&ifp->tmp_list);
		if (ifp->ifpub) {
			in6_ifa_put(ifp->ifpub);
			ifp->ifpub = NULL;
L
Linus Torvalds 已提交
1109
		}
1110
		__in6_ifa_put(ifp);
L
Linus Torvalds 已提交
1111 1112
	}

1113 1114
	if (ifp->flags & IFA_F_PERMANENT && !(ifp->flags & IFA_F_NOPREFIXROUTE))
		action = check_cleanup_prefix_route(ifp, &expires);
1115

1116 1117 1118 1119
	list_del_init(&ifp->if_list);
	__in6_ifa_put(ifp);

	write_unlock_bh(&ifp->idev->lock);
L
Linus Torvalds 已提交
1120

1121
	addrconf_del_dad_work(ifp);
1122

L
Linus Torvalds 已提交
1123 1124
	ipv6_ifa_notify(RTM_DELADDR, ifp);

1125
	inet6addr_notifier_call_chain(NETDEV_DOWN, ifp);
L
Linus Torvalds 已提交
1126

1127 1128 1129
	if (action != CLEANUP_PREFIX_RT_NOP) {
		cleanup_prefix_route(ifp, expires,
			action == CLEANUP_PREFIX_RT_DEL);
L
Linus Torvalds 已提交
1130 1131
	}

1132 1133
	/* clean up prefsrc entries */
	rt6_remove_prefsrc(ifp);
1134
out:
L
Linus Torvalds 已提交
1135 1136 1137 1138 1139 1140 1141
	in6_ifa_put(ifp);
}

static int ipv6_create_tempaddr(struct inet6_ifaddr *ifp, struct inet6_ifaddr *ift)
{
	struct inet6_dev *idev = ifp->idev;
	struct in6_addr addr, *tmpaddr;
1142
	unsigned long tmp_prefered_lft, tmp_valid_lft, tmp_tstamp, age;
1143
	unsigned long regen_advance;
L
Linus Torvalds 已提交
1144 1145
	int tmp_plen;
	int ret = 0;
1146
	u32 addr_flags;
1147
	unsigned long now = jiffies;
L
Linus Torvalds 已提交
1148

1149
	write_lock_bh(&idev->lock);
L
Linus Torvalds 已提交
1150 1151 1152 1153 1154 1155 1156 1157 1158 1159 1160
	if (ift) {
		spin_lock_bh(&ift->lock);
		memcpy(&addr.s6_addr[8], &ift->addr.s6_addr[8], 8);
		spin_unlock_bh(&ift->lock);
		tmpaddr = &addr;
	} else {
		tmpaddr = NULL;
	}
retry:
	in6_dev_hold(idev);
	if (idev->cnf.use_tempaddr <= 0) {
1161
		write_unlock_bh(&idev->lock);
1162
		pr_info("%s: use_tempaddr is disabled\n", __func__);
L
Linus Torvalds 已提交
1163 1164 1165 1166 1167 1168 1169 1170
		in6_dev_put(idev);
		ret = -1;
		goto out;
	}
	spin_lock_bh(&ifp->lock);
	if (ifp->regen_count++ >= idev->cnf.regen_max_retry) {
		idev->cnf.use_tempaddr = -1;	/*XXX*/
		spin_unlock_bh(&ifp->lock);
1171
		write_unlock_bh(&idev->lock);
1172 1173
		pr_warn("%s: regeneration time exceeded - disabled temporary address support\n",
			__func__);
L
Linus Torvalds 已提交
1174 1175 1176 1177 1178 1179
		in6_dev_put(idev);
		ret = -1;
		goto out;
	}
	in6_ifa_hold(ifp);
	memcpy(addr.s6_addr, ifp->addr.s6_addr, 8);
S
Sorin Dumitru 已提交
1180
	__ipv6_try_regen_rndid(idev, tmpaddr);
L
Linus Torvalds 已提交
1181
	memcpy(&addr.s6_addr[8], idev->rndid, 8);
1182
	age = (now - ifp->tstamp) / HZ;
L
Linus Torvalds 已提交
1183 1184
	tmp_valid_lft = min_t(__u32,
			      ifp->valid_lft,
1185
			      idev->cnf.temp_valid_lft + age);
1186 1187
	tmp_prefered_lft = min_t(__u32,
				 ifp->prefered_lft,
1188
				 idev->cnf.temp_prefered_lft + age -
1189
				 idev->cnf.max_desync_factor);
L
Linus Torvalds 已提交
1190 1191 1192 1193
	tmp_plen = ifp->prefix_len;
	tmp_tstamp = ifp->tstamp;
	spin_unlock_bh(&ifp->lock);

1194
	regen_advance = idev->cnf.regen_max_retry *
1195 1196
			idev->cnf.dad_transmits *
			NEIGH_VAR(idev->nd_parms, RETRANS_TIME) / HZ;
1197
	write_unlock_bh(&idev->lock);
1198

1199 1200 1201 1202
	/* A temporary address is created only if this calculated Preferred
	 * Lifetime is greater than REGEN_ADVANCE time units.  In particular,
	 * an implementation must not create a temporary address with a zero
	 * Preferred Lifetime.
1203 1204
	 * Use age calculation as in addrconf_verify to avoid unnecessary
	 * temporary addresses being generated.
1205
	 */
1206 1207
	age = (now - tmp_tstamp + ADDRCONF_TIMER_FUZZ_MINUS) / HZ;
	if (tmp_prefered_lft <= regen_advance + age) {
1208 1209 1210 1211 1212 1213
		in6_ifa_put(ifp);
		in6_dev_put(idev);
		ret = -1;
		goto out;
	}

1214 1215 1216 1217 1218
	addr_flags = IFA_F_TEMPORARY;
	/* set in addrconf_prefix_rcv() */
	if (ifp->flags & IFA_F_OPTIMISTIC)
		addr_flags |= IFA_F_OPTIMISTIC;

1219 1220 1221 1222
	ift = ipv6_add_addr(idev, &addr, NULL, tmp_plen,
			    ipv6_addr_scope(&addr), addr_flags,
			    tmp_valid_lft, tmp_prefered_lft);
	if (IS_ERR(ift)) {
L
Linus Torvalds 已提交
1223 1224
		in6_ifa_put(ifp);
		in6_dev_put(idev);
1225
		pr_info("%s: retry temporary address regeneration\n", __func__);
L
Linus Torvalds 已提交
1226
		tmpaddr = &addr;
1227
		write_lock_bh(&idev->lock);
L
Linus Torvalds 已提交
1228 1229 1230 1231 1232
		goto retry;
	}

	spin_lock_bh(&ift->lock);
	ift->ifpub = ifp;
1233
	ift->cstamp = now;
L
Linus Torvalds 已提交
1234 1235 1236
	ift->tstamp = tmp_tstamp;
	spin_unlock_bh(&ift->lock);

1237
	addrconf_dad_start(ift);
L
Linus Torvalds 已提交
1238 1239 1240 1241 1242 1243 1244
	in6_ifa_put(ift);
	in6_dev_put(idev);
out:
	return ret;
}

/*
1245
 *	Choose an appropriate source address (RFC3484)
L
Linus Torvalds 已提交
1246
 */
1247 1248 1249 1250 1251 1252 1253 1254 1255 1256 1257 1258 1259
enum {
	IPV6_SADDR_RULE_INIT = 0,
	IPV6_SADDR_RULE_LOCAL,
	IPV6_SADDR_RULE_SCOPE,
	IPV6_SADDR_RULE_PREFERRED,
#ifdef CONFIG_IPV6_MIP6
	IPV6_SADDR_RULE_HOA,
#endif
	IPV6_SADDR_RULE_OIF,
	IPV6_SADDR_RULE_LABEL,
	IPV6_SADDR_RULE_PRIVACY,
	IPV6_SADDR_RULE_ORCHID,
	IPV6_SADDR_RULE_PREFIX,
1260 1261 1262
#ifdef CONFIG_IPV6_OPTIMISTIC_DAD
	IPV6_SADDR_RULE_NOT_OPTIMISTIC,
#endif
1263 1264 1265
	IPV6_SADDR_RULE_MAX
};

1266
struct ipv6_saddr_score {
1267 1268 1269 1270 1271 1272
	int			rule;
	int			addr_type;
	struct inet6_ifaddr	*ifa;
	DECLARE_BITMAP(scorebits, IPV6_SADDR_RULE_MAX);
	int			scopedist;
	int			matchlen;
1273 1274
};

1275
struct ipv6_saddr_dst {
1276
	const struct in6_addr *addr;
1277 1278 1279
	int ifindex;
	int scope;
	int label;
1280
	unsigned int prefs;
1281
};
1282

D
Dave Jones 已提交
1283
static inline int ipv6_saddr_preferred(int type)
L
Linus Torvalds 已提交
1284
{
U
Ulrich Weber 已提交
1285
	if (type & (IPV6_ADDR_MAPPED|IPV6_ADDR_COMPATv4|IPV6_ADDR_LOOPBACK))
1286 1287
		return 1;
	return 0;
L
Linus Torvalds 已提交
1288 1289
}

1290 1291 1292 1293 1294 1295 1296 1297 1298
static inline bool ipv6_use_optimistic_addr(struct inet6_dev *idev)
{
#ifdef CONFIG_IPV6_OPTIMISTIC_DAD
	return idev && idev->cnf.optimistic_dad && idev->cnf.use_optimistic;
#else
	return false;
#endif
}

1299 1300
static int ipv6_get_saddr_eval(struct net *net,
			       struct ipv6_saddr_score *score,
1301 1302 1303 1304 1305 1306 1307 1308 1309 1310 1311 1312 1313 1314 1315 1316 1317 1318 1319 1320 1321 1322 1323 1324 1325 1326 1327 1328 1329 1330 1331 1332 1333 1334 1335 1336 1337 1338 1339
			       struct ipv6_saddr_dst *dst,
			       int i)
{
	int ret;

	if (i <= score->rule) {
		switch (i) {
		case IPV6_SADDR_RULE_SCOPE:
			ret = score->scopedist;
			break;
		case IPV6_SADDR_RULE_PREFIX:
			ret = score->matchlen;
			break;
		default:
			ret = !!test_bit(i, score->scorebits);
		}
		goto out;
	}

	switch (i) {
	case IPV6_SADDR_RULE_INIT:
		/* Rule 0: remember if hiscore is not ready yet */
		ret = !!score->ifa;
		break;
	case IPV6_SADDR_RULE_LOCAL:
		/* Rule 1: Prefer same address */
		ret = ipv6_addr_equal(&score->ifa->addr, dst->addr);
		break;
	case IPV6_SADDR_RULE_SCOPE:
		/* Rule 2: Prefer appropriate scope
		 *
		 *      ret
		 *       ^
		 *    -1 |  d 15
		 *    ---+--+-+---> scope
		 *       |
		 *       |             d is scope of the destination.
		 *  B-d  |  \
		 *       |   \      <- smaller scope is better if
1340
		 *  B-15 |    \        if scope is enough for destination.
1341 1342 1343 1344 1345 1346 1347 1348 1349 1350 1351 1352 1353 1354 1355 1356 1357 1358
		 *       |             ret = B - scope (-1 <= scope >= d <= 15).
		 * d-C-1 | /
		 *       |/         <- greater is better
		 *   -C  /             if scope is not enough for destination.
		 *      /|             ret = scope - C (-1 <= d < scope <= 15).
		 *
		 * d - C - 1 < B -15 (for all -1 <= d <= 15).
		 * C > d + 14 - B >= 15 + 14 - B = 29 - B.
		 * Assume B = 0 and we get C > 29.
		 */
		ret = __ipv6_addr_src_scope(score->addr_type);
		if (ret >= dst->scope)
			ret = -ret;
		else
			ret -= 128;	/* 30 is enough */
		score->scopedist = ret;
		break;
	case IPV6_SADDR_RULE_PREFERRED:
1359
	    {
1360
		/* Rule 3: Avoid deprecated and optimistic addresses */
1361 1362 1363 1364
		u8 avoid = IFA_F_DEPRECATED;

		if (!ipv6_use_optimistic_addr(score->ifa->idev))
			avoid |= IFA_F_OPTIMISTIC;
1365
		ret = ipv6_saddr_preferred(score->addr_type) ||
1366
		      !(score->ifa->flags & avoid);
1367
		break;
1368
	    }
1369 1370
#ifdef CONFIG_IPV6_MIP6
	case IPV6_SADDR_RULE_HOA:
1371
	    {
1372
		/* Rule 4: Prefer home address */
1373 1374
		int prefhome = !(dst->prefs & IPV6_PREFER_SRC_COA);
		ret = !(score->ifa->flags & IFA_F_HOMEADDRESS) ^ prefhome;
1375
		break;
1376
	    }
1377 1378 1379 1380 1381 1382 1383 1384
#endif
	case IPV6_SADDR_RULE_OIF:
		/* Rule 5: Prefer outgoing interface */
		ret = (!dst->ifindex ||
		       dst->ifindex == score->ifa->idev->dev->ifindex);
		break;
	case IPV6_SADDR_RULE_LABEL:
		/* Rule 6: Prefer matching label */
1385 1386
		ret = ipv6_addr_label(net,
				      &score->ifa->addr, score->addr_type,
1387 1388 1389
				      score->ifa->idev->dev->ifindex) == dst->label;
		break;
	case IPV6_SADDR_RULE_PRIVACY:
1390
	    {
1391
		/* Rule 7: Prefer public address
L
Lucas De Marchi 已提交
1392
		 * Note: prefer temporary address if use_tempaddr >= 2
1393
		 */
1394 1395 1396 1397
		int preftmp = dst->prefs & (IPV6_PREFER_SRC_PUBLIC|IPV6_PREFER_SRC_TMP) ?
				!!(dst->prefs & IPV6_PREFER_SRC_TMP) :
				score->ifa->idev->cnf.use_tempaddr >= 2;
		ret = (!(score->ifa->flags & IFA_F_TEMPORARY)) ^ preftmp;
1398
		break;
1399
	    }
1400 1401 1402 1403 1404 1405 1406 1407 1408
	case IPV6_SADDR_RULE_ORCHID:
		/* Rule 8-: Prefer ORCHID vs ORCHID or
		 *	    non-ORCHID vs non-ORCHID
		 */
		ret = !(ipv6_addr_orchid(&score->ifa->addr) ^
			ipv6_addr_orchid(dst->addr));
		break;
	case IPV6_SADDR_RULE_PREFIX:
		/* Rule 8: Use longest matching prefix */
1409 1410 1411 1412
		ret = ipv6_addr_diff(&score->ifa->addr, dst->addr);
		if (ret > score->ifa->prefix_len)
			ret = score->ifa->prefix_len;
		score->matchlen = ret;
1413
		break;
1414 1415 1416 1417 1418 1419 1420 1421
#ifdef CONFIG_IPV6_OPTIMISTIC_DAD
	case IPV6_SADDR_RULE_NOT_OPTIMISTIC:
		/* Optimistic addresses still have lower precedence than other
		 * preferred addresses.
		 */
		ret = !(score->ifa->flags & IFA_F_OPTIMISTIC);
		break;
#endif
1422 1423 1424 1425 1426 1427 1428 1429 1430 1431 1432
	default:
		ret = 0;
	}

	if (ret)
		__set_bit(i, score->scorebits);
	score->rule = i;
out:
	return ret;
}

1433 1434 1435 1436 1437
static int __ipv6_dev_get_saddr(struct net *net,
				struct ipv6_saddr_dst *dst,
				struct inet6_dev *idev,
				struct ipv6_saddr_score *scores,
				int hiscore_idx)
1438
{
1439
	struct ipv6_saddr_score *score = &scores[1 - hiscore_idx], *hiscore = &scores[hiscore_idx];
1440 1441 1442 1443 1444 1445 1446 1447 1448 1449 1450 1451 1452 1453 1454 1455 1456 1457 1458 1459 1460 1461 1462 1463 1464 1465 1466 1467 1468 1469 1470 1471 1472 1473 1474 1475 1476 1477 1478 1479 1480 1481 1482 1483 1484 1485 1486 1487 1488 1489 1490 1491 1492 1493 1494 1495 1496 1497

	read_lock_bh(&idev->lock);
	list_for_each_entry(score->ifa, &idev->addr_list, if_list) {
		int i;

		/*
		 * - Tentative Address (RFC2462 section 5.4)
		 *  - A tentative address is not considered
		 *    "assigned to an interface" in the traditional
		 *    sense, unless it is also flagged as optimistic.
		 * - Candidate Source Address (section 4)
		 *  - In any case, anycast addresses, multicast
		 *    addresses, and the unspecified address MUST
		 *    NOT be included in a candidate set.
		 */
		if ((score->ifa->flags & IFA_F_TENTATIVE) &&
		    (!(score->ifa->flags & IFA_F_OPTIMISTIC)))
			continue;

		score->addr_type = __ipv6_addr_type(&score->ifa->addr);

		if (unlikely(score->addr_type == IPV6_ADDR_ANY ||
			     score->addr_type & IPV6_ADDR_MULTICAST)) {
			net_dbg_ratelimited("ADDRCONF: unspecified / multicast address assigned as unicast address on %s",
					    idev->dev->name);
			continue;
		}

		score->rule = -1;
		bitmap_zero(score->scorebits, IPV6_SADDR_RULE_MAX);

		for (i = 0; i < IPV6_SADDR_RULE_MAX; i++) {
			int minihiscore, miniscore;

			minihiscore = ipv6_get_saddr_eval(net, hiscore, dst, i);
			miniscore = ipv6_get_saddr_eval(net, score, dst, i);

			if (minihiscore > miniscore) {
				if (i == IPV6_SADDR_RULE_SCOPE &&
				    score->scopedist > 0) {
					/*
					 * special case:
					 * each remaining entry
					 * has too small (not enough)
					 * scope, because ifa entries
					 * are sorted by their scope
					 * values.
					 */
					goto out;
				}
				break;
			} else if (minihiscore < miniscore) {
				if (hiscore->ifa)
					in6_ifa_put(hiscore->ifa);

				in6_ifa_hold(score->ifa);

				swap(hiscore, score);
1498
				hiscore_idx = 1 - hiscore_idx;
1499 1500 1501 1502 1503 1504 1505 1506 1507 1508

				/* restore our iterator */
				score->ifa = hiscore->ifa;

				break;
			}
		}
	}
out:
	read_unlock_bh(&idev->lock);
1509
	return hiscore_idx;
1510 1511
}

1512
int ipv6_dev_get_saddr(struct net *net, const struct net_device *dst_dev,
1513
		       const struct in6_addr *daddr, unsigned int prefs,
1514
		       struct in6_addr *saddr)
L
Linus Torvalds 已提交
1515
{
1516
	struct ipv6_saddr_score scores[2], *hiscore;
1517
	struct ipv6_saddr_dst dst;
1518
	struct inet6_dev *idev;
1519
	struct net_device *dev;
1520
	int dst_type;
1521
	bool use_oif_addr = false;
1522
	int hiscore_idx = 0;
L
Linus Torvalds 已提交
1523

1524 1525 1526 1527
	dst_type = __ipv6_addr_type(daddr);
	dst.addr = daddr;
	dst.ifindex = dst_dev ? dst_dev->ifindex : 0;
	dst.scope = __ipv6_addr_src_scope(dst_type);
1528
	dst.label = ipv6_addr_label(net, daddr, dst_type, dst.ifindex);
1529
	dst.prefs = prefs;
1530

1531 1532
	scores[hiscore_idx].rule = -1;
	scores[hiscore_idx].ifa = NULL;
L
Linus Torvalds 已提交
1533

1534
	rcu_read_lock();
L
Linus Torvalds 已提交
1535

1536 1537 1538 1539 1540 1541 1542 1543 1544 1545 1546
	/* Candidate Source Address (section 4)
	 *  - multicast and link-local destination address,
	 *    the set of candidate source address MUST only
	 *    include addresses assigned to interfaces
	 *    belonging to the same link as the outgoing
	 *    interface.
	 * (- For site-local destination addresses, the
	 *    set of candidate source addresses MUST only
	 *    include addresses assigned to interfaces
	 *    belonging to the same site as the outgoing
	 *    interface.)
1547 1548 1549 1550
	 *  - "It is RECOMMENDED that the candidate source addresses
	 *    be the set of unicast addresses assigned to the
	 *    interface that will be used to send to the destination
	 *    (the 'outgoing' interface)." (RFC 6724)
1551 1552
	 */
	if (dst_dev) {
1553
		idev = __in6_dev_get(dst_dev);
1554
		if ((dst_type & IPV6_ADDR_MULTICAST) ||
1555 1556
		    dst.scope <= IPV6_ADDR_SCOPE_LINKLOCAL ||
		    (idev && idev->cnf.use_oif_addrs_only)) {
1557 1558 1559
			use_oif_addr = true;
		}
	}
1560

1561
	if (use_oif_addr) {
1562
		if (idev)
1563
			hiscore_idx = __ipv6_dev_get_saddr(net, &dst, idev, scores, hiscore_idx);
1564 1565 1566 1567
	} else {
		for_each_netdev_rcu(net, dev) {
			idev = __in6_dev_get(dev);
			if (!idev)
1568
				continue;
1569
			hiscore_idx = __ipv6_dev_get_saddr(net, &dst, idev, scores, hiscore_idx);
L
Linus Torvalds 已提交
1570 1571
		}
	}
1572
	rcu_read_unlock();
L
Linus Torvalds 已提交
1573

1574
	hiscore = &scores[hiscore_idx];
1575
	if (!hiscore->ifa)
1576
		return -EADDRNOTAVAIL;
1577

A
Alexey Dobriyan 已提交
1578
	*saddr = hiscore->ifa->addr;
1579
	in6_ifa_put(hiscore->ifa);
1580
	return 0;
L
Linus Torvalds 已提交
1581
}
1582
EXPORT_SYMBOL(ipv6_dev_get_saddr);
L
Linus Torvalds 已提交
1583

1584
int __ipv6_get_lladdr(struct inet6_dev *idev, struct in6_addr *addr,
1585
		      u32 banned_flags)
1586 1587 1588 1589
{
	struct inet6_ifaddr *ifp;
	int err = -EADDRNOTAVAIL;

1590 1591 1592
	list_for_each_entry_reverse(ifp, &idev->addr_list, if_list) {
		if (ifp->scope > IFA_LINK)
			break;
1593 1594 1595 1596 1597 1598 1599 1600 1601 1602
		if (ifp->scope == IFA_LINK &&
		    !(ifp->flags & banned_flags)) {
			*addr = ifp->addr;
			err = 0;
			break;
		}
	}
	return err;
}

1603
int ipv6_get_lladdr(struct net_device *dev, struct in6_addr *addr,
1604
		    u32 banned_flags)
L
Linus Torvalds 已提交
1605 1606 1607 1608
{
	struct inet6_dev *idev;
	int err = -EADDRNOTAVAIL;

1609
	rcu_read_lock();
S
Stephen Hemminger 已提交
1610 1611
	idev = __in6_dev_get(dev);
	if (idev) {
L
Linus Torvalds 已提交
1612
		read_lock_bh(&idev->lock);
1613
		err = __ipv6_get_lladdr(idev, addr, banned_flags);
L
Linus Torvalds 已提交
1614 1615
		read_unlock_bh(&idev->lock);
	}
1616
	rcu_read_unlock();
L
Linus Torvalds 已提交
1617 1618 1619 1620 1621 1622 1623 1624 1625
	return err;
}

static int ipv6_count_addresses(struct inet6_dev *idev)
{
	int cnt = 0;
	struct inet6_ifaddr *ifp;

	read_lock_bh(&idev->lock);
1626
	list_for_each_entry(ifp, &idev->addr_list, if_list)
L
Linus Torvalds 已提交
1627 1628 1629 1630 1631
		cnt++;
	read_unlock_bh(&idev->lock);
	return cnt;
}

1632
int ipv6_chk_addr(struct net *net, const struct in6_addr *addr,
1633
		  const struct net_device *dev, int strict)
1634 1635 1636 1637 1638 1639 1640 1641
{
	return ipv6_chk_addr_and_flags(net, addr, dev, strict, IFA_F_TENTATIVE);
}
EXPORT_SYMBOL(ipv6_chk_addr);

int ipv6_chk_addr_and_flags(struct net *net, const struct in6_addr *addr,
			    const struct net_device *dev, int strict,
			    u32 banned_flags)
L
Linus Torvalds 已提交
1642
{
1643
	struct inet6_ifaddr *ifp;
E
Eric Dumazet 已提交
1644
	unsigned int hash = inet6_addr_hash(addr);
1645
	u32 ifp_flags;
L
Linus Torvalds 已提交
1646

1647
	rcu_read_lock_bh();
1648
	hlist_for_each_entry_rcu(ifp, &inet6_addr_lst[hash], addr_lst) {
1649
		if (!net_eq(dev_net(ifp->idev->dev), net))
1650
			continue;
1651 1652 1653 1654 1655 1656
		/* Decouple optimistic from tentative for evaluation here.
		 * Ban optimistic addresses explicitly, when required.
		 */
		ifp_flags = (ifp->flags&IFA_F_OPTIMISTIC)
			    ? (ifp->flags&~IFA_F_TENTATIVE)
			    : ifp->flags;
L
Linus Torvalds 已提交
1657
		if (ipv6_addr_equal(&ifp->addr, addr) &&
1658
		    !(ifp_flags&banned_flags) &&
1659
		    (!dev || ifp->idev->dev == dev ||
1660 1661 1662
		     !(ifp->scope&(IFA_LINK|IFA_HOST) || strict))) {
			rcu_read_unlock_bh();
			return 1;
L
Linus Torvalds 已提交
1663 1664
		}
	}
1665

1666 1667
	rcu_read_unlock_bh();
	return 0;
L
Linus Torvalds 已提交
1668
}
1669
EXPORT_SYMBOL(ipv6_chk_addr_and_flags);
1670

1671 1672
static bool ipv6_chk_same_addr(struct net *net, const struct in6_addr *addr,
			       struct net_device *dev)
L
Linus Torvalds 已提交
1673
{
E
Eric Dumazet 已提交
1674
	unsigned int hash = inet6_addr_hash(addr);
1675
	struct inet6_ifaddr *ifp;
L
Linus Torvalds 已提交
1676

1677
	hlist_for_each_entry(ifp, &inet6_addr_lst[hash], addr_lst) {
1678
		if (!net_eq(dev_net(ifp->idev->dev), net))
1679
			continue;
L
Linus Torvalds 已提交
1680
		if (ipv6_addr_equal(&ifp->addr, addr)) {
1681
			if (!dev || ifp->idev->dev == dev)
1682
				return true;
L
Linus Torvalds 已提交
1683 1684
		}
	}
1685
	return false;
L
Linus Torvalds 已提交
1686 1687
}

1688 1689 1690 1691 1692 1693 1694 1695 1696 1697 1698 1699 1700 1701 1702 1703 1704 1705 1706 1707 1708 1709 1710 1711 1712 1713 1714
/* Compares an address/prefix_len with addresses on device @dev.
 * If one is found it returns true.
 */
bool ipv6_chk_custom_prefix(const struct in6_addr *addr,
	const unsigned int prefix_len, struct net_device *dev)
{
	struct inet6_dev *idev;
	struct inet6_ifaddr *ifa;
	bool ret = false;

	rcu_read_lock();
	idev = __in6_dev_get(dev);
	if (idev) {
		read_lock_bh(&idev->lock);
		list_for_each_entry(ifa, &idev->addr_list, if_list) {
			ret = ipv6_prefix_equal(addr, &ifa->addr, prefix_len);
			if (ret)
				break;
		}
		read_unlock_bh(&idev->lock);
	}
	rcu_read_unlock();

	return ret;
}
EXPORT_SYMBOL(ipv6_chk_custom_prefix);

1715
int ipv6_chk_prefix(const struct in6_addr *addr, struct net_device *dev)
1716 1717 1718 1719 1720 1721 1722 1723 1724 1725
{
	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);
1726
		list_for_each_entry(ifa, &idev->addr_list, if_list) {
1727 1728 1729 1730 1731 1732 1733 1734 1735 1736 1737 1738
			onlink = ipv6_prefix_equal(addr, &ifa->addr,
						   ifa->prefix_len);
			if (onlink)
				break;
		}
		read_unlock_bh(&idev->lock);
	}
	rcu_read_unlock();
	return onlink;
}
EXPORT_SYMBOL(ipv6_chk_prefix);

1739
struct inet6_ifaddr *ipv6_get_ifaddr(struct net *net, const struct in6_addr *addr,
1740
				     struct net_device *dev, int strict)
L
Linus Torvalds 已提交
1741
{
1742
	struct inet6_ifaddr *ifp, *result = NULL;
E
Eric Dumazet 已提交
1743
	unsigned int hash = inet6_addr_hash(addr);
L
Linus Torvalds 已提交
1744

1745
	rcu_read_lock_bh();
1746
	hlist_for_each_entry_rcu_bh(ifp, &inet6_addr_lst[hash], addr_lst) {
1747
		if (!net_eq(dev_net(ifp->idev->dev), net))
1748
			continue;
L
Linus Torvalds 已提交
1749
		if (ipv6_addr_equal(&ifp->addr, addr)) {
1750
			if (!dev || ifp->idev->dev == dev ||
L
Linus Torvalds 已提交
1751
			    !(ifp->scope&(IFA_LINK|IFA_HOST) || strict)) {
1752
				result = ifp;
L
Linus Torvalds 已提交
1753 1754 1755 1756 1757
				in6_ifa_hold(ifp);
				break;
			}
		}
	}
1758
	rcu_read_unlock_bh();
L
Linus Torvalds 已提交
1759

1760
	return result;
L
Linus Torvalds 已提交
1761 1762 1763 1764
}

/* Gets referenced address, destroys ifaddr */

B
Brian Haley 已提交
1765
static void addrconf_dad_stop(struct inet6_ifaddr *ifp, int dad_failed)
L
Linus Torvalds 已提交
1766 1767 1768
{
	if (ifp->flags&IFA_F_PERMANENT) {
		spin_lock_bh(&ifp->lock);
1769
		addrconf_del_dad_work(ifp);
L
Linus Torvalds 已提交
1770
		ifp->flags |= IFA_F_TENTATIVE;
B
Brian Haley 已提交
1771 1772
		if (dad_failed)
			ifp->flags |= IFA_F_DADFAILED;
L
Linus Torvalds 已提交
1773
		spin_unlock_bh(&ifp->lock);
1774 1775
		if (dad_failed)
			ipv6_ifa_notify(0, ifp);
L
Linus Torvalds 已提交
1776 1777 1778 1779 1780 1781 1782 1783 1784 1785 1786 1787 1788 1789
		in6_ifa_put(ifp);
	} else if (ifp->flags&IFA_F_TEMPORARY) {
		struct inet6_ifaddr *ifpub;
		spin_lock_bh(&ifp->lock);
		ifpub = ifp->ifpub;
		if (ifpub) {
			in6_ifa_hold(ifpub);
			spin_unlock_bh(&ifp->lock);
			ipv6_create_tempaddr(ifpub, ifp);
			in6_ifa_put(ifpub);
		} else {
			spin_unlock_bh(&ifp->lock);
		}
		ipv6_del_addr(ifp);
1790
	} else {
L
Linus Torvalds 已提交
1791
		ipv6_del_addr(ifp);
1792
	}
L
Linus Torvalds 已提交
1793 1794
}

H
Herbert Xu 已提交
1795 1796 1797 1798
static int addrconf_dad_end(struct inet6_ifaddr *ifp)
{
	int err = -ENOENT;

1799
	spin_lock_bh(&ifp->lock);
H
Herbert Xu 已提交
1800 1801 1802 1803
	if (ifp->state == INET6_IFADDR_STATE_DAD) {
		ifp->state = INET6_IFADDR_STATE_POSTDAD;
		err = 0;
	}
1804
	spin_unlock_bh(&ifp->lock);
H
Herbert Xu 已提交
1805 1806 1807 1808

	return err;
}

1809 1810
void addrconf_dad_failure(struct inet6_ifaddr *ifp)
{
1811
	struct in6_addr addr;
1812
	struct inet6_dev *idev = ifp->idev;
1813
	struct net *net = dev_net(ifp->idev->dev);
1814

1815 1816
	if (addrconf_dad_end(ifp)) {
		in6_ifa_put(ifp);
H
Herbert Xu 已提交
1817
		return;
1818
	}
H
Herbert Xu 已提交
1819

1820 1821
	net_info_ratelimited("%s: IPv6 duplicate address %pI6c detected!\n",
			     ifp->idev->dev->name, &ifp->addr);
1822

1823 1824 1825 1826 1827 1828 1829 1830 1831 1832 1833
	spin_lock_bh(&ifp->lock);

	if (ifp->flags & IFA_F_STABLE_PRIVACY) {
		int scope = ifp->scope;
		u32 flags = ifp->flags;
		struct in6_addr new_addr;
		struct inet6_ifaddr *ifp2;
		u32 valid_lft, preferred_lft;
		int pfxlen = ifp->prefix_len;
		int retries = ifp->stable_privacy_retry + 1;

1834
		if (retries > net->ipv6.sysctl.idgen_retries) {
1835 1836 1837 1838 1839 1840 1841 1842 1843 1844 1845 1846 1847 1848
			net_info_ratelimited("%s: privacy stable address generation failed because of DAD conflicts!\n",
					     ifp->idev->dev->name);
			goto errdad;
		}

		new_addr = ifp->addr;
		if (ipv6_generate_stable_address(&new_addr, retries,
						 idev))
			goto errdad;

		valid_lft = ifp->valid_lft;
		preferred_lft = ifp->prefered_lft;

		spin_unlock_bh(&ifp->lock);
1849

1850 1851 1852 1853 1854 1855 1856 1857 1858 1859 1860 1861 1862 1863 1864 1865 1866 1867 1868
		if (idev->cnf.max_addresses &&
		    ipv6_count_addresses(idev) >=
		    idev->cnf.max_addresses)
			goto lock_errdad;

		net_info_ratelimited("%s: generating new stable privacy address because of DAD conflict\n",
				     ifp->idev->dev->name);

		ifp2 = ipv6_add_addr(idev, &new_addr, NULL, pfxlen,
				     scope, flags, valid_lft,
				     preferred_lft);
		if (IS_ERR(ifp2))
			goto lock_errdad;

		spin_lock_bh(&ifp2->lock);
		ifp2->stable_privacy_retry = retries;
		ifp2->state = INET6_IFADDR_STATE_PREDAD;
		spin_unlock_bh(&ifp2->lock);

1869
		addrconf_mod_dad_work(ifp2, net->ipv6.sysctl.idgen_delay);
1870 1871 1872 1873
		in6_ifa_put(ifp2);
lock_errdad:
		spin_lock_bh(&ifp->lock);
	} else if (idev->cnf.accept_dad > 1 && !idev->cnf.disable_ipv6) {
1874 1875 1876 1877 1878 1879 1880
		addr.s6_addr32[0] = htonl(0xfe800000);
		addr.s6_addr32[1] = 0;

		if (!ipv6_generate_eui64(addr.s6_addr + 8, idev->dev) &&
		    ipv6_addr_equal(&ifp->addr, &addr)) {
			/* DAD failed for link-local based on MAC address */
			idev->cnf.disable_ipv6 = 1;
1881

1882
			pr_info("%s: IPv6 being disabled!\n",
1883
				ifp->idev->dev->name);
1884 1885 1886
		}
	}

1887
errdad:
1888 1889
	/* transition from _POSTDAD to _ERRDAD */
	ifp->state = INET6_IFADDR_STATE_ERRDAD;
1890
	spin_unlock_bh(&ifp->lock);
1891 1892

	addrconf_mod_dad_work(ifp, 0);
1893
}
L
Linus Torvalds 已提交
1894

1895 1896
/* Join to solicited addr multicast group.
 * caller must hold RTNL */
1897
void addrconf_join_solict(struct net_device *dev, const struct in6_addr *addr)
L
Linus Torvalds 已提交
1898 1899 1900 1901 1902 1903 1904 1905 1906 1907
{
	struct in6_addr maddr;

	if (dev->flags&(IFF_LOOPBACK|IFF_NOARP))
		return;

	addrconf_addr_solict_mult(addr, &maddr);
	ipv6_dev_mc_inc(dev, &maddr);
}

1908
/* caller must hold RTNL */
1909
void addrconf_leave_solict(struct inet6_dev *idev, const struct in6_addr *addr)
L
Linus Torvalds 已提交
1910 1911 1912 1913 1914 1915 1916 1917 1918 1919
{
	struct in6_addr maddr;

	if (idev->dev->flags&(IFF_LOOPBACK|IFF_NOARP))
		return;

	addrconf_addr_solict_mult(addr, &maddr);
	__ipv6_dev_mc_dec(idev, &maddr);
}

1920
/* caller must hold RTNL */
1921
static void addrconf_join_anycast(struct inet6_ifaddr *ifp)
L
Linus Torvalds 已提交
1922 1923
{
	struct in6_addr addr;
1924

1925
	if (ifp->prefix_len >= 127) /* RFC 6164 */
1926
		return;
L
Linus Torvalds 已提交
1927 1928 1929
	ipv6_addr_prefix(&addr, &ifp->addr, ifp->prefix_len);
	if (ipv6_addr_any(&addr))
		return;
W
WANG Cong 已提交
1930
	__ipv6_dev_ac_inc(ifp->idev, &addr);
L
Linus Torvalds 已提交
1931 1932
}

1933
/* caller must hold RTNL */
1934
static void addrconf_leave_anycast(struct inet6_ifaddr *ifp)
L
Linus Torvalds 已提交
1935 1936
{
	struct in6_addr addr;
1937

1938
	if (ifp->prefix_len >= 127) /* RFC 6164 */
1939
		return;
L
Linus Torvalds 已提交
1940 1941 1942 1943 1944 1945
	ipv6_addr_prefix(&addr, &ifp->addr, ifp->prefix_len);
	if (ipv6_addr_any(&addr))
		return;
	__ipv6_dev_ac_dec(ifp->idev, &addr);
}

1946 1947 1948 1949 1950 1951 1952 1953 1954 1955 1956 1957 1958 1959 1960 1961 1962 1963 1964 1965 1966 1967 1968 1969 1970 1971 1972 1973 1974 1975 1976
static int addrconf_ifid_eui48(u8 *eui, struct net_device *dev)
{
	if (dev->addr_len != ETH_ALEN)
		return -1;
	memcpy(eui, dev->dev_addr, 3);
	memcpy(eui + 5, dev->dev_addr + 3, 3);

	/*
	 * The zSeries OSA network cards can be shared among various
	 * OS instances, but the OSA cards have only one MAC address.
	 * This leads to duplicate address conflicts in conjunction
	 * with IPv6 if more than one instance uses the same card.
	 *
	 * The driver for these cards can deliver a unique 16-bit
	 * identifier for each instance sharing the same card.  It is
	 * placed instead of 0xFFFE in the interface identifier.  The
	 * "u" bit of the interface identifier is not inverted in this
	 * case.  Hence the resulting interface identifier has local
	 * scope according to RFC2373.
	 */
	if (dev->dev_id) {
		eui[3] = (dev->dev_id >> 8) & 0xFF;
		eui[4] = dev->dev_id & 0xFF;
	} else {
		eui[3] = 0xFF;
		eui[4] = 0xFE;
		eui[0] ^= 2;
	}
	return 0;
}

1977 1978 1979 1980 1981
static int addrconf_ifid_eui64(u8 *eui, struct net_device *dev)
{
	if (dev->addr_len != IEEE802154_ADDR_LEN)
		return -1;
	memcpy(eui, dev->dev_addr, 8);
1982
	eui[0] ^= 2;
1983 1984 1985
	return 0;
}

1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999
static int addrconf_ifid_ieee1394(u8 *eui, struct net_device *dev)
{
	union fwnet_hwaddr *ha;

	if (dev->addr_len != FWNET_ALEN)
		return -1;

	ha = (union fwnet_hwaddr *)dev->dev_addr;

	memcpy(eui, &ha->uc.uniq_id, sizeof(ha->uc.uniq_id));
	eui[0] ^= 2;
	return 0;
}

2000 2001 2002 2003 2004 2005
static int addrconf_ifid_arcnet(u8 *eui, struct net_device *dev)
{
	/* XXX: inherit EUI-64 from other interface -- yoshfuji */
	if (dev->addr_len != ARCNET_ALEN)
		return -1;
	memset(eui, 0, 7);
2006
	eui[7] = *(u8 *)dev->dev_addr;
2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018
	return 0;
}

static int addrconf_ifid_infiniband(u8 *eui, struct net_device *dev)
{
	if (dev->addr_len != INFINIBAND_ALEN)
		return -1;
	memcpy(eui, dev->dev_addr + 12, 8);
	eui[0] |= 2;
	return 0;
}

2019
static int __ipv6_isatap_ifid(u8 *eui, __be32 addr)
2020
{
2021 2022
	if (addr == 0)
		return -1;
2023 2024 2025 2026 2027 2028 2029 2030 2031 2032 2033 2034 2035 2036 2037 2038 2039 2040 2041 2042
	eui[0] = (ipv4_is_zeronet(addr) || ipv4_is_private_10(addr) ||
		  ipv4_is_loopback(addr) || ipv4_is_linklocal_169(addr) ||
		  ipv4_is_private_172(addr) || ipv4_is_test_192(addr) ||
		  ipv4_is_anycast_6to4(addr) || ipv4_is_private_192(addr) ||
		  ipv4_is_test_198(addr) || ipv4_is_multicast(addr) ||
		  ipv4_is_lbcast(addr)) ? 0x00 : 0x02;
	eui[1] = 0;
	eui[2] = 0x5E;
	eui[3] = 0xFE;
	memcpy(eui + 4, &addr, 4);
	return 0;
}

static int addrconf_ifid_sit(u8 *eui, struct net_device *dev)
{
	if (dev->priv_flags & IFF_ISATAP)
		return __ipv6_isatap_ifid(eui, *(__be32 *)dev->dev_addr);
	return -1;
}

2043 2044 2045 2046 2047
static int addrconf_ifid_gre(u8 *eui, struct net_device *dev)
{
	return __ipv6_isatap_ifid(eui, *(__be32 *)dev->dev_addr);
}

2048 2049 2050 2051 2052 2053 2054 2055 2056 2057
static int addrconf_ifid_ip6tnl(u8 *eui, struct net_device *dev)
{
	memcpy(eui, dev->perm_addr, 3);
	memcpy(eui + 5, dev->perm_addr + 3, 3);
	eui[3] = 0xFF;
	eui[4] = 0xFE;
	eui[0] ^= 2;
	return 0;
}

L
Linus Torvalds 已提交
2058 2059 2060 2061 2062
static int ipv6_generate_eui64(u8 *eui, struct net_device *dev)
{
	switch (dev->type) {
	case ARPHRD_ETHER:
	case ARPHRD_FDDI:
2063
		return addrconf_ifid_eui48(eui, dev);
L
Linus Torvalds 已提交
2064
	case ARPHRD_ARCNET:
2065
		return addrconf_ifid_arcnet(eui, dev);
L
Linus Torvalds 已提交
2066
	case ARPHRD_INFINIBAND:
2067
		return addrconf_ifid_infiniband(eui, dev);
F
Fred L. Templin 已提交
2068
	case ARPHRD_SIT:
2069
		return addrconf_ifid_sit(eui, dev);
2070 2071
	case ARPHRD_IPGRE:
		return addrconf_ifid_gre(eui, dev);
2072
	case ARPHRD_6LOWPAN:
2073 2074
	case ARPHRD_IEEE802154:
		return addrconf_ifid_eui64(eui, dev);
2075 2076
	case ARPHRD_IEEE1394:
		return addrconf_ifid_ieee1394(eui, dev);
2077 2078
	case ARPHRD_TUNNEL6:
		return addrconf_ifid_ip6tnl(eui, dev);
L
Linus Torvalds 已提交
2079 2080 2081 2082 2083 2084 2085 2086 2087 2088
	}
	return -1;
}

static int ipv6_inherit_eui64(u8 *eui, struct inet6_dev *idev)
{
	int err = -1;
	struct inet6_ifaddr *ifp;

	read_lock_bh(&idev->lock);
2089 2090 2091
	list_for_each_entry_reverse(ifp, &idev->addr_list, if_list) {
		if (ifp->scope > IFA_LINK)
			break;
L
Linus Torvalds 已提交
2092 2093 2094 2095 2096 2097 2098 2099 2100 2101 2102
		if (ifp->scope == IFA_LINK && !(ifp->flags&IFA_F_TENTATIVE)) {
			memcpy(eui, ifp->addr.s6_addr+8, 8);
			err = 0;
			break;
		}
	}
	read_unlock_bh(&idev->lock);
	return err;
}

/* (re)generation of randomized interface identifier (RFC 3041 3.2, 3.5) */
S
Sorin Dumitru 已提交
2103
static void __ipv6_regen_rndid(struct inet6_dev *idev)
L
Linus Torvalds 已提交
2104 2105
{
regen:
2106
	get_random_bytes(idev->rndid, sizeof(idev->rndid));
L
Linus Torvalds 已提交
2107 2108 2109 2110 2111 2112 2113 2114
	idev->rndid[0] &= ~0x02;

	/*
	 * <draft-ietf-ipngwg-temp-addresses-v2-00.txt>:
	 * check if generated address is not inappropriate
	 *
	 *  - Reserved subnet anycast (RFC 2526)
	 *	11111101 11....11 1xxxxxxx
F
Fred L. Templin 已提交
2115
	 *  - ISATAP (RFC4214) 6.1
L
Linus Torvalds 已提交
2116 2117 2118 2119
	 *	00-00-5E-FE-xx-xx-xx-xx
	 *  - value 0
	 *  - XXX: already assigned to an address on the device
	 */
2120
	if (idev->rndid[0] == 0xfd &&
L
Linus Torvalds 已提交
2121 2122 2123 2124 2125 2126 2127 2128 2129 2130 2131 2132 2133 2134 2135 2136
	    (idev->rndid[1]&idev->rndid[2]&idev->rndid[3]&idev->rndid[4]&idev->rndid[5]&idev->rndid[6]) == 0xff &&
	    (idev->rndid[7]&0x80))
		goto regen;
	if ((idev->rndid[0]|idev->rndid[1]) == 0) {
		if (idev->rndid[2] == 0x5e && idev->rndid[3] == 0xfe)
			goto regen;
		if ((idev->rndid[2]|idev->rndid[3]|idev->rndid[4]|idev->rndid[5]|idev->rndid[6]|idev->rndid[7]) == 0x00)
			goto regen;
	}
}

static void ipv6_regen_rndid(unsigned long data)
{
	struct inet6_dev *idev = (struct inet6_dev *) data;
	unsigned long expires;

2137
	rcu_read_lock_bh();
L
Linus Torvalds 已提交
2138 2139 2140 2141 2142
	write_lock_bh(&idev->lock);

	if (idev->dead)
		goto out;

S
Sorin Dumitru 已提交
2143
	__ipv6_regen_rndid(idev);
2144

L
Linus Torvalds 已提交
2145
	expires = jiffies +
2146
		idev->cnf.temp_prefered_lft * HZ -
J
Jiri Pirko 已提交
2147 2148
		idev->cnf.regen_max_retry * idev->cnf.dad_transmits *
		NEIGH_VAR(idev->nd_parms, RETRANS_TIME) -
2149
		idev->cnf.max_desync_factor * HZ;
L
Linus Torvalds 已提交
2150
	if (time_before(expires, jiffies)) {
2151 2152
		pr_warn("%s: too short regeneration interval; timer disabled for %s\n",
			__func__, idev->dev->name);
L
Linus Torvalds 已提交
2153 2154 2155 2156 2157 2158 2159 2160
		goto out;
	}

	if (!mod_timer(&idev->regen_timer, expires))
		in6_dev_hold(idev);

out:
	write_unlock_bh(&idev->lock);
2161
	rcu_read_unlock_bh();
L
Linus Torvalds 已提交
2162 2163 2164
	in6_dev_put(idev);
}

S
Sorin Dumitru 已提交
2165
static void  __ipv6_try_regen_rndid(struct inet6_dev *idev, struct in6_addr *tmpaddr)
2166
{
L
Linus Torvalds 已提交
2167
	if (tmpaddr && memcmp(idev->rndid, &tmpaddr->s6_addr[8], 8) == 0)
S
Sorin Dumitru 已提交
2168
		__ipv6_regen_rndid(idev);
L
Linus Torvalds 已提交
2169 2170 2171 2172 2173 2174 2175 2176
}

/*
 *	Add prefix route.
 */

static void
addrconf_prefix_route(struct in6_addr *pfx, int plen, struct net_device *dev,
J
Jamal Hadi Salim 已提交
2177
		      unsigned long expires, u32 flags)
L
Linus Torvalds 已提交
2178
{
2179 2180 2181 2182 2183 2184 2185
	struct fib6_config cfg = {
		.fc_table = RT6_TABLE_PREFIX,
		.fc_metric = IP6_RT_PRIO_ADDRCONF,
		.fc_ifindex = dev->ifindex,
		.fc_expires = expires,
		.fc_dst_len = plen,
		.fc_flags = RTF_UP | flags,
2186
		.fc_nlinfo.nl_net = dev_net(dev),
2187
		.fc_protocol = RTPROT_KERNEL,
2188
	};
L
Linus Torvalds 已提交
2189

A
Alexey Dobriyan 已提交
2190
	cfg.fc_dst = *pfx;
L
Linus Torvalds 已提交
2191 2192 2193 2194 2195

	/* Prevent useless cloning on PtP SIT.
	   This thing is done here expecting that the whole
	   class of non-broadcast devices need not cloning.
	 */
A
Amerigo Wang 已提交
2196
#if IS_ENABLED(CONFIG_IPV6_SIT)
2197 2198
	if (dev->type == ARPHRD_SIT && (dev->flags & IFF_POINTOPOINT))
		cfg.fc_flags |= RTF_NONEXTHOP;
2199
#endif
L
Linus Torvalds 已提交
2200

2201
	ip6_route_add(&cfg);
L
Linus Torvalds 已提交
2202 2203
}

2204 2205 2206 2207 2208 2209 2210 2211 2212 2213 2214

static struct rt6_info *addrconf_get_prefix_route(const struct in6_addr *pfx,
						  int plen,
						  const struct net_device *dev,
						  u32 flags, u32 noflags)
{
	struct fib6_node *fn;
	struct rt6_info *rt = NULL;
	struct fib6_table *table;

	table = fib6_get_table(dev_net(dev), RT6_TABLE_PREFIX);
2215
	if (!table)
2216 2217
		return NULL;

2218
	read_lock_bh(&table->tb6_lock);
2219 2220 2221
	fn = fib6_locate(&table->tb6_root, pfx, plen, NULL, 0);
	if (!fn)
		goto out;
2222 2223

	noflags |= RTF_CACHE;
2224
	for (rt = fn->leaf; rt; rt = rt->dst.rt6_next) {
2225
		if (rt->dst.dev->ifindex != dev->ifindex)
2226 2227 2228
			continue;
		if ((rt->rt6i_flags & flags) != flags)
			continue;
2229
		if ((rt->rt6i_flags & noflags) != 0)
2230 2231 2232 2233 2234
			continue;
		dst_hold(&rt->dst);
		break;
	}
out:
2235
	read_unlock_bh(&table->tb6_lock);
2236 2237 2238 2239
	return rt;
}


L
Linus Torvalds 已提交
2240 2241 2242 2243
/* Create "default" multicast route to the interface */

static void addrconf_add_mroute(struct net_device *dev)
{
2244 2245 2246 2247 2248 2249
	struct fib6_config cfg = {
		.fc_table = RT6_TABLE_LOCAL,
		.fc_metric = IP6_RT_PRIO_ADDRCONF,
		.fc_ifindex = dev->ifindex,
		.fc_dst_len = 8,
		.fc_flags = RTF_UP,
2250
		.fc_nlinfo.nl_net = dev_net(dev),
2251 2252 2253 2254 2255
	};

	ipv6_addr_set(&cfg.fc_dst, htonl(0xFF000000), 0, 0, 0);

	ip6_route_add(&cfg);
L
Linus Torvalds 已提交
2256 2257 2258 2259 2260 2261 2262 2263
}

static struct inet6_dev *addrconf_add_dev(struct net_device *dev)
{
	struct inet6_dev *idev;

	ASSERT_RTNL();

S
Stephen Hemminger 已提交
2264 2265
	idev = ipv6_find_idev(dev);
	if (!idev)
2266 2267 2268 2269
		return ERR_PTR(-ENOBUFS);

	if (idev->cnf.disable_ipv6)
		return ERR_PTR(-EACCES);
L
Linus Torvalds 已提交
2270 2271

	/* Add default multicast route */
2272 2273
	if (!(dev->flags & IFF_LOOPBACK))
		addrconf_add_mroute(dev);
L
Linus Torvalds 已提交
2274 2275 2276 2277

	return idev;
}

2278 2279 2280 2281 2282 2283 2284 2285 2286 2287 2288 2289 2290 2291 2292 2293 2294 2295 2296 2297 2298 2299 2300 2301 2302 2303 2304 2305 2306 2307 2308 2309 2310 2311 2312 2313 2314 2315 2316 2317 2318 2319 2320 2321 2322 2323 2324 2325 2326 2327 2328 2329 2330 2331 2332 2333 2334 2335 2336 2337 2338 2339 2340 2341 2342 2343 2344
static void manage_tempaddrs(struct inet6_dev *idev,
			     struct inet6_ifaddr *ifp,
			     __u32 valid_lft, __u32 prefered_lft,
			     bool create, unsigned long now)
{
	u32 flags;
	struct inet6_ifaddr *ift;

	read_lock_bh(&idev->lock);
	/* update all temporary addresses in the list */
	list_for_each_entry(ift, &idev->tempaddr_list, tmp_list) {
		int age, max_valid, max_prefered;

		if (ifp != ift->ifpub)
			continue;

		/* RFC 4941 section 3.3:
		 * If a received option will extend the lifetime of a public
		 * address, the lifetimes of temporary addresses should
		 * be extended, subject to the overall constraint that no
		 * temporary addresses should ever remain "valid" or "preferred"
		 * for a time longer than (TEMP_VALID_LIFETIME) or
		 * (TEMP_PREFERRED_LIFETIME - DESYNC_FACTOR), respectively.
		 */
		age = (now - ift->cstamp) / HZ;
		max_valid = idev->cnf.temp_valid_lft - age;
		if (max_valid < 0)
			max_valid = 0;

		max_prefered = idev->cnf.temp_prefered_lft -
			       idev->cnf.max_desync_factor - age;
		if (max_prefered < 0)
			max_prefered = 0;

		if (valid_lft > max_valid)
			valid_lft = max_valid;

		if (prefered_lft > max_prefered)
			prefered_lft = max_prefered;

		spin_lock(&ift->lock);
		flags = ift->flags;
		ift->valid_lft = valid_lft;
		ift->prefered_lft = prefered_lft;
		ift->tstamp = now;
		if (prefered_lft > 0)
			ift->flags &= ~IFA_F_DEPRECATED;

		spin_unlock(&ift->lock);
		if (!(flags&IFA_F_TENTATIVE))
			ipv6_ifa_notify(0, ift);
	}

	if ((create || list_empty(&idev->tempaddr_list)) &&
	    idev->cnf.use_tempaddr > 0) {
		/* When a new public address is created as described
		 * in [ADDRCONF], also create a new temporary address.
		 * Also create a temporary address if it's enabled but
		 * no temporary address currently exists.
		 */
		read_unlock_bh(&idev->lock);
		ipv6_create_tempaddr(ifp, NULL);
	} else {
		read_unlock_bh(&idev->lock);
	}
}

2345
void addrconf_prefix_rcv(struct net_device *dev, u8 *opt, int len, bool sllao)
L
Linus Torvalds 已提交
2346 2347 2348 2349 2350
{
	struct prefix_info *pinfo;
	__u32 valid_lft;
	__u32 prefered_lft;
	int addr_type;
2351
	u32 addr_flags = 0;
L
Linus Torvalds 已提交
2352
	struct inet6_dev *in6_dev;
2353
	struct net *net = dev_net(dev);
L
Linus Torvalds 已提交
2354 2355

	pinfo = (struct prefix_info *) opt;
2356

L
Linus Torvalds 已提交
2357
	if (len < sizeof(struct prefix_info)) {
2358
		ADBG("addrconf: prefix option too short\n");
L
Linus Torvalds 已提交
2359 2360
		return;
	}
2361

L
Linus Torvalds 已提交
2362 2363 2364 2365 2366 2367 2368 2369 2370 2371 2372 2373 2374
	/*
	 *	Validation checks ([ADDRCONF], page 19)
	 */

	addr_type = ipv6_addr_type(&pinfo->prefix);

	if (addr_type & (IPV6_ADDR_MULTICAST|IPV6_ADDR_LINKLOCAL))
		return;

	valid_lft = ntohl(pinfo->valid);
	prefered_lft = ntohl(pinfo->prefered);

	if (prefered_lft > valid_lft) {
2375
		net_warn_ratelimited("addrconf: prefix option has invalid lifetime\n");
L
Linus Torvalds 已提交
2376 2377 2378 2379 2380
		return;
	}

	in6_dev = in6_dev_get(dev);

2381
	if (!in6_dev) {
2382 2383
		net_dbg_ratelimited("addrconf: device %s not configured\n",
				    dev->name);
L
Linus Torvalds 已提交
2384 2385 2386 2387 2388 2389 2390 2391 2392
		return;
	}

	/*
	 *	Two things going on here:
	 *	1) Add routes for on-link prefixes
	 *	2) Configure prefixes with the auto flag set
	 */

2393 2394 2395 2396
	if (pinfo->onlink) {
		struct rt6_info *rt;
		unsigned long rt_expires;

2397 2398 2399 2400 2401
		/* Avoid arithmetic overflow. Really, we could
		 * save rt_expires in seconds, likely valid_lft,
		 * but it would require division in fib gc, that it
		 * not good.
		 */
2402 2403 2404 2405
		if (HZ > USER_HZ)
			rt_expires = addrconf_timeout_fixup(valid_lft, HZ);
		else
			rt_expires = addrconf_timeout_fixup(valid_lft, USER_HZ);
Y
YOSHIFUJI Hideaki 已提交
2406

2407 2408
		if (addrconf_finite_timeout(rt_expires))
			rt_expires *= HZ;
L
Linus Torvalds 已提交
2409

2410 2411 2412 2413 2414
		rt = addrconf_get_prefix_route(&pinfo->prefix,
					       pinfo->prefix_len,
					       dev,
					       RTF_ADDRCONF | RTF_PREFIX_RT,
					       RTF_GATEWAY | RTF_DEFAULT);
L
Linus Torvalds 已提交
2415

2416
		if (rt) {
2417 2418 2419 2420
			/* Autoconf prefix route */
			if (valid_lft == 0) {
				ip6_del_rt(rt);
				rt = NULL;
2421
			} else if (addrconf_finite_timeout(rt_expires)) {
2422
				/* not infinity */
2423
				rt6_set_expires(rt, jiffies + rt_expires);
2424
			} else {
2425
				rt6_clean_expires(rt);
L
Linus Torvalds 已提交
2426 2427
			}
		} else if (valid_lft) {
2428
			clock_t expires = 0;
2429 2430
			int flags = RTF_ADDRCONF | RTF_PREFIX_RT;
			if (addrconf_finite_timeout(rt_expires)) {
2431 2432 2433 2434
				/* not infinity */
				flags |= RTF_EXPIRES;
				expires = jiffies_to_clock_t(rt_expires);
			}
L
Linus Torvalds 已提交
2435
			addrconf_prefix_route(&pinfo->prefix, pinfo->prefix_len,
2436
					      dev, expires, flags);
L
Linus Torvalds 已提交
2437
		}
A
Amerigo Wang 已提交
2438
		ip6_rt_put(rt);
L
Linus Torvalds 已提交
2439 2440 2441 2442 2443
	}

	/* Try to figure out our local address for this prefix */

	if (pinfo->autoconf && in6_dev->cnf.autoconf) {
2444
		struct inet6_ifaddr *ifp;
L
Linus Torvalds 已提交
2445 2446
		struct in6_addr addr;
		int create = 0, update_lft = 0;
2447
		bool tokenized = false;
L
Linus Torvalds 已提交
2448 2449 2450

		if (pinfo->prefix_len == 64) {
			memcpy(&addr, &pinfo->prefix, 8);
2451 2452 2453 2454 2455 2456

			if (!ipv6_addr_any(&in6_dev->token)) {
				read_lock_bh(&in6_dev->lock);
				memcpy(addr.s6_addr + 8,
				       in6_dev->token.s6_addr + 8, 8);
				read_unlock_bh(&in6_dev->lock);
2457
				tokenized = true;
2458 2459 2460 2461
			} else if (in6_dev->addr_gen_mode ==
				   IN6_ADDR_GEN_MODE_STABLE_PRIVACY &&
				   !ipv6_generate_stable_address(&addr, 0,
								 in6_dev)) {
2462
				addr_flags |= IFA_F_STABLE_PRIVACY;
2463
				goto ok;
2464 2465
			} else if (ipv6_generate_eui64(addr.s6_addr + 8, dev) &&
				   ipv6_inherit_eui64(addr.s6_addr + 8, in6_dev)) {
L
Linus Torvalds 已提交
2466 2467 2468 2469 2470
				in6_dev_put(in6_dev);
				return;
			}
			goto ok;
		}
2471 2472
		net_dbg_ratelimited("IPv6 addrconf: prefix with wrong length %d\n",
				    pinfo->prefix_len);
L
Linus Torvalds 已提交
2473 2474 2475 2476 2477
		in6_dev_put(in6_dev);
		return;

ok:

2478
		ifp = ipv6_get_ifaddr(net, &addr, dev, 1);
L
Linus Torvalds 已提交
2479

2480
		if (!ifp && valid_lft) {
L
Linus Torvalds 已提交
2481
			int max_addresses = in6_dev->cnf.max_addresses;
2482 2483 2484

#ifdef CONFIG_IPV6_OPTIMISTIC_DAD
			if (in6_dev->cnf.optimistic_dad &&
2485
			    !net->ipv6.devconf_all->forwarding && sllao)
2486 2487
				addr_flags = IFA_F_OPTIMISTIC;
#endif
L
Linus Torvalds 已提交
2488 2489 2490 2491 2492 2493

			/* Do not allow to create too much of autoconfigured
			 * addresses; this would be too easy way to crash kernel.
			 */
			if (!max_addresses ||
			    ipv6_count_addresses(in6_dev) < max_addresses)
2494 2495
				ifp = ipv6_add_addr(in6_dev, &addr, NULL,
						    pinfo->prefix_len,
2496
						    addr_type&IPV6_ADDR_SCOPE_MASK,
2497 2498
						    addr_flags, valid_lft,
						    prefered_lft);
L
Linus Torvalds 已提交
2499

2500
			if (IS_ERR_OR_NULL(ifp)) {
L
Linus Torvalds 已提交
2501 2502 2503 2504
				in6_dev_put(in6_dev);
				return;
			}

2505 2506
			update_lft = 0;
			create = 1;
2507 2508
			spin_lock_bh(&ifp->lock);
			ifp->flags |= IFA_F_MANAGETEMPADDR;
L
Linus Torvalds 已提交
2509
			ifp->cstamp = jiffies;
2510
			ifp->tokenized = tokenized;
2511
			spin_unlock_bh(&ifp->lock);
2512
			addrconf_dad_start(ifp);
L
Linus Torvalds 已提交
2513 2514 2515
		}

		if (ifp) {
2516
			u32 flags;
L
Linus Torvalds 已提交
2517 2518 2519 2520
			unsigned long now;
			u32 stored_lft;

			/* update lifetime (RFC2462 5.5.3 e) */
2521
			spin_lock_bh(&ifp->lock);
L
Linus Torvalds 已提交
2522 2523 2524 2525 2526
			now = jiffies;
			if (ifp->valid_lft > (now - ifp->tstamp) / HZ)
				stored_lft = ifp->valid_lft - (now - ifp->tstamp) / HZ;
			else
				stored_lft = 0;
2527
			if (!update_lft && !create && stored_lft) {
2528 2529
				const u32 minimum_lft = min_t(u32,
					stored_lft, MIN_VALID_LIFETIME);
2530 2531 2532 2533 2534 2535 2536 2537 2538 2539 2540 2541 2542
				valid_lft = max(valid_lft, minimum_lft);

				/* RFC4862 Section 5.5.3e:
				 * "Note that the preferred lifetime of the
				 *  corresponding address is always reset to
				 *  the Preferred Lifetime in the received
				 *  Prefix Information option, regardless of
				 *  whether the valid lifetime is also reset or
				 *  ignored."
				 *
				 * So we should always update prefered_lft here.
				 */
				update_lft = 1;
L
Linus Torvalds 已提交
2543 2544 2545 2546 2547 2548 2549 2550
			}

			if (update_lft) {
				ifp->valid_lft = valid_lft;
				ifp->prefered_lft = prefered_lft;
				ifp->tstamp = now;
				flags = ifp->flags;
				ifp->flags &= ~IFA_F_DEPRECATED;
2551
				spin_unlock_bh(&ifp->lock);
L
Linus Torvalds 已提交
2552 2553 2554 2555

				if (!(flags&IFA_F_TENTATIVE))
					ipv6_ifa_notify(0, ifp);
			} else
2556
				spin_unlock_bh(&ifp->lock);
L
Linus Torvalds 已提交
2557

2558 2559
			manage_tempaddrs(in6_dev, ifp, valid_lft, prefered_lft,
					 create, now);
2560

L
Linus Torvalds 已提交
2561
			in6_ifa_put(ifp);
2562
			addrconf_verify();
L
Linus Torvalds 已提交
2563 2564 2565 2566 2567 2568 2569 2570 2571 2572 2573
		}
	}
	inet6_prefix_notify(RTM_NEWPREFIX, in6_dev, pinfo);
	in6_dev_put(in6_dev);
}

/*
 *	Set destination address.
 *	Special case for SIT interfaces where we create a new "virtual"
 *	device.
 */
2574
int addrconf_set_dstaddr(struct net *net, void __user *arg)
L
Linus Torvalds 已提交
2575 2576 2577 2578 2579 2580 2581 2582 2583 2584 2585
{
	struct in6_ifreq ireq;
	struct net_device *dev;
	int err = -EINVAL;

	rtnl_lock();

	err = -EFAULT;
	if (copy_from_user(&ireq, arg, sizeof(struct in6_ifreq)))
		goto err_exit;

2586
	dev = __dev_get_by_index(net, ireq.ifr6_ifindex);
L
Linus Torvalds 已提交
2587 2588

	err = -ENODEV;
2589
	if (!dev)
L
Linus Torvalds 已提交
2590 2591
		goto err_exit;

A
Amerigo Wang 已提交
2592
#if IS_ENABLED(CONFIG_IPV6_SIT)
L
Linus Torvalds 已提交
2593
	if (dev->type == ARPHRD_SIT) {
2594
		const struct net_device_ops *ops = dev->netdev_ops;
L
Linus Torvalds 已提交
2595 2596 2597 2598 2599 2600 2601 2602 2603 2604 2605 2606 2607 2608
		struct ifreq ifr;
		struct ip_tunnel_parm p;

		err = -EADDRNOTAVAIL;
		if (!(ipv6_addr_type(&ireq.ifr6_addr) & IPV6_ADDR_COMPATv4))
			goto err_exit;

		memset(&p, 0, sizeof(p));
		p.iph.daddr = ireq.ifr6_addr.s6_addr32[3];
		p.iph.saddr = 0;
		p.iph.version = 4;
		p.iph.ihl = 5;
		p.iph.protocol = IPPROTO_IPV6;
		p.iph.ttl = 64;
2609
		ifr.ifr_ifru.ifru_data = (__force void __user *)&p;
L
Linus Torvalds 已提交
2610

2611 2612 2613 2614 2615 2616 2617 2618
		if (ops->ndo_do_ioctl) {
			mm_segment_t oldfs = get_fs();

			set_fs(KERNEL_DS);
			err = ops->ndo_do_ioctl(dev, &ifr, SIOCADDTUNNEL);
			set_fs(oldfs);
		} else
			err = -EOPNOTSUPP;
L
Linus Torvalds 已提交
2619 2620 2621

		if (err == 0) {
			err = -ENOBUFS;
2622 2623
			dev = __dev_get_by_name(net, p.name);
			if (!dev)
L
Linus Torvalds 已提交
2624 2625 2626 2627
				goto err_exit;
			err = dev_open(dev);
		}
	}
2628
#endif
L
Linus Torvalds 已提交
2629 2630 2631 2632 2633 2634

err_exit:
	rtnl_unlock();
	return err;
}

2635 2636 2637 2638 2639 2640 2641 2642 2643
static int ipv6_mc_config(struct sock *sk, bool join,
			  const struct in6_addr *addr, int ifindex)
{
	int ret;

	ASSERT_RTNL();

	lock_sock(sk);
	if (join)
2644
		ret = ipv6_sock_mc_join(sk, ifindex, addr);
2645
	else
2646
		ret = ipv6_sock_mc_drop(sk, ifindex, addr);
2647 2648 2649 2650 2651
	release_sock(sk);

	return ret;
}

L
Linus Torvalds 已提交
2652 2653 2654
/*
 *	Manual configuration of address on an interface
 */
2655 2656
static int inet6_addr_add(struct net *net, int ifindex,
			  const struct in6_addr *pfx,
2657
			  const struct in6_addr *peer_pfx,
2658 2659
			  unsigned int plen, __u32 ifa_flags,
			  __u32 prefered_lft, __u32 valid_lft)
L
Linus Torvalds 已提交
2660 2661 2662 2663
{
	struct inet6_ifaddr *ifp;
	struct inet6_dev *idev;
	struct net_device *dev;
2664 2665
	unsigned long timeout;
	clock_t expires;
L
Linus Torvalds 已提交
2666
	int scope;
2667
	u32 flags;
L
Linus Torvalds 已提交
2668 2669

	ASSERT_RTNL();
2670

2671 2672 2673
	if (plen > 128)
		return -EINVAL;

2674 2675 2676 2677
	/* check the lifetime */
	if (!valid_lft || prefered_lft > valid_lft)
		return -EINVAL;

2678 2679 2680
	if (ifa_flags & IFA_F_MANAGETEMPADDR && plen != 64)
		return -EINVAL;

2681 2682
	dev = __dev_get_by_index(net, ifindex);
	if (!dev)
L
Linus Torvalds 已提交
2683
		return -ENODEV;
2684

2685 2686 2687
	idev = addrconf_add_dev(dev);
	if (IS_ERR(idev))
		return PTR_ERR(idev);
L
Linus Torvalds 已提交
2688

2689 2690 2691 2692 2693 2694 2695 2696
	if (ifa_flags & IFA_F_MCAUTOJOIN) {
		int ret = ipv6_mc_config(net->ipv6.mc_autojoin_sk,
					 true, pfx, ifindex);

		if (ret < 0)
			return ret;
	}

L
Linus Torvalds 已提交
2697 2698
	scope = ipv6_addr_scope(pfx);

2699 2700 2701 2702
	timeout = addrconf_timeout_fixup(valid_lft, HZ);
	if (addrconf_finite_timeout(timeout)) {
		expires = jiffies_to_clock_t(timeout * HZ);
		valid_lft = timeout;
2703
		flags = RTF_EXPIRES;
2704 2705 2706 2707
	} else {
		expires = 0;
		flags = 0;
		ifa_flags |= IFA_F_PERMANENT;
2708
	}
2709

2710 2711 2712 2713 2714 2715
	timeout = addrconf_timeout_fixup(prefered_lft, HZ);
	if (addrconf_finite_timeout(timeout)) {
		if (timeout == 0)
			ifa_flags |= IFA_F_DEPRECATED;
		prefered_lft = timeout;
	}
2716

2717 2718
	ifp = ipv6_add_addr(idev, pfx, peer_pfx, plen, scope, ifa_flags,
			    valid_lft, prefered_lft);
2719

L
Linus Torvalds 已提交
2720
	if (!IS_ERR(ifp)) {
2721 2722 2723 2724 2725
		if (!(ifa_flags & IFA_F_NOPREFIXROUTE)) {
			addrconf_prefix_route(&ifp->addr, ifp->prefix_len, dev,
					      expires, flags);
		}

2726 2727 2728 2729 2730
		/*
		 * Note that section 3.1 of RFC 4429 indicates
		 * that the Optimistic flag should not be set for
		 * manually configured addresses
		 */
2731
		addrconf_dad_start(ifp);
2732 2733 2734
		if (ifa_flags & IFA_F_MANAGETEMPADDR)
			manage_tempaddrs(idev, ifp, valid_lft, prefered_lft,
					 true, jiffies);
L
Linus Torvalds 已提交
2735
		in6_ifa_put(ifp);
2736
		addrconf_verify_rtnl();
L
Linus Torvalds 已提交
2737
		return 0;
2738 2739 2740
	} else if (ifa_flags & IFA_F_MCAUTOJOIN) {
		ipv6_mc_config(net->ipv6.mc_autojoin_sk,
			       false, pfx, ifindex);
L
Linus Torvalds 已提交
2741 2742 2743 2744 2745
	}

	return PTR_ERR(ifp);
}

2746 2747
static int inet6_addr_del(struct net *net, int ifindex, u32 ifa_flags,
			  const struct in6_addr *pfx, unsigned int plen)
L
Linus Torvalds 已提交
2748 2749 2750 2751
{
	struct inet6_ifaddr *ifp;
	struct inet6_dev *idev;
	struct net_device *dev;
2752

2753 2754 2755
	if (plen > 128)
		return -EINVAL;

2756 2757
	dev = __dev_get_by_index(net, ifindex);
	if (!dev)
L
Linus Torvalds 已提交
2758 2759
		return -ENODEV;

2760
	idev = __in6_dev_get(dev);
2761
	if (!idev)
L
Linus Torvalds 已提交
2762 2763 2764
		return -ENXIO;

	read_lock_bh(&idev->lock);
2765
	list_for_each_entry(ifp, &idev->addr_list, if_list) {
L
Linus Torvalds 已提交
2766 2767 2768 2769
		if (ifp->prefix_len == plen &&
		    ipv6_addr_equal(pfx, &ifp->addr)) {
			in6_ifa_hold(ifp);
			read_unlock_bh(&idev->lock);
2770

2771 2772 2773 2774
			if (!(ifp->flags & IFA_F_TEMPORARY) &&
			    (ifa_flags & IFA_F_MANAGETEMPADDR))
				manage_tempaddrs(idev, ifp, 0, 0, false,
						 jiffies);
L
Linus Torvalds 已提交
2775
			ipv6_del_addr(ifp);
2776
			addrconf_verify_rtnl();
2777 2778 2779 2780
			if (ipv6_addr_is_multicast(pfx)) {
				ipv6_mc_config(net->ipv6.mc_autojoin_sk,
					       false, pfx, dev->ifindex);
			}
L
Linus Torvalds 已提交
2781 2782 2783 2784 2785 2786 2787 2788
			return 0;
		}
	}
	read_unlock_bh(&idev->lock);
	return -EADDRNOTAVAIL;
}


2789
int addrconf_add_ifaddr(struct net *net, void __user *arg)
L
Linus Torvalds 已提交
2790 2791 2792
{
	struct in6_ifreq ireq;
	int err;
2793

2794
	if (!ns_capable(net->user_ns, CAP_NET_ADMIN))
L
Linus Torvalds 已提交
2795
		return -EPERM;
2796

L
Linus Torvalds 已提交
2797 2798 2799 2800
	if (copy_from_user(&ireq, arg, sizeof(struct in6_ifreq)))
		return -EFAULT;

	rtnl_lock();
2801
	err = inet6_addr_add(net, ireq.ifr6_ifindex, &ireq.ifr6_addr, NULL,
2802 2803
			     ireq.ifr6_prefixlen, IFA_F_PERMANENT,
			     INFINITY_LIFE_TIME, INFINITY_LIFE_TIME);
L
Linus Torvalds 已提交
2804 2805 2806 2807
	rtnl_unlock();
	return err;
}

2808
int addrconf_del_ifaddr(struct net *net, void __user *arg)
L
Linus Torvalds 已提交
2809 2810 2811
{
	struct in6_ifreq ireq;
	int err;
2812

2813
	if (!ns_capable(net->user_ns, CAP_NET_ADMIN))
L
Linus Torvalds 已提交
2814 2815 2816 2817 2818 2819
		return -EPERM;

	if (copy_from_user(&ireq, arg, sizeof(struct in6_ifreq)))
		return -EFAULT;

	rtnl_lock();
2820
	err = inet6_addr_del(net, ireq.ifr6_ifindex, 0, &ireq.ifr6_addr,
2821
			     ireq.ifr6_prefixlen);
L
Linus Torvalds 已提交
2822 2823 2824 2825
	rtnl_unlock();
	return err;
}

2826 2827 2828 2829 2830
static void add_addr(struct inet6_dev *idev, const struct in6_addr *addr,
		     int plen, int scope)
{
	struct inet6_ifaddr *ifp;

2831
	ifp = ipv6_add_addr(idev, addr, NULL, plen,
2832 2833
			    scope, IFA_F_PERMANENT,
			    INFINITY_LIFE_TIME, INFINITY_LIFE_TIME);
2834 2835 2836 2837 2838 2839 2840 2841 2842
	if (!IS_ERR(ifp)) {
		spin_lock_bh(&ifp->lock);
		ifp->flags &= ~IFA_F_TENTATIVE;
		spin_unlock_bh(&ifp->lock);
		ipv6_ifa_notify(RTM_NEWADDR, ifp);
		in6_ifa_put(ifp);
	}
}

A
Amerigo Wang 已提交
2843
#if IS_ENABLED(CONFIG_IPV6_SIT)
L
Linus Torvalds 已提交
2844 2845 2846 2847
static void sit_add_v4_addrs(struct inet6_dev *idev)
{
	struct in6_addr addr;
	struct net_device *dev;
2848
	struct net *net = dev_net(idev->dev);
2849
	int scope, plen;
2850
	u32 pflags = 0;
L
Linus Torvalds 已提交
2851 2852 2853 2854 2855 2856 2857 2858 2859

	ASSERT_RTNL();

	memset(&addr, 0, sizeof(struct in6_addr));
	memcpy(&addr.s6_addr32[3], idev->dev->dev_addr, 4);

	if (idev->dev->flags&IFF_POINTOPOINT) {
		addr.s6_addr32[0] = htonl(0xfe800000);
		scope = IFA_LINK;
2860
		plen = 64;
L
Linus Torvalds 已提交
2861 2862
	} else {
		scope = IPV6_ADDR_COMPATv4;
2863
		plen = 96;
2864
		pflags |= RTF_NONEXTHOP;
L
Linus Torvalds 已提交
2865 2866 2867
	}

	if (addr.s6_addr32[3]) {
2868
		add_addr(idev, &addr, plen, scope);
2869
		addrconf_prefix_route(&addr, plen, idev->dev, 0, pflags);
L
Linus Torvalds 已提交
2870 2871 2872
		return;
	}

2873
	for_each_netdev(net, dev) {
2874
		struct in_device *in_dev = __in_dev_get_rtnl(dev);
L
Linus Torvalds 已提交
2875
		if (in_dev && (dev->flags & IFF_UP)) {
2876
			struct in_ifaddr *ifa;
L
Linus Torvalds 已提交
2877 2878 2879 2880 2881 2882 2883 2884 2885 2886 2887 2888 2889 2890 2891

			int flag = scope;

			for (ifa = in_dev->ifa_list; ifa; ifa = ifa->ifa_next) {

				addr.s6_addr32[3] = ifa->ifa_local;

				if (ifa->ifa_scope == RT_SCOPE_LINK)
					continue;
				if (ifa->ifa_scope >= RT_SCOPE_HOST) {
					if (idev->dev->flags&IFF_POINTOPOINT)
						continue;
					flag |= IFA_HOST;
				}

2892
				add_addr(idev, &addr, plen, flag);
2893 2894
				addrconf_prefix_route(&addr, plen, idev->dev, 0,
						      pflags);
L
Linus Torvalds 已提交
2895 2896
			}
		}
2897
	}
L
Linus Torvalds 已提交
2898
}
2899
#endif
L
Linus Torvalds 已提交
2900 2901 2902 2903

static void init_loopback(struct net_device *dev)
{
	struct inet6_dev  *idev;
2904 2905 2906
	struct net_device *sp_dev;
	struct inet6_ifaddr *sp_ifa;
	struct rt6_info *sp_rt;
L
Linus Torvalds 已提交
2907 2908 2909 2910 2911

	/* ::1 */

	ASSERT_RTNL();

2912
	idev = ipv6_find_idev(dev);
2913
	if (!idev) {
2914
		pr_debug("%s: add_dev failed\n", __func__);
L
Linus Torvalds 已提交
2915 2916 2917
		return;
	}

2918
	add_addr(idev, &in6addr_loopback, 128, IFA_HOST);
2919 2920 2921 2922 2923 2924 2925 2926 2927 2928 2929 2930 2931 2932 2933 2934

	/* Add routes to other interface's IPv6 addresses */
	for_each_netdev(dev_net(dev), sp_dev) {
		if (!strcmp(sp_dev->name, dev->name))
			continue;

		idev = __in6_dev_get(sp_dev);
		if (!idev)
			continue;

		read_lock_bh(&idev->lock);
		list_for_each_entry(sp_ifa, &idev->addr_list, if_list) {

			if (sp_ifa->flags & (IFA_F_DADFAILED | IFA_F_TENTATIVE))
				continue;

2935 2936 2937 2938 2939 2940 2941 2942 2943 2944 2945 2946
			if (sp_ifa->rt) {
				/* This dst has been added to garbage list when
				 * lo device down, release this obsolete dst and
				 * reallocate a new router for ifa.
				 */
				if (sp_ifa->rt->dst.obsolete > 0) {
					ip6_rt_put(sp_ifa->rt);
					sp_ifa->rt = NULL;
				} else {
					continue;
				}
			}
2947

2948
			sp_rt = addrconf_dst_alloc(idev, &sp_ifa->addr, false);
2949 2950

			/* Failure cases are ignored */
2951 2952
			if (!IS_ERR(sp_rt)) {
				sp_ifa->rt = sp_rt;
2953
				ip6_ins_rt(sp_rt);
2954
			}
2955 2956 2957
		}
		read_unlock_bh(&idev->lock);
	}
L
Linus Torvalds 已提交
2958 2959
}

2960 2961
static void addrconf_add_linklocal(struct inet6_dev *idev,
				   const struct in6_addr *addr, u32 flags)
L
Linus Torvalds 已提交
2962
{
2963
	struct inet6_ifaddr *ifp;
2964
	u32 addr_flags = flags | IFA_F_PERMANENT;
2965 2966 2967

#ifdef CONFIG_IPV6_OPTIMISTIC_DAD
	if (idev->cnf.optimistic_dad &&
2968
	    !dev_net(idev->dev)->ipv6.devconf_all->forwarding)
2969 2970
		addr_flags |= IFA_F_OPTIMISTIC;
#endif
L
Linus Torvalds 已提交
2971

2972 2973
	ifp = ipv6_add_addr(idev, addr, NULL, 64, IFA_LINK, addr_flags,
			    INFINITY_LIFE_TIME, INFINITY_LIFE_TIME);
L
Linus Torvalds 已提交
2974
	if (!IS_ERR(ifp)) {
2975
		addrconf_prefix_route(&ifp->addr, ifp->prefix_len, idev->dev, 0, 0);
2976
		addrconf_dad_start(ifp);
L
Linus Torvalds 已提交
2977 2978 2979 2980
		in6_ifa_put(ifp);
	}
}

2981 2982 2983 2984 2985 2986 2987 2988 2989 2990 2991 2992 2993 2994 2995 2996 2997 2998 2999 3000 3001 3002 3003 3004 3005 3006 3007 3008
static bool ipv6_reserved_interfaceid(struct in6_addr address)
{
	if ((address.s6_addr32[2] | address.s6_addr32[3]) == 0)
		return true;

	if (address.s6_addr32[2] == htonl(0x02005eff) &&
	    ((address.s6_addr32[3] & htonl(0xfe000000)) == htonl(0xfe000000)))
		return true;

	if (address.s6_addr32[2] == htonl(0xfdffffff) &&
	    ((address.s6_addr32[3] & htonl(0xffffff80)) == htonl(0xffffff80)))
		return true;

	return false;
}

static int ipv6_generate_stable_address(struct in6_addr *address,
					u8 dad_count,
					const struct inet6_dev *idev)
{
	static DEFINE_SPINLOCK(lock);
	static __u32 digest[SHA_DIGEST_WORDS];
	static __u32 workspace[SHA_WORKSPACE_WORDS];

	static union {
		char __data[SHA_MESSAGE_BYTES];
		struct {
			struct in6_addr secret;
3009
			__be32 prefix[2];
3010 3011 3012 3013 3014 3015 3016 3017 3018 3019 3020 3021 3022 3023 3024 3025 3026 3027 3028 3029 3030 3031 3032 3033 3034
			unsigned char hwaddr[MAX_ADDR_LEN];
			u8 dad_count;
		} __packed;
	} data;

	struct in6_addr secret;
	struct in6_addr temp;
	struct net *net = dev_net(idev->dev);

	BUILD_BUG_ON(sizeof(data.__data) != sizeof(data));

	if (idev->cnf.stable_secret.initialized)
		secret = idev->cnf.stable_secret.secret;
	else if (net->ipv6.devconf_dflt->stable_secret.initialized)
		secret = net->ipv6.devconf_dflt->stable_secret.secret;
	else
		return -1;

retry:
	spin_lock_bh(&lock);

	sha_init(digest);
	memset(&data, 0, sizeof(data));
	memset(workspace, 0, sizeof(workspace));
	memcpy(data.hwaddr, idev->dev->perm_addr, idev->dev->addr_len);
3035 3036
	data.prefix[0] = address->s6_addr32[0];
	data.prefix[1] = address->s6_addr32[1];
3037 3038 3039 3040 3041 3042
	data.secret = secret;
	data.dad_count = dad_count;

	sha_transform(digest, data.__data, workspace);

	temp = *address;
3043 3044
	temp.s6_addr32[2] = (__force __be32)digest[0];
	temp.s6_addr32[3] = (__force __be32)digest[1];
3045 3046 3047 3048 3049

	spin_unlock_bh(&lock);

	if (ipv6_reserved_interfaceid(temp)) {
		dad_count++;
3050
		if (dad_count > dev_net(idev->dev)->ipv6.sysctl.idgen_retries)
3051 3052 3053 3054 3055 3056 3057 3058
			return -1;
		goto retry;
	}

	*address = temp;
	return 0;
}

3059 3060
static void addrconf_addr_gen(struct inet6_dev *idev, bool prefix_route)
{
3061 3062 3063
	struct in6_addr addr;

	ipv6_addr_set(&addr, htonl(0xFE800000), 0, 0, 0);
3064

3065 3066
	if (idev->addr_gen_mode == IN6_ADDR_GEN_MODE_STABLE_PRIVACY) {
		if (!ipv6_generate_stable_address(&addr, 0, idev))
3067 3068
			addrconf_add_linklocal(idev, &addr,
					       IFA_F_STABLE_PRIVACY);
3069 3070 3071
		else if (prefix_route)
			addrconf_prefix_route(&addr, 64, idev->dev, 0, 0);
	} else if (idev->addr_gen_mode == IN6_ADDR_GEN_MODE_EUI64) {
3072 3073 3074 3075 3076
		/* addrconf_add_linklocal also adds a prefix_route and we
		 * only need to care about prefix routes if ipv6_generate_eui64
		 * couldn't generate one.
		 */
		if (ipv6_generate_eui64(addr.s6_addr + 8, idev->dev) == 0)
3077
			addrconf_add_linklocal(idev, &addr, 0);
3078 3079 3080 3081 3082
		else if (prefix_route)
			addrconf_prefix_route(&addr, 64, idev->dev, 0, 0);
	}
}

L
Linus Torvalds 已提交
3083 3084
static void addrconf_dev_config(struct net_device *dev)
{
3085
	struct inet6_dev *idev;
L
Linus Torvalds 已提交
3086 3087 3088

	ASSERT_RTNL();

3089 3090 3091
	if ((dev->type != ARPHRD_ETHER) &&
	    (dev->type != ARPHRD_FDDI) &&
	    (dev->type != ARPHRD_ARCNET) &&
3092
	    (dev->type != ARPHRD_INFINIBAND) &&
3093
	    (dev->type != ARPHRD_IEEE802154) &&
3094
	    (dev->type != ARPHRD_IEEE1394) &&
3095 3096
	    (dev->type != ARPHRD_TUNNEL6) &&
	    (dev->type != ARPHRD_6LOWPAN)) {
3097 3098 3099 3100
		/* Alas, we support only Ethernet autoconfiguration. */
		return;
	}

L
Linus Torvalds 已提交
3101
	idev = addrconf_add_dev(dev);
3102
	if (IS_ERR(idev))
L
Linus Torvalds 已提交
3103 3104
		return;

3105
	addrconf_addr_gen(idev, false);
L
Linus Torvalds 已提交
3106 3107
}

A
Amerigo Wang 已提交
3108
#if IS_ENABLED(CONFIG_IPV6_SIT)
L
Linus Torvalds 已提交
3109 3110 3111 3112 3113 3114
static void addrconf_sit_config(struct net_device *dev)
{
	struct inet6_dev *idev;

	ASSERT_RTNL();

3115 3116 3117
	/*
	 * Configure the tunnel with one of our IPv4
	 * addresses... we should configure all of
L
Linus Torvalds 已提交
3118 3119 3120
	 * our v4 addrs in the tunnel
	 */

3121
	idev = ipv6_find_idev(dev);
3122
	if (!idev) {
3123
		pr_debug("%s: add_dev failed\n", __func__);
L
Linus Torvalds 已提交
3124 3125 3126
		return;
	}

F
Fred L. Templin 已提交
3127
	if (dev->priv_flags & IFF_ISATAP) {
3128
		addrconf_addr_gen(idev, false);
F
Fred L. Templin 已提交
3129 3130 3131
		return;
	}

L
Linus Torvalds 已提交
3132 3133
	sit_add_v4_addrs(idev);

3134
	if (dev->flags&IFF_POINTOPOINT)
L
Linus Torvalds 已提交
3135 3136
		addrconf_add_mroute(dev);
}
3137
#endif
L
Linus Torvalds 已提交
3138

A
Amerigo Wang 已提交
3139
#if IS_ENABLED(CONFIG_NET_IPGRE)
3140 3141 3142 3143 3144 3145
static void addrconf_gre_config(struct net_device *dev)
{
	struct inet6_dev *idev;

	ASSERT_RTNL();

3146
	idev = ipv6_find_idev(dev);
3147
	if (!idev) {
3148
		pr_debug("%s: add_dev failed\n", __func__);
3149 3150 3151
		return;
	}

3152
	addrconf_addr_gen(idev, true);
3153 3154 3155
}
#endif

3156
static int addrconf_notify(struct notifier_block *this, unsigned long event,
3157
			   void *ptr)
L
Linus Torvalds 已提交
3158
{
3159
	struct net_device *dev = netdev_notifier_info_to_dev(ptr);
3160
	struct inet6_dev *idev = __in6_dev_get(dev);
3161
	int run_pending = 0;
3162
	int err;
L
Linus Torvalds 已提交
3163

S
Stephen Hemminger 已提交
3164
	switch (event) {
3165
	case NETDEV_REGISTER:
3166
		if (!idev && dev->mtu >= IPV6_MIN_MTU) {
3167
			idev = ipv6_add_dev(dev);
3168 3169
			if (IS_ERR(idev))
				return notifier_from_errno(PTR_ERR(idev));
3170 3171
		}
		break;
S
Stephen Hemminger 已提交
3172

L
Linus Torvalds 已提交
3173
	case NETDEV_UP:
3174
	case NETDEV_CHANGE:
3175 3176 3177
		if (dev->flags & IFF_SLAVE)
			break;

3178 3179 3180
		if (idev && idev->cnf.disable_ipv6)
			break;

3181
		if (event == NETDEV_UP) {
3182
			if (!addrconf_qdisc_ok(dev)) {
3183
				/* device is not ready yet. */
3184
				pr_info("ADDRCONF(NETDEV_UP): %s: link is not ready\n",
3185 3186 3187
					dev->name);
				break;
			}
3188

3189 3190 3191
			if (!idev && dev->mtu >= IPV6_MIN_MTU)
				idev = ipv6_add_dev(dev);

3192
			if (!IS_ERR_OR_NULL(idev)) {
3193
				idev->if_flags |= IF_READY;
3194 3195
				run_pending = 1;
			}
3196
		} else {
3197
			if (!addrconf_qdisc_ok(dev)) {
3198 3199 3200 3201 3202
				/* device is still not ready. */
				break;
			}

			if (idev) {
S
Stephen Hemminger 已提交
3203
				if (idev->if_flags & IF_READY)
3204 3205 3206 3207 3208
					/* device is already configured. */
					break;
				idev->if_flags |= IF_READY;
			}

3209 3210
			pr_info("ADDRCONF(NETDEV_CHANGE): %s: link becomes ready\n",
				dev->name);
3211

3212
			run_pending = 1;
3213 3214
		}

S
Stephen Hemminger 已提交
3215
		switch (dev->type) {
A
Amerigo Wang 已提交
3216
#if IS_ENABLED(CONFIG_IPV6_SIT)
L
Linus Torvalds 已提交
3217 3218 3219
		case ARPHRD_SIT:
			addrconf_sit_config(dev);
			break;
3220
#endif
A
Amerigo Wang 已提交
3221
#if IS_ENABLED(CONFIG_NET_IPGRE)
3222 3223 3224
		case ARPHRD_IPGRE:
			addrconf_gre_config(dev);
			break;
3225
#endif
L
Linus Torvalds 已提交
3226 3227 3228 3229 3230 3231 3232
		case ARPHRD_LOOPBACK:
			init_loopback(dev);
			break;

		default:
			addrconf_dev_config(dev);
			break;
3233
		}
S
Stephen Hemminger 已提交
3234

3235
		if (!IS_ERR_OR_NULL(idev)) {
3236 3237 3238
			if (run_pending)
				addrconf_dad_run(idev);

S
Stephen Hemminger 已提交
3239 3240 3241 3242
			/*
			 * If the MTU changed during the interface down,
			 * when the interface up, the changed MTU must be
			 * reflected in the idev as well as routers.
L
Linus Torvalds 已提交
3243
			 */
S
Stephen Hemminger 已提交
3244 3245
			if (idev->cnf.mtu6 != dev->mtu &&
			    dev->mtu >= IPV6_MIN_MTU) {
L
Linus Torvalds 已提交
3246 3247 3248 3249 3250
				rt6_mtu_change(dev, dev->mtu);
				idev->cnf.mtu6 = dev->mtu;
			}
			idev->tstamp = jiffies;
			inet6_ifinfo_notify(RTM_NEWLINK, idev);
S
Stephen Hemminger 已提交
3251 3252 3253 3254

			/*
			 * If the changed mtu during down is lower than
			 * IPV6_MIN_MTU stop IPv6 on this interface.
L
Linus Torvalds 已提交
3255 3256
			 */
			if (dev->mtu < IPV6_MIN_MTU)
S
Stephen Hemminger 已提交
3257
				addrconf_ifdown(dev, 1);
L
Linus Torvalds 已提交
3258 3259 3260 3261
		}
		break;

	case NETDEV_CHANGEMTU:
3262
		if (idev && dev->mtu >= IPV6_MIN_MTU) {
L
Linus Torvalds 已提交
3263 3264 3265 3266 3267
			rt6_mtu_change(dev, dev->mtu);
			idev->cnf.mtu6 = dev->mtu;
			break;
		}

3268 3269
		if (!idev && dev->mtu >= IPV6_MIN_MTU) {
			idev = ipv6_add_dev(dev);
3270
			if (!IS_ERR(idev))
3271 3272 3273
				break;
		}

S
Stephen Hemminger 已提交
3274
		/*
3275
		 * if MTU under IPV6_MIN_MTU.
S
Stephen Hemminger 已提交
3276 3277
		 * Stop IPv6 on this interface.
		 */
L
Linus Torvalds 已提交
3278 3279 3280 3281 3282 3283 3284 3285

	case NETDEV_DOWN:
	case NETDEV_UNREGISTER:
		/*
		 *	Remove all addresses from this interface.
		 */
		addrconf_ifdown(dev, event != NETDEV_DOWN);
		break;
3286

L
Linus Torvalds 已提交
3287 3288
	case NETDEV_CHANGENAME:
		if (idev) {
3289
			snmp6_unregister_dev(idev);
3290
			addrconf_sysctl_unregister(idev);
3291
			err = addrconf_sysctl_register(idev);
3292 3293
			if (err)
				return notifier_from_errno(err);
3294 3295 3296 3297 3298
			err = snmp6_register_dev(idev);
			if (err) {
				addrconf_sysctl_unregister(idev);
				return notifier_from_errno(err);
			}
3299
		}
L
Linus Torvalds 已提交
3300
		break;
S
Stephen Hemminger 已提交
3301

3302 3303 3304
	case NETDEV_PRE_TYPE_CHANGE:
	case NETDEV_POST_TYPE_CHANGE:
		addrconf_type_change(dev, event);
3305
		break;
3306
	}
L
Linus Torvalds 已提交
3307 3308 3309 3310 3311 3312 3313 3314 3315 3316 3317

	return NOTIFY_OK;
}

/*
 *	addrconf module should be notified of a device going up
 */
static struct notifier_block ipv6_dev_notf = {
	.notifier_call = addrconf_notify,
};

3318
static void addrconf_type_change(struct net_device *dev, unsigned long event)
3319 3320 3321 3322 3323 3324
{
	struct inet6_dev *idev;
	ASSERT_RTNL();

	idev = __in6_dev_get(dev);

3325
	if (event == NETDEV_POST_TYPE_CHANGE)
3326
		ipv6_mc_remap(idev);
3327
	else if (event == NETDEV_PRE_TYPE_CHANGE)
3328 3329 3330
		ipv6_mc_unmap(idev);
}

L
Linus Torvalds 已提交
3331 3332
static int addrconf_ifdown(struct net_device *dev, int how)
{
3333
	struct net *net = dev_net(dev);
3334 3335
	struct inet6_dev *idev;
	struct inet6_ifaddr *ifa;
3336
	int state, i;
L
Linus Torvalds 已提交
3337 3338 3339

	ASSERT_RTNL();

3340 3341
	rt6_ifdown(net, dev);
	neigh_ifdown(&nd_tbl, dev);
L
Linus Torvalds 已提交
3342 3343

	idev = __in6_dev_get(dev);
3344
	if (!idev)
L
Linus Torvalds 已提交
3345 3346
		return -ENODEV;

S
Stephen Hemminger 已提交
3347 3348 3349
	/*
	 * Step 1: remove reference to ipv6 device from parent device.
	 *	   Do not dev_put!
L
Linus Torvalds 已提交
3350
	 */
3351
	if (how) {
L
Linus Torvalds 已提交
3352
		idev->dead = 1;
3353 3354

		/* protected by rtnl_lock */
3355
		RCU_INIT_POINTER(dev->ip6_ptr, NULL);
L
Linus Torvalds 已提交
3356 3357 3358 3359 3360 3361

		/* Step 1.5: remove snmp6 entry */
		snmp6_unregister_dev(idev);

	}

3362 3363 3364 3365 3366
	/* Step 2: clear hash table */
	for (i = 0; i < IN6_ADDR_HSIZE; i++) {
		struct hlist_head *h = &inet6_addr_lst[i];

		spin_lock_bh(&addrconf_hash_lock);
3367
restart:
3368
		hlist_for_each_entry_rcu(ifa, h, addr_lst) {
3369 3370
			if (ifa->idev == idev) {
				hlist_del_init_rcu(&ifa->addr_lst);
3371
				addrconf_del_dad_work(ifa);
3372 3373 3374 3375 3376 3377
				goto restart;
			}
		}
		spin_unlock_bh(&addrconf_hash_lock);
	}

L
Linus Torvalds 已提交
3378 3379
	write_lock_bh(&idev->lock);

3380 3381
	addrconf_del_rs_timer(idev);

S
Stephen Hemminger 已提交
3382
	/* Step 2: clear flags for stateless addrconf */
3383
	if (!how)
3384
		idev->if_flags &= ~(IF_RS_SENT|IF_RA_RCVD|IF_READY);
L
Linus Torvalds 已提交
3385

3386
	if (how && del_timer(&idev->regen_timer))
L
Linus Torvalds 已提交
3387 3388
		in6_dev_put(idev);

S
Stephen Hemminger 已提交
3389
	/* Step 3: clear tempaddr list */
3390 3391 3392 3393
	while (!list_empty(&idev->tempaddr_list)) {
		ifa = list_first_entry(&idev->tempaddr_list,
				       struct inet6_ifaddr, tmp_list);
		list_del(&ifa->tmp_list);
L
Linus Torvalds 已提交
3394 3395 3396 3397 3398 3399 3400 3401 3402 3403 3404
		write_unlock_bh(&idev->lock);
		spin_lock_bh(&ifa->lock);

		if (ifa->ifpub) {
			in6_ifa_put(ifa->ifpub);
			ifa->ifpub = NULL;
		}
		spin_unlock_bh(&ifa->lock);
		in6_ifa_put(ifa);
		write_lock_bh(&idev->lock);
	}
3405

3406 3407 3408
	while (!list_empty(&idev->addr_list)) {
		ifa = list_first_entry(&idev->addr_list,
				       struct inet6_ifaddr, if_list);
3409
		addrconf_del_dad_work(ifa);
3410

3411
		list_del(&ifa->if_list);
3412

3413
		write_unlock_bh(&idev->lock);
3414

3415
		spin_lock_bh(&ifa->lock);
3416 3417
		state = ifa->state;
		ifa->state = INET6_IFADDR_STATE_DEAD;
3418
		spin_unlock_bh(&ifa->lock);
3419

3420 3421
		if (state != INET6_IFADDR_STATE_DEAD) {
			__ipv6_ifa_notify(RTM_DELADDR, ifa);
3422
			inet6addr_notifier_call_chain(NETDEV_DOWN, ifa);
3423
		}
3424
		in6_ifa_put(ifa);
3425

3426 3427
		write_lock_bh(&idev->lock);
	}
3428

L
Linus Torvalds 已提交
3429 3430
	write_unlock_bh(&idev->lock);

3431 3432 3433
	/* Step 5: Discard anycast and multicast list */
	if (how) {
		ipv6_ac_destroy_dev(idev);
L
Linus Torvalds 已提交
3434
		ipv6_mc_destroy_dev(idev);
3435
	} else {
L
Linus Torvalds 已提交
3436
		ipv6_mc_down(idev);
3437
	}
L
Linus Torvalds 已提交
3438 3439

	idev->tstamp = jiffies;
3440

S
Stephen Hemminger 已提交
3441
	/* Last: Shot the device (if unregistered) */
3442
	if (how) {
3443
		addrconf_sysctl_unregister(idev);
L
Linus Torvalds 已提交
3444 3445 3446 3447 3448 3449 3450 3451 3452
		neigh_parms_release(&nd_tbl, idev->nd_parms);
		neigh_ifdown(&nd_tbl, dev);
		in6_dev_put(idev);
	}
	return 0;
}

static void addrconf_rs_timer(unsigned long data)
{
3453
	struct inet6_dev *idev = (struct inet6_dev *)data;
3454
	struct net_device *dev = idev->dev;
3455
	struct in6_addr lladdr;
L
Linus Torvalds 已提交
3456

3457
	write_lock(&idev->lock);
S
stephen hemminger 已提交
3458
	if (idev->dead || !(idev->if_flags & IF_READY))
L
Linus Torvalds 已提交
3459 3460
		goto out;

3461
	if (!ipv6_accept_ra(idev))
S
stephen hemminger 已提交
3462 3463 3464 3465
		goto out;

	/* Announcement received after solicitation was sent */
	if (idev->if_flags & IF_RA_RCVD)
L
Linus Torvalds 已提交
3466 3467
		goto out;

3468
	if (idev->rs_probes++ < idev->cnf.rtr_solicits) {
3469 3470 3471
		write_unlock(&idev->lock);
		if (!ipv6_get_lladdr(dev, &lladdr, IFA_F_TENTATIVE))
			ndisc_send_rs(dev, &lladdr,
3472 3473
				      &in6addr_linklocal_allrouters);
		else
3474
			goto put;
L
Linus Torvalds 已提交
3475

3476
		write_lock(&idev->lock);
3477 3478 3479 3480 3481
		/* The wait after the last probe can be shorter */
		addrconf_mod_rs_timer(idev, (idev->rs_probes ==
					     idev->cnf.rtr_solicits) ?
				      idev->cnf.rtr_solicit_delay :
				      idev->cnf.rtr_solicit_interval);
L
Linus Torvalds 已提交
3482 3483 3484 3485 3486
	} else {
		/*
		 * Note: we do not support deprecated "all on-link"
		 * assumption any longer.
		 */
3487
		pr_debug("%s: no IPv6 routers present\n", idev->dev->name);
L
Linus Torvalds 已提交
3488 3489 3490
	}

out:
3491
	write_unlock(&idev->lock);
3492
put:
3493
	in6_dev_put(idev);
L
Linus Torvalds 已提交
3494 3495 3496 3497 3498
}

/*
 *	Duplicate Address Detection
 */
3499 3500 3501 3502 3503
static void addrconf_dad_kick(struct inet6_ifaddr *ifp)
{
	unsigned long rand_num;
	struct inet6_dev *idev = ifp->idev;

3504 3505 3506
	if (ifp->flags & IFA_F_OPTIMISTIC)
		rand_num = 0;
	else
3507
		rand_num = prandom_u32() % (idev->cnf.rtr_solicit_delay ? : 1);
3508

3509
	ifp->dad_probes = idev->cnf.dad_transmits;
3510
	addrconf_mod_dad_work(ifp, rand_num);
3511 3512
}

3513
static void addrconf_dad_begin(struct inet6_ifaddr *ifp)
L
Linus Torvalds 已提交
3514 3515 3516 3517 3518 3519
{
	struct inet6_dev *idev = ifp->idev;
	struct net_device *dev = idev->dev;

	addrconf_join_solict(dev, &ifp->addr);

3520
	prandom_seed((__force u32) ifp->addr.s6_addr32[3]);
L
Linus Torvalds 已提交
3521 3522

	read_lock_bh(&idev->lock);
3523
	spin_lock(&ifp->lock);
3524
	if (ifp->state == INET6_IFADDR_STATE_DEAD)
L
Linus Torvalds 已提交
3525 3526 3527
		goto out;

	if (dev->flags&(IFF_NOARP|IFF_LOOPBACK) ||
3528
	    idev->cnf.accept_dad < 1 ||
3529 3530
	    !(ifp->flags&IFA_F_TENTATIVE) ||
	    ifp->flags & IFA_F_NODAD) {
B
Brian Haley 已提交
3531
		ifp->flags &= ~(IFA_F_TENTATIVE|IFA_F_OPTIMISTIC|IFA_F_DADFAILED);
3532
		spin_unlock(&ifp->lock);
L
Linus Torvalds 已提交
3533 3534 3535 3536 3537 3538
		read_unlock_bh(&idev->lock);

		addrconf_dad_completed(ifp);
		return;
	}

3539
	if (!(idev->if_flags & IF_READY)) {
3540
		spin_unlock(&ifp->lock);
3541
		read_unlock_bh(&idev->lock);
3542
		/*
3543
		 * If the device is not ready:
3544 3545 3546 3547
		 * - keep it tentative if it is a permanent address.
		 * - otherwise, kill it.
		 */
		in6_ifa_hold(ifp);
B
Brian Haley 已提交
3548
		addrconf_dad_stop(ifp, 0);
3549
		return;
3550
	}
3551 3552 3553 3554 3555

	/*
	 * Optimistic nodes can start receiving
	 * Frames right away
	 */
3556
	if (ifp->flags & IFA_F_OPTIMISTIC) {
3557
		ip6_ins_rt(ifp->rt);
3558 3559 3560 3561 3562 3563 3564
		if (ipv6_use_optimistic_addr(idev)) {
			/* Because optimistic nodes can use this address,
			 * notify listeners. If DAD fails, RTM_DELADDR is sent.
			 */
			ipv6_ifa_notify(RTM_NEWADDR, ifp);
		}
	}
3565

3566
	addrconf_dad_kick(ifp);
L
Linus Torvalds 已提交
3567
out:
3568
	spin_unlock(&ifp->lock);
L
Linus Torvalds 已提交
3569 3570 3571
	read_unlock_bh(&idev->lock);
}

3572
static void addrconf_dad_start(struct inet6_ifaddr *ifp)
L
Linus Torvalds 已提交
3573
{
3574 3575
	bool begin_dad = false;

3576
	spin_lock_bh(&ifp->lock);
3577 3578 3579 3580
	if (ifp->state != INET6_IFADDR_STATE_DEAD) {
		ifp->state = INET6_IFADDR_STATE_PREDAD;
		begin_dad = true;
	}
3581
	spin_unlock_bh(&ifp->lock);
3582 3583 3584 3585 3586 3587 3588 3589 3590 3591

	if (begin_dad)
		addrconf_mod_dad_work(ifp, 0);
}

static void addrconf_dad_work(struct work_struct *w)
{
	struct inet6_ifaddr *ifp = container_of(to_delayed_work(w),
						struct inet6_ifaddr,
						dad_work);
L
Linus Torvalds 已提交
3592 3593 3594
	struct inet6_dev *idev = ifp->idev;
	struct in6_addr mcaddr;

3595 3596 3597 3598 3599 3600 3601 3602
	enum {
		DAD_PROCESS,
		DAD_BEGIN,
		DAD_ABORT,
	} action = DAD_PROCESS;

	rtnl_lock();

3603
	spin_lock_bh(&ifp->lock);
3604 3605 3606 3607 3608 3609 3610
	if (ifp->state == INET6_IFADDR_STATE_PREDAD) {
		action = DAD_BEGIN;
		ifp->state = INET6_IFADDR_STATE_DAD;
	} else if (ifp->state == INET6_IFADDR_STATE_ERRDAD) {
		action = DAD_ABORT;
		ifp->state = INET6_IFADDR_STATE_POSTDAD;
	}
3611
	spin_unlock_bh(&ifp->lock);
3612 3613 3614 3615 3616 3617 3618 3619 3620

	if (action == DAD_BEGIN) {
		addrconf_dad_begin(ifp);
		goto out;
	} else if (action == DAD_ABORT) {
		addrconf_dad_stop(ifp, 1);
		goto out;
	}

3621
	if (!ifp->dad_probes && addrconf_dad_end(ifp))
H
Herbert Xu 已提交
3622 3623
		goto out;

3624
	write_lock_bh(&idev->lock);
3625
	if (idev->dead || !(idev->if_flags & IF_READY)) {
3626
		write_unlock_bh(&idev->lock);
L
Linus Torvalds 已提交
3627 3628
		goto out;
	}
3629 3630

	spin_lock(&ifp->lock);
3631 3632
	if (ifp->state == INET6_IFADDR_STATE_DEAD) {
		spin_unlock(&ifp->lock);
3633
		write_unlock_bh(&idev->lock);
3634 3635 3636
		goto out;
	}

3637
	if (ifp->dad_probes == 0) {
L
Linus Torvalds 已提交
3638 3639 3640 3641
		/*
		 * DAD was successful
		 */

B
Brian Haley 已提交
3642
		ifp->flags &= ~(IFA_F_TENTATIVE|IFA_F_OPTIMISTIC|IFA_F_DADFAILED);
3643
		spin_unlock(&ifp->lock);
3644
		write_unlock_bh(&idev->lock);
L
Linus Torvalds 已提交
3645 3646 3647 3648 3649 3650

		addrconf_dad_completed(ifp);

		goto out;
	}

3651
	ifp->dad_probes--;
3652 3653
	addrconf_mod_dad_work(ifp,
			      NEIGH_VAR(ifp->idev->nd_parms, RETRANS_TIME));
3654
	spin_unlock(&ifp->lock);
3655
	write_unlock_bh(&idev->lock);
L
Linus Torvalds 已提交
3656 3657 3658

	/* send a neighbour solicitation for our addr */
	addrconf_addr_solict_mult(&ifp->addr, &mcaddr);
3659
	ndisc_send_ns(ifp->idev->dev, NULL, &ifp->addr, &mcaddr, &in6addr_any);
L
Linus Torvalds 已提交
3660 3661
out:
	in6_ifa_put(ifp);
3662
	rtnl_unlock();
L
Linus Torvalds 已提交
3663 3664
}

3665 3666 3667 3668 3669 3670
/* ifp->idev must be at least read locked */
static bool ipv6_lonely_lladdr(struct inet6_ifaddr *ifp)
{
	struct inet6_ifaddr *ifpiter;
	struct inet6_dev *idev = ifp->idev;

3671 3672 3673
	list_for_each_entry_reverse(ifpiter, &idev->addr_list, if_list) {
		if (ifpiter->scope > IFA_LINK)
			break;
3674 3675 3676 3677 3678 3679 3680 3681 3682
		if (ifp != ifpiter && ifpiter->scope == IFA_LINK &&
		    (ifpiter->flags & (IFA_F_PERMANENT|IFA_F_TENTATIVE|
				       IFA_F_OPTIMISTIC|IFA_F_DADFAILED)) ==
		    IFA_F_PERMANENT)
			return false;
	}
	return true;
}

L
Linus Torvalds 已提交
3683 3684
static void addrconf_dad_completed(struct inet6_ifaddr *ifp)
{
S
Stephen Hemminger 已提交
3685
	struct net_device *dev = ifp->idev->dev;
3686
	struct in6_addr lladdr;
3687
	bool send_rs, send_mld;
3688

3689
	addrconf_del_dad_work(ifp);
L
Linus Torvalds 已提交
3690 3691 3692 3693 3694 3695 3696

	/*
	 *	Configure the address for reception. Now it is valid.
	 */

	ipv6_ifa_notify(RTM_NEWADDR, ifp);

3697 3698
	/* If added prefix is link local and we are prepared to process
	   router advertisements, start sending router solicitations.
L
Linus Torvalds 已提交
3699 3700
	 */

3701
	read_lock_bh(&ifp->idev->lock);
3702
	send_mld = ifp->scope == IFA_LINK && ipv6_lonely_lladdr(ifp);
3703 3704
	send_rs = send_mld &&
		  ipv6_accept_ra(ifp->idev) &&
3705
		  ifp->idev->cnf.rtr_solicits > 0 &&
3706
		  (dev->flags&IFF_LOOPBACK) == 0;
3707 3708
	read_unlock_bh(&ifp->idev->lock);

3709 3710 3711 3712 3713 3714
	/* While dad is in progress mld report's source address is in6_addrany.
	 * Resend with proper ll now.
	 */
	if (send_mld)
		ipv6_mc_dad_complete(ifp->idev);

3715
	if (send_rs) {
L
Linus Torvalds 已提交
3716 3717 3718 3719 3720
		/*
		 *	If a host as already performed a random delay
		 *	[...] as part of DAD [...] there is no need
		 *	to delay again before sending the first RS
		 */
3721
		if (ipv6_get_lladdr(dev, &lladdr, IFA_F_TENTATIVE))
3722
			return;
3723
		ndisc_send_rs(dev, &lladdr, &in6addr_linklocal_allrouters);
L
Linus Torvalds 已提交
3724

3725 3726 3727
		write_lock_bh(&ifp->idev->lock);
		spin_lock(&ifp->lock);
		ifp->idev->rs_probes = 1;
L
Linus Torvalds 已提交
3728
		ifp->idev->if_flags |= IF_RS_SENT;
3729 3730 3731 3732
		addrconf_mod_rs_timer(ifp->idev,
				      ifp->idev->cnf.rtr_solicit_interval);
		spin_unlock(&ifp->lock);
		write_unlock_bh(&ifp->idev->lock);
L
Linus Torvalds 已提交
3733 3734 3735
	}
}

S
Stephen Hemminger 已提交
3736 3737
static void addrconf_dad_run(struct inet6_dev *idev)
{
3738 3739 3740
	struct inet6_ifaddr *ifp;

	read_lock_bh(&idev->lock);
3741
	list_for_each_entry(ifp, &idev->addr_list, if_list) {
3742
		spin_lock(&ifp->lock);
H
Herbert Xu 已提交
3743 3744 3745
		if (ifp->flags & IFA_F_TENTATIVE &&
		    ifp->state == INET6_IFADDR_STATE_DAD)
			addrconf_dad_kick(ifp);
3746
		spin_unlock(&ifp->lock);
3747 3748 3749 3750
	}
	read_unlock_bh(&idev->lock);
}

L
Linus Torvalds 已提交
3751 3752
#ifdef CONFIG_PROC_FS
struct if6_iter_state {
3753
	struct seq_net_private p;
L
Linus Torvalds 已提交
3754
	int bucket;
3755
	int offset;
L
Linus Torvalds 已提交
3756 3757
};

3758
static struct inet6_ifaddr *if6_get_first(struct seq_file *seq, loff_t pos)
L
Linus Torvalds 已提交
3759 3760 3761
{
	struct inet6_ifaddr *ifa = NULL;
	struct if6_iter_state *state = seq->private;
3762
	struct net *net = seq_file_net(seq);
3763
	int p = 0;
L
Linus Torvalds 已提交
3764

3765 3766 3767 3768 3769 3770 3771
	/* initial bucket if pos is 0 */
	if (pos == 0) {
		state->bucket = 0;
		state->offset = 0;
	}

	for (; state->bucket < IN6_ADDR_HSIZE; ++state->bucket) {
3772
		hlist_for_each_entry_rcu_bh(ifa, &inet6_addr_lst[state->bucket],
3773
					 addr_lst) {
3774 3775
			if (!net_eq(dev_net(ifa->idev->dev), net))
				continue;
3776 3777 3778 3779 3780 3781
			/* sync with offset */
			if (p < state->offset) {
				p++;
				continue;
			}
			state->offset++;
3782
			return ifa;
3783 3784 3785 3786 3787
		}

		/* prepare for next bucket */
		state->offset = 0;
		p = 0;
L
Linus Torvalds 已提交
3788
	}
3789
	return NULL;
L
Linus Torvalds 已提交
3790 3791
}

3792 3793
static struct inet6_ifaddr *if6_get_next(struct seq_file *seq,
					 struct inet6_ifaddr *ifa)
L
Linus Torvalds 已提交
3794 3795
{
	struct if6_iter_state *state = seq->private;
3796
	struct net *net = seq_file_net(seq);
L
Linus Torvalds 已提交
3797

3798
	hlist_for_each_entry_continue_rcu_bh(ifa, addr_lst) {
3799 3800
		if (!net_eq(dev_net(ifa->idev->dev), net))
			continue;
3801
		state->offset++;
3802
		return ifa;
3803
	}
3804

3805
	while (++state->bucket < IN6_ADDR_HSIZE) {
3806
		state->offset = 0;
3807
		hlist_for_each_entry_rcu_bh(ifa,
3808
				     &inet6_addr_lst[state->bucket], addr_lst) {
3809 3810
			if (!net_eq(dev_net(ifa->idev->dev), net))
				continue;
3811
			state->offset++;
3812
			return ifa;
3813
		}
L
Linus Torvalds 已提交
3814
	}
3815

3816
	return NULL;
L
Linus Torvalds 已提交
3817 3818 3819
}

static void *if6_seq_start(struct seq_file *seq, loff_t *pos)
E
Eric Dumazet 已提交
3820
	__acquires(rcu_bh)
L
Linus Torvalds 已提交
3821
{
3822
	rcu_read_lock_bh();
3823
	return if6_get_first(seq, *pos);
L
Linus Torvalds 已提交
3824 3825 3826 3827 3828 3829 3830 3831 3832 3833 3834 3835
}

static void *if6_seq_next(struct seq_file *seq, void *v, loff_t *pos)
{
	struct inet6_ifaddr *ifa;

	ifa = if6_get_next(seq, v);
	++*pos;
	return ifa;
}

static void if6_seq_stop(struct seq_file *seq, void *v)
E
Eric Dumazet 已提交
3836
	__releases(rcu_bh)
L
Linus Torvalds 已提交
3837
{
3838
	rcu_read_unlock_bh();
L
Linus Torvalds 已提交
3839 3840 3841 3842 3843
}

static int if6_seq_show(struct seq_file *seq, void *v)
{
	struct inet6_ifaddr *ifp = (struct inet6_ifaddr *)v;
3844
	seq_printf(seq, "%pi6 %02x %02x %02x %02x %8s\n",
3845
		   &ifp->addr,
L
Linus Torvalds 已提交
3846 3847 3848
		   ifp->idev->dev->ifindex,
		   ifp->prefix_len,
		   ifp->scope,
3849
		   (u8) ifp->flags,
L
Linus Torvalds 已提交
3850 3851 3852 3853
		   ifp->idev->dev->name);
	return 0;
}

3854
static const struct seq_operations if6_seq_ops = {
L
Linus Torvalds 已提交
3855 3856 3857 3858 3859 3860 3861 3862
	.start	= if6_seq_start,
	.next	= if6_seq_next,
	.show	= if6_seq_show,
	.stop	= if6_seq_stop,
};

static int if6_seq_open(struct inode *inode, struct file *file)
{
3863 3864
	return seq_open_net(inode, file, &if6_seq_ops,
			    sizeof(struct if6_iter_state));
L
Linus Torvalds 已提交
3865 3866
}

3867
static const struct file_operations if6_fops = {
L
Linus Torvalds 已提交
3868 3869 3870 3871
	.owner		= THIS_MODULE,
	.open		= if6_seq_open,
	.read		= seq_read,
	.llseek		= seq_lseek,
3872
	.release	= seq_release_net,
L
Linus Torvalds 已提交
3873 3874
};

3875
static int __net_init if6_proc_net_init(struct net *net)
L
Linus Torvalds 已提交
3876
{
3877
	if (!proc_create("if_inet6", S_IRUGO, net->proc_net, &if6_fops))
L
Linus Torvalds 已提交
3878 3879 3880 3881
		return -ENOMEM;
	return 0;
}

3882
static void __net_exit if6_proc_net_exit(struct net *net)
3883
{
3884
	remove_proc_entry("if_inet6", net->proc_net);
3885 3886 3887
}

static struct pernet_operations if6_proc_net_ops = {
3888 3889
	.init = if6_proc_net_init,
	.exit = if6_proc_net_exit,
3890 3891 3892 3893 3894 3895 3896
};

int __init if6_proc_init(void)
{
	return register_pernet_subsys(&if6_proc_net_ops);
}

L
Linus Torvalds 已提交
3897 3898
void if6_proc_exit(void)
{
3899
	unregister_pernet_subsys(&if6_proc_net_ops);
L
Linus Torvalds 已提交
3900 3901 3902
}
#endif	/* CONFIG_PROC_FS */

A
Amerigo Wang 已提交
3903
#if IS_ENABLED(CONFIG_IPV6_MIP6)
3904
/* Check if address is a home address configured on any interface. */
3905
int ipv6_chk_home_addr(struct net *net, const struct in6_addr *addr)
3906 3907
{
	int ret = 0;
3908
	struct inet6_ifaddr *ifp = NULL;
E
Eric Dumazet 已提交
3909
	unsigned int hash = inet6_addr_hash(addr);
3910

3911
	rcu_read_lock_bh();
3912
	hlist_for_each_entry_rcu_bh(ifp, &inet6_addr_lst[hash], addr_lst) {
3913
		if (!net_eq(dev_net(ifp->idev->dev), net))
3914
			continue;
3915
		if (ipv6_addr_equal(&ifp->addr, addr) &&
3916 3917 3918 3919 3920
		    (ifp->flags & IFA_F_HOMEADDRESS)) {
			ret = 1;
			break;
		}
	}
3921
	rcu_read_unlock_bh();
3922 3923 3924 3925
	return ret;
}
#endif

L
Linus Torvalds 已提交
3926 3927 3928 3929
/*
 *	Periodic address status verification
 */

3930
static void addrconf_verify_rtnl(void)
L
Linus Torvalds 已提交
3931
{
3932
	unsigned long now, next, next_sec, next_sched;
L
Linus Torvalds 已提交
3933 3934 3935
	struct inet6_ifaddr *ifp;
	int i;

3936 3937
	ASSERT_RTNL();

3938
	rcu_read_lock_bh();
L
Linus Torvalds 已提交
3939
	now = jiffies;
3940
	next = round_jiffies_up(now + ADDR_CHECK_FREQUENCY);
L
Linus Torvalds 已提交
3941

3942
	cancel_delayed_work(&addr_chk_work);
L
Linus Torvalds 已提交
3943

S
Stephen Hemminger 已提交
3944
	for (i = 0; i < IN6_ADDR_HSIZE; i++) {
L
Linus Torvalds 已提交
3945
restart:
3946
		hlist_for_each_entry_rcu_bh(ifp, &inet6_addr_lst[i], addr_lst) {
L
Linus Torvalds 已提交
3947 3948
			unsigned long age;

3949 3950 3951 3952 3953 3954
			/* When setting preferred_lft to a value not zero or
			 * infinity, while valid_lft is infinity
			 * IFA_F_PERMANENT has a non-infinity life time.
			 */
			if ((ifp->flags & IFA_F_PERMANENT) &&
			    (ifp->prefered_lft == INFINITY_LIFE_TIME))
L
Linus Torvalds 已提交
3955 3956 3957
				continue;

			spin_lock(&ifp->lock);
3958 3959
			/* We try to batch several events at once. */
			age = (now - ifp->tstamp + ADDRCONF_TIMER_FUZZ_MINUS) / HZ;
L
Linus Torvalds 已提交
3960

3961 3962
			if (ifp->valid_lft != INFINITY_LIFE_TIME &&
			    age >= ifp->valid_lft) {
L
Linus Torvalds 已提交
3963 3964 3965 3966
				spin_unlock(&ifp->lock);
				in6_ifa_hold(ifp);
				ipv6_del_addr(ifp);
				goto restart;
3967 3968 3969
			} else if (ifp->prefered_lft == INFINITY_LIFE_TIME) {
				spin_unlock(&ifp->lock);
				continue;
L
Linus Torvalds 已提交
3970
			} else if (age >= ifp->prefered_lft) {
3971
				/* jiffies - ifp->tstamp > age >= ifp->prefered_lft */
L
Linus Torvalds 已提交
3972 3973 3974 3975 3976 3977 3978
				int deprecate = 0;

				if (!(ifp->flags&IFA_F_DEPRECATED)) {
					deprecate = 1;
					ifp->flags |= IFA_F_DEPRECATED;
				}

3979 3980
				if ((ifp->valid_lft != INFINITY_LIFE_TIME) &&
				    (time_before(ifp->tstamp + ifp->valid_lft * HZ, next)))
L
Linus Torvalds 已提交
3981 3982 3983 3984 3985 3986 3987 3988 3989 3990 3991 3992 3993
					next = ifp->tstamp + ifp->valid_lft * HZ;

				spin_unlock(&ifp->lock);

				if (deprecate) {
					in6_ifa_hold(ifp);

					ipv6_ifa_notify(0, ifp);
					in6_ifa_put(ifp);
					goto restart;
				}
			} else if ((ifp->flags&IFA_F_TEMPORARY) &&
				   !(ifp->flags&IFA_F_TENTATIVE)) {
3994 3995
				unsigned long regen_advance = ifp->idev->cnf.regen_max_retry *
					ifp->idev->cnf.dad_transmits *
J
Jiri Pirko 已提交
3996
					NEIGH_VAR(ifp->idev->nd_parms, RETRANS_TIME) / HZ;
3997

L
Linus Torvalds 已提交
3998 3999 4000 4001 4002 4003 4004 4005 4006
				if (age >= ifp->prefered_lft - regen_advance) {
					struct inet6_ifaddr *ifpub = ifp->ifpub;
					if (time_before(ifp->tstamp + ifp->prefered_lft * HZ, next))
						next = ifp->tstamp + ifp->prefered_lft * HZ;
					if (!ifp->regen_count && ifpub) {
						ifp->regen_count++;
						in6_ifa_hold(ifp);
						in6_ifa_hold(ifpub);
						spin_unlock(&ifp->lock);
4007

4008 4009 4010
						spin_lock(&ifpub->lock);
						ifpub->regen_count = 0;
						spin_unlock(&ifpub->lock);
L
Linus Torvalds 已提交
4011 4012 4013 4014 4015 4016 4017 4018 4019 4020 4021 4022 4023 4024 4025 4026 4027
						ipv6_create_tempaddr(ifpub, ifp);
						in6_ifa_put(ifpub);
						in6_ifa_put(ifp);
						goto restart;
					}
				} else if (time_before(ifp->tstamp + ifp->prefered_lft * HZ - regen_advance * HZ, next))
					next = ifp->tstamp + ifp->prefered_lft * HZ - regen_advance * HZ;
				spin_unlock(&ifp->lock);
			} else {
				/* ifp->prefered_lft <= ifp->valid_lft */
				if (time_before(ifp->tstamp + ifp->prefered_lft * HZ, next))
					next = ifp->tstamp + ifp->prefered_lft * HZ;
				spin_unlock(&ifp->lock);
			}
		}
	}

4028 4029 4030 4031 4032 4033 4034 4035 4036 4037 4038
	next_sec = round_jiffies_up(next);
	next_sched = next;

	/* If rounded timeout is accurate enough, accept it. */
	if (time_before(next_sec, next + ADDRCONF_TIMER_FUZZ))
		next_sched = next_sec;

	/* And minimum interval is ADDRCONF_TIMER_FUZZ_MAX. */
	if (time_before(next_sched, jiffies + ADDRCONF_TIMER_FUZZ_MAX))
		next_sched = jiffies + ADDRCONF_TIMER_FUZZ_MAX;

4039 4040
	ADBG(KERN_DEBUG "now = %lu, schedule = %lu, rounded schedule = %lu => %lu\n",
	      now, next, next_sec, next_sched);
4041
	mod_delayed_work(addrconf_wq, &addr_chk_work, next_sched - now);
4042
	rcu_read_unlock_bh();
L
Linus Torvalds 已提交
4043 4044
}

4045 4046 4047 4048 4049 4050 4051 4052 4053 4054 4055 4056
static void addrconf_verify_work(struct work_struct *w)
{
	rtnl_lock();
	addrconf_verify_rtnl();
	rtnl_unlock();
}

static void addrconf_verify(void)
{
	mod_delayed_work(addrconf_wq, &addr_chk_work, 0);
}

4057 4058
static struct in6_addr *extract_addr(struct nlattr *addr, struct nlattr *local,
				     struct in6_addr **peer_pfx)
4059 4060 4061
{
	struct in6_addr *pfx = NULL;

4062 4063
	*peer_pfx = NULL;

4064 4065 4066 4067 4068
	if (addr)
		pfx = nla_data(addr);

	if (local) {
		if (pfx && nla_memcmp(local, pfx, sizeof(*pfx)))
4069 4070
			*peer_pfx = pfx;
		pfx = nla_data(local);
4071 4072 4073 4074 4075
	}

	return pfx;
}

4076
static const struct nla_policy ifa_ipv6_policy[IFA_MAX+1] = {
4077 4078 4079
	[IFA_ADDRESS]		= { .len = sizeof(struct in6_addr) },
	[IFA_LOCAL]		= { .len = sizeof(struct in6_addr) },
	[IFA_CACHEINFO]		= { .len = sizeof(struct ifa_cacheinfo) },
4080
	[IFA_FLAGS]		= { .len = sizeof(u32) },
4081 4082
};

L
Linus Torvalds 已提交
4083
static int
4084
inet6_rtm_deladdr(struct sk_buff *skb, struct nlmsghdr *nlh)
L
Linus Torvalds 已提交
4085
{
4086
	struct net *net = sock_net(skb->sk);
4087 4088
	struct ifaddrmsg *ifm;
	struct nlattr *tb[IFA_MAX+1];
4089
	struct in6_addr *pfx, *peer_pfx;
4090
	u32 ifa_flags;
4091
	int err;
L
Linus Torvalds 已提交
4092

4093 4094 4095 4096 4097
	err = nlmsg_parse(nlh, sizeof(*ifm), tb, IFA_MAX, ifa_ipv6_policy);
	if (err < 0)
		return err;

	ifm = nlmsg_data(nlh);
4098
	pfx = extract_addr(tb[IFA_ADDRESS], tb[IFA_LOCAL], &peer_pfx);
4099
	if (!pfx)
L
Linus Torvalds 已提交
4100 4101
		return -EINVAL;

4102 4103 4104 4105 4106 4107 4108
	ifa_flags = tb[IFA_FLAGS] ? nla_get_u32(tb[IFA_FLAGS]) : ifm->ifa_flags;

	/* We ignore other flags so far. */
	ifa_flags &= IFA_F_MANAGETEMPADDR;

	return inet6_addr_del(net, ifm->ifa_index, ifa_flags, pfx,
			      ifm->ifa_prefixlen);
L
Linus Torvalds 已提交
4109 4110
}

4111
static int inet6_addr_modify(struct inet6_ifaddr *ifp, u32 ifa_flags,
4112
			     u32 prefered_lft, u32 valid_lft)
4113
{
4114 4115
	u32 flags;
	clock_t expires;
4116
	unsigned long timeout;
4117
	bool was_managetempaddr;
4118
	bool had_prefixroute;
4119

4120 4121
	ASSERT_RTNL();

4122 4123 4124
	if (!valid_lft || (prefered_lft > valid_lft))
		return -EINVAL;

4125 4126 4127 4128
	if (ifa_flags & IFA_F_MANAGETEMPADDR &&
	    (ifp->flags & IFA_F_TEMPORARY || ifp->prefix_len != 64))
		return -EINVAL;

4129 4130 4131 4132
	timeout = addrconf_timeout_fixup(valid_lft, HZ);
	if (addrconf_finite_timeout(timeout)) {
		expires = jiffies_to_clock_t(timeout * HZ);
		valid_lft = timeout;
4133
		flags = RTF_EXPIRES;
4134 4135 4136 4137
	} else {
		expires = 0;
		flags = 0;
		ifa_flags |= IFA_F_PERMANENT;
4138
	}
4139

4140 4141 4142 4143 4144 4145
	timeout = addrconf_timeout_fixup(prefered_lft, HZ);
	if (addrconf_finite_timeout(timeout)) {
		if (timeout == 0)
			ifa_flags |= IFA_F_DEPRECATED;
		prefered_lft = timeout;
	}
4146 4147

	spin_lock_bh(&ifp->lock);
4148
	was_managetempaddr = ifp->flags & IFA_F_MANAGETEMPADDR;
4149 4150
	had_prefixroute = ifp->flags & IFA_F_PERMANENT &&
			  !(ifp->flags & IFA_F_NOPREFIXROUTE);
4151
	ifp->flags &= ~(IFA_F_DEPRECATED | IFA_F_PERMANENT | IFA_F_NODAD |
4152 4153
			IFA_F_HOMEADDRESS | IFA_F_MANAGETEMPADDR |
			IFA_F_NOPREFIXROUTE);
4154
	ifp->flags |= ifa_flags;
4155 4156 4157 4158 4159 4160 4161 4162
	ifp->tstamp = jiffies;
	ifp->valid_lft = valid_lft;
	ifp->prefered_lft = prefered_lft;

	spin_unlock_bh(&ifp->lock);
	if (!(ifp->flags&IFA_F_TENTATIVE))
		ipv6_ifa_notify(0, ifp);

4163 4164 4165
	if (!(ifa_flags & IFA_F_NOPREFIXROUTE)) {
		addrconf_prefix_route(&ifp->addr, ifp->prefix_len, ifp->idev->dev,
				      expires, flags);
4166 4167 4168 4169 4170 4171 4172 4173 4174 4175 4176 4177
	} else if (had_prefixroute) {
		enum cleanup_prefix_rt_t action;
		unsigned long rt_expires;

		write_lock_bh(&ifp->idev->lock);
		action = check_cleanup_prefix_route(ifp, &rt_expires);
		write_unlock_bh(&ifp->idev->lock);

		if (action != CLEANUP_PREFIX_RT_NOP) {
			cleanup_prefix_route(ifp, rt_expires,
				action == CLEANUP_PREFIX_RT_DEL);
		}
4178
	}
4179 4180 4181 4182 4183 4184 4185 4186

	if (was_managetempaddr || ifp->flags & IFA_F_MANAGETEMPADDR) {
		if (was_managetempaddr && !(ifp->flags & IFA_F_MANAGETEMPADDR))
			valid_lft = prefered_lft = 0;
		manage_tempaddrs(ifp->idev, ifp, valid_lft, prefered_lft,
				 !was_managetempaddr, jiffies);
	}

4187
	addrconf_verify_rtnl();
4188 4189 4190 4191

	return 0;
}

L
Linus Torvalds 已提交
4192
static int
4193
inet6_rtm_newaddr(struct sk_buff *skb, struct nlmsghdr *nlh)
L
Linus Torvalds 已提交
4194
{
4195
	struct net *net = sock_net(skb->sk);
4196 4197
	struct ifaddrmsg *ifm;
	struct nlattr *tb[IFA_MAX+1];
4198
	struct in6_addr *pfx, *peer_pfx;
4199 4200
	struct inet6_ifaddr *ifa;
	struct net_device *dev;
4201
	u32 valid_lft = INFINITY_LIFE_TIME, preferred_lft = INFINITY_LIFE_TIME;
4202
	u32 ifa_flags;
4203
	int err;
L
Linus Torvalds 已提交
4204

4205 4206 4207 4208 4209
	err = nlmsg_parse(nlh, sizeof(*ifm), tb, IFA_MAX, ifa_ipv6_policy);
	if (err < 0)
		return err;

	ifm = nlmsg_data(nlh);
4210
	pfx = extract_addr(tb[IFA_ADDRESS], tb[IFA_LOCAL], &peer_pfx);
4211
	if (!pfx)
L
Linus Torvalds 已提交
4212 4213
		return -EINVAL;

4214
	if (tb[IFA_CACHEINFO]) {
4215
		struct ifa_cacheinfo *ci;
4216 4217

		ci = nla_data(tb[IFA_CACHEINFO]);
4218
		valid_lft = ci->ifa_valid;
4219 4220 4221 4222
		preferred_lft = ci->ifa_prefered;
	} else {
		preferred_lft = INFINITY_LIFE_TIME;
		valid_lft = INFINITY_LIFE_TIME;
4223 4224
	}

4225
	dev =  __dev_get_by_index(net, ifm->ifa_index);
4226
	if (!dev)
4227 4228
		return -ENODEV;

4229 4230
	ifa_flags = tb[IFA_FLAGS] ? nla_get_u32(tb[IFA_FLAGS]) : ifm->ifa_flags;

4231
	/* We ignore other flags so far. */
4232
	ifa_flags &= IFA_F_NODAD | IFA_F_HOMEADDRESS | IFA_F_MANAGETEMPADDR |
4233
		     IFA_F_NOPREFIXROUTE | IFA_F_MCAUTOJOIN;
4234

4235
	ifa = ipv6_get_ifaddr(net, pfx, dev, 1);
4236
	if (!ifa) {
4237 4238
		/*
		 * It would be best to check for !NLM_F_CREATE here but
4239
		 * userspace already relies on not having to provide this.
4240
		 */
4241
		return inet6_addr_add(net, ifm->ifa_index, pfx, peer_pfx,
4242 4243
				      ifm->ifa_prefixlen, ifa_flags,
				      preferred_lft, valid_lft);
4244 4245
	}

4246 4247 4248 4249
	if (nlh->nlmsg_flags & NLM_F_EXCL ||
	    !(nlh->nlmsg_flags & NLM_F_REPLACE))
		err = -EEXIST;
	else
4250
		err = inet6_addr_modify(ifa, ifa_flags, preferred_lft, valid_lft);
4251 4252 4253 4254

	in6_ifa_put(ifa);

	return err;
L
Linus Torvalds 已提交
4255 4256
}

4257
static void put_ifaddrmsg(struct nlmsghdr *nlh, u8 prefixlen, u32 flags,
4258 4259 4260 4261 4262 4263 4264 4265 4266 4267 4268 4269
			  u8 scope, int ifindex)
{
	struct ifaddrmsg *ifm;

	ifm = nlmsg_data(nlh);
	ifm->ifa_family = AF_INET6;
	ifm->ifa_prefixlen = prefixlen;
	ifm->ifa_flags = flags;
	ifm->ifa_scope = scope;
	ifm->ifa_index = ifindex;
}

4270 4271 4272 4273 4274
static int put_cacheinfo(struct sk_buff *skb, unsigned long cstamp,
			 unsigned long tstamp, u32 preferred, u32 valid)
{
	struct ifa_cacheinfo ci;

4275 4276
	ci.cstamp = cstamp_delta(cstamp);
	ci.tstamp = cstamp_delta(tstamp);
4277 4278 4279 4280 4281 4282
	ci.ifa_prefered = preferred;
	ci.ifa_valid = valid;

	return nla_put(skb, IFA_CACHEINFO, sizeof(ci), &ci);
}

4283 4284 4285 4286 4287 4288 4289 4290 4291 4292 4293 4294
static inline int rt_scope(int ifa_scope)
{
	if (ifa_scope & IFA_HOST)
		return RT_SCOPE_HOST;
	else if (ifa_scope & IFA_LINK)
		return RT_SCOPE_LINK;
	else if (ifa_scope & IFA_SITE)
		return RT_SCOPE_SITE;
	else
		return RT_SCOPE_UNIVERSE;
}

4295 4296
static inline int inet6_ifaddr_msgsize(void)
{
4297
	return NLMSG_ALIGN(sizeof(struct ifaddrmsg))
4298
	       + nla_total_size(16) /* IFA_LOCAL */
4299
	       + nla_total_size(16) /* IFA_ADDRESS */
4300 4301
	       + nla_total_size(sizeof(struct ifa_cacheinfo))
	       + nla_total_size(4)  /* IFA_FLAGS */;
4302
}
4303

L
Linus Torvalds 已提交
4304
static int inet6_fill_ifaddr(struct sk_buff *skb, struct inet6_ifaddr *ifa,
4305
			     u32 portid, u32 seq, int event, unsigned int flags)
L
Linus Torvalds 已提交
4306 4307
{
	struct nlmsghdr  *nlh;
4308
	u32 preferred, valid;
L
Linus Torvalds 已提交
4309

4310
	nlh = nlmsg_put(skb, portid, seq, event, sizeof(struct ifaddrmsg), flags);
4311
	if (!nlh)
4312
		return -EMSGSIZE;
4313

4314 4315 4316
	put_ifaddrmsg(nlh, ifa->prefix_len, ifa->flags, rt_scope(ifa->scope),
		      ifa->idev->dev->ifindex);

4317 4318
	if (!((ifa->flags&IFA_F_PERMANENT) &&
	      (ifa->prefered_lft == INFINITY_LIFE_TIME))) {
4319 4320 4321
		preferred = ifa->prefered_lft;
		valid = ifa->valid_lft;
		if (preferred != INFINITY_LIFE_TIME) {
L
Linus Torvalds 已提交
4322
			long tval = (jiffies - ifa->tstamp)/HZ;
4323 4324 4325 4326
			if (preferred > tval)
				preferred -= tval;
			else
				preferred = 0;
4327 4328 4329 4330 4331 4332
			if (valid != INFINITY_LIFE_TIME) {
				if (valid > tval)
					valid -= tval;
				else
					valid = 0;
			}
L
Linus Torvalds 已提交
4333 4334
		}
	} else {
4335 4336 4337 4338
		preferred = INFINITY_LIFE_TIME;
		valid = INFINITY_LIFE_TIME;
	}

C
Cong Wang 已提交
4339
	if (!ipv6_addr_any(&ifa->peer_addr)) {
4340 4341
		if (nla_put_in6_addr(skb, IFA_LOCAL, &ifa->addr) < 0 ||
		    nla_put_in6_addr(skb, IFA_ADDRESS, &ifa->peer_addr) < 0)
4342 4343
			goto error;
	} else
4344
		if (nla_put_in6_addr(skb, IFA_ADDRESS, &ifa->addr) < 0)
4345 4346 4347 4348
			goto error;

	if (put_cacheinfo(skb, ifa->cstamp, ifa->tstamp, preferred, valid) < 0)
		goto error;
L
Linus Torvalds 已提交
4349

4350 4351 4352
	if (nla_put_u32(skb, IFA_FLAGS, ifa->flags) < 0)
		goto error;

4353 4354
	nlmsg_end(skb, nlh);
	return 0;
4355 4356 4357 4358

error:
	nlmsg_cancel(skb, nlh);
	return -EMSGSIZE;
L
Linus Torvalds 已提交
4359 4360 4361
}

static int inet6_fill_ifmcaddr(struct sk_buff *skb, struct ifmcaddr6 *ifmca,
4362
				u32 portid, u32 seq, int event, u16 flags)
L
Linus Torvalds 已提交
4363 4364
{
	struct nlmsghdr  *nlh;
4365 4366 4367 4368 4369
	u8 scope = RT_SCOPE_UNIVERSE;
	int ifindex = ifmca->idev->dev->ifindex;

	if (ipv6_addr_scope(&ifmca->mca_addr) & IFA_SITE)
		scope = RT_SCOPE_SITE;
L
Linus Torvalds 已提交
4370

4371
	nlh = nlmsg_put(skb, portid, seq, event, sizeof(struct ifaddrmsg), flags);
4372
	if (!nlh)
4373
		return -EMSGSIZE;
4374

4375
	put_ifaddrmsg(nlh, 128, IFA_F_PERMANENT, scope, ifindex);
4376
	if (nla_put_in6_addr(skb, IFA_MULTICAST, &ifmca->mca_addr) < 0 ||
4377
	    put_cacheinfo(skb, ifmca->mca_cstamp, ifmca->mca_tstamp,
4378 4379 4380 4381
			  INFINITY_LIFE_TIME, INFINITY_LIFE_TIME) < 0) {
		nlmsg_cancel(skb, nlh);
		return -EMSGSIZE;
	}
4382

4383 4384
	nlmsg_end(skb, nlh);
	return 0;
L
Linus Torvalds 已提交
4385 4386 4387
}

static int inet6_fill_ifacaddr(struct sk_buff *skb, struct ifacaddr6 *ifaca,
4388
				u32 portid, u32 seq, int event, unsigned int flags)
L
Linus Torvalds 已提交
4389 4390
{
	struct nlmsghdr  *nlh;
4391 4392 4393 4394 4395
	u8 scope = RT_SCOPE_UNIVERSE;
	int ifindex = ifaca->aca_idev->dev->ifindex;

	if (ipv6_addr_scope(&ifaca->aca_addr) & IFA_SITE)
		scope = RT_SCOPE_SITE;
L
Linus Torvalds 已提交
4396

4397
	nlh = nlmsg_put(skb, portid, seq, event, sizeof(struct ifaddrmsg), flags);
4398
	if (!nlh)
4399
		return -EMSGSIZE;
4400

4401
	put_ifaddrmsg(nlh, 128, IFA_F_PERMANENT, scope, ifindex);
4402
	if (nla_put_in6_addr(skb, IFA_ANYCAST, &ifaca->aca_addr) < 0 ||
4403
	    put_cacheinfo(skb, ifaca->aca_cstamp, ifaca->aca_tstamp,
4404 4405 4406 4407
			  INFINITY_LIFE_TIME, INFINITY_LIFE_TIME) < 0) {
		nlmsg_cancel(skb, nlh);
		return -EMSGSIZE;
	}
L
Linus Torvalds 已提交
4408

4409 4410
	nlmsg_end(skb, nlh);
	return 0;
L
Linus Torvalds 已提交
4411 4412
}

S
Stephen Hemminger 已提交
4413
enum addr_type_t {
L
Linus Torvalds 已提交
4414 4415 4416 4417 4418
	UNICAST_ADDR,
	MULTICAST_ADDR,
	ANYCAST_ADDR,
};

E
Eric Dumazet 已提交
4419 4420 4421 4422 4423 4424 4425 4426 4427 4428 4429 4430
/* called with rcu_read_lock() */
static int in6_dump_addrs(struct inet6_dev *idev, struct sk_buff *skb,
			  struct netlink_callback *cb, enum addr_type_t type,
			  int s_ip_idx, int *p_ip_idx)
{
	struct ifmcaddr6 *ifmca;
	struct ifacaddr6 *ifaca;
	int err = 1;
	int ip_idx = *p_ip_idx;

	read_lock_bh(&idev->lock);
	switch (type) {
4431 4432 4433
	case UNICAST_ADDR: {
		struct inet6_ifaddr *ifa;

E
Eric Dumazet 已提交
4434
		/* unicast address incl. temp addr */
4435 4436
		list_for_each_entry(ifa, &idev->addr_list, if_list) {
			if (++ip_idx < s_ip_idx)
E
Eric Dumazet 已提交
4437 4438
				continue;
			err = inet6_fill_ifaddr(skb, ifa,
4439
						NETLINK_CB(cb->skb).portid,
E
Eric Dumazet 已提交
4440 4441 4442
						cb->nlh->nlmsg_seq,
						RTM_NEWADDR,
						NLM_F_MULTI);
4443
			if (err < 0)
E
Eric Dumazet 已提交
4444
				break;
4445
			nl_dump_check_consistent(cb, nlmsg_hdr(skb));
E
Eric Dumazet 已提交
4446 4447
		}
		break;
4448
	}
E
Eric Dumazet 已提交
4449 4450 4451 4452 4453 4454 4455
	case MULTICAST_ADDR:
		/* multicast address */
		for (ifmca = idev->mc_list; ifmca;
		     ifmca = ifmca->next, ip_idx++) {
			if (ip_idx < s_ip_idx)
				continue;
			err = inet6_fill_ifmcaddr(skb, ifmca,
4456
						  NETLINK_CB(cb->skb).portid,
E
Eric Dumazet 已提交
4457 4458 4459
						  cb->nlh->nlmsg_seq,
						  RTM_GETMULTICAST,
						  NLM_F_MULTI);
4460
			if (err < 0)
E
Eric Dumazet 已提交
4461 4462 4463 4464 4465 4466 4467 4468 4469 4470
				break;
		}
		break;
	case ANYCAST_ADDR:
		/* anycast address */
		for (ifaca = idev->ac_list; ifaca;
		     ifaca = ifaca->aca_next, ip_idx++) {
			if (ip_idx < s_ip_idx)
				continue;
			err = inet6_fill_ifacaddr(skb, ifaca,
4471
						  NETLINK_CB(cb->skb).portid,
E
Eric Dumazet 已提交
4472 4473 4474
						  cb->nlh->nlmsg_seq,
						  RTM_GETANYCAST,
						  NLM_F_MULTI);
4475
			if (err < 0)
E
Eric Dumazet 已提交
4476 4477 4478 4479 4480 4481 4482 4483 4484 4485 4486
				break;
		}
		break;
	default:
		break;
	}
	read_unlock_bh(&idev->lock);
	*p_ip_idx = ip_idx;
	return err;
}

L
Linus Torvalds 已提交
4487 4488 4489
static int inet6_dump_addr(struct sk_buff *skb, struct netlink_callback *cb,
			   enum addr_type_t type)
{
E
Eric Dumazet 已提交
4490 4491
	struct net *net = sock_net(skb->sk);
	int h, s_h;
L
Linus Torvalds 已提交
4492 4493 4494
	int idx, ip_idx;
	int s_idx, s_ip_idx;
	struct net_device *dev;
E
Eric Dumazet 已提交
4495 4496
	struct inet6_dev *idev;
	struct hlist_head *head;
4497

E
Eric Dumazet 已提交
4498 4499 4500
	s_h = cb->args[0];
	s_idx = idx = cb->args[1];
	s_ip_idx = ip_idx = cb->args[2];
4501

E
Eric Dumazet 已提交
4502
	rcu_read_lock();
4503
	cb->seq = atomic_read(&net->ipv6.dev_addr_genid) ^ net->dev_base_seq;
E
Eric Dumazet 已提交
4504 4505 4506
	for (h = s_h; h < NETDEV_HASHENTRIES; h++, s_idx = 0) {
		idx = 0;
		head = &net->dev_index_head[h];
4507
		hlist_for_each_entry_rcu(dev, head, index_hlist) {
E
Eric Dumazet 已提交
4508 4509
			if (idx < s_idx)
				goto cont;
4510
			if (h > s_h || idx > s_idx)
E
Eric Dumazet 已提交
4511 4512
				s_ip_idx = 0;
			ip_idx = 0;
S
Stephen Hemminger 已提交
4513 4514
			idev = __in6_dev_get(dev);
			if (!idev)
E
Eric Dumazet 已提交
4515 4516 4517
				goto cont;

			if (in6_dump_addrs(idev, skb, cb, type,
4518
					   s_ip_idx, &ip_idx) < 0)
E
Eric Dumazet 已提交
4519
				goto done;
4520
cont:
E
Eric Dumazet 已提交
4521 4522
			idx++;
		}
L
Linus Torvalds 已提交
4523
	}
E
Eric Dumazet 已提交
4524 4525 4526 4527 4528 4529
done:
	rcu_read_unlock();
	cb->args[0] = h;
	cb->args[1] = idx;
	cb->args[2] = ip_idx;

L
Linus Torvalds 已提交
4530 4531 4532 4533 4534 4535
	return skb->len;
}

static int inet6_dump_ifaddr(struct sk_buff *skb, struct netlink_callback *cb)
{
	enum addr_type_t type = UNICAST_ADDR;
4536

L
Linus Torvalds 已提交
4537 4538 4539 4540 4541 4542
	return inet6_dump_addr(skb, cb, type);
}

static int inet6_dump_ifmcaddr(struct sk_buff *skb, struct netlink_callback *cb)
{
	enum addr_type_t type = MULTICAST_ADDR;
4543

L
Linus Torvalds 已提交
4544 4545 4546 4547 4548 4549 4550
	return inet6_dump_addr(skb, cb, type);
}


static int inet6_dump_ifacaddr(struct sk_buff *skb, struct netlink_callback *cb)
{
	enum addr_type_t type = ANYCAST_ADDR;
4551

L
Linus Torvalds 已提交
4552 4553 4554
	return inet6_dump_addr(skb, cb, type);
}

4555
static int inet6_rtm_getaddr(struct sk_buff *in_skb, struct nlmsghdr *nlh)
4556
{
4557
	struct net *net = sock_net(in_skb->sk);
4558 4559
	struct ifaddrmsg *ifm;
	struct nlattr *tb[IFA_MAX+1];
4560
	struct in6_addr *addr = NULL, *peer;
4561 4562 4563 4564 4565
	struct net_device *dev = NULL;
	struct inet6_ifaddr *ifa;
	struct sk_buff *skb;
	int err;

4566 4567 4568 4569
	err = nlmsg_parse(nlh, sizeof(*ifm), tb, IFA_MAX, ifa_ipv6_policy);
	if (err < 0)
		goto errout;

4570
	addr = extract_addr(tb[IFA_ADDRESS], tb[IFA_LOCAL], &peer);
4571
	if (!addr) {
4572 4573
		err = -EINVAL;
		goto errout;
4574 4575
	}

4576
	ifm = nlmsg_data(nlh);
4577
	if (ifm->ifa_index)
4578
		dev = __dev_get_by_index(net, ifm->ifa_index);
4579

S
Stephen Hemminger 已提交
4580 4581
	ifa = ipv6_get_ifaddr(net, addr, dev, 1);
	if (!ifa) {
4582 4583 4584
		err = -EADDRNOTAVAIL;
		goto errout;
	}
4585

S
Stephen Hemminger 已提交
4586 4587
	skb = nlmsg_new(inet6_ifaddr_msgsize(), GFP_KERNEL);
	if (!skb) {
4588
		err = -ENOBUFS;
4589
		goto errout_ifa;
4590 4591
	}

4592
	err = inet6_fill_ifaddr(skb, ifa, NETLINK_CB(in_skb).portid,
4593
				nlh->nlmsg_seq, RTM_NEWADDR, 0);
4594 4595 4596 4597 4598 4599
	if (err < 0) {
		/* -EMSGSIZE implies BUG in inet6_ifaddr_msgsize() */
		WARN_ON(err == -EMSGSIZE);
		kfree_skb(skb);
		goto errout_ifa;
	}
4600
	err = rtnl_unicast(skb, net, NETLINK_CB(in_skb).portid);
4601
errout_ifa:
4602
	in6_ifa_put(ifa);
4603
errout:
4604 4605 4606
	return err;
}

L
Linus Torvalds 已提交
4607 4608 4609
static void inet6_ifa_notify(int event, struct inet6_ifaddr *ifa)
{
	struct sk_buff *skb;
4610
	struct net *net = dev_net(ifa->idev->dev);
4611
	int err = -ENOBUFS;
L
Linus Torvalds 已提交
4612

4613
	skb = nlmsg_new(inet6_ifaddr_msgsize(), GFP_ATOMIC);
4614
	if (!skb)
4615 4616 4617
		goto errout;

	err = inet6_fill_ifaddr(skb, ifa, 0, 0, event, 0);
4618 4619 4620 4621 4622 4623
	if (err < 0) {
		/* -EMSGSIZE implies BUG in inet6_ifaddr_msgsize() */
		WARN_ON(err == -EMSGSIZE);
		kfree_skb(skb);
		goto errout;
	}
4624 4625
	rtnl_notify(skb, net, 0, RTNLGRP_IPV6_IFADDR, NULL, GFP_ATOMIC);
	return;
4626 4627
errout:
	if (err < 0)
4628
		rtnl_set_sk_err(net, RTNLGRP_IPV6_IFADDR, err);
L
Linus Torvalds 已提交
4629 4630
}

D
Dave Jones 已提交
4631
static inline void ipv6_store_devconf(struct ipv6_devconf *cnf,
L
Linus Torvalds 已提交
4632 4633
				__s32 *array, int bytes)
{
4634 4635
	BUG_ON(bytes < (DEVCONF_MAX * 4));

L
Linus Torvalds 已提交
4636 4637 4638 4639 4640 4641 4642 4643 4644
	memset(array, 0, bytes);
	array[DEVCONF_FORWARDING] = cnf->forwarding;
	array[DEVCONF_HOPLIMIT] = cnf->hop_limit;
	array[DEVCONF_MTU6] = cnf->mtu6;
	array[DEVCONF_ACCEPT_RA] = cnf->accept_ra;
	array[DEVCONF_ACCEPT_REDIRECTS] = cnf->accept_redirects;
	array[DEVCONF_AUTOCONF] = cnf->autoconf;
	array[DEVCONF_DAD_TRANSMITS] = cnf->dad_transmits;
	array[DEVCONF_RTR_SOLICITS] = cnf->rtr_solicits;
4645 4646 4647 4648
	array[DEVCONF_RTR_SOLICIT_INTERVAL] =
		jiffies_to_msecs(cnf->rtr_solicit_interval);
	array[DEVCONF_RTR_SOLICIT_DELAY] =
		jiffies_to_msecs(cnf->rtr_solicit_delay);
L
Linus Torvalds 已提交
4649
	array[DEVCONF_FORCE_MLD_VERSION] = cnf->force_mld_version;
4650 4651 4652 4653
	array[DEVCONF_MLDV1_UNSOLICITED_REPORT_INTERVAL] =
		jiffies_to_msecs(cnf->mldv1_unsolicited_report_interval);
	array[DEVCONF_MLDV2_UNSOLICITED_REPORT_INTERVAL] =
		jiffies_to_msecs(cnf->mldv2_unsolicited_report_interval);
L
Linus Torvalds 已提交
4654 4655 4656 4657 4658 4659
	array[DEVCONF_USE_TEMPADDR] = cnf->use_tempaddr;
	array[DEVCONF_TEMP_VALID_LFT] = cnf->temp_valid_lft;
	array[DEVCONF_TEMP_PREFERED_LFT] = cnf->temp_prefered_lft;
	array[DEVCONF_REGEN_MAX_RETRY] = cnf->regen_max_retry;
	array[DEVCONF_MAX_DESYNC_FACTOR] = cnf->max_desync_factor;
	array[DEVCONF_MAX_ADDRESSES] = cnf->max_addresses;
4660
	array[DEVCONF_ACCEPT_RA_DEFRTR] = cnf->accept_ra_defrtr;
4661
	array[DEVCONF_ACCEPT_RA_MIN_HOP_LIMIT] = cnf->accept_ra_min_hop_limit;
4662
	array[DEVCONF_ACCEPT_RA_PINFO] = cnf->accept_ra_pinfo;
4663 4664
#ifdef CONFIG_IPV6_ROUTER_PREF
	array[DEVCONF_ACCEPT_RA_RTR_PREF] = cnf->accept_ra_rtr_pref;
4665 4666
	array[DEVCONF_RTR_PROBE_INTERVAL] =
		jiffies_to_msecs(cnf->rtr_probe_interval);
N
Neil Horman 已提交
4667
#ifdef CONFIG_IPV6_ROUTE_INFO
4668 4669
	array[DEVCONF_ACCEPT_RA_RT_INFO_MAX_PLEN] = cnf->accept_ra_rt_info_max_plen;
#endif
4670
#endif
4671
	array[DEVCONF_PROXY_NDP] = cnf->proxy_ndp;
4672
	array[DEVCONF_ACCEPT_SOURCE_ROUTE] = cnf->accept_source_route;
4673 4674
#ifdef CONFIG_IPV6_OPTIMISTIC_DAD
	array[DEVCONF_OPTIMISTIC_DAD] = cnf->optimistic_dad;
4675
	array[DEVCONF_USE_OPTIMISTIC] = cnf->use_optimistic;
4676
#endif
4677 4678 4679
#ifdef CONFIG_IPV6_MROUTE
	array[DEVCONF_MC_FORWARDING] = cnf->mc_forwarding;
#endif
4680
	array[DEVCONF_DISABLE_IPV6] = cnf->disable_ipv6;
4681
	array[DEVCONF_ACCEPT_DAD] = cnf->accept_dad;
4682
	array[DEVCONF_FORCE_TLLAO] = cnf->force_tllao;
4683
	array[DEVCONF_NDISC_NOTIFY] = cnf->ndisc_notify;
4684
	array[DEVCONF_SUPPRESS_FRAG_NDISC] = cnf->suppress_frag_ndisc;
4685
	array[DEVCONF_ACCEPT_RA_FROM_LOCAL] = cnf->accept_ra_from_local;
4686
	array[DEVCONF_ACCEPT_RA_MTU] = cnf->accept_ra_mtu;
4687
	array[DEVCONF_IGNORE_ROUTES_WITH_LINKDOWN] = cnf->ignore_routes_with_linkdown;
4688
	/* we omit DEVCONF_STABLE_SECRET for now */
4689
	array[DEVCONF_USE_OIF_ADDRS_ONLY] = cnf->use_oif_addrs_only;
L
Linus Torvalds 已提交
4690 4691
}

T
Thomas Graf 已提交
4692 4693 4694 4695 4696 4697
static inline size_t inet6_ifla6_size(void)
{
	return nla_total_size(4) /* IFLA_INET6_FLAGS */
	     + nla_total_size(sizeof(struct ifla_cacheinfo))
	     + nla_total_size(DEVCONF_MAX * 4) /* IFLA_INET6_CONF */
	     + nla_total_size(IPSTATS_MIB_MAX * 8) /* IFLA_INET6_STATS */
4698 4699
	     + nla_total_size(ICMP6_MIB_MAX * 8) /* IFLA_INET6_ICMP6STATS */
	     + nla_total_size(sizeof(struct in6_addr)); /* IFLA_INET6_TOKEN */
T
Thomas Graf 已提交
4700 4701
}

4702 4703 4704 4705 4706 4707 4708
static inline size_t inet6_if_nlmsg_size(void)
{
	return NLMSG_ALIGN(sizeof(struct ifinfomsg))
	       + nla_total_size(IFNAMSIZ) /* IFLA_IFNAME */
	       + nla_total_size(MAX_ADDR_LEN) /* IFLA_ADDRESS */
	       + nla_total_size(4) /* IFLA_MTU */
	       + nla_total_size(4) /* IFLA_LINK */
4709
	       + nla_total_size(1) /* IFLA_OPERSTATE */
T
Thomas Graf 已提交
4710
	       + nla_total_size(inet6_ifla6_size()); /* IFLA_PROTINFO */
4711
}
4712

4713
static inline void __snmp6_fill_statsdev(u64 *stats, atomic_long_t *mib,
4714
				      int items, int bytes)
4715 4716 4717 4718 4719 4720 4721 4722
{
	int i;
	int pad = bytes - sizeof(u64) * items;
	BUG_ON(pad < 0);

	/* Use put_unaligned() because stats may not be aligned for u64. */
	put_unaligned(items, &stats[0]);
	for (i = 1; i < items; i++)
4723
		put_unaligned(atomic_long_read(&mib[i]), &stats[i]);
4724 4725 4726 4727

	memset(&stats[items], 0, pad);
}

W
WANG Cong 已提交
4728
static inline void __snmp6_fill_stats64(u64 *stats, void __percpu *mib,
4729 4730 4731 4732 4733 4734 4735 4736 4737 4738 4739 4740 4741 4742
				      int items, int bytes, size_t syncpoff)
{
	int i;
	int pad = bytes - sizeof(u64) * items;
	BUG_ON(pad < 0);

	/* Use put_unaligned() because stats may not be aligned for u64. */
	put_unaligned(items, &stats[0]);
	for (i = 1; i < items; i++)
		put_unaligned(snmp_fold_field64(mib, i, syncpoff), &stats[i]);

	memset(&stats[items], 0, pad);
}

4743 4744 4745
static void snmp6_fill_stats(u64 *stats, struct inet6_dev *idev, int attrtype,
			     int bytes)
{
S
Stephen Hemminger 已提交
4746
	switch (attrtype) {
4747
	case IFLA_INET6_STATS:
W
WANG Cong 已提交
4748
		__snmp6_fill_stats64(stats, idev->stats.ipv6,
4749
				     IPSTATS_MIB_MAX, bytes, offsetof(struct ipstats_mib, syncp));
4750 4751
		break;
	case IFLA_INET6_ICMP6STATS:
4752
		__snmp6_fill_statsdev(stats, idev->stats.icmpv6dev->mibs, ICMP6_MIB_MAX, bytes);
4753 4754 4755 4756
		break;
	}
}

T
Thomas Graf 已提交
4757 4758 4759 4760 4761
static int inet6_fill_ifla6_attrs(struct sk_buff *skb, struct inet6_dev *idev)
{
	struct nlattr *nla;
	struct ifla_cacheinfo ci;

D
David S. Miller 已提交
4762 4763
	if (nla_put_u32(skb, IFLA_INET6_FLAGS, idev->if_flags))
		goto nla_put_failure;
T
Thomas Graf 已提交
4764
	ci.max_reasm_len = IPV6_MAXPLEN;
4765 4766
	ci.tstamp = cstamp_delta(idev->tstamp);
	ci.reachable_time = jiffies_to_msecs(idev->nd_parms->reachable_time);
J
Jiri Pirko 已提交
4767
	ci.retrans_time = jiffies_to_msecs(NEIGH_VAR(idev->nd_parms, RETRANS_TIME));
D
David S. Miller 已提交
4768 4769
	if (nla_put(skb, IFLA_INET6_CACHEINFO, sizeof(ci), &ci))
		goto nla_put_failure;
T
Thomas Graf 已提交
4770
	nla = nla_reserve(skb, IFLA_INET6_CONF, DEVCONF_MAX * sizeof(s32));
4771
	if (!nla)
T
Thomas Graf 已提交
4772 4773 4774 4775 4776 4777
		goto nla_put_failure;
	ipv6_store_devconf(&idev->cnf, nla_data(nla), nla_len(nla));

	/* XXX - MC not implemented */

	nla = nla_reserve(skb, IFLA_INET6_STATS, IPSTATS_MIB_MAX * sizeof(u64));
4778
	if (!nla)
T
Thomas Graf 已提交
4779 4780 4781 4782
		goto nla_put_failure;
	snmp6_fill_stats(nla_data(nla), idev, IFLA_INET6_STATS, nla_len(nla));

	nla = nla_reserve(skb, IFLA_INET6_ICMP6STATS, ICMP6_MIB_MAX * sizeof(u64));
4783
	if (!nla)
T
Thomas Graf 已提交
4784 4785 4786
		goto nla_put_failure;
	snmp6_fill_stats(nla_data(nla), idev, IFLA_INET6_ICMP6STATS, nla_len(nla));

4787
	nla = nla_reserve(skb, IFLA_INET6_TOKEN, sizeof(struct in6_addr));
4788
	if (!nla)
4789
		goto nla_put_failure;
4790 4791 4792 4793

	if (nla_put_u8(skb, IFLA_INET6_ADDR_GEN_MODE, idev->addr_gen_mode))
		goto nla_put_failure;

4794 4795 4796 4797
	read_lock_bh(&idev->lock);
	memcpy(nla_data(nla), idev->token.s6_addr, nla_len(nla));
	read_unlock_bh(&idev->lock);

T
Thomas Graf 已提交
4798 4799 4800 4801 4802 4803 4804 4805 4806 4807 4808 4809 4810 4811 4812 4813 4814 4815 4816 4817 4818 4819 4820 4821 4822 4823 4824
	return 0;

nla_put_failure:
	return -EMSGSIZE;
}

static size_t inet6_get_link_af_size(const struct net_device *dev)
{
	if (!__in6_dev_get(dev))
		return 0;

	return inet6_ifla6_size();
}

static int inet6_fill_link_af(struct sk_buff *skb, const struct net_device *dev)
{
	struct inet6_dev *idev = __in6_dev_get(dev);

	if (!idev)
		return -ENODATA;

	if (inet6_fill_ifla6_attrs(skb, idev) < 0)
		return -EMSGSIZE;

	return 0;
}

4825 4826 4827 4828
static int inet6_set_iftoken(struct inet6_dev *idev, struct in6_addr *token)
{
	struct inet6_ifaddr *ifp;
	struct net_device *dev = idev->dev;
4829
	bool update_rs = false;
4830
	struct in6_addr ll_addr;
4831

4832 4833
	ASSERT_RTNL();

4834
	if (!token)
4835 4836 4837 4838 4839 4840 4841 4842 4843 4844 4845 4846 4847 4848 4849 4850 4851
		return -EINVAL;
	if (ipv6_addr_any(token))
		return -EINVAL;
	if (dev->flags & (IFF_LOOPBACK | IFF_NOARP))
		return -EINVAL;
	if (!ipv6_accept_ra(idev))
		return -EINVAL;
	if (idev->cnf.rtr_solicits <= 0)
		return -EINVAL;

	write_lock_bh(&idev->lock);

	BUILD_BUG_ON(sizeof(token->s6_addr) != 16);
	memcpy(idev->token.s6_addr + 8, token->s6_addr + 8, 8);

	write_unlock_bh(&idev->lock);

4852 4853 4854
	if (!idev->dead && (idev->if_flags & IF_READY) &&
	    !ipv6_get_lladdr(dev, &ll_addr, IFA_F_TENTATIVE |
			     IFA_F_OPTIMISTIC)) {
4855 4856 4857 4858 4859 4860 4861

		/* If we're not ready, then normal ifup will take care
		 * of this. Otherwise, we need to request our rs here.
		 */
		ndisc_send_rs(dev, &ll_addr, &in6addr_linklocal_allrouters);
		update_rs = true;
	}
4862 4863

	write_lock_bh(&idev->lock);
4864

4865
	if (update_rs) {
4866
		idev->if_flags |= IF_RS_SENT;
4867 4868 4869
		idev->rs_probes = 1;
		addrconf_mod_rs_timer(idev, idev->cnf.rtr_solicit_interval);
	}
4870 4871 4872 4873

	/* Well, that's kinda nasty ... */
	list_for_each_entry(ifp, &idev->addr_list, if_list) {
		spin_lock(&ifp->lock);
4874
		if (ifp->tokenized) {
4875 4876 4877 4878 4879 4880 4881
			ifp->valid_lft = 0;
			ifp->prefered_lft = 0;
		}
		spin_unlock(&ifp->lock);
	}

	write_unlock_bh(&idev->lock);
4882
	inet6_ifinfo_notify(RTM_NEWLINK, idev);
4883
	addrconf_verify_rtnl();
4884 4885 4886
	return 0;
}

4887 4888 4889 4890 4891 4892 4893 4894 4895 4896 4897 4898 4899 4900 4901 4902
static const struct nla_policy inet6_af_policy[IFLA_INET6_MAX + 1] = {
	[IFLA_INET6_ADDR_GEN_MODE]	= { .type = NLA_U8 },
	[IFLA_INET6_TOKEN]		= { .len = sizeof(struct in6_addr) },
};

static int inet6_validate_link_af(const struct net_device *dev,
				  const struct nlattr *nla)
{
	struct nlattr *tb[IFLA_INET6_MAX + 1];

	if (dev && !__in6_dev_get(dev))
		return -EAFNOSUPPORT;

	return nla_parse_nested(tb, IFLA_INET6_MAX, nla, inet6_af_policy);
}

4903 4904 4905 4906 4907 4908 4909 4910 4911 4912 4913 4914
static int inet6_set_link_af(struct net_device *dev, const struct nlattr *nla)
{
	int err = -EINVAL;
	struct inet6_dev *idev = __in6_dev_get(dev);
	struct nlattr *tb[IFLA_INET6_MAX + 1];

	if (!idev)
		return -EAFNOSUPPORT;

	if (nla_parse_nested(tb, IFLA_INET6_MAX, nla, NULL) < 0)
		BUG();

4915
	if (tb[IFLA_INET6_TOKEN]) {
4916
		err = inet6_set_iftoken(idev, nla_data(tb[IFLA_INET6_TOKEN]));
4917 4918 4919 4920 4921 4922 4923 4924
		if (err)
			return err;
	}

	if (tb[IFLA_INET6_ADDR_GEN_MODE]) {
		u8 mode = nla_get_u8(tb[IFLA_INET6_ADDR_GEN_MODE]);

		if (mode != IN6_ADDR_GEN_MODE_EUI64 &&
4925 4926 4927 4928 4929 4930 4931
		    mode != IN6_ADDR_GEN_MODE_NONE &&
		    mode != IN6_ADDR_GEN_MODE_STABLE_PRIVACY)
			return -EINVAL;

		if (mode == IN6_ADDR_GEN_MODE_STABLE_PRIVACY &&
		    !idev->cnf.stable_secret.initialized &&
		    !dev_net(dev)->ipv6.devconf_dflt->stable_secret.initialized)
4932
			return -EINVAL;
4933

4934 4935 4936
		idev->addr_gen_mode = mode;
		err = 0;
	}
4937 4938 4939 4940

	return err;
}

4941
static int inet6_fill_ifinfo(struct sk_buff *skb, struct inet6_dev *idev,
4942
			     u32 portid, u32 seq, int event, unsigned int flags)
L
Linus Torvalds 已提交
4943
{
4944 4945 4946 4947 4948
	struct net_device *dev = idev->dev;
	struct ifinfomsg *hdr;
	struct nlmsghdr *nlh;
	void *protoinfo;

4949
	nlh = nlmsg_put(skb, portid, seq, event, sizeof(*hdr), flags);
4950
	if (!nlh)
4951
		return -EMSGSIZE;
4952 4953 4954 4955 4956 4957 4958 4959 4960

	hdr = nlmsg_data(nlh);
	hdr->ifi_family = AF_INET6;
	hdr->__ifi_pad = 0;
	hdr->ifi_type = dev->type;
	hdr->ifi_index = dev->ifindex;
	hdr->ifi_flags = dev_get_flags(dev);
	hdr->ifi_change = 0;

D
David S. Miller 已提交
4961 4962 4963 4964
	if (nla_put_string(skb, IFLA_IFNAME, dev->name) ||
	    (dev->addr_len &&
	     nla_put(skb, IFLA_ADDRESS, dev->addr_len, dev->dev_addr)) ||
	    nla_put_u32(skb, IFLA_MTU, dev->mtu) ||
4965
	    (dev->ifindex != dev_get_iflink(dev) &&
4966 4967 4968
	     nla_put_u32(skb, IFLA_LINK, dev_get_iflink(dev))) ||
	    nla_put_u8(skb, IFLA_OPERSTATE,
		       netif_running(dev) ? dev->operstate : IF_OPER_DOWN))
D
David S. Miller 已提交
4969
		goto nla_put_failure;
4970
	protoinfo = nla_nest_start(skb, IFLA_PROTINFO);
4971
	if (!protoinfo)
4972
		goto nla_put_failure;
L
Linus Torvalds 已提交
4973

T
Thomas Graf 已提交
4974
	if (inet6_fill_ifla6_attrs(skb, idev) < 0)
4975
		goto nla_put_failure;
L
Linus Torvalds 已提交
4976

4977
	nla_nest_end(skb, protoinfo);
4978 4979
	nlmsg_end(skb, nlh);
	return 0;
L
Linus Torvalds 已提交
4980

4981
nla_put_failure:
4982 4983
	nlmsg_cancel(skb, nlh);
	return -EMSGSIZE;
L
Linus Torvalds 已提交
4984 4985 4986 4987
}

static int inet6_dump_ifinfo(struct sk_buff *skb, struct netlink_callback *cb)
{
4988
	struct net *net = sock_net(skb->sk);
E
Eric Dumazet 已提交
4989
	int h, s_h;
4990
	int idx = 0, s_idx;
L
Linus Torvalds 已提交
4991 4992
	struct net_device *dev;
	struct inet6_dev *idev;
E
Eric Dumazet 已提交
4993
	struct hlist_head *head;
L
Linus Torvalds 已提交
4994

E
Eric Dumazet 已提交
4995 4996 4997 4998 4999 5000 5001
	s_h = cb->args[0];
	s_idx = cb->args[1];

	rcu_read_lock();
	for (h = s_h; h < NETDEV_HASHENTRIES; h++, s_idx = 0) {
		idx = 0;
		head = &net->dev_index_head[h];
5002
		hlist_for_each_entry_rcu(dev, head, index_hlist) {
E
Eric Dumazet 已提交
5003 5004 5005 5006 5007 5008
			if (idx < s_idx)
				goto cont;
			idev = __in6_dev_get(dev);
			if (!idev)
				goto cont;
			if (inet6_fill_ifinfo(skb, idev,
5009
					      NETLINK_CB(cb->skb).portid,
E
Eric Dumazet 已提交
5010
					      cb->nlh->nlmsg_seq,
5011
					      RTM_NEWLINK, NLM_F_MULTI) < 0)
E
Eric Dumazet 已提交
5012
				goto out;
5013
cont:
E
Eric Dumazet 已提交
5014 5015
			idx++;
		}
L
Linus Torvalds 已提交
5016
	}
E
Eric Dumazet 已提交
5017 5018 5019 5020
out:
	rcu_read_unlock();
	cb->args[1] = idx;
	cb->args[0] = h;
L
Linus Torvalds 已提交
5021 5022 5023 5024 5025 5026 5027

	return skb->len;
}

void inet6_ifinfo_notify(int event, struct inet6_dev *idev)
{
	struct sk_buff *skb;
5028
	struct net *net = dev_net(idev->dev);
5029
	int err = -ENOBUFS;
5030

5031
	skb = nlmsg_new(inet6_if_nlmsg_size(), GFP_ATOMIC);
5032
	if (!skb)
5033 5034 5035
		goto errout;

	err = inet6_fill_ifinfo(skb, idev, 0, 0, event, 0);
5036 5037 5038 5039 5040 5041
	if (err < 0) {
		/* -EMSGSIZE implies BUG in inet6_if_nlmsg_size() */
		WARN_ON(err == -EMSGSIZE);
		kfree_skb(skb);
		goto errout;
	}
5042
	rtnl_notify(skb, net, 0, RTNLGRP_IPV6_IFINFO, NULL, GFP_ATOMIC);
5043
	return;
5044 5045
errout:
	if (err < 0)
5046
		rtnl_set_sk_err(net, RTNLGRP_IPV6_IFINFO, err);
L
Linus Torvalds 已提交
5047 5048
}

5049 5050 5051 5052 5053 5054
static inline size_t inet6_prefix_nlmsg_size(void)
{
	return NLMSG_ALIGN(sizeof(struct prefixmsg))
	       + nla_total_size(sizeof(struct in6_addr))
	       + nla_total_size(sizeof(struct prefix_cacheinfo));
}
5055

L
Linus Torvalds 已提交
5056
static int inet6_fill_prefix(struct sk_buff *skb, struct inet6_dev *idev,
5057
			     struct prefix_info *pinfo, u32 portid, u32 seq,
5058
			     int event, unsigned int flags)
L
Linus Torvalds 已提交
5059
{
5060 5061
	struct prefixmsg *pmsg;
	struct nlmsghdr *nlh;
L
Linus Torvalds 已提交
5062 5063
	struct prefix_cacheinfo	ci;

5064
	nlh = nlmsg_put(skb, portid, seq, event, sizeof(*pmsg), flags);
5065
	if (!nlh)
5066
		return -EMSGSIZE;
5067 5068

	pmsg = nlmsg_data(nlh);
L
Linus Torvalds 已提交
5069
	pmsg->prefix_family = AF_INET6;
5070 5071
	pmsg->prefix_pad1 = 0;
	pmsg->prefix_pad2 = 0;
L
Linus Torvalds 已提交
5072 5073 5074
	pmsg->prefix_ifindex = idev->dev->ifindex;
	pmsg->prefix_len = pinfo->prefix_len;
	pmsg->prefix_type = pinfo->type;
5075
	pmsg->prefix_pad3 = 0;
L
Linus Torvalds 已提交
5076 5077 5078 5079 5080 5081
	pmsg->prefix_flags = 0;
	if (pinfo->onlink)
		pmsg->prefix_flags |= IF_PREFIX_ONLINK;
	if (pinfo->autoconf)
		pmsg->prefix_flags |= IF_PREFIX_AUTOCONF;

D
David S. Miller 已提交
5082 5083
	if (nla_put(skb, PREFIX_ADDRESS, sizeof(pinfo->prefix), &pinfo->prefix))
		goto nla_put_failure;
L
Linus Torvalds 已提交
5084 5085
	ci.preferred_time = ntohl(pinfo->prefered);
	ci.valid_time = ntohl(pinfo->valid);
D
David S. Miller 已提交
5086 5087
	if (nla_put(skb, PREFIX_CACHEINFO, sizeof(ci), &ci))
		goto nla_put_failure;
5088 5089
	nlmsg_end(skb, nlh);
	return 0;
L
Linus Torvalds 已提交
5090

5091
nla_put_failure:
5092 5093
	nlmsg_cancel(skb, nlh);
	return -EMSGSIZE;
L
Linus Torvalds 已提交
5094 5095
}

5096
static void inet6_prefix_notify(int event, struct inet6_dev *idev,
L
Linus Torvalds 已提交
5097 5098 5099
			 struct prefix_info *pinfo)
{
	struct sk_buff *skb;
5100
	struct net *net = dev_net(idev->dev);
5101
	int err = -ENOBUFS;
L
Linus Torvalds 已提交
5102

5103
	skb = nlmsg_new(inet6_prefix_nlmsg_size(), GFP_ATOMIC);
5104
	if (!skb)
5105 5106 5107
		goto errout;

	err = inet6_fill_prefix(skb, idev, pinfo, 0, 0, event, 0);
5108 5109 5110 5111 5112 5113
	if (err < 0) {
		/* -EMSGSIZE implies BUG in inet6_prefix_nlmsg_size() */
		WARN_ON(err == -EMSGSIZE);
		kfree_skb(skb);
		goto errout;
	}
5114 5115
	rtnl_notify(skb, net, 0, RTNLGRP_IPV6_PREFIX, NULL, GFP_ATOMIC);
	return;
5116 5117
errout:
	if (err < 0)
5118
		rtnl_set_sk_err(net, RTNLGRP_IPV6_PREFIX, err);
L
Linus Torvalds 已提交
5119 5120 5121 5122
}

static void __ipv6_ifa_notify(int event, struct inet6_ifaddr *ifp)
{
5123 5124
	struct net *net = dev_net(ifp->idev->dev);

5125 5126 5127
	if (event)
		ASSERT_RTNL();

L
Linus Torvalds 已提交
5128 5129 5130 5131
	inet6_ifa_notify(event ? : RTM_NEWADDR, ifp);

	switch (event) {
	case RTM_NEWADDR:
5132 5133 5134 5135 5136 5137 5138 5139
		/*
		 * If the address was optimistic
		 * we inserted the route at the start of
		 * our DAD process, so we don't need
		 * to do it again
		 */
		if (!(ifp->rt->rt6i_node))
			ip6_ins_rt(ifp->rt);
L
Linus Torvalds 已提交
5140 5141
		if (ifp->idev->cnf.forwarding)
			addrconf_join_anycast(ifp);
C
Cong Wang 已提交
5142
		if (!ipv6_addr_any(&ifp->peer_addr))
5143 5144
			addrconf_prefix_route(&ifp->peer_addr, 128,
					      ifp->idev->dev, 0, 0);
L
Linus Torvalds 已提交
5145 5146 5147 5148 5149
		break;
	case RTM_DELADDR:
		if (ifp->idev->cnf.forwarding)
			addrconf_leave_anycast(ifp);
		addrconf_leave_solict(ifp->idev, &ifp->addr);
C
Cong Wang 已提交
5150
		if (!ipv6_addr_any(&ifp->peer_addr)) {
5151 5152
			struct rt6_info *rt;

5153 5154
			rt = addrconf_get_prefix_route(&ifp->peer_addr, 128,
						       ifp->idev->dev, 0, 0);
5155 5156
			if (rt && ip6_del_rt(rt))
				dst_free(&rt->dst);
5157
		}
5158
		dst_hold(&ifp->rt->dst);
5159

5160
		if (ip6_del_rt(ifp->rt))
5161
			dst_free(&ifp->rt->dst);
H
Hannes Frederic Sowa 已提交
5162 5163

		rt_genid_bump_ipv6(net);
L
Linus Torvalds 已提交
5164 5165
		break;
	}
5166
	atomic_inc(&net->ipv6.dev_addr_genid);
L
Linus Torvalds 已提交
5167 5168 5169 5170
}

static void ipv6_ifa_notify(int event, struct inet6_ifaddr *ifp)
{
5171
	rcu_read_lock_bh();
L
Linus Torvalds 已提交
5172 5173
	if (likely(ifp->idev->dead == 0))
		__ipv6_ifa_notify(event, ifp);
5174
	rcu_read_unlock_bh();
L
Linus Torvalds 已提交
5175 5176 5177 5178 5179
}

#ifdef CONFIG_SYSCTL

static
5180
int addrconf_sysctl_forward(struct ctl_table *ctl, int write,
L
Linus Torvalds 已提交
5181 5182 5183 5184
			   void __user *buffer, size_t *lenp, loff_t *ppos)
{
	int *valp = ctl->data;
	int val = *valp;
E
Eric W. Biederman 已提交
5185
	loff_t pos = *ppos;
5186
	struct ctl_table lctl;
L
Linus Torvalds 已提交
5187 5188
	int ret;

5189 5190 5191 5192 5193 5194 5195 5196
	/*
	 * ctl->data points to idev->cnf.forwarding, we should
	 * not modify it until we get the rtnl lock.
	 */
	lctl = *ctl;
	lctl.data = &val;

	ret = proc_dointvec(&lctl, write, buffer, lenp, ppos);
L
Linus Torvalds 已提交
5197

5198
	if (write)
5199
		ret = addrconf_fixup_forwarding(ctl, valp, val);
E
Eric W. Biederman 已提交
5200 5201
	if (ret)
		*ppos = pos;
5202
	return ret;
L
Linus Torvalds 已提交
5203 5204
}

5205 5206 5207 5208 5209 5210 5211 5212 5213 5214 5215 5216 5217 5218 5219
static
int addrconf_sysctl_mtu(struct ctl_table *ctl, int write,
			void __user *buffer, size_t *lenp, loff_t *ppos)
{
	struct inet6_dev *idev = ctl->extra1;
	int min_mtu = IPV6_MIN_MTU;
	struct ctl_table lctl;

	lctl = *ctl;
	lctl.extra1 = &min_mtu;
	lctl.extra2 = idev ? &idev->dev->mtu : NULL;

	return proc_dointvec_minmax(&lctl, write, buffer, lenp, ppos);
}

5220 5221
static void dev_disable_change(struct inet6_dev *idev)
{
5222 5223
	struct netdev_notifier_info info;

5224 5225 5226
	if (!idev || !idev->dev)
		return;

5227
	netdev_notifier_info_init(&info, idev->dev);
5228
	if (idev->cnf.disable_ipv6)
5229
		addrconf_notify(NULL, NETDEV_DOWN, &info);
5230
	else
5231
		addrconf_notify(NULL, NETDEV_UP, &info);
5232 5233 5234 5235 5236 5237 5238
}

static void addrconf_disable_change(struct net *net, __s32 newf)
{
	struct net_device *dev;
	struct inet6_dev *idev;

5239 5240
	rcu_read_lock();
	for_each_netdev_rcu(net, dev) {
5241 5242 5243 5244 5245 5246 5247 5248
		idev = __in6_dev_get(dev);
		if (idev) {
			int changed = (!idev->cnf.disable_ipv6) ^ (!newf);
			idev->cnf.disable_ipv6 = newf;
			if (changed)
				dev_disable_change(idev);
		}
	}
5249
	rcu_read_unlock();
5250 5251
}

5252
static int addrconf_disable_ipv6(struct ctl_table *table, int *p, int newf)
5253 5254
{
	struct net *net;
5255 5256 5257 5258
	int old;

	if (!rtnl_trylock())
		return restart_syscall();
5259 5260

	net = (struct net *)table->extra2;
5261 5262
	old = *p;
	*p = newf;
5263

5264 5265
	if (p == &net->ipv6.devconf_dflt->disable_ipv6) {
		rtnl_unlock();
5266
		return 0;
E
Eric W. Biederman 已提交
5267
	}
5268 5269 5270 5271

	if (p == &net->ipv6.devconf_all->disable_ipv6) {
		net->ipv6.devconf_dflt->disable_ipv6 = newf;
		addrconf_disable_change(net, newf);
5272
	} else if ((!newf) ^ (!old))
5273 5274 5275 5276 5277 5278 5279
		dev_disable_change((struct inet6_dev *)table->extra1);

	rtnl_unlock();
	return 0;
}

static
5280
int addrconf_sysctl_disable(struct ctl_table *ctl, int write,
5281 5282 5283 5284
			    void __user *buffer, size_t *lenp, loff_t *ppos)
{
	int *valp = ctl->data;
	int val = *valp;
E
Eric W. Biederman 已提交
5285
	loff_t pos = *ppos;
5286
	struct ctl_table lctl;
5287 5288
	int ret;

5289 5290 5291 5292 5293 5294 5295 5296
	/*
	 * ctl->data points to idev->cnf.disable_ipv6, we should
	 * not modify it until we get the rtnl lock.
	 */
	lctl = *ctl;
	lctl.data = &val;

	ret = proc_dointvec(&lctl, write, buffer, lenp, ppos);
5297 5298 5299

	if (write)
		ret = addrconf_disable_ipv6(ctl, valp, val);
E
Eric W. Biederman 已提交
5300 5301
	if (ret)
		*ppos = pos;
5302 5303 5304
	return ret;
}

5305 5306 5307 5308 5309 5310 5311 5312 5313 5314 5315 5316 5317 5318 5319 5320 5321 5322 5323 5324 5325 5326 5327 5328 5329 5330 5331 5332 5333 5334 5335 5336 5337 5338 5339 5340 5341 5342 5343
static
int addrconf_sysctl_proxy_ndp(struct ctl_table *ctl, int write,
			      void __user *buffer, size_t *lenp, loff_t *ppos)
{
	int *valp = ctl->data;
	int ret;
	int old, new;

	old = *valp;
	ret = proc_dointvec(ctl, write, buffer, lenp, ppos);
	new = *valp;

	if (write && old != new) {
		struct net *net = ctl->extra2;

		if (!rtnl_trylock())
			return restart_syscall();

		if (valp == &net->ipv6.devconf_dflt->proxy_ndp)
			inet6_netconf_notify_devconf(net, NETCONFA_PROXY_NEIGH,
						     NETCONFA_IFINDEX_DEFAULT,
						     net->ipv6.devconf_dflt);
		else if (valp == &net->ipv6.devconf_all->proxy_ndp)
			inet6_netconf_notify_devconf(net, NETCONFA_PROXY_NEIGH,
						     NETCONFA_IFINDEX_ALL,
						     net->ipv6.devconf_all);
		else {
			struct inet6_dev *idev = ctl->extra1;

			inet6_netconf_notify_devconf(net, NETCONFA_PROXY_NEIGH,
						     idev->dev->ifindex,
						     &idev->cnf);
		}
		rtnl_unlock();
	}

	return ret;
}

5344 5345 5346 5347 5348 5349 5350 5351
static int addrconf_sysctl_stable_secret(struct ctl_table *ctl, int write,
					 void __user *buffer, size_t *lenp,
					 loff_t *ppos)
{
	int err;
	struct in6_addr addr;
	char str[IPV6_MAX_STRLEN];
	struct ctl_table lctl = *ctl;
5352
	struct net *net = ctl->extra2;
5353 5354
	struct ipv6_stable_secret *secret = ctl->data;

5355 5356 5357
	if (&net->ipv6.devconf_all->stable_secret == ctl->data)
		return -EIO;

5358 5359 5360 5361 5362 5363 5364 5365 5366 5367 5368 5369 5370 5371 5372 5373 5374 5375 5376 5377 5378 5379 5380 5381 5382 5383 5384 5385 5386 5387 5388 5389
	lctl.maxlen = IPV6_MAX_STRLEN;
	lctl.data = str;

	if (!rtnl_trylock())
		return restart_syscall();

	if (!write && !secret->initialized) {
		err = -EIO;
		goto out;
	}

	if (!write) {
		err = snprintf(str, sizeof(str), "%pI6",
			       &secret->secret);
		if (err >= sizeof(str)) {
			err = -EIO;
			goto out;
		}
	}

	err = proc_dostring(&lctl, write, buffer, lenp, ppos);
	if (err || !write)
		goto out;

	if (in6_pton(str, -1, addr.in6_u.u6_addr8, -1, NULL) != 1) {
		err = -EIO;
		goto out;
	}

	secret->initialized = true;
	secret->secret = addr;

5390 5391 5392 5393 5394 5395 5396 5397 5398 5399 5400 5401 5402 5403 5404 5405 5406
	if (&net->ipv6.devconf_dflt->stable_secret == ctl->data) {
		struct net_device *dev;

		for_each_netdev(net, dev) {
			struct inet6_dev *idev = __in6_dev_get(dev);

			if (idev) {
				idev->addr_gen_mode =
					IN6_ADDR_GEN_MODE_STABLE_PRIVACY;
			}
		}
	} else {
		struct inet6_dev *idev = ctl->extra1;

		idev->addr_gen_mode = IN6_ADDR_GEN_MODE_STABLE_PRIVACY;
	}

5407 5408 5409 5410 5411
out:
	rtnl_unlock();

	return err;
}
5412

5413 5414 5415 5416 5417 5418 5419 5420 5421 5422 5423 5424 5425 5426 5427 5428 5429 5430 5431 5432 5433 5434 5435 5436 5437 5438 5439 5440
static
int addrconf_sysctl_ignore_routes_with_linkdown(struct ctl_table *ctl,
						int write,
						void __user *buffer,
						size_t *lenp,
						loff_t *ppos)
{
	int *valp = ctl->data;
	int val = *valp;
	loff_t pos = *ppos;
	struct ctl_table lctl;
	int ret;

	/* ctl->data points to idev->cnf.ignore_routes_when_linkdown
	 * we should not modify it until we get the rtnl lock.
	 */
	lctl = *ctl;
	lctl.data = &val;

	ret = proc_dointvec(&lctl, write, buffer, lenp, ppos);

	if (write)
		ret = addrconf_fixup_linkdown(ctl, valp, val);
	if (ret)
		*ppos = pos;
	return ret;
}

L
Linus Torvalds 已提交
5441 5442 5443
static struct addrconf_sysctl_table
{
	struct ctl_table_header *sysctl_header;
5444
	struct ctl_table addrconf_vars[DEVCONF_MAX+1];
5445
} addrconf_sysctl __read_mostly = {
L
Linus Torvalds 已提交
5446 5447
	.sysctl_header = NULL,
	.addrconf_vars = {
5448
		{
S
Stephen Hemminger 已提交
5449 5450 5451 5452 5453
			.procname	= "forwarding",
			.data		= &ipv6_devconf.forwarding,
			.maxlen		= sizeof(int),
			.mode		= 0644,
			.proc_handler	= addrconf_sysctl_forward,
L
Linus Torvalds 已提交
5454 5455
		},
		{
S
Stephen Hemminger 已提交
5456 5457 5458 5459 5460
			.procname	= "hop_limit",
			.data		= &ipv6_devconf.hop_limit,
			.maxlen		= sizeof(int),
			.mode		= 0644,
			.proc_handler	= proc_dointvec,
L
Linus Torvalds 已提交
5461 5462
		},
		{
S
Stephen Hemminger 已提交
5463 5464 5465 5466
			.procname	= "mtu",
			.data		= &ipv6_devconf.mtu6,
			.maxlen		= sizeof(int),
			.mode		= 0644,
5467
			.proc_handler	= addrconf_sysctl_mtu,
L
Linus Torvalds 已提交
5468 5469
		},
		{
S
Stephen Hemminger 已提交
5470 5471 5472 5473 5474
			.procname	= "accept_ra",
			.data		= &ipv6_devconf.accept_ra,
			.maxlen		= sizeof(int),
			.mode		= 0644,
			.proc_handler	= proc_dointvec,
L
Linus Torvalds 已提交
5475 5476
		},
		{
S
Stephen Hemminger 已提交
5477 5478 5479 5480 5481
			.procname	= "accept_redirects",
			.data		= &ipv6_devconf.accept_redirects,
			.maxlen		= sizeof(int),
			.mode		= 0644,
			.proc_handler	= proc_dointvec,
L
Linus Torvalds 已提交
5482 5483
		},
		{
S
Stephen Hemminger 已提交
5484 5485 5486 5487 5488
			.procname	= "autoconf",
			.data		= &ipv6_devconf.autoconf,
			.maxlen		= sizeof(int),
			.mode		= 0644,
			.proc_handler	= proc_dointvec,
L
Linus Torvalds 已提交
5489 5490
		},
		{
S
Stephen Hemminger 已提交
5491 5492 5493 5494 5495
			.procname	= "dad_transmits",
			.data		= &ipv6_devconf.dad_transmits,
			.maxlen		= sizeof(int),
			.mode		= 0644,
			.proc_handler	= proc_dointvec,
L
Linus Torvalds 已提交
5496 5497
		},
		{
S
Stephen Hemminger 已提交
5498 5499 5500 5501 5502
			.procname	= "router_solicitations",
			.data		= &ipv6_devconf.rtr_solicits,
			.maxlen		= sizeof(int),
			.mode		= 0644,
			.proc_handler	= proc_dointvec,
L
Linus Torvalds 已提交
5503 5504
		},
		{
S
Stephen Hemminger 已提交
5505 5506 5507 5508 5509
			.procname	= "router_solicitation_interval",
			.data		= &ipv6_devconf.rtr_solicit_interval,
			.maxlen		= sizeof(int),
			.mode		= 0644,
			.proc_handler	= proc_dointvec_jiffies,
L
Linus Torvalds 已提交
5510 5511
		},
		{
S
Stephen Hemminger 已提交
5512 5513 5514 5515 5516
			.procname	= "router_solicitation_delay",
			.data		= &ipv6_devconf.rtr_solicit_delay,
			.maxlen		= sizeof(int),
			.mode		= 0644,
			.proc_handler	= proc_dointvec_jiffies,
L
Linus Torvalds 已提交
5517 5518
		},
		{
S
Stephen Hemminger 已提交
5519 5520 5521 5522 5523
			.procname	= "force_mld_version",
			.data		= &ipv6_devconf.force_mld_version,
			.maxlen		= sizeof(int),
			.mode		= 0644,
			.proc_handler	= proc_dointvec,
L
Linus Torvalds 已提交
5524
		},
5525 5526 5527 5528 5529 5530 5531 5532 5533 5534 5535 5536 5537 5538 5539 5540
		{
			.procname	= "mldv1_unsolicited_report_interval",
			.data		=
				&ipv6_devconf.mldv1_unsolicited_report_interval,
			.maxlen		= sizeof(int),
			.mode		= 0644,
			.proc_handler	= proc_dointvec_ms_jiffies,
		},
		{
			.procname	= "mldv2_unsolicited_report_interval",
			.data		=
				&ipv6_devconf.mldv2_unsolicited_report_interval,
			.maxlen		= sizeof(int),
			.mode		= 0644,
			.proc_handler	= proc_dointvec_ms_jiffies,
		},
L
Linus Torvalds 已提交
5541
		{
S
Stephen Hemminger 已提交
5542 5543 5544 5545 5546
			.procname	= "use_tempaddr",
			.data		= &ipv6_devconf.use_tempaddr,
			.maxlen		= sizeof(int),
			.mode		= 0644,
			.proc_handler	= proc_dointvec,
L
Linus Torvalds 已提交
5547 5548
		},
		{
S
Stephen Hemminger 已提交
5549 5550 5551 5552 5553
			.procname	= "temp_valid_lft",
			.data		= &ipv6_devconf.temp_valid_lft,
			.maxlen		= sizeof(int),
			.mode		= 0644,
			.proc_handler	= proc_dointvec,
L
Linus Torvalds 已提交
5554 5555
		},
		{
S
Stephen Hemminger 已提交
5556 5557 5558 5559 5560
			.procname	= "temp_prefered_lft",
			.data		= &ipv6_devconf.temp_prefered_lft,
			.maxlen		= sizeof(int),
			.mode		= 0644,
			.proc_handler	= proc_dointvec,
L
Linus Torvalds 已提交
5561 5562
		},
		{
S
Stephen Hemminger 已提交
5563 5564 5565 5566 5567
			.procname	= "regen_max_retry",
			.data		= &ipv6_devconf.regen_max_retry,
			.maxlen		= sizeof(int),
			.mode		= 0644,
			.proc_handler	= proc_dointvec,
L
Linus Torvalds 已提交
5568 5569
		},
		{
S
Stephen Hemminger 已提交
5570 5571 5572 5573 5574
			.procname	= "max_desync_factor",
			.data		= &ipv6_devconf.max_desync_factor,
			.maxlen		= sizeof(int),
			.mode		= 0644,
			.proc_handler	= proc_dointvec,
L
Linus Torvalds 已提交
5575 5576
		},
		{
S
Stephen Hemminger 已提交
5577 5578 5579 5580 5581
			.procname	= "max_addresses",
			.data		= &ipv6_devconf.max_addresses,
			.maxlen		= sizeof(int),
			.mode		= 0644,
			.proc_handler	= proc_dointvec,
L
Linus Torvalds 已提交
5582
		},
5583
		{
S
Stephen Hemminger 已提交
5584 5585 5586 5587 5588
			.procname	= "accept_ra_defrtr",
			.data		= &ipv6_devconf.accept_ra_defrtr,
			.maxlen		= sizeof(int),
			.mode		= 0644,
			.proc_handler	= proc_dointvec,
5589
		},
5590 5591 5592 5593 5594 5595 5596
		{
			.procname	= "accept_ra_min_hop_limit",
			.data		= &ipv6_devconf.accept_ra_min_hop_limit,
			.maxlen		= sizeof(int),
			.mode		= 0644,
			.proc_handler	= proc_dointvec,
		},
5597
		{
S
Stephen Hemminger 已提交
5598 5599 5600 5601 5602
			.procname	= "accept_ra_pinfo",
			.data		= &ipv6_devconf.accept_ra_pinfo,
			.maxlen		= sizeof(int),
			.mode		= 0644,
			.proc_handler	= proc_dointvec,
5603
		},
5604 5605
#ifdef CONFIG_IPV6_ROUTER_PREF
		{
S
Stephen Hemminger 已提交
5606 5607 5608 5609 5610
			.procname	= "accept_ra_rtr_pref",
			.data		= &ipv6_devconf.accept_ra_rtr_pref,
			.maxlen		= sizeof(int),
			.mode		= 0644,
			.proc_handler	= proc_dointvec,
5611
		},
5612
		{
S
Stephen Hemminger 已提交
5613 5614 5615 5616 5617
			.procname	= "router_probe_interval",
			.data		= &ipv6_devconf.rtr_probe_interval,
			.maxlen		= sizeof(int),
			.mode		= 0644,
			.proc_handler	= proc_dointvec_jiffies,
5618
		},
N
Neil Horman 已提交
5619
#ifdef CONFIG_IPV6_ROUTE_INFO
5620
		{
S
Stephen Hemminger 已提交
5621 5622 5623 5624 5625
			.procname	= "accept_ra_rt_info_max_plen",
			.data		= &ipv6_devconf.accept_ra_rt_info_max_plen,
			.maxlen		= sizeof(int),
			.mode		= 0644,
			.proc_handler	= proc_dointvec,
5626 5627
		},
#endif
5628
#endif
5629
		{
S
Stephen Hemminger 已提交
5630 5631 5632 5633
			.procname	= "proxy_ndp",
			.data		= &ipv6_devconf.proxy_ndp,
			.maxlen		= sizeof(int),
			.mode		= 0644,
5634
			.proc_handler	= addrconf_sysctl_proxy_ndp,
5635
		},
5636
		{
S
Stephen Hemminger 已提交
5637 5638 5639 5640 5641
			.procname	= "accept_source_route",
			.data		= &ipv6_devconf.accept_source_route,
			.maxlen		= sizeof(int),
			.mode		= 0644,
			.proc_handler	= proc_dointvec,
5642
		},
5643 5644
#ifdef CONFIG_IPV6_OPTIMISTIC_DAD
		{
S
Stephen Hemminger 已提交
5645 5646 5647 5648 5649
			.procname       = "optimistic_dad",
			.data           = &ipv6_devconf.optimistic_dad,
			.maxlen         = sizeof(int),
			.mode           = 0644,
			.proc_handler   = proc_dointvec,
5650 5651

		},
5652 5653 5654 5655 5656 5657 5658 5659
		{
			.procname       = "use_optimistic",
			.data           = &ipv6_devconf.use_optimistic,
			.maxlen         = sizeof(int),
			.mode           = 0644,
			.proc_handler   = proc_dointvec,

		},
5660 5661 5662
#endif
#ifdef CONFIG_IPV6_MROUTE
		{
S
Stephen Hemminger 已提交
5663 5664 5665 5666 5667
			.procname	= "mc_forwarding",
			.data		= &ipv6_devconf.mc_forwarding,
			.maxlen		= sizeof(int),
			.mode		= 0444,
			.proc_handler	= proc_dointvec,
5668
		},
5669
#endif
5670
		{
S
Stephen Hemminger 已提交
5671 5672 5673 5674 5675
			.procname	= "disable_ipv6",
			.data		= &ipv6_devconf.disable_ipv6,
			.maxlen		= sizeof(int),
			.mode		= 0644,
			.proc_handler	= addrconf_sysctl_disable,
5676
		},
5677
		{
S
Stephen Hemminger 已提交
5678 5679 5680 5681 5682
			.procname	= "accept_dad",
			.data		= &ipv6_devconf.accept_dad,
			.maxlen		= sizeof(int),
			.mode		= 0644,
			.proc_handler	= proc_dointvec,
5683
		},
5684 5685 5686 5687 5688 5689 5690
		{
			.procname       = "force_tllao",
			.data           = &ipv6_devconf.force_tllao,
			.maxlen         = sizeof(int),
			.mode           = 0644,
			.proc_handler   = proc_dointvec
		},
5691 5692 5693 5694 5695 5696 5697
		{
			.procname       = "ndisc_notify",
			.data           = &ipv6_devconf.ndisc_notify,
			.maxlen         = sizeof(int),
			.mode           = 0644,
			.proc_handler   = proc_dointvec
		},
5698 5699 5700 5701 5702 5703 5704
		{
			.procname	= "suppress_frag_ndisc",
			.data		= &ipv6_devconf.suppress_frag_ndisc,
			.maxlen		= sizeof(int),
			.mode		= 0644,
			.proc_handler	= proc_dointvec
		},
5705 5706 5707 5708 5709 5710 5711
		{
			.procname	= "accept_ra_from_local",
			.data		= &ipv6_devconf.accept_ra_from_local,
			.maxlen		= sizeof(int),
			.mode		= 0644,
			.proc_handler	= proc_dointvec,
		},
5712 5713 5714 5715 5716 5717 5718
		{
			.procname	= "accept_ra_mtu",
			.data		= &ipv6_devconf.accept_ra_mtu,
			.maxlen		= sizeof(int),
			.mode		= 0644,
			.proc_handler	= proc_dointvec,
		},
5719 5720 5721 5722 5723 5724 5725
		{
			.procname	= "stable_secret",
			.data		= &ipv6_devconf.stable_secret,
			.maxlen		= IPV6_MAX_STRLEN,
			.mode		= 0600,
			.proc_handler	= addrconf_sysctl_stable_secret,
		},
5726 5727 5728 5729 5730 5731
		{
			.procname       = "use_oif_addrs_only",
			.data           = &ipv6_devconf.use_oif_addrs_only,
			.maxlen         = sizeof(int),
			.mode           = 0644,
			.proc_handler   = proc_dointvec,
5732 5733 5734 5735 5736 5737 5738
		},
		{
			.procname	= "ignore_routes_with_linkdown",
			.data		= &ipv6_devconf.ignore_routes_with_linkdown,
			.maxlen		= sizeof(int),
			.mode		= 0644,
			.proc_handler	= addrconf_sysctl_ignore_routes_with_linkdown,
5739
		},
L
Linus Torvalds 已提交
5740
		{
5741
			/* sentinel */
L
Linus Torvalds 已提交
5742 5743 5744 5745
		}
	},
};

5746
static int __addrconf_sysctl_register(struct net *net, char *dev_name,
5747
		struct inet6_dev *idev, struct ipv6_devconf *p)
L
Linus Torvalds 已提交
5748 5749 5750
{
	int i;
	struct addrconf_sysctl_table *t;
5751
	char path[sizeof("net/ipv6/conf/") + IFNAMSIZ];
5752

A
Arnaldo Carvalho de Melo 已提交
5753
	t = kmemdup(&addrconf_sysctl, sizeof(*t), GFP_KERNEL);
5754
	if (!t)
5755 5756
		goto out;

S
Stephen Hemminger 已提交
5757
	for (i = 0; t->addrconf_vars[i].data; i++) {
S
Stephen Hemminger 已提交
5758
		t->addrconf_vars[i].data += (char *)p - (char *)&ipv6_devconf;
L
Linus Torvalds 已提交
5759
		t->addrconf_vars[i].extra1 = idev; /* embedded; no ref */
5760
		t->addrconf_vars[i].extra2 = net;
L
Linus Torvalds 已提交
5761 5762
	}

5763
	snprintf(path, sizeof(path), "net/ipv6/conf/%s", dev_name);
L
Linus Torvalds 已提交
5764

5765
	t->sysctl_header = register_net_sysctl(net, path, t->addrconf_vars);
5766
	if (!t->sysctl_header)
5767
		goto free;
5768 5769

	p->sysctl = t;
5770
	return 0;
L
Linus Torvalds 已提交
5771

5772
free:
L
Linus Torvalds 已提交
5773
	kfree(t);
5774
out:
5775
	return -ENOBUFS;
L
Linus Torvalds 已提交
5776 5777
}

5778 5779 5780 5781
static void __addrconf_sysctl_unregister(struct ipv6_devconf *p)
{
	struct addrconf_sysctl_table *t;

5782
	if (!p->sysctl)
5783 5784 5785 5786
		return;

	t = p->sysctl;
	p->sysctl = NULL;
5787
	unregister_net_sysctl_table(t->sysctl_header);
5788 5789 5790
	kfree(t);
}

5791
static int addrconf_sysctl_register(struct inet6_dev *idev)
5792
{
5793 5794 5795 5796 5797 5798 5799 5800 5801 5802 5803 5804 5805 5806 5807
	int err;

	if (!sysctl_dev_name_is_allowed(idev->dev->name))
		return -EINVAL;

	err = neigh_sysctl_register(idev->dev, idev->nd_parms,
				    &ndisc_ifinfo_sysctl_change);
	if (err)
		return err;
	err = __addrconf_sysctl_register(dev_net(idev->dev), idev->dev->name,
					 idev, &idev->cnf);
	if (err)
		neigh_sysctl_unregister(idev->nd_parms);

	return err;
5808 5809
}

5810
static void addrconf_sysctl_unregister(struct inet6_dev *idev)
L
Linus Torvalds 已提交
5811
{
5812 5813
	__addrconf_sysctl_unregister(&idev->cnf);
	neigh_sysctl_unregister(idev->nd_parms);
L
Linus Torvalds 已提交
5814 5815 5816 5817 5818
}


#endif

5819
static int __net_init addrconf_init_net(struct net *net)
5820
{
5821
	int err = -ENOMEM;
5822 5823
	struct ipv6_devconf *all, *dflt;

5824
	all = kmemdup(&ipv6_devconf, sizeof(ipv6_devconf), GFP_KERNEL);
5825
	if (!all)
5826
		goto err_alloc_all;
5827

5828
	dflt = kmemdup(&ipv6_devconf_dflt, sizeof(ipv6_devconf_dflt), GFP_KERNEL);
5829
	if (!dflt)
5830
		goto err_alloc_dflt;
5831

5832 5833 5834
	/* these will be inherited by all namespaces */
	dflt->autoconf = ipv6_defaults.autoconf;
	dflt->disable_ipv6 = ipv6_defaults.disable_ipv6;
5835

5836 5837 5838
	dflt->stable_secret.initialized = false;
	all->stable_secret.initialized = false;

5839 5840 5841 5842
	net->ipv6.devconf_all = all;
	net->ipv6.devconf_dflt = dflt;

#ifdef CONFIG_SYSCTL
5843
	err = __addrconf_sysctl_register(net, "all", NULL, all);
5844 5845 5846
	if (err < 0)
		goto err_reg_all;

5847
	err = __addrconf_sysctl_register(net, "default", NULL, dflt);
5848 5849 5850 5851 5852 5853 5854 5855 5856 5857 5858 5859 5860 5861 5862 5863 5864
	if (err < 0)
		goto err_reg_dflt;
#endif
	return 0;

#ifdef CONFIG_SYSCTL
err_reg_dflt:
	__addrconf_sysctl_unregister(all);
err_reg_all:
	kfree(dflt);
#endif
err_alloc_dflt:
	kfree(all);
err_alloc_all:
	return err;
}

5865
static void __net_exit addrconf_exit_net(struct net *net)
5866 5867 5868 5869 5870
{
#ifdef CONFIG_SYSCTL
	__addrconf_sysctl_unregister(net->ipv6.devconf_dflt);
	__addrconf_sysctl_unregister(net->ipv6.devconf_all);
#endif
Z
zhuyj 已提交
5871 5872
	kfree(net->ipv6.devconf_dflt);
	kfree(net->ipv6.devconf_all);
5873 5874 5875 5876 5877 5878 5879
}

static struct pernet_operations addrconf_ops = {
	.init = addrconf_init_net,
	.exit = addrconf_exit_net,
};

5880
static struct rtnl_af_ops inet6_ops __read_mostly = {
T
Thomas Graf 已提交
5881 5882 5883
	.family		  = AF_INET6,
	.fill_link_af	  = inet6_fill_link_af,
	.get_link_af_size = inet6_get_link_af_size,
5884
	.validate_link_af = inet6_validate_link_af,
5885
	.set_link_af	  = inet6_set_link_af,
T
Thomas Graf 已提交
5886 5887
};

L
Linus Torvalds 已提交
5888 5889 5890 5891 5892 5893
/*
 *	Init / cleanup code
 */

int __init addrconf_init(void)
{
5894
	struct inet6_dev *idev;
5895
	int i, err;
5896

S
Stephen Hemminger 已提交
5897 5898
	err = ipv6_addr_label_init();
	if (err < 0) {
5899 5900
		pr_crit("%s: cannot initialize default policy table: %d\n",
			__func__, err);
5901
		goto out;
5902
	}
L
Linus Torvalds 已提交
5903

5904 5905 5906
	err = register_pernet_subsys(&addrconf_ops);
	if (err < 0)
		goto out_addrlabel;
5907

5908 5909 5910 5911 5912 5913
	addrconf_wq = create_workqueue("ipv6_addrconf");
	if (!addrconf_wq) {
		err = -ENOMEM;
		goto out_nowq;
	}

L
Linus Torvalds 已提交
5914 5915 5916 5917 5918 5919 5920 5921 5922 5923 5924 5925 5926 5927 5928 5929 5930 5931 5932
	/* The addrconf netdev notifier requires that loopback_dev
	 * has it's ipv6 private information allocated and setup
	 * before it can bring up and give link-local addresses
	 * to other devices which are up.
	 *
	 * Unfortunately, loopback_dev is not necessarily the first
	 * entry in the global dev_base list of net devices.  In fact,
	 * it is likely to be the very last entry on that list.
	 * So this causes the notifier registry below to try and
	 * give link-local addresses to all devices besides loopback_dev
	 * first, then loopback_dev, which cases all the non-loopback_dev
	 * devices to fail to get a link-local address.
	 *
	 * So, as a temporary fix, allocate the ipv6 structure for
	 * loopback_dev first by hand.
	 * Longer term, all of the dependencies ipv6 has upon the loopback
	 * device and it being up should be removed.
	 */
	rtnl_lock();
5933
	idev = ipv6_add_dev(init_net.loopback_dev);
L
Linus Torvalds 已提交
5934
	rtnl_unlock();
5935 5936
	if (IS_ERR(idev)) {
		err = PTR_ERR(idev);
5937
		goto errlo;
5938
	}
L
Linus Torvalds 已提交
5939

5940 5941 5942
	for (i = 0; i < IN6_ADDR_HSIZE; i++)
		INIT_HLIST_HEAD(&inet6_addr_lst[i]);

L
Linus Torvalds 已提交
5943 5944
	register_netdevice_notifier(&ipv6_dev_notf);

5945
	addrconf_verify();
5946

5947
	rtnl_af_register(&inet6_ops);
T
Thomas Graf 已提交
5948

5949 5950
	err = __rtnl_register(PF_INET6, RTM_GETLINK, NULL, inet6_dump_ifinfo,
			      NULL);
5951 5952 5953 5954
	if (err < 0)
		goto errout;

	/* Only the first call to __rtnl_register can fail */
5955 5956 5957 5958 5959 5960 5961 5962
	__rtnl_register(PF_INET6, RTM_NEWADDR, inet6_rtm_newaddr, NULL, NULL);
	__rtnl_register(PF_INET6, RTM_DELADDR, inet6_rtm_deladdr, NULL, NULL);
	__rtnl_register(PF_INET6, RTM_GETADDR, inet6_rtm_getaddr,
			inet6_dump_ifaddr, NULL);
	__rtnl_register(PF_INET6, RTM_GETMULTICAST, NULL,
			inet6_dump_ifmcaddr, NULL);
	__rtnl_register(PF_INET6, RTM_GETANYCAST, NULL,
			inet6_dump_ifacaddr, NULL);
5963
	__rtnl_register(PF_INET6, RTM_GETNETCONF, inet6_netconf_get_devconf,
5964
			inet6_netconf_dump_devconf, NULL);
5965

5966 5967
	ipv6_addr_label_rtnl_register();

L
Linus Torvalds 已提交
5968
	return 0;
5969
errout:
T
Thomas Graf 已提交
5970
	rtnl_af_unregister(&inet6_ops);
5971
	unregister_netdevice_notifier(&ipv6_dev_notf);
5972
errlo:
5973 5974
	destroy_workqueue(addrconf_wq);
out_nowq:
5975
	unregister_pernet_subsys(&addrconf_ops);
5976 5977 5978
out_addrlabel:
	ipv6_addr_label_cleanup();
out:
5979
	return err;
L
Linus Torvalds 已提交
5980 5981
}

5982
void addrconf_cleanup(void)
L
Linus Torvalds 已提交
5983
{
5984
	struct net_device *dev;
L
Linus Torvalds 已提交
5985 5986 5987
	int i;

	unregister_netdevice_notifier(&ipv6_dev_notf);
5988
	unregister_pernet_subsys(&addrconf_ops);
5989
	ipv6_addr_label_cleanup();
L
Linus Torvalds 已提交
5990 5991 5992

	rtnl_lock();

T
Thomas Graf 已提交
5993 5994
	__rtnl_af_unregister(&inet6_ops);

5995 5996 5997 5998 5999 6000 6001 6002
	/* clean dev list */
	for_each_netdev(&init_net, dev) {
		if (__in6_dev_get(dev) == NULL)
			continue;
		addrconf_ifdown(dev, 1);
	}
	addrconf_ifdown(init_net.loopback_dev, 2);

L
Linus Torvalds 已提交
6003 6004 6005
	/*
	 *	Check hash table.
	 */
6006
	spin_lock_bh(&addrconf_hash_lock);
6007 6008
	for (i = 0; i < IN6_ADDR_HSIZE; i++)
		WARN_ON(!hlist_empty(&inet6_addr_lst[i]));
6009
	spin_unlock_bh(&addrconf_hash_lock);
6010
	cancel_delayed_work(&addr_chk_work);
L
Linus Torvalds 已提交
6011
	rtnl_unlock();
6012 6013

	destroy_workqueue(addrconf_wq);
L
Linus Torvalds 已提交
6014
}