addrconf.c 164.3 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>
46
#include <linux/sched/signal.h>
L
Linus Torvalds 已提交
47 48 49
#include <linux/socket.h>
#include <linux/sockios.h>
#include <linux/net.h>
50
#include <linux/inet.h>
L
Linus Torvalds 已提交
51 52
#include <linux/in6.h>
#include <linux/netdevice.h>
53
#include <linux/if_addr.h>
L
Linus Torvalds 已提交
54 55 56 57 58 59
#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>
60
#include <linux/slab.h>
L
Linus Torvalds 已提交
61 62 63
#ifdef CONFIG_SYSCTL
#include <linux/sysctl.h>
#endif
64
#include <linux/capability.h>
L
Linus Torvalds 已提交
65 66
#include <linux/delay.h>
#include <linux/notifier.h>
67
#include <linux/string.h>
E
Eric Dumazet 已提交
68
#include <linux/hash.h>
L
Linus Torvalds 已提交
69

70
#include <net/net_namespace.h>
L
Linus Torvalds 已提交
71 72 73
#include <net/sock.h>
#include <net/snmp.h>

74
#include <net/6lowpan.h>
75
#include <net/firewire.h>
L
Linus Torvalds 已提交
76 77 78 79 80 81 82
#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>
83
#include <net/netlink.h>
84
#include <net/pkt_sched.h>
D
David Ahern 已提交
85
#include <net/l3mdev.h>
L
Linus Torvalds 已提交
86 87
#include <linux/if_tunnel.h>
#include <linux/rtnetlink.h>
88
#include <linux/netconf.h>
L
Linus Torvalds 已提交
89
#include <linux/random.h>
S
Stephen Hemminger 已提交
90
#include <linux/uaccess.h>
91
#include <asm/unaligned.h>
L
Linus Torvalds 已提交
92 93 94

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

#define	INFINITY_LIFE_TIME	0xFFFFFFFF
98

99 100 101
#define IPV6_MAX_STRLEN \
	sizeof("ffff:ffff:ffff:ffff:ffff:ffff:255.255.255.255")

102 103 104 105
static inline u32 cstamp_delta(unsigned long cstamp)
{
	return (cstamp - INITIAL_JIFFIES) * 100UL / HZ;
}
L
Linus Torvalds 已提交
106

107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127
static inline s32 rfc3315_s14_backoff_init(s32 irt)
{
	/* multiply 'initial retransmission time' by 0.9 .. 1.1 */
	u64 tmp = (900000 + prandom_u32() % 200001) * (u64)irt;
	do_div(tmp, 1000000);
	return (s32)tmp;
}

static inline s32 rfc3315_s14_backoff_update(s32 rt, s32 mrt)
{
	/* multiply 'retransmission timeout' by 1.9 .. 2.1 */
	u64 tmp = (1900000 + prandom_u32() % 200001) * (u64)rt;
	do_div(tmp, 1000000);
	if ((s32)tmp > mrt) {
		/* multiply 'maximum retransmission time' by 0.9 .. 1.1 */
		tmp = (900000 + prandom_u32() % 200001) * (u64)mrt;
		do_div(tmp, 1000000);
	}
	return (s32)tmp;
}

L
Linus Torvalds 已提交
128
#ifdef CONFIG_SYSCTL
129
static int addrconf_sysctl_register(struct inet6_dev *idev);
130 131
static void addrconf_sysctl_unregister(struct inet6_dev *idev);
#else
132
static inline int addrconf_sysctl_register(struct inet6_dev *idev)
133
{
134
	return 0;
135 136 137 138 139
}

static inline void addrconf_sysctl_unregister(struct inet6_dev *idev)
{
}
L
Linus Torvalds 已提交
140 141
#endif

142 143
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
144

145
static int ipv6_generate_eui64(u8 *eui, struct net_device *dev);
146
static int ipv6_count_addresses(const struct inet6_dev *idev);
147 148 149
static int ipv6_generate_stable_address(struct in6_addr *addr,
					u8 dad_count,
					const struct inet6_dev *idev);
L
Linus Torvalds 已提交
150

151 152
#define IN6_ADDR_HSIZE_SHIFT	8
#define IN6_ADDR_HSIZE		(1 << IN6_ADDR_HSIZE_SHIFT)
L
Linus Torvalds 已提交
153 154 155
/*
 *	Configured unicast address hash table
 */
156
static struct hlist_head inet6_addr_lst[IN6_ADDR_HSIZE];
157
static DEFINE_SPINLOCK(addrconf_hash_lock);
L
Linus Torvalds 已提交
158

159 160 161
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
162

163 164
static struct workqueue_struct *addrconf_wq;
static DECLARE_DELAYED_WORK(addr_chk_work, addrconf_verify_work);
L
Linus Torvalds 已提交
165 166 167 168

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

169 170
static void addrconf_type_change(struct net_device *dev,
				 unsigned long event);
L
Linus Torvalds 已提交
171 172
static int addrconf_ifdown(struct net_device *dev, int how);

173
static struct fib6_info *addrconf_get_prefix_route(const struct in6_addr *pfx,
174 175 176 177
						  int plen,
						  const struct net_device *dev,
						  u32 flags, u32 noflags);

178
static void addrconf_dad_start(struct inet6_ifaddr *ifp);
179
static void addrconf_dad_work(struct work_struct *w);
180 181
static void addrconf_dad_completed(struct inet6_ifaddr *ifp, bool bump_id,
				   bool send_na);
182
static void addrconf_dad_run(struct inet6_dev *idev);
183
static void addrconf_rs_timer(struct timer_list *t);
L
Linus Torvalds 已提交
184 185 186
static void __ipv6_ifa_notify(int event, struct inet6_ifaddr *ifa);
static void ipv6_ifa_notify(int event, struct inet6_ifaddr *ifa);

187
static void inet6_prefix_notify(int event, struct inet6_dev *idev,
L
Linus Torvalds 已提交
188 189
				struct prefix_info *pinfo);

190
static struct ipv6_devconf ipv6_devconf __read_mostly = {
L
Linus Torvalds 已提交
191 192 193 194 195 196 197
	.forwarding		= 0,
	.hop_limit		= IPV6_DEFAULT_HOPLIMIT,
	.mtu6			= IPV6_MIN_MTU,
	.accept_ra		= 1,
	.accept_redirects	= 1,
	.autoconf		= 1,
	.force_mld_version	= 0,
198 199
	.mldv1_unsolicited_report_interval = 10 * HZ,
	.mldv2_unsolicited_report_interval = HZ,
L
Linus Torvalds 已提交
200 201 202
	.dad_transmits		= 1,
	.rtr_solicits		= MAX_RTR_SOLICITATIONS,
	.rtr_solicit_interval	= RTR_SOLICITATION_INTERVAL,
203
	.rtr_solicit_max_interval = RTR_SOLICITATION_MAX_INTERVAL,
L
Linus Torvalds 已提交
204
	.rtr_solicit_delay	= MAX_RTR_SOLICITATION_DELAY,
205
	.use_tempaddr		= 0,
L
Linus Torvalds 已提交
206 207 208 209 210
	.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,
211
	.accept_ra_defrtr	= 1,
212
	.accept_ra_from_local	= 0,
213
	.accept_ra_min_hop_limit= 1,
214
	.accept_ra_pinfo	= 1,
215 216
#ifdef CONFIG_IPV6_ROUTER_PREF
	.accept_ra_rtr_pref	= 1,
217
	.rtr_probe_interval	= 60 * HZ,
218
#ifdef CONFIG_IPV6_ROUTE_INFO
219
	.accept_ra_rt_info_min_plen = 0,
220 221
	.accept_ra_rt_info_max_plen = 0,
#endif
222
#endif
223
	.proxy_ndp		= 0,
224
	.accept_source_route	= 0,	/* we do not accept RH0 by default. */
225
	.disable_ipv6		= 0,
226
	.accept_dad		= 0,
227
	.suppress_frag_ndisc	= 1,
228
	.accept_ra_mtu		= 1,
229 230
	.stable_secret		= {
		.initialized = false,
231 232
	},
	.use_oif_addrs_only	= 0,
233
	.ignore_routes_with_linkdown = 0,
234
	.keep_addr_on_down	= 0,
235
	.seg6_enabled		= 0,
236 237 238
#ifdef CONFIG_IPV6_SEG6_HMAC
	.seg6_require_hmac	= 0,
#endif
239
	.enhanced_dad           = 1,
240
	.addr_gen_mode		= IN6_ADDR_GEN_MODE_EUI64,
241
	.disable_policy		= 0,
L
Linus Torvalds 已提交
242 243
};

244
static struct ipv6_devconf ipv6_devconf_dflt __read_mostly = {
L
Linus Torvalds 已提交
245 246 247 248 249 250
	.forwarding		= 0,
	.hop_limit		= IPV6_DEFAULT_HOPLIMIT,
	.mtu6			= IPV6_MIN_MTU,
	.accept_ra		= 1,
	.accept_redirects	= 1,
	.autoconf		= 1,
251 252 253
	.force_mld_version	= 0,
	.mldv1_unsolicited_report_interval = 10 * HZ,
	.mldv2_unsolicited_report_interval = HZ,
L
Linus Torvalds 已提交
254 255 256
	.dad_transmits		= 1,
	.rtr_solicits		= MAX_RTR_SOLICITATIONS,
	.rtr_solicit_interval	= RTR_SOLICITATION_INTERVAL,
257
	.rtr_solicit_max_interval = RTR_SOLICITATION_MAX_INTERVAL,
L
Linus Torvalds 已提交
258 259 260 261 262 263 264
	.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,
265
	.accept_ra_defrtr	= 1,
266
	.accept_ra_from_local	= 0,
267
	.accept_ra_min_hop_limit= 1,
268
	.accept_ra_pinfo	= 1,
269 270
#ifdef CONFIG_IPV6_ROUTER_PREF
	.accept_ra_rtr_pref	= 1,
271
	.rtr_probe_interval	= 60 * HZ,
272
#ifdef CONFIG_IPV6_ROUTE_INFO
273
	.accept_ra_rt_info_min_plen = 0,
274 275
	.accept_ra_rt_info_max_plen = 0,
#endif
276
#endif
277
	.proxy_ndp		= 0,
278
	.accept_source_route	= 0,	/* we do not accept RH0 by default. */
279
	.disable_ipv6		= 0,
280
	.accept_dad		= 1,
281
	.suppress_frag_ndisc	= 1,
282
	.accept_ra_mtu		= 1,
283 284 285
	.stable_secret		= {
		.initialized = false,
	},
286
	.use_oif_addrs_only	= 0,
287
	.ignore_routes_with_linkdown = 0,
288
	.keep_addr_on_down	= 0,
289
	.seg6_enabled		= 0,
290 291 292
#ifdef CONFIG_IPV6_SEG6_HMAC
	.seg6_require_hmac	= 0,
#endif
293
	.enhanced_dad           = 1,
294
	.addr_gen_mode		= IN6_ADDR_GEN_MODE_EUI64,
295
	.disable_policy		= 0,
L
Linus Torvalds 已提交
296 297
};

298 299
/* Check if link is ready: is it up and is a valid qdisc available */
static inline bool addrconf_link_ready(const struct net_device *dev)
300
{
301
	return netif_oper_up(dev) && !qdisc_tx_is_noop(dev);
302 303
}

304
static void addrconf_del_rs_timer(struct inet6_dev *idev)
L
Linus Torvalds 已提交
305
{
306 307 308 309
	if (del_timer(&idev->rs_timer))
		__in6_dev_put(idev);
}

310
static void addrconf_del_dad_work(struct inet6_ifaddr *ifp)
311
{
312
	if (cancel_delayed_work(&ifp->dad_work))
L
Linus Torvalds 已提交
313 314 315
		__in6_ifa_put(ifp);
}

316 317 318 319 320 321 322
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
323

324 325
static void addrconf_mod_dad_work(struct inet6_ifaddr *ifp,
				   unsigned long delay)
L
Linus Torvalds 已提交
326
{
327 328 329
	in6_ifa_hold(ifp);
	if (mod_delayed_work(addrconf_wq, &ifp->dad_work, delay))
		in6_ifa_put(ifp);
L
Linus Torvalds 已提交
330 331
}

332 333
static int snmp6_alloc_dev(struct inet6_dev *idev)
{
334 335
	int i;

W
WANG Cong 已提交
336 337
	idev->stats.ipv6 = alloc_percpu(struct ipstats_mib);
	if (!idev->stats.ipv6)
338
		goto err_ip;
339 340 341

	for_each_possible_cpu(i) {
		struct ipstats_mib *addrconf_stats;
W
WANG Cong 已提交
342
		addrconf_stats = per_cpu_ptr(idev->stats.ipv6, i);
343 344 345 346
		u64_stats_init(&addrconf_stats->syncp);
	}


347 348 349
	idev->stats.icmpv6dev = kzalloc(sizeof(struct icmpv6_mib_device),
					GFP_KERNEL);
	if (!idev->stats.icmpv6dev)
350
		goto err_icmp;
351 352 353
	idev->stats.icmpv6msgdev = kzalloc(sizeof(struct icmpv6msg_mib_device),
					   GFP_KERNEL);
	if (!idev->stats.icmpv6msgdev)
354
		goto err_icmpmsg;
355 356 357

	return 0;

358
err_icmpmsg:
359
	kfree(idev->stats.icmpv6dev);
360
err_icmp:
W
WANG Cong 已提交
361
	free_percpu(idev->stats.ipv6);
362
err_ip:
363
	return -ENOMEM;
364 365
}

366
static struct inet6_dev *ipv6_add_dev(struct net_device *dev)
L
Linus Torvalds 已提交
367 368
{
	struct inet6_dev *ndev;
369
	int err = -ENOMEM;
L
Linus Torvalds 已提交
370 371 372 373

	ASSERT_RTNL();

	if (dev->mtu < IPV6_MIN_MTU)
374
		return ERR_PTR(-EINVAL);
L
Linus Torvalds 已提交
375

376
	ndev = kzalloc(sizeof(struct inet6_dev), GFP_KERNEL);
377
	if (!ndev)
378
		return ERR_PTR(err);
379 380 381

	rwlock_init(&ndev->lock);
	ndev->dev = dev;
382
	INIT_LIST_HEAD(&ndev->addr_list);
383
	timer_setup(&ndev->rs_timer, addrconf_rs_timer, 0);
384
	memcpy(&ndev->cnf, dev_net(dev)->ipv6.devconf_dflt, sizeof(ndev->cnf));
385 386

	if (ndev->cnf.stable_secret.initialized)
387
		ndev->cnf.addr_gen_mode = IN6_ADDR_GEN_MODE_STABLE_PRIVACY;
388
	else
389
		ndev->cnf.addr_gen_mode = ipv6_devconf_dflt.addr_gen_mode;
390

391 392
	ndev->cnf.mtu6 = dev->mtu;
	ndev->nd_parms = neigh_parms_alloc(dev, &nd_tbl);
393
	if (!ndev->nd_parms) {
394
		kfree(ndev);
395
		return ERR_PTR(err);
396
	}
397 398
	if (ndev->cnf.forwarding)
		dev_disable_lro(dev);
399 400
	/* We refer to the device */
	dev_hold(dev);
L
Linus Torvalds 已提交
401

402
	if (snmp6_alloc_dev(ndev) < 0) {
403 404
		netdev_dbg(dev, "%s: cannot allocate memory for statistics\n",
			   __func__);
405
		neigh_parms_release(&nd_tbl, ndev->nd_parms);
R
Roy Li 已提交
406 407
		dev_put(dev);
		kfree(ndev);
408
		return ERR_PTR(err);
409
	}
L
Linus Torvalds 已提交
410

411
	if (snmp6_register_dev(ndev) < 0) {
412 413
		netdev_dbg(dev, "%s: cannot create /proc/net/dev_snmp6/%s\n",
			   __func__, dev->name);
414
		goto err_release;
415 416
	}

417
	/* One reference from device. */
418
	refcount_set(&ndev->refcnt, 1);
L
Linus Torvalds 已提交
419

420 421 422
	if (dev->flags & (IFF_NOARP | IFF_LOOPBACK))
		ndev->cnf.accept_dad = -1;

A
Amerigo Wang 已提交
423
#if IS_ENABLED(CONFIG_IPV6_SIT)
424
	if (dev->type == ARPHRD_SIT && (dev->priv_flags & IFF_ISATAP)) {
425
		pr_info("%s: Disabled Multicast RS\n", dev->name);
426 427 428 429
		ndev->cnf.rtr_solicits = 0;
	}
#endif

430
	INIT_LIST_HEAD(&ndev->tempaddr_list);
J
Jiri Bohac 已提交
431
	ndev->desync_factor = U32_MAX;
432 433
	if ((dev->flags&IFF_LOOPBACK) ||
	    dev->type == ARPHRD_TUNNEL ||
434
	    dev->type == ARPHRD_TUNNEL6 ||
435 436
	    dev->type == ARPHRD_SIT ||
	    dev->type == ARPHRD_NONE) {
437
		ndev->cnf.use_tempaddr = -1;
438 439
	} else
		ipv6_regen_rndid(ndev);
440

441
	ndev->token = in6addr_any;
L
Linus Torvalds 已提交
442

443
	if (netif_running(dev) && addrconf_link_ready(dev))
444 445
		ndev->if_flags |= IF_READY;

446 447
	ipv6_mc_init_dev(ndev);
	ndev->tstamp = jiffies;
448 449 450
	err = addrconf_sysctl_register(ndev);
	if (err) {
		ipv6_mc_destroy_dev(ndev);
451
		snmp6_unregister_dev(ndev);
452 453
		goto err_release;
	}
454
	/* protected by rtnl_lock */
455
	rcu_assign_pointer(dev->ip6_ptr, ndev);
456

457 458 459
	/* Join interface-local all-node multicast group */
	ipv6_dev_mc_inc(dev, &in6addr_interfacelocal_allnodes);

460
	/* Join all-node multicast group */
461
	ipv6_dev_mc_inc(dev, &in6addr_linklocal_allnodes);
462

463
	/* Join all-router multicast group if forwarding is set */
L
Li Wei 已提交
464
	if (ndev->cnf.forwarding && (dev->flags & IFF_MULTICAST))
465 466
		ipv6_dev_mc_inc(dev, &in6addr_linklocal_allrouters);

L
Linus Torvalds 已提交
467
	return ndev;
468 469 470 471 472 473

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
474 475
}

476
static struct inet6_dev *ipv6_find_idev(struct net_device *dev)
L
Linus Torvalds 已提交
477 478 479 480 481
{
	struct inet6_dev *idev;

	ASSERT_RTNL();

S
Stephen Hemminger 已提交
482 483 484
	idev = __in6_dev_get(dev);
	if (!idev) {
		idev = ipv6_add_dev(dev);
485
		if (IS_ERR(idev))
L
Linus Torvalds 已提交
486 487
			return NULL;
	}
488

L
Linus Torvalds 已提交
489 490 491 492 493
	if (dev->flags&IFF_UP)
		ipv6_mc_up(idev);
	return idev;
}

494 495 496 497
static int inet6_netconf_msgsize_devconf(int type)
{
	int size =  NLMSG_ALIGN(sizeof(struct netconfmsg))
		    + nla_total_size(4);	/* NETCONFA_IFINDEX */
498
	bool all = false;
499

500 501 502 503
	if (type == NETCONFA_ALL)
		all = true;

	if (all || type == NETCONFA_FORWARDING)
504
		size += nla_total_size(4);
505
#ifdef CONFIG_IPV6_MROUTE
506
	if (all || type == NETCONFA_MC_FORWARDING)
507
		size += nla_total_size(4);
508
#endif
509
	if (all || type == NETCONFA_PROXY_NEIGH)
510
		size += nla_total_size(4);
511

512
	if (all || type == NETCONFA_IGNORE_ROUTES_WITH_LINKDOWN)
513 514
		size += nla_total_size(4);

515 516 517 518 519 520 521 522 523 524
	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;
525
	bool all = false;
526 527 528

	nlh = nlmsg_put(skb, portid, seq, event, sizeof(struct netconfmsg),
			flags);
529
	if (!nlh)
530 531
		return -EMSGSIZE;

532 533 534
	if (type == NETCONFA_ALL)
		all = true;

535 536 537 538 539 540
	ncm = nlmsg_data(nlh);
	ncm->ncm_family = AF_INET6;

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

541 542 543
	if (!devconf)
		goto out;

544
	if ((all || type == NETCONFA_FORWARDING) &&
545 546
	    nla_put_s32(skb, NETCONFA_FORWARDING, devconf->forwarding) < 0)
		goto nla_put_failure;
547
#ifdef CONFIG_IPV6_MROUTE
548
	if ((all || type == NETCONFA_MC_FORWARDING) &&
549 550 551
	    nla_put_s32(skb, NETCONFA_MC_FORWARDING,
			devconf->mc_forwarding) < 0)
		goto nla_put_failure;
552
#endif
553
	if ((all || type == NETCONFA_PROXY_NEIGH) &&
554 555 556
	    nla_put_s32(skb, NETCONFA_PROXY_NEIGH, devconf->proxy_ndp) < 0)
		goto nla_put_failure;

557
	if ((all || type == NETCONFA_IGNORE_ROUTES_WITH_LINKDOWN) &&
558 559 560 561
	    nla_put_s32(skb, NETCONFA_IGNORE_ROUTES_WITH_LINKDOWN,
			devconf->ignore_routes_with_linkdown) < 0)
		goto nla_put_failure;

562
out:
563 564
	nlmsg_end(skb, nlh);
	return 0;
565 566 567 568 569 570

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

571 572
void inet6_netconf_notify_devconf(struct net *net, int event, int type,
				  int ifindex, struct ipv6_devconf *devconf)
573 574 575 576
{
	struct sk_buff *skb;
	int err = -ENOBUFS;

577
	skb = nlmsg_new(inet6_netconf_msgsize_devconf(type), GFP_KERNEL);
578
	if (!skb)
579 580 581
		goto errout;

	err = inet6_netconf_fill_devconf(skb, ifindex, devconf, 0, 0,
582
					 event, 0, type);
583 584 585 586 587 588
	if (err < 0) {
		/* -EMSGSIZE implies BUG in inet6_netconf_msgsize_devconf() */
		WARN_ON(err == -EMSGSIZE);
		kfree_skb(skb);
		goto errout;
	}
589
	rtnl_notify(skb, net, 0, RTNLGRP_IPV6_NETCONF, NULL, GFP_KERNEL);
590 591
	return;
errout:
592
	rtnl_set_sk_err(net, RTNLGRP_IPV6_NETCONF, err);
593 594
}

595 596 597
static const struct nla_policy devconf_ipv6_policy[NETCONFA_MAX+1] = {
	[NETCONFA_IFINDEX]	= { .len = sizeof(int) },
	[NETCONFA_FORWARDING]	= { .len = sizeof(int) },
598
	[NETCONFA_PROXY_NEIGH]	= { .len = sizeof(int) },
599
	[NETCONFA_IGNORE_ROUTES_WITH_LINKDOWN]	= { .len = sizeof(int) },
600 601 602
};

static int inet6_netconf_get_devconf(struct sk_buff *in_skb,
603 604
				     struct nlmsghdr *nlh,
				     struct netlink_ext_ack *extack)
605 606 607
{
	struct net *net = sock_net(in_skb->sk);
	struct nlattr *tb[NETCONFA_MAX+1];
608 609
	struct inet6_dev *in6_dev = NULL;
	struct net_device *dev = NULL;
610 611 612 613 614 615 616
	struct netconfmsg *ncm;
	struct sk_buff *skb;
	struct ipv6_devconf *devconf;
	int ifindex;
	int err;

	err = nlmsg_parse(nlh, sizeof(*ncm), tb, NETCONFA_MAX,
617
			  devconf_ipv6_policy, extack);
618
	if (err < 0)
619
		return err;
620 621

	if (!tb[NETCONFA_IFINDEX])
622
		return -EINVAL;
623

624
	err = -EINVAL;
625 626 627 628 629 630 631 632 633
	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:
634
		dev = dev_get_by_index(net, ifindex);
635
		if (!dev)
636 637
			return -EINVAL;
		in6_dev = in6_dev_get(dev);
638
		if (!in6_dev)
639 640 641 642 643 644
			goto errout;
		devconf = &in6_dev->cnf;
		break;
	}

	err = -ENOBUFS;
645
	skb = nlmsg_new(inet6_netconf_msgsize_devconf(NETCONFA_ALL), GFP_KERNEL);
646
	if (!skb)
647 648 649 650 651
		goto errout;

	err = inet6_netconf_fill_devconf(skb, ifindex, devconf,
					 NETLINK_CB(in_skb).portid,
					 nlh->nlmsg_seq, RTM_NEWNETCONF, 0,
652
					 NETCONFA_ALL);
653 654 655 656 657 658 659 660
	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:
661 662 663 664
	if (in6_dev)
		in6_dev_put(in6_dev);
	if (dev)
		dev_put(dev);
665 666 667
	return err;
}

668 669 670 671 672 673 674 675 676 677 678 679 680 681 682 683 684
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();
685 686
		cb->seq = atomic_read(&net->ipv6.dev_addr_genid) ^
			  net->dev_base_seq;
687 688 689 690 691 692 693 694 695 696 697 698 699
		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,
700
						       NETCONFA_ALL) < 0) {
701 702 703
				rcu_read_unlock();
				goto done;
			}
704
			nl_dump_check_consistent(cb, nlmsg_hdr(skb));
705 706 707 708 709 710 711 712 713 714 715
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,
716
					       NETCONFA_ALL) < 0)
717 718 719 720 721 722 723 724 725 726
			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,
727
					       NETCONFA_ALL) < 0)
728 729 730 731 732 733 734 735 736 737 738
			goto done;
		else
			h++;
	}
done:
	cb->args[0] = h;
	cb->args[1] = idx;

	return skb->len;
}

L
Linus Torvalds 已提交
739 740 741 742 743 744 745 746 747
#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;
748 749
	if (idev->cnf.forwarding)
		dev_disable_lro(dev);
A
Amerigo Wang 已提交
750
	if (dev->flags & IFF_MULTICAST) {
751
		if (idev->cnf.forwarding) {
752
			ipv6_dev_mc_inc(dev, &in6addr_linklocal_allrouters);
753 754 755
			ipv6_dev_mc_inc(dev, &in6addr_interfacelocal_allrouters);
			ipv6_dev_mc_inc(dev, &in6addr_sitelocal_allrouters);
		} else {
756
			ipv6_dev_mc_dec(dev, &in6addr_linklocal_allrouters);
757 758 759
			ipv6_dev_mc_dec(dev, &in6addr_interfacelocal_allrouters);
			ipv6_dev_mc_dec(dev, &in6addr_sitelocal_allrouters);
		}
L
Linus Torvalds 已提交
760
	}
761 762

	list_for_each_entry(ifa, &idev->addr_list, if_list) {
M
Michal Wrobel 已提交
763 764
		if (ifa->flags&IFA_F_TENTATIVE)
			continue;
L
Linus Torvalds 已提交
765 766 767 768 769
		if (idev->cnf.forwarding)
			addrconf_join_anycast(ifa);
		else
			addrconf_leave_anycast(ifa);
	}
770 771
	inet6_netconf_notify_devconf(dev_net(dev), RTM_NEWNETCONF,
				     NETCONFA_FORWARDING,
772
				     dev->ifindex, &idev->cnf);
L
Linus Torvalds 已提交
773 774 775
}


776
static void addrconf_forward_change(struct net *net, __s32 newf)
L
Linus Torvalds 已提交
777 778 779 780
{
	struct net_device *dev;
	struct inet6_dev *idev;

781
	for_each_netdev(net, dev) {
L
Linus Torvalds 已提交
782 783
		idev = __in6_dev_get(dev);
		if (idev) {
784 785
			int changed = (!idev->cnf.forwarding) ^ (!newf);
			idev->cnf.forwarding = newf;
L
Linus Torvalds 已提交
786 787 788 789 790
			if (changed)
				dev_forward_change(idev);
		}
	}
}
791

792
static int addrconf_fixup_forwarding(struct ctl_table *table, int *p, int newf)
793
{
794
	struct net *net;
795 796 797 798
	int old;

	if (!rtnl_trylock())
		return restart_syscall();
799 800

	net = (struct net *)table->extra2;
801 802
	old = *p;
	*p = newf;
803

804
	if (p == &net->ipv6.devconf_dflt->forwarding) {
805
		if ((!newf) ^ (!old))
806 807
			inet6_netconf_notify_devconf(net, RTM_NEWNETCONF,
						     NETCONFA_FORWARDING,
808 809
						     NETCONFA_IFINDEX_DEFAULT,
						     net->ipv6.devconf_dflt);
810 811
		rtnl_unlock();
		return 0;
E
Eric W. Biederman 已提交
812
	}
813

814
	if (p == &net->ipv6.devconf_all->forwarding) {
815 816
		int old_dflt = net->ipv6.devconf_dflt->forwarding;

817
		net->ipv6.devconf_dflt->forwarding = newf;
818
		if ((!newf) ^ (!old_dflt))
819 820
			inet6_netconf_notify_devconf(net, RTM_NEWNETCONF,
						     NETCONFA_FORWARDING,
821 822 823
						     NETCONFA_IFINDEX_DEFAULT,
						     net->ipv6.devconf_dflt);

824
		addrconf_forward_change(net, newf);
825
		if ((!newf) ^ (!old))
826 827
			inet6_netconf_notify_devconf(net, RTM_NEWNETCONF,
						     NETCONFA_FORWARDING,
828 829
						     NETCONFA_IFINDEX_ALL,
						     net->ipv6.devconf_all);
830
	} else if ((!newf) ^ (!old))
831
		dev_forward_change((struct inet6_dev *)table->extra1);
832
	rtnl_unlock();
833

834
	if (newf)
835
		rt6_purge_dflt_routers(net);
836
	return 1;
837
}
838 839 840 841 842 843 844 845 846 847 848 849 850 851

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),
852
							     RTM_NEWNETCONF,
853 854 855 856 857 858 859 860 861 862 863 864 865 866 867 868 869 870 871 872 873 874
							     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,
875
						     RTM_NEWNETCONF,
876 877 878 879 880 881 882 883 884 885 886 887
						     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,
888
						     RTM_NEWNETCONF,
889 890 891 892 893 894 895 896 897
						     NETCONFA_IGNORE_ROUTES_WITH_LINKDOWN,
						     NETCONFA_IFINDEX_ALL,
						     net->ipv6.devconf_all);
	}
	rtnl_unlock();

	return 1;
}

L
Linus Torvalds 已提交
898 899
#endif

900
/* Nobody refers to this ifaddr, destroy it */
L
Linus Torvalds 已提交
901 902
void inet6_ifa_finish_destroy(struct inet6_ifaddr *ifp)
{
903
	WARN_ON(!hlist_unhashed(&ifp->addr_lst));
904

L
Linus Torvalds 已提交
905
#ifdef NET_REFCNT_DEBUG
906
	pr_debug("%s\n", __func__);
L
Linus Torvalds 已提交
907 908 909 910
#endif

	in6_dev_put(ifp->idev);

911 912 913
	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
914

915
	if (ifp->state != INET6_IFADDR_STATE_DEAD) {
916
		pr_warn("Freeing alive inet6 address %p\n", ifp);
L
Linus Torvalds 已提交
917 918 919
		return;
	}

920
	kfree_rcu(ifp, rcu);
L
Linus Torvalds 已提交
921 922
}

B
Brian Haley 已提交
923 924 925
static void
ipv6_link_dev_addr(struct inet6_dev *idev, struct inet6_ifaddr *ifp)
{
926
	struct list_head *p;
927
	int ifp_scope = ipv6_addr_src_scope(&ifp->addr);
B
Brian Haley 已提交
928 929 930 931 932

	/*
	 * Each device address list is sorted in order of scope -
	 * global before linklocal.
	 */
933 934 935
	list_for_each(p, &idev->addr_list) {
		struct inet6_ifaddr *ifa
			= list_entry(p, struct inet6_ifaddr, if_list);
936
		if (ifp_scope >= ipv6_addr_src_scope(&ifa->addr))
B
Brian Haley 已提交
937 938 939
			break;
	}

940
	list_add_tail_rcu(&ifp->if_list, p);
B
Brian Haley 已提交
941 942
}

943
static u32 inet6_addr_hash(const struct net *net, const struct in6_addr *addr)
944
{
945 946 947
	u32 val = ipv6_addr_hash(addr) ^ net_hash_mix(net);

	return hash_32(val, IN6_ADDR_HSIZE_SHIFT);
948 949
}

950
static bool ipv6_chk_same_addr(struct net *net, const struct in6_addr *addr,
951
			       struct net_device *dev, unsigned int hash)
952 953 954 955 956 957 958 959 960 961 962 963 964 965
{
	struct inet6_ifaddr *ifp;

	hlist_for_each_entry(ifp, &inet6_addr_lst[hash], addr_lst) {
		if (!net_eq(dev_net(ifp->idev->dev), net))
			continue;
		if (ipv6_addr_equal(&ifp->addr, addr)) {
			if (!dev || ifp->idev->dev == dev)
				return true;
		}
	}
	return false;
}

966 967
static int ipv6_add_addr_hash(struct net_device *dev, struct inet6_ifaddr *ifa)
{
968
	unsigned int hash = inet6_addr_hash(dev_net(dev), &ifa->addr);
969 970 971 972 973
	int err = 0;

	spin_lock(&addrconf_hash_lock);

	/* Ignore adding duplicate addresses on an interface */
974
	if (ipv6_chk_same_addr(dev_net(dev), &ifa->addr, dev, hash)) {
975
		netdev_dbg(dev, "ipv6_add_addr: already assigned\n");
976
		err = -EEXIST;
977 978
	} else {
		hlist_add_head_rcu(&ifa->addr_lst, &inet6_addr_lst[hash]);
979 980 981 982 983 984 985
	}

	spin_unlock(&addrconf_hash_lock);

	return err;
}

L
Linus Torvalds 已提交
986 987 988
/* On success it returns ifp with increased reference count */

static struct inet6_ifaddr *
989
ipv6_add_addr(struct inet6_dev *idev, struct ifa6_config *cfg,
990
	      bool can_block, struct netlink_ext_ack *extack)
L
Linus Torvalds 已提交
991
{
992
	gfp_t gfp_flags = can_block ? GFP_KERNEL : GFP_ATOMIC;
993
	int addr_type = ipv6_addr_type(cfg->pfx);
994
	struct net *net = dev_net(idev->dev);
L
Linus Torvalds 已提交
995
	struct inet6_ifaddr *ifa = NULL;
996
	struct fib6_info *f6i = NULL;
L
Linus Torvalds 已提交
997
	int err = 0;
998 999 1000 1001 1002 1003

	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
1004 1005 1006

	if (idev->dead) {
		err = -ENODEV;			/*XXX*/
1007
		goto out;
L
Linus Torvalds 已提交
1008 1009
	}

1010
	if (idev->cnf.disable_ipv6) {
1011
		err = -EACCES;
1012
		goto out;
1013 1014
	}

1015 1016 1017 1018 1019
	/* validator notifier needs to be blocking;
	 * do not call in atomic context
	 */
	if (can_block) {
		struct in6_validator_info i6vi = {
1020
			.i6vi_addr = *cfg->pfx,
1021
			.i6vi_dev = idev,
1022
			.extack = extack,
1023 1024 1025 1026 1027 1028 1029
		};

		err = inet6addr_validator_notifier_call_chain(NETDEV_UP, &i6vi);
		err = notifier_to_errno(err);
		if (err < 0)
			goto out;
	}
L
Linus Torvalds 已提交
1030

1031
	ifa = kzalloc(sizeof(*ifa), gfp_flags);
1032
	if (!ifa) {
L
Linus Torvalds 已提交
1033 1034 1035 1036
		err = -ENOBUFS;
		goto out;
	}

1037
	f6i = addrconf_f6i_alloc(net, idev, cfg->pfx, false, gfp_flags);
1038 1039 1040
	if (IS_ERR(f6i)) {
		err = PTR_ERR(f6i);
		f6i = NULL;
L
Linus Torvalds 已提交
1041 1042 1043
		goto out;
	}

1044 1045
	if (net->ipv6.devconf_all->disable_policy ||
	    idev->cnf.disable_policy)
1046
		f6i->dst_nopolicy = true;
1047

1048 1049
	neigh_parms_data_state_setall(idev->nd_parms);

1050 1051 1052
	ifa->addr = *cfg->pfx;
	if (cfg->peer_pfx)
		ifa->peer_addr = *cfg->peer_pfx;
L
Linus Torvalds 已提交
1053 1054

	spin_lock_init(&ifa->lock);
1055
	INIT_DELAYED_WORK(&ifa->dad_work, addrconf_dad_work);
1056
	INIT_HLIST_NODE(&ifa->addr_lst);
1057 1058
	ifa->scope = cfg->scope;
	ifa->prefix_len = cfg->plen;
1059
	ifa->rt_priority = cfg->rt_priority;
1060
	ifa->flags = cfg->ifa_flags;
1061
	/* No need to add the TENTATIVE flag for addresses with NODAD */
1062
	if (!(cfg->ifa_flags & IFA_F_NODAD))
1063
		ifa->flags |= IFA_F_TENTATIVE;
1064 1065
	ifa->valid_lft = cfg->valid_lft;
	ifa->prefered_lft = cfg->preferred_lft;
L
Linus Torvalds 已提交
1066
	ifa->cstamp = ifa->tstamp = jiffies;
1067
	ifa->tokenized = false;
L
Linus Torvalds 已提交
1068

1069
	ifa->rt = f6i;
1070

L
Linus Torvalds 已提交
1071
	ifa->idev = idev;
1072 1073
	in6_dev_hold(idev);

L
Linus Torvalds 已提交
1074
	/* For caller */
1075
	refcount_set(&ifa->refcnt, 1);
L
Linus Torvalds 已提交
1076

1077
	rcu_read_lock_bh();
L
Linus Torvalds 已提交
1078

1079 1080 1081 1082 1083
	err = ipv6_add_addr_hash(idev->dev, ifa);
	if (err < 0) {
		rcu_read_unlock_bh();
		goto out;
	}
L
Linus Torvalds 已提交
1084 1085

	write_lock(&idev->lock);
1086

L
Linus Torvalds 已提交
1087
	/* Add to inet6_dev unicast addr list. */
B
Brian Haley 已提交
1088
	ipv6_link_dev_addr(idev, ifa);
L
Linus Torvalds 已提交
1089 1090

	if (ifa->flags&IFA_F_TEMPORARY) {
1091
		list_add(&ifa->tmp_list, &idev->tempaddr_list);
L
Linus Torvalds 已提交
1092 1093 1094 1095 1096
		in6_ifa_hold(ifa);
	}

	in6_ifa_hold(ifa);
	write_unlock(&idev->lock);
1097

1098
	rcu_read_unlock_bh();
L
Linus Torvalds 已提交
1099

1100 1101 1102
	inet6addr_notifier_call_chain(NETDEV_UP, ifa);
out:
	if (unlikely(err < 0)) {
1103
		fib6_info_release(f6i);
1104

1105 1106 1107 1108 1109
		if (ifa) {
			if (ifa->idev)
				in6_dev_put(ifa->idev);
			kfree(ifa);
		}
L
Linus Torvalds 已提交
1110 1111 1112 1113 1114 1115
		ifa = ERR_PTR(err);
	}

	return ifa;
}

1116 1117 1118 1119 1120 1121 1122 1123 1124 1125 1126 1127 1128 1129 1130 1131 1132 1133 1134 1135 1136 1137 1138 1139 1140 1141 1142 1143 1144 1145 1146 1147 1148 1149 1150 1151 1152 1153 1154 1155 1156 1157 1158 1159 1160 1161 1162 1163 1164 1165 1166 1167 1168 1169 1170 1171 1172 1173 1174 1175 1176 1177 1178 1179
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)
{
1180
	struct fib6_info *f6i;
1181

1182
	f6i = addrconf_get_prefix_route(&ifp->addr,
1183 1184 1185
				       ifp->prefix_len,
				       ifp->idev->dev,
				       0, RTF_GATEWAY | RTF_DEFAULT);
1186
	if (f6i) {
1187
		if (del_rt)
1188
			ip6_del_rt(dev_net(ifp->idev->dev), f6i);
1189
		else {
1190 1191 1192
			if (!(f6i->fib6_flags & RTF_EXPIRES))
				fib6_set_expires(f6i, expires);
			fib6_info_release(f6i);
1193 1194 1195 1196 1197
		}
	}
}


L
Linus Torvalds 已提交
1198 1199 1200 1201
/* This function wants to get referenced ifp and releases it before return */

static void ipv6_del_addr(struct inet6_ifaddr *ifp)
{
1202
	int state;
1203 1204
	enum cleanup_prefix_rt_t action = CLEANUP_PREFIX_RT_NOP;
	unsigned long expires;
L
Linus Torvalds 已提交
1205

1206 1207
	ASSERT_RTNL();

1208
	spin_lock_bh(&ifp->lock);
1209
	state = ifp->state;
1210
	ifp->state = INET6_IFADDR_STATE_DEAD;
1211
	spin_unlock_bh(&ifp->lock);
1212 1213 1214

	if (state == INET6_IFADDR_STATE_DEAD)
		goto out;
L
Linus Torvalds 已提交
1215

1216 1217 1218
	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
1219

1220
	write_lock_bh(&ifp->idev->lock);
1221

L
Linus Torvalds 已提交
1222
	if (ifp->flags&IFA_F_TEMPORARY) {
1223 1224 1225 1226
		list_del(&ifp->tmp_list);
		if (ifp->ifpub) {
			in6_ifa_put(ifp->ifpub);
			ifp->ifpub = NULL;
L
Linus Torvalds 已提交
1227
		}
1228
		__in6_ifa_put(ifp);
L
Linus Torvalds 已提交
1229 1230
	}

1231 1232
	if (ifp->flags & IFA_F_PERMANENT && !(ifp->flags & IFA_F_NOPREFIXROUTE))
		action = check_cleanup_prefix_route(ifp, &expires);
1233

1234
	list_del_rcu(&ifp->if_list);
1235 1236 1237
	__in6_ifa_put(ifp);

	write_unlock_bh(&ifp->idev->lock);
L
Linus Torvalds 已提交
1238

1239
	addrconf_del_dad_work(ifp);
1240

L
Linus Torvalds 已提交
1241 1242
	ipv6_ifa_notify(RTM_DELADDR, ifp);

1243
	inet6addr_notifier_call_chain(NETDEV_DOWN, ifp);
L
Linus Torvalds 已提交
1244

1245 1246 1247
	if (action != CLEANUP_PREFIX_RT_NOP) {
		cleanup_prefix_route(ifp, expires,
			action == CLEANUP_PREFIX_RT_DEL);
L
Linus Torvalds 已提交
1248 1249
	}

1250 1251
	/* clean up prefsrc entries */
	rt6_remove_prefsrc(ifp);
1252
out:
L
Linus Torvalds 已提交
1253 1254 1255
	in6_ifa_put(ifp);
}

E
Eric Dumazet 已提交
1256 1257 1258
static int ipv6_create_tempaddr(struct inet6_ifaddr *ifp,
				struct inet6_ifaddr *ift,
				bool block)
L
Linus Torvalds 已提交
1259 1260 1261
{
	struct inet6_dev *idev = ifp->idev;
	struct in6_addr addr, *tmpaddr;
1262
	unsigned long tmp_tstamp, age;
1263
	unsigned long regen_advance;
1264
	struct ifa6_config cfg;
L
Linus Torvalds 已提交
1265
	int ret = 0;
1266
	unsigned long now = jiffies;
J
Jiri Bohac 已提交
1267
	long max_desync_factor;
1268
	s32 cnf_temp_preferred_lft;
L
Linus Torvalds 已提交
1269

1270
	write_lock_bh(&idev->lock);
L
Linus Torvalds 已提交
1271 1272 1273 1274 1275 1276 1277 1278 1279 1280 1281
	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) {
1282
		write_unlock_bh(&idev->lock);
1283
		pr_info("%s: use_tempaddr is disabled\n", __func__);
L
Linus Torvalds 已提交
1284 1285 1286 1287 1288 1289 1290 1291
		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);
1292
		write_unlock_bh(&idev->lock);
1293 1294
		pr_warn("%s: regeneration time exceeded - disabled temporary address support\n",
			__func__);
L
Linus Torvalds 已提交
1295 1296 1297 1298 1299 1300
		in6_dev_put(idev);
		ret = -1;
		goto out;
	}
	in6_ifa_hold(ifp);
	memcpy(addr.s6_addr, ifp->addr.s6_addr, 8);
1301
	ipv6_try_regen_rndid(idev, tmpaddr);
L
Linus Torvalds 已提交
1302
	memcpy(&addr.s6_addr[8], idev->rndid, 8);
1303
	age = (now - ifp->tstamp) / HZ;
J
Jiri Bohac 已提交
1304 1305 1306 1307 1308 1309 1310 1311

	regen_advance = idev->cnf.regen_max_retry *
			idev->cnf.dad_transmits *
			NEIGH_VAR(idev->nd_parms, RETRANS_TIME) / HZ;

	/* recalculate max_desync_factor each time and update
	 * idev->desync_factor if it's larger
	 */
1312
	cnf_temp_preferred_lft = READ_ONCE(idev->cnf.temp_prefered_lft);
J
Jiri Bohac 已提交
1313 1314
	max_desync_factor = min_t(__u32,
				  idev->cnf.max_desync_factor,
1315
				  cnf_temp_preferred_lft - regen_advance);
J
Jiri Bohac 已提交
1316 1317 1318 1319 1320 1321 1322 1323 1324 1325 1326

	if (unlikely(idev->desync_factor > max_desync_factor)) {
		if (max_desync_factor > 0) {
			get_random_bytes(&idev->desync_factor,
					 sizeof(idev->desync_factor));
			idev->desync_factor %= max_desync_factor;
		} else {
			idev->desync_factor = 0;
		}
	}

1327
	memset(&cfg, 0, sizeof(cfg));
1328
	cfg.valid_lft = min_t(__u32, ifp->valid_lft,
1329
			      idev->cnf.temp_valid_lft + age);
1330 1331 1332 1333
	cfg.preferred_lft = cnf_temp_preferred_lft + age - idev->desync_factor;
	cfg.preferred_lft = min_t(__u32, ifp->prefered_lft, cfg.preferred_lft);

	cfg.plen = ifp->prefix_len;
L
Linus Torvalds 已提交
1334 1335 1336
	tmp_tstamp = ifp->tstamp;
	spin_unlock_bh(&ifp->lock);

1337
	write_unlock_bh(&idev->lock);
1338

1339 1340 1341 1342
	/* 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.
1343 1344
	 * Use age calculation as in addrconf_verify to avoid unnecessary
	 * temporary addresses being generated.
1345
	 */
1346
	age = (now - tmp_tstamp + ADDRCONF_TIMER_FUZZ_MINUS) / HZ;
1347
	if (cfg.preferred_lft <= regen_advance + age) {
1348 1349 1350 1351 1352 1353
		in6_ifa_put(ifp);
		in6_dev_put(idev);
		ret = -1;
		goto out;
	}

1354
	cfg.ifa_flags = IFA_F_TEMPORARY;
1355 1356
	/* set in addrconf_prefix_rcv() */
	if (ifp->flags & IFA_F_OPTIMISTIC)
1357
		cfg.ifa_flags |= IFA_F_OPTIMISTIC;
1358

1359 1360
	cfg.pfx = &addr;
	cfg.scope = ipv6_addr_scope(cfg.pfx);
1361

1362
	ift = ipv6_add_addr(idev, &cfg, block, NULL);
1363
	if (IS_ERR(ift)) {
L
Linus Torvalds 已提交
1364 1365
		in6_ifa_put(ifp);
		in6_dev_put(idev);
1366
		pr_info("%s: retry temporary address regeneration\n", __func__);
L
Linus Torvalds 已提交
1367
		tmpaddr = &addr;
1368
		write_lock_bh(&idev->lock);
L
Linus Torvalds 已提交
1369 1370 1371 1372 1373
		goto retry;
	}

	spin_lock_bh(&ift->lock);
	ift->ifpub = ifp;
1374
	ift->cstamp = now;
L
Linus Torvalds 已提交
1375 1376 1377
	ift->tstamp = tmp_tstamp;
	spin_unlock_bh(&ift->lock);

1378
	addrconf_dad_start(ift);
L
Linus Torvalds 已提交
1379 1380 1381 1382 1383 1384 1385
	in6_ifa_put(ift);
	in6_dev_put(idev);
out:
	return ret;
}

/*
1386
 *	Choose an appropriate source address (RFC3484)
L
Linus Torvalds 已提交
1387
 */
1388 1389 1390 1391 1392 1393 1394 1395 1396 1397 1398 1399 1400
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,
1401 1402 1403
#ifdef CONFIG_IPV6_OPTIMISTIC_DAD
	IPV6_SADDR_RULE_NOT_OPTIMISTIC,
#endif
1404 1405 1406
	IPV6_SADDR_RULE_MAX
};

1407
struct ipv6_saddr_score {
1408 1409 1410 1411 1412 1413
	int			rule;
	int			addr_type;
	struct inet6_ifaddr	*ifa;
	DECLARE_BITMAP(scorebits, IPV6_SADDR_RULE_MAX);
	int			scopedist;
	int			matchlen;
1414 1415
};

1416
struct ipv6_saddr_dst {
1417
	const struct in6_addr *addr;
1418 1419 1420
	int ifindex;
	int scope;
	int label;
1421
	unsigned int prefs;
1422
};
1423

D
Dave Jones 已提交
1424
static inline int ipv6_saddr_preferred(int type)
L
Linus Torvalds 已提交
1425
{
U
Ulrich Weber 已提交
1426
	if (type & (IPV6_ADDR_MAPPED|IPV6_ADDR_COMPATv4|IPV6_ADDR_LOOPBACK))
1427 1428
		return 1;
	return 0;
L
Linus Torvalds 已提交
1429 1430
}

1431 1432
static bool ipv6_use_optimistic_addr(struct net *net,
				     struct inet6_dev *idev)
1433 1434
{
#ifdef CONFIG_IPV6_OPTIMISTIC_DAD
1435 1436 1437 1438 1439 1440 1441 1442
	if (!idev)
		return false;
	if (!net->ipv6.devconf_all->optimistic_dad && !idev->cnf.optimistic_dad)
		return false;
	if (!net->ipv6.devconf_all->use_optimistic && !idev->cnf.use_optimistic)
		return false;

	return true;
1443 1444 1445 1446 1447
#else
	return false;
#endif
}

1448 1449 1450 1451 1452 1453 1454 1455 1456 1457 1458 1459 1460 1461 1462
static bool ipv6_allow_optimistic_dad(struct net *net,
				      struct inet6_dev *idev)
{
#ifdef CONFIG_IPV6_OPTIMISTIC_DAD
	if (!idev)
		return false;
	if (!net->ipv6.devconf_all->optimistic_dad && !idev->cnf.optimistic_dad)
		return false;

	return true;
#else
	return false;
#endif
}

1463 1464
static int ipv6_get_saddr_eval(struct net *net,
			       struct ipv6_saddr_score *score,
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 1498 1499 1500 1501 1502 1503
			       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
1504
		 *  B-15 |    \        if scope is enough for destination.
1505 1506 1507 1508 1509 1510 1511 1512 1513 1514 1515 1516 1517 1518 1519 1520 1521 1522
		 *       |             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:
1523
	    {
1524
		/* Rule 3: Avoid deprecated and optimistic addresses */
1525 1526
		u8 avoid = IFA_F_DEPRECATED;

1527
		if (!ipv6_use_optimistic_addr(net, score->ifa->idev))
1528
			avoid |= IFA_F_OPTIMISTIC;
1529
		ret = ipv6_saddr_preferred(score->addr_type) ||
1530
		      !(score->ifa->flags & avoid);
1531
		break;
1532
	    }
1533 1534
#ifdef CONFIG_IPV6_MIP6
	case IPV6_SADDR_RULE_HOA:
1535
	    {
1536
		/* Rule 4: Prefer home address */
1537 1538
		int prefhome = !(dst->prefs & IPV6_PREFER_SRC_COA);
		ret = !(score->ifa->flags & IFA_F_HOMEADDRESS) ^ prefhome;
1539
		break;
1540
	    }
1541 1542 1543 1544 1545 1546 1547 1548
#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 */
1549 1550
		ret = ipv6_addr_label(net,
				      &score->ifa->addr, score->addr_type,
1551 1552 1553
				      score->ifa->idev->dev->ifindex) == dst->label;
		break;
	case IPV6_SADDR_RULE_PRIVACY:
1554
	    {
1555
		/* Rule 7: Prefer public address
L
Lucas De Marchi 已提交
1556
		 * Note: prefer temporary address if use_tempaddr >= 2
1557
		 */
1558 1559 1560 1561
		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;
1562
		break;
1563
	    }
1564 1565 1566 1567 1568 1569 1570 1571 1572
	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 */
1573 1574 1575 1576
		ret = ipv6_addr_diff(&score->ifa->addr, dst->addr);
		if (ret > score->ifa->prefix_len)
			ret = score->ifa->prefix_len;
		score->matchlen = ret;
1577
		break;
1578 1579 1580 1581 1582 1583 1584 1585
#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
1586 1587 1588 1589 1590 1591 1592 1593 1594 1595 1596
	default:
		ret = 0;
	}

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

1597 1598 1599 1600 1601
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)
1602
{
1603
	struct ipv6_saddr_score *score = &scores[1 - hiscore_idx], *hiscore = &scores[hiscore_idx];
1604

1605
	list_for_each_entry_rcu(score->ifa, &idev->addr_list, if_list) {
1606 1607 1608 1609 1610 1611 1612 1613 1614 1615 1616 1617 1618 1619 1620 1621 1622 1623 1624 1625 1626 1627 1628 1629 1630 1631 1632 1633 1634 1635 1636 1637 1638 1639 1640 1641 1642 1643 1644 1645 1646 1647 1648 1649 1650 1651 1652 1653 1654 1655
		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) {
				swap(hiscore, score);
1656
				hiscore_idx = 1 - hiscore_idx;
1657 1658 1659 1660 1661 1662 1663 1664 1665

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

				break;
			}
		}
	}
out:
1666
	return hiscore_idx;
1667 1668
}

1669 1670 1671 1672 1673 1674 1675 1676 1677 1678 1679 1680 1681 1682 1683 1684 1685 1686 1687 1688 1689 1690
static int ipv6_get_saddr_master(struct net *net,
				 const struct net_device *dst_dev,
				 const struct net_device *master,
				 struct ipv6_saddr_dst *dst,
				 struct ipv6_saddr_score *scores,
				 int hiscore_idx)
{
	struct inet6_dev *idev;

	idev = __in6_dev_get(dst_dev);
	if (idev)
		hiscore_idx = __ipv6_dev_get_saddr(net, dst, idev,
						   scores, hiscore_idx);

	idev = __in6_dev_get(master);
	if (idev)
		hiscore_idx = __ipv6_dev_get_saddr(net, dst, idev,
						   scores, hiscore_idx);

	return hiscore_idx;
}

1691
int ipv6_dev_get_saddr(struct net *net, const struct net_device *dst_dev,
1692
		       const struct in6_addr *daddr, unsigned int prefs,
1693
		       struct in6_addr *saddr)
L
Linus Torvalds 已提交
1694
{
1695
	struct ipv6_saddr_score scores[2], *hiscore;
1696
	struct ipv6_saddr_dst dst;
1697
	struct inet6_dev *idev;
1698
	struct net_device *dev;
1699
	int dst_type;
1700
	bool use_oif_addr = false;
1701
	int hiscore_idx = 0;
1702
	int ret = 0;
L
Linus Torvalds 已提交
1703

1704 1705 1706 1707
	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);
1708
	dst.label = ipv6_addr_label(net, daddr, dst_type, dst.ifindex);
1709
	dst.prefs = prefs;
1710

1711 1712
	scores[hiscore_idx].rule = -1;
	scores[hiscore_idx].ifa = NULL;
L
Linus Torvalds 已提交
1713

1714
	rcu_read_lock();
L
Linus Torvalds 已提交
1715

1716 1717 1718 1719 1720 1721 1722 1723 1724 1725 1726
	/* 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.)
1727 1728 1729 1730
	 *  - "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)
1731 1732
	 */
	if (dst_dev) {
1733
		idev = __in6_dev_get(dst_dev);
1734
		if ((dst_type & IPV6_ADDR_MULTICAST) ||
1735 1736
		    dst.scope <= IPV6_ADDR_SCOPE_LINKLOCAL ||
		    (idev && idev->cnf.use_oif_addrs_only)) {
1737 1738 1739
			use_oif_addr = true;
		}
	}
1740

1741
	if (use_oif_addr) {
1742
		if (idev)
1743
			hiscore_idx = __ipv6_dev_get_saddr(net, &dst, idev, scores, hiscore_idx);
1744
	} else {
1745 1746 1747 1748 1749 1750 1751 1752 1753 1754 1755 1756 1757 1758 1759 1760 1761 1762 1763
		const struct net_device *master;
		int master_idx = 0;

		/* if dst_dev exists and is enslaved to an L3 device, then
		 * prefer addresses from dst_dev and then the master over
		 * any other enslaved devices in the L3 domain.
		 */
		master = l3mdev_master_dev_rcu(dst_dev);
		if (master) {
			master_idx = master->ifindex;

			hiscore_idx = ipv6_get_saddr_master(net, dst_dev,
							    master, &dst,
							    scores, hiscore_idx);

			if (scores[hiscore_idx].ifa)
				goto out;
		}

1764
		for_each_netdev_rcu(net, dev) {
1765 1766 1767 1768 1769
			/* only consider addresses on devices in the
			 * same L3 domain
			 */
			if (l3mdev_master_ifindex_rcu(dev) != master_idx)
				continue;
1770 1771
			idev = __in6_dev_get(dev);
			if (!idev)
1772
				continue;
1773
			hiscore_idx = __ipv6_dev_get_saddr(net, &dst, idev, scores, hiscore_idx);
L
Linus Torvalds 已提交
1774 1775
		}
	}
1776 1777

out:
1778
	hiscore = &scores[hiscore_idx];
1779
	if (!hiscore->ifa)
1780 1781 1782
		ret = -EADDRNOTAVAIL;
	else
		*saddr = hiscore->ifa->addr;
1783

1784 1785
	rcu_read_unlock();
	return ret;
L
Linus Torvalds 已提交
1786
}
1787
EXPORT_SYMBOL(ipv6_dev_get_saddr);
L
Linus Torvalds 已提交
1788

1789
int __ipv6_get_lladdr(struct inet6_dev *idev, struct in6_addr *addr,
1790
		      u32 banned_flags)
1791 1792 1793 1794
{
	struct inet6_ifaddr *ifp;
	int err = -EADDRNOTAVAIL;

1795 1796 1797
	list_for_each_entry_reverse(ifp, &idev->addr_list, if_list) {
		if (ifp->scope > IFA_LINK)
			break;
1798 1799 1800 1801 1802 1803 1804 1805 1806 1807
		if (ifp->scope == IFA_LINK &&
		    !(ifp->flags & banned_flags)) {
			*addr = ifp->addr;
			err = 0;
			break;
		}
	}
	return err;
}

1808
int ipv6_get_lladdr(struct net_device *dev, struct in6_addr *addr,
1809
		    u32 banned_flags)
L
Linus Torvalds 已提交
1810 1811 1812 1813
{
	struct inet6_dev *idev;
	int err = -EADDRNOTAVAIL;

1814
	rcu_read_lock();
S
Stephen Hemminger 已提交
1815 1816
	idev = __in6_dev_get(dev);
	if (idev) {
L
Linus Torvalds 已提交
1817
		read_lock_bh(&idev->lock);
1818
		err = __ipv6_get_lladdr(idev, addr, banned_flags);
L
Linus Torvalds 已提交
1819 1820
		read_unlock_bh(&idev->lock);
	}
1821
	rcu_read_unlock();
L
Linus Torvalds 已提交
1822 1823 1824
	return err;
}

1825
static int ipv6_count_addresses(const struct inet6_dev *idev)
L
Linus Torvalds 已提交
1826
{
1827
	const struct inet6_ifaddr *ifp;
L
Linus Torvalds 已提交
1828 1829
	int cnt = 0;

1830 1831
	rcu_read_lock();
	list_for_each_entry_rcu(ifp, &idev->addr_list, if_list)
L
Linus Torvalds 已提交
1832
		cnt++;
1833
	rcu_read_unlock();
L
Linus Torvalds 已提交
1834 1835 1836
	return cnt;
}

1837
int ipv6_chk_addr(struct net *net, const struct in6_addr *addr,
1838
		  const struct net_device *dev, int strict)
1839
{
1840 1841
	return ipv6_chk_addr_and_flags(net, addr, dev, !dev,
				       strict, IFA_F_TENTATIVE);
1842 1843 1844
}
EXPORT_SYMBOL(ipv6_chk_addr);

1845 1846 1847 1848 1849 1850 1851 1852 1853
/* device argument is used to find the L3 domain of interest. If
 * skip_dev_check is set, then the ifp device is not checked against
 * the passed in dev argument. So the 2 cases for addresses checks are:
 *   1. does the address exist in the L3 domain that dev is part of
 *      (skip_dev_check = true), or
 *
 *   2. does the address exist on the specific device
 *      (skip_dev_check = false)
 */
1854
int ipv6_chk_addr_and_flags(struct net *net, const struct in6_addr *addr,
1855 1856
			    const struct net_device *dev, bool skip_dev_check,
			    int strict, u32 banned_flags)
L
Linus Torvalds 已提交
1857
{
1858
	unsigned int hash = inet6_addr_hash(net, addr);
1859
	const struct net_device *l3mdev;
1860
	struct inet6_ifaddr *ifp;
1861
	u32 ifp_flags;
L
Linus Torvalds 已提交
1862

1863
	rcu_read_lock();
1864

1865
	l3mdev = l3mdev_master_dev_rcu(dev);
1866 1867 1868
	if (skip_dev_check)
		dev = NULL;

1869
	hlist_for_each_entry_rcu(ifp, &inet6_addr_lst[hash], addr_lst) {
1870
		if (!net_eq(dev_net(ifp->idev->dev), net))
1871
			continue;
1872 1873 1874 1875

		if (l3mdev_master_dev_rcu(ifp->idev->dev) != l3mdev)
			continue;

1876 1877 1878 1879 1880 1881
		/* 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
1882
		if (ipv6_addr_equal(&ifp->addr, addr) &&
1883
		    !(ifp_flags&banned_flags) &&
1884
		    (!dev || ifp->idev->dev == dev ||
1885
		     !(ifp->scope&(IFA_LINK|IFA_HOST) || strict))) {
1886
			rcu_read_unlock();
1887
			return 1;
L
Linus Torvalds 已提交
1888 1889
		}
	}
1890

1891
	rcu_read_unlock();
1892
	return 0;
L
Linus Torvalds 已提交
1893
}
1894
EXPORT_SYMBOL(ipv6_chk_addr_and_flags);
1895

L
Linus Torvalds 已提交
1896

1897 1898 1899 1900 1901 1902
/* 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)
{
1903 1904
	const struct inet6_ifaddr *ifa;
	const struct inet6_dev *idev;
1905 1906 1907 1908 1909
	bool ret = false;

	rcu_read_lock();
	idev = __in6_dev_get(dev);
	if (idev) {
1910
		list_for_each_entry_rcu(ifa, &idev->addr_list, if_list) {
1911 1912 1913 1914 1915 1916 1917 1918 1919 1920 1921
			ret = ipv6_prefix_equal(addr, &ifa->addr, prefix_len);
			if (ret)
				break;
		}
	}
	rcu_read_unlock();

	return ret;
}
EXPORT_SYMBOL(ipv6_chk_custom_prefix);

1922
int ipv6_chk_prefix(const struct in6_addr *addr, struct net_device *dev)
1923
{
1924 1925
	const struct inet6_ifaddr *ifa;
	const struct inet6_dev *idev;
1926 1927 1928 1929 1930 1931
	int	onlink;

	onlink = 0;
	rcu_read_lock();
	idev = __in6_dev_get(dev);
	if (idev) {
1932
		list_for_each_entry_rcu(ifa, &idev->addr_list, if_list) {
1933 1934 1935 1936 1937 1938 1939 1940 1941 1942 1943
			onlink = ipv6_prefix_equal(addr, &ifa->addr,
						   ifa->prefix_len);
			if (onlink)
				break;
		}
	}
	rcu_read_unlock();
	return onlink;
}
EXPORT_SYMBOL(ipv6_chk_prefix);

1944
struct inet6_ifaddr *ipv6_get_ifaddr(struct net *net, const struct in6_addr *addr,
1945
				     struct net_device *dev, int strict)
L
Linus Torvalds 已提交
1946
{
1947
	unsigned int hash = inet6_addr_hash(net, addr);
1948
	struct inet6_ifaddr *ifp, *result = NULL;
L
Linus Torvalds 已提交
1949

1950 1951
	rcu_read_lock();
	hlist_for_each_entry_rcu(ifp, &inet6_addr_lst[hash], addr_lst) {
1952
		if (!net_eq(dev_net(ifp->idev->dev), net))
1953
			continue;
L
Linus Torvalds 已提交
1954
		if (ipv6_addr_equal(&ifp->addr, addr)) {
1955
			if (!dev || ifp->idev->dev == dev ||
L
Linus Torvalds 已提交
1956
			    !(ifp->scope&(IFA_LINK|IFA_HOST) || strict)) {
1957
				result = ifp;
L
Linus Torvalds 已提交
1958 1959 1960 1961 1962
				in6_ifa_hold(ifp);
				break;
			}
		}
	}
1963
	rcu_read_unlock();
L
Linus Torvalds 已提交
1964

1965
	return result;
L
Linus Torvalds 已提交
1966 1967 1968 1969
}

/* Gets referenced address, destroys ifaddr */

B
Brian Haley 已提交
1970
static void addrconf_dad_stop(struct inet6_ifaddr *ifp, int dad_failed)
L
Linus Torvalds 已提交
1971
{
1972 1973 1974
	if (dad_failed)
		ifp->flags |= IFA_F_DADFAILED;

1975
	if (ifp->flags&IFA_F_TEMPORARY) {
L
Linus Torvalds 已提交
1976 1977 1978 1979 1980 1981
		struct inet6_ifaddr *ifpub;
		spin_lock_bh(&ifp->lock);
		ifpub = ifp->ifpub;
		if (ifpub) {
			in6_ifa_hold(ifpub);
			spin_unlock_bh(&ifp->lock);
E
Eric Dumazet 已提交
1982
			ipv6_create_tempaddr(ifpub, ifp, true);
L
Linus Torvalds 已提交
1983 1984 1985 1986 1987
			in6_ifa_put(ifpub);
		} else {
			spin_unlock_bh(&ifp->lock);
		}
		ipv6_del_addr(ifp);
1988 1989 1990 1991
	} else if (ifp->flags&IFA_F_PERMANENT || !dad_failed) {
		spin_lock_bh(&ifp->lock);
		addrconf_del_dad_work(ifp);
		ifp->flags |= IFA_F_TENTATIVE;
1992 1993
		if (dad_failed)
			ifp->flags &= ~IFA_F_OPTIMISTIC;
1994 1995 1996 1997
		spin_unlock_bh(&ifp->lock);
		if (dad_failed)
			ipv6_ifa_notify(0, ifp);
		in6_ifa_put(ifp);
1998
	} else {
L
Linus Torvalds 已提交
1999
		ipv6_del_addr(ifp);
2000
	}
L
Linus Torvalds 已提交
2001 2002
}

H
Herbert Xu 已提交
2003 2004 2005 2006
static int addrconf_dad_end(struct inet6_ifaddr *ifp)
{
	int err = -ENOENT;

2007
	spin_lock_bh(&ifp->lock);
H
Herbert Xu 已提交
2008 2009 2010 2011
	if (ifp->state == INET6_IFADDR_STATE_DAD) {
		ifp->state = INET6_IFADDR_STATE_POSTDAD;
		err = 0;
	}
2012
	spin_unlock_bh(&ifp->lock);
H
Herbert Xu 已提交
2013 2014 2015 2016

	return err;
}

2017
void addrconf_dad_failure(struct sk_buff *skb, struct inet6_ifaddr *ifp)
2018
{
2019
	struct inet6_dev *idev = ifp->idev;
2020
	struct net *net = dev_net(ifp->idev->dev);
2021

2022 2023
	if (addrconf_dad_end(ifp)) {
		in6_ifa_put(ifp);
H
Herbert Xu 已提交
2024
		return;
2025
	}
H
Herbert Xu 已提交
2026

2027 2028
	net_info_ratelimited("%s: IPv6 duplicate address %pI6c used by %pM detected!\n",
			     ifp->idev->dev->name, &ifp->addr, eth_hdr(skb)->h_source);
2029

2030 2031 2032 2033 2034 2035
	spin_lock_bh(&ifp->lock);

	if (ifp->flags & IFA_F_STABLE_PRIVACY) {
		struct in6_addr new_addr;
		struct inet6_ifaddr *ifp2;
		int retries = ifp->stable_privacy_retry + 1;
2036 2037 2038 2039 2040 2041 2042 2043
		struct ifa6_config cfg = {
			.pfx = &new_addr,
			.plen = ifp->prefix_len,
			.ifa_flags = ifp->flags,
			.valid_lft = ifp->valid_lft,
			.preferred_lft = ifp->prefered_lft,
			.scope = ifp->scope,
		};
2044

2045
		if (retries > net->ipv6.sysctl.idgen_retries) {
2046 2047 2048 2049 2050 2051 2052 2053 2054 2055 2056
			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;

		spin_unlock_bh(&ifp->lock);
2057

2058 2059 2060 2061 2062 2063 2064 2065
		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);

2066
		ifp2 = ipv6_add_addr(idev, &cfg, false, NULL);
2067 2068 2069 2070 2071 2072 2073 2074
		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);

2075
		addrconf_mod_dad_work(ifp2, net->ipv6.sysctl.idgen_delay);
2076 2077 2078
		in6_ifa_put(ifp2);
lock_errdad:
		spin_lock_bh(&ifp->lock);
2079 2080
	}

2081
errdad:
2082 2083
	/* transition from _POSTDAD to _ERRDAD */
	ifp->state = INET6_IFADDR_STATE_ERRDAD;
2084
	spin_unlock_bh(&ifp->lock);
2085 2086

	addrconf_mod_dad_work(ifp, 0);
2087
	in6_ifa_put(ifp);
2088
}
L
Linus Torvalds 已提交
2089

2090 2091
/* Join to solicited addr multicast group.
 * caller must hold RTNL */
2092
void addrconf_join_solict(struct net_device *dev, const struct in6_addr *addr)
L
Linus Torvalds 已提交
2093 2094 2095 2096 2097 2098 2099 2100 2101 2102
{
	struct in6_addr maddr;

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

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

2103
/* caller must hold RTNL */
2104
void addrconf_leave_solict(struct inet6_dev *idev, const struct in6_addr *addr)
L
Linus Torvalds 已提交
2105 2106 2107 2108 2109 2110 2111 2112 2113 2114
{
	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);
}

2115
/* caller must hold RTNL */
2116
static void addrconf_join_anycast(struct inet6_ifaddr *ifp)
L
Linus Torvalds 已提交
2117 2118
{
	struct in6_addr addr;
2119

2120
	if (ifp->prefix_len >= 127) /* RFC 6164 */
2121
		return;
L
Linus Torvalds 已提交
2122 2123 2124
	ipv6_addr_prefix(&addr, &ifp->addr, ifp->prefix_len);
	if (ipv6_addr_any(&addr))
		return;
W
WANG Cong 已提交
2125
	__ipv6_dev_ac_inc(ifp->idev, &addr);
L
Linus Torvalds 已提交
2126 2127
}

2128
/* caller must hold RTNL */
2129
static void addrconf_leave_anycast(struct inet6_ifaddr *ifp)
L
Linus Torvalds 已提交
2130 2131
{
	struct in6_addr addr;
2132

2133
	if (ifp->prefix_len >= 127) /* RFC 6164 */
2134
		return;
L
Linus Torvalds 已提交
2135 2136 2137 2138 2139 2140
	ipv6_addr_prefix(&addr, &ifp->addr, ifp->prefix_len);
	if (ipv6_addr_any(&addr))
		return;
	__ipv6_dev_ac_dec(ifp->idev, &addr);
}

2141
static int addrconf_ifid_6lowpan(u8 *eui, struct net_device *dev)
2142
{
2143 2144
	switch (dev->addr_len) {
	case ETH_ALEN:
2145 2146 2147 2148 2149
		memcpy(eui, dev->dev_addr, 3);
		eui[3] = 0xFF;
		eui[4] = 0xFE;
		memcpy(eui + 5, dev->dev_addr + 3, 3);
		break;
2150 2151 2152 2153 2154
	case EUI64_ADDR_LEN:
		memcpy(eui, dev->dev_addr, EUI64_ADDR_LEN);
		eui[0] ^= 2;
		break;
	default:
2155
		return -1;
2156 2157
	}

2158 2159 2160
	return 0;
}

2161 2162 2163 2164 2165 2166 2167 2168 2169 2170 2171 2172 2173 2174
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;
}

2175 2176 2177 2178 2179 2180
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);
2181
	eui[7] = *(u8 *)dev->dev_addr;
2182 2183 2184 2185 2186 2187 2188 2189 2190 2191 2192 2193
	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;
}

2194
static int __ipv6_isatap_ifid(u8 *eui, __be32 addr)
2195
{
2196 2197
	if (addr == 0)
		return -1;
2198 2199 2200 2201 2202 2203 2204 2205 2206 2207 2208 2209 2210 2211 2212 2213 2214 2215 2216 2217
	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;
}

2218 2219 2220 2221 2222
static int addrconf_ifid_gre(u8 *eui, struct net_device *dev)
{
	return __ipv6_isatap_ifid(eui, *(__be32 *)dev->dev_addr);
}

2223 2224 2225 2226 2227 2228 2229 2230 2231 2232
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
2233 2234 2235 2236 2237
static int ipv6_generate_eui64(u8 *eui, struct net_device *dev)
{
	switch (dev->type) {
	case ARPHRD_ETHER:
	case ARPHRD_FDDI:
2238
		return addrconf_ifid_eui48(eui, dev);
L
Linus Torvalds 已提交
2239
	case ARPHRD_ARCNET:
2240
		return addrconf_ifid_arcnet(eui, dev);
L
Linus Torvalds 已提交
2241
	case ARPHRD_INFINIBAND:
2242
		return addrconf_ifid_infiniband(eui, dev);
F
Fred L. Templin 已提交
2243
	case ARPHRD_SIT:
2244
		return addrconf_ifid_sit(eui, dev);
2245
	case ARPHRD_IPGRE:
2246
	case ARPHRD_TUNNEL:
2247
		return addrconf_ifid_gre(eui, dev);
2248
	case ARPHRD_6LOWPAN:
2249
		return addrconf_ifid_6lowpan(eui, dev);
2250 2251
	case ARPHRD_IEEE1394:
		return addrconf_ifid_ieee1394(eui, dev);
2252
	case ARPHRD_TUNNEL6:
2253
	case ARPHRD_IP6GRE:
2254
	case ARPHRD_RAWIP:
2255
		return addrconf_ifid_ip6tnl(eui, dev);
L
Linus Torvalds 已提交
2256 2257 2258 2259 2260 2261 2262 2263 2264 2265
	}
	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);
2266 2267 2268
	list_for_each_entry_reverse(ifp, &idev->addr_list, if_list) {
		if (ifp->scope > IFA_LINK)
			break;
L
Linus Torvalds 已提交
2269 2270 2271 2272 2273 2274 2275 2276 2277 2278 2279
		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) */
2280
static void ipv6_regen_rndid(struct inet6_dev *idev)
L
Linus Torvalds 已提交
2281 2282
{
regen:
2283
	get_random_bytes(idev->rndid, sizeof(idev->rndid));
L
Linus Torvalds 已提交
2284 2285 2286 2287 2288 2289 2290 2291
	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
2292
	 *  - ISATAP (RFC4214) 6.1
L
Linus Torvalds 已提交
2293 2294 2295 2296
	 *	00-00-5E-FE-xx-xx-xx-xx
	 *  - value 0
	 *  - XXX: already assigned to an address on the device
	 */
2297
	if (idev->rndid[0] == 0xfd &&
L
Linus Torvalds 已提交
2298 2299 2300 2301 2302 2303 2304 2305 2306 2307 2308
	    (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;
	}
}

2309
static void  ipv6_try_regen_rndid(struct inet6_dev *idev, struct in6_addr *tmpaddr)
2310
{
L
Linus Torvalds 已提交
2311
	if (tmpaddr && memcmp(idev->rndid, &tmpaddr->s6_addr[8], 8) == 0)
2312
		ipv6_regen_rndid(idev);
L
Linus Torvalds 已提交
2313 2314 2315 2316 2317 2318 2319
}

/*
 *	Add prefix route.
 */

static void
2320 2321 2322
addrconf_prefix_route(struct in6_addr *pfx, int plen, u32 metric,
		      struct net_device *dev, unsigned long expires,
		      u32 flags, gfp_t gfp_flags)
L
Linus Torvalds 已提交
2323
{
2324
	struct fib6_config cfg = {
D
David Ahern 已提交
2325
		.fc_table = l3mdev_fib_table(dev) ? : RT6_TABLE_PREFIX,
2326
		.fc_metric = metric ? : IP6_RT_PRIO_ADDRCONF,
2327 2328 2329 2330
		.fc_ifindex = dev->ifindex,
		.fc_expires = expires,
		.fc_dst_len = plen,
		.fc_flags = RTF_UP | flags,
2331
		.fc_nlinfo.nl_net = dev_net(dev),
2332
		.fc_protocol = RTPROT_KERNEL,
2333
		.fc_type = RTN_UNICAST,
2334
	};
L
Linus Torvalds 已提交
2335

A
Alexey Dobriyan 已提交
2336
	cfg.fc_dst = *pfx;
L
Linus Torvalds 已提交
2337 2338 2339 2340 2341

	/* 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
2342
#if IS_ENABLED(CONFIG_IPV6_SIT)
2343 2344
	if (dev->type == ARPHRD_SIT && (dev->flags & IFF_POINTOPOINT))
		cfg.fc_flags |= RTF_NONEXTHOP;
2345
#endif
L
Linus Torvalds 已提交
2346

2347
	ip6_route_add(&cfg, gfp_flags, NULL);
L
Linus Torvalds 已提交
2348 2349
}

2350

2351
static struct fib6_info *addrconf_get_prefix_route(const struct in6_addr *pfx,
2352 2353 2354 2355 2356
						  int plen,
						  const struct net_device *dev,
						  u32 flags, u32 noflags)
{
	struct fib6_node *fn;
2357
	struct fib6_info *rt = NULL;
2358
	struct fib6_table *table;
D
David Ahern 已提交
2359
	u32 tb_id = l3mdev_fib_table(dev) ? : RT6_TABLE_PREFIX;
2360

D
David Ahern 已提交
2361
	table = fib6_get_table(dev_net(dev), tb_id);
2362
	if (!table)
2363 2364
		return NULL;

2365
	rcu_read_lock();
2366
	fn = fib6_locate(&table->tb6_root, pfx, plen, NULL, 0, true);
2367 2368
	if (!fn)
		goto out;
2369

2370
	for_each_fib6_node_rt_rcu(fn) {
2371
		if (rt->fib6_nh.nh_dev->ifindex != dev->ifindex)
2372
			continue;
2373
		if ((rt->fib6_flags & flags) != flags)
2374
			continue;
2375
		if ((rt->fib6_flags & noflags) != 0)
2376
			continue;
2377
		fib6_info_hold(rt);
2378 2379 2380
		break;
	}
out:
2381
	rcu_read_unlock();
2382 2383 2384 2385
	return rt;
}


L
Linus Torvalds 已提交
2386 2387 2388 2389
/* Create "default" multicast route to the interface */

static void addrconf_add_mroute(struct net_device *dev)
{
2390
	struct fib6_config cfg = {
D
David Ahern 已提交
2391
		.fc_table = l3mdev_fib_table(dev) ? : RT6_TABLE_LOCAL,
2392 2393 2394 2395
		.fc_metric = IP6_RT_PRIO_ADDRCONF,
		.fc_ifindex = dev->ifindex,
		.fc_dst_len = 8,
		.fc_flags = RTF_UP,
2396
		.fc_type = RTN_UNICAST,
2397
		.fc_nlinfo.nl_net = dev_net(dev),
2398 2399 2400 2401
	};

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

2402
	ip6_route_add(&cfg, GFP_ATOMIC, NULL);
L
Linus Torvalds 已提交
2403 2404 2405 2406 2407 2408 2409 2410
}

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

	ASSERT_RTNL();

S
Stephen Hemminger 已提交
2411 2412
	idev = ipv6_find_idev(dev);
	if (!idev)
2413 2414 2415 2416
		return ERR_PTR(-ENOBUFS);

	if (idev->cnf.disable_ipv6)
		return ERR_PTR(-EACCES);
L
Linus Torvalds 已提交
2417 2418

	/* Add default multicast route */
2419
	if (!(dev->flags & IFF_LOOPBACK) && !netif_is_l3_master(dev))
2420
		addrconf_add_mroute(dev);
L
Linus Torvalds 已提交
2421 2422 2423 2424

	return idev;
}

2425 2426 2427 2428 2429 2430 2431 2432 2433 2434 2435 2436 2437 2438 2439 2440 2441 2442 2443 2444 2445 2446 2447 2448 2449 2450 2451 2452 2453 2454
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 -
J
Jiri Bohac 已提交
2455
			       idev->desync_factor - age;
2456 2457 2458 2459 2460 2461 2462 2463 2464 2465 2466 2467 2468 2469 2470 2471 2472 2473 2474 2475 2476 2477 2478 2479 2480 2481 2482 2483 2484 2485
		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);
E
Eric Dumazet 已提交
2486
		ipv6_create_tempaddr(ifp, NULL, false);
2487 2488 2489 2490 2491
	} else {
		read_unlock_bh(&idev->lock);
	}
}

2492 2493
static bool is_addr_mode_generate_stable(struct inet6_dev *idev)
{
2494 2495
	return idev->cnf.addr_gen_mode == IN6_ADDR_GEN_MODE_STABLE_PRIVACY ||
	       idev->cnf.addr_gen_mode == IN6_ADDR_GEN_MODE_RANDOM;
2496 2497
}

A
Alexander Aring 已提交
2498 2499 2500 2501 2502 2503
int addrconf_prefix_rcv_add_addr(struct net *net, struct net_device *dev,
				 const struct prefix_info *pinfo,
				 struct inet6_dev *in6_dev,
				 const struct in6_addr *addr, int addr_type,
				 u32 addr_flags, bool sllao, bool tokenized,
				 __u32 valid_lft, u32 prefered_lft)
2504 2505 2506 2507 2508 2509
{
	struct inet6_ifaddr *ifp = ipv6_get_ifaddr(net, addr, dev, 1);
	int create = 0, update_lft = 0;

	if (!ifp && valid_lft) {
		int max_addresses = in6_dev->cnf.max_addresses;
2510 2511 2512 2513 2514 2515 2516 2517
		struct ifa6_config cfg = {
			.pfx = addr,
			.plen = pinfo->prefix_len,
			.ifa_flags = addr_flags,
			.valid_lft = valid_lft,
			.preferred_lft = prefered_lft,
			.scope = addr_type & IPV6_ADDR_SCOPE_MASK,
		};
2518 2519

#ifdef CONFIG_IPV6_OPTIMISTIC_DAD
2520 2521
		if ((net->ipv6.devconf_all->optimistic_dad ||
		     in6_dev->cnf.optimistic_dad) &&
2522
		    !net->ipv6.devconf_all->forwarding && sllao)
2523
			cfg.ifa_flags |= IFA_F_OPTIMISTIC;
2524 2525 2526 2527 2528 2529 2530
#endif

		/* 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)
2531
			ifp = ipv6_add_addr(in6_dev, &cfg, false, NULL);
2532 2533 2534 2535 2536 2537 2538 2539 2540 2541 2542 2543 2544 2545 2546 2547 2548 2549 2550 2551 2552 2553 2554 2555 2556

		if (IS_ERR_OR_NULL(ifp))
			return -1;

		create = 1;
		spin_lock_bh(&ifp->lock);
		ifp->flags |= IFA_F_MANAGETEMPADDR;
		ifp->cstamp = jiffies;
		ifp->tokenized = tokenized;
		spin_unlock_bh(&ifp->lock);
		addrconf_dad_start(ifp);
	}

	if (ifp) {
		u32 flags;
		unsigned long now;
		u32 stored_lft;

		/* update lifetime (RFC2462 5.5.3 e) */
		spin_lock_bh(&ifp->lock);
		now = jiffies;
		if (ifp->valid_lft > (now - ifp->tstamp) / HZ)
			stored_lft = ifp->valid_lft - (now - ifp->tstamp) / HZ;
		else
			stored_lft = 0;
2557
		if (!create && stored_lft) {
2558 2559 2560 2561 2562 2563 2564 2565 2566 2567 2568 2569 2570 2571 2572 2573 2574 2575 2576 2577 2578 2579 2580 2581 2582 2583 2584 2585 2586 2587 2588 2589 2590 2591 2592 2593 2594 2595 2596
			const u32 minimum_lft = min_t(u32,
				stored_lft, MIN_VALID_LIFETIME);
			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;
		}

		if (update_lft) {
			ifp->valid_lft = valid_lft;
			ifp->prefered_lft = prefered_lft;
			ifp->tstamp = now;
			flags = ifp->flags;
			ifp->flags &= ~IFA_F_DEPRECATED;
			spin_unlock_bh(&ifp->lock);

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

		manage_tempaddrs(in6_dev, ifp, valid_lft, prefered_lft,
				 create, now);

		in6_ifa_put(ifp);
		addrconf_verify();
	}

	return 0;
}
A
Alexander Aring 已提交
2597
EXPORT_SYMBOL_GPL(addrconf_prefix_rcv_add_addr);
2598

2599
void addrconf_prefix_rcv(struct net_device *dev, u8 *opt, int len, bool sllao)
L
Linus Torvalds 已提交
2600 2601 2602 2603
{
	struct prefix_info *pinfo;
	__u32 valid_lft;
	__u32 prefered_lft;
2604
	int addr_type, err;
2605
	u32 addr_flags = 0;
L
Linus Torvalds 已提交
2606
	struct inet6_dev *in6_dev;
2607
	struct net *net = dev_net(dev);
L
Linus Torvalds 已提交
2608 2609

	pinfo = (struct prefix_info *) opt;
2610

L
Linus Torvalds 已提交
2611
	if (len < sizeof(struct prefix_info)) {
2612
		netdev_dbg(dev, "addrconf: prefix option too short\n");
L
Linus Torvalds 已提交
2613 2614
		return;
	}
2615

L
Linus Torvalds 已提交
2616 2617 2618 2619 2620 2621 2622 2623 2624 2625 2626 2627 2628
	/*
	 *	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) {
2629
		net_warn_ratelimited("addrconf: prefix option has invalid lifetime\n");
L
Linus Torvalds 已提交
2630 2631 2632 2633 2634
		return;
	}

	in6_dev = in6_dev_get(dev);

2635
	if (!in6_dev) {
2636 2637
		net_dbg_ratelimited("addrconf: device %s not configured\n",
				    dev->name);
L
Linus Torvalds 已提交
2638 2639 2640 2641 2642 2643 2644 2645 2646
		return;
	}

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

2647
	if (pinfo->onlink) {
2648
		struct fib6_info *rt;
2649 2650
		unsigned long rt_expires;

2651 2652 2653 2654 2655
		/* 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.
		 */
2656 2657 2658 2659
		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
2660

2661 2662
		if (addrconf_finite_timeout(rt_expires))
			rt_expires *= HZ;
L
Linus Torvalds 已提交
2663

2664 2665 2666 2667 2668
		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
2669

2670
		if (rt) {
2671 2672
			/* Autoconf prefix route */
			if (valid_lft == 0) {
2673
				ip6_del_rt(net, rt);
2674
				rt = NULL;
2675
			} else if (addrconf_finite_timeout(rt_expires)) {
2676
				/* not infinity */
2677
				fib6_set_expires(rt, jiffies + rt_expires);
2678
			} else {
2679
				fib6_clean_expires(rt);
L
Linus Torvalds 已提交
2680 2681
			}
		} else if (valid_lft) {
2682
			clock_t expires = 0;
2683 2684
			int flags = RTF_ADDRCONF | RTF_PREFIX_RT;
			if (addrconf_finite_timeout(rt_expires)) {
2685 2686 2687 2688
				/* not infinity */
				flags |= RTF_EXPIRES;
				expires = jiffies_to_clock_t(rt_expires);
			}
L
Linus Torvalds 已提交
2689
			addrconf_prefix_route(&pinfo->prefix, pinfo->prefix_len,
2690 2691
					      0, dev, expires, flags,
					      GFP_ATOMIC);
L
Linus Torvalds 已提交
2692
		}
2693
		fib6_info_release(rt);
L
Linus Torvalds 已提交
2694 2695 2696 2697 2698 2699
	}

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

	if (pinfo->autoconf && in6_dev->cnf.autoconf) {
		struct in6_addr addr;
2700
		bool tokenized = false, dev_addr_generated = false;
L
Linus Torvalds 已提交
2701 2702 2703

		if (pinfo->prefix_len == 64) {
			memcpy(&addr, &pinfo->prefix, 8);
2704 2705 2706 2707 2708 2709

			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);
2710
				tokenized = true;
2711
			} else if (is_addr_mode_generate_stable(in6_dev) &&
2712 2713
				   !ipv6_generate_stable_address(&addr, 0,
								 in6_dev)) {
2714
				addr_flags |= IFA_F_STABLE_PRIVACY;
2715
				goto ok;
2716 2717
			} else if (ipv6_generate_eui64(addr.s6_addr + 8, dev) &&
				   ipv6_inherit_eui64(addr.s6_addr + 8, in6_dev)) {
2718
				goto put;
2719 2720
			} else {
				dev_addr_generated = true;
L
Linus Torvalds 已提交
2721 2722 2723
			}
			goto ok;
		}
2724 2725
		net_dbg_ratelimited("IPv6 addrconf: prefix with wrong length %d\n",
				    pinfo->prefix_len);
2726
		goto put;
L
Linus Torvalds 已提交
2727 2728

ok:
2729 2730 2731 2732 2733 2734 2735
		err = addrconf_prefix_rcv_add_addr(net, dev, pinfo, in6_dev,
						   &addr, addr_type,
						   addr_flags, sllao,
						   tokenized, valid_lft,
						   prefered_lft);
		if (err)
			goto put;
2736 2737 2738 2739 2740 2741 2742 2743 2744

		/* Ignore error case here because previous prefix add addr was
		 * successful which will be notified.
		 */
		ndisc_ops_prefix_rcv_add_addr(net, dev, pinfo, in6_dev, &addr,
					      addr_type, addr_flags, sllao,
					      tokenized, valid_lft,
					      prefered_lft,
					      dev_addr_generated);
L
Linus Torvalds 已提交
2745 2746
	}
	inet6_prefix_notify(RTM_NEWPREFIX, in6_dev, pinfo);
2747
put:
L
Linus Torvalds 已提交
2748 2749 2750 2751 2752 2753 2754 2755
	in6_dev_put(in6_dev);
}

/*
 *	Set destination address.
 *	Special case for SIT interfaces where we create a new "virtual"
 *	device.
 */
2756
int addrconf_set_dstaddr(struct net *net, void __user *arg)
L
Linus Torvalds 已提交
2757 2758 2759 2760 2761 2762 2763 2764 2765 2766 2767
{
	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;

2768
	dev = __dev_get_by_index(net, ireq.ifr6_ifindex);
L
Linus Torvalds 已提交
2769 2770

	err = -ENODEV;
2771
	if (!dev)
L
Linus Torvalds 已提交
2772 2773
		goto err_exit;

A
Amerigo Wang 已提交
2774
#if IS_ENABLED(CONFIG_IPV6_SIT)
L
Linus Torvalds 已提交
2775
	if (dev->type == ARPHRD_SIT) {
2776
		const struct net_device_ops *ops = dev->netdev_ops;
L
Linus Torvalds 已提交
2777 2778 2779 2780 2781 2782 2783 2784 2785 2786 2787 2788 2789 2790
		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;
2791
		ifr.ifr_ifru.ifru_data = (__force void __user *)&p;
L
Linus Torvalds 已提交
2792

2793 2794 2795 2796 2797 2798 2799 2800
		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
2801 2802 2803

		if (err == 0) {
			err = -ENOBUFS;
2804 2805
			dev = __dev_get_by_name(net, p.name);
			if (!dev)
L
Linus Torvalds 已提交
2806 2807 2808 2809
				goto err_exit;
			err = dev_open(dev);
		}
	}
2810
#endif
L
Linus Torvalds 已提交
2811 2812 2813 2814 2815 2816

err_exit:
	rtnl_unlock();
	return err;
}

2817 2818 2819 2820 2821 2822 2823 2824 2825
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)
2826
		ret = ipv6_sock_mc_join(sk, ifindex, addr);
2827
	else
2828
		ret = ipv6_sock_mc_drop(sk, ifindex, addr);
2829 2830 2831 2832 2833
	release_sock(sk);

	return ret;
}

L
Linus Torvalds 已提交
2834 2835 2836
/*
 *	Manual configuration of address on an interface
 */
2837
static int inet6_addr_add(struct net *net, int ifindex,
2838
			  struct ifa6_config *cfg,
2839
			  struct netlink_ext_ack *extack)
L
Linus Torvalds 已提交
2840 2841 2842 2843
{
	struct inet6_ifaddr *ifp;
	struct inet6_dev *idev;
	struct net_device *dev;
2844 2845
	unsigned long timeout;
	clock_t expires;
2846
	u32 flags;
L
Linus Torvalds 已提交
2847 2848

	ASSERT_RTNL();
2849

2850
	if (cfg->plen > 128)
2851 2852
		return -EINVAL;

2853
	/* check the lifetime */
2854
	if (!cfg->valid_lft || cfg->preferred_lft > cfg->valid_lft)
2855 2856
		return -EINVAL;

2857
	if (cfg->ifa_flags & IFA_F_MANAGETEMPADDR && cfg->plen != 64)
2858 2859
		return -EINVAL;

2860 2861
	dev = __dev_get_by_index(net, ifindex);
	if (!dev)
L
Linus Torvalds 已提交
2862
		return -ENODEV;
2863

2864 2865 2866
	idev = addrconf_add_dev(dev);
	if (IS_ERR(idev))
		return PTR_ERR(idev);
L
Linus Torvalds 已提交
2867

2868
	if (cfg->ifa_flags & IFA_F_MCAUTOJOIN) {
2869
		int ret = ipv6_mc_config(net->ipv6.mc_autojoin_sk,
2870
					 true, cfg->pfx, ifindex);
2871 2872 2873 2874 2875

		if (ret < 0)
			return ret;
	}

2876
	cfg->scope = ipv6_addr_scope(cfg->pfx);
L
Linus Torvalds 已提交
2877

2878
	timeout = addrconf_timeout_fixup(cfg->valid_lft, HZ);
2879 2880
	if (addrconf_finite_timeout(timeout)) {
		expires = jiffies_to_clock_t(timeout * HZ);
2881
		cfg->valid_lft = timeout;
2882
		flags = RTF_EXPIRES;
2883 2884 2885
	} else {
		expires = 0;
		flags = 0;
2886
		cfg->ifa_flags |= IFA_F_PERMANENT;
2887
	}
2888

2889
	timeout = addrconf_timeout_fixup(cfg->preferred_lft, HZ);
2890 2891
	if (addrconf_finite_timeout(timeout)) {
		if (timeout == 0)
2892 2893
			cfg->ifa_flags |= IFA_F_DEPRECATED;
		cfg->preferred_lft = timeout;
2894
	}
2895

2896
	ifp = ipv6_add_addr(idev, cfg, true, extack);
L
Linus Torvalds 已提交
2897
	if (!IS_ERR(ifp)) {
2898
		if (!(cfg->ifa_flags & IFA_F_NOPREFIXROUTE)) {
2899 2900 2901
			addrconf_prefix_route(&ifp->addr, ifp->prefix_len,
					      ifp->rt_priority, dev, expires,
					      flags, GFP_KERNEL);
2902 2903
		}

2904 2905 2906 2907 2908
		/* Send a netlink notification if DAD is enabled and
		 * optimistic flag is not set
		 */
		if (!(ifp->flags & (IFA_F_OPTIMISTIC | IFA_F_NODAD)))
			ipv6_ifa_notify(0, ifp);
2909 2910 2911 2912 2913
		/*
		 * Note that section 3.1 of RFC 4429 indicates
		 * that the Optimistic flag should not be set for
		 * manually configured addresses
		 */
2914
		addrconf_dad_start(ifp);
2915 2916 2917
		if (cfg->ifa_flags & IFA_F_MANAGETEMPADDR)
			manage_tempaddrs(idev, ifp, cfg->valid_lft,
					 cfg->preferred_lft, true, jiffies);
L
Linus Torvalds 已提交
2918
		in6_ifa_put(ifp);
2919
		addrconf_verify_rtnl();
L
Linus Torvalds 已提交
2920
		return 0;
2921 2922 2923
	} else if (cfg->ifa_flags & IFA_F_MCAUTOJOIN) {
		ipv6_mc_config(net->ipv6.mc_autojoin_sk, false,
			       cfg->pfx, ifindex);
L
Linus Torvalds 已提交
2924 2925 2926 2927 2928
	}

	return PTR_ERR(ifp);
}

2929 2930
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
2931 2932 2933 2934
{
	struct inet6_ifaddr *ifp;
	struct inet6_dev *idev;
	struct net_device *dev;
2935

2936 2937 2938
	if (plen > 128)
		return -EINVAL;

2939 2940
	dev = __dev_get_by_index(net, ifindex);
	if (!dev)
L
Linus Torvalds 已提交
2941 2942
		return -ENODEV;

2943
	idev = __in6_dev_get(dev);
2944
	if (!idev)
L
Linus Torvalds 已提交
2945 2946 2947
		return -ENXIO;

	read_lock_bh(&idev->lock);
2948
	list_for_each_entry(ifp, &idev->addr_list, if_list) {
L
Linus Torvalds 已提交
2949 2950 2951 2952
		if (ifp->prefix_len == plen &&
		    ipv6_addr_equal(pfx, &ifp->addr)) {
			in6_ifa_hold(ifp);
			read_unlock_bh(&idev->lock);
2953

2954 2955 2956 2957
			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
2958
			ipv6_del_addr(ifp);
2959
			addrconf_verify_rtnl();
2960 2961 2962 2963
			if (ipv6_addr_is_multicast(pfx)) {
				ipv6_mc_config(net->ipv6.mc_autojoin_sk,
					       false, pfx, dev->ifindex);
			}
L
Linus Torvalds 已提交
2964 2965 2966 2967 2968 2969 2970 2971
			return 0;
		}
	}
	read_unlock_bh(&idev->lock);
	return -EADDRNOTAVAIL;
}


2972
int addrconf_add_ifaddr(struct net *net, void __user *arg)
L
Linus Torvalds 已提交
2973
{
2974 2975 2976 2977 2978
	struct ifa6_config cfg = {
		.ifa_flags = IFA_F_PERMANENT,
		.preferred_lft = INFINITY_LIFE_TIME,
		.valid_lft = INFINITY_LIFE_TIME,
	};
L
Linus Torvalds 已提交
2979 2980
	struct in6_ifreq ireq;
	int err;
2981

2982
	if (!ns_capable(net->user_ns, CAP_NET_ADMIN))
L
Linus Torvalds 已提交
2983
		return -EPERM;
2984

L
Linus Torvalds 已提交
2985 2986 2987
	if (copy_from_user(&ireq, arg, sizeof(struct in6_ifreq)))
		return -EFAULT;

2988 2989 2990
	cfg.pfx = &ireq.ifr6_addr;
	cfg.plen = ireq.ifr6_prefixlen;

L
Linus Torvalds 已提交
2991
	rtnl_lock();
2992
	err = inet6_addr_add(net, ireq.ifr6_ifindex, &cfg, NULL);
L
Linus Torvalds 已提交
2993 2994 2995 2996
	rtnl_unlock();
	return err;
}

2997
int addrconf_del_ifaddr(struct net *net, void __user *arg)
L
Linus Torvalds 已提交
2998 2999 3000
{
	struct in6_ifreq ireq;
	int err;
3001

3002
	if (!ns_capable(net->user_ns, CAP_NET_ADMIN))
L
Linus Torvalds 已提交
3003 3004 3005 3006 3007 3008
		return -EPERM;

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

	rtnl_lock();
3009
	err = inet6_addr_del(net, ireq.ifr6_ifindex, 0, &ireq.ifr6_addr,
3010
			     ireq.ifr6_prefixlen);
L
Linus Torvalds 已提交
3011 3012 3013 3014
	rtnl_unlock();
	return err;
}

3015 3016 3017 3018
static void add_addr(struct inet6_dev *idev, const struct in6_addr *addr,
		     int plen, int scope)
{
	struct inet6_ifaddr *ifp;
3019 3020 3021 3022 3023 3024 3025 3026
	struct ifa6_config cfg = {
		.pfx = addr,
		.plen = plen,
		.ifa_flags = IFA_F_PERMANENT,
		.valid_lft = INFINITY_LIFE_TIME,
		.preferred_lft = INFINITY_LIFE_TIME,
		.scope = scope
	};
3027

3028
	ifp = ipv6_add_addr(idev, &cfg, true, NULL);
3029 3030 3031 3032
	if (!IS_ERR(ifp)) {
		spin_lock_bh(&ifp->lock);
		ifp->flags &= ~IFA_F_TENTATIVE;
		spin_unlock_bh(&ifp->lock);
3033
		rt_genid_bump_ipv6(dev_net(idev->dev));
3034 3035 3036 3037 3038
		ipv6_ifa_notify(RTM_NEWADDR, ifp);
		in6_ifa_put(ifp);
	}
}

A
Amerigo Wang 已提交
3039
#if IS_ENABLED(CONFIG_IPV6_SIT)
L
Linus Torvalds 已提交
3040 3041 3042 3043
static void sit_add_v4_addrs(struct inet6_dev *idev)
{
	struct in6_addr addr;
	struct net_device *dev;
3044
	struct net *net = dev_net(idev->dev);
3045
	int scope, plen;
3046
	u32 pflags = 0;
L
Linus Torvalds 已提交
3047 3048 3049 3050 3051 3052 3053 3054 3055

	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;
3056
		plen = 64;
L
Linus Torvalds 已提交
3057 3058
	} else {
		scope = IPV6_ADDR_COMPATv4;
3059
		plen = 96;
3060
		pflags |= RTF_NONEXTHOP;
L
Linus Torvalds 已提交
3061 3062 3063
	}

	if (addr.s6_addr32[3]) {
3064
		add_addr(idev, &addr, plen, scope);
3065
		addrconf_prefix_route(&addr, plen, 0, idev->dev, 0, pflags,
3066
				      GFP_ATOMIC);
L
Linus Torvalds 已提交
3067 3068 3069
		return;
	}

3070
	for_each_netdev(net, dev) {
3071
		struct in_device *in_dev = __in_dev_get_rtnl(dev);
L
Linus Torvalds 已提交
3072
		if (in_dev && (dev->flags & IFF_UP)) {
3073
			struct in_ifaddr *ifa;
L
Linus Torvalds 已提交
3074 3075 3076 3077 3078 3079 3080 3081 3082 3083 3084 3085 3086 3087 3088

			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;
				}

3089
				add_addr(idev, &addr, plen, flag);
3090 3091
				addrconf_prefix_route(&addr, plen, 0, idev->dev,
						      0, pflags, GFP_ATOMIC);
L
Linus Torvalds 已提交
3092 3093
			}
		}
3094
	}
L
Linus Torvalds 已提交
3095
}
3096
#endif
L
Linus Torvalds 已提交
3097 3098 3099 3100 3101 3102 3103 3104 3105

static void init_loopback(struct net_device *dev)
{
	struct inet6_dev  *idev;

	/* ::1 */

	ASSERT_RTNL();

3106
	idev = ipv6_find_idev(dev);
3107
	if (!idev) {
3108
		pr_debug("%s: add_dev failed\n", __func__);
L
Linus Torvalds 已提交
3109 3110 3111
		return;
	}

3112
	add_addr(idev, &in6addr_loopback, 128, IFA_HOST);
L
Linus Torvalds 已提交
3113 3114
}

3115 3116
void addrconf_add_linklocal(struct inet6_dev *idev,
			    const struct in6_addr *addr, u32 flags)
L
Linus Torvalds 已提交
3117
{
3118 3119 3120 3121 3122 3123 3124 3125
	struct ifa6_config cfg = {
		.pfx = addr,
		.plen = 64,
		.ifa_flags = flags | IFA_F_PERMANENT,
		.valid_lft = INFINITY_LIFE_TIME,
		.preferred_lft = INFINITY_LIFE_TIME,
		.scope = IFA_LINK
	};
3126
	struct inet6_ifaddr *ifp;
3127 3128

#ifdef CONFIG_IPV6_OPTIMISTIC_DAD
3129 3130
	if ((dev_net(idev->dev)->ipv6.devconf_all->optimistic_dad ||
	     idev->cnf.optimistic_dad) &&
3131
	    !dev_net(idev->dev)->ipv6.devconf_all->forwarding)
3132
		cfg.ifa_flags |= IFA_F_OPTIMISTIC;
3133
#endif
L
Linus Torvalds 已提交
3134

3135
	ifp = ipv6_add_addr(idev, &cfg, true, NULL);
L
Linus Torvalds 已提交
3136
	if (!IS_ERR(ifp)) {
3137
		addrconf_prefix_route(&ifp->addr, ifp->prefix_len, 0, idev->dev,
3138
				      0, 0, GFP_ATOMIC);
3139
		addrconf_dad_start(ifp);
L
Linus Torvalds 已提交
3140 3141 3142
		in6_ifa_put(ifp);
	}
}
3143
EXPORT_SYMBOL_GPL(addrconf_add_linklocal);
L
Linus Torvalds 已提交
3144

3145 3146 3147 3148 3149 3150 3151 3152 3153 3154 3155 3156 3157 3158 3159 3160 3161 3162 3163 3164 3165 3166 3167 3168 3169 3170 3171 3172
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;
3173
			__be32 prefix[2];
3174 3175 3176 3177 3178 3179 3180 3181 3182 3183 3184 3185 3186 3187 3188 3189 3190 3191 3192 3193 3194 3195 3196 3197 3198
			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);
3199 3200
	data.prefix[0] = address->s6_addr32[0];
	data.prefix[1] = address->s6_addr32[1];
3201 3202 3203 3204 3205 3206
	data.secret = secret;
	data.dad_count = dad_count;

	sha_transform(digest, data.__data, workspace);

	temp = *address;
3207 3208
	temp.s6_addr32[2] = (__force __be32)digest[0];
	temp.s6_addr32[3] = (__force __be32)digest[1];
3209 3210 3211 3212 3213

	spin_unlock_bh(&lock);

	if (ipv6_reserved_interfaceid(temp)) {
		dad_count++;
3214
		if (dad_count > dev_net(idev->dev)->ipv6.sysctl.idgen_retries)
3215 3216 3217 3218 3219 3220 3221 3222
			return -1;
		goto retry;
	}

	*address = temp;
	return 0;
}

3223 3224 3225 3226 3227 3228 3229 3230 3231 3232 3233
static void ipv6_gen_mode_random_init(struct inet6_dev *idev)
{
	struct ipv6_stable_secret *s = &idev->cnf.stable_secret;

	if (s->initialized)
		return;
	s = &idev->cnf.stable_secret;
	get_random_bytes(&s->secret, sizeof(s->secret));
	s->initialized = true;
}

3234 3235
static void addrconf_addr_gen(struct inet6_dev *idev, bool prefix_route)
{
3236 3237
	struct in6_addr addr;

D
David Ahern 已提交
3238 3239 3240 3241
	/* no link local addresses on L3 master devices */
	if (netif_is_l3_master(idev->dev))
		return;

3242
	ipv6_addr_set(&addr, htonl(0xFE800000), 0, 0, 0);
3243

3244
	switch (idev->cnf.addr_gen_mode) {
3245 3246 3247 3248
	case IN6_ADDR_GEN_MODE_RANDOM:
		ipv6_gen_mode_random_init(idev);
		/* fallthrough */
	case IN6_ADDR_GEN_MODE_STABLE_PRIVACY:
3249
		if (!ipv6_generate_stable_address(&addr, 0, idev))
3250 3251
			addrconf_add_linklocal(idev, &addr,
					       IFA_F_STABLE_PRIVACY);
3252
		else if (prefix_route)
3253
			addrconf_prefix_route(&addr, 64, 0, idev->dev,
3254
					      0, 0, GFP_KERNEL);
3255 3256
		break;
	case IN6_ADDR_GEN_MODE_EUI64:
3257 3258 3259 3260 3261
		/* 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)
3262
			addrconf_add_linklocal(idev, &addr, 0);
3263
		else if (prefix_route)
3264
			addrconf_prefix_route(&addr, 64, 0, idev->dev,
3265
					      0, 0, GFP_KERNEL);
3266 3267 3268 3269 3270
		break;
	case IN6_ADDR_GEN_MODE_NONE:
	default:
		/* will not add any link local address */
		break;
3271 3272 3273
	}
}

L
Linus Torvalds 已提交
3274 3275
static void addrconf_dev_config(struct net_device *dev)
{
3276
	struct inet6_dev *idev;
L
Linus Torvalds 已提交
3277 3278 3279

	ASSERT_RTNL();

3280 3281 3282
	if ((dev->type != ARPHRD_ETHER) &&
	    (dev->type != ARPHRD_FDDI) &&
	    (dev->type != ARPHRD_ARCNET) &&
3283
	    (dev->type != ARPHRD_INFINIBAND) &&
3284
	    (dev->type != ARPHRD_IEEE1394) &&
3285
	    (dev->type != ARPHRD_TUNNEL6) &&
3286
	    (dev->type != ARPHRD_6LOWPAN) &&
3287 3288 3289
	    (dev->type != ARPHRD_IP6GRE) &&
	    (dev->type != ARPHRD_IPGRE) &&
	    (dev->type != ARPHRD_TUNNEL) &&
3290 3291
	    (dev->type != ARPHRD_NONE) &&
	    (dev->type != ARPHRD_RAWIP)) {
3292 3293 3294 3295
		/* Alas, we support only Ethernet autoconfiguration. */
		return;
	}

L
Linus Torvalds 已提交
3296
	idev = addrconf_add_dev(dev);
3297
	if (IS_ERR(idev))
L
Linus Torvalds 已提交
3298 3299
		return;

3300 3301
	/* this device type has no EUI support */
	if (dev->type == ARPHRD_NONE &&
3302 3303
	    idev->cnf.addr_gen_mode == IN6_ADDR_GEN_MODE_EUI64)
		idev->cnf.addr_gen_mode = IN6_ADDR_GEN_MODE_RANDOM;
3304

3305
	addrconf_addr_gen(idev, false);
L
Linus Torvalds 已提交
3306 3307
}

A
Amerigo Wang 已提交
3308
#if IS_ENABLED(CONFIG_IPV6_SIT)
L
Linus Torvalds 已提交
3309 3310 3311 3312 3313 3314
static void addrconf_sit_config(struct net_device *dev)
{
	struct inet6_dev *idev;

	ASSERT_RTNL();

3315 3316 3317
	/*
	 * Configure the tunnel with one of our IPv4
	 * addresses... we should configure all of
L
Linus Torvalds 已提交
3318 3319 3320
	 * our v4 addrs in the tunnel
	 */

3321
	idev = ipv6_find_idev(dev);
3322
	if (!idev) {
3323
		pr_debug("%s: add_dev failed\n", __func__);
L
Linus Torvalds 已提交
3324 3325 3326
		return;
	}

F
Fred L. Templin 已提交
3327
	if (dev->priv_flags & IFF_ISATAP) {
3328
		addrconf_addr_gen(idev, false);
F
Fred L. Templin 已提交
3329 3330 3331
		return;
	}

L
Linus Torvalds 已提交
3332 3333
	sit_add_v4_addrs(idev);

3334
	if (dev->flags&IFF_POINTOPOINT)
L
Linus Torvalds 已提交
3335 3336
		addrconf_add_mroute(dev);
}
3337
#endif
L
Linus Torvalds 已提交
3338

A
Amerigo Wang 已提交
3339
#if IS_ENABLED(CONFIG_NET_IPGRE)
3340 3341 3342 3343 3344 3345
static void addrconf_gre_config(struct net_device *dev)
{
	struct inet6_dev *idev;

	ASSERT_RTNL();

3346
	idev = ipv6_find_idev(dev);
3347
	if (!idev) {
3348
		pr_debug("%s: add_dev failed\n", __func__);
3349 3350 3351
		return;
	}

3352
	addrconf_addr_gen(idev, true);
3353 3354
	if (dev->flags & IFF_POINTOPOINT)
		addrconf_add_mroute(dev);
3355 3356 3357
}
#endif

3358 3359
static int fixup_permanent_addr(struct net *net,
				struct inet6_dev *idev,
3360 3361
				struct inet6_ifaddr *ifp)
{
3362
	/* !fib6_node means the host route was removed from the
3363 3364 3365
	 * FIB, for example, if 'lo' device is taken down. In that
	 * case regenerate the host route.
	 */
3366
	if (!ifp->rt || !ifp->rt->fib6_node) {
3367
		struct fib6_info *f6i, *prev;
3368

3369 3370 3371 3372
		f6i = addrconf_f6i_alloc(net, idev, &ifp->addr, false,
					 GFP_ATOMIC);
		if (IS_ERR(f6i))
			return PTR_ERR(f6i);
3373

3374 3375 3376
		/* ifp->rt can be accessed outside of rtnl */
		spin_lock(&ifp->lock);
		prev = ifp->rt;
3377
		ifp->rt = f6i;
3378 3379
		spin_unlock(&ifp->lock);

3380
		fib6_info_release(prev);
3381 3382 3383 3384
	}

	if (!(ifp->flags & IFA_F_NOPREFIXROUTE)) {
		addrconf_prefix_route(&ifp->addr, ifp->prefix_len,
3385 3386
				      ifp->rt_priority, idev->dev, 0, 0,
				      GFP_ATOMIC);
3387 3388
	}

3389 3390
	if (ifp->state == INET6_IFADDR_STATE_PREDAD)
		addrconf_dad_start(ifp);
3391 3392 3393 3394

	return 0;
}

3395
static void addrconf_permanent_addr(struct net *net, struct net_device *dev)
3396 3397 3398 3399 3400 3401 3402 3403 3404 3405 3406 3407
{
	struct inet6_ifaddr *ifp, *tmp;
	struct inet6_dev *idev;

	idev = __in6_dev_get(dev);
	if (!idev)
		return;

	write_lock_bh(&idev->lock);

	list_for_each_entry_safe(ifp, tmp, &idev->addr_list, if_list) {
		if ((ifp->flags & IFA_F_PERMANENT) &&
3408
		    fixup_permanent_addr(net, idev, ifp) < 0) {
3409
			write_unlock_bh(&idev->lock);
3410
			in6_ifa_hold(ifp);
3411 3412 3413 3414 3415 3416 3417 3418 3419 3420 3421
			ipv6_del_addr(ifp);
			write_lock_bh(&idev->lock);

			net_info_ratelimited("%s: Failed to add prefix route for address %pI6c; dropping\n",
					     idev->dev->name, &ifp->addr);
		}
	}

	write_unlock_bh(&idev->lock);
}

3422
static int addrconf_notify(struct notifier_block *this, unsigned long event,
3423
			   void *ptr)
L
Linus Torvalds 已提交
3424
{
3425
	struct net_device *dev = netdev_notifier_info_to_dev(ptr);
3426
	struct netdev_notifier_changeupper_info *info;
3427
	struct inet6_dev *idev = __in6_dev_get(dev);
3428
	struct net *net = dev_net(dev);
3429
	int run_pending = 0;
3430
	int err;
L
Linus Torvalds 已提交
3431

S
Stephen Hemminger 已提交
3432
	switch (event) {
3433
	case NETDEV_REGISTER:
3434
		if (!idev && dev->mtu >= IPV6_MIN_MTU) {
3435
			idev = ipv6_add_dev(dev);
3436 3437
			if (IS_ERR(idev))
				return notifier_from_errno(PTR_ERR(idev));
3438 3439
		}
		break;
S
Stephen Hemminger 已提交
3440

3441 3442 3443
	case NETDEV_CHANGEMTU:
		/* if MTU under IPV6_MIN_MTU stop IPv6 on this interface. */
		if (dev->mtu < IPV6_MIN_MTU) {
3444
			addrconf_ifdown(dev, dev != net->loopback_dev);
3445 3446 3447 3448 3449 3450 3451 3452 3453 3454 3455 3456 3457 3458 3459 3460 3461 3462 3463 3464 3465 3466
			break;
		}

		if (idev) {
			rt6_mtu_change(dev, dev->mtu);
			idev->cnf.mtu6 = dev->mtu;
			break;
		}

		/* allocate new idev */
		idev = ipv6_add_dev(dev);
		if (IS_ERR(idev))
			break;

		/* device is still not ready */
		if (!(idev->if_flags & IF_READY))
			break;

		run_pending = 1;

		/* fall through */

L
Linus Torvalds 已提交
3467
	case NETDEV_UP:
3468
	case NETDEV_CHANGE:
3469 3470 3471
		if (dev->flags & IFF_SLAVE)
			break;

3472 3473 3474
		if (idev && idev->cnf.disable_ipv6)
			break;

3475
		if (event == NETDEV_UP) {
3476
			/* restore routes for permanent addresses */
3477
			addrconf_permanent_addr(net, dev);
3478

3479
			if (!addrconf_link_ready(dev)) {
3480
				/* device is not ready yet. */
3481
				pr_info("ADDRCONF(NETDEV_UP): %s: link is not ready\n",
3482 3483 3484
					dev->name);
				break;
			}
3485

3486 3487 3488
			if (!idev && dev->mtu >= IPV6_MIN_MTU)
				idev = ipv6_add_dev(dev);

3489
			if (!IS_ERR_OR_NULL(idev)) {
3490
				idev->if_flags |= IF_READY;
3491 3492
				run_pending = 1;
			}
3493
		} else if (event == NETDEV_CHANGE) {
3494
			if (!addrconf_link_ready(dev)) {
3495
				/* device is still not ready. */
3496
				rt6_sync_down_dev(dev, event);
3497 3498 3499 3500
				break;
			}

			if (idev) {
3501 3502 3503 3504 3505 3506 3507
				if (idev->if_flags & IF_READY) {
					/* device is already configured -
					 * but resend MLD reports, we might
					 * have roamed and need to update
					 * multicast snooping switches
					 */
					ipv6_mc_up(idev);
3508
					rt6_sync_up(dev, RTNH_F_LINKDOWN);
3509
					break;
3510
				}
3511 3512 3513
				idev->if_flags |= IF_READY;
			}

3514 3515
			pr_info("ADDRCONF(NETDEV_CHANGE): %s: link becomes ready\n",
				dev->name);
3516

3517
			run_pending = 1;
3518 3519
		}

S
Stephen Hemminger 已提交
3520
		switch (dev->type) {
A
Amerigo Wang 已提交
3521
#if IS_ENABLED(CONFIG_IPV6_SIT)
L
Linus Torvalds 已提交
3522 3523 3524
		case ARPHRD_SIT:
			addrconf_sit_config(dev);
			break;
3525
#endif
A
Amerigo Wang 已提交
3526
#if IS_ENABLED(CONFIG_NET_IPGRE)
3527 3528 3529
		case ARPHRD_IPGRE:
			addrconf_gre_config(dev);
			break;
3530
#endif
L
Linus Torvalds 已提交
3531 3532 3533 3534 3535 3536 3537
		case ARPHRD_LOOPBACK:
			init_loopback(dev);
			break;

		default:
			addrconf_dev_config(dev);
			break;
3538
		}
S
Stephen Hemminger 已提交
3539

3540
		if (!IS_ERR_OR_NULL(idev)) {
3541 3542 3543
			if (run_pending)
				addrconf_dad_run(idev);

3544 3545 3546
			/* Device has an address by now */
			rt6_sync_up(dev, RTNH_F_DEAD);

S
Stephen Hemminger 已提交
3547 3548 3549 3550
			/*
			 * 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
3551
			 */
S
Stephen Hemminger 已提交
3552 3553
			if (idev->cnf.mtu6 != dev->mtu &&
			    dev->mtu >= IPV6_MIN_MTU) {
L
Linus Torvalds 已提交
3554 3555 3556 3557 3558
				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
3559 3560 3561 3562

			/*
			 * If the changed mtu during down is lower than
			 * IPV6_MIN_MTU stop IPv6 on this interface.
L
Linus Torvalds 已提交
3563 3564
			 */
			if (dev->mtu < IPV6_MIN_MTU)
3565
				addrconf_ifdown(dev, dev != net->loopback_dev);
L
Linus Torvalds 已提交
3566 3567 3568 3569 3570 3571 3572 3573 3574 3575
		}
		break;

	case NETDEV_DOWN:
	case NETDEV_UNREGISTER:
		/*
		 *	Remove all addresses from this interface.
		 */
		addrconf_ifdown(dev, event != NETDEV_DOWN);
		break;
3576

L
Linus Torvalds 已提交
3577 3578
	case NETDEV_CHANGENAME:
		if (idev) {
3579
			snmp6_unregister_dev(idev);
3580
			addrconf_sysctl_unregister(idev);
3581
			err = addrconf_sysctl_register(idev);
3582 3583
			if (err)
				return notifier_from_errno(err);
3584 3585 3586 3587 3588
			err = snmp6_register_dev(idev);
			if (err) {
				addrconf_sysctl_unregister(idev);
				return notifier_from_errno(err);
			}
3589
		}
L
Linus Torvalds 已提交
3590
		break;
S
Stephen Hemminger 已提交
3591

3592 3593
	case NETDEV_PRE_TYPE_CHANGE:
	case NETDEV_POST_TYPE_CHANGE:
3594 3595
		if (idev)
			addrconf_type_change(dev, event);
3596
		break;
3597 3598 3599 3600 3601 3602 3603 3604 3605

	case NETDEV_CHANGEUPPER:
		info = ptr;

		/* flush all routes if dev is linked to or unlinked from
		 * an L3 master device (e.g., VRF)
		 */
		if (info->upper_dev && netif_is_l3_master(info->upper_dev))
			addrconf_ifdown(dev, 0);
3606
	}
L
Linus Torvalds 已提交
3607 3608 3609 3610 3611 3612 3613 3614 3615

	return NOTIFY_OK;
}

/*
 *	addrconf module should be notified of a device going up
 */
static struct notifier_block ipv6_dev_notf = {
	.notifier_call = addrconf_notify,
3616
	.priority = ADDRCONF_NOTIFY_PRIORITY,
L
Linus Torvalds 已提交
3617 3618
};

3619
static void addrconf_type_change(struct net_device *dev, unsigned long event)
3620 3621 3622 3623 3624 3625
{
	struct inet6_dev *idev;
	ASSERT_RTNL();

	idev = __in6_dev_get(dev);

3626
	if (event == NETDEV_POST_TYPE_CHANGE)
3627
		ipv6_mc_remap(idev);
3628
	else if (event == NETDEV_PRE_TYPE_CHANGE)
3629 3630 3631
		ipv6_mc_unmap(idev);
}

3632 3633 3634 3635 3636 3637
static bool addr_is_local(const struct in6_addr *addr)
{
	return ipv6_addr_type(addr) &
		(IPV6_ADDR_LINKLOCAL | IPV6_ADDR_LOOPBACK);
}

L
Linus Torvalds 已提交
3638 3639
static int addrconf_ifdown(struct net_device *dev, int how)
{
3640
	unsigned long event = how ? NETDEV_UNREGISTER : NETDEV_DOWN;
3641
	struct net *net = dev_net(dev);
3642
	struct inet6_dev *idev;
3643
	struct inet6_ifaddr *ifa, *tmp;
3644
	bool keep_addr = false;
3645
	int state, i;
L
Linus Torvalds 已提交
3646 3647 3648

	ASSERT_RTNL();

3649
	rt6_disable_ip(dev, event);
L
Linus Torvalds 已提交
3650 3651

	idev = __in6_dev_get(dev);
3652
	if (!idev)
L
Linus Torvalds 已提交
3653 3654
		return -ENODEV;

S
Stephen Hemminger 已提交
3655 3656 3657
	/*
	 * Step 1: remove reference to ipv6 device from parent device.
	 *	   Do not dev_put!
L
Linus Torvalds 已提交
3658
	 */
3659
	if (how) {
L
Linus Torvalds 已提交
3660
		idev->dead = 1;
3661 3662

		/* protected by rtnl_lock */
3663
		RCU_INIT_POINTER(dev->ip6_ptr, NULL);
L
Linus Torvalds 已提交
3664 3665 3666 3667 3668 3669

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

	}

3670 3671 3672
	/* combine the user config with event to determine if permanent
	 * addresses are to be removed from address hash table
	 */
3673 3674 3675 3676 3677 3678 3679 3680 3681
	if (!how && !idev->cnf.disable_ipv6) {
		/* aggregate the system setting and interface setting */
		int _keep_addr = net->ipv6.devconf_all->keep_addr_on_down;

		if (!_keep_addr)
			_keep_addr = idev->cnf.keep_addr_on_down;

		keep_addr = (_keep_addr > 0);
	}
3682

3683 3684 3685 3686 3687
	/* 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);
3688
restart:
3689
		hlist_for_each_entry_rcu(ifa, h, addr_lst) {
3690
			if (ifa->idev == idev) {
3691
				addrconf_del_dad_work(ifa);
3692 3693 3694 3695
				/* combined flag + permanent flag decide if
				 * address is retained on a down event
				 */
				if (!keep_addr ||
3696 3697
				    !(ifa->flags & IFA_F_PERMANENT) ||
				    addr_is_local(&ifa->addr)) {
3698 3699 3700
					hlist_del_init_rcu(&ifa->addr_lst);
					goto restart;
				}
3701 3702 3703 3704 3705
			}
		}
		spin_unlock_bh(&addrconf_hash_lock);
	}

L
Linus Torvalds 已提交
3706 3707
	write_lock_bh(&idev->lock);

3708 3709
	addrconf_del_rs_timer(idev);

S
Stephen Hemminger 已提交
3710
	/* Step 2: clear flags for stateless addrconf */
3711
	if (!how)
3712
		idev->if_flags &= ~(IF_RS_SENT|IF_RA_RCVD|IF_READY);
L
Linus Torvalds 已提交
3713

S
Stephen Hemminger 已提交
3714
	/* Step 3: clear tempaddr list */
3715 3716 3717 3718
	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
3719 3720 3721 3722 3723 3724 3725 3726 3727 3728 3729
		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);
	}
3730

3731
	list_for_each_entry_safe(ifa, tmp, &idev->addr_list, if_list) {
3732
		struct fib6_info *rt = NULL;
3733
		bool keep;
3734

3735
		addrconf_del_dad_work(ifa);
3736

3737 3738 3739
		keep = keep_addr && (ifa->flags & IFA_F_PERMANENT) &&
			!addr_is_local(&ifa->addr);

3740
		write_unlock_bh(&idev->lock);
3741
		spin_lock_bh(&ifa->lock);
3742

3743
		if (keep) {
3744 3745
			/* set state to skip the notifier below */
			state = INET6_IFADDR_STATE_DEAD;
3746
			ifa->state = INET6_IFADDR_STATE_PREDAD;
3747 3748
			if (!(ifa->flags & IFA_F_NODAD))
				ifa->flags |= IFA_F_TENTATIVE;
3749 3750 3751

			rt = ifa->rt;
			ifa->rt = NULL;
3752 3753 3754 3755 3756
		} else {
			state = ifa->state;
			ifa->state = INET6_IFADDR_STATE_DEAD;
		}

3757
		spin_unlock_bh(&ifa->lock);
3758

3759
		if (rt)
3760
			ip6_del_rt(net, rt);
3761

3762 3763
		if (state != INET6_IFADDR_STATE_DEAD) {
			__ipv6_ifa_notify(RTM_DELADDR, ifa);
3764
			inet6addr_notifier_call_chain(NETDEV_DOWN, ifa);
3765 3766 3767 3768
		} else {
			if (idev->cnf.forwarding)
				addrconf_leave_anycast(ifa);
			addrconf_leave_solict(ifa->idev, &ifa->addr);
3769
		}
3770

3771
		write_lock_bh(&idev->lock);
3772 3773 3774 3775
		if (!keep) {
			list_del_rcu(&ifa->if_list);
			in6_ifa_put(ifa);
		}
3776
	}
3777

L
Linus Torvalds 已提交
3778 3779
	write_unlock_bh(&idev->lock);

3780 3781 3782
	/* Step 5: Discard anycast and multicast list */
	if (how) {
		ipv6_ac_destroy_dev(idev);
L
Linus Torvalds 已提交
3783
		ipv6_mc_destroy_dev(idev);
3784
	} else {
L
Linus Torvalds 已提交
3785
		ipv6_mc_down(idev);
3786
	}
L
Linus Torvalds 已提交
3787 3788

	idev->tstamp = jiffies;
3789

S
Stephen Hemminger 已提交
3790
	/* Last: Shot the device (if unregistered) */
3791
	if (how) {
3792
		addrconf_sysctl_unregister(idev);
L
Linus Torvalds 已提交
3793 3794 3795 3796 3797 3798 3799
		neigh_parms_release(&nd_tbl, idev->nd_parms);
		neigh_ifdown(&nd_tbl, dev);
		in6_dev_put(idev);
	}
	return 0;
}

3800
static void addrconf_rs_timer(struct timer_list *t)
L
Linus Torvalds 已提交
3801
{
3802
	struct inet6_dev *idev = from_timer(idev, t, rs_timer);
3803
	struct net_device *dev = idev->dev;
3804
	struct in6_addr lladdr;
L
Linus Torvalds 已提交
3805

3806
	write_lock(&idev->lock);
S
stephen hemminger 已提交
3807
	if (idev->dead || !(idev->if_flags & IF_READY))
L
Linus Torvalds 已提交
3808 3809
		goto out;

3810
	if (!ipv6_accept_ra(idev))
S
stephen hemminger 已提交
3811 3812 3813 3814
		goto out;

	/* Announcement received after solicitation was sent */
	if (idev->if_flags & IF_RA_RCVD)
L
Linus Torvalds 已提交
3815 3816
		goto out;

3817
	if (idev->rs_probes++ < idev->cnf.rtr_solicits || idev->cnf.rtr_solicits < 0) {
3818 3819 3820
		write_unlock(&idev->lock);
		if (!ipv6_get_lladdr(dev, &lladdr, IFA_F_TENTATIVE))
			ndisc_send_rs(dev, &lladdr,
3821 3822
				      &in6addr_linklocal_allrouters);
		else
3823
			goto put;
L
Linus Torvalds 已提交
3824

3825
		write_lock(&idev->lock);
3826 3827
		idev->rs_interval = rfc3315_s14_backoff_update(
			idev->rs_interval, idev->cnf.rtr_solicit_max_interval);
3828 3829 3830 3831
		/* 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 :
3832
				      idev->rs_interval);
L
Linus Torvalds 已提交
3833 3834 3835 3836 3837
	} else {
		/*
		 * Note: we do not support deprecated "all on-link"
		 * assumption any longer.
		 */
3838
		pr_debug("%s: no IPv6 routers present\n", idev->dev->name);
L
Linus Torvalds 已提交
3839 3840 3841
	}

out:
3842
	write_unlock(&idev->lock);
3843
put:
3844
	in6_dev_put(idev);
L
Linus Torvalds 已提交
3845 3846 3847 3848 3849
}

/*
 *	Duplicate Address Detection
 */
3850 3851 3852 3853
static void addrconf_dad_kick(struct inet6_ifaddr *ifp)
{
	unsigned long rand_num;
	struct inet6_dev *idev = ifp->idev;
3854
	u64 nonce;
3855

3856 3857 3858
	if (ifp->flags & IFA_F_OPTIMISTIC)
		rand_num = 0;
	else
3859
		rand_num = prandom_u32() % (idev->cnf.rtr_solicit_delay ? : 1);
3860

3861 3862 3863 3864 3865 3866 3867 3868
	nonce = 0;
	if (idev->cnf.enhanced_dad ||
	    dev_net(idev->dev)->ipv6.devconf_all->enhanced_dad) {
		do
			get_random_bytes(&nonce, 6);
		while (nonce == 0);
	}
	ifp->dad_nonce = nonce;
3869
	ifp->dad_probes = idev->cnf.dad_transmits;
3870
	addrconf_mod_dad_work(ifp, rand_num);
3871 3872
}

3873
static void addrconf_dad_begin(struct inet6_ifaddr *ifp)
L
Linus Torvalds 已提交
3874 3875 3876
{
	struct inet6_dev *idev = ifp->idev;
	struct net_device *dev = idev->dev;
3877
	bool bump_id, notify = false;
3878
	struct net *net;
L
Linus Torvalds 已提交
3879 3880 3881

	addrconf_join_solict(dev, &ifp->addr);

3882
	prandom_seed((__force u32) ifp->addr.s6_addr32[3]);
L
Linus Torvalds 已提交
3883 3884

	read_lock_bh(&idev->lock);
3885
	spin_lock(&ifp->lock);
3886
	if (ifp->state == INET6_IFADDR_STATE_DEAD)
L
Linus Torvalds 已提交
3887 3888
		goto out;

3889
	net = dev_net(dev);
L
Linus Torvalds 已提交
3890
	if (dev->flags&(IFF_NOARP|IFF_LOOPBACK) ||
3891
	    (net->ipv6.devconf_all->accept_dad < 1 &&
3892
	     idev->cnf.accept_dad < 1) ||
3893 3894
	    !(ifp->flags&IFA_F_TENTATIVE) ||
	    ifp->flags & IFA_F_NODAD) {
3895 3896 3897 3898 3899
		bool send_na = false;

		if (ifp->flags & IFA_F_TENTATIVE &&
		    !(ifp->flags & IFA_F_OPTIMISTIC))
			send_na = true;
3900
		bump_id = ifp->flags & IFA_F_TENTATIVE;
B
Brian Haley 已提交
3901
		ifp->flags &= ~(IFA_F_TENTATIVE|IFA_F_OPTIMISTIC|IFA_F_DADFAILED);
3902
		spin_unlock(&ifp->lock);
L
Linus Torvalds 已提交
3903 3904
		read_unlock_bh(&idev->lock);

3905
		addrconf_dad_completed(ifp, bump_id, send_na);
L
Linus Torvalds 已提交
3906 3907 3908
		return;
	}

3909
	if (!(idev->if_flags & IF_READY)) {
3910
		spin_unlock(&ifp->lock);
3911
		read_unlock_bh(&idev->lock);
3912
		/*
3913
		 * If the device is not ready:
3914 3915 3916 3917
		 * - keep it tentative if it is a permanent address.
		 * - otherwise, kill it.
		 */
		in6_ifa_hold(ifp);
B
Brian Haley 已提交
3918
		addrconf_dad_stop(ifp, 0);
3919
		return;
3920
	}
3921 3922 3923 3924 3925

	/*
	 * Optimistic nodes can start receiving
	 * Frames right away
	 */
3926
	if (ifp->flags & IFA_F_OPTIMISTIC) {
3927 3928
		ip6_ins_rt(net, ifp->rt);
		if (ipv6_use_optimistic_addr(net, idev)) {
3929 3930 3931
			/* Because optimistic nodes can use this address,
			 * notify listeners. If DAD fails, RTM_DELADDR is sent.
			 */
3932
			notify = true;
3933 3934
		}
	}
3935

3936
	addrconf_dad_kick(ifp);
L
Linus Torvalds 已提交
3937
out:
3938
	spin_unlock(&ifp->lock);
L
Linus Torvalds 已提交
3939
	read_unlock_bh(&idev->lock);
3940 3941
	if (notify)
		ipv6_ifa_notify(RTM_NEWADDR, ifp);
L
Linus Torvalds 已提交
3942 3943
}

3944
static void addrconf_dad_start(struct inet6_ifaddr *ifp)
L
Linus Torvalds 已提交
3945
{
3946 3947
	bool begin_dad = false;

3948
	spin_lock_bh(&ifp->lock);
3949 3950 3951 3952
	if (ifp->state != INET6_IFADDR_STATE_DEAD) {
		ifp->state = INET6_IFADDR_STATE_PREDAD;
		begin_dad = true;
	}
3953
	spin_unlock_bh(&ifp->lock);
3954 3955 3956 3957 3958 3959 3960 3961 3962 3963

	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
3964
	struct inet6_dev *idev = ifp->idev;
3965
	bool bump_id, disable_ipv6 = false;
L
Linus Torvalds 已提交
3966 3967
	struct in6_addr mcaddr;

3968 3969 3970 3971 3972 3973 3974 3975
	enum {
		DAD_PROCESS,
		DAD_BEGIN,
		DAD_ABORT,
	} action = DAD_PROCESS;

	rtnl_lock();

3976
	spin_lock_bh(&ifp->lock);
3977 3978 3979 3980 3981 3982
	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;
3983

3984 3985 3986
		if ((dev_net(idev->dev)->ipv6.devconf_all->accept_dad > 1 ||
		     idev->cnf.accept_dad > 1) &&
		    !idev->cnf.disable_ipv6 &&
3987 3988 3989 3990 3991 3992 3993 3994 3995 3996 3997 3998 3999 4000 4001 4002
		    !(ifp->flags & IFA_F_STABLE_PRIVACY)) {
			struct in6_addr addr;

			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 */
				idev->cnf.disable_ipv6 = 1;

				pr_info("%s: IPv6 being disabled!\n",
					ifp->idev->dev->name);
				disable_ipv6 = true;
			}
		}
4003
	}
4004
	spin_unlock_bh(&ifp->lock);
4005 4006 4007 4008 4009

	if (action == DAD_BEGIN) {
		addrconf_dad_begin(ifp);
		goto out;
	} else if (action == DAD_ABORT) {
4010
		in6_ifa_hold(ifp);
4011
		addrconf_dad_stop(ifp, 1);
4012 4013
		if (disable_ipv6)
			addrconf_ifdown(idev->dev, 0);
4014 4015 4016
		goto out;
	}

4017
	if (!ifp->dad_probes && addrconf_dad_end(ifp))
H
Herbert Xu 已提交
4018 4019
		goto out;

4020
	write_lock_bh(&idev->lock);
4021
	if (idev->dead || !(idev->if_flags & IF_READY)) {
4022
		write_unlock_bh(&idev->lock);
L
Linus Torvalds 已提交
4023 4024
		goto out;
	}
4025 4026

	spin_lock(&ifp->lock);
4027 4028
	if (ifp->state == INET6_IFADDR_STATE_DEAD) {
		spin_unlock(&ifp->lock);
4029
		write_unlock_bh(&idev->lock);
4030 4031 4032
		goto out;
	}

4033
	if (ifp->dad_probes == 0) {
4034 4035
		bool send_na = false;

L
Linus Torvalds 已提交
4036 4037 4038 4039
		/*
		 * DAD was successful
		 */

4040 4041 4042
		if (ifp->flags & IFA_F_TENTATIVE &&
		    !(ifp->flags & IFA_F_OPTIMISTIC))
			send_na = true;
4043
		bump_id = ifp->flags & IFA_F_TENTATIVE;
B
Brian Haley 已提交
4044
		ifp->flags &= ~(IFA_F_TENTATIVE|IFA_F_OPTIMISTIC|IFA_F_DADFAILED);
4045
		spin_unlock(&ifp->lock);
4046
		write_unlock_bh(&idev->lock);
L
Linus Torvalds 已提交
4047

4048
		addrconf_dad_completed(ifp, bump_id, send_na);
L
Linus Torvalds 已提交
4049 4050 4051 4052

		goto out;
	}

4053
	ifp->dad_probes--;
4054 4055
	addrconf_mod_dad_work(ifp,
			      NEIGH_VAR(ifp->idev->nd_parms, RETRANS_TIME));
4056
	spin_unlock(&ifp->lock);
4057
	write_unlock_bh(&idev->lock);
L
Linus Torvalds 已提交
4058 4059 4060

	/* send a neighbour solicitation for our addr */
	addrconf_addr_solict_mult(&ifp->addr, &mcaddr);
4061 4062
	ndisc_send_ns(ifp->idev->dev, &ifp->addr, &mcaddr, &in6addr_any,
		      ifp->dad_nonce);
L
Linus Torvalds 已提交
4063 4064
out:
	in6_ifa_put(ifp);
4065
	rtnl_unlock();
L
Linus Torvalds 已提交
4066 4067
}

4068 4069 4070 4071 4072 4073
/* 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;

4074 4075 4076
	list_for_each_entry_reverse(ifpiter, &idev->addr_list, if_list) {
		if (ifpiter->scope > IFA_LINK)
			break;
4077 4078 4079 4080 4081 4082 4083 4084 4085
		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;
}

4086 4087
static void addrconf_dad_completed(struct inet6_ifaddr *ifp, bool bump_id,
				   bool send_na)
L
Linus Torvalds 已提交
4088
{
S
Stephen Hemminger 已提交
4089
	struct net_device *dev = ifp->idev->dev;
4090
	struct in6_addr lladdr;
4091
	bool send_rs, send_mld;
4092

4093
	addrconf_del_dad_work(ifp);
L
Linus Torvalds 已提交
4094 4095 4096 4097 4098 4099 4100

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

	ipv6_ifa_notify(RTM_NEWADDR, ifp);

4101 4102
	/* If added prefix is link local and we are prepared to process
	   router advertisements, start sending router solicitations.
L
Linus Torvalds 已提交
4103 4104
	 */

4105
	read_lock_bh(&ifp->idev->lock);
4106
	send_mld = ifp->scope == IFA_LINK && ipv6_lonely_lladdr(ifp);
4107 4108
	send_rs = send_mld &&
		  ipv6_accept_ra(ifp->idev) &&
4109
		  ifp->idev->cnf.rtr_solicits != 0 &&
4110
		  (dev->flags&IFF_LOOPBACK) == 0;
4111 4112
	read_unlock_bh(&ifp->idev->lock);

4113 4114 4115 4116 4117 4118
	/* 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);

4119 4120 4121 4122 4123 4124 4125 4126 4127 4128
	/* send unsolicited NA if enabled */
	if (send_na &&
	    (ifp->idev->cnf.ndisc_notify ||
	     dev_net(dev)->ipv6.devconf_all->ndisc_notify)) {
		ndisc_send_na(dev, &in6addr_linklocal_allnodes, &ifp->addr,
			      /*router=*/ !!ifp->idev->cnf.forwarding,
			      /*solicited=*/ false, /*override=*/ true,
			      /*inc_opt=*/ true);
	}

4129
	if (send_rs) {
L
Linus Torvalds 已提交
4130 4131 4132 4133 4134
		/*
		 *	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
		 */
4135
		if (ipv6_get_lladdr(dev, &lladdr, IFA_F_TENTATIVE))
4136
			return;
4137
		ndisc_send_rs(dev, &lladdr, &in6addr_linklocal_allrouters);
L
Linus Torvalds 已提交
4138

4139 4140
		write_lock_bh(&ifp->idev->lock);
		spin_lock(&ifp->lock);
4141 4142
		ifp->idev->rs_interval = rfc3315_s14_backoff_init(
			ifp->idev->cnf.rtr_solicit_interval);
4143
		ifp->idev->rs_probes = 1;
L
Linus Torvalds 已提交
4144
		ifp->idev->if_flags |= IF_RS_SENT;
4145
		addrconf_mod_rs_timer(ifp->idev, ifp->idev->rs_interval);
4146 4147
		spin_unlock(&ifp->lock);
		write_unlock_bh(&ifp->idev->lock);
L
Linus Torvalds 已提交
4148
	}
4149 4150 4151

	if (bump_id)
		rt_genid_bump_ipv6(dev_net(dev));
4152 4153 4154 4155 4156 4157

	/* Make sure that a new temporary address will be created
	 * before this temporary address becomes deprecated.
	 */
	if (ifp->flags & IFA_F_TEMPORARY)
		addrconf_verify_rtnl();
L
Linus Torvalds 已提交
4158 4159
}

S
Stephen Hemminger 已提交
4160 4161
static void addrconf_dad_run(struct inet6_dev *idev)
{
4162 4163 4164
	struct inet6_ifaddr *ifp;

	read_lock_bh(&idev->lock);
4165
	list_for_each_entry(ifp, &idev->addr_list, if_list) {
4166
		spin_lock(&ifp->lock);
H
Herbert Xu 已提交
4167 4168 4169
		if (ifp->flags & IFA_F_TENTATIVE &&
		    ifp->state == INET6_IFADDR_STATE_DAD)
			addrconf_dad_kick(ifp);
4170
		spin_unlock(&ifp->lock);
4171 4172 4173 4174
	}
	read_unlock_bh(&idev->lock);
}

L
Linus Torvalds 已提交
4175 4176
#ifdef CONFIG_PROC_FS
struct if6_iter_state {
4177
	struct seq_net_private p;
L
Linus Torvalds 已提交
4178
	int bucket;
4179
	int offset;
L
Linus Torvalds 已提交
4180 4181
};

4182
static struct inet6_ifaddr *if6_get_first(struct seq_file *seq, loff_t pos)
L
Linus Torvalds 已提交
4183 4184
{
	struct if6_iter_state *state = seq->private;
4185
	struct net *net = seq_file_net(seq);
4186
	struct inet6_ifaddr *ifa = NULL;
4187
	int p = 0;
L
Linus Torvalds 已提交
4188

4189 4190 4191 4192 4193 4194 4195
	/* initial bucket if pos is 0 */
	if (pos == 0) {
		state->bucket = 0;
		state->offset = 0;
	}

	for (; state->bucket < IN6_ADDR_HSIZE; ++state->bucket) {
4196
		hlist_for_each_entry_rcu(ifa, &inet6_addr_lst[state->bucket],
4197
					 addr_lst) {
4198 4199
			if (!net_eq(dev_net(ifa->idev->dev), net))
				continue;
4200 4201 4202 4203 4204 4205
			/* sync with offset */
			if (p < state->offset) {
				p++;
				continue;
			}
			state->offset++;
4206
			return ifa;
4207 4208 4209 4210 4211
		}

		/* prepare for next bucket */
		state->offset = 0;
		p = 0;
L
Linus Torvalds 已提交
4212
	}
4213
	return NULL;
L
Linus Torvalds 已提交
4214 4215
}

4216 4217
static struct inet6_ifaddr *if6_get_next(struct seq_file *seq,
					 struct inet6_ifaddr *ifa)
L
Linus Torvalds 已提交
4218 4219
{
	struct if6_iter_state *state = seq->private;
4220
	struct net *net = seq_file_net(seq);
L
Linus Torvalds 已提交
4221

4222
	hlist_for_each_entry_continue_rcu(ifa, addr_lst) {
4223 4224
		if (!net_eq(dev_net(ifa->idev->dev), net))
			continue;
4225
		state->offset++;
4226
		return ifa;
4227
	}
4228

4229
	while (++state->bucket < IN6_ADDR_HSIZE) {
4230
		state->offset = 0;
4231
		hlist_for_each_entry_rcu(ifa,
4232
				     &inet6_addr_lst[state->bucket], addr_lst) {
4233 4234
			if (!net_eq(dev_net(ifa->idev->dev), net))
				continue;
4235
			state->offset++;
4236
			return ifa;
4237
		}
L
Linus Torvalds 已提交
4238
	}
4239

4240
	return NULL;
L
Linus Torvalds 已提交
4241 4242 4243
}

static void *if6_seq_start(struct seq_file *seq, loff_t *pos)
4244
	__acquires(rcu)
L
Linus Torvalds 已提交
4245
{
4246
	rcu_read_lock();
4247
	return if6_get_first(seq, *pos);
L
Linus Torvalds 已提交
4248 4249 4250 4251 4252 4253 4254 4255 4256 4257 4258 4259
}

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)
4260
	__releases(rcu)
L
Linus Torvalds 已提交
4261
{
4262
	rcu_read_unlock();
L
Linus Torvalds 已提交
4263 4264 4265 4266 4267
}

static int if6_seq_show(struct seq_file *seq, void *v)
{
	struct inet6_ifaddr *ifp = (struct inet6_ifaddr *)v;
4268
	seq_printf(seq, "%pi6 %02x %02x %02x %02x %8s\n",
4269
		   &ifp->addr,
L
Linus Torvalds 已提交
4270 4271 4272
		   ifp->idev->dev->ifindex,
		   ifp->prefix_len,
		   ifp->scope,
4273
		   (u8) ifp->flags,
L
Linus Torvalds 已提交
4274 4275 4276 4277
		   ifp->idev->dev->name);
	return 0;
}

4278
static const struct seq_operations if6_seq_ops = {
L
Linus Torvalds 已提交
4279 4280 4281 4282 4283 4284
	.start	= if6_seq_start,
	.next	= if6_seq_next,
	.show	= if6_seq_show,
	.stop	= if6_seq_stop,
};

4285
static int __net_init if6_proc_net_init(struct net *net)
L
Linus Torvalds 已提交
4286
{
4287 4288
	if (!proc_create_net("if_inet6", 0444, net->proc_net, &if6_seq_ops,
			sizeof(struct if6_iter_state)))
L
Linus Torvalds 已提交
4289 4290 4291 4292
		return -ENOMEM;
	return 0;
}

4293
static void __net_exit if6_proc_net_exit(struct net *net)
4294
{
4295
	remove_proc_entry("if_inet6", net->proc_net);
4296 4297 4298
}

static struct pernet_operations if6_proc_net_ops = {
4299 4300
	.init = if6_proc_net_init,
	.exit = if6_proc_net_exit,
4301 4302 4303 4304 4305 4306 4307
};

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

L
Linus Torvalds 已提交
4308 4309
void if6_proc_exit(void)
{
4310
	unregister_pernet_subsys(&if6_proc_net_ops);
L
Linus Torvalds 已提交
4311 4312 4313
}
#endif	/* CONFIG_PROC_FS */

A
Amerigo Wang 已提交
4314
#if IS_ENABLED(CONFIG_IPV6_MIP6)
4315
/* Check if address is a home address configured on any interface. */
4316
int ipv6_chk_home_addr(struct net *net, const struct in6_addr *addr)
4317
{
4318
	unsigned int hash = inet6_addr_hash(net, addr);
4319
	struct inet6_ifaddr *ifp = NULL;
4320
	int ret = 0;
4321

4322 4323
	rcu_read_lock();
	hlist_for_each_entry_rcu(ifp, &inet6_addr_lst[hash], addr_lst) {
4324
		if (!net_eq(dev_net(ifp->idev->dev), net))
4325
			continue;
4326
		if (ipv6_addr_equal(&ifp->addr, addr) &&
4327 4328 4329 4330 4331
		    (ifp->flags & IFA_F_HOMEADDRESS)) {
			ret = 1;
			break;
		}
	}
4332
	rcu_read_unlock();
4333 4334 4335 4336
	return ret;
}
#endif

L
Linus Torvalds 已提交
4337 4338 4339 4340
/*
 *	Periodic address status verification
 */

4341
static void addrconf_verify_rtnl(void)
L
Linus Torvalds 已提交
4342
{
4343
	unsigned long now, next, next_sec, next_sched;
L
Linus Torvalds 已提交
4344 4345 4346
	struct inet6_ifaddr *ifp;
	int i;

4347 4348
	ASSERT_RTNL();

4349
	rcu_read_lock_bh();
L
Linus Torvalds 已提交
4350
	now = jiffies;
4351
	next = round_jiffies_up(now + ADDR_CHECK_FREQUENCY);
L
Linus Torvalds 已提交
4352

4353
	cancel_delayed_work(&addr_chk_work);
L
Linus Torvalds 已提交
4354

S
Stephen Hemminger 已提交
4355
	for (i = 0; i < IN6_ADDR_HSIZE; i++) {
L
Linus Torvalds 已提交
4356
restart:
4357
		hlist_for_each_entry_rcu_bh(ifp, &inet6_addr_lst[i], addr_lst) {
L
Linus Torvalds 已提交
4358 4359
			unsigned long age;

4360 4361 4362 4363 4364 4365
			/* 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
4366 4367 4368
				continue;

			spin_lock(&ifp->lock);
4369 4370
			/* We try to batch several events at once. */
			age = (now - ifp->tstamp + ADDRCONF_TIMER_FUZZ_MINUS) / HZ;
L
Linus Torvalds 已提交
4371

4372 4373
			if (ifp->valid_lft != INFINITY_LIFE_TIME &&
			    age >= ifp->valid_lft) {
L
Linus Torvalds 已提交
4374 4375 4376 4377
				spin_unlock(&ifp->lock);
				in6_ifa_hold(ifp);
				ipv6_del_addr(ifp);
				goto restart;
4378 4379 4380
			} else if (ifp->prefered_lft == INFINITY_LIFE_TIME) {
				spin_unlock(&ifp->lock);
				continue;
L
Linus Torvalds 已提交
4381
			} else if (age >= ifp->prefered_lft) {
4382
				/* jiffies - ifp->tstamp > age >= ifp->prefered_lft */
L
Linus Torvalds 已提交
4383 4384 4385 4386 4387 4388 4389
				int deprecate = 0;

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

4390 4391
				if ((ifp->valid_lft != INFINITY_LIFE_TIME) &&
				    (time_before(ifp->tstamp + ifp->valid_lft * HZ, next)))
L
Linus Torvalds 已提交
4392 4393 4394 4395 4396 4397 4398 4399 4400 4401 4402 4403 4404
					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)) {
4405 4406
				unsigned long regen_advance = ifp->idev->cnf.regen_max_retry *
					ifp->idev->cnf.dad_transmits *
J
Jiri Pirko 已提交
4407
					NEIGH_VAR(ifp->idev->nd_parms, RETRANS_TIME) / HZ;
4408

L
Linus Torvalds 已提交
4409 4410 4411 4412 4413 4414 4415 4416 4417
				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);
4418

4419 4420 4421
						spin_lock(&ifpub->lock);
						ifpub->regen_count = 0;
						spin_unlock(&ifpub->lock);
4422
						rcu_read_unlock_bh();
E
Eric Dumazet 已提交
4423
						ipv6_create_tempaddr(ifpub, ifp, true);
L
Linus Torvalds 已提交
4424 4425
						in6_ifa_put(ifpub);
						in6_ifa_put(ifp);
4426
						rcu_read_lock_bh();
L
Linus Torvalds 已提交
4427 4428 4429 4430 4431 4432 4433 4434 4435 4436 4437 4438 4439 4440
						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);
			}
		}
	}

4441 4442 4443 4444 4445 4446 4447 4448 4449 4450 4451
	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;

4452 4453
	pr_debug("now = %lu, schedule = %lu, rounded schedule = %lu => %lu\n",
		 now, next, next_sec, next_sched);
4454
	mod_delayed_work(addrconf_wq, &addr_chk_work, next_sched - now);
4455
	rcu_read_unlock_bh();
L
Linus Torvalds 已提交
4456 4457
}

4458 4459 4460 4461 4462 4463 4464 4465 4466 4467 4468 4469
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);
}

4470 4471
static struct in6_addr *extract_addr(struct nlattr *addr, struct nlattr *local,
				     struct in6_addr **peer_pfx)
4472 4473 4474
{
	struct in6_addr *pfx = NULL;

4475 4476
	*peer_pfx = NULL;

4477 4478 4479 4480 4481
	if (addr)
		pfx = nla_data(addr);

	if (local) {
		if (pfx && nla_memcmp(local, pfx, sizeof(*pfx)))
4482 4483
			*peer_pfx = pfx;
		pfx = nla_data(local);
4484 4485 4486 4487 4488
	}

	return pfx;
}

4489
static const struct nla_policy ifa_ipv6_policy[IFA_MAX+1] = {
4490 4491 4492
	[IFA_ADDRESS]		= { .len = sizeof(struct in6_addr) },
	[IFA_LOCAL]		= { .len = sizeof(struct in6_addr) },
	[IFA_CACHEINFO]		= { .len = sizeof(struct ifa_cacheinfo) },
4493
	[IFA_FLAGS]		= { .len = sizeof(u32) },
4494
	[IFA_RT_PRIORITY]	= { .len = sizeof(u32) },
4495 4496
};

L
Linus Torvalds 已提交
4497
static int
4498 4499
inet6_rtm_deladdr(struct sk_buff *skb, struct nlmsghdr *nlh,
		  struct netlink_ext_ack *extack)
L
Linus Torvalds 已提交
4500
{
4501
	struct net *net = sock_net(skb->sk);
4502 4503
	struct ifaddrmsg *ifm;
	struct nlattr *tb[IFA_MAX+1];
4504
	struct in6_addr *pfx, *peer_pfx;
4505
	u32 ifa_flags;
4506
	int err;
L
Linus Torvalds 已提交
4507

4508
	err = nlmsg_parse(nlh, sizeof(*ifm), tb, IFA_MAX, ifa_ipv6_policy,
4509
			  extack);
4510 4511 4512 4513
	if (err < 0)
		return err;

	ifm = nlmsg_data(nlh);
4514
	pfx = extract_addr(tb[IFA_ADDRESS], tb[IFA_LOCAL], &peer_pfx);
4515
	if (!pfx)
L
Linus Torvalds 已提交
4516 4517
		return -EINVAL;

4518 4519 4520 4521 4522 4523 4524
	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
4525 4526
}

4527 4528 4529 4530
static int modify_prefix_route(struct inet6_ifaddr *ifp,
			       unsigned long expires, u32 flags)
{
	struct fib6_info *f6i;
4531
	u32 prio;
4532 4533 4534 4535 4536 4537 4538 4539

	f6i = addrconf_get_prefix_route(&ifp->addr,
					ifp->prefix_len,
					ifp->idev->dev,
					0, RTF_GATEWAY | RTF_DEFAULT);
	if (!f6i)
		return -ENOENT;

4540 4541 4542 4543 4544
	prio = ifp->rt_priority ? : IP6_RT_PRIO_ADDRCONF;
	if (f6i->fib6_metric != prio) {
		/* delete old one */
		ip6_del_rt(dev_net(ifp->idev->dev), f6i);

4545 4546 4547 4548 4549 4550 4551 4552 4553 4554 4555 4556 4557 4558 4559 4560
		/* add new one */
		addrconf_prefix_route(&ifp->addr, ifp->prefix_len,
				      ifp->rt_priority, ifp->idev->dev,
				      expires, flags, GFP_KERNEL);
	} else {
		if (!expires)
			fib6_clean_expires(f6i);
		else
			fib6_set_expires(f6i, expires);

		fib6_info_release(f6i);
	}

	return 0;
}

4561
static int inet6_addr_modify(struct inet6_ifaddr *ifp, struct ifa6_config *cfg)
4562
{
4563 4564
	u32 flags;
	clock_t expires;
4565
	unsigned long timeout;
4566
	bool was_managetempaddr;
4567
	bool had_prefixroute;
4568

4569 4570
	ASSERT_RTNL();

4571
	if (!cfg->valid_lft || cfg->preferred_lft > cfg->valid_lft)
4572 4573
		return -EINVAL;

4574
	if (cfg->ifa_flags & IFA_F_MANAGETEMPADDR &&
4575 4576 4577
	    (ifp->flags & IFA_F_TEMPORARY || ifp->prefix_len != 64))
		return -EINVAL;

4578
	if (!(ifp->flags & IFA_F_TENTATIVE) || ifp->flags & IFA_F_DADFAILED)
4579
		cfg->ifa_flags &= ~IFA_F_OPTIMISTIC;
4580

4581
	timeout = addrconf_timeout_fixup(cfg->valid_lft, HZ);
4582 4583
	if (addrconf_finite_timeout(timeout)) {
		expires = jiffies_to_clock_t(timeout * HZ);
4584
		cfg->valid_lft = timeout;
4585
		flags = RTF_EXPIRES;
4586 4587 4588
	} else {
		expires = 0;
		flags = 0;
4589
		cfg->ifa_flags |= IFA_F_PERMANENT;
4590
	}
4591

4592
	timeout = addrconf_timeout_fixup(cfg->preferred_lft, HZ);
4593 4594
	if (addrconf_finite_timeout(timeout)) {
		if (timeout == 0)
4595 4596
			cfg->ifa_flags |= IFA_F_DEPRECATED;
		cfg->preferred_lft = timeout;
4597
	}
4598 4599

	spin_lock_bh(&ifp->lock);
4600
	was_managetempaddr = ifp->flags & IFA_F_MANAGETEMPADDR;
4601 4602
	had_prefixroute = ifp->flags & IFA_F_PERMANENT &&
			  !(ifp->flags & IFA_F_NOPREFIXROUTE);
4603
	ifp->flags &= ~(IFA_F_DEPRECATED | IFA_F_PERMANENT | IFA_F_NODAD |
4604 4605
			IFA_F_HOMEADDRESS | IFA_F_MANAGETEMPADDR |
			IFA_F_NOPREFIXROUTE);
4606
	ifp->flags |= cfg->ifa_flags;
4607
	ifp->tstamp = jiffies;
4608 4609
	ifp->valid_lft = cfg->valid_lft;
	ifp->prefered_lft = cfg->preferred_lft;
4610

4611 4612
	if (cfg->rt_priority && cfg->rt_priority != ifp->rt_priority)
		ifp->rt_priority = cfg->rt_priority;
4613 4614 4615 4616 4617

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

4618
	if (!(cfg->ifa_flags & IFA_F_NOPREFIXROUTE)) {
4619 4620 4621 4622 4623 4624 4625 4626 4627 4628 4629
		int rc = -ENOENT;

		if (had_prefixroute)
			rc = modify_prefix_route(ifp, expires, flags);

		/* prefix route could have been deleted; if so restore it */
		if (rc == -ENOENT) {
			addrconf_prefix_route(&ifp->addr, ifp->prefix_len,
					      ifp->rt_priority, ifp->idev->dev,
					      expires, flags, GFP_KERNEL);
		}
4630 4631 4632 4633 4634 4635 4636 4637 4638 4639 4640 4641
	} 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);
		}
4642
	}
4643 4644

	if (was_managetempaddr || ifp->flags & IFA_F_MANAGETEMPADDR) {
4645 4646 4647 4648 4649 4650 4651 4652
		if (was_managetempaddr &&
		    !(ifp->flags & IFA_F_MANAGETEMPADDR)) {
			cfg->valid_lft = 0;
			cfg->preferred_lft = 0;
		}
		manage_tempaddrs(ifp->idev, ifp, cfg->valid_lft,
				 cfg->preferred_lft, !was_managetempaddr,
				 jiffies);
4653 4654
	}

4655
	addrconf_verify_rtnl();
4656 4657 4658 4659

	return 0;
}

L
Linus Torvalds 已提交
4660
static int
4661 4662
inet6_rtm_newaddr(struct sk_buff *skb, struct nlmsghdr *nlh,
		  struct netlink_ext_ack *extack)
L
Linus Torvalds 已提交
4663
{
4664
	struct net *net = sock_net(skb->sk);
4665 4666
	struct ifaddrmsg *ifm;
	struct nlattr *tb[IFA_MAX+1];
4667
	struct in6_addr *peer_pfx;
4668 4669
	struct inet6_ifaddr *ifa;
	struct net_device *dev;
4670
	struct inet6_dev *idev;
4671
	struct ifa6_config cfg;
4672
	int err;
L
Linus Torvalds 已提交
4673

4674
	err = nlmsg_parse(nlh, sizeof(*ifm), tb, IFA_MAX, ifa_ipv6_policy,
4675
			  extack);
4676 4677 4678
	if (err < 0)
		return err;

4679 4680
	memset(&cfg, 0, sizeof(cfg));

4681
	ifm = nlmsg_data(nlh);
4682 4683
	cfg.pfx = extract_addr(tb[IFA_ADDRESS], tb[IFA_LOCAL], &peer_pfx);
	if (!cfg.pfx)
L
Linus Torvalds 已提交
4684 4685
		return -EINVAL;

4686 4687
	cfg.peer_pfx = peer_pfx;
	cfg.plen = ifm->ifa_prefixlen;
4688 4689 4690
	if (tb[IFA_RT_PRIORITY])
		cfg.rt_priority = nla_get_u32(tb[IFA_RT_PRIORITY]);

4691 4692 4693
	cfg.valid_lft = INFINITY_LIFE_TIME;
	cfg.preferred_lft = INFINITY_LIFE_TIME;

4694
	if (tb[IFA_CACHEINFO]) {
4695
		struct ifa_cacheinfo *ci;
4696 4697

		ci = nla_data(tb[IFA_CACHEINFO]);
4698 4699
		cfg.valid_lft = ci->ifa_valid;
		cfg.preferred_lft = ci->ifa_prefered;
4700 4701
	}

4702
	dev =  __dev_get_by_index(net, ifm->ifa_index);
4703
	if (!dev)
4704 4705
		return -ENODEV;

4706 4707 4708 4709
	if (tb[IFA_FLAGS])
		cfg.ifa_flags = nla_get_u32(tb[IFA_FLAGS]);
	else
		cfg.ifa_flags = ifm->ifa_flags;
4710

4711
	/* We ignore other flags so far. */
4712 4713 4714
	cfg.ifa_flags &= IFA_F_NODAD | IFA_F_HOMEADDRESS |
			 IFA_F_MANAGETEMPADDR | IFA_F_NOPREFIXROUTE |
			 IFA_F_MCAUTOJOIN | IFA_F_OPTIMISTIC;
4715 4716 4717 4718 4719 4720

	idev = ipv6_find_idev(dev);
	if (IS_ERR(idev))
		return PTR_ERR(idev);

	if (!ipv6_allow_optimistic_dad(net, idev))
4721
		cfg.ifa_flags &= ~IFA_F_OPTIMISTIC;
4722

4723 4724
	if (cfg.ifa_flags & IFA_F_NODAD &&
	    cfg.ifa_flags & IFA_F_OPTIMISTIC) {
4725 4726 4727
		NL_SET_ERR_MSG(extack, "IFA_F_NODAD and IFA_F_OPTIMISTIC are mutually exclusive");
		return -EINVAL;
	}
4728

4729
	ifa = ipv6_get_ifaddr(net, cfg.pfx, dev, 1);
4730
	if (!ifa) {
4731 4732
		/*
		 * It would be best to check for !NLM_F_CREATE here but
4733
		 * userspace already relies on not having to provide this.
4734
		 */
4735
		return inet6_addr_add(net, ifm->ifa_index, &cfg, extack);
4736 4737
	}

4738 4739 4740 4741
	if (nlh->nlmsg_flags & NLM_F_EXCL ||
	    !(nlh->nlmsg_flags & NLM_F_REPLACE))
		err = -EEXIST;
	else
4742
		err = inet6_addr_modify(ifa, &cfg);
4743 4744 4745 4746

	in6_ifa_put(ifa);

	return err;
L
Linus Torvalds 已提交
4747 4748
}

4749
static void put_ifaddrmsg(struct nlmsghdr *nlh, u8 prefixlen, u32 flags,
4750 4751 4752 4753 4754 4755 4756 4757 4758 4759 4760 4761
			  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;
}

4762 4763 4764 4765 4766
static int put_cacheinfo(struct sk_buff *skb, unsigned long cstamp,
			 unsigned long tstamp, u32 preferred, u32 valid)
{
	struct ifa_cacheinfo ci;

4767 4768
	ci.cstamp = cstamp_delta(cstamp);
	ci.tstamp = cstamp_delta(tstamp);
4769 4770 4771 4772 4773 4774
	ci.ifa_prefered = preferred;
	ci.ifa_valid = valid;

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

4775 4776 4777 4778 4779 4780 4781 4782 4783 4784 4785 4786
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;
}

4787 4788
static inline int inet6_ifaddr_msgsize(void)
{
4789
	return NLMSG_ALIGN(sizeof(struct ifaddrmsg))
4790
	       + nla_total_size(16) /* IFA_LOCAL */
4791
	       + nla_total_size(16) /* IFA_ADDRESS */
4792
	       + nla_total_size(sizeof(struct ifa_cacheinfo))
4793 4794
	       + nla_total_size(4)  /* IFA_FLAGS */
	       + nla_total_size(4)  /* IFA_RT_PRIORITY */;
4795
}
4796

L
Linus Torvalds 已提交
4797
static int inet6_fill_ifaddr(struct sk_buff *skb, struct inet6_ifaddr *ifa,
4798
			     u32 portid, u32 seq, int event, unsigned int flags)
L
Linus Torvalds 已提交
4799 4800
{
	struct nlmsghdr  *nlh;
4801
	u32 preferred, valid;
L
Linus Torvalds 已提交
4802

4803
	nlh = nlmsg_put(skb, portid, seq, event, sizeof(struct ifaddrmsg), flags);
4804
	if (!nlh)
4805
		return -EMSGSIZE;
4806

4807 4808 4809
	put_ifaddrmsg(nlh, ifa->prefix_len, ifa->flags, rt_scope(ifa->scope),
		      ifa->idev->dev->ifindex);

4810 4811
	if (!((ifa->flags&IFA_F_PERMANENT) &&
	      (ifa->prefered_lft == INFINITY_LIFE_TIME))) {
4812 4813 4814
		preferred = ifa->prefered_lft;
		valid = ifa->valid_lft;
		if (preferred != INFINITY_LIFE_TIME) {
L
Linus Torvalds 已提交
4815
			long tval = (jiffies - ifa->tstamp)/HZ;
4816 4817 4818 4819
			if (preferred > tval)
				preferred -= tval;
			else
				preferred = 0;
4820 4821 4822 4823 4824 4825
			if (valid != INFINITY_LIFE_TIME) {
				if (valid > tval)
					valid -= tval;
				else
					valid = 0;
			}
L
Linus Torvalds 已提交
4826 4827
		}
	} else {
4828 4829 4830 4831
		preferred = INFINITY_LIFE_TIME;
		valid = INFINITY_LIFE_TIME;
	}

C
Cong Wang 已提交
4832
	if (!ipv6_addr_any(&ifa->peer_addr)) {
4833 4834
		if (nla_put_in6_addr(skb, IFA_LOCAL, &ifa->addr) < 0 ||
		    nla_put_in6_addr(skb, IFA_ADDRESS, &ifa->peer_addr) < 0)
4835 4836
			goto error;
	} else
4837
		if (nla_put_in6_addr(skb, IFA_ADDRESS, &ifa->addr) < 0)
4838 4839
			goto error;

4840 4841 4842 4843
	if (ifa->rt_priority &&
	    nla_put_u32(skb, IFA_RT_PRIORITY, ifa->rt_priority))
		goto error;

4844 4845
	if (put_cacheinfo(skb, ifa->cstamp, ifa->tstamp, preferred, valid) < 0)
		goto error;
L
Linus Torvalds 已提交
4846

4847 4848 4849
	if (nla_put_u32(skb, IFA_FLAGS, ifa->flags) < 0)
		goto error;

4850 4851
	nlmsg_end(skb, nlh);
	return 0;
4852 4853 4854 4855

error:
	nlmsg_cancel(skb, nlh);
	return -EMSGSIZE;
L
Linus Torvalds 已提交
4856 4857 4858
}

static int inet6_fill_ifmcaddr(struct sk_buff *skb, struct ifmcaddr6 *ifmca,
4859
				u32 portid, u32 seq, int event, u16 flags)
L
Linus Torvalds 已提交
4860 4861
{
	struct nlmsghdr  *nlh;
4862 4863 4864 4865 4866
	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
4867

4868
	nlh = nlmsg_put(skb, portid, seq, event, sizeof(struct ifaddrmsg), flags);
4869
	if (!nlh)
4870
		return -EMSGSIZE;
4871

4872
	put_ifaddrmsg(nlh, 128, IFA_F_PERMANENT, scope, ifindex);
4873
	if (nla_put_in6_addr(skb, IFA_MULTICAST, &ifmca->mca_addr) < 0 ||
4874
	    put_cacheinfo(skb, ifmca->mca_cstamp, ifmca->mca_tstamp,
4875 4876 4877 4878
			  INFINITY_LIFE_TIME, INFINITY_LIFE_TIME) < 0) {
		nlmsg_cancel(skb, nlh);
		return -EMSGSIZE;
	}
4879

4880 4881
	nlmsg_end(skb, nlh);
	return 0;
L
Linus Torvalds 已提交
4882 4883 4884
}

static int inet6_fill_ifacaddr(struct sk_buff *skb, struct ifacaddr6 *ifaca,
4885
				u32 portid, u32 seq, int event, unsigned int flags)
L
Linus Torvalds 已提交
4886
{
D
David Ahern 已提交
4887 4888
	struct net_device *dev = fib6_info_nh_dev(ifaca->aca_rt);
	int ifindex = dev ? dev->ifindex : 1;
L
Linus Torvalds 已提交
4889
	struct nlmsghdr  *nlh;
4890 4891 4892 4893
	u8 scope = RT_SCOPE_UNIVERSE;

	if (ipv6_addr_scope(&ifaca->aca_addr) & IFA_SITE)
		scope = RT_SCOPE_SITE;
L
Linus Torvalds 已提交
4894

4895
	nlh = nlmsg_put(skb, portid, seq, event, sizeof(struct ifaddrmsg), flags);
4896
	if (!nlh)
4897
		return -EMSGSIZE;
4898

4899
	put_ifaddrmsg(nlh, 128, IFA_F_PERMANENT, scope, ifindex);
4900
	if (nla_put_in6_addr(skb, IFA_ANYCAST, &ifaca->aca_addr) < 0 ||
4901
	    put_cacheinfo(skb, ifaca->aca_cstamp, ifaca->aca_tstamp,
4902 4903 4904 4905
			  INFINITY_LIFE_TIME, INFINITY_LIFE_TIME) < 0) {
		nlmsg_cancel(skb, nlh);
		return -EMSGSIZE;
	}
L
Linus Torvalds 已提交
4906

4907 4908
	nlmsg_end(skb, nlh);
	return 0;
L
Linus Torvalds 已提交
4909 4910
}

S
Stephen Hemminger 已提交
4911
enum addr_type_t {
L
Linus Torvalds 已提交
4912 4913 4914 4915 4916
	UNICAST_ADDR,
	MULTICAST_ADDR,
	ANYCAST_ADDR,
};

E
Eric Dumazet 已提交
4917 4918 4919 4920 4921 4922 4923 4924 4925 4926 4927 4928
/* 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) {
4929 4930 4931
	case UNICAST_ADDR: {
		struct inet6_ifaddr *ifa;

E
Eric Dumazet 已提交
4932
		/* unicast address incl. temp addr */
4933 4934
		list_for_each_entry(ifa, &idev->addr_list, if_list) {
			if (++ip_idx < s_ip_idx)
E
Eric Dumazet 已提交
4935 4936
				continue;
			err = inet6_fill_ifaddr(skb, ifa,
4937
						NETLINK_CB(cb->skb).portid,
E
Eric Dumazet 已提交
4938 4939 4940
						cb->nlh->nlmsg_seq,
						RTM_NEWADDR,
						NLM_F_MULTI);
4941
			if (err < 0)
E
Eric Dumazet 已提交
4942
				break;
4943
			nl_dump_check_consistent(cb, nlmsg_hdr(skb));
E
Eric Dumazet 已提交
4944 4945
		}
		break;
4946
	}
E
Eric Dumazet 已提交
4947 4948 4949 4950 4951 4952 4953
	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,
4954
						  NETLINK_CB(cb->skb).portid,
E
Eric Dumazet 已提交
4955 4956 4957
						  cb->nlh->nlmsg_seq,
						  RTM_GETMULTICAST,
						  NLM_F_MULTI);
4958
			if (err < 0)
E
Eric Dumazet 已提交
4959 4960 4961 4962 4963 4964 4965 4966 4967 4968
				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,
4969
						  NETLINK_CB(cb->skb).portid,
E
Eric Dumazet 已提交
4970 4971 4972
						  cb->nlh->nlmsg_seq,
						  RTM_GETANYCAST,
						  NLM_F_MULTI);
4973
			if (err < 0)
E
Eric Dumazet 已提交
4974 4975 4976 4977 4978 4979 4980 4981 4982 4983 4984
				break;
		}
		break;
	default:
		break;
	}
	read_unlock_bh(&idev->lock);
	*p_ip_idx = ip_idx;
	return err;
}

L
Linus Torvalds 已提交
4985 4986 4987
static int inet6_dump_addr(struct sk_buff *skb, struct netlink_callback *cb,
			   enum addr_type_t type)
{
E
Eric Dumazet 已提交
4988 4989
	struct net *net = sock_net(skb->sk);
	int h, s_h;
L
Linus Torvalds 已提交
4990 4991 4992
	int idx, ip_idx;
	int s_idx, s_ip_idx;
	struct net_device *dev;
E
Eric Dumazet 已提交
4993 4994
	struct inet6_dev *idev;
	struct hlist_head *head;
4995

E
Eric Dumazet 已提交
4996 4997 4998
	s_h = cb->args[0];
	s_idx = idx = cb->args[1];
	s_ip_idx = ip_idx = cb->args[2];
4999

E
Eric Dumazet 已提交
5000
	rcu_read_lock();
5001
	cb->seq = atomic_read(&net->ipv6.dev_addr_genid) ^ net->dev_base_seq;
E
Eric Dumazet 已提交
5002 5003 5004
	for (h = s_h; h < NETDEV_HASHENTRIES; h++, s_idx = 0) {
		idx = 0;
		head = &net->dev_index_head[h];
5005
		hlist_for_each_entry_rcu(dev, head, index_hlist) {
E
Eric Dumazet 已提交
5006 5007
			if (idx < s_idx)
				goto cont;
5008
			if (h > s_h || idx > s_idx)
E
Eric Dumazet 已提交
5009 5010
				s_ip_idx = 0;
			ip_idx = 0;
S
Stephen Hemminger 已提交
5011 5012
			idev = __in6_dev_get(dev);
			if (!idev)
E
Eric Dumazet 已提交
5013 5014 5015
				goto cont;

			if (in6_dump_addrs(idev, skb, cb, type,
5016
					   s_ip_idx, &ip_idx) < 0)
E
Eric Dumazet 已提交
5017
				goto done;
5018
cont:
E
Eric Dumazet 已提交
5019 5020
			idx++;
		}
L
Linus Torvalds 已提交
5021
	}
E
Eric Dumazet 已提交
5022 5023 5024 5025 5026 5027
done:
	rcu_read_unlock();
	cb->args[0] = h;
	cb->args[1] = idx;
	cb->args[2] = ip_idx;

L
Linus Torvalds 已提交
5028 5029 5030 5031 5032 5033
	return skb->len;
}

static int inet6_dump_ifaddr(struct sk_buff *skb, struct netlink_callback *cb)
{
	enum addr_type_t type = UNICAST_ADDR;
5034

L
Linus Torvalds 已提交
5035 5036 5037 5038 5039 5040
	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;
5041

L
Linus Torvalds 已提交
5042 5043 5044 5045 5046 5047 5048
	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;
5049

L
Linus Torvalds 已提交
5050 5051 5052
	return inet6_dump_addr(skb, cb, type);
}

5053 5054
static int inet6_rtm_getaddr(struct sk_buff *in_skb, struct nlmsghdr *nlh,
			     struct netlink_ext_ack *extack)
5055
{
5056
	struct net *net = sock_net(in_skb->sk);
5057 5058
	struct ifaddrmsg *ifm;
	struct nlattr *tb[IFA_MAX+1];
5059
	struct in6_addr *addr = NULL, *peer;
5060 5061 5062 5063 5064
	struct net_device *dev = NULL;
	struct inet6_ifaddr *ifa;
	struct sk_buff *skb;
	int err;

5065
	err = nlmsg_parse(nlh, sizeof(*ifm), tb, IFA_MAX, ifa_ipv6_policy,
5066
			  extack);
5067
	if (err < 0)
5068
		return err;
5069

5070
	addr = extract_addr(tb[IFA_ADDRESS], tb[IFA_LOCAL], &peer);
5071 5072
	if (!addr)
		return -EINVAL;
5073

5074
	ifm = nlmsg_data(nlh);
5075
	if (ifm->ifa_index)
5076
		dev = dev_get_by_index(net, ifm->ifa_index);
5077

S
Stephen Hemminger 已提交
5078 5079
	ifa = ipv6_get_ifaddr(net, addr, dev, 1);
	if (!ifa) {
5080 5081 5082
		err = -EADDRNOTAVAIL;
		goto errout;
	}
5083

S
Stephen Hemminger 已提交
5084 5085
	skb = nlmsg_new(inet6_ifaddr_msgsize(), GFP_KERNEL);
	if (!skb) {
5086
		err = -ENOBUFS;
5087
		goto errout_ifa;
5088 5089
	}

5090
	err = inet6_fill_ifaddr(skb, ifa, NETLINK_CB(in_skb).portid,
5091
				nlh->nlmsg_seq, RTM_NEWADDR, 0);
5092 5093 5094 5095 5096 5097
	if (err < 0) {
		/* -EMSGSIZE implies BUG in inet6_ifaddr_msgsize() */
		WARN_ON(err == -EMSGSIZE);
		kfree_skb(skb);
		goto errout_ifa;
	}
5098
	err = rtnl_unicast(skb, net, NETLINK_CB(in_skb).portid);
5099
errout_ifa:
5100
	in6_ifa_put(ifa);
5101
errout:
5102 5103
	if (dev)
		dev_put(dev);
5104 5105 5106
	return err;
}

L
Linus Torvalds 已提交
5107 5108 5109
static void inet6_ifa_notify(int event, struct inet6_ifaddr *ifa)
{
	struct sk_buff *skb;
5110
	struct net *net = dev_net(ifa->idev->dev);
5111
	int err = -ENOBUFS;
L
Linus Torvalds 已提交
5112

5113
	skb = nlmsg_new(inet6_ifaddr_msgsize(), GFP_ATOMIC);
5114
	if (!skb)
5115 5116 5117
		goto errout;

	err = inet6_fill_ifaddr(skb, ifa, 0, 0, event, 0);
5118 5119 5120 5121 5122 5123
	if (err < 0) {
		/* -EMSGSIZE implies BUG in inet6_ifaddr_msgsize() */
		WARN_ON(err == -EMSGSIZE);
		kfree_skb(skb);
		goto errout;
	}
5124 5125
	rtnl_notify(skb, net, 0, RTNLGRP_IPV6_IFADDR, NULL, GFP_ATOMIC);
	return;
5126 5127
errout:
	if (err < 0)
5128
		rtnl_set_sk_err(net, RTNLGRP_IPV6_IFADDR, err);
L
Linus Torvalds 已提交
5129 5130
}

D
Dave Jones 已提交
5131
static inline void ipv6_store_devconf(struct ipv6_devconf *cnf,
L
Linus Torvalds 已提交
5132 5133
				__s32 *array, int bytes)
{
5134 5135
	BUG_ON(bytes < (DEVCONF_MAX * 4));

L
Linus Torvalds 已提交
5136 5137 5138 5139 5140 5141 5142 5143 5144
	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;
5145 5146
	array[DEVCONF_RTR_SOLICIT_INTERVAL] =
		jiffies_to_msecs(cnf->rtr_solicit_interval);
5147 5148
	array[DEVCONF_RTR_SOLICIT_MAX_INTERVAL] =
		jiffies_to_msecs(cnf->rtr_solicit_max_interval);
5149 5150
	array[DEVCONF_RTR_SOLICIT_DELAY] =
		jiffies_to_msecs(cnf->rtr_solicit_delay);
L
Linus Torvalds 已提交
5151
	array[DEVCONF_FORCE_MLD_VERSION] = cnf->force_mld_version;
5152 5153 5154 5155
	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
5156 5157 5158 5159 5160 5161
	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;
5162
	array[DEVCONF_ACCEPT_RA_DEFRTR] = cnf->accept_ra_defrtr;
5163
	array[DEVCONF_ACCEPT_RA_MIN_HOP_LIMIT] = cnf->accept_ra_min_hop_limit;
5164
	array[DEVCONF_ACCEPT_RA_PINFO] = cnf->accept_ra_pinfo;
5165 5166
#ifdef CONFIG_IPV6_ROUTER_PREF
	array[DEVCONF_ACCEPT_RA_RTR_PREF] = cnf->accept_ra_rtr_pref;
5167 5168
	array[DEVCONF_RTR_PROBE_INTERVAL] =
		jiffies_to_msecs(cnf->rtr_probe_interval);
N
Neil Horman 已提交
5169
#ifdef CONFIG_IPV6_ROUTE_INFO
5170
	array[DEVCONF_ACCEPT_RA_RT_INFO_MIN_PLEN] = cnf->accept_ra_rt_info_min_plen;
5171 5172
	array[DEVCONF_ACCEPT_RA_RT_INFO_MAX_PLEN] = cnf->accept_ra_rt_info_max_plen;
#endif
5173
#endif
5174
	array[DEVCONF_PROXY_NDP] = cnf->proxy_ndp;
5175
	array[DEVCONF_ACCEPT_SOURCE_ROUTE] = cnf->accept_source_route;
5176 5177
#ifdef CONFIG_IPV6_OPTIMISTIC_DAD
	array[DEVCONF_OPTIMISTIC_DAD] = cnf->optimistic_dad;
5178
	array[DEVCONF_USE_OPTIMISTIC] = cnf->use_optimistic;
5179
#endif
5180 5181 5182
#ifdef CONFIG_IPV6_MROUTE
	array[DEVCONF_MC_FORWARDING] = cnf->mc_forwarding;
#endif
5183
	array[DEVCONF_DISABLE_IPV6] = cnf->disable_ipv6;
5184
	array[DEVCONF_ACCEPT_DAD] = cnf->accept_dad;
5185
	array[DEVCONF_FORCE_TLLAO] = cnf->force_tllao;
5186
	array[DEVCONF_NDISC_NOTIFY] = cnf->ndisc_notify;
5187
	array[DEVCONF_SUPPRESS_FRAG_NDISC] = cnf->suppress_frag_ndisc;
5188
	array[DEVCONF_ACCEPT_RA_FROM_LOCAL] = cnf->accept_ra_from_local;
5189
	array[DEVCONF_ACCEPT_RA_MTU] = cnf->accept_ra_mtu;
5190
	array[DEVCONF_IGNORE_ROUTES_WITH_LINKDOWN] = cnf->ignore_routes_with_linkdown;
5191
	/* we omit DEVCONF_STABLE_SECRET for now */
5192
	array[DEVCONF_USE_OIF_ADDRS_ONLY] = cnf->use_oif_addrs_only;
5193
	array[DEVCONF_DROP_UNICAST_IN_L2_MULTICAST] = cnf->drop_unicast_in_l2_multicast;
5194
	array[DEVCONF_DROP_UNSOLICITED_NA] = cnf->drop_unsolicited_na;
5195
	array[DEVCONF_KEEP_ADDR_ON_DOWN] = cnf->keep_addr_on_down;
5196
	array[DEVCONF_SEG6_ENABLED] = cnf->seg6_enabled;
5197 5198 5199
#ifdef CONFIG_IPV6_SEG6_HMAC
	array[DEVCONF_SEG6_REQUIRE_HMAC] = cnf->seg6_require_hmac;
#endif
5200
	array[DEVCONF_ENHANCED_DAD] = cnf->enhanced_dad;
5201
	array[DEVCONF_ADDR_GEN_MODE] = cnf->addr_gen_mode;
5202
	array[DEVCONF_DISABLE_POLICY] = cnf->disable_policy;
5203
	array[DEVCONF_NDISC_TCLASS] = cnf->ndisc_tclass;
L
Linus Torvalds 已提交
5204 5205
}

T
Thomas Graf 已提交
5206 5207 5208 5209 5210 5211
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 */
5212 5213
	     + 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
5214 5215
}

5216 5217 5218 5219 5220 5221 5222
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 */
5223
	       + nla_total_size(1) /* IFLA_OPERSTATE */
T
Thomas Graf 已提交
5224
	       + nla_total_size(inet6_ifla6_size()); /* IFLA_PROTINFO */
5225
}
5226

5227
static inline void __snmp6_fill_statsdev(u64 *stats, atomic_long_t *mib,
5228
					int bytes)
5229 5230
{
	int i;
5231
	int pad = bytes - sizeof(u64) * ICMP6_MIB_MAX;
5232 5233 5234
	BUG_ON(pad < 0);

	/* Use put_unaligned() because stats may not be aligned for u64. */
5235 5236
	put_unaligned(ICMP6_MIB_MAX, &stats[0]);
	for (i = 1; i < ICMP6_MIB_MAX; i++)
5237
		put_unaligned(atomic_long_read(&mib[i]), &stats[i]);
5238

5239
	memset(&stats[ICMP6_MIB_MAX], 0, pad);
5240 5241
}

W
WANG Cong 已提交
5242
static inline void __snmp6_fill_stats64(u64 *stats, void __percpu *mib,
5243
					int bytes, size_t syncpoff)
5244
{
5245 5246 5247 5248
	int i, c;
	u64 buff[IPSTATS_MIB_MAX];
	int pad = bytes - sizeof(u64) * IPSTATS_MIB_MAX;

5249 5250
	BUG_ON(pad < 0);

5251 5252
	memset(buff, 0, sizeof(buff));
	buff[0] = IPSTATS_MIB_MAX;
5253

5254 5255 5256 5257 5258 5259 5260
	for_each_possible_cpu(c) {
		for (i = 1; i < IPSTATS_MIB_MAX; i++)
			buff[i] += snmp_get_cpu_field64(mib, c, i, syncpoff);
	}

	memcpy(stats, buff, IPSTATS_MIB_MAX * sizeof(u64));
	memset(&stats[IPSTATS_MIB_MAX], 0, pad);
5261 5262
}

5263 5264 5265
static void snmp6_fill_stats(u64 *stats, struct inet6_dev *idev, int attrtype,
			     int bytes)
{
S
Stephen Hemminger 已提交
5266
	switch (attrtype) {
5267
	case IFLA_INET6_STATS:
5268 5269
		__snmp6_fill_stats64(stats, idev->stats.ipv6, bytes,
				     offsetof(struct ipstats_mib, syncp));
5270 5271
		break;
	case IFLA_INET6_ICMP6STATS:
5272
		__snmp6_fill_statsdev(stats, idev->stats.icmpv6dev->mibs, bytes);
5273 5274 5275 5276
		break;
	}
}

5277 5278
static int inet6_fill_ifla6_attrs(struct sk_buff *skb, struct inet6_dev *idev,
				  u32 ext_filter_mask)
T
Thomas Graf 已提交
5279 5280 5281 5282
{
	struct nlattr *nla;
	struct ifla_cacheinfo ci;

D
David S. Miller 已提交
5283 5284
	if (nla_put_u32(skb, IFLA_INET6_FLAGS, idev->if_flags))
		goto nla_put_failure;
T
Thomas Graf 已提交
5285
	ci.max_reasm_len = IPV6_MAXPLEN;
5286 5287
	ci.tstamp = cstamp_delta(idev->tstamp);
	ci.reachable_time = jiffies_to_msecs(idev->nd_parms->reachable_time);
J
Jiri Pirko 已提交
5288
	ci.retrans_time = jiffies_to_msecs(NEIGH_VAR(idev->nd_parms, RETRANS_TIME));
D
David S. Miller 已提交
5289 5290
	if (nla_put(skb, IFLA_INET6_CACHEINFO, sizeof(ci), &ci))
		goto nla_put_failure;
T
Thomas Graf 已提交
5291
	nla = nla_reserve(skb, IFLA_INET6_CONF, DEVCONF_MAX * sizeof(s32));
5292
	if (!nla)
T
Thomas Graf 已提交
5293 5294 5295 5296 5297
		goto nla_put_failure;
	ipv6_store_devconf(&idev->cnf, nla_data(nla), nla_len(nla));

	/* XXX - MC not implemented */

5298 5299 5300
	if (ext_filter_mask & RTEXT_FILTER_SKIP_STATS)
		return 0;

T
Thomas Graf 已提交
5301
	nla = nla_reserve(skb, IFLA_INET6_STATS, IPSTATS_MIB_MAX * sizeof(u64));
5302
	if (!nla)
T
Thomas Graf 已提交
5303 5304 5305 5306
		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));
5307
	if (!nla)
T
Thomas Graf 已提交
5308 5309 5310
		goto nla_put_failure;
	snmp6_fill_stats(nla_data(nla), idev, IFLA_INET6_ICMP6STATS, nla_len(nla));

5311
	nla = nla_reserve(skb, IFLA_INET6_TOKEN, sizeof(struct in6_addr));
5312
	if (!nla)
5313
		goto nla_put_failure;
5314

5315
	if (nla_put_u8(skb, IFLA_INET6_ADDR_GEN_MODE, idev->cnf.addr_gen_mode))
5316 5317
		goto nla_put_failure;

5318 5319 5320 5321
	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
5322 5323 5324 5325 5326 5327
	return 0;

nla_put_failure:
	return -EMSGSIZE;
}

5328 5329
static size_t inet6_get_link_af_size(const struct net_device *dev,
				     u32 ext_filter_mask)
T
Thomas Graf 已提交
5330 5331 5332 5333 5334 5335 5336
{
	if (!__in6_dev_get(dev))
		return 0;

	return inet6_ifla6_size();
}

5337 5338
static int inet6_fill_link_af(struct sk_buff *skb, const struct net_device *dev,
			      u32 ext_filter_mask)
T
Thomas Graf 已提交
5339 5340 5341 5342 5343 5344
{
	struct inet6_dev *idev = __in6_dev_get(dev);

	if (!idev)
		return -ENODATA;

5345
	if (inet6_fill_ifla6_attrs(skb, idev, ext_filter_mask) < 0)
T
Thomas Graf 已提交
5346 5347 5348 5349 5350
		return -EMSGSIZE;

	return 0;
}

5351 5352 5353 5354
static int inet6_set_iftoken(struct inet6_dev *idev, struct in6_addr *token)
{
	struct inet6_ifaddr *ifp;
	struct net_device *dev = idev->dev;
5355
	bool clear_token, update_rs = false;
5356
	struct in6_addr ll_addr;
5357

5358 5359
	ASSERT_RTNL();

5360
	if (!token)
5361 5362 5363 5364 5365
		return -EINVAL;
	if (dev->flags & (IFF_LOOPBACK | IFF_NOARP))
		return -EINVAL;
	if (!ipv6_accept_ra(idev))
		return -EINVAL;
5366
	if (idev->cnf.rtr_solicits == 0)
5367 5368 5369 5370 5371 5372 5373 5374 5375
		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);

5376 5377 5378 5379
	clear_token = ipv6_addr_any(token);
	if (clear_token)
		goto update_lft;

5380 5381 5382
	if (!idev->dead && (idev->if_flags & IF_READY) &&
	    !ipv6_get_lladdr(dev, &ll_addr, IFA_F_TENTATIVE |
			     IFA_F_OPTIMISTIC)) {
5383 5384 5385 5386 5387 5388
		/* 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;
	}
5389

5390
update_lft:
5391
	write_lock_bh(&idev->lock);
5392

5393
	if (update_rs) {
5394
		idev->if_flags |= IF_RS_SENT;
5395 5396
		idev->rs_interval = rfc3315_s14_backoff_init(
			idev->cnf.rtr_solicit_interval);
5397
		idev->rs_probes = 1;
5398
		addrconf_mod_rs_timer(idev, idev->rs_interval);
5399
	}
5400 5401 5402 5403

	/* Well, that's kinda nasty ... */
	list_for_each_entry(ifp, &idev->addr_list, if_list) {
		spin_lock(&ifp->lock);
5404
		if (ifp->tokenized) {
5405 5406 5407 5408 5409 5410 5411
			ifp->valid_lft = 0;
			ifp->prefered_lft = 0;
		}
		spin_unlock(&ifp->lock);
	}

	write_unlock_bh(&idev->lock);
5412
	inet6_ifinfo_notify(RTM_NEWLINK, idev);
5413
	addrconf_verify_rtnl();
5414 5415 5416
	return 0;
}

5417 5418 5419 5420 5421 5422 5423 5424 5425 5426 5427 5428 5429
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;

5430 5431
	return nla_parse_nested(tb, IFLA_INET6_MAX, nla, inet6_af_policy,
				NULL);
5432 5433
}

5434 5435 5436 5437 5438 5439 5440 5441 5442 5443 5444 5445 5446 5447 5448 5449 5450 5451 5452 5453
static int check_addr_gen_mode(int mode)
{
	if (mode != IN6_ADDR_GEN_MODE_EUI64 &&
	    mode != IN6_ADDR_GEN_MODE_NONE &&
	    mode != IN6_ADDR_GEN_MODE_STABLE_PRIVACY &&
	    mode != IN6_ADDR_GEN_MODE_RANDOM)
		return -EINVAL;
	return 1;
}

static int check_stable_privacy(struct inet6_dev *idev, struct net *net,
				int mode)
{
	if (mode == IN6_ADDR_GEN_MODE_STABLE_PRIVACY &&
	    !idev->cnf.stable_secret.initialized &&
	    !net->ipv6.devconf_dflt->stable_secret.initialized)
		return -EINVAL;
	return 1;
}

5454 5455 5456 5457 5458 5459 5460 5461 5462
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;

5463
	if (nla_parse_nested(tb, IFLA_INET6_MAX, nla, NULL, NULL) < 0)
5464 5465
		BUG();

5466
	if (tb[IFLA_INET6_TOKEN]) {
5467
		err = inet6_set_iftoken(idev, nla_data(tb[IFLA_INET6_TOKEN]));
5468 5469 5470 5471 5472 5473 5474
		if (err)
			return err;
	}

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

5475 5476
		if (check_addr_gen_mode(mode) < 0 ||
		    check_stable_privacy(idev, dev_net(dev), mode) < 0)
5477
			return -EINVAL;
5478

5479
		idev->cnf.addr_gen_mode = mode;
5480 5481
		err = 0;
	}
5482 5483 5484 5485

	return err;
}

5486
static int inet6_fill_ifinfo(struct sk_buff *skb, struct inet6_dev *idev,
5487
			     u32 portid, u32 seq, int event, unsigned int flags)
L
Linus Torvalds 已提交
5488
{
5489 5490 5491 5492 5493
	struct net_device *dev = idev->dev;
	struct ifinfomsg *hdr;
	struct nlmsghdr *nlh;
	void *protoinfo;

5494
	nlh = nlmsg_put(skb, portid, seq, event, sizeof(*hdr), flags);
5495
	if (!nlh)
5496
		return -EMSGSIZE;
5497 5498 5499 5500 5501 5502 5503 5504 5505

	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
5506 5507 5508 5509
	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) ||
5510
	    (dev->ifindex != dev_get_iflink(dev) &&
5511 5512 5513
	     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
5514
		goto nla_put_failure;
5515
	protoinfo = nla_nest_start(skb, IFLA_PROTINFO);
5516
	if (!protoinfo)
5517
		goto nla_put_failure;
L
Linus Torvalds 已提交
5518

5519
	if (inet6_fill_ifla6_attrs(skb, idev, 0) < 0)
5520
		goto nla_put_failure;
L
Linus Torvalds 已提交
5521

5522
	nla_nest_end(skb, protoinfo);
5523 5524
	nlmsg_end(skb, nlh);
	return 0;
L
Linus Torvalds 已提交
5525

5526
nla_put_failure:
5527 5528
	nlmsg_cancel(skb, nlh);
	return -EMSGSIZE;
L
Linus Torvalds 已提交
5529 5530 5531 5532
}

static int inet6_dump_ifinfo(struct sk_buff *skb, struct netlink_callback *cb)
{
5533
	struct net *net = sock_net(skb->sk);
E
Eric Dumazet 已提交
5534
	int h, s_h;
5535
	int idx = 0, s_idx;
L
Linus Torvalds 已提交
5536 5537
	struct net_device *dev;
	struct inet6_dev *idev;
E
Eric Dumazet 已提交
5538
	struct hlist_head *head;
L
Linus Torvalds 已提交
5539

E
Eric Dumazet 已提交
5540 5541 5542 5543 5544 5545 5546
	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];
5547
		hlist_for_each_entry_rcu(dev, head, index_hlist) {
E
Eric Dumazet 已提交
5548 5549 5550 5551 5552 5553
			if (idx < s_idx)
				goto cont;
			idev = __in6_dev_get(dev);
			if (!idev)
				goto cont;
			if (inet6_fill_ifinfo(skb, idev,
5554
					      NETLINK_CB(cb->skb).portid,
E
Eric Dumazet 已提交
5555
					      cb->nlh->nlmsg_seq,
5556
					      RTM_NEWLINK, NLM_F_MULTI) < 0)
E
Eric Dumazet 已提交
5557
				goto out;
5558
cont:
E
Eric Dumazet 已提交
5559 5560
			idx++;
		}
L
Linus Torvalds 已提交
5561
	}
E
Eric Dumazet 已提交
5562 5563 5564 5565
out:
	rcu_read_unlock();
	cb->args[1] = idx;
	cb->args[0] = h;
L
Linus Torvalds 已提交
5566 5567 5568 5569 5570 5571 5572

	return skb->len;
}

void inet6_ifinfo_notify(int event, struct inet6_dev *idev)
{
	struct sk_buff *skb;
5573
	struct net *net = dev_net(idev->dev);
5574
	int err = -ENOBUFS;
5575

5576
	skb = nlmsg_new(inet6_if_nlmsg_size(), GFP_ATOMIC);
5577
	if (!skb)
5578 5579 5580
		goto errout;

	err = inet6_fill_ifinfo(skb, idev, 0, 0, event, 0);
5581 5582 5583 5584 5585 5586
	if (err < 0) {
		/* -EMSGSIZE implies BUG in inet6_if_nlmsg_size() */
		WARN_ON(err == -EMSGSIZE);
		kfree_skb(skb);
		goto errout;
	}
5587
	rtnl_notify(skb, net, 0, RTNLGRP_IPV6_IFINFO, NULL, GFP_ATOMIC);
5588
	return;
5589 5590
errout:
	if (err < 0)
5591
		rtnl_set_sk_err(net, RTNLGRP_IPV6_IFINFO, err);
L
Linus Torvalds 已提交
5592 5593
}

5594 5595 5596 5597 5598 5599
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));
}
5600

L
Linus Torvalds 已提交
5601
static int inet6_fill_prefix(struct sk_buff *skb, struct inet6_dev *idev,
5602
			     struct prefix_info *pinfo, u32 portid, u32 seq,
5603
			     int event, unsigned int flags)
L
Linus Torvalds 已提交
5604
{
5605 5606
	struct prefixmsg *pmsg;
	struct nlmsghdr *nlh;
L
Linus Torvalds 已提交
5607 5608
	struct prefix_cacheinfo	ci;

5609
	nlh = nlmsg_put(skb, portid, seq, event, sizeof(*pmsg), flags);
5610
	if (!nlh)
5611
		return -EMSGSIZE;
5612 5613

	pmsg = nlmsg_data(nlh);
L
Linus Torvalds 已提交
5614
	pmsg->prefix_family = AF_INET6;
5615 5616
	pmsg->prefix_pad1 = 0;
	pmsg->prefix_pad2 = 0;
L
Linus Torvalds 已提交
5617 5618 5619
	pmsg->prefix_ifindex = idev->dev->ifindex;
	pmsg->prefix_len = pinfo->prefix_len;
	pmsg->prefix_type = pinfo->type;
5620
	pmsg->prefix_pad3 = 0;
L
Linus Torvalds 已提交
5621 5622 5623 5624 5625 5626
	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
5627 5628
	if (nla_put(skb, PREFIX_ADDRESS, sizeof(pinfo->prefix), &pinfo->prefix))
		goto nla_put_failure;
L
Linus Torvalds 已提交
5629 5630
	ci.preferred_time = ntohl(pinfo->prefered);
	ci.valid_time = ntohl(pinfo->valid);
D
David S. Miller 已提交
5631 5632
	if (nla_put(skb, PREFIX_CACHEINFO, sizeof(ci), &ci))
		goto nla_put_failure;
5633 5634
	nlmsg_end(skb, nlh);
	return 0;
L
Linus Torvalds 已提交
5635

5636
nla_put_failure:
5637 5638
	nlmsg_cancel(skb, nlh);
	return -EMSGSIZE;
L
Linus Torvalds 已提交
5639 5640
}

5641
static void inet6_prefix_notify(int event, struct inet6_dev *idev,
L
Linus Torvalds 已提交
5642 5643 5644
			 struct prefix_info *pinfo)
{
	struct sk_buff *skb;
5645
	struct net *net = dev_net(idev->dev);
5646
	int err = -ENOBUFS;
L
Linus Torvalds 已提交
5647

5648
	skb = nlmsg_new(inet6_prefix_nlmsg_size(), GFP_ATOMIC);
5649
	if (!skb)
5650 5651 5652
		goto errout;

	err = inet6_fill_prefix(skb, idev, pinfo, 0, 0, event, 0);
5653 5654 5655 5656 5657 5658
	if (err < 0) {
		/* -EMSGSIZE implies BUG in inet6_prefix_nlmsg_size() */
		WARN_ON(err == -EMSGSIZE);
		kfree_skb(skb);
		goto errout;
	}
5659 5660
	rtnl_notify(skb, net, 0, RTNLGRP_IPV6_PREFIX, NULL, GFP_ATOMIC);
	return;
5661 5662
errout:
	if (err < 0)
5663
		rtnl_set_sk_err(net, RTNLGRP_IPV6_PREFIX, err);
L
Linus Torvalds 已提交
5664 5665 5666 5667
}

static void __ipv6_ifa_notify(int event, struct inet6_ifaddr *ifp)
{
5668 5669
	struct net *net = dev_net(ifp->idev->dev);

5670 5671 5672
	if (event)
		ASSERT_RTNL();

L
Linus Torvalds 已提交
5673 5674 5675 5676
	inet6_ifa_notify(event ? : RTM_NEWADDR, ifp);

	switch (event) {
	case RTM_NEWADDR:
5677 5678 5679 5680 5681 5682
		/*
		 * 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
		 */
5683
		if (!rcu_access_pointer(ifp->rt->fib6_node))
5684
			ip6_ins_rt(net, ifp->rt);
L
Linus Torvalds 已提交
5685 5686
		if (ifp->idev->cnf.forwarding)
			addrconf_join_anycast(ifp);
C
Cong Wang 已提交
5687
		if (!ipv6_addr_any(&ifp->peer_addr))
5688
			addrconf_prefix_route(&ifp->peer_addr, 128, 0,
5689
					      ifp->idev->dev, 0, 0,
5690
					      GFP_ATOMIC);
L
Linus Torvalds 已提交
5691 5692 5693 5694 5695
		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
5696
		if (!ipv6_addr_any(&ifp->peer_addr)) {
5697
			struct fib6_info *rt;
5698

5699 5700
			rt = addrconf_get_prefix_route(&ifp->peer_addr, 128,
						       ifp->idev->dev, 0, 0);
M
Martin KaFai Lau 已提交
5701
			if (rt)
5702
				ip6_del_rt(net, rt);
5703
		}
5704
		if (ifp->rt) {
5705 5706
			ip6_del_rt(net, ifp->rt);
			ifp->rt = NULL;
5707
		}
H
Hannes Frederic Sowa 已提交
5708
		rt_genid_bump_ipv6(net);
L
Linus Torvalds 已提交
5709 5710
		break;
	}
5711
	atomic_inc(&net->ipv6.dev_addr_genid);
L
Linus Torvalds 已提交
5712 5713 5714 5715
}

static void ipv6_ifa_notify(int event, struct inet6_ifaddr *ifp)
{
5716
	rcu_read_lock_bh();
L
Linus Torvalds 已提交
5717 5718
	if (likely(ifp->idev->dead == 0))
		__ipv6_ifa_notify(event, ifp);
5719
	rcu_read_unlock_bh();
L
Linus Torvalds 已提交
5720 5721 5722 5723 5724
}

#ifdef CONFIG_SYSCTL

static
5725
int addrconf_sysctl_forward(struct ctl_table *ctl, int write,
L
Linus Torvalds 已提交
5726 5727 5728 5729
			   void __user *buffer, size_t *lenp, loff_t *ppos)
{
	int *valp = ctl->data;
	int val = *valp;
E
Eric W. Biederman 已提交
5730
	loff_t pos = *ppos;
5731
	struct ctl_table lctl;
L
Linus Torvalds 已提交
5732 5733
	int ret;

5734 5735 5736 5737 5738 5739 5740 5741
	/*
	 * 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
5742

5743
	if (write)
5744
		ret = addrconf_fixup_forwarding(ctl, valp, val);
E
Eric W. Biederman 已提交
5745 5746
	if (ret)
		*ppos = pos;
5747
	return ret;
L
Linus Torvalds 已提交
5748 5749
}

5750 5751 5752 5753 5754 5755 5756 5757 5758 5759 5760 5761 5762 5763 5764
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);
}

5765 5766
static void dev_disable_change(struct inet6_dev *idev)
{
5767 5768
	struct netdev_notifier_info info;

5769 5770 5771
	if (!idev || !idev->dev)
		return;

5772
	netdev_notifier_info_init(&info, idev->dev);
5773
	if (idev->cnf.disable_ipv6)
5774
		addrconf_notify(NULL, NETDEV_DOWN, &info);
5775
	else
5776
		addrconf_notify(NULL, NETDEV_UP, &info);
5777 5778 5779 5780 5781 5782 5783
}

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

5784
	for_each_netdev(net, dev) {
5785 5786 5787 5788 5789 5790 5791 5792 5793 5794
		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);
		}
	}
}

5795
static int addrconf_disable_ipv6(struct ctl_table *table, int *p, int newf)
5796 5797
{
	struct net *net;
5798 5799 5800 5801
	int old;

	if (!rtnl_trylock())
		return restart_syscall();
5802 5803

	net = (struct net *)table->extra2;
5804 5805
	old = *p;
	*p = newf;
5806

5807 5808
	if (p == &net->ipv6.devconf_dflt->disable_ipv6) {
		rtnl_unlock();
5809
		return 0;
E
Eric W. Biederman 已提交
5810
	}
5811 5812 5813 5814

	if (p == &net->ipv6.devconf_all->disable_ipv6) {
		net->ipv6.devconf_dflt->disable_ipv6 = newf;
		addrconf_disable_change(net, newf);
5815
	} else if ((!newf) ^ (!old))
5816 5817 5818 5819 5820 5821 5822
		dev_disable_change((struct inet6_dev *)table->extra1);

	rtnl_unlock();
	return 0;
}

static
5823
int addrconf_sysctl_disable(struct ctl_table *ctl, int write,
5824 5825 5826 5827
			    void __user *buffer, size_t *lenp, loff_t *ppos)
{
	int *valp = ctl->data;
	int val = *valp;
E
Eric W. Biederman 已提交
5828
	loff_t pos = *ppos;
5829
	struct ctl_table lctl;
5830 5831
	int ret;

5832 5833 5834 5835 5836 5837 5838 5839
	/*
	 * 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);
5840 5841 5842

	if (write)
		ret = addrconf_disable_ipv6(ctl, valp, val);
E
Eric W. Biederman 已提交
5843 5844
	if (ret)
		*ppos = pos;
5845 5846 5847
	return ret;
}

5848 5849 5850 5851 5852 5853 5854 5855 5856 5857 5858 5859 5860 5861 5862 5863 5864 5865 5866
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)
5867 5868
			inet6_netconf_notify_devconf(net, RTM_NEWNETCONF,
						     NETCONFA_PROXY_NEIGH,
5869 5870 5871
						     NETCONFA_IFINDEX_DEFAULT,
						     net->ipv6.devconf_dflt);
		else if (valp == &net->ipv6.devconf_all->proxy_ndp)
5872 5873
			inet6_netconf_notify_devconf(net, RTM_NEWNETCONF,
						     NETCONFA_PROXY_NEIGH,
5874 5875 5876 5877 5878
						     NETCONFA_IFINDEX_ALL,
						     net->ipv6.devconf_all);
		else {
			struct inet6_dev *idev = ctl->extra1;

5879 5880
			inet6_netconf_notify_devconf(net, RTM_NEWNETCONF,
						     NETCONFA_PROXY_NEIGH,
5881 5882 5883 5884 5885 5886 5887 5888 5889
						     idev->dev->ifindex,
						     &idev->cnf);
		}
		rtnl_unlock();
	}

	return ret;
}

5890 5891 5892 5893 5894
static int addrconf_sysctl_addr_gen_mode(struct ctl_table *ctl, int write,
					 void __user *buffer, size_t *lenp,
					 loff_t *ppos)
{
	int ret = 0;
5895
	u32 new_val;
5896 5897
	struct inet6_dev *idev = (struct inet6_dev *)ctl->extra1;
	struct net *net = (struct net *)ctl->extra2;
5898 5899 5900 5901 5902
	struct ctl_table tmp = {
		.data = &new_val,
		.maxlen = sizeof(new_val),
		.mode = ctl->mode,
	};
5903

5904 5905 5906
	if (!rtnl_trylock())
		return restart_syscall();

5907
	new_val = *((u32 *)ctl->data);
5908

5909 5910 5911
	ret = proc_douintvec(&tmp, write, buffer, lenp, ppos);
	if (ret != 0)
		goto out;
5912

5913
	if (write) {
5914 5915 5916 5917
		if (check_addr_gen_mode(new_val) < 0) {
			ret = -EINVAL;
			goto out;
		}
5918

5919
		if (idev) {
5920 5921 5922 5923
			if (check_stable_privacy(idev, net, new_val) < 0) {
				ret = -EINVAL;
				goto out;
			}
5924 5925 5926 5927 5928 5929

			if (idev->cnf.addr_gen_mode != new_val) {
				idev->cnf.addr_gen_mode = new_val;
				addrconf_dev_config(idev->dev);
			}
		}
5930 5931

		*((u32 *)ctl->data) = new_val;
5932 5933
	}

5934 5935 5936
out:
	rtnl_unlock();

5937 5938 5939
	return ret;
}

5940 5941 5942 5943 5944 5945 5946 5947
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;
5948
	struct net *net = ctl->extra2;
5949 5950
	struct ipv6_stable_secret *secret = ctl->data;

5951 5952 5953
	if (&net->ipv6.devconf_all->stable_secret == ctl->data)
		return -EIO;

5954 5955 5956 5957 5958 5959 5960 5961 5962 5963 5964
	lctl.maxlen = IPV6_MAX_STRLEN;
	lctl.data = str;

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

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

5965 5966 5967 5968
	err = snprintf(str, sizeof(str), "%pI6", &secret->secret);
	if (err >= sizeof(str)) {
		err = -EIO;
		goto out;
5969 5970 5971 5972 5973 5974 5975 5976 5977 5978 5979 5980 5981 5982
	}

	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;

5983 5984 5985 5986 5987 5988 5989
	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) {
5990
				idev->cnf.addr_gen_mode =
5991 5992 5993 5994 5995 5996
					IN6_ADDR_GEN_MODE_STABLE_PRIVACY;
			}
		}
	} else {
		struct inet6_dev *idev = ctl->extra1;

5997
		idev->cnf.addr_gen_mode = IN6_ADDR_GEN_MODE_STABLE_PRIVACY;
5998 5999
	}

6000 6001 6002 6003 6004
out:
	rtnl_unlock();

	return err;
}
6005

6006 6007 6008 6009 6010 6011 6012 6013 6014 6015 6016 6017 6018 6019 6020 6021 6022 6023 6024 6025 6026 6027 6028 6029 6030 6031 6032 6033
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;
}

6034 6035 6036 6037 6038 6039 6040 6041 6042 6043 6044 6045 6046 6047 6048 6049 6050 6051 6052 6053
static
void addrconf_set_nopolicy(struct rt6_info *rt, int action)
{
	if (rt) {
		if (action)
			rt->dst.flags |= DST_NOPOLICY;
		else
			rt->dst.flags &= ~DST_NOPOLICY;
	}
}

static
void addrconf_disable_policy_idev(struct inet6_dev *idev, int val)
{
	struct inet6_ifaddr *ifa;

	read_lock_bh(&idev->lock);
	list_for_each_entry(ifa, &idev->addr_list, if_list) {
		spin_lock(&ifa->lock);
		if (ifa->rt) {
6054
			struct fib6_info *rt = ifa->rt;
6055 6056
			int cpu;

6057
			rcu_read_lock();
6058
			ifa->rt->dst_nopolicy = val ? true : false;
6059 6060 6061 6062 6063 6064 6065 6066
			if (rt->rt6i_pcpu) {
				for_each_possible_cpu(cpu) {
					struct rt6_info **rtp;

					rtp = per_cpu_ptr(rt->rt6i_pcpu, cpu);
					addrconf_set_nopolicy(*rtp, val);
				}
			}
6067
			rcu_read_unlock();
6068 6069 6070 6071 6072 6073 6074 6075 6076 6077 6078 6079 6080 6081 6082 6083 6084 6085 6086 6087 6088 6089 6090 6091 6092 6093 6094 6095 6096 6097 6098 6099 6100 6101 6102 6103 6104 6105 6106 6107 6108 6109 6110 6111 6112 6113 6114 6115 6116 6117 6118 6119 6120 6121 6122 6123 6124 6125 6126 6127 6128 6129 6130 6131
		}
		spin_unlock(&ifa->lock);
	}
	read_unlock_bh(&idev->lock);
}

static
int addrconf_disable_policy(struct ctl_table *ctl, int *valp, int val)
{
	struct inet6_dev *idev;
	struct net *net;

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

	*valp = val;

	net = (struct net *)ctl->extra2;
	if (valp == &net->ipv6.devconf_dflt->disable_policy) {
		rtnl_unlock();
		return 0;
	}

	if (valp == &net->ipv6.devconf_all->disable_policy)  {
		struct net_device *dev;

		for_each_netdev(net, dev) {
			idev = __in6_dev_get(dev);
			if (idev)
				addrconf_disable_policy_idev(idev, val);
		}
	} else {
		idev = (struct inet6_dev *)ctl->extra1;
		addrconf_disable_policy_idev(idev, val);
	}

	rtnl_unlock();
	return 0;
}

static
int addrconf_sysctl_disable_policy(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;

	lctl = *ctl;
	lctl.data = &val;
	ret = proc_dointvec(&lctl, write, buffer, lenp, ppos);

	if (write && (*valp != val))
		ret = addrconf_disable_policy(ctl, valp, val);

	if (ret)
		*ppos = pos;

	return ret;
}

6132
static int minus_one = -1;
6133
static const int zero = 0;
6134 6135 6136
static const int one = 1;
static const int two_five_five = 255;

6137
static const struct ctl_table addrconf_sysctl[] = {
6138 6139 6140 6141 6142 6143 6144 6145 6146 6147 6148 6149
	{
		.procname	= "forwarding",
		.data		= &ipv6_devconf.forwarding,
		.maxlen		= sizeof(int),
		.mode		= 0644,
		.proc_handler	= addrconf_sysctl_forward,
	},
	{
		.procname	= "hop_limit",
		.data		= &ipv6_devconf.hop_limit,
		.maxlen		= sizeof(int),
		.mode		= 0644,
6150 6151 6152
		.proc_handler	= proc_dointvec_minmax,
		.extra1		= (void *)&one,
		.extra2		= (void *)&two_five_five,
6153 6154 6155 6156 6157 6158 6159 6160 6161 6162 6163 6164 6165 6166 6167 6168 6169 6170 6171 6172 6173 6174 6175 6176 6177 6178 6179 6180 6181 6182 6183 6184 6185 6186 6187 6188 6189 6190 6191 6192 6193
	},
	{
		.procname	= "mtu",
		.data		= &ipv6_devconf.mtu6,
		.maxlen		= sizeof(int),
		.mode		= 0644,
		.proc_handler	= addrconf_sysctl_mtu,
	},
	{
		.procname	= "accept_ra",
		.data		= &ipv6_devconf.accept_ra,
		.maxlen		= sizeof(int),
		.mode		= 0644,
		.proc_handler	= proc_dointvec,
	},
	{
		.procname	= "accept_redirects",
		.data		= &ipv6_devconf.accept_redirects,
		.maxlen		= sizeof(int),
		.mode		= 0644,
		.proc_handler	= proc_dointvec,
	},
	{
		.procname	= "autoconf",
		.data		= &ipv6_devconf.autoconf,
		.maxlen		= sizeof(int),
		.mode		= 0644,
		.proc_handler	= proc_dointvec,
	},
	{
		.procname	= "dad_transmits",
		.data		= &ipv6_devconf.dad_transmits,
		.maxlen		= sizeof(int),
		.mode		= 0644,
		.proc_handler	= proc_dointvec,
	},
	{
		.procname	= "router_solicitations",
		.data		= &ipv6_devconf.rtr_solicits,
		.maxlen		= sizeof(int),
		.mode		= 0644,
6194 6195
		.proc_handler	= proc_dointvec_minmax,
		.extra1		= &minus_one,
6196 6197 6198 6199 6200 6201 6202 6203
	},
	{
		.procname	= "router_solicitation_interval",
		.data		= &ipv6_devconf.rtr_solicit_interval,
		.maxlen		= sizeof(int),
		.mode		= 0644,
		.proc_handler	= proc_dointvec_jiffies,
	},
6204 6205 6206 6207 6208 6209 6210
	{
		.procname	= "router_solicitation_max_interval",
		.data		= &ipv6_devconf.rtr_solicit_max_interval,
		.maxlen		= sizeof(int),
		.mode		= 0644,
		.proc_handler	= proc_dointvec_jiffies,
	},
6211 6212 6213 6214 6215 6216 6217 6218 6219 6220 6221 6222 6223 6224 6225 6226 6227 6228 6229 6230 6231 6232 6233 6234 6235 6236 6237 6238 6239 6240 6241 6242 6243 6244 6245 6246 6247 6248 6249 6250 6251 6252 6253 6254 6255 6256 6257 6258 6259 6260 6261 6262 6263 6264 6265 6266 6267 6268 6269 6270 6271 6272 6273 6274 6275 6276 6277 6278 6279 6280 6281 6282 6283 6284 6285 6286 6287 6288 6289 6290 6291 6292 6293 6294 6295 6296 6297 6298 6299 6300 6301 6302 6303
	{
		.procname	= "router_solicitation_delay",
		.data		= &ipv6_devconf.rtr_solicit_delay,
		.maxlen		= sizeof(int),
		.mode		= 0644,
		.proc_handler	= proc_dointvec_jiffies,
	},
	{
		.procname	= "force_mld_version",
		.data		= &ipv6_devconf.force_mld_version,
		.maxlen		= sizeof(int),
		.mode		= 0644,
		.proc_handler	= proc_dointvec,
	},
	{
		.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,
	},
	{
		.procname	= "use_tempaddr",
		.data		= &ipv6_devconf.use_tempaddr,
		.maxlen		= sizeof(int),
		.mode		= 0644,
		.proc_handler	= proc_dointvec,
	},
	{
		.procname	= "temp_valid_lft",
		.data		= &ipv6_devconf.temp_valid_lft,
		.maxlen		= sizeof(int),
		.mode		= 0644,
		.proc_handler	= proc_dointvec,
	},
	{
		.procname	= "temp_prefered_lft",
		.data		= &ipv6_devconf.temp_prefered_lft,
		.maxlen		= sizeof(int),
		.mode		= 0644,
		.proc_handler	= proc_dointvec,
	},
	{
		.procname	= "regen_max_retry",
		.data		= &ipv6_devconf.regen_max_retry,
		.maxlen		= sizeof(int),
		.mode		= 0644,
		.proc_handler	= proc_dointvec,
	},
	{
		.procname	= "max_desync_factor",
		.data		= &ipv6_devconf.max_desync_factor,
		.maxlen		= sizeof(int),
		.mode		= 0644,
		.proc_handler	= proc_dointvec,
	},
	{
		.procname	= "max_addresses",
		.data		= &ipv6_devconf.max_addresses,
		.maxlen		= sizeof(int),
		.mode		= 0644,
		.proc_handler	= proc_dointvec,
	},
	{
		.procname	= "accept_ra_defrtr",
		.data		= &ipv6_devconf.accept_ra_defrtr,
		.maxlen		= sizeof(int),
		.mode		= 0644,
		.proc_handler	= proc_dointvec,
	},
	{
		.procname	= "accept_ra_min_hop_limit",
		.data		= &ipv6_devconf.accept_ra_min_hop_limit,
		.maxlen		= sizeof(int),
		.mode		= 0644,
		.proc_handler	= proc_dointvec,
	},
	{
		.procname	= "accept_ra_pinfo",
		.data		= &ipv6_devconf.accept_ra_pinfo,
		.maxlen		= sizeof(int),
		.mode		= 0644,
		.proc_handler	= proc_dointvec,
	},
6304
#ifdef CONFIG_IPV6_ROUTER_PREF
6305 6306 6307 6308 6309 6310 6311 6312 6313 6314 6315 6316 6317 6318
	{
		.procname	= "accept_ra_rtr_pref",
		.data		= &ipv6_devconf.accept_ra_rtr_pref,
		.maxlen		= sizeof(int),
		.mode		= 0644,
		.proc_handler	= proc_dointvec,
	},
	{
		.procname	= "router_probe_interval",
		.data		= &ipv6_devconf.rtr_probe_interval,
		.maxlen		= sizeof(int),
		.mode		= 0644,
		.proc_handler	= proc_dointvec_jiffies,
	},
N
Neil Horman 已提交
6319
#ifdef CONFIG_IPV6_ROUTE_INFO
6320 6321 6322 6323 6324 6325 6326
	{
		.procname	= "accept_ra_rt_info_min_plen",
		.data		= &ipv6_devconf.accept_ra_rt_info_min_plen,
		.maxlen		= sizeof(int),
		.mode		= 0644,
		.proc_handler	= proc_dointvec,
	},
6327 6328 6329 6330 6331 6332 6333
	{
		.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,
	},
6334
#endif
6335
#endif
6336 6337 6338 6339 6340 6341 6342 6343 6344 6345 6346 6347 6348 6349
	{
		.procname	= "proxy_ndp",
		.data		= &ipv6_devconf.proxy_ndp,
		.maxlen		= sizeof(int),
		.mode		= 0644,
		.proc_handler	= addrconf_sysctl_proxy_ndp,
	},
	{
		.procname	= "accept_source_route",
		.data		= &ipv6_devconf.accept_source_route,
		.maxlen		= sizeof(int),
		.mode		= 0644,
		.proc_handler	= proc_dointvec,
	},
6350
#ifdef CONFIG_IPV6_OPTIMISTIC_DAD
6351 6352 6353 6354 6355 6356 6357 6358 6359 6360 6361 6362 6363 6364
	{
		.procname	= "optimistic_dad",
		.data		= &ipv6_devconf.optimistic_dad,
		.maxlen		= sizeof(int),
		.mode		= 0644,
		.proc_handler   = proc_dointvec,
	},
	{
		.procname	= "use_optimistic",
		.data		= &ipv6_devconf.use_optimistic,
		.maxlen		= sizeof(int),
		.mode		= 0644,
		.proc_handler	= proc_dointvec,
	},
6365 6366
#endif
#ifdef CONFIG_IPV6_MROUTE
6367 6368 6369 6370 6371 6372 6373
	{
		.procname	= "mc_forwarding",
		.data		= &ipv6_devconf.mc_forwarding,
		.maxlen		= sizeof(int),
		.mode		= 0444,
		.proc_handler	= proc_dointvec,
	},
6374
#endif
6375 6376 6377 6378 6379 6380 6381 6382 6383 6384 6385 6386 6387 6388 6389 6390 6391 6392 6393 6394 6395 6396 6397 6398 6399 6400 6401 6402 6403 6404 6405 6406 6407 6408 6409 6410 6411 6412 6413 6414 6415 6416 6417 6418 6419 6420 6421 6422 6423 6424 6425 6426 6427 6428 6429 6430 6431 6432 6433 6434 6435 6436 6437 6438 6439 6440 6441 6442 6443 6444 6445 6446 6447 6448 6449 6450 6451 6452 6453 6454 6455 6456 6457 6458 6459 6460 6461 6462 6463 6464 6465 6466
	{
		.procname	= "disable_ipv6",
		.data		= &ipv6_devconf.disable_ipv6,
		.maxlen		= sizeof(int),
		.mode		= 0644,
		.proc_handler	= addrconf_sysctl_disable,
	},
	{
		.procname	= "accept_dad",
		.data		= &ipv6_devconf.accept_dad,
		.maxlen		= sizeof(int),
		.mode		= 0644,
		.proc_handler	= proc_dointvec,
	},
	{
		.procname	= "force_tllao",
		.data		= &ipv6_devconf.force_tllao,
		.maxlen		= sizeof(int),
		.mode		= 0644,
		.proc_handler	= proc_dointvec
	},
	{
		.procname	= "ndisc_notify",
		.data		= &ipv6_devconf.ndisc_notify,
		.maxlen		= sizeof(int),
		.mode		= 0644,
		.proc_handler	= proc_dointvec
	},
	{
		.procname	= "suppress_frag_ndisc",
		.data		= &ipv6_devconf.suppress_frag_ndisc,
		.maxlen		= sizeof(int),
		.mode		= 0644,
		.proc_handler	= proc_dointvec
	},
	{
		.procname	= "accept_ra_from_local",
		.data		= &ipv6_devconf.accept_ra_from_local,
		.maxlen		= sizeof(int),
		.mode		= 0644,
		.proc_handler	= proc_dointvec,
	},
	{
		.procname	= "accept_ra_mtu",
		.data		= &ipv6_devconf.accept_ra_mtu,
		.maxlen		= sizeof(int),
		.mode		= 0644,
		.proc_handler	= proc_dointvec,
	},
	{
		.procname	= "stable_secret",
		.data		= &ipv6_devconf.stable_secret,
		.maxlen		= IPV6_MAX_STRLEN,
		.mode		= 0600,
		.proc_handler	= addrconf_sysctl_stable_secret,
	},
	{
		.procname	= "use_oif_addrs_only",
		.data		= &ipv6_devconf.use_oif_addrs_only,
		.maxlen		= sizeof(int),
		.mode		= 0644,
		.proc_handler	= proc_dointvec,
	},
	{
		.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,
	},
	{
		.procname	= "drop_unicast_in_l2_multicast",
		.data		= &ipv6_devconf.drop_unicast_in_l2_multicast,
		.maxlen		= sizeof(int),
		.mode		= 0644,
		.proc_handler	= proc_dointvec,
	},
	{
		.procname	= "drop_unsolicited_na",
		.data		= &ipv6_devconf.drop_unsolicited_na,
		.maxlen		= sizeof(int),
		.mode		= 0644,
		.proc_handler	= proc_dointvec,
	},
	{
		.procname	= "keep_addr_on_down",
		.data		= &ipv6_devconf.keep_addr_on_down,
		.maxlen		= sizeof(int),
		.mode		= 0644,
		.proc_handler	= proc_dointvec,

	},
6467 6468 6469 6470 6471 6472 6473
	{
		.procname	= "seg6_enabled",
		.data		= &ipv6_devconf.seg6_enabled,
		.maxlen		= sizeof(int),
		.mode		= 0644,
		.proc_handler	= proc_dointvec,
	},
6474 6475 6476 6477 6478 6479 6480 6481 6482
#ifdef CONFIG_IPV6_SEG6_HMAC
	{
		.procname	= "seg6_require_hmac",
		.data		= &ipv6_devconf.seg6_require_hmac,
		.maxlen		= sizeof(int),
		.mode		= 0644,
		.proc_handler	= proc_dointvec,
	},
#endif
6483 6484 6485 6486 6487 6488 6489
	{
		.procname       = "enhanced_dad",
		.data           = &ipv6_devconf.enhanced_dad,
		.maxlen         = sizeof(int),
		.mode           = 0644,
		.proc_handler   = proc_dointvec,
	},
6490 6491 6492 6493 6494 6495 6496
	{
		.procname		= "addr_gen_mode",
		.data			= &ipv6_devconf.addr_gen_mode,
		.maxlen			= sizeof(int),
		.mode			= 0644,
		.proc_handler	= addrconf_sysctl_addr_gen_mode,
	},
6497 6498 6499 6500 6501 6502 6503
	{
		.procname       = "disable_policy",
		.data           = &ipv6_devconf.disable_policy,
		.maxlen         = sizeof(int),
		.mode           = 0644,
		.proc_handler   = addrconf_sysctl_disable_policy,
	},
6504 6505 6506 6507 6508 6509 6510 6511 6512
	{
		.procname	= "ndisc_tclass",
		.data		= &ipv6_devconf.ndisc_tclass,
		.maxlen		= sizeof(int),
		.mode		= 0644,
		.proc_handler	= proc_dointvec_minmax,
		.extra1		= (void *)&zero,
		.extra2		= (void *)&two_five_five,
	},
6513 6514 6515
	{
		/* sentinel */
	}
L
Linus Torvalds 已提交
6516 6517
};

6518
static int __addrconf_sysctl_register(struct net *net, char *dev_name,
6519
		struct inet6_dev *idev, struct ipv6_devconf *p)
L
Linus Torvalds 已提交
6520
{
6521
	int i, ifindex;
6522
	struct ctl_table *table;
6523
	char path[sizeof("net/ipv6/conf/") + IFNAMSIZ];
6524

6525 6526
	table = kmemdup(addrconf_sysctl, sizeof(addrconf_sysctl), GFP_KERNEL);
	if (!table)
6527 6528
		goto out;

6529 6530
	for (i = 0; table[i].data; i++) {
		table[i].data += (char *)p - (char *)&ipv6_devconf;
6531 6532 6533 6534 6535 6536 6537 6538
		/* If one of these is already set, then it is not safe to
		 * overwrite either of them: this makes proc_dointvec_minmax
		 * usable.
		 */
		if (!table[i].extra1 && !table[i].extra2) {
			table[i].extra1 = idev; /* embedded; no ref */
			table[i].extra2 = net;
		}
L
Linus Torvalds 已提交
6539 6540
	}

6541
	snprintf(path, sizeof(path), "net/ipv6/conf/%s", dev_name);
L
Linus Torvalds 已提交
6542

6543 6544
	p->sysctl_header = register_net_sysctl(net, path, table);
	if (!p->sysctl_header)
6545
		goto free;
6546

6547 6548 6549 6550 6551 6552
	if (!strcmp(dev_name, "all"))
		ifindex = NETCONFA_IFINDEX_ALL;
	else if (!strcmp(dev_name, "default"))
		ifindex = NETCONFA_IFINDEX_DEFAULT;
	else
		ifindex = idev->dev->ifindex;
6553 6554
	inet6_netconf_notify_devconf(net, RTM_NEWNETCONF, NETCONFA_ALL,
				     ifindex, p);
6555
	return 0;
L
Linus Torvalds 已提交
6556

6557
free:
6558
	kfree(table);
6559
out:
6560
	return -ENOBUFS;
L
Linus Torvalds 已提交
6561 6562
}

6563 6564
static void __addrconf_sysctl_unregister(struct net *net,
					 struct ipv6_devconf *p, int ifindex)
6565
{
6566
	struct ctl_table *table;
6567

6568
	if (!p->sysctl_header)
6569 6570
		return;

6571 6572 6573 6574
	table = p->sysctl_header->ctl_table_arg;
	unregister_net_sysctl_table(p->sysctl_header);
	p->sysctl_header = NULL;
	kfree(table);
6575 6576

	inet6_netconf_notify_devconf(net, RTM_DELNETCONF, 0, ifindex, NULL);
6577 6578
}

6579
static int addrconf_sysctl_register(struct inet6_dev *idev)
6580
{
6581 6582 6583 6584 6585 6586 6587 6588 6589 6590 6591 6592 6593 6594 6595
	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;
6596 6597
}

6598
static void addrconf_sysctl_unregister(struct inet6_dev *idev)
L
Linus Torvalds 已提交
6599
{
6600 6601
	__addrconf_sysctl_unregister(dev_net(idev->dev), &idev->cnf,
				     idev->dev->ifindex);
6602
	neigh_sysctl_unregister(idev->nd_parms);
L
Linus Torvalds 已提交
6603 6604 6605 6606 6607
}


#endif

6608
static int __net_init addrconf_init_net(struct net *net)
6609
{
6610
	int err = -ENOMEM;
6611 6612
	struct ipv6_devconf *all, *dflt;

6613
	all = kmemdup(&ipv6_devconf, sizeof(ipv6_devconf), GFP_KERNEL);
6614
	if (!all)
6615
		goto err_alloc_all;
6616

6617
	dflt = kmemdup(&ipv6_devconf_dflt, sizeof(ipv6_devconf_dflt), GFP_KERNEL);
6618
	if (!dflt)
6619
		goto err_alloc_dflt;
6620

6621 6622 6623
	/* these will be inherited by all namespaces */
	dflt->autoconf = ipv6_defaults.autoconf;
	dflt->disable_ipv6 = ipv6_defaults.disable_ipv6;
6624

6625 6626 6627
	dflt->stable_secret.initialized = false;
	all->stable_secret.initialized = false;

6628 6629 6630 6631
	net->ipv6.devconf_all = all;
	net->ipv6.devconf_dflt = dflt;

#ifdef CONFIG_SYSCTL
6632
	err = __addrconf_sysctl_register(net, "all", NULL, all);
6633 6634 6635
	if (err < 0)
		goto err_reg_all;

6636
	err = __addrconf_sysctl_register(net, "default", NULL, dflt);
6637 6638 6639 6640 6641 6642 6643
	if (err < 0)
		goto err_reg_dflt;
#endif
	return 0;

#ifdef CONFIG_SYSCTL
err_reg_dflt:
6644
	__addrconf_sysctl_unregister(net, all, NETCONFA_IFINDEX_ALL);
6645 6646 6647 6648 6649 6650 6651 6652 6653
err_reg_all:
	kfree(dflt);
#endif
err_alloc_dflt:
	kfree(all);
err_alloc_all:
	return err;
}

6654
static void __net_exit addrconf_exit_net(struct net *net)
6655 6656
{
#ifdef CONFIG_SYSCTL
6657 6658 6659 6660
	__addrconf_sysctl_unregister(net, net->ipv6.devconf_dflt,
				     NETCONFA_IFINDEX_DEFAULT);
	__addrconf_sysctl_unregister(net, net->ipv6.devconf_all,
				     NETCONFA_IFINDEX_ALL);
6661
#endif
Z
zhuyj 已提交
6662 6663
	kfree(net->ipv6.devconf_dflt);
	kfree(net->ipv6.devconf_all);
6664 6665 6666 6667 6668 6669 6670
}

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

6671
static struct rtnl_af_ops inet6_ops __read_mostly = {
T
Thomas Graf 已提交
6672 6673 6674
	.family		  = AF_INET6,
	.fill_link_af	  = inet6_fill_link_af,
	.get_link_af_size = inet6_get_link_af_size,
6675
	.validate_link_af = inet6_validate_link_af,
6676
	.set_link_af	  = inet6_set_link_af,
T
Thomas Graf 已提交
6677 6678
};

L
Linus Torvalds 已提交
6679 6680 6681 6682 6683 6684
/*
 *	Init / cleanup code
 */

int __init addrconf_init(void)
{
6685
	struct inet6_dev *idev;
6686
	int i, err;
6687

S
Stephen Hemminger 已提交
6688 6689
	err = ipv6_addr_label_init();
	if (err < 0) {
6690 6691
		pr_crit("%s: cannot initialize default policy table: %d\n",
			__func__, err);
6692
		goto out;
6693
	}
L
Linus Torvalds 已提交
6694

6695 6696 6697
	err = register_pernet_subsys(&addrconf_ops);
	if (err < 0)
		goto out_addrlabel;
6698

6699 6700 6701 6702 6703 6704
	addrconf_wq = create_workqueue("ipv6_addrconf");
	if (!addrconf_wq) {
		err = -ENOMEM;
		goto out_nowq;
	}

L
Linus Torvalds 已提交
6705 6706 6707 6708 6709 6710 6711 6712 6713 6714 6715 6716 6717 6718 6719 6720 6721 6722 6723
	/* 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();
6724
	idev = ipv6_add_dev(init_net.loopback_dev);
L
Linus Torvalds 已提交
6725
	rtnl_unlock();
6726 6727
	if (IS_ERR(idev)) {
		err = PTR_ERR(idev);
6728
		goto errlo;
6729
	}
L
Linus Torvalds 已提交
6730

6731 6732
	ip6_route_init_special_entries();

6733 6734 6735
	for (i = 0; i < IN6_ADDR_HSIZE; i++)
		INIT_HLIST_HEAD(&inet6_addr_lst[i]);

L
Linus Torvalds 已提交
6736 6737
	register_netdevice_notifier(&ipv6_dev_notf);

6738
	addrconf_verify();
6739

6740
	rtnl_af_register(&inet6_ops);
T
Thomas Graf 已提交
6741

6742 6743
	err = rtnl_register_module(THIS_MODULE, PF_INET6, RTM_GETLINK,
				   NULL, inet6_dump_ifinfo, 0);
6744 6745 6746
	if (err < 0)
		goto errout;

6747 6748 6749 6750 6751 6752 6753 6754 6755 6756 6757 6758 6759 6760 6761 6762 6763 6764 6765 6766 6767 6768 6769 6770 6771 6772 6773
	err = rtnl_register_module(THIS_MODULE, PF_INET6, RTM_NEWADDR,
				   inet6_rtm_newaddr, NULL, 0);
	if (err < 0)
		goto errout;
	err = rtnl_register_module(THIS_MODULE, PF_INET6, RTM_DELADDR,
				   inet6_rtm_deladdr, NULL, 0);
	if (err < 0)
		goto errout;
	err = rtnl_register_module(THIS_MODULE, PF_INET6, RTM_GETADDR,
				   inet6_rtm_getaddr, inet6_dump_ifaddr,
				   RTNL_FLAG_DOIT_UNLOCKED);
	if (err < 0)
		goto errout;
	err = rtnl_register_module(THIS_MODULE, PF_INET6, RTM_GETMULTICAST,
				   NULL, inet6_dump_ifmcaddr, 0);
	if (err < 0)
		goto errout;
	err = rtnl_register_module(THIS_MODULE, PF_INET6, RTM_GETANYCAST,
				   NULL, inet6_dump_ifacaddr, 0);
	if (err < 0)
		goto errout;
	err = rtnl_register_module(THIS_MODULE, PF_INET6, RTM_GETNETCONF,
				   inet6_netconf_get_devconf,
				   inet6_netconf_dump_devconf,
				   RTNL_FLAG_DOIT_UNLOCKED);
	if (err < 0)
		goto errout;
6774 6775 6776
	err = ipv6_addr_label_rtnl_register();
	if (err < 0)
		goto errout;
6777

L
Linus Torvalds 已提交
6778
	return 0;
6779
errout:
6780
	rtnl_unregister_all(PF_INET6);
T
Thomas Graf 已提交
6781
	rtnl_af_unregister(&inet6_ops);
6782
	unregister_netdevice_notifier(&ipv6_dev_notf);
6783
errlo:
6784 6785
	destroy_workqueue(addrconf_wq);
out_nowq:
6786
	unregister_pernet_subsys(&addrconf_ops);
6787 6788 6789
out_addrlabel:
	ipv6_addr_label_cleanup();
out:
6790
	return err;
L
Linus Torvalds 已提交
6791 6792
}

6793
void addrconf_cleanup(void)
L
Linus Torvalds 已提交
6794
{
6795
	struct net_device *dev;
L
Linus Torvalds 已提交
6796 6797 6798
	int i;

	unregister_netdevice_notifier(&ipv6_dev_notf);
6799
	unregister_pernet_subsys(&addrconf_ops);
6800
	ipv6_addr_label_cleanup();
L
Linus Torvalds 已提交
6801

6802
	rtnl_af_unregister(&inet6_ops);
L
Linus Torvalds 已提交
6803

6804
	rtnl_lock();
T
Thomas Graf 已提交
6805

6806 6807 6808 6809 6810 6811 6812 6813
	/* 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
6814 6815 6816
	/*
	 *	Check hash table.
	 */
6817
	spin_lock_bh(&addrconf_hash_lock);
6818 6819
	for (i = 0; i < IN6_ADDR_HSIZE; i++)
		WARN_ON(!hlist_empty(&inet6_addr_lst[i]));
6820
	spin_unlock_bh(&addrconf_hash_lock);
6821
	cancel_delayed_work(&addr_chk_work);
L
Linus Torvalds 已提交
6822
	rtnl_unlock();
6823 6824

	destroy_workqueue(addrconf_wq);
L
Linus Torvalds 已提交
6825
}