xref: /kernel/linux/linux-5.10/net/ipv6/addrconf.c (revision 8c2ecf20)
1// SPDX-License-Identifier: GPL-2.0-or-later
2/*
3 *	IPv6 Address [auto]configuration
4 *	Linux INET6 implementation
5 *
6 *	Authors:
7 *	Pedro Roque		<roque@di.fc.ul.pt>
8 *	Alexey Kuznetsov	<kuznet@ms2.inr.ac.ru>
9 */
10
11/*
12 *	Changes:
13 *
14 *	Janos Farkas			:	delete timer on ifdown
15 *	<chexum@bankinf.banki.hu>
16 *	Andi Kleen			:	kill double kfree on module
17 *						unload.
18 *	Maciej W. Rozycki		:	FDDI support
19 *	sekiya@USAGI			:	Don't send too many RS
20 *						packets.
21 *	yoshfuji@USAGI			:       Fixed interval between DAD
22 *						packets.
23 *	YOSHIFUJI Hideaki @USAGI	:	improved accuracy of
24 *						address validation timer.
25 *	YOSHIFUJI Hideaki @USAGI	:	Privacy Extensions (RFC3041)
26 *						support.
27 *	Yuji SEKIYA @USAGI		:	Don't assign a same IPv6
28 *						address on a same interface.
29 *	YOSHIFUJI Hideaki @USAGI	:	ARCnet support
30 *	YOSHIFUJI Hideaki @USAGI	:	convert /proc/net/if_inet6 to
31 *						seq_file.
32 *	YOSHIFUJI Hideaki @USAGI	:	improved source address
33 *						selection; consider scope,
34 *						status etc.
35 */
36
37#define pr_fmt(fmt) "IPv6: " fmt
38
39#include <linux/errno.h>
40#include <linux/types.h>
41#include <linux/kernel.h>
42#include <linux/sched/signal.h>
43#include <linux/socket.h>
44#include <linux/sockios.h>
45#include <linux/net.h>
46#include <linux/inet.h>
47#include <linux/in6.h>
48#include <linux/netdevice.h>
49#include <linux/if_addr.h>
50#include <linux/if_arp.h>
51#include <linux/if_arcnet.h>
52#include <linux/if_infiniband.h>
53#include <linux/route.h>
54#include <linux/inetdevice.h>
55#include <linux/init.h>
56#include <linux/slab.h>
57#ifdef CONFIG_SYSCTL
58#include <linux/sysctl.h>
59#endif
60#include <linux/capability.h>
61#include <linux/delay.h>
62#include <linux/notifier.h>
63#include <linux/string.h>
64#include <linux/hash.h>
65
66#include <net/net_namespace.h>
67#include <net/sock.h>
68#include <net/snmp.h>
69
70#include <net/6lowpan.h>
71#include <net/firewire.h>
72#include <net/ipv6.h>
73#include <net/protocol.h>
74#include <net/ndisc.h>
75#include <net/ip6_route.h>
76#include <net/addrconf.h>
77#include <net/tcp.h>
78#include <net/ip.h>
79#include <net/netlink.h>
80#include <net/pkt_sched.h>
81#include <net/l3mdev.h>
82#include <linux/if_tunnel.h>
83#include <linux/rtnetlink.h>
84#include <linux/netconf.h>
85#include <linux/random.h>
86#include <linux/uaccess.h>
87#include <asm/unaligned.h>
88
89#include <linux/proc_fs.h>
90#include <linux/seq_file.h>
91#include <linux/export.h>
92
93#define	INFINITY_LIFE_TIME	0xFFFFFFFF
94
95#define IPV6_MAX_STRLEN \
96	sizeof("ffff:ffff:ffff:ffff:ffff:ffff:255.255.255.255")
97
98static inline u32 cstamp_delta(unsigned long cstamp)
99{
100	return (cstamp - INITIAL_JIFFIES) * 100UL / HZ;
101}
102
103static inline s32 rfc3315_s14_backoff_init(s32 irt)
104{
105	/* multiply 'initial retransmission time' by 0.9 .. 1.1 */
106	u64 tmp = (900000 + prandom_u32() % 200001) * (u64)irt;
107	do_div(tmp, 1000000);
108	return (s32)tmp;
109}
110
111static inline s32 rfc3315_s14_backoff_update(s32 rt, s32 mrt)
112{
113	/* multiply 'retransmission timeout' by 1.9 .. 2.1 */
114	u64 tmp = (1900000 + prandom_u32() % 200001) * (u64)rt;
115	do_div(tmp, 1000000);
116	if ((s32)tmp > mrt) {
117		/* multiply 'maximum retransmission time' by 0.9 .. 1.1 */
118		tmp = (900000 + prandom_u32() % 200001) * (u64)mrt;
119		do_div(tmp, 1000000);
120	}
121	return (s32)tmp;
122}
123
124#ifdef CONFIG_SYSCTL
125static int addrconf_sysctl_register(struct inet6_dev *idev);
126static void addrconf_sysctl_unregister(struct inet6_dev *idev);
127#else
128static inline int addrconf_sysctl_register(struct inet6_dev *idev)
129{
130	return 0;
131}
132
133static inline void addrconf_sysctl_unregister(struct inet6_dev *idev)
134{
135}
136#endif
137
138static void ipv6_gen_rnd_iid(struct in6_addr *addr);
139
140static int ipv6_generate_eui64(u8 *eui, struct net_device *dev);
141static int ipv6_count_addresses(const struct inet6_dev *idev);
142static int ipv6_generate_stable_address(struct in6_addr *addr,
143					u8 dad_count,
144					const struct inet6_dev *idev);
145
146#define IN6_ADDR_HSIZE_SHIFT	8
147#define IN6_ADDR_HSIZE		(1 << IN6_ADDR_HSIZE_SHIFT)
148/*
149 *	Configured unicast address hash table
150 */
151static struct hlist_head inet6_addr_lst[IN6_ADDR_HSIZE];
152static DEFINE_SPINLOCK(addrconf_hash_lock);
153
154static void addrconf_verify(void);
155static void addrconf_verify_rtnl(void);
156static void addrconf_verify_work(struct work_struct *);
157
158static struct workqueue_struct *addrconf_wq;
159static DECLARE_DELAYED_WORK(addr_chk_work, addrconf_verify_work);
160
161static void addrconf_join_anycast(struct inet6_ifaddr *ifp);
162static void addrconf_leave_anycast(struct inet6_ifaddr *ifp);
163
164static void addrconf_type_change(struct net_device *dev,
165				 unsigned long event);
166static int addrconf_ifdown(struct net_device *dev, bool unregister);
167
168static struct fib6_info *addrconf_get_prefix_route(const struct in6_addr *pfx,
169						  int plen,
170						  const struct net_device *dev,
171						  u32 flags, u32 noflags,
172						  bool no_gw);
173
174static void addrconf_dad_start(struct inet6_ifaddr *ifp);
175static void addrconf_dad_work(struct work_struct *w);
176static void addrconf_dad_completed(struct inet6_ifaddr *ifp, bool bump_id,
177				   bool send_na);
178static void addrconf_dad_run(struct inet6_dev *idev, bool restart);
179static void addrconf_rs_timer(struct timer_list *t);
180static void __ipv6_ifa_notify(int event, struct inet6_ifaddr *ifa);
181static void ipv6_ifa_notify(int event, struct inet6_ifaddr *ifa);
182
183static void inet6_prefix_notify(int event, struct inet6_dev *idev,
184				struct prefix_info *pinfo);
185
186static struct ipv6_devconf ipv6_devconf __read_mostly = {
187	.forwarding		= 0,
188	.hop_limit		= IPV6_DEFAULT_HOPLIMIT,
189	.mtu6			= IPV6_MIN_MTU,
190	.accept_ra		= 1,
191	.accept_redirects	= 1,
192	.autoconf		= 1,
193	.force_mld_version	= 0,
194	.mldv1_unsolicited_report_interval = 10 * HZ,
195	.mldv2_unsolicited_report_interval = HZ,
196	.dad_transmits		= 1,
197	.rtr_solicits		= MAX_RTR_SOLICITATIONS,
198	.rtr_solicit_interval	= RTR_SOLICITATION_INTERVAL,
199	.rtr_solicit_max_interval = RTR_SOLICITATION_MAX_INTERVAL,
200	.rtr_solicit_delay	= MAX_RTR_SOLICITATION_DELAY,
201	.use_tempaddr		= 0,
202	.temp_valid_lft		= TEMP_VALID_LIFETIME,
203	.temp_prefered_lft	= TEMP_PREFERRED_LIFETIME,
204	.regen_max_retry	= REGEN_MAX_RETRY,
205	.max_desync_factor	= MAX_DESYNC_FACTOR,
206	.max_addresses		= IPV6_MAX_ADDRESSES,
207	.accept_ra_defrtr	= 1,
208	.accept_ra_from_local	= 0,
209	.accept_ra_min_hop_limit= 1,
210	.accept_ra_min_lft	= 0,
211	.accept_ra_pinfo	= 1,
212#ifdef CONFIG_IPV6_ROUTER_PREF
213	.accept_ra_rtr_pref	= 1,
214	.rtr_probe_interval	= 60 * HZ,
215#ifdef CONFIG_IPV6_ROUTE_INFO
216	.accept_ra_rt_info_min_plen = 0,
217	.accept_ra_rt_info_max_plen = 0,
218#endif
219#endif
220	.proxy_ndp		= 0,
221	.accept_source_route	= 0,	/* we do not accept RH0 by default. */
222	.disable_ipv6		= 0,
223	.accept_dad		= 0,
224	.suppress_frag_ndisc	= 1,
225	.accept_ra_mtu		= 1,
226	.stable_secret		= {
227		.initialized = false,
228	},
229	.use_oif_addrs_only	= 0,
230	.ignore_routes_with_linkdown = 0,
231	.keep_addr_on_down	= 0,
232	.seg6_enabled		= 0,
233#ifdef CONFIG_IPV6_SEG6_HMAC
234	.seg6_require_hmac	= 0,
235#endif
236	.enhanced_dad           = 1,
237	.addr_gen_mode		= IN6_ADDR_GEN_MODE_EUI64,
238	.disable_policy		= 0,
239	.rpl_seg_enabled	= 0,
240};
241
242static struct ipv6_devconf ipv6_devconf_dflt __read_mostly = {
243	.forwarding		= 0,
244	.hop_limit		= IPV6_DEFAULT_HOPLIMIT,
245	.mtu6			= IPV6_MIN_MTU,
246	.accept_ra		= 1,
247	.accept_redirects	= 1,
248	.autoconf		= 1,
249	.force_mld_version	= 0,
250	.mldv1_unsolicited_report_interval = 10 * HZ,
251	.mldv2_unsolicited_report_interval = HZ,
252	.dad_transmits		= 1,
253	.rtr_solicits		= MAX_RTR_SOLICITATIONS,
254	.rtr_solicit_interval	= RTR_SOLICITATION_INTERVAL,
255	.rtr_solicit_max_interval = RTR_SOLICITATION_MAX_INTERVAL,
256	.rtr_solicit_delay	= MAX_RTR_SOLICITATION_DELAY,
257	.use_tempaddr		= 0,
258	.temp_valid_lft		= TEMP_VALID_LIFETIME,
259	.temp_prefered_lft	= TEMP_PREFERRED_LIFETIME,
260	.regen_max_retry	= REGEN_MAX_RETRY,
261	.max_desync_factor	= MAX_DESYNC_FACTOR,
262	.max_addresses		= IPV6_MAX_ADDRESSES,
263	.accept_ra_defrtr	= 1,
264	.accept_ra_from_local	= 0,
265	.accept_ra_min_hop_limit= 1,
266	.accept_ra_min_lft	= 0,
267	.accept_ra_pinfo	= 1,
268#ifdef CONFIG_IPV6_ROUTER_PREF
269	.accept_ra_rtr_pref	= 1,
270	.rtr_probe_interval	= 60 * HZ,
271#ifdef CONFIG_IPV6_ROUTE_INFO
272	.accept_ra_rt_info_min_plen = 0,
273	.accept_ra_rt_info_max_plen = 0,
274#endif
275#endif
276	.proxy_ndp		= 0,
277	.accept_source_route	= 0,	/* we do not accept RH0 by default. */
278	.disable_ipv6		= 0,
279	.accept_dad		= 1,
280	.suppress_frag_ndisc	= 1,
281	.accept_ra_mtu		= 1,
282	.stable_secret		= {
283		.initialized = false,
284	},
285	.use_oif_addrs_only	= 0,
286	.ignore_routes_with_linkdown = 0,
287	.keep_addr_on_down	= 0,
288	.seg6_enabled		= 0,
289#ifdef CONFIG_IPV6_SEG6_HMAC
290	.seg6_require_hmac	= 0,
291#endif
292	.enhanced_dad           = 1,
293	.addr_gen_mode		= IN6_ADDR_GEN_MODE_EUI64,
294	.disable_policy		= 0,
295	.rpl_seg_enabled	= 0,
296};
297
298/* Check if link is ready: is it up and is a valid qdisc available */
299static inline bool addrconf_link_ready(const struct net_device *dev)
300{
301	return netif_oper_up(dev) && !qdisc_tx_is_noop(dev);
302}
303
304static void addrconf_del_rs_timer(struct inet6_dev *idev)
305{
306	if (del_timer(&idev->rs_timer))
307		__in6_dev_put(idev);
308}
309
310static void addrconf_del_dad_work(struct inet6_ifaddr *ifp)
311{
312	if (cancel_delayed_work(&ifp->dad_work))
313		__in6_ifa_put(ifp);
314}
315
316static void addrconf_mod_rs_timer(struct inet6_dev *idev,
317				  unsigned long when)
318{
319	if (!mod_timer(&idev->rs_timer, jiffies + when))
320		in6_dev_hold(idev);
321}
322
323static void addrconf_mod_dad_work(struct inet6_ifaddr *ifp,
324				   unsigned long delay)
325{
326	in6_ifa_hold(ifp);
327	if (mod_delayed_work(addrconf_wq, &ifp->dad_work, delay))
328		in6_ifa_put(ifp);
329}
330
331static int snmp6_alloc_dev(struct inet6_dev *idev)
332{
333	int i;
334
335	idev->stats.ipv6 = alloc_percpu(struct ipstats_mib);
336	if (!idev->stats.ipv6)
337		goto err_ip;
338
339	for_each_possible_cpu(i) {
340		struct ipstats_mib *addrconf_stats;
341		addrconf_stats = per_cpu_ptr(idev->stats.ipv6, i);
342		u64_stats_init(&addrconf_stats->syncp);
343	}
344
345
346	idev->stats.icmpv6dev = kzalloc(sizeof(struct icmpv6_mib_device),
347					GFP_KERNEL);
348	if (!idev->stats.icmpv6dev)
349		goto err_icmp;
350	idev->stats.icmpv6msgdev = kzalloc(sizeof(struct icmpv6msg_mib_device),
351					   GFP_KERNEL);
352	if (!idev->stats.icmpv6msgdev)
353		goto err_icmpmsg;
354
355	return 0;
356
357err_icmpmsg:
358	kfree(idev->stats.icmpv6dev);
359err_icmp:
360	free_percpu(idev->stats.ipv6);
361err_ip:
362	return -ENOMEM;
363}
364
365static struct inet6_dev *ipv6_add_dev(struct net_device *dev)
366{
367	struct inet6_dev *ndev;
368	int err = -ENOMEM;
369
370	ASSERT_RTNL();
371
372	if (dev->mtu < IPV6_MIN_MTU)
373		return ERR_PTR(-EINVAL);
374
375	ndev = kzalloc(sizeof(struct inet6_dev), GFP_KERNEL);
376	if (!ndev)
377		return ERR_PTR(err);
378
379	rwlock_init(&ndev->lock);
380	ndev->dev = dev;
381	INIT_LIST_HEAD(&ndev->addr_list);
382	timer_setup(&ndev->rs_timer, addrconf_rs_timer, 0);
383	memcpy(&ndev->cnf, dev_net(dev)->ipv6.devconf_dflt, sizeof(ndev->cnf));
384
385	if (ndev->cnf.stable_secret.initialized)
386		ndev->cnf.addr_gen_mode = IN6_ADDR_GEN_MODE_STABLE_PRIVACY;
387
388	ndev->cnf.mtu6 = dev->mtu;
389	ndev->nd_parms = neigh_parms_alloc(dev, &nd_tbl);
390	if (!ndev->nd_parms) {
391		kfree(ndev);
392		return ERR_PTR(err);
393	}
394	if (ndev->cnf.forwarding)
395		dev_disable_lro(dev);
396	/* We refer to the device */
397	dev_hold(dev);
398
399	if (snmp6_alloc_dev(ndev) < 0) {
400		netdev_dbg(dev, "%s: cannot allocate memory for statistics\n",
401			   __func__);
402		neigh_parms_release(&nd_tbl, ndev->nd_parms);
403		dev_put(dev);
404		kfree(ndev);
405		return ERR_PTR(err);
406	}
407
408	if (snmp6_register_dev(ndev) < 0) {
409		netdev_dbg(dev, "%s: cannot create /proc/net/dev_snmp6/%s\n",
410			   __func__, dev->name);
411		goto err_release;
412	}
413
414	/* One reference from device. */
415	refcount_set(&ndev->refcnt, 1);
416
417	if (dev->flags & (IFF_NOARP | IFF_LOOPBACK))
418		ndev->cnf.accept_dad = -1;
419
420#if IS_ENABLED(CONFIG_IPV6_SIT)
421	if (dev->type == ARPHRD_SIT && (dev->priv_flags & IFF_ISATAP)) {
422		pr_info("%s: Disabled Multicast RS\n", dev->name);
423		ndev->cnf.rtr_solicits = 0;
424	}
425#endif
426
427	INIT_LIST_HEAD(&ndev->tempaddr_list);
428	ndev->desync_factor = U32_MAX;
429	if ((dev->flags&IFF_LOOPBACK) ||
430	    dev->type == ARPHRD_TUNNEL ||
431	    dev->type == ARPHRD_TUNNEL6 ||
432	    dev->type == ARPHRD_SIT ||
433	    dev->type == ARPHRD_NONE) {
434		ndev->cnf.use_tempaddr = -1;
435	}
436
437	ndev->token = in6addr_any;
438
439	if (netif_running(dev) && addrconf_link_ready(dev))
440		ndev->if_flags |= IF_READY;
441
442	ipv6_mc_init_dev(ndev);
443	ndev->tstamp = jiffies;
444	err = addrconf_sysctl_register(ndev);
445	if (err) {
446		ipv6_mc_destroy_dev(ndev);
447		snmp6_unregister_dev(ndev);
448		goto err_release;
449	}
450	/* protected by rtnl_lock */
451	rcu_assign_pointer(dev->ip6_ptr, ndev);
452
453	/* Join interface-local all-node multicast group */
454	ipv6_dev_mc_inc(dev, &in6addr_interfacelocal_allnodes);
455
456	/* Join all-node multicast group */
457	ipv6_dev_mc_inc(dev, &in6addr_linklocal_allnodes);
458
459	/* Join all-router multicast group if forwarding is set */
460	if (ndev->cnf.forwarding && (dev->flags & IFF_MULTICAST))
461		ipv6_dev_mc_inc(dev, &in6addr_linklocal_allrouters);
462
463	return ndev;
464
465err_release:
466	neigh_parms_release(&nd_tbl, ndev->nd_parms);
467	ndev->dead = 1;
468	in6_dev_finish_destroy(ndev);
469	return ERR_PTR(err);
470}
471
472static struct inet6_dev *ipv6_find_idev(struct net_device *dev)
473{
474	struct inet6_dev *idev;
475
476	ASSERT_RTNL();
477
478	idev = __in6_dev_get(dev);
479	if (!idev) {
480		idev = ipv6_add_dev(dev);
481		if (IS_ERR(idev))
482			return idev;
483	}
484
485	if (dev->flags&IFF_UP)
486		ipv6_mc_up(idev);
487	return idev;
488}
489
490static int inet6_netconf_msgsize_devconf(int type)
491{
492	int size =  NLMSG_ALIGN(sizeof(struct netconfmsg))
493		    + nla_total_size(4);	/* NETCONFA_IFINDEX */
494	bool all = false;
495
496	if (type == NETCONFA_ALL)
497		all = true;
498
499	if (all || type == NETCONFA_FORWARDING)
500		size += nla_total_size(4);
501#ifdef CONFIG_IPV6_MROUTE
502	if (all || type == NETCONFA_MC_FORWARDING)
503		size += nla_total_size(4);
504#endif
505	if (all || type == NETCONFA_PROXY_NEIGH)
506		size += nla_total_size(4);
507
508	if (all || type == NETCONFA_IGNORE_ROUTES_WITH_LINKDOWN)
509		size += nla_total_size(4);
510
511	return size;
512}
513
514static int inet6_netconf_fill_devconf(struct sk_buff *skb, int ifindex,
515				      struct ipv6_devconf *devconf, u32 portid,
516				      u32 seq, int event, unsigned int flags,
517				      int type)
518{
519	struct nlmsghdr  *nlh;
520	struct netconfmsg *ncm;
521	bool all = false;
522
523	nlh = nlmsg_put(skb, portid, seq, event, sizeof(struct netconfmsg),
524			flags);
525	if (!nlh)
526		return -EMSGSIZE;
527
528	if (type == NETCONFA_ALL)
529		all = true;
530
531	ncm = nlmsg_data(nlh);
532	ncm->ncm_family = AF_INET6;
533
534	if (nla_put_s32(skb, NETCONFA_IFINDEX, ifindex) < 0)
535		goto nla_put_failure;
536
537	if (!devconf)
538		goto out;
539
540	if ((all || type == NETCONFA_FORWARDING) &&
541	    nla_put_s32(skb, NETCONFA_FORWARDING, devconf->forwarding) < 0)
542		goto nla_put_failure;
543#ifdef CONFIG_IPV6_MROUTE
544	if ((all || type == NETCONFA_MC_FORWARDING) &&
545	    nla_put_s32(skb, NETCONFA_MC_FORWARDING,
546			atomic_read(&devconf->mc_forwarding)) < 0)
547		goto nla_put_failure;
548#endif
549	if ((all || type == NETCONFA_PROXY_NEIGH) &&
550	    nla_put_s32(skb, NETCONFA_PROXY_NEIGH, devconf->proxy_ndp) < 0)
551		goto nla_put_failure;
552
553	if ((all || type == NETCONFA_IGNORE_ROUTES_WITH_LINKDOWN) &&
554	    nla_put_s32(skb, NETCONFA_IGNORE_ROUTES_WITH_LINKDOWN,
555			devconf->ignore_routes_with_linkdown) < 0)
556		goto nla_put_failure;
557
558out:
559	nlmsg_end(skb, nlh);
560	return 0;
561
562nla_put_failure:
563	nlmsg_cancel(skb, nlh);
564	return -EMSGSIZE;
565}
566
567void inet6_netconf_notify_devconf(struct net *net, int event, int type,
568				  int ifindex, struct ipv6_devconf *devconf)
569{
570	struct sk_buff *skb;
571	int err = -ENOBUFS;
572
573	skb = nlmsg_new(inet6_netconf_msgsize_devconf(type), GFP_KERNEL);
574	if (!skb)
575		goto errout;
576
577	err = inet6_netconf_fill_devconf(skb, ifindex, devconf, 0, 0,
578					 event, 0, type);
579	if (err < 0) {
580		/* -EMSGSIZE implies BUG in inet6_netconf_msgsize_devconf() */
581		WARN_ON(err == -EMSGSIZE);
582		kfree_skb(skb);
583		goto errout;
584	}
585	rtnl_notify(skb, net, 0, RTNLGRP_IPV6_NETCONF, NULL, GFP_KERNEL);
586	return;
587errout:
588	rtnl_set_sk_err(net, RTNLGRP_IPV6_NETCONF, err);
589}
590
591static const struct nla_policy devconf_ipv6_policy[NETCONFA_MAX+1] = {
592	[NETCONFA_IFINDEX]	= { .len = sizeof(int) },
593	[NETCONFA_FORWARDING]	= { .len = sizeof(int) },
594	[NETCONFA_PROXY_NEIGH]	= { .len = sizeof(int) },
595	[NETCONFA_IGNORE_ROUTES_WITH_LINKDOWN]	= { .len = sizeof(int) },
596};
597
598static int inet6_netconf_valid_get_req(struct sk_buff *skb,
599				       const struct nlmsghdr *nlh,
600				       struct nlattr **tb,
601				       struct netlink_ext_ack *extack)
602{
603	int i, err;
604
605	if (nlh->nlmsg_len < nlmsg_msg_size(sizeof(struct netconfmsg))) {
606		NL_SET_ERR_MSG_MOD(extack, "Invalid header for netconf get request");
607		return -EINVAL;
608	}
609
610	if (!netlink_strict_get_check(skb))
611		return nlmsg_parse_deprecated(nlh, sizeof(struct netconfmsg),
612					      tb, NETCONFA_MAX,
613					      devconf_ipv6_policy, extack);
614
615	err = nlmsg_parse_deprecated_strict(nlh, sizeof(struct netconfmsg),
616					    tb, NETCONFA_MAX,
617					    devconf_ipv6_policy, extack);
618	if (err)
619		return err;
620
621	for (i = 0; i <= NETCONFA_MAX; i++) {
622		if (!tb[i])
623			continue;
624
625		switch (i) {
626		case NETCONFA_IFINDEX:
627			break;
628		default:
629			NL_SET_ERR_MSG_MOD(extack, "Unsupported attribute in netconf get request");
630			return -EINVAL;
631		}
632	}
633
634	return 0;
635}
636
637static int inet6_netconf_get_devconf(struct sk_buff *in_skb,
638				     struct nlmsghdr *nlh,
639				     struct netlink_ext_ack *extack)
640{
641	struct net *net = sock_net(in_skb->sk);
642	struct nlattr *tb[NETCONFA_MAX+1];
643	struct inet6_dev *in6_dev = NULL;
644	struct net_device *dev = NULL;
645	struct sk_buff *skb;
646	struct ipv6_devconf *devconf;
647	int ifindex;
648	int err;
649
650	err = inet6_netconf_valid_get_req(in_skb, nlh, tb, extack);
651	if (err < 0)
652		return err;
653
654	if (!tb[NETCONFA_IFINDEX])
655		return -EINVAL;
656
657	err = -EINVAL;
658	ifindex = nla_get_s32(tb[NETCONFA_IFINDEX]);
659	switch (ifindex) {
660	case NETCONFA_IFINDEX_ALL:
661		devconf = net->ipv6.devconf_all;
662		break;
663	case NETCONFA_IFINDEX_DEFAULT:
664		devconf = net->ipv6.devconf_dflt;
665		break;
666	default:
667		dev = dev_get_by_index(net, ifindex);
668		if (!dev)
669			return -EINVAL;
670		in6_dev = in6_dev_get(dev);
671		if (!in6_dev)
672			goto errout;
673		devconf = &in6_dev->cnf;
674		break;
675	}
676
677	err = -ENOBUFS;
678	skb = nlmsg_new(inet6_netconf_msgsize_devconf(NETCONFA_ALL), GFP_KERNEL);
679	if (!skb)
680		goto errout;
681
682	err = inet6_netconf_fill_devconf(skb, ifindex, devconf,
683					 NETLINK_CB(in_skb).portid,
684					 nlh->nlmsg_seq, RTM_NEWNETCONF, 0,
685					 NETCONFA_ALL);
686	if (err < 0) {
687		/* -EMSGSIZE implies BUG in inet6_netconf_msgsize_devconf() */
688		WARN_ON(err == -EMSGSIZE);
689		kfree_skb(skb);
690		goto errout;
691	}
692	err = rtnl_unicast(skb, net, NETLINK_CB(in_skb).portid);
693errout:
694	if (in6_dev)
695		in6_dev_put(in6_dev);
696	if (dev)
697		dev_put(dev);
698	return err;
699}
700
701static int inet6_netconf_dump_devconf(struct sk_buff *skb,
702				      struct netlink_callback *cb)
703{
704	const struct nlmsghdr *nlh = cb->nlh;
705	struct net *net = sock_net(skb->sk);
706	int h, s_h;
707	int idx, s_idx;
708	struct net_device *dev;
709	struct inet6_dev *idev;
710	struct hlist_head *head;
711
712	if (cb->strict_check) {
713		struct netlink_ext_ack *extack = cb->extack;
714		struct netconfmsg *ncm;
715
716		if (nlh->nlmsg_len < nlmsg_msg_size(sizeof(*ncm))) {
717			NL_SET_ERR_MSG_MOD(extack, "Invalid header for netconf dump request");
718			return -EINVAL;
719		}
720
721		if (nlmsg_attrlen(nlh, sizeof(*ncm))) {
722			NL_SET_ERR_MSG_MOD(extack, "Invalid data after header in netconf dump request");
723			return -EINVAL;
724		}
725	}
726
727	s_h = cb->args[0];
728	s_idx = idx = cb->args[1];
729
730	for (h = s_h; h < NETDEV_HASHENTRIES; h++, s_idx = 0) {
731		idx = 0;
732		head = &net->dev_index_head[h];
733		rcu_read_lock();
734		cb->seq = atomic_read(&net->ipv6.dev_addr_genid) ^
735			  net->dev_base_seq;
736		hlist_for_each_entry_rcu(dev, head, index_hlist) {
737			if (idx < s_idx)
738				goto cont;
739			idev = __in6_dev_get(dev);
740			if (!idev)
741				goto cont;
742
743			if (inet6_netconf_fill_devconf(skb, dev->ifindex,
744						       &idev->cnf,
745						       NETLINK_CB(cb->skb).portid,
746						       nlh->nlmsg_seq,
747						       RTM_NEWNETCONF,
748						       NLM_F_MULTI,
749						       NETCONFA_ALL) < 0) {
750				rcu_read_unlock();
751				goto done;
752			}
753			nl_dump_check_consistent(cb, nlmsg_hdr(skb));
754cont:
755			idx++;
756		}
757		rcu_read_unlock();
758	}
759	if (h == NETDEV_HASHENTRIES) {
760		if (inet6_netconf_fill_devconf(skb, NETCONFA_IFINDEX_ALL,
761					       net->ipv6.devconf_all,
762					       NETLINK_CB(cb->skb).portid,
763					       nlh->nlmsg_seq,
764					       RTM_NEWNETCONF, NLM_F_MULTI,
765					       NETCONFA_ALL) < 0)
766			goto done;
767		else
768			h++;
769	}
770	if (h == NETDEV_HASHENTRIES + 1) {
771		if (inet6_netconf_fill_devconf(skb, NETCONFA_IFINDEX_DEFAULT,
772					       net->ipv6.devconf_dflt,
773					       NETLINK_CB(cb->skb).portid,
774					       nlh->nlmsg_seq,
775					       RTM_NEWNETCONF, NLM_F_MULTI,
776					       NETCONFA_ALL) < 0)
777			goto done;
778		else
779			h++;
780	}
781done:
782	cb->args[0] = h;
783	cb->args[1] = idx;
784
785	return skb->len;
786}
787
788#ifdef CONFIG_SYSCTL
789static void dev_forward_change(struct inet6_dev *idev)
790{
791	struct net_device *dev;
792	struct inet6_ifaddr *ifa;
793	LIST_HEAD(tmp_addr_list);
794
795	if (!idev)
796		return;
797	dev = idev->dev;
798	if (idev->cnf.forwarding)
799		dev_disable_lro(dev);
800	if (dev->flags & IFF_MULTICAST) {
801		if (idev->cnf.forwarding) {
802			ipv6_dev_mc_inc(dev, &in6addr_linklocal_allrouters);
803			ipv6_dev_mc_inc(dev, &in6addr_interfacelocal_allrouters);
804			ipv6_dev_mc_inc(dev, &in6addr_sitelocal_allrouters);
805		} else {
806			ipv6_dev_mc_dec(dev, &in6addr_linklocal_allrouters);
807			ipv6_dev_mc_dec(dev, &in6addr_interfacelocal_allrouters);
808			ipv6_dev_mc_dec(dev, &in6addr_sitelocal_allrouters);
809		}
810	}
811
812	read_lock_bh(&idev->lock);
813	list_for_each_entry(ifa, &idev->addr_list, if_list) {
814		if (ifa->flags&IFA_F_TENTATIVE)
815			continue;
816		list_add_tail(&ifa->if_list_aux, &tmp_addr_list);
817	}
818	read_unlock_bh(&idev->lock);
819
820	while (!list_empty(&tmp_addr_list)) {
821		ifa = list_first_entry(&tmp_addr_list,
822				       struct inet6_ifaddr, if_list_aux);
823		list_del(&ifa->if_list_aux);
824		if (idev->cnf.forwarding)
825			addrconf_join_anycast(ifa);
826		else
827			addrconf_leave_anycast(ifa);
828	}
829
830	inet6_netconf_notify_devconf(dev_net(dev), RTM_NEWNETCONF,
831				     NETCONFA_FORWARDING,
832				     dev->ifindex, &idev->cnf);
833}
834
835
836static void addrconf_forward_change(struct net *net, __s32 newf)
837{
838	struct net_device *dev;
839	struct inet6_dev *idev;
840
841	for_each_netdev(net, dev) {
842		idev = __in6_dev_get(dev);
843		if (idev) {
844			int changed = (!idev->cnf.forwarding) ^ (!newf);
845			idev->cnf.forwarding = newf;
846			if (changed)
847				dev_forward_change(idev);
848		}
849	}
850}
851
852static int addrconf_fixup_forwarding(struct ctl_table *table, int *p, int newf)
853{
854	struct net *net;
855	int old;
856
857	if (!rtnl_trylock())
858		return restart_syscall();
859
860	net = (struct net *)table->extra2;
861	old = *p;
862	*p = newf;
863
864	if (p == &net->ipv6.devconf_dflt->forwarding) {
865		if ((!newf) ^ (!old))
866			inet6_netconf_notify_devconf(net, RTM_NEWNETCONF,
867						     NETCONFA_FORWARDING,
868						     NETCONFA_IFINDEX_DEFAULT,
869						     net->ipv6.devconf_dflt);
870		rtnl_unlock();
871		return 0;
872	}
873
874	if (p == &net->ipv6.devconf_all->forwarding) {
875		int old_dflt = net->ipv6.devconf_dflt->forwarding;
876
877		net->ipv6.devconf_dflt->forwarding = newf;
878		if ((!newf) ^ (!old_dflt))
879			inet6_netconf_notify_devconf(net, RTM_NEWNETCONF,
880						     NETCONFA_FORWARDING,
881						     NETCONFA_IFINDEX_DEFAULT,
882						     net->ipv6.devconf_dflt);
883
884		addrconf_forward_change(net, newf);
885		if ((!newf) ^ (!old))
886			inet6_netconf_notify_devconf(net, RTM_NEWNETCONF,
887						     NETCONFA_FORWARDING,
888						     NETCONFA_IFINDEX_ALL,
889						     net->ipv6.devconf_all);
890	} else if ((!newf) ^ (!old))
891		dev_forward_change((struct inet6_dev *)table->extra1);
892	rtnl_unlock();
893
894	if (newf)
895		rt6_purge_dflt_routers(net);
896	return 1;
897}
898
899static void addrconf_linkdown_change(struct net *net, __s32 newf)
900{
901	struct net_device *dev;
902	struct inet6_dev *idev;
903
904	for_each_netdev(net, dev) {
905		idev = __in6_dev_get(dev);
906		if (idev) {
907			int changed = (!idev->cnf.ignore_routes_with_linkdown) ^ (!newf);
908
909			idev->cnf.ignore_routes_with_linkdown = newf;
910			if (changed)
911				inet6_netconf_notify_devconf(dev_net(dev),
912							     RTM_NEWNETCONF,
913							     NETCONFA_IGNORE_ROUTES_WITH_LINKDOWN,
914							     dev->ifindex,
915							     &idev->cnf);
916		}
917	}
918}
919
920static int addrconf_fixup_linkdown(struct ctl_table *table, int *p, int newf)
921{
922	struct net *net;
923	int old;
924
925	if (!rtnl_trylock())
926		return restart_syscall();
927
928	net = (struct net *)table->extra2;
929	old = *p;
930	*p = newf;
931
932	if (p == &net->ipv6.devconf_dflt->ignore_routes_with_linkdown) {
933		if ((!newf) ^ (!old))
934			inet6_netconf_notify_devconf(net,
935						     RTM_NEWNETCONF,
936						     NETCONFA_IGNORE_ROUTES_WITH_LINKDOWN,
937						     NETCONFA_IFINDEX_DEFAULT,
938						     net->ipv6.devconf_dflt);
939		rtnl_unlock();
940		return 0;
941	}
942
943	if (p == &net->ipv6.devconf_all->ignore_routes_with_linkdown) {
944		net->ipv6.devconf_dflt->ignore_routes_with_linkdown = newf;
945		addrconf_linkdown_change(net, newf);
946		if ((!newf) ^ (!old))
947			inet6_netconf_notify_devconf(net,
948						     RTM_NEWNETCONF,
949						     NETCONFA_IGNORE_ROUTES_WITH_LINKDOWN,
950						     NETCONFA_IFINDEX_ALL,
951						     net->ipv6.devconf_all);
952	}
953	rtnl_unlock();
954
955	return 1;
956}
957
958#endif
959
960/* Nobody refers to this ifaddr, destroy it */
961void inet6_ifa_finish_destroy(struct inet6_ifaddr *ifp)
962{
963	WARN_ON(!hlist_unhashed(&ifp->addr_lst));
964
965#ifdef NET_REFCNT_DEBUG
966	pr_debug("%s\n", __func__);
967#endif
968
969	in6_dev_put(ifp->idev);
970
971	if (cancel_delayed_work(&ifp->dad_work))
972		pr_notice("delayed DAD work was pending while freeing ifa=%p\n",
973			  ifp);
974
975	if (ifp->state != INET6_IFADDR_STATE_DEAD) {
976		pr_warn("Freeing alive inet6 address %p\n", ifp);
977		return;
978	}
979
980	kfree_rcu(ifp, rcu);
981}
982
983static void
984ipv6_link_dev_addr(struct inet6_dev *idev, struct inet6_ifaddr *ifp)
985{
986	struct list_head *p;
987	int ifp_scope = ipv6_addr_src_scope(&ifp->addr);
988
989	/*
990	 * Each device address list is sorted in order of scope -
991	 * global before linklocal.
992	 */
993	list_for_each(p, &idev->addr_list) {
994		struct inet6_ifaddr *ifa
995			= list_entry(p, struct inet6_ifaddr, if_list);
996		if (ifp_scope >= ipv6_addr_src_scope(&ifa->addr))
997			break;
998	}
999
1000	list_add_tail_rcu(&ifp->if_list, p);
1001}
1002
1003static u32 inet6_addr_hash(const struct net *net, const struct in6_addr *addr)
1004{
1005	u32 val = ipv6_addr_hash(addr) ^ net_hash_mix(net);
1006
1007	return hash_32(val, IN6_ADDR_HSIZE_SHIFT);
1008}
1009
1010static bool ipv6_chk_same_addr(struct net *net, const struct in6_addr *addr,
1011			       struct net_device *dev, unsigned int hash)
1012{
1013	struct inet6_ifaddr *ifp;
1014
1015	hlist_for_each_entry(ifp, &inet6_addr_lst[hash], addr_lst) {
1016		if (!net_eq(dev_net(ifp->idev->dev), net))
1017			continue;
1018		if (ipv6_addr_equal(&ifp->addr, addr)) {
1019			if (!dev || ifp->idev->dev == dev)
1020				return true;
1021		}
1022	}
1023	return false;
1024}
1025
1026static int ipv6_add_addr_hash(struct net_device *dev, struct inet6_ifaddr *ifa)
1027{
1028	unsigned int hash = inet6_addr_hash(dev_net(dev), &ifa->addr);
1029	int err = 0;
1030
1031	spin_lock(&addrconf_hash_lock);
1032
1033	/* Ignore adding duplicate addresses on an interface */
1034	if (ipv6_chk_same_addr(dev_net(dev), &ifa->addr, dev, hash)) {
1035		netdev_dbg(dev, "ipv6_add_addr: already assigned\n");
1036		err = -EEXIST;
1037	} else {
1038		hlist_add_head_rcu(&ifa->addr_lst, &inet6_addr_lst[hash]);
1039	}
1040
1041	spin_unlock(&addrconf_hash_lock);
1042
1043	return err;
1044}
1045
1046/* On success it returns ifp with increased reference count */
1047
1048static struct inet6_ifaddr *
1049ipv6_add_addr(struct inet6_dev *idev, struct ifa6_config *cfg,
1050	      bool can_block, struct netlink_ext_ack *extack)
1051{
1052	gfp_t gfp_flags = can_block ? GFP_KERNEL : GFP_ATOMIC;
1053	int addr_type = ipv6_addr_type(cfg->pfx);
1054	struct net *net = dev_net(idev->dev);
1055	struct inet6_ifaddr *ifa = NULL;
1056	struct fib6_info *f6i = NULL;
1057	int err = 0;
1058
1059	if (addr_type == IPV6_ADDR_ANY ||
1060	    (addr_type & IPV6_ADDR_MULTICAST &&
1061	     !(cfg->ifa_flags & IFA_F_MCAUTOJOIN)) ||
1062	    (!(idev->dev->flags & IFF_LOOPBACK) &&
1063	     !netif_is_l3_master(idev->dev) &&
1064	     addr_type & IPV6_ADDR_LOOPBACK))
1065		return ERR_PTR(-EADDRNOTAVAIL);
1066
1067	if (idev->dead) {
1068		err = -ENODEV;			/*XXX*/
1069		goto out;
1070	}
1071
1072	if (idev->cnf.disable_ipv6) {
1073		err = -EACCES;
1074		goto out;
1075	}
1076
1077	/* validator notifier needs to be blocking;
1078	 * do not call in atomic context
1079	 */
1080	if (can_block) {
1081		struct in6_validator_info i6vi = {
1082			.i6vi_addr = *cfg->pfx,
1083			.i6vi_dev = idev,
1084			.extack = extack,
1085		};
1086
1087		err = inet6addr_validator_notifier_call_chain(NETDEV_UP, &i6vi);
1088		err = notifier_to_errno(err);
1089		if (err < 0)
1090			goto out;
1091	}
1092
1093	ifa = kzalloc(sizeof(*ifa), gfp_flags);
1094	if (!ifa) {
1095		err = -ENOBUFS;
1096		goto out;
1097	}
1098
1099	f6i = addrconf_f6i_alloc(net, idev, cfg->pfx, false, gfp_flags);
1100	if (IS_ERR(f6i)) {
1101		err = PTR_ERR(f6i);
1102		f6i = NULL;
1103		goto out;
1104	}
1105
1106	neigh_parms_data_state_setall(idev->nd_parms);
1107
1108	ifa->addr = *cfg->pfx;
1109	if (cfg->peer_pfx)
1110		ifa->peer_addr = *cfg->peer_pfx;
1111
1112	spin_lock_init(&ifa->lock);
1113	INIT_DELAYED_WORK(&ifa->dad_work, addrconf_dad_work);
1114	INIT_HLIST_NODE(&ifa->addr_lst);
1115	ifa->scope = cfg->scope;
1116	ifa->prefix_len = cfg->plen;
1117	ifa->rt_priority = cfg->rt_priority;
1118	ifa->flags = cfg->ifa_flags;
1119	/* No need to add the TENTATIVE flag for addresses with NODAD */
1120	if (!(cfg->ifa_flags & IFA_F_NODAD))
1121		ifa->flags |= IFA_F_TENTATIVE;
1122	ifa->valid_lft = cfg->valid_lft;
1123	ifa->prefered_lft = cfg->preferred_lft;
1124	ifa->cstamp = ifa->tstamp = jiffies;
1125	ifa->tokenized = false;
1126
1127	ifa->rt = f6i;
1128
1129	ifa->idev = idev;
1130	in6_dev_hold(idev);
1131
1132	/* For caller */
1133	refcount_set(&ifa->refcnt, 1);
1134
1135	rcu_read_lock_bh();
1136
1137	err = ipv6_add_addr_hash(idev->dev, ifa);
1138	if (err < 0) {
1139		rcu_read_unlock_bh();
1140		goto out;
1141	}
1142
1143	write_lock(&idev->lock);
1144
1145	/* Add to inet6_dev unicast addr list. */
1146	ipv6_link_dev_addr(idev, ifa);
1147
1148	if (ifa->flags&IFA_F_TEMPORARY) {
1149		list_add(&ifa->tmp_list, &idev->tempaddr_list);
1150		in6_ifa_hold(ifa);
1151	}
1152
1153	in6_ifa_hold(ifa);
1154	write_unlock(&idev->lock);
1155
1156	rcu_read_unlock_bh();
1157
1158	inet6addr_notifier_call_chain(NETDEV_UP, ifa);
1159out:
1160	if (unlikely(err < 0)) {
1161		fib6_info_release(f6i);
1162
1163		if (ifa) {
1164			if (ifa->idev)
1165				in6_dev_put(ifa->idev);
1166			kfree(ifa);
1167		}
1168		ifa = ERR_PTR(err);
1169	}
1170
1171	return ifa;
1172}
1173
1174enum cleanup_prefix_rt_t {
1175	CLEANUP_PREFIX_RT_NOP,    /* no cleanup action for prefix route */
1176	CLEANUP_PREFIX_RT_DEL,    /* delete the prefix route */
1177	CLEANUP_PREFIX_RT_EXPIRE, /* update the lifetime of the prefix route */
1178};
1179
1180/*
1181 * Check, whether the prefix for ifp would still need a prefix route
1182 * after deleting ifp. The function returns one of the CLEANUP_PREFIX_RT_*
1183 * constants.
1184 *
1185 * 1) we don't purge prefix if address was not permanent.
1186 *    prefix is managed by its own lifetime.
1187 * 2) we also don't purge, if the address was IFA_F_NOPREFIXROUTE.
1188 * 3) if there are no addresses, delete prefix.
1189 * 4) if there are still other permanent address(es),
1190 *    corresponding prefix is still permanent.
1191 * 5) if there are still other addresses with IFA_F_NOPREFIXROUTE,
1192 *    don't purge the prefix, assume user space is managing it.
1193 * 6) otherwise, update prefix lifetime to the
1194 *    longest valid lifetime among the corresponding
1195 *    addresses on the device.
1196 *    Note: subsequent RA will update lifetime.
1197 **/
1198static enum cleanup_prefix_rt_t
1199check_cleanup_prefix_route(struct inet6_ifaddr *ifp, unsigned long *expires)
1200{
1201	struct inet6_ifaddr *ifa;
1202	struct inet6_dev *idev = ifp->idev;
1203	unsigned long lifetime;
1204	enum cleanup_prefix_rt_t action = CLEANUP_PREFIX_RT_DEL;
1205
1206	*expires = jiffies;
1207
1208	list_for_each_entry(ifa, &idev->addr_list, if_list) {
1209		if (ifa == ifp)
1210			continue;
1211		if (ifa->prefix_len != ifp->prefix_len ||
1212		    !ipv6_prefix_equal(&ifa->addr, &ifp->addr,
1213				       ifp->prefix_len))
1214			continue;
1215		if (ifa->flags & (IFA_F_PERMANENT | IFA_F_NOPREFIXROUTE))
1216			return CLEANUP_PREFIX_RT_NOP;
1217
1218		action = CLEANUP_PREFIX_RT_EXPIRE;
1219
1220		spin_lock(&ifa->lock);
1221
1222		lifetime = addrconf_timeout_fixup(ifa->valid_lft, HZ);
1223		/*
1224		 * Note: Because this address is
1225		 * not permanent, lifetime <
1226		 * LONG_MAX / HZ here.
1227		 */
1228		if (time_before(*expires, ifa->tstamp + lifetime * HZ))
1229			*expires = ifa->tstamp + lifetime * HZ;
1230		spin_unlock(&ifa->lock);
1231	}
1232
1233	return action;
1234}
1235
1236static void
1237cleanup_prefix_route(struct inet6_ifaddr *ifp, unsigned long expires,
1238		     bool del_rt, bool del_peer)
1239{
1240	struct fib6_info *f6i;
1241
1242	f6i = addrconf_get_prefix_route(del_peer ? &ifp->peer_addr : &ifp->addr,
1243					ifp->prefix_len,
1244					ifp->idev->dev, 0, RTF_DEFAULT, true);
1245	if (f6i) {
1246		if (del_rt)
1247			ip6_del_rt(dev_net(ifp->idev->dev), f6i, false);
1248		else {
1249			if (!(f6i->fib6_flags & RTF_EXPIRES))
1250				fib6_set_expires(f6i, expires);
1251			fib6_info_release(f6i);
1252		}
1253	}
1254}
1255
1256
1257/* This function wants to get referenced ifp and releases it before return */
1258
1259static void ipv6_del_addr(struct inet6_ifaddr *ifp)
1260{
1261	int state;
1262	enum cleanup_prefix_rt_t action = CLEANUP_PREFIX_RT_NOP;
1263	unsigned long expires;
1264
1265	ASSERT_RTNL();
1266
1267	spin_lock_bh(&ifp->lock);
1268	state = ifp->state;
1269	ifp->state = INET6_IFADDR_STATE_DEAD;
1270	spin_unlock_bh(&ifp->lock);
1271
1272	if (state == INET6_IFADDR_STATE_DEAD)
1273		goto out;
1274
1275	spin_lock_bh(&addrconf_hash_lock);
1276	hlist_del_init_rcu(&ifp->addr_lst);
1277	spin_unlock_bh(&addrconf_hash_lock);
1278
1279	write_lock_bh(&ifp->idev->lock);
1280
1281	if (ifp->flags&IFA_F_TEMPORARY) {
1282		list_del(&ifp->tmp_list);
1283		if (ifp->ifpub) {
1284			in6_ifa_put(ifp->ifpub);
1285			ifp->ifpub = NULL;
1286		}
1287		__in6_ifa_put(ifp);
1288	}
1289
1290	if (ifp->flags & IFA_F_PERMANENT && !(ifp->flags & IFA_F_NOPREFIXROUTE))
1291		action = check_cleanup_prefix_route(ifp, &expires);
1292
1293	list_del_rcu(&ifp->if_list);
1294	__in6_ifa_put(ifp);
1295
1296	write_unlock_bh(&ifp->idev->lock);
1297
1298	addrconf_del_dad_work(ifp);
1299
1300	ipv6_ifa_notify(RTM_DELADDR, ifp);
1301
1302	inet6addr_notifier_call_chain(NETDEV_DOWN, ifp);
1303
1304	if (action != CLEANUP_PREFIX_RT_NOP) {
1305		cleanup_prefix_route(ifp, expires,
1306			action == CLEANUP_PREFIX_RT_DEL, false);
1307	}
1308
1309	/* clean up prefsrc entries */
1310	rt6_remove_prefsrc(ifp);
1311out:
1312	in6_ifa_put(ifp);
1313}
1314
1315static int ipv6_create_tempaddr(struct inet6_ifaddr *ifp, bool block)
1316{
1317	struct inet6_dev *idev = ifp->idev;
1318	unsigned long tmp_tstamp, age;
1319	unsigned long regen_advance;
1320	unsigned long now = jiffies;
1321	s32 cnf_temp_preferred_lft;
1322	struct inet6_ifaddr *ift;
1323	struct ifa6_config cfg;
1324	long max_desync_factor;
1325	struct in6_addr addr;
1326	int ret = 0;
1327
1328	write_lock_bh(&idev->lock);
1329
1330retry:
1331	in6_dev_hold(idev);
1332	if (idev->cnf.use_tempaddr <= 0) {
1333		write_unlock_bh(&idev->lock);
1334		pr_info("%s: use_tempaddr is disabled\n", __func__);
1335		in6_dev_put(idev);
1336		ret = -1;
1337		goto out;
1338	}
1339	spin_lock_bh(&ifp->lock);
1340	if (ifp->regen_count++ >= idev->cnf.regen_max_retry) {
1341		idev->cnf.use_tempaddr = -1;	/*XXX*/
1342		spin_unlock_bh(&ifp->lock);
1343		write_unlock_bh(&idev->lock);
1344		pr_warn("%s: regeneration time exceeded - disabled temporary address support\n",
1345			__func__);
1346		in6_dev_put(idev);
1347		ret = -1;
1348		goto out;
1349	}
1350	in6_ifa_hold(ifp);
1351	memcpy(addr.s6_addr, ifp->addr.s6_addr, 8);
1352	ipv6_gen_rnd_iid(&addr);
1353
1354	age = (now - ifp->tstamp) / HZ;
1355
1356	regen_advance = idev->cnf.regen_max_retry *
1357			idev->cnf.dad_transmits *
1358			max(NEIGH_VAR(idev->nd_parms, RETRANS_TIME), HZ/100) / HZ;
1359
1360	/* recalculate max_desync_factor each time and update
1361	 * idev->desync_factor if it's larger
1362	 */
1363	cnf_temp_preferred_lft = READ_ONCE(idev->cnf.temp_prefered_lft);
1364	max_desync_factor = min_t(long,
1365				  idev->cnf.max_desync_factor,
1366				  cnf_temp_preferred_lft - regen_advance);
1367
1368	if (unlikely(idev->desync_factor > max_desync_factor)) {
1369		if (max_desync_factor > 0) {
1370			get_random_bytes(&idev->desync_factor,
1371					 sizeof(idev->desync_factor));
1372			idev->desync_factor %= max_desync_factor;
1373		} else {
1374			idev->desync_factor = 0;
1375		}
1376	}
1377
1378	memset(&cfg, 0, sizeof(cfg));
1379	cfg.valid_lft = min_t(__u32, ifp->valid_lft,
1380			      idev->cnf.temp_valid_lft + age);
1381	cfg.preferred_lft = cnf_temp_preferred_lft + age - idev->desync_factor;
1382	cfg.preferred_lft = min_t(__u32, ifp->prefered_lft, cfg.preferred_lft);
1383
1384	cfg.plen = ifp->prefix_len;
1385	tmp_tstamp = ifp->tstamp;
1386	spin_unlock_bh(&ifp->lock);
1387
1388	write_unlock_bh(&idev->lock);
1389
1390	/* A temporary address is created only if this calculated Preferred
1391	 * Lifetime is greater than REGEN_ADVANCE time units.  In particular,
1392	 * an implementation must not create a temporary address with a zero
1393	 * Preferred Lifetime.
1394	 * Use age calculation as in addrconf_verify to avoid unnecessary
1395	 * temporary addresses being generated.
1396	 */
1397	age = (now - tmp_tstamp + ADDRCONF_TIMER_FUZZ_MINUS) / HZ;
1398	if (cfg.preferred_lft <= regen_advance + age) {
1399		in6_ifa_put(ifp);
1400		in6_dev_put(idev);
1401		ret = -1;
1402		goto out;
1403	}
1404
1405	cfg.ifa_flags = IFA_F_TEMPORARY;
1406	/* set in addrconf_prefix_rcv() */
1407	if (ifp->flags & IFA_F_OPTIMISTIC)
1408		cfg.ifa_flags |= IFA_F_OPTIMISTIC;
1409
1410	cfg.pfx = &addr;
1411	cfg.scope = ipv6_addr_scope(cfg.pfx);
1412
1413	ift = ipv6_add_addr(idev, &cfg, block, NULL);
1414	if (IS_ERR(ift)) {
1415		in6_ifa_put(ifp);
1416		in6_dev_put(idev);
1417		pr_info("%s: retry temporary address regeneration\n", __func__);
1418		write_lock_bh(&idev->lock);
1419		goto retry;
1420	}
1421
1422	spin_lock_bh(&ift->lock);
1423	ift->ifpub = ifp;
1424	ift->cstamp = now;
1425	ift->tstamp = tmp_tstamp;
1426	spin_unlock_bh(&ift->lock);
1427
1428	addrconf_dad_start(ift);
1429	in6_ifa_put(ift);
1430	in6_dev_put(idev);
1431out:
1432	return ret;
1433}
1434
1435/*
1436 *	Choose an appropriate source address (RFC3484)
1437 */
1438enum {
1439	IPV6_SADDR_RULE_INIT = 0,
1440	IPV6_SADDR_RULE_LOCAL,
1441	IPV6_SADDR_RULE_SCOPE,
1442	IPV6_SADDR_RULE_PREFERRED,
1443#ifdef CONFIG_IPV6_MIP6
1444	IPV6_SADDR_RULE_HOA,
1445#endif
1446	IPV6_SADDR_RULE_OIF,
1447	IPV6_SADDR_RULE_LABEL,
1448	IPV6_SADDR_RULE_PRIVACY,
1449	IPV6_SADDR_RULE_ORCHID,
1450	IPV6_SADDR_RULE_PREFIX,
1451#ifdef CONFIG_IPV6_OPTIMISTIC_DAD
1452	IPV6_SADDR_RULE_NOT_OPTIMISTIC,
1453#endif
1454	IPV6_SADDR_RULE_MAX
1455};
1456
1457struct ipv6_saddr_score {
1458	int			rule;
1459	int			addr_type;
1460	struct inet6_ifaddr	*ifa;
1461	DECLARE_BITMAP(scorebits, IPV6_SADDR_RULE_MAX);
1462	int			scopedist;
1463	int			matchlen;
1464};
1465
1466struct ipv6_saddr_dst {
1467	const struct in6_addr *addr;
1468	int ifindex;
1469	int scope;
1470	int label;
1471	unsigned int prefs;
1472};
1473
1474static inline int ipv6_saddr_preferred(int type)
1475{
1476	if (type & (IPV6_ADDR_MAPPED|IPV6_ADDR_COMPATv4|IPV6_ADDR_LOOPBACK))
1477		return 1;
1478	return 0;
1479}
1480
1481static bool ipv6_use_optimistic_addr(struct net *net,
1482				     struct inet6_dev *idev)
1483{
1484#ifdef CONFIG_IPV6_OPTIMISTIC_DAD
1485	if (!idev)
1486		return false;
1487	if (!net->ipv6.devconf_all->optimistic_dad && !idev->cnf.optimistic_dad)
1488		return false;
1489	if (!net->ipv6.devconf_all->use_optimistic && !idev->cnf.use_optimistic)
1490		return false;
1491
1492	return true;
1493#else
1494	return false;
1495#endif
1496}
1497
1498static bool ipv6_allow_optimistic_dad(struct net *net,
1499				      struct inet6_dev *idev)
1500{
1501#ifdef CONFIG_IPV6_OPTIMISTIC_DAD
1502	if (!idev)
1503		return false;
1504	if (!net->ipv6.devconf_all->optimistic_dad && !idev->cnf.optimistic_dad)
1505		return false;
1506
1507	return true;
1508#else
1509	return false;
1510#endif
1511}
1512
1513static int ipv6_get_saddr_eval(struct net *net,
1514			       struct ipv6_saddr_score *score,
1515			       struct ipv6_saddr_dst *dst,
1516			       int i)
1517{
1518	int ret;
1519
1520	if (i <= score->rule) {
1521		switch (i) {
1522		case IPV6_SADDR_RULE_SCOPE:
1523			ret = score->scopedist;
1524			break;
1525		case IPV6_SADDR_RULE_PREFIX:
1526			ret = score->matchlen;
1527			break;
1528		default:
1529			ret = !!test_bit(i, score->scorebits);
1530		}
1531		goto out;
1532	}
1533
1534	switch (i) {
1535	case IPV6_SADDR_RULE_INIT:
1536		/* Rule 0: remember if hiscore is not ready yet */
1537		ret = !!score->ifa;
1538		break;
1539	case IPV6_SADDR_RULE_LOCAL:
1540		/* Rule 1: Prefer same address */
1541		ret = ipv6_addr_equal(&score->ifa->addr, dst->addr);
1542		break;
1543	case IPV6_SADDR_RULE_SCOPE:
1544		/* Rule 2: Prefer appropriate scope
1545		 *
1546		 *      ret
1547		 *       ^
1548		 *    -1 |  d 15
1549		 *    ---+--+-+---> scope
1550		 *       |
1551		 *       |             d is scope of the destination.
1552		 *  B-d  |  \
1553		 *       |   \      <- smaller scope is better if
1554		 *  B-15 |    \        if scope is enough for destination.
1555		 *       |             ret = B - scope (-1 <= scope >= d <= 15).
1556		 * d-C-1 | /
1557		 *       |/         <- greater is better
1558		 *   -C  /             if scope is not enough for destination.
1559		 *      /|             ret = scope - C (-1 <= d < scope <= 15).
1560		 *
1561		 * d - C - 1 < B -15 (for all -1 <= d <= 15).
1562		 * C > d + 14 - B >= 15 + 14 - B = 29 - B.
1563		 * Assume B = 0 and we get C > 29.
1564		 */
1565		ret = __ipv6_addr_src_scope(score->addr_type);
1566		if (ret >= dst->scope)
1567			ret = -ret;
1568		else
1569			ret -= 128;	/* 30 is enough */
1570		score->scopedist = ret;
1571		break;
1572	case IPV6_SADDR_RULE_PREFERRED:
1573	    {
1574		/* Rule 3: Avoid deprecated and optimistic addresses */
1575		u8 avoid = IFA_F_DEPRECATED;
1576
1577		if (!ipv6_use_optimistic_addr(net, score->ifa->idev))
1578			avoid |= IFA_F_OPTIMISTIC;
1579		ret = ipv6_saddr_preferred(score->addr_type) ||
1580		      !(score->ifa->flags & avoid);
1581		break;
1582	    }
1583#ifdef CONFIG_IPV6_MIP6
1584	case IPV6_SADDR_RULE_HOA:
1585	    {
1586		/* Rule 4: Prefer home address */
1587		int prefhome = !(dst->prefs & IPV6_PREFER_SRC_COA);
1588		ret = !(score->ifa->flags & IFA_F_HOMEADDRESS) ^ prefhome;
1589		break;
1590	    }
1591#endif
1592	case IPV6_SADDR_RULE_OIF:
1593		/* Rule 5: Prefer outgoing interface */
1594		ret = (!dst->ifindex ||
1595		       dst->ifindex == score->ifa->idev->dev->ifindex);
1596		break;
1597	case IPV6_SADDR_RULE_LABEL:
1598		/* Rule 6: Prefer matching label */
1599		ret = ipv6_addr_label(net,
1600				      &score->ifa->addr, score->addr_type,
1601				      score->ifa->idev->dev->ifindex) == dst->label;
1602		break;
1603	case IPV6_SADDR_RULE_PRIVACY:
1604	    {
1605		/* Rule 7: Prefer public address
1606		 * Note: prefer temporary address if use_tempaddr >= 2
1607		 */
1608		int preftmp = dst->prefs & (IPV6_PREFER_SRC_PUBLIC|IPV6_PREFER_SRC_TMP) ?
1609				!!(dst->prefs & IPV6_PREFER_SRC_TMP) :
1610				score->ifa->idev->cnf.use_tempaddr >= 2;
1611		ret = (!(score->ifa->flags & IFA_F_TEMPORARY)) ^ preftmp;
1612		break;
1613	    }
1614	case IPV6_SADDR_RULE_ORCHID:
1615		/* Rule 8-: Prefer ORCHID vs ORCHID or
1616		 *	    non-ORCHID vs non-ORCHID
1617		 */
1618		ret = !(ipv6_addr_orchid(&score->ifa->addr) ^
1619			ipv6_addr_orchid(dst->addr));
1620		break;
1621	case IPV6_SADDR_RULE_PREFIX:
1622		/* Rule 8: Use longest matching prefix */
1623		ret = ipv6_addr_diff(&score->ifa->addr, dst->addr);
1624		if (ret > score->ifa->prefix_len)
1625			ret = score->ifa->prefix_len;
1626		score->matchlen = ret;
1627		break;
1628#ifdef CONFIG_IPV6_OPTIMISTIC_DAD
1629	case IPV6_SADDR_RULE_NOT_OPTIMISTIC:
1630		/* Optimistic addresses still have lower precedence than other
1631		 * preferred addresses.
1632		 */
1633		ret = !(score->ifa->flags & IFA_F_OPTIMISTIC);
1634		break;
1635#endif
1636	default:
1637		ret = 0;
1638	}
1639
1640	if (ret)
1641		__set_bit(i, score->scorebits);
1642	score->rule = i;
1643out:
1644	return ret;
1645}
1646
1647static int __ipv6_dev_get_saddr(struct net *net,
1648				struct ipv6_saddr_dst *dst,
1649				struct inet6_dev *idev,
1650				struct ipv6_saddr_score *scores,
1651				int hiscore_idx)
1652{
1653	struct ipv6_saddr_score *score = &scores[1 - hiscore_idx], *hiscore = &scores[hiscore_idx];
1654
1655	list_for_each_entry_rcu(score->ifa, &idev->addr_list, if_list) {
1656		int i;
1657
1658		/*
1659		 * - Tentative Address (RFC2462 section 5.4)
1660		 *  - A tentative address is not considered
1661		 *    "assigned to an interface" in the traditional
1662		 *    sense, unless it is also flagged as optimistic.
1663		 * - Candidate Source Address (section 4)
1664		 *  - In any case, anycast addresses, multicast
1665		 *    addresses, and the unspecified address MUST
1666		 *    NOT be included in a candidate set.
1667		 */
1668		if ((score->ifa->flags & IFA_F_TENTATIVE) &&
1669		    (!(score->ifa->flags & IFA_F_OPTIMISTIC)))
1670			continue;
1671
1672		score->addr_type = __ipv6_addr_type(&score->ifa->addr);
1673
1674		if (unlikely(score->addr_type == IPV6_ADDR_ANY ||
1675			     score->addr_type & IPV6_ADDR_MULTICAST)) {
1676			net_dbg_ratelimited("ADDRCONF: unspecified / multicast address assigned as unicast address on %s",
1677					    idev->dev->name);
1678			continue;
1679		}
1680
1681		score->rule = -1;
1682		bitmap_zero(score->scorebits, IPV6_SADDR_RULE_MAX);
1683
1684		for (i = 0; i < IPV6_SADDR_RULE_MAX; i++) {
1685			int minihiscore, miniscore;
1686
1687			minihiscore = ipv6_get_saddr_eval(net, hiscore, dst, i);
1688			miniscore = ipv6_get_saddr_eval(net, score, dst, i);
1689
1690			if (minihiscore > miniscore) {
1691				if (i == IPV6_SADDR_RULE_SCOPE &&
1692				    score->scopedist > 0) {
1693					/*
1694					 * special case:
1695					 * each remaining entry
1696					 * has too small (not enough)
1697					 * scope, because ifa entries
1698					 * are sorted by their scope
1699					 * values.
1700					 */
1701					goto out;
1702				}
1703				break;
1704			} else if (minihiscore < miniscore) {
1705				swap(hiscore, score);
1706				hiscore_idx = 1 - hiscore_idx;
1707
1708				/* restore our iterator */
1709				score->ifa = hiscore->ifa;
1710
1711				break;
1712			}
1713		}
1714	}
1715out:
1716	return hiscore_idx;
1717}
1718
1719static int ipv6_get_saddr_master(struct net *net,
1720				 const struct net_device *dst_dev,
1721				 const struct net_device *master,
1722				 struct ipv6_saddr_dst *dst,
1723				 struct ipv6_saddr_score *scores,
1724				 int hiscore_idx)
1725{
1726	struct inet6_dev *idev;
1727
1728	idev = __in6_dev_get(dst_dev);
1729	if (idev)
1730		hiscore_idx = __ipv6_dev_get_saddr(net, dst, idev,
1731						   scores, hiscore_idx);
1732
1733	idev = __in6_dev_get(master);
1734	if (idev)
1735		hiscore_idx = __ipv6_dev_get_saddr(net, dst, idev,
1736						   scores, hiscore_idx);
1737
1738	return hiscore_idx;
1739}
1740
1741int ipv6_dev_get_saddr(struct net *net, const struct net_device *dst_dev,
1742		       const struct in6_addr *daddr, unsigned int prefs,
1743		       struct in6_addr *saddr)
1744{
1745	struct ipv6_saddr_score scores[2], *hiscore;
1746	struct ipv6_saddr_dst dst;
1747	struct inet6_dev *idev;
1748	struct net_device *dev;
1749	int dst_type;
1750	bool use_oif_addr = false;
1751	int hiscore_idx = 0;
1752	int ret = 0;
1753
1754	dst_type = __ipv6_addr_type(daddr);
1755	dst.addr = daddr;
1756	dst.ifindex = dst_dev ? dst_dev->ifindex : 0;
1757	dst.scope = __ipv6_addr_src_scope(dst_type);
1758	dst.label = ipv6_addr_label(net, daddr, dst_type, dst.ifindex);
1759	dst.prefs = prefs;
1760
1761	scores[hiscore_idx].rule = -1;
1762	scores[hiscore_idx].ifa = NULL;
1763
1764	rcu_read_lock();
1765
1766	/* Candidate Source Address (section 4)
1767	 *  - multicast and link-local destination address,
1768	 *    the set of candidate source address MUST only
1769	 *    include addresses assigned to interfaces
1770	 *    belonging to the same link as the outgoing
1771	 *    interface.
1772	 * (- For site-local destination addresses, the
1773	 *    set of candidate source addresses MUST only
1774	 *    include addresses assigned to interfaces
1775	 *    belonging to the same site as the outgoing
1776	 *    interface.)
1777	 *  - "It is RECOMMENDED that the candidate source addresses
1778	 *    be the set of unicast addresses assigned to the
1779	 *    interface that will be used to send to the destination
1780	 *    (the 'outgoing' interface)." (RFC 6724)
1781	 */
1782	if (dst_dev) {
1783		idev = __in6_dev_get(dst_dev);
1784		if ((dst_type & IPV6_ADDR_MULTICAST) ||
1785		    dst.scope <= IPV6_ADDR_SCOPE_LINKLOCAL ||
1786		    (idev && idev->cnf.use_oif_addrs_only)) {
1787			use_oif_addr = true;
1788		}
1789	}
1790
1791	if (use_oif_addr) {
1792		if (idev)
1793			hiscore_idx = __ipv6_dev_get_saddr(net, &dst, idev, scores, hiscore_idx);
1794	} else {
1795		const struct net_device *master;
1796		int master_idx = 0;
1797
1798		/* if dst_dev exists and is enslaved to an L3 device, then
1799		 * prefer addresses from dst_dev and then the master over
1800		 * any other enslaved devices in the L3 domain.
1801		 */
1802		master = l3mdev_master_dev_rcu(dst_dev);
1803		if (master) {
1804			master_idx = master->ifindex;
1805
1806			hiscore_idx = ipv6_get_saddr_master(net, dst_dev,
1807							    master, &dst,
1808							    scores, hiscore_idx);
1809
1810			if (scores[hiscore_idx].ifa)
1811				goto out;
1812		}
1813
1814		for_each_netdev_rcu(net, dev) {
1815			/* only consider addresses on devices in the
1816			 * same L3 domain
1817			 */
1818			if (l3mdev_master_ifindex_rcu(dev) != master_idx)
1819				continue;
1820			idev = __in6_dev_get(dev);
1821			if (!idev)
1822				continue;
1823			hiscore_idx = __ipv6_dev_get_saddr(net, &dst, idev, scores, hiscore_idx);
1824		}
1825	}
1826
1827out:
1828	hiscore = &scores[hiscore_idx];
1829	if (!hiscore->ifa)
1830		ret = -EADDRNOTAVAIL;
1831	else
1832		*saddr = hiscore->ifa->addr;
1833
1834	rcu_read_unlock();
1835	return ret;
1836}
1837EXPORT_SYMBOL(ipv6_dev_get_saddr);
1838
1839int __ipv6_get_lladdr(struct inet6_dev *idev, struct in6_addr *addr,
1840		      u32 banned_flags)
1841{
1842	struct inet6_ifaddr *ifp;
1843	int err = -EADDRNOTAVAIL;
1844
1845	list_for_each_entry_reverse(ifp, &idev->addr_list, if_list) {
1846		if (ifp->scope > IFA_LINK)
1847			break;
1848		if (ifp->scope == IFA_LINK &&
1849		    !(ifp->flags & banned_flags)) {
1850			*addr = ifp->addr;
1851			err = 0;
1852			break;
1853		}
1854	}
1855	return err;
1856}
1857
1858int ipv6_get_lladdr(struct net_device *dev, struct in6_addr *addr,
1859		    u32 banned_flags)
1860{
1861	struct inet6_dev *idev;
1862	int err = -EADDRNOTAVAIL;
1863
1864	rcu_read_lock();
1865	idev = __in6_dev_get(dev);
1866	if (idev) {
1867		read_lock_bh(&idev->lock);
1868		err = __ipv6_get_lladdr(idev, addr, banned_flags);
1869		read_unlock_bh(&idev->lock);
1870	}
1871	rcu_read_unlock();
1872	return err;
1873}
1874
1875static int ipv6_count_addresses(const struct inet6_dev *idev)
1876{
1877	const struct inet6_ifaddr *ifp;
1878	int cnt = 0;
1879
1880	rcu_read_lock();
1881	list_for_each_entry_rcu(ifp, &idev->addr_list, if_list)
1882		cnt++;
1883	rcu_read_unlock();
1884	return cnt;
1885}
1886
1887int ipv6_chk_addr(struct net *net, const struct in6_addr *addr,
1888		  const struct net_device *dev, int strict)
1889{
1890	return ipv6_chk_addr_and_flags(net, addr, dev, !dev,
1891				       strict, IFA_F_TENTATIVE);
1892}
1893EXPORT_SYMBOL(ipv6_chk_addr);
1894
1895/* device argument is used to find the L3 domain of interest. If
1896 * skip_dev_check is set, then the ifp device is not checked against
1897 * the passed in dev argument. So the 2 cases for addresses checks are:
1898 *   1. does the address exist in the L3 domain that dev is part of
1899 *      (skip_dev_check = true), or
1900 *
1901 *   2. does the address exist on the specific device
1902 *      (skip_dev_check = false)
1903 */
1904static struct net_device *
1905__ipv6_chk_addr_and_flags(struct net *net, const struct in6_addr *addr,
1906			  const struct net_device *dev, bool skip_dev_check,
1907			  int strict, u32 banned_flags)
1908{
1909	unsigned int hash = inet6_addr_hash(net, addr);
1910	struct net_device *l3mdev, *ndev;
1911	struct inet6_ifaddr *ifp;
1912	u32 ifp_flags;
1913
1914	rcu_read_lock();
1915
1916	l3mdev = l3mdev_master_dev_rcu(dev);
1917	if (skip_dev_check)
1918		dev = NULL;
1919
1920	hlist_for_each_entry_rcu(ifp, &inet6_addr_lst[hash], addr_lst) {
1921		ndev = ifp->idev->dev;
1922		if (!net_eq(dev_net(ndev), net))
1923			continue;
1924
1925		if (l3mdev_master_dev_rcu(ndev) != l3mdev)
1926			continue;
1927
1928		/* Decouple optimistic from tentative for evaluation here.
1929		 * Ban optimistic addresses explicitly, when required.
1930		 */
1931		ifp_flags = (ifp->flags&IFA_F_OPTIMISTIC)
1932			    ? (ifp->flags&~IFA_F_TENTATIVE)
1933			    : ifp->flags;
1934		if (ipv6_addr_equal(&ifp->addr, addr) &&
1935		    !(ifp_flags&banned_flags) &&
1936		    (!dev || ndev == dev ||
1937		     !(ifp->scope&(IFA_LINK|IFA_HOST) || strict))) {
1938			rcu_read_unlock();
1939			return ndev;
1940		}
1941	}
1942
1943	rcu_read_unlock();
1944	return NULL;
1945}
1946
1947int ipv6_chk_addr_and_flags(struct net *net, const struct in6_addr *addr,
1948			    const struct net_device *dev, bool skip_dev_check,
1949			    int strict, u32 banned_flags)
1950{
1951	return __ipv6_chk_addr_and_flags(net, addr, dev, skip_dev_check,
1952					 strict, banned_flags) ? 1 : 0;
1953}
1954EXPORT_SYMBOL(ipv6_chk_addr_and_flags);
1955
1956
1957/* Compares an address/prefix_len with addresses on device @dev.
1958 * If one is found it returns true.
1959 */
1960bool ipv6_chk_custom_prefix(const struct in6_addr *addr,
1961	const unsigned int prefix_len, struct net_device *dev)
1962{
1963	const struct inet6_ifaddr *ifa;
1964	const struct inet6_dev *idev;
1965	bool ret = false;
1966
1967	rcu_read_lock();
1968	idev = __in6_dev_get(dev);
1969	if (idev) {
1970		list_for_each_entry_rcu(ifa, &idev->addr_list, if_list) {
1971			ret = ipv6_prefix_equal(addr, &ifa->addr, prefix_len);
1972			if (ret)
1973				break;
1974		}
1975	}
1976	rcu_read_unlock();
1977
1978	return ret;
1979}
1980EXPORT_SYMBOL(ipv6_chk_custom_prefix);
1981
1982int ipv6_chk_prefix(const struct in6_addr *addr, struct net_device *dev)
1983{
1984	const struct inet6_ifaddr *ifa;
1985	const struct inet6_dev *idev;
1986	int	onlink;
1987
1988	onlink = 0;
1989	rcu_read_lock();
1990	idev = __in6_dev_get(dev);
1991	if (idev) {
1992		list_for_each_entry_rcu(ifa, &idev->addr_list, if_list) {
1993			onlink = ipv6_prefix_equal(addr, &ifa->addr,
1994						   ifa->prefix_len);
1995			if (onlink)
1996				break;
1997		}
1998	}
1999	rcu_read_unlock();
2000	return onlink;
2001}
2002EXPORT_SYMBOL(ipv6_chk_prefix);
2003
2004/**
2005 * ipv6_dev_find - find the first device with a given source address.
2006 * @net: the net namespace
2007 * @addr: the source address
2008 *
2009 * The caller should be protected by RCU, or RTNL.
2010 */
2011struct net_device *ipv6_dev_find(struct net *net, const struct in6_addr *addr,
2012				 struct net_device *dev)
2013{
2014	return __ipv6_chk_addr_and_flags(net, addr, dev, !dev, 1,
2015					 IFA_F_TENTATIVE);
2016}
2017EXPORT_SYMBOL(ipv6_dev_find);
2018
2019struct inet6_ifaddr *ipv6_get_ifaddr(struct net *net, const struct in6_addr *addr,
2020				     struct net_device *dev, int strict)
2021{
2022	unsigned int hash = inet6_addr_hash(net, addr);
2023	struct inet6_ifaddr *ifp, *result = NULL;
2024
2025	rcu_read_lock();
2026	hlist_for_each_entry_rcu(ifp, &inet6_addr_lst[hash], addr_lst) {
2027		if (!net_eq(dev_net(ifp->idev->dev), net))
2028			continue;
2029		if (ipv6_addr_equal(&ifp->addr, addr)) {
2030			if (!dev || ifp->idev->dev == dev ||
2031			    !(ifp->scope&(IFA_LINK|IFA_HOST) || strict)) {
2032				if (in6_ifa_hold_safe(ifp)) {
2033					result = ifp;
2034					break;
2035				}
2036			}
2037		}
2038	}
2039	rcu_read_unlock();
2040
2041	return result;
2042}
2043
2044/* Gets referenced address, destroys ifaddr */
2045
2046static void addrconf_dad_stop(struct inet6_ifaddr *ifp, int dad_failed)
2047{
2048	if (dad_failed)
2049		ifp->flags |= IFA_F_DADFAILED;
2050
2051	if (ifp->flags&IFA_F_TEMPORARY) {
2052		struct inet6_ifaddr *ifpub;
2053		spin_lock_bh(&ifp->lock);
2054		ifpub = ifp->ifpub;
2055		if (ifpub) {
2056			in6_ifa_hold(ifpub);
2057			spin_unlock_bh(&ifp->lock);
2058			ipv6_create_tempaddr(ifpub, true);
2059			in6_ifa_put(ifpub);
2060		} else {
2061			spin_unlock_bh(&ifp->lock);
2062		}
2063		ipv6_del_addr(ifp);
2064	} else if (ifp->flags&IFA_F_PERMANENT || !dad_failed) {
2065		spin_lock_bh(&ifp->lock);
2066		addrconf_del_dad_work(ifp);
2067		ifp->flags |= IFA_F_TENTATIVE;
2068		if (dad_failed)
2069			ifp->flags &= ~IFA_F_OPTIMISTIC;
2070		spin_unlock_bh(&ifp->lock);
2071		if (dad_failed)
2072			ipv6_ifa_notify(0, ifp);
2073		in6_ifa_put(ifp);
2074	} else {
2075		ipv6_del_addr(ifp);
2076	}
2077}
2078
2079static int addrconf_dad_end(struct inet6_ifaddr *ifp)
2080{
2081	int err = -ENOENT;
2082
2083	spin_lock_bh(&ifp->lock);
2084	if (ifp->state == INET6_IFADDR_STATE_DAD) {
2085		ifp->state = INET6_IFADDR_STATE_POSTDAD;
2086		err = 0;
2087	}
2088	spin_unlock_bh(&ifp->lock);
2089
2090	return err;
2091}
2092
2093void addrconf_dad_failure(struct sk_buff *skb, struct inet6_ifaddr *ifp)
2094{
2095	struct inet6_dev *idev = ifp->idev;
2096	struct net *net = dev_net(ifp->idev->dev);
2097
2098	if (addrconf_dad_end(ifp)) {
2099		in6_ifa_put(ifp);
2100		return;
2101	}
2102
2103	net_info_ratelimited("%s: IPv6 duplicate address %pI6c used by %pM detected!\n",
2104			     ifp->idev->dev->name, &ifp->addr, eth_hdr(skb)->h_source);
2105
2106	spin_lock_bh(&ifp->lock);
2107
2108	if (ifp->flags & IFA_F_STABLE_PRIVACY) {
2109		struct in6_addr new_addr;
2110		struct inet6_ifaddr *ifp2;
2111		int retries = ifp->stable_privacy_retry + 1;
2112		struct ifa6_config cfg = {
2113			.pfx = &new_addr,
2114			.plen = ifp->prefix_len,
2115			.ifa_flags = ifp->flags,
2116			.valid_lft = ifp->valid_lft,
2117			.preferred_lft = ifp->prefered_lft,
2118			.scope = ifp->scope,
2119		};
2120
2121		if (retries > net->ipv6.sysctl.idgen_retries) {
2122			net_info_ratelimited("%s: privacy stable address generation failed because of DAD conflicts!\n",
2123					     ifp->idev->dev->name);
2124			goto errdad;
2125		}
2126
2127		new_addr = ifp->addr;
2128		if (ipv6_generate_stable_address(&new_addr, retries,
2129						 idev))
2130			goto errdad;
2131
2132		spin_unlock_bh(&ifp->lock);
2133
2134		if (idev->cnf.max_addresses &&
2135		    ipv6_count_addresses(idev) >=
2136		    idev->cnf.max_addresses)
2137			goto lock_errdad;
2138
2139		net_info_ratelimited("%s: generating new stable privacy address because of DAD conflict\n",
2140				     ifp->idev->dev->name);
2141
2142		ifp2 = ipv6_add_addr(idev, &cfg, false, NULL);
2143		if (IS_ERR(ifp2))
2144			goto lock_errdad;
2145
2146		spin_lock_bh(&ifp2->lock);
2147		ifp2->stable_privacy_retry = retries;
2148		ifp2->state = INET6_IFADDR_STATE_PREDAD;
2149		spin_unlock_bh(&ifp2->lock);
2150
2151		addrconf_mod_dad_work(ifp2, net->ipv6.sysctl.idgen_delay);
2152		in6_ifa_put(ifp2);
2153lock_errdad:
2154		spin_lock_bh(&ifp->lock);
2155	}
2156
2157errdad:
2158	/* transition from _POSTDAD to _ERRDAD */
2159	ifp->state = INET6_IFADDR_STATE_ERRDAD;
2160	spin_unlock_bh(&ifp->lock);
2161
2162	addrconf_mod_dad_work(ifp, 0);
2163	in6_ifa_put(ifp);
2164}
2165
2166/* Join to solicited addr multicast group.
2167 * caller must hold RTNL */
2168void addrconf_join_solict(struct net_device *dev, const struct in6_addr *addr)
2169{
2170	struct in6_addr maddr;
2171
2172	if (dev->flags&(IFF_LOOPBACK|IFF_NOARP))
2173		return;
2174
2175	addrconf_addr_solict_mult(addr, &maddr);
2176	ipv6_dev_mc_inc(dev, &maddr);
2177}
2178
2179/* caller must hold RTNL */
2180void addrconf_leave_solict(struct inet6_dev *idev, const struct in6_addr *addr)
2181{
2182	struct in6_addr maddr;
2183
2184	if (idev->dev->flags&(IFF_LOOPBACK|IFF_NOARP))
2185		return;
2186
2187	addrconf_addr_solict_mult(addr, &maddr);
2188	__ipv6_dev_mc_dec(idev, &maddr);
2189}
2190
2191/* caller must hold RTNL */
2192static void addrconf_join_anycast(struct inet6_ifaddr *ifp)
2193{
2194	struct in6_addr addr;
2195
2196	if (ifp->prefix_len >= 127) /* RFC 6164 */
2197		return;
2198	ipv6_addr_prefix(&addr, &ifp->addr, ifp->prefix_len);
2199	if (ipv6_addr_any(&addr))
2200		return;
2201	__ipv6_dev_ac_inc(ifp->idev, &addr);
2202}
2203
2204/* caller must hold RTNL */
2205static void addrconf_leave_anycast(struct inet6_ifaddr *ifp)
2206{
2207	struct in6_addr addr;
2208
2209	if (ifp->prefix_len >= 127) /* RFC 6164 */
2210		return;
2211	ipv6_addr_prefix(&addr, &ifp->addr, ifp->prefix_len);
2212	if (ipv6_addr_any(&addr))
2213		return;
2214	__ipv6_dev_ac_dec(ifp->idev, &addr);
2215}
2216
2217static int addrconf_ifid_6lowpan(u8 *eui, struct net_device *dev)
2218{
2219	switch (dev->addr_len) {
2220	case ETH_ALEN:
2221		memcpy(eui, dev->dev_addr, 3);
2222		eui[3] = 0xFF;
2223		eui[4] = 0xFE;
2224		memcpy(eui + 5, dev->dev_addr + 3, 3);
2225		break;
2226	case EUI64_ADDR_LEN:
2227		memcpy(eui, dev->dev_addr, EUI64_ADDR_LEN);
2228		eui[0] ^= 2;
2229		break;
2230	default:
2231		return -1;
2232	}
2233
2234	return 0;
2235}
2236
2237static int addrconf_ifid_ieee1394(u8 *eui, struct net_device *dev)
2238{
2239	union fwnet_hwaddr *ha;
2240
2241	if (dev->addr_len != FWNET_ALEN)
2242		return -1;
2243
2244	ha = (union fwnet_hwaddr *)dev->dev_addr;
2245
2246	memcpy(eui, &ha->uc.uniq_id, sizeof(ha->uc.uniq_id));
2247	eui[0] ^= 2;
2248	return 0;
2249}
2250
2251static int addrconf_ifid_arcnet(u8 *eui, struct net_device *dev)
2252{
2253	/* XXX: inherit EUI-64 from other interface -- yoshfuji */
2254	if (dev->addr_len != ARCNET_ALEN)
2255		return -1;
2256	memset(eui, 0, 7);
2257	eui[7] = *(u8 *)dev->dev_addr;
2258	return 0;
2259}
2260
2261static int addrconf_ifid_infiniband(u8 *eui, struct net_device *dev)
2262{
2263	if (dev->addr_len != INFINIBAND_ALEN)
2264		return -1;
2265	memcpy(eui, dev->dev_addr + 12, 8);
2266	eui[0] |= 2;
2267	return 0;
2268}
2269
2270static int __ipv6_isatap_ifid(u8 *eui, __be32 addr)
2271{
2272	if (addr == 0)
2273		return -1;
2274	eui[0] = (ipv4_is_zeronet(addr) || ipv4_is_private_10(addr) ||
2275		  ipv4_is_loopback(addr) || ipv4_is_linklocal_169(addr) ||
2276		  ipv4_is_private_172(addr) || ipv4_is_test_192(addr) ||
2277		  ipv4_is_anycast_6to4(addr) || ipv4_is_private_192(addr) ||
2278		  ipv4_is_test_198(addr) || ipv4_is_multicast(addr) ||
2279		  ipv4_is_lbcast(addr)) ? 0x00 : 0x02;
2280	eui[1] = 0;
2281	eui[2] = 0x5E;
2282	eui[3] = 0xFE;
2283	memcpy(eui + 4, &addr, 4);
2284	return 0;
2285}
2286
2287static int addrconf_ifid_sit(u8 *eui, struct net_device *dev)
2288{
2289	if (dev->priv_flags & IFF_ISATAP)
2290		return __ipv6_isatap_ifid(eui, *(__be32 *)dev->dev_addr);
2291	return -1;
2292}
2293
2294static int addrconf_ifid_gre(u8 *eui, struct net_device *dev)
2295{
2296	return __ipv6_isatap_ifid(eui, *(__be32 *)dev->dev_addr);
2297}
2298
2299static int addrconf_ifid_ip6tnl(u8 *eui, struct net_device *dev)
2300{
2301	memcpy(eui, dev->perm_addr, 3);
2302	memcpy(eui + 5, dev->perm_addr + 3, 3);
2303	eui[3] = 0xFF;
2304	eui[4] = 0xFE;
2305	eui[0] ^= 2;
2306	return 0;
2307}
2308
2309static int ipv6_generate_eui64(u8 *eui, struct net_device *dev)
2310{
2311	switch (dev->type) {
2312	case ARPHRD_ETHER:
2313	case ARPHRD_FDDI:
2314		return addrconf_ifid_eui48(eui, dev);
2315	case ARPHRD_ARCNET:
2316		return addrconf_ifid_arcnet(eui, dev);
2317	case ARPHRD_INFINIBAND:
2318		return addrconf_ifid_infiniband(eui, dev);
2319	case ARPHRD_SIT:
2320		return addrconf_ifid_sit(eui, dev);
2321	case ARPHRD_IPGRE:
2322	case ARPHRD_TUNNEL:
2323		return addrconf_ifid_gre(eui, dev);
2324	case ARPHRD_6LOWPAN:
2325		return addrconf_ifid_6lowpan(eui, dev);
2326	case ARPHRD_IEEE1394:
2327		return addrconf_ifid_ieee1394(eui, dev);
2328	case ARPHRD_TUNNEL6:
2329	case ARPHRD_IP6GRE:
2330	case ARPHRD_RAWIP:
2331		return addrconf_ifid_ip6tnl(eui, dev);
2332	}
2333	return -1;
2334}
2335
2336static int ipv6_inherit_eui64(u8 *eui, struct inet6_dev *idev)
2337{
2338	int err = -1;
2339	struct inet6_ifaddr *ifp;
2340
2341	read_lock_bh(&idev->lock);
2342	list_for_each_entry_reverse(ifp, &idev->addr_list, if_list) {
2343		if (ifp->scope > IFA_LINK)
2344			break;
2345		if (ifp->scope == IFA_LINK && !(ifp->flags&IFA_F_TENTATIVE)) {
2346			memcpy(eui, ifp->addr.s6_addr+8, 8);
2347			err = 0;
2348			break;
2349		}
2350	}
2351	read_unlock_bh(&idev->lock);
2352	return err;
2353}
2354
2355/* Generation of a randomized Interface Identifier
2356 * draft-ietf-6man-rfc4941bis, Section 3.3.1
2357 */
2358
2359static void ipv6_gen_rnd_iid(struct in6_addr *addr)
2360{
2361regen:
2362	get_random_bytes(&addr->s6_addr[8], 8);
2363
2364	/* <draft-ietf-6man-rfc4941bis-08.txt>, Section 3.3.1:
2365	 * check if generated address is not inappropriate:
2366	 *
2367	 * - Reserved IPv6 Interface Identifers
2368	 * - XXX: already assigned to an address on the device
2369	 */
2370
2371	/* Subnet-router anycast: 0000:0000:0000:0000 */
2372	if (!(addr->s6_addr32[2] | addr->s6_addr32[3]))
2373		goto regen;
2374
2375	/* IANA Ethernet block: 0200:5EFF:FE00:0000-0200:5EFF:FE00:5212
2376	 * Proxy Mobile IPv6:   0200:5EFF:FE00:5213
2377	 * IANA Ethernet block: 0200:5EFF:FE00:5214-0200:5EFF:FEFF:FFFF
2378	 */
2379	if (ntohl(addr->s6_addr32[2]) == 0x02005eff &&
2380	    (ntohl(addr->s6_addr32[3]) & 0Xff000000) == 0xfe000000)
2381		goto regen;
2382
2383	/* Reserved subnet anycast addresses */
2384	if (ntohl(addr->s6_addr32[2]) == 0xfdffffff &&
2385	    ntohl(addr->s6_addr32[3]) >= 0Xffffff80)
2386		goto regen;
2387}
2388
2389/*
2390 *	Add prefix route.
2391 */
2392
2393static void
2394addrconf_prefix_route(struct in6_addr *pfx, int plen, u32 metric,
2395		      struct net_device *dev, unsigned long expires,
2396		      u32 flags, gfp_t gfp_flags)
2397{
2398	struct fib6_config cfg = {
2399		.fc_table = l3mdev_fib_table(dev) ? : RT6_TABLE_PREFIX,
2400		.fc_metric = metric ? : IP6_RT_PRIO_ADDRCONF,
2401		.fc_ifindex = dev->ifindex,
2402		.fc_expires = expires,
2403		.fc_dst_len = plen,
2404		.fc_flags = RTF_UP | flags,
2405		.fc_nlinfo.nl_net = dev_net(dev),
2406		.fc_protocol = RTPROT_KERNEL,
2407		.fc_type = RTN_UNICAST,
2408	};
2409
2410	cfg.fc_dst = *pfx;
2411
2412	/* Prevent useless cloning on PtP SIT.
2413	   This thing is done here expecting that the whole
2414	   class of non-broadcast devices need not cloning.
2415	 */
2416#if IS_ENABLED(CONFIG_IPV6_SIT)
2417	if (dev->type == ARPHRD_SIT && (dev->flags & IFF_POINTOPOINT))
2418		cfg.fc_flags |= RTF_NONEXTHOP;
2419#endif
2420
2421	ip6_route_add(&cfg, gfp_flags, NULL);
2422}
2423
2424
2425static struct fib6_info *addrconf_get_prefix_route(const struct in6_addr *pfx,
2426						  int plen,
2427						  const struct net_device *dev,
2428						  u32 flags, u32 noflags,
2429						  bool no_gw)
2430{
2431	struct fib6_node *fn;
2432	struct fib6_info *rt = NULL;
2433	struct fib6_table *table;
2434	u32 tb_id = l3mdev_fib_table(dev) ? : RT6_TABLE_PREFIX;
2435
2436	table = fib6_get_table(dev_net(dev), tb_id);
2437	if (!table)
2438		return NULL;
2439
2440	rcu_read_lock();
2441	fn = fib6_locate(&table->tb6_root, pfx, plen, NULL, 0, true);
2442	if (!fn)
2443		goto out;
2444
2445	for_each_fib6_node_rt_rcu(fn) {
2446		/* prefix routes only use builtin fib6_nh */
2447		if (rt->nh)
2448			continue;
2449
2450		if (rt->fib6_nh->fib_nh_dev->ifindex != dev->ifindex)
2451			continue;
2452		if (no_gw && rt->fib6_nh->fib_nh_gw_family)
2453			continue;
2454		if ((rt->fib6_flags & flags) != flags)
2455			continue;
2456		if ((rt->fib6_flags & noflags) != 0)
2457			continue;
2458		if (!fib6_info_hold_safe(rt))
2459			continue;
2460		break;
2461	}
2462out:
2463	rcu_read_unlock();
2464	return rt;
2465}
2466
2467
2468/* Create "default" multicast route to the interface */
2469
2470static void addrconf_add_mroute(struct net_device *dev)
2471{
2472	struct fib6_config cfg = {
2473		.fc_table = l3mdev_fib_table(dev) ? : RT6_TABLE_LOCAL,
2474		.fc_metric = IP6_RT_PRIO_ADDRCONF,
2475		.fc_ifindex = dev->ifindex,
2476		.fc_dst_len = 8,
2477		.fc_flags = RTF_UP,
2478		.fc_type = RTN_MULTICAST,
2479		.fc_nlinfo.nl_net = dev_net(dev),
2480		.fc_protocol = RTPROT_KERNEL,
2481	};
2482
2483	ipv6_addr_set(&cfg.fc_dst, htonl(0xFF000000), 0, 0, 0);
2484
2485	ip6_route_add(&cfg, GFP_KERNEL, NULL);
2486}
2487
2488static struct inet6_dev *addrconf_add_dev(struct net_device *dev)
2489{
2490	struct inet6_dev *idev;
2491
2492	ASSERT_RTNL();
2493
2494	idev = ipv6_find_idev(dev);
2495	if (IS_ERR(idev))
2496		return idev;
2497
2498	if (idev->cnf.disable_ipv6)
2499		return ERR_PTR(-EACCES);
2500
2501	/* Add default multicast route */
2502	if (!(dev->flags & IFF_LOOPBACK) && !netif_is_l3_master(dev))
2503		addrconf_add_mroute(dev);
2504
2505	return idev;
2506}
2507
2508static void manage_tempaddrs(struct inet6_dev *idev,
2509			     struct inet6_ifaddr *ifp,
2510			     __u32 valid_lft, __u32 prefered_lft,
2511			     bool create, unsigned long now)
2512{
2513	u32 flags;
2514	struct inet6_ifaddr *ift;
2515
2516	read_lock_bh(&idev->lock);
2517	/* update all temporary addresses in the list */
2518	list_for_each_entry(ift, &idev->tempaddr_list, tmp_list) {
2519		int age, max_valid, max_prefered;
2520
2521		if (ifp != ift->ifpub)
2522			continue;
2523
2524		/* RFC 4941 section 3.3:
2525		 * If a received option will extend the lifetime of a public
2526		 * address, the lifetimes of temporary addresses should
2527		 * be extended, subject to the overall constraint that no
2528		 * temporary addresses should ever remain "valid" or "preferred"
2529		 * for a time longer than (TEMP_VALID_LIFETIME) or
2530		 * (TEMP_PREFERRED_LIFETIME - DESYNC_FACTOR), respectively.
2531		 */
2532		age = (now - ift->cstamp) / HZ;
2533		max_valid = idev->cnf.temp_valid_lft - age;
2534		if (max_valid < 0)
2535			max_valid = 0;
2536
2537		max_prefered = idev->cnf.temp_prefered_lft -
2538			       idev->desync_factor - age;
2539		if (max_prefered < 0)
2540			max_prefered = 0;
2541
2542		if (valid_lft > max_valid)
2543			valid_lft = max_valid;
2544
2545		if (prefered_lft > max_prefered)
2546			prefered_lft = max_prefered;
2547
2548		spin_lock(&ift->lock);
2549		flags = ift->flags;
2550		ift->valid_lft = valid_lft;
2551		ift->prefered_lft = prefered_lft;
2552		ift->tstamp = now;
2553		if (prefered_lft > 0)
2554			ift->flags &= ~IFA_F_DEPRECATED;
2555
2556		spin_unlock(&ift->lock);
2557		if (!(flags&IFA_F_TENTATIVE))
2558			ipv6_ifa_notify(0, ift);
2559	}
2560
2561	/* Also create a temporary address if it's enabled but no temporary
2562	 * address currently exists.
2563	 * However, we get called with valid_lft == 0, prefered_lft == 0, create == false
2564	 * as part of cleanup (ie. deleting the mngtmpaddr).
2565	 * We don't want that to result in creating a new temporary ip address.
2566	 */
2567	if (list_empty(&idev->tempaddr_list) && (valid_lft || prefered_lft))
2568		create = true;
2569
2570	if (create && idev->cnf.use_tempaddr > 0) {
2571		/* When a new public address is created as described
2572		 * in [ADDRCONF], also create a new temporary address.
2573		 */
2574		read_unlock_bh(&idev->lock);
2575		ipv6_create_tempaddr(ifp, false);
2576	} else {
2577		read_unlock_bh(&idev->lock);
2578	}
2579}
2580
2581static bool is_addr_mode_generate_stable(struct inet6_dev *idev)
2582{
2583	return idev->cnf.addr_gen_mode == IN6_ADDR_GEN_MODE_STABLE_PRIVACY ||
2584	       idev->cnf.addr_gen_mode == IN6_ADDR_GEN_MODE_RANDOM;
2585}
2586
2587int addrconf_prefix_rcv_add_addr(struct net *net, struct net_device *dev,
2588				 const struct prefix_info *pinfo,
2589				 struct inet6_dev *in6_dev,
2590				 const struct in6_addr *addr, int addr_type,
2591				 u32 addr_flags, bool sllao, bool tokenized,
2592				 __u32 valid_lft, u32 prefered_lft)
2593{
2594	struct inet6_ifaddr *ifp = ipv6_get_ifaddr(net, addr, dev, 1);
2595	int create = 0, update_lft = 0;
2596
2597	if (!ifp && valid_lft) {
2598		int max_addresses = in6_dev->cnf.max_addresses;
2599		struct ifa6_config cfg = {
2600			.pfx = addr,
2601			.plen = pinfo->prefix_len,
2602			.ifa_flags = addr_flags,
2603			.valid_lft = valid_lft,
2604			.preferred_lft = prefered_lft,
2605			.scope = addr_type & IPV6_ADDR_SCOPE_MASK,
2606		};
2607
2608#ifdef CONFIG_IPV6_OPTIMISTIC_DAD
2609		if ((net->ipv6.devconf_all->optimistic_dad ||
2610		     in6_dev->cnf.optimistic_dad) &&
2611		    !net->ipv6.devconf_all->forwarding && sllao)
2612			cfg.ifa_flags |= IFA_F_OPTIMISTIC;
2613#endif
2614
2615		/* Do not allow to create too much of autoconfigured
2616		 * addresses; this would be too easy way to crash kernel.
2617		 */
2618		if (!max_addresses ||
2619		    ipv6_count_addresses(in6_dev) < max_addresses)
2620			ifp = ipv6_add_addr(in6_dev, &cfg, false, NULL);
2621
2622		if (IS_ERR_OR_NULL(ifp))
2623			return -1;
2624
2625		create = 1;
2626		spin_lock_bh(&ifp->lock);
2627		ifp->flags |= IFA_F_MANAGETEMPADDR;
2628		ifp->cstamp = jiffies;
2629		ifp->tokenized = tokenized;
2630		spin_unlock_bh(&ifp->lock);
2631		addrconf_dad_start(ifp);
2632	}
2633
2634	if (ifp) {
2635		u32 flags;
2636		unsigned long now;
2637		u32 stored_lft;
2638
2639		/* update lifetime (RFC2462 5.5.3 e) */
2640		spin_lock_bh(&ifp->lock);
2641		now = jiffies;
2642		if (ifp->valid_lft > (now - ifp->tstamp) / HZ)
2643			stored_lft = ifp->valid_lft - (now - ifp->tstamp) / HZ;
2644		else
2645			stored_lft = 0;
2646		if (!create && stored_lft) {
2647			const u32 minimum_lft = min_t(u32,
2648				stored_lft, MIN_VALID_LIFETIME);
2649			valid_lft = max(valid_lft, minimum_lft);
2650
2651			/* RFC4862 Section 5.5.3e:
2652			 * "Note that the preferred lifetime of the
2653			 *  corresponding address is always reset to
2654			 *  the Preferred Lifetime in the received
2655			 *  Prefix Information option, regardless of
2656			 *  whether the valid lifetime is also reset or
2657			 *  ignored."
2658			 *
2659			 * So we should always update prefered_lft here.
2660			 */
2661			update_lft = 1;
2662		}
2663
2664		if (update_lft) {
2665			ifp->valid_lft = valid_lft;
2666			ifp->prefered_lft = prefered_lft;
2667			ifp->tstamp = now;
2668			flags = ifp->flags;
2669			ifp->flags &= ~IFA_F_DEPRECATED;
2670			spin_unlock_bh(&ifp->lock);
2671
2672			if (!(flags&IFA_F_TENTATIVE))
2673				ipv6_ifa_notify(0, ifp);
2674		} else
2675			spin_unlock_bh(&ifp->lock);
2676
2677		manage_tempaddrs(in6_dev, ifp, valid_lft, prefered_lft,
2678				 create, now);
2679
2680		in6_ifa_put(ifp);
2681		addrconf_verify();
2682	}
2683
2684	return 0;
2685}
2686EXPORT_SYMBOL_GPL(addrconf_prefix_rcv_add_addr);
2687
2688void addrconf_prefix_rcv(struct net_device *dev, u8 *opt, int len, bool sllao)
2689{
2690	struct prefix_info *pinfo;
2691	__u32 valid_lft;
2692	__u32 prefered_lft;
2693	int addr_type, err;
2694	u32 addr_flags = 0;
2695	struct inet6_dev *in6_dev;
2696	struct net *net = dev_net(dev);
2697
2698	pinfo = (struct prefix_info *) opt;
2699
2700	if (len < sizeof(struct prefix_info)) {
2701		netdev_dbg(dev, "addrconf: prefix option too short\n");
2702		return;
2703	}
2704
2705	/*
2706	 *	Validation checks ([ADDRCONF], page 19)
2707	 */
2708
2709	addr_type = ipv6_addr_type(&pinfo->prefix);
2710
2711	if (addr_type & (IPV6_ADDR_MULTICAST|IPV6_ADDR_LINKLOCAL))
2712		return;
2713
2714	valid_lft = ntohl(pinfo->valid);
2715	prefered_lft = ntohl(pinfo->prefered);
2716
2717	if (prefered_lft > valid_lft) {
2718		net_warn_ratelimited("addrconf: prefix option has invalid lifetime\n");
2719		return;
2720	}
2721
2722	in6_dev = in6_dev_get(dev);
2723
2724	if (!in6_dev) {
2725		net_dbg_ratelimited("addrconf: device %s not configured\n",
2726				    dev->name);
2727		return;
2728	}
2729
2730	if (valid_lft != 0 && valid_lft < in6_dev->cnf.accept_ra_min_lft)
2731		goto put;
2732
2733	/*
2734	 *	Two things going on here:
2735	 *	1) Add routes for on-link prefixes
2736	 *	2) Configure prefixes with the auto flag set
2737	 */
2738
2739	if (pinfo->onlink) {
2740		struct fib6_info *rt;
2741		unsigned long rt_expires;
2742
2743		/* Avoid arithmetic overflow. Really, we could
2744		 * save rt_expires in seconds, likely valid_lft,
2745		 * but it would require division in fib gc, that it
2746		 * not good.
2747		 */
2748		if (HZ > USER_HZ)
2749			rt_expires = addrconf_timeout_fixup(valid_lft, HZ);
2750		else
2751			rt_expires = addrconf_timeout_fixup(valid_lft, USER_HZ);
2752
2753		if (addrconf_finite_timeout(rt_expires))
2754			rt_expires *= HZ;
2755
2756		rt = addrconf_get_prefix_route(&pinfo->prefix,
2757					       pinfo->prefix_len,
2758					       dev,
2759					       RTF_ADDRCONF | RTF_PREFIX_RT,
2760					       RTF_DEFAULT, true);
2761
2762		if (rt) {
2763			/* Autoconf prefix route */
2764			if (valid_lft == 0) {
2765				ip6_del_rt(net, rt, false);
2766				rt = NULL;
2767			} else if (addrconf_finite_timeout(rt_expires)) {
2768				/* not infinity */
2769				fib6_set_expires(rt, jiffies + rt_expires);
2770			} else {
2771				fib6_clean_expires(rt);
2772			}
2773		} else if (valid_lft) {
2774			clock_t expires = 0;
2775			int flags = RTF_ADDRCONF | RTF_PREFIX_RT;
2776			if (addrconf_finite_timeout(rt_expires)) {
2777				/* not infinity */
2778				flags |= RTF_EXPIRES;
2779				expires = jiffies_to_clock_t(rt_expires);
2780			}
2781			addrconf_prefix_route(&pinfo->prefix, pinfo->prefix_len,
2782					      0, dev, expires, flags,
2783					      GFP_ATOMIC);
2784		}
2785		fib6_info_release(rt);
2786	}
2787
2788	/* Try to figure out our local address for this prefix */
2789
2790	if (pinfo->autoconf && in6_dev->cnf.autoconf) {
2791		struct in6_addr addr;
2792		bool tokenized = false, dev_addr_generated = false;
2793
2794		if (pinfo->prefix_len == 64) {
2795			memcpy(&addr, &pinfo->prefix, 8);
2796
2797			if (!ipv6_addr_any(&in6_dev->token)) {
2798				read_lock_bh(&in6_dev->lock);
2799				memcpy(addr.s6_addr + 8,
2800				       in6_dev->token.s6_addr + 8, 8);
2801				read_unlock_bh(&in6_dev->lock);
2802				tokenized = true;
2803			} else if (is_addr_mode_generate_stable(in6_dev) &&
2804				   !ipv6_generate_stable_address(&addr, 0,
2805								 in6_dev)) {
2806				addr_flags |= IFA_F_STABLE_PRIVACY;
2807				goto ok;
2808			} else if (ipv6_generate_eui64(addr.s6_addr + 8, dev) &&
2809				   ipv6_inherit_eui64(addr.s6_addr + 8, in6_dev)) {
2810				goto put;
2811			} else {
2812				dev_addr_generated = true;
2813			}
2814			goto ok;
2815		}
2816		net_dbg_ratelimited("IPv6 addrconf: prefix with wrong length %d\n",
2817				    pinfo->prefix_len);
2818		goto put;
2819
2820ok:
2821		err = addrconf_prefix_rcv_add_addr(net, dev, pinfo, in6_dev,
2822						   &addr, addr_type,
2823						   addr_flags, sllao,
2824						   tokenized, valid_lft,
2825						   prefered_lft);
2826		if (err)
2827			goto put;
2828
2829		/* Ignore error case here because previous prefix add addr was
2830		 * successful which will be notified.
2831		 */
2832		ndisc_ops_prefix_rcv_add_addr(net, dev, pinfo, in6_dev, &addr,
2833					      addr_type, addr_flags, sllao,
2834					      tokenized, valid_lft,
2835					      prefered_lft,
2836					      dev_addr_generated);
2837	}
2838	inet6_prefix_notify(RTM_NEWPREFIX, in6_dev, pinfo);
2839put:
2840	in6_dev_put(in6_dev);
2841}
2842
2843static int addrconf_set_sit_dstaddr(struct net *net, struct net_device *dev,
2844		struct in6_ifreq *ireq)
2845{
2846	struct ip_tunnel_parm p = { };
2847	int err;
2848
2849	if (!(ipv6_addr_type(&ireq->ifr6_addr) & IPV6_ADDR_COMPATv4))
2850		return -EADDRNOTAVAIL;
2851
2852	p.iph.daddr = ireq->ifr6_addr.s6_addr32[3];
2853	p.iph.version = 4;
2854	p.iph.ihl = 5;
2855	p.iph.protocol = IPPROTO_IPV6;
2856	p.iph.ttl = 64;
2857
2858	if (!dev->netdev_ops->ndo_tunnel_ctl)
2859		return -EOPNOTSUPP;
2860	err = dev->netdev_ops->ndo_tunnel_ctl(dev, &p, SIOCADDTUNNEL);
2861	if (err)
2862		return err;
2863
2864	dev = __dev_get_by_name(net, p.name);
2865	if (!dev)
2866		return -ENOBUFS;
2867	return dev_open(dev, NULL);
2868}
2869
2870/*
2871 *	Set destination address.
2872 *	Special case for SIT interfaces where we create a new "virtual"
2873 *	device.
2874 */
2875int addrconf_set_dstaddr(struct net *net, void __user *arg)
2876{
2877	struct net_device *dev;
2878	struct in6_ifreq ireq;
2879	int err = -ENODEV;
2880
2881	if (!IS_ENABLED(CONFIG_IPV6_SIT))
2882		return -ENODEV;
2883	if (copy_from_user(&ireq, arg, sizeof(struct in6_ifreq)))
2884		return -EFAULT;
2885
2886	rtnl_lock();
2887	dev = __dev_get_by_index(net, ireq.ifr6_ifindex);
2888	if (dev && dev->type == ARPHRD_SIT)
2889		err = addrconf_set_sit_dstaddr(net, dev, &ireq);
2890	rtnl_unlock();
2891	return err;
2892}
2893
2894static int ipv6_mc_config(struct sock *sk, bool join,
2895			  const struct in6_addr *addr, int ifindex)
2896{
2897	int ret;
2898
2899	ASSERT_RTNL();
2900
2901	lock_sock(sk);
2902	if (join)
2903		ret = ipv6_sock_mc_join(sk, ifindex, addr);
2904	else
2905		ret = ipv6_sock_mc_drop(sk, ifindex, addr);
2906	release_sock(sk);
2907
2908	return ret;
2909}
2910
2911/*
2912 *	Manual configuration of address on an interface
2913 */
2914static int inet6_addr_add(struct net *net, int ifindex,
2915			  struct ifa6_config *cfg,
2916			  struct netlink_ext_ack *extack)
2917{
2918	struct inet6_ifaddr *ifp;
2919	struct inet6_dev *idev;
2920	struct net_device *dev;
2921	unsigned long timeout;
2922	clock_t expires;
2923	u32 flags;
2924
2925	ASSERT_RTNL();
2926
2927	if (cfg->plen > 128)
2928		return -EINVAL;
2929
2930	/* check the lifetime */
2931	if (!cfg->valid_lft || cfg->preferred_lft > cfg->valid_lft)
2932		return -EINVAL;
2933
2934	if (cfg->ifa_flags & IFA_F_MANAGETEMPADDR && cfg->plen != 64)
2935		return -EINVAL;
2936
2937	dev = __dev_get_by_index(net, ifindex);
2938	if (!dev)
2939		return -ENODEV;
2940
2941	idev = addrconf_add_dev(dev);
2942	if (IS_ERR(idev))
2943		return PTR_ERR(idev);
2944
2945	if (cfg->ifa_flags & IFA_F_MCAUTOJOIN) {
2946		int ret = ipv6_mc_config(net->ipv6.mc_autojoin_sk,
2947					 true, cfg->pfx, ifindex);
2948
2949		if (ret < 0)
2950			return ret;
2951	}
2952
2953	cfg->scope = ipv6_addr_scope(cfg->pfx);
2954
2955	timeout = addrconf_timeout_fixup(cfg->valid_lft, HZ);
2956	if (addrconf_finite_timeout(timeout)) {
2957		expires = jiffies_to_clock_t(timeout * HZ);
2958		cfg->valid_lft = timeout;
2959		flags = RTF_EXPIRES;
2960	} else {
2961		expires = 0;
2962		flags = 0;
2963		cfg->ifa_flags |= IFA_F_PERMANENT;
2964	}
2965
2966	timeout = addrconf_timeout_fixup(cfg->preferred_lft, HZ);
2967	if (addrconf_finite_timeout(timeout)) {
2968		if (timeout == 0)
2969			cfg->ifa_flags |= IFA_F_DEPRECATED;
2970		cfg->preferred_lft = timeout;
2971	}
2972
2973	ifp = ipv6_add_addr(idev, cfg, true, extack);
2974	if (!IS_ERR(ifp)) {
2975		if (!(cfg->ifa_flags & IFA_F_NOPREFIXROUTE)) {
2976			addrconf_prefix_route(&ifp->addr, ifp->prefix_len,
2977					      ifp->rt_priority, dev, expires,
2978					      flags, GFP_KERNEL);
2979		}
2980
2981		/* Send a netlink notification if DAD is enabled and
2982		 * optimistic flag is not set
2983		 */
2984		if (!(ifp->flags & (IFA_F_OPTIMISTIC | IFA_F_NODAD)))
2985			ipv6_ifa_notify(0, ifp);
2986		/*
2987		 * Note that section 3.1 of RFC 4429 indicates
2988		 * that the Optimistic flag should not be set for
2989		 * manually configured addresses
2990		 */
2991		addrconf_dad_start(ifp);
2992		if (cfg->ifa_flags & IFA_F_MANAGETEMPADDR)
2993			manage_tempaddrs(idev, ifp, cfg->valid_lft,
2994					 cfg->preferred_lft, true, jiffies);
2995		in6_ifa_put(ifp);
2996		addrconf_verify_rtnl();
2997		return 0;
2998	} else if (cfg->ifa_flags & IFA_F_MCAUTOJOIN) {
2999		ipv6_mc_config(net->ipv6.mc_autojoin_sk, false,
3000			       cfg->pfx, ifindex);
3001	}
3002
3003	return PTR_ERR(ifp);
3004}
3005
3006static int inet6_addr_del(struct net *net, int ifindex, u32 ifa_flags,
3007			  const struct in6_addr *pfx, unsigned int plen)
3008{
3009	struct inet6_ifaddr *ifp;
3010	struct inet6_dev *idev;
3011	struct net_device *dev;
3012
3013	if (plen > 128)
3014		return -EINVAL;
3015
3016	dev = __dev_get_by_index(net, ifindex);
3017	if (!dev)
3018		return -ENODEV;
3019
3020	idev = __in6_dev_get(dev);
3021	if (!idev)
3022		return -ENXIO;
3023
3024	read_lock_bh(&idev->lock);
3025	list_for_each_entry(ifp, &idev->addr_list, if_list) {
3026		if (ifp->prefix_len == plen &&
3027		    ipv6_addr_equal(pfx, &ifp->addr)) {
3028			in6_ifa_hold(ifp);
3029			read_unlock_bh(&idev->lock);
3030
3031			if (!(ifp->flags & IFA_F_TEMPORARY) &&
3032			    (ifa_flags & IFA_F_MANAGETEMPADDR))
3033				manage_tempaddrs(idev, ifp, 0, 0, false,
3034						 jiffies);
3035			ipv6_del_addr(ifp);
3036			addrconf_verify_rtnl();
3037			if (ipv6_addr_is_multicast(pfx)) {
3038				ipv6_mc_config(net->ipv6.mc_autojoin_sk,
3039					       false, pfx, dev->ifindex);
3040			}
3041			return 0;
3042		}
3043	}
3044	read_unlock_bh(&idev->lock);
3045	return -EADDRNOTAVAIL;
3046}
3047
3048
3049int addrconf_add_ifaddr(struct net *net, void __user *arg)
3050{
3051	struct ifa6_config cfg = {
3052		.ifa_flags = IFA_F_PERMANENT,
3053		.preferred_lft = INFINITY_LIFE_TIME,
3054		.valid_lft = INFINITY_LIFE_TIME,
3055	};
3056	struct in6_ifreq ireq;
3057	int err;
3058
3059	if (!ns_capable(net->user_ns, CAP_NET_ADMIN))
3060		return -EPERM;
3061
3062	if (copy_from_user(&ireq, arg, sizeof(struct in6_ifreq)))
3063		return -EFAULT;
3064
3065	cfg.pfx = &ireq.ifr6_addr;
3066	cfg.plen = ireq.ifr6_prefixlen;
3067
3068	rtnl_lock();
3069	err = inet6_addr_add(net, ireq.ifr6_ifindex, &cfg, NULL);
3070	rtnl_unlock();
3071	return err;
3072}
3073
3074int addrconf_del_ifaddr(struct net *net, void __user *arg)
3075{
3076	struct in6_ifreq ireq;
3077	int err;
3078
3079	if (!ns_capable(net->user_ns, CAP_NET_ADMIN))
3080		return -EPERM;
3081
3082	if (copy_from_user(&ireq, arg, sizeof(struct in6_ifreq)))
3083		return -EFAULT;
3084
3085	rtnl_lock();
3086	err = inet6_addr_del(net, ireq.ifr6_ifindex, 0, &ireq.ifr6_addr,
3087			     ireq.ifr6_prefixlen);
3088	rtnl_unlock();
3089	return err;
3090}
3091
3092static void add_addr(struct inet6_dev *idev, const struct in6_addr *addr,
3093		     int plen, int scope)
3094{
3095	struct inet6_ifaddr *ifp;
3096	struct ifa6_config cfg = {
3097		.pfx = addr,
3098		.plen = plen,
3099		.ifa_flags = IFA_F_PERMANENT,
3100		.valid_lft = INFINITY_LIFE_TIME,
3101		.preferred_lft = INFINITY_LIFE_TIME,
3102		.scope = scope
3103	};
3104
3105	ifp = ipv6_add_addr(idev, &cfg, true, NULL);
3106	if (!IS_ERR(ifp)) {
3107		spin_lock_bh(&ifp->lock);
3108		ifp->flags &= ~IFA_F_TENTATIVE;
3109		spin_unlock_bh(&ifp->lock);
3110		rt_genid_bump_ipv6(dev_net(idev->dev));
3111		ipv6_ifa_notify(RTM_NEWADDR, ifp);
3112		in6_ifa_put(ifp);
3113	}
3114}
3115
3116#if IS_ENABLED(CONFIG_IPV6_SIT)
3117static void sit_add_v4_addrs(struct inet6_dev *idev)
3118{
3119	struct in6_addr addr;
3120	struct net_device *dev;
3121	struct net *net = dev_net(idev->dev);
3122	int scope, plen;
3123	u32 pflags = 0;
3124
3125	ASSERT_RTNL();
3126
3127	memset(&addr, 0, sizeof(struct in6_addr));
3128	memcpy(&addr.s6_addr32[3], idev->dev->dev_addr, 4);
3129
3130	if (idev->dev->flags&IFF_POINTOPOINT) {
3131		if (idev->cnf.addr_gen_mode == IN6_ADDR_GEN_MODE_NONE)
3132			return;
3133
3134		addr.s6_addr32[0] = htonl(0xfe800000);
3135		scope = IFA_LINK;
3136		plen = 64;
3137	} else {
3138		scope = IPV6_ADDR_COMPATv4;
3139		plen = 96;
3140		pflags |= RTF_NONEXTHOP;
3141	}
3142
3143	if (addr.s6_addr32[3]) {
3144		add_addr(idev, &addr, plen, scope);
3145		addrconf_prefix_route(&addr, plen, 0, idev->dev, 0, pflags,
3146				      GFP_KERNEL);
3147		return;
3148	}
3149
3150	for_each_netdev(net, dev) {
3151		struct in_device *in_dev = __in_dev_get_rtnl(dev);
3152		if (in_dev && (dev->flags & IFF_UP)) {
3153			struct in_ifaddr *ifa;
3154			int flag = scope;
3155
3156			in_dev_for_each_ifa_rtnl(ifa, in_dev) {
3157				addr.s6_addr32[3] = ifa->ifa_local;
3158
3159				if (ifa->ifa_scope == RT_SCOPE_LINK)
3160					continue;
3161				if (ifa->ifa_scope >= RT_SCOPE_HOST) {
3162					if (idev->dev->flags&IFF_POINTOPOINT)
3163						continue;
3164					flag |= IFA_HOST;
3165				}
3166
3167				add_addr(idev, &addr, plen, flag);
3168				addrconf_prefix_route(&addr, plen, 0, idev->dev,
3169						      0, pflags, GFP_KERNEL);
3170			}
3171		}
3172	}
3173}
3174#endif
3175
3176static void init_loopback(struct net_device *dev)
3177{
3178	struct inet6_dev  *idev;
3179
3180	/* ::1 */
3181
3182	ASSERT_RTNL();
3183
3184	idev = ipv6_find_idev(dev);
3185	if (IS_ERR(idev)) {
3186		pr_debug("%s: add_dev failed\n", __func__);
3187		return;
3188	}
3189
3190	add_addr(idev, &in6addr_loopback, 128, IFA_HOST);
3191}
3192
3193void addrconf_add_linklocal(struct inet6_dev *idev,
3194			    const struct in6_addr *addr, u32 flags)
3195{
3196	struct ifa6_config cfg = {
3197		.pfx = addr,
3198		.plen = 64,
3199		.ifa_flags = flags | IFA_F_PERMANENT,
3200		.valid_lft = INFINITY_LIFE_TIME,
3201		.preferred_lft = INFINITY_LIFE_TIME,
3202		.scope = IFA_LINK
3203	};
3204	struct inet6_ifaddr *ifp;
3205
3206#ifdef CONFIG_IPV6_OPTIMISTIC_DAD
3207	if ((dev_net(idev->dev)->ipv6.devconf_all->optimistic_dad ||
3208	     idev->cnf.optimistic_dad) &&
3209	    !dev_net(idev->dev)->ipv6.devconf_all->forwarding)
3210		cfg.ifa_flags |= IFA_F_OPTIMISTIC;
3211#endif
3212
3213	ifp = ipv6_add_addr(idev, &cfg, true, NULL);
3214	if (!IS_ERR(ifp)) {
3215		addrconf_prefix_route(&ifp->addr, ifp->prefix_len, 0, idev->dev,
3216				      0, 0, GFP_ATOMIC);
3217		addrconf_dad_start(ifp);
3218		in6_ifa_put(ifp);
3219	}
3220}
3221EXPORT_SYMBOL_GPL(addrconf_add_linklocal);
3222
3223static bool ipv6_reserved_interfaceid(struct in6_addr address)
3224{
3225	if ((address.s6_addr32[2] | address.s6_addr32[3]) == 0)
3226		return true;
3227
3228	if (address.s6_addr32[2] == htonl(0x02005eff) &&
3229	    ((address.s6_addr32[3] & htonl(0xfe000000)) == htonl(0xfe000000)))
3230		return true;
3231
3232	if (address.s6_addr32[2] == htonl(0xfdffffff) &&
3233	    ((address.s6_addr32[3] & htonl(0xffffff80)) == htonl(0xffffff80)))
3234		return true;
3235
3236	return false;
3237}
3238
3239static int ipv6_generate_stable_address(struct in6_addr *address,
3240					u8 dad_count,
3241					const struct inet6_dev *idev)
3242{
3243	static DEFINE_SPINLOCK(lock);
3244	static __u32 digest[SHA1_DIGEST_WORDS];
3245	static __u32 workspace[SHA1_WORKSPACE_WORDS];
3246
3247	static union {
3248		char __data[SHA1_BLOCK_SIZE];
3249		struct {
3250			struct in6_addr secret;
3251			__be32 prefix[2];
3252			unsigned char hwaddr[MAX_ADDR_LEN];
3253			u8 dad_count;
3254		} __packed;
3255	} data;
3256
3257	struct in6_addr secret;
3258	struct in6_addr temp;
3259	struct net *net = dev_net(idev->dev);
3260
3261	BUILD_BUG_ON(sizeof(data.__data) != sizeof(data));
3262
3263	if (idev->cnf.stable_secret.initialized)
3264		secret = idev->cnf.stable_secret.secret;
3265	else if (net->ipv6.devconf_dflt->stable_secret.initialized)
3266		secret = net->ipv6.devconf_dflt->stable_secret.secret;
3267	else
3268		return -1;
3269
3270retry:
3271	spin_lock_bh(&lock);
3272
3273	sha1_init(digest);
3274	memset(&data, 0, sizeof(data));
3275	memset(workspace, 0, sizeof(workspace));
3276	memcpy(data.hwaddr, idev->dev->perm_addr, idev->dev->addr_len);
3277	data.prefix[0] = address->s6_addr32[0];
3278	data.prefix[1] = address->s6_addr32[1];
3279	data.secret = secret;
3280	data.dad_count = dad_count;
3281
3282	sha1_transform(digest, data.__data, workspace);
3283
3284	temp = *address;
3285	temp.s6_addr32[2] = (__force __be32)digest[0];
3286	temp.s6_addr32[3] = (__force __be32)digest[1];
3287
3288	spin_unlock_bh(&lock);
3289
3290	if (ipv6_reserved_interfaceid(temp)) {
3291		dad_count++;
3292		if (dad_count > dev_net(idev->dev)->ipv6.sysctl.idgen_retries)
3293			return -1;
3294		goto retry;
3295	}
3296
3297	*address = temp;
3298	return 0;
3299}
3300
3301static void ipv6_gen_mode_random_init(struct inet6_dev *idev)
3302{
3303	struct ipv6_stable_secret *s = &idev->cnf.stable_secret;
3304
3305	if (s->initialized)
3306		return;
3307	s = &idev->cnf.stable_secret;
3308	get_random_bytes(&s->secret, sizeof(s->secret));
3309	s->initialized = true;
3310}
3311
3312static void addrconf_addr_gen(struct inet6_dev *idev, bool prefix_route)
3313{
3314	struct in6_addr addr;
3315
3316	/* no link local addresses on L3 master devices */
3317	if (netif_is_l3_master(idev->dev))
3318		return;
3319
3320	/* no link local addresses on devices flagged as slaves */
3321	if (idev->dev->flags & IFF_SLAVE)
3322		return;
3323
3324	ipv6_addr_set(&addr, htonl(0xFE800000), 0, 0, 0);
3325
3326	switch (idev->cnf.addr_gen_mode) {
3327	case IN6_ADDR_GEN_MODE_RANDOM:
3328		ipv6_gen_mode_random_init(idev);
3329		fallthrough;
3330	case IN6_ADDR_GEN_MODE_STABLE_PRIVACY:
3331		if (!ipv6_generate_stable_address(&addr, 0, idev))
3332			addrconf_add_linklocal(idev, &addr,
3333					       IFA_F_STABLE_PRIVACY);
3334		else if (prefix_route)
3335			addrconf_prefix_route(&addr, 64, 0, idev->dev,
3336					      0, 0, GFP_KERNEL);
3337		break;
3338	case IN6_ADDR_GEN_MODE_EUI64:
3339		/* addrconf_add_linklocal also adds a prefix_route and we
3340		 * only need to care about prefix routes if ipv6_generate_eui64
3341		 * couldn't generate one.
3342		 */
3343		if (ipv6_generate_eui64(addr.s6_addr + 8, idev->dev) == 0)
3344			addrconf_add_linklocal(idev, &addr, 0);
3345		else if (prefix_route)
3346			addrconf_prefix_route(&addr, 64, 0, idev->dev,
3347					      0, 0, GFP_KERNEL);
3348		break;
3349	case IN6_ADDR_GEN_MODE_NONE:
3350	default:
3351		/* will not add any link local address */
3352		break;
3353	}
3354}
3355
3356static void addrconf_dev_config(struct net_device *dev)
3357{
3358	struct inet6_dev *idev;
3359
3360	ASSERT_RTNL();
3361
3362	if ((dev->type != ARPHRD_ETHER) &&
3363	    (dev->type != ARPHRD_FDDI) &&
3364	    (dev->type != ARPHRD_ARCNET) &&
3365	    (dev->type != ARPHRD_INFINIBAND) &&
3366	    (dev->type != ARPHRD_IEEE1394) &&
3367	    (dev->type != ARPHRD_TUNNEL6) &&
3368	    (dev->type != ARPHRD_6LOWPAN) &&
3369	    (dev->type != ARPHRD_IP6GRE) &&
3370	    (dev->type != ARPHRD_IPGRE) &&
3371	    (dev->type != ARPHRD_TUNNEL) &&
3372	    (dev->type != ARPHRD_NONE) &&
3373	    (dev->type != ARPHRD_RAWIP)) {
3374		/* Alas, we support only Ethernet autoconfiguration. */
3375		idev = __in6_dev_get(dev);
3376		if (!IS_ERR_OR_NULL(idev) && dev->flags & IFF_UP &&
3377		    dev->flags & IFF_MULTICAST)
3378			ipv6_mc_up(idev);
3379		return;
3380	}
3381
3382	idev = addrconf_add_dev(dev);
3383	if (IS_ERR(idev))
3384		return;
3385
3386	/* this device type has no EUI support */
3387	if (dev->type == ARPHRD_NONE &&
3388	    idev->cnf.addr_gen_mode == IN6_ADDR_GEN_MODE_EUI64)
3389		idev->cnf.addr_gen_mode = IN6_ADDR_GEN_MODE_RANDOM;
3390
3391	addrconf_addr_gen(idev, false);
3392}
3393
3394#if IS_ENABLED(CONFIG_IPV6_SIT)
3395static void addrconf_sit_config(struct net_device *dev)
3396{
3397	struct inet6_dev *idev;
3398
3399	ASSERT_RTNL();
3400
3401	/*
3402	 * Configure the tunnel with one of our IPv4
3403	 * addresses... we should configure all of
3404	 * our v4 addrs in the tunnel
3405	 */
3406
3407	idev = ipv6_find_idev(dev);
3408	if (IS_ERR(idev)) {
3409		pr_debug("%s: add_dev failed\n", __func__);
3410		return;
3411	}
3412
3413	if (dev->priv_flags & IFF_ISATAP) {
3414		addrconf_addr_gen(idev, false);
3415		return;
3416	}
3417
3418	sit_add_v4_addrs(idev);
3419
3420	if (dev->flags&IFF_POINTOPOINT)
3421		addrconf_add_mroute(dev);
3422}
3423#endif
3424
3425#if IS_ENABLED(CONFIG_NET_IPGRE)
3426static void addrconf_gre_config(struct net_device *dev)
3427{
3428	struct inet6_dev *idev;
3429
3430	ASSERT_RTNL();
3431
3432	idev = ipv6_find_idev(dev);
3433	if (IS_ERR(idev)) {
3434		pr_debug("%s: add_dev failed\n", __func__);
3435		return;
3436	}
3437
3438	addrconf_addr_gen(idev, true);
3439	if (dev->flags & IFF_POINTOPOINT)
3440		addrconf_add_mroute(dev);
3441}
3442#endif
3443
3444static int fixup_permanent_addr(struct net *net,
3445				struct inet6_dev *idev,
3446				struct inet6_ifaddr *ifp)
3447{
3448	/* !fib6_node means the host route was removed from the
3449	 * FIB, for example, if 'lo' device is taken down. In that
3450	 * case regenerate the host route.
3451	 */
3452	if (!ifp->rt || !ifp->rt->fib6_node) {
3453		struct fib6_info *f6i, *prev;
3454
3455		f6i = addrconf_f6i_alloc(net, idev, &ifp->addr, false,
3456					 GFP_ATOMIC);
3457		if (IS_ERR(f6i))
3458			return PTR_ERR(f6i);
3459
3460		/* ifp->rt can be accessed outside of rtnl */
3461		spin_lock(&ifp->lock);
3462		prev = ifp->rt;
3463		ifp->rt = f6i;
3464		spin_unlock(&ifp->lock);
3465
3466		fib6_info_release(prev);
3467	}
3468
3469	if (!(ifp->flags & IFA_F_NOPREFIXROUTE)) {
3470		addrconf_prefix_route(&ifp->addr, ifp->prefix_len,
3471				      ifp->rt_priority, idev->dev, 0, 0,
3472				      GFP_ATOMIC);
3473	}
3474
3475	if (ifp->state == INET6_IFADDR_STATE_PREDAD)
3476		addrconf_dad_start(ifp);
3477
3478	return 0;
3479}
3480
3481static void addrconf_permanent_addr(struct net *net, struct net_device *dev)
3482{
3483	struct inet6_ifaddr *ifp, *tmp;
3484	struct inet6_dev *idev;
3485
3486	idev = __in6_dev_get(dev);
3487	if (!idev)
3488		return;
3489
3490	write_lock_bh(&idev->lock);
3491
3492	list_for_each_entry_safe(ifp, tmp, &idev->addr_list, if_list) {
3493		if ((ifp->flags & IFA_F_PERMANENT) &&
3494		    fixup_permanent_addr(net, idev, ifp) < 0) {
3495			write_unlock_bh(&idev->lock);
3496			in6_ifa_hold(ifp);
3497			ipv6_del_addr(ifp);
3498			write_lock_bh(&idev->lock);
3499
3500			net_info_ratelimited("%s: Failed to add prefix route for address %pI6c; dropping\n",
3501					     idev->dev->name, &ifp->addr);
3502		}
3503	}
3504
3505	write_unlock_bh(&idev->lock);
3506}
3507
3508static int addrconf_notify(struct notifier_block *this, unsigned long event,
3509			   void *ptr)
3510{
3511	struct net_device *dev = netdev_notifier_info_to_dev(ptr);
3512	struct netdev_notifier_change_info *change_info;
3513	struct netdev_notifier_changeupper_info *info;
3514	struct inet6_dev *idev = __in6_dev_get(dev);
3515	struct net *net = dev_net(dev);
3516	int run_pending = 0;
3517	int err;
3518
3519	switch (event) {
3520	case NETDEV_REGISTER:
3521		if (!idev && dev->mtu >= IPV6_MIN_MTU) {
3522			idev = ipv6_add_dev(dev);
3523			if (IS_ERR(idev))
3524				return notifier_from_errno(PTR_ERR(idev));
3525		}
3526		break;
3527
3528	case NETDEV_CHANGEMTU:
3529		/* if MTU under IPV6_MIN_MTU stop IPv6 on this interface. */
3530		if (dev->mtu < IPV6_MIN_MTU) {
3531			addrconf_ifdown(dev, dev != net->loopback_dev);
3532			break;
3533		}
3534
3535		if (idev) {
3536			rt6_mtu_change(dev, dev->mtu);
3537			idev->cnf.mtu6 = dev->mtu;
3538			break;
3539		}
3540
3541		/* allocate new idev */
3542		idev = ipv6_add_dev(dev);
3543		if (IS_ERR(idev))
3544			break;
3545
3546		/* device is still not ready */
3547		if (!(idev->if_flags & IF_READY))
3548			break;
3549
3550		run_pending = 1;
3551		fallthrough;
3552	case NETDEV_UP:
3553	case NETDEV_CHANGE:
3554		if (dev->flags & IFF_SLAVE)
3555			break;
3556
3557		if (idev && idev->cnf.disable_ipv6)
3558			break;
3559
3560		if (event == NETDEV_UP) {
3561			/* restore routes for permanent addresses */
3562			addrconf_permanent_addr(net, dev);
3563
3564			if (!addrconf_link_ready(dev)) {
3565				/* device is not ready yet. */
3566				pr_debug("ADDRCONF(NETDEV_UP): %s: link is not ready\n",
3567					 dev->name);
3568				break;
3569			}
3570
3571			if (!idev && dev->mtu >= IPV6_MIN_MTU)
3572				idev = ipv6_add_dev(dev);
3573
3574			if (!IS_ERR_OR_NULL(idev)) {
3575				idev->if_flags |= IF_READY;
3576				run_pending = 1;
3577			}
3578		} else if (event == NETDEV_CHANGE) {
3579			if (!addrconf_link_ready(dev)) {
3580				/* device is still not ready. */
3581				rt6_sync_down_dev(dev, event);
3582				break;
3583			}
3584
3585			if (!IS_ERR_OR_NULL(idev)) {
3586				if (idev->if_flags & IF_READY) {
3587					/* device is already configured -
3588					 * but resend MLD reports, we might
3589					 * have roamed and need to update
3590					 * multicast snooping switches
3591					 */
3592					ipv6_mc_up(idev);
3593					change_info = ptr;
3594					if (change_info->flags_changed & IFF_NOARP)
3595						addrconf_dad_run(idev, true);
3596					rt6_sync_up(dev, RTNH_F_LINKDOWN);
3597					break;
3598				}
3599				idev->if_flags |= IF_READY;
3600			}
3601
3602			pr_info("ADDRCONF(NETDEV_CHANGE): %s: link becomes ready\n",
3603				dev->name);
3604
3605			run_pending = 1;
3606		}
3607
3608		switch (dev->type) {
3609#if IS_ENABLED(CONFIG_IPV6_SIT)
3610		case ARPHRD_SIT:
3611			addrconf_sit_config(dev);
3612			break;
3613#endif
3614#if IS_ENABLED(CONFIG_NET_IPGRE)
3615		case ARPHRD_IPGRE:
3616			addrconf_gre_config(dev);
3617			break;
3618#endif
3619		case ARPHRD_LOOPBACK:
3620			init_loopback(dev);
3621			break;
3622
3623		default:
3624			addrconf_dev_config(dev);
3625			break;
3626		}
3627
3628		if (!IS_ERR_OR_NULL(idev)) {
3629			if (run_pending)
3630				addrconf_dad_run(idev, false);
3631
3632			/* Device has an address by now */
3633			rt6_sync_up(dev, RTNH_F_DEAD);
3634
3635			/*
3636			 * If the MTU changed during the interface down,
3637			 * when the interface up, the changed MTU must be
3638			 * reflected in the idev as well as routers.
3639			 */
3640			if (idev->cnf.mtu6 != dev->mtu &&
3641			    dev->mtu >= IPV6_MIN_MTU) {
3642				rt6_mtu_change(dev, dev->mtu);
3643				idev->cnf.mtu6 = dev->mtu;
3644			}
3645			idev->tstamp = jiffies;
3646			inet6_ifinfo_notify(RTM_NEWLINK, idev);
3647
3648			/*
3649			 * If the changed mtu during down is lower than
3650			 * IPV6_MIN_MTU stop IPv6 on this interface.
3651			 */
3652			if (dev->mtu < IPV6_MIN_MTU)
3653				addrconf_ifdown(dev, dev != net->loopback_dev);
3654		}
3655		break;
3656
3657	case NETDEV_DOWN:
3658	case NETDEV_UNREGISTER:
3659		/*
3660		 *	Remove all addresses from this interface.
3661		 */
3662		addrconf_ifdown(dev, event != NETDEV_DOWN);
3663		break;
3664
3665	case NETDEV_CHANGENAME:
3666		if (idev) {
3667			snmp6_unregister_dev(idev);
3668			addrconf_sysctl_unregister(idev);
3669			err = addrconf_sysctl_register(idev);
3670			if (err)
3671				return notifier_from_errno(err);
3672			err = snmp6_register_dev(idev);
3673			if (err) {
3674				addrconf_sysctl_unregister(idev);
3675				return notifier_from_errno(err);
3676			}
3677		}
3678		break;
3679
3680	case NETDEV_PRE_TYPE_CHANGE:
3681	case NETDEV_POST_TYPE_CHANGE:
3682		if (idev)
3683			addrconf_type_change(dev, event);
3684		break;
3685
3686	case NETDEV_CHANGEUPPER:
3687		info = ptr;
3688
3689		/* flush all routes if dev is linked to or unlinked from
3690		 * an L3 master device (e.g., VRF)
3691		 */
3692		if (info->upper_dev && netif_is_l3_master(info->upper_dev))
3693			addrconf_ifdown(dev, false);
3694	}
3695
3696	return NOTIFY_OK;
3697}
3698
3699/*
3700 *	addrconf module should be notified of a device going up
3701 */
3702static struct notifier_block ipv6_dev_notf = {
3703	.notifier_call = addrconf_notify,
3704	.priority = ADDRCONF_NOTIFY_PRIORITY,
3705};
3706
3707static void addrconf_type_change(struct net_device *dev, unsigned long event)
3708{
3709	struct inet6_dev *idev;
3710	ASSERT_RTNL();
3711
3712	idev = __in6_dev_get(dev);
3713
3714	if (event == NETDEV_POST_TYPE_CHANGE)
3715		ipv6_mc_remap(idev);
3716	else if (event == NETDEV_PRE_TYPE_CHANGE)
3717		ipv6_mc_unmap(idev);
3718}
3719
3720static bool addr_is_local(const struct in6_addr *addr)
3721{
3722	return ipv6_addr_type(addr) &
3723		(IPV6_ADDR_LINKLOCAL | IPV6_ADDR_LOOPBACK);
3724}
3725
3726static int addrconf_ifdown(struct net_device *dev, bool unregister)
3727{
3728	unsigned long event = unregister ? NETDEV_UNREGISTER : NETDEV_DOWN;
3729	struct net *net = dev_net(dev);
3730	struct inet6_dev *idev;
3731	struct inet6_ifaddr *ifa;
3732	LIST_HEAD(tmp_addr_list);
3733	bool keep_addr = false;
3734	bool was_ready;
3735	int state, i;
3736
3737	ASSERT_RTNL();
3738
3739	rt6_disable_ip(dev, event);
3740
3741	idev = __in6_dev_get(dev);
3742	if (!idev)
3743		return -ENODEV;
3744
3745	/*
3746	 * Step 1: remove reference to ipv6 device from parent device.
3747	 *	   Do not dev_put!
3748	 */
3749	if (unregister) {
3750		idev->dead = 1;
3751
3752		/* protected by rtnl_lock */
3753		RCU_INIT_POINTER(dev->ip6_ptr, NULL);
3754
3755		/* Step 1.5: remove snmp6 entry */
3756		snmp6_unregister_dev(idev);
3757
3758	}
3759
3760	/* combine the user config with event to determine if permanent
3761	 * addresses are to be removed from address hash table
3762	 */
3763	if (!unregister && !idev->cnf.disable_ipv6) {
3764		/* aggregate the system setting and interface setting */
3765		int _keep_addr = net->ipv6.devconf_all->keep_addr_on_down;
3766
3767		if (!_keep_addr)
3768			_keep_addr = idev->cnf.keep_addr_on_down;
3769
3770		keep_addr = (_keep_addr > 0);
3771	}
3772
3773	/* Step 2: clear hash table */
3774	for (i = 0; i < IN6_ADDR_HSIZE; i++) {
3775		struct hlist_head *h = &inet6_addr_lst[i];
3776
3777		spin_lock_bh(&addrconf_hash_lock);
3778restart:
3779		hlist_for_each_entry_rcu(ifa, h, addr_lst) {
3780			if (ifa->idev == idev) {
3781				addrconf_del_dad_work(ifa);
3782				/* combined flag + permanent flag decide if
3783				 * address is retained on a down event
3784				 */
3785				if (!keep_addr ||
3786				    !(ifa->flags & IFA_F_PERMANENT) ||
3787				    addr_is_local(&ifa->addr)) {
3788					hlist_del_init_rcu(&ifa->addr_lst);
3789					goto restart;
3790				}
3791			}
3792		}
3793		spin_unlock_bh(&addrconf_hash_lock);
3794	}
3795
3796	write_lock_bh(&idev->lock);
3797
3798	addrconf_del_rs_timer(idev);
3799
3800	/* Step 2: clear flags for stateless addrconf, repeated down
3801	 *         detection
3802	 */
3803	was_ready = idev->if_flags & IF_READY;
3804	if (!unregister)
3805		idev->if_flags &= ~(IF_RS_SENT|IF_RA_RCVD|IF_READY);
3806
3807	/* Step 3: clear tempaddr list */
3808	while (!list_empty(&idev->tempaddr_list)) {
3809		ifa = list_first_entry(&idev->tempaddr_list,
3810				       struct inet6_ifaddr, tmp_list);
3811		list_del(&ifa->tmp_list);
3812		write_unlock_bh(&idev->lock);
3813		spin_lock_bh(&ifa->lock);
3814
3815		if (ifa->ifpub) {
3816			in6_ifa_put(ifa->ifpub);
3817			ifa->ifpub = NULL;
3818		}
3819		spin_unlock_bh(&ifa->lock);
3820		in6_ifa_put(ifa);
3821		write_lock_bh(&idev->lock);
3822	}
3823
3824	list_for_each_entry(ifa, &idev->addr_list, if_list)
3825		list_add_tail(&ifa->if_list_aux, &tmp_addr_list);
3826	write_unlock_bh(&idev->lock);
3827
3828	while (!list_empty(&tmp_addr_list)) {
3829		struct fib6_info *rt = NULL;
3830		bool keep;
3831
3832		ifa = list_first_entry(&tmp_addr_list,
3833				       struct inet6_ifaddr, if_list_aux);
3834		list_del(&ifa->if_list_aux);
3835
3836		addrconf_del_dad_work(ifa);
3837
3838		keep = keep_addr && (ifa->flags & IFA_F_PERMANENT) &&
3839			!addr_is_local(&ifa->addr);
3840
3841		spin_lock_bh(&ifa->lock);
3842
3843		if (keep) {
3844			/* set state to skip the notifier below */
3845			state = INET6_IFADDR_STATE_DEAD;
3846			ifa->state = INET6_IFADDR_STATE_PREDAD;
3847			if (!(ifa->flags & IFA_F_NODAD))
3848				ifa->flags |= IFA_F_TENTATIVE;
3849
3850			rt = ifa->rt;
3851			ifa->rt = NULL;
3852		} else {
3853			state = ifa->state;
3854			ifa->state = INET6_IFADDR_STATE_DEAD;
3855		}
3856
3857		spin_unlock_bh(&ifa->lock);
3858
3859		if (rt)
3860			ip6_del_rt(net, rt, false);
3861
3862		if (state != INET6_IFADDR_STATE_DEAD) {
3863			__ipv6_ifa_notify(RTM_DELADDR, ifa);
3864			inet6addr_notifier_call_chain(NETDEV_DOWN, ifa);
3865		} else {
3866			if (idev->cnf.forwarding)
3867				addrconf_leave_anycast(ifa);
3868			addrconf_leave_solict(ifa->idev, &ifa->addr);
3869		}
3870
3871		if (!keep) {
3872			write_lock_bh(&idev->lock);
3873			list_del_rcu(&ifa->if_list);
3874			write_unlock_bh(&idev->lock);
3875			in6_ifa_put(ifa);
3876		}
3877	}
3878
3879	/* Step 5: Discard anycast and multicast list */
3880	if (unregister) {
3881		ipv6_ac_destroy_dev(idev);
3882		ipv6_mc_destroy_dev(idev);
3883	} else if (was_ready) {
3884		ipv6_mc_down(idev);
3885	}
3886
3887	idev->tstamp = jiffies;
3888
3889	/* Last: Shot the device (if unregistered) */
3890	if (unregister) {
3891		addrconf_sysctl_unregister(idev);
3892		neigh_parms_release(&nd_tbl, idev->nd_parms);
3893		neigh_ifdown(&nd_tbl, dev);
3894		in6_dev_put(idev);
3895	}
3896	return 0;
3897}
3898
3899static void addrconf_rs_timer(struct timer_list *t)
3900{
3901	struct inet6_dev *idev = from_timer(idev, t, rs_timer);
3902	struct net_device *dev = idev->dev;
3903	struct in6_addr lladdr;
3904
3905	write_lock(&idev->lock);
3906	if (idev->dead || !(idev->if_flags & IF_READY))
3907		goto out;
3908
3909	if (!ipv6_accept_ra(idev))
3910		goto out;
3911
3912	/* Announcement received after solicitation was sent */
3913	if (idev->if_flags & IF_RA_RCVD)
3914		goto out;
3915
3916	if (idev->rs_probes++ < idev->cnf.rtr_solicits || idev->cnf.rtr_solicits < 0) {
3917		write_unlock(&idev->lock);
3918		if (!ipv6_get_lladdr(dev, &lladdr, IFA_F_TENTATIVE))
3919			ndisc_send_rs(dev, &lladdr,
3920				      &in6addr_linklocal_allrouters);
3921		else
3922			goto put;
3923
3924		write_lock(&idev->lock);
3925		idev->rs_interval = rfc3315_s14_backoff_update(
3926			idev->rs_interval, idev->cnf.rtr_solicit_max_interval);
3927		/* The wait after the last probe can be shorter */
3928		addrconf_mod_rs_timer(idev, (idev->rs_probes ==
3929					     idev->cnf.rtr_solicits) ?
3930				      idev->cnf.rtr_solicit_delay :
3931				      idev->rs_interval);
3932	} else {
3933		/*
3934		 * Note: we do not support deprecated "all on-link"
3935		 * assumption any longer.
3936		 */
3937		pr_debug("%s: no IPv6 routers present\n", idev->dev->name);
3938	}
3939
3940out:
3941	write_unlock(&idev->lock);
3942put:
3943	in6_dev_put(idev);
3944}
3945
3946/*
3947 *	Duplicate Address Detection
3948 */
3949static void addrconf_dad_kick(struct inet6_ifaddr *ifp)
3950{
3951	unsigned long rand_num;
3952	struct inet6_dev *idev = ifp->idev;
3953	u64 nonce;
3954
3955	if (ifp->flags & IFA_F_OPTIMISTIC)
3956		rand_num = 0;
3957	else
3958		rand_num = prandom_u32() % (idev->cnf.rtr_solicit_delay ? : 1);
3959
3960	nonce = 0;
3961	if (idev->cnf.enhanced_dad ||
3962	    dev_net(idev->dev)->ipv6.devconf_all->enhanced_dad) {
3963		do
3964			get_random_bytes(&nonce, 6);
3965		while (nonce == 0);
3966	}
3967	ifp->dad_nonce = nonce;
3968	ifp->dad_probes = idev->cnf.dad_transmits;
3969	addrconf_mod_dad_work(ifp, rand_num);
3970}
3971
3972static void addrconf_dad_begin(struct inet6_ifaddr *ifp)
3973{
3974	struct inet6_dev *idev = ifp->idev;
3975	struct net_device *dev = idev->dev;
3976	bool bump_id, notify = false;
3977	struct net *net;
3978
3979	addrconf_join_solict(dev, &ifp->addr);
3980
3981	prandom_seed((__force u32) ifp->addr.s6_addr32[3]);
3982
3983	read_lock_bh(&idev->lock);
3984	spin_lock(&ifp->lock);
3985	if (ifp->state == INET6_IFADDR_STATE_DEAD)
3986		goto out;
3987
3988	net = dev_net(dev);
3989	if (dev->flags&(IFF_NOARP|IFF_LOOPBACK) ||
3990	    (net->ipv6.devconf_all->accept_dad < 1 &&
3991	     idev->cnf.accept_dad < 1) ||
3992	    !(ifp->flags&IFA_F_TENTATIVE) ||
3993	    ifp->flags & IFA_F_NODAD) {
3994		bool send_na = false;
3995
3996		if (ifp->flags & IFA_F_TENTATIVE &&
3997		    !(ifp->flags & IFA_F_OPTIMISTIC))
3998			send_na = true;
3999		bump_id = ifp->flags & IFA_F_TENTATIVE;
4000		ifp->flags &= ~(IFA_F_TENTATIVE|IFA_F_OPTIMISTIC|IFA_F_DADFAILED);
4001		spin_unlock(&ifp->lock);
4002		read_unlock_bh(&idev->lock);
4003
4004		addrconf_dad_completed(ifp, bump_id, send_na);
4005		return;
4006	}
4007
4008	if (!(idev->if_flags & IF_READY)) {
4009		spin_unlock(&ifp->lock);
4010		read_unlock_bh(&idev->lock);
4011		/*
4012		 * If the device is not ready:
4013		 * - keep it tentative if it is a permanent address.
4014		 * - otherwise, kill it.
4015		 */
4016		in6_ifa_hold(ifp);
4017		addrconf_dad_stop(ifp, 0);
4018		return;
4019	}
4020
4021	/*
4022	 * Optimistic nodes can start receiving
4023	 * Frames right away
4024	 */
4025	if (ifp->flags & IFA_F_OPTIMISTIC) {
4026		ip6_ins_rt(net, ifp->rt);
4027		if (ipv6_use_optimistic_addr(net, idev)) {
4028			/* Because optimistic nodes can use this address,
4029			 * notify listeners. If DAD fails, RTM_DELADDR is sent.
4030			 */
4031			notify = true;
4032		}
4033	}
4034
4035	addrconf_dad_kick(ifp);
4036out:
4037	spin_unlock(&ifp->lock);
4038	read_unlock_bh(&idev->lock);
4039	if (notify)
4040		ipv6_ifa_notify(RTM_NEWADDR, ifp);
4041}
4042
4043static void addrconf_dad_start(struct inet6_ifaddr *ifp)
4044{
4045	bool begin_dad = false;
4046
4047	spin_lock_bh(&ifp->lock);
4048	if (ifp->state != INET6_IFADDR_STATE_DEAD) {
4049		ifp->state = INET6_IFADDR_STATE_PREDAD;
4050		begin_dad = true;
4051	}
4052	spin_unlock_bh(&ifp->lock);
4053
4054	if (begin_dad)
4055		addrconf_mod_dad_work(ifp, 0);
4056}
4057
4058static void addrconf_dad_work(struct work_struct *w)
4059{
4060	struct inet6_ifaddr *ifp = container_of(to_delayed_work(w),
4061						struct inet6_ifaddr,
4062						dad_work);
4063	struct inet6_dev *idev = ifp->idev;
4064	bool bump_id, disable_ipv6 = false;
4065	struct in6_addr mcaddr;
4066
4067	enum {
4068		DAD_PROCESS,
4069		DAD_BEGIN,
4070		DAD_ABORT,
4071	} action = DAD_PROCESS;
4072
4073	rtnl_lock();
4074
4075	spin_lock_bh(&ifp->lock);
4076	if (ifp->state == INET6_IFADDR_STATE_PREDAD) {
4077		action = DAD_BEGIN;
4078		ifp->state = INET6_IFADDR_STATE_DAD;
4079	} else if (ifp->state == INET6_IFADDR_STATE_ERRDAD) {
4080		action = DAD_ABORT;
4081		ifp->state = INET6_IFADDR_STATE_POSTDAD;
4082
4083		if ((dev_net(idev->dev)->ipv6.devconf_all->accept_dad > 1 ||
4084		     idev->cnf.accept_dad > 1) &&
4085		    !idev->cnf.disable_ipv6 &&
4086		    !(ifp->flags & IFA_F_STABLE_PRIVACY)) {
4087			struct in6_addr addr;
4088
4089			addr.s6_addr32[0] = htonl(0xfe800000);
4090			addr.s6_addr32[1] = 0;
4091
4092			if (!ipv6_generate_eui64(addr.s6_addr + 8, idev->dev) &&
4093			    ipv6_addr_equal(&ifp->addr, &addr)) {
4094				/* DAD failed for link-local based on MAC */
4095				idev->cnf.disable_ipv6 = 1;
4096
4097				pr_info("%s: IPv6 being disabled!\n",
4098					ifp->idev->dev->name);
4099				disable_ipv6 = true;
4100			}
4101		}
4102	}
4103	spin_unlock_bh(&ifp->lock);
4104
4105	if (action == DAD_BEGIN) {
4106		addrconf_dad_begin(ifp);
4107		goto out;
4108	} else if (action == DAD_ABORT) {
4109		in6_ifa_hold(ifp);
4110		addrconf_dad_stop(ifp, 1);
4111		if (disable_ipv6)
4112			addrconf_ifdown(idev->dev, false);
4113		goto out;
4114	}
4115
4116	if (!ifp->dad_probes && addrconf_dad_end(ifp))
4117		goto out;
4118
4119	write_lock_bh(&idev->lock);
4120	if (idev->dead || !(idev->if_flags & IF_READY)) {
4121		write_unlock_bh(&idev->lock);
4122		goto out;
4123	}
4124
4125	spin_lock(&ifp->lock);
4126	if (ifp->state == INET6_IFADDR_STATE_DEAD) {
4127		spin_unlock(&ifp->lock);
4128		write_unlock_bh(&idev->lock);
4129		goto out;
4130	}
4131
4132	if (ifp->dad_probes == 0) {
4133		bool send_na = false;
4134
4135		/*
4136		 * DAD was successful
4137		 */
4138
4139		if (ifp->flags & IFA_F_TENTATIVE &&
4140		    !(ifp->flags & IFA_F_OPTIMISTIC))
4141			send_na = true;
4142		bump_id = ifp->flags & IFA_F_TENTATIVE;
4143		ifp->flags &= ~(IFA_F_TENTATIVE|IFA_F_OPTIMISTIC|IFA_F_DADFAILED);
4144		spin_unlock(&ifp->lock);
4145		write_unlock_bh(&idev->lock);
4146
4147		addrconf_dad_completed(ifp, bump_id, send_na);
4148
4149		goto out;
4150	}
4151
4152	ifp->dad_probes--;
4153	addrconf_mod_dad_work(ifp,
4154			      max(NEIGH_VAR(ifp->idev->nd_parms, RETRANS_TIME),
4155				  HZ/100));
4156	spin_unlock(&ifp->lock);
4157	write_unlock_bh(&idev->lock);
4158
4159	/* send a neighbour solicitation for our addr */
4160	addrconf_addr_solict_mult(&ifp->addr, &mcaddr);
4161	ndisc_send_ns(ifp->idev->dev, &ifp->addr, &mcaddr, &in6addr_any,
4162		      ifp->dad_nonce);
4163out:
4164	in6_ifa_put(ifp);
4165	rtnl_unlock();
4166}
4167
4168/* ifp->idev must be at least read locked */
4169static bool ipv6_lonely_lladdr(struct inet6_ifaddr *ifp)
4170{
4171	struct inet6_ifaddr *ifpiter;
4172	struct inet6_dev *idev = ifp->idev;
4173
4174	list_for_each_entry_reverse(ifpiter, &idev->addr_list, if_list) {
4175		if (ifpiter->scope > IFA_LINK)
4176			break;
4177		if (ifp != ifpiter && ifpiter->scope == IFA_LINK &&
4178		    (ifpiter->flags & (IFA_F_PERMANENT|IFA_F_TENTATIVE|
4179				       IFA_F_OPTIMISTIC|IFA_F_DADFAILED)) ==
4180		    IFA_F_PERMANENT)
4181			return false;
4182	}
4183	return true;
4184}
4185
4186static void addrconf_dad_completed(struct inet6_ifaddr *ifp, bool bump_id,
4187				   bool send_na)
4188{
4189	struct net_device *dev = ifp->idev->dev;
4190	struct in6_addr lladdr;
4191	bool send_rs, send_mld;
4192
4193	addrconf_del_dad_work(ifp);
4194
4195	/*
4196	 *	Configure the address for reception. Now it is valid.
4197	 */
4198
4199	ipv6_ifa_notify(RTM_NEWADDR, ifp);
4200
4201	/* If added prefix is link local and we are prepared to process
4202	   router advertisements, start sending router solicitations.
4203	 */
4204
4205	read_lock_bh(&ifp->idev->lock);
4206	send_mld = ifp->scope == IFA_LINK && ipv6_lonely_lladdr(ifp);
4207	send_rs = send_mld &&
4208		  ipv6_accept_ra(ifp->idev) &&
4209		  ifp->idev->cnf.rtr_solicits != 0 &&
4210		  (dev->flags & IFF_LOOPBACK) == 0 &&
4211		  (dev->type != ARPHRD_TUNNEL);
4212	read_unlock_bh(&ifp->idev->lock);
4213
4214	/* While dad is in progress mld report's source address is in6_addrany.
4215	 * Resend with proper ll now.
4216	 */
4217	if (send_mld)
4218		ipv6_mc_dad_complete(ifp->idev);
4219
4220	/* send unsolicited NA if enabled */
4221	if (send_na &&
4222	    (ifp->idev->cnf.ndisc_notify ||
4223	     dev_net(dev)->ipv6.devconf_all->ndisc_notify)) {
4224		ndisc_send_na(dev, &in6addr_linklocal_allnodes, &ifp->addr,
4225			      /*router=*/ !!ifp->idev->cnf.forwarding,
4226			      /*solicited=*/ false, /*override=*/ true,
4227			      /*inc_opt=*/ true);
4228	}
4229
4230	if (send_rs) {
4231		/*
4232		 *	If a host as already performed a random delay
4233		 *	[...] as part of DAD [...] there is no need
4234		 *	to delay again before sending the first RS
4235		 */
4236		if (ipv6_get_lladdr(dev, &lladdr, IFA_F_TENTATIVE))
4237			return;
4238		ndisc_send_rs(dev, &lladdr, &in6addr_linklocal_allrouters);
4239
4240		write_lock_bh(&ifp->idev->lock);
4241		spin_lock(&ifp->lock);
4242		ifp->idev->rs_interval = rfc3315_s14_backoff_init(
4243			ifp->idev->cnf.rtr_solicit_interval);
4244		ifp->idev->rs_probes = 1;
4245		ifp->idev->if_flags |= IF_RS_SENT;
4246		addrconf_mod_rs_timer(ifp->idev, ifp->idev->rs_interval);
4247		spin_unlock(&ifp->lock);
4248		write_unlock_bh(&ifp->idev->lock);
4249	}
4250
4251	if (bump_id)
4252		rt_genid_bump_ipv6(dev_net(dev));
4253
4254	/* Make sure that a new temporary address will be created
4255	 * before this temporary address becomes deprecated.
4256	 */
4257	if (ifp->flags & IFA_F_TEMPORARY)
4258		addrconf_verify_rtnl();
4259}
4260
4261static void addrconf_dad_run(struct inet6_dev *idev, bool restart)
4262{
4263	struct inet6_ifaddr *ifp;
4264
4265	read_lock_bh(&idev->lock);
4266	list_for_each_entry(ifp, &idev->addr_list, if_list) {
4267		spin_lock(&ifp->lock);
4268		if ((ifp->flags & IFA_F_TENTATIVE &&
4269		     ifp->state == INET6_IFADDR_STATE_DAD) || restart) {
4270			if (restart)
4271				ifp->state = INET6_IFADDR_STATE_PREDAD;
4272			addrconf_dad_kick(ifp);
4273		}
4274		spin_unlock(&ifp->lock);
4275	}
4276	read_unlock_bh(&idev->lock);
4277}
4278
4279#ifdef CONFIG_PROC_FS
4280struct if6_iter_state {
4281	struct seq_net_private p;
4282	int bucket;
4283	int offset;
4284};
4285
4286static struct inet6_ifaddr *if6_get_first(struct seq_file *seq, loff_t pos)
4287{
4288	struct if6_iter_state *state = seq->private;
4289	struct net *net = seq_file_net(seq);
4290	struct inet6_ifaddr *ifa = NULL;
4291	int p = 0;
4292
4293	/* initial bucket if pos is 0 */
4294	if (pos == 0) {
4295		state->bucket = 0;
4296		state->offset = 0;
4297	}
4298
4299	for (; state->bucket < IN6_ADDR_HSIZE; ++state->bucket) {
4300		hlist_for_each_entry_rcu(ifa, &inet6_addr_lst[state->bucket],
4301					 addr_lst) {
4302			if (!net_eq(dev_net(ifa->idev->dev), net))
4303				continue;
4304			/* sync with offset */
4305			if (p < state->offset) {
4306				p++;
4307				continue;
4308			}
4309			return ifa;
4310		}
4311
4312		/* prepare for next bucket */
4313		state->offset = 0;
4314		p = 0;
4315	}
4316	return NULL;
4317}
4318
4319static struct inet6_ifaddr *if6_get_next(struct seq_file *seq,
4320					 struct inet6_ifaddr *ifa)
4321{
4322	struct if6_iter_state *state = seq->private;
4323	struct net *net = seq_file_net(seq);
4324
4325	hlist_for_each_entry_continue_rcu(ifa, addr_lst) {
4326		if (!net_eq(dev_net(ifa->idev->dev), net))
4327			continue;
4328		state->offset++;
4329		return ifa;
4330	}
4331
4332	state->offset = 0;
4333	while (++state->bucket < IN6_ADDR_HSIZE) {
4334		hlist_for_each_entry_rcu(ifa,
4335				     &inet6_addr_lst[state->bucket], addr_lst) {
4336			if (!net_eq(dev_net(ifa->idev->dev), net))
4337				continue;
4338			return ifa;
4339		}
4340	}
4341
4342	return NULL;
4343}
4344
4345static void *if6_seq_start(struct seq_file *seq, loff_t *pos)
4346	__acquires(rcu)
4347{
4348	rcu_read_lock();
4349	return if6_get_first(seq, *pos);
4350}
4351
4352static void *if6_seq_next(struct seq_file *seq, void *v, loff_t *pos)
4353{
4354	struct inet6_ifaddr *ifa;
4355
4356	ifa = if6_get_next(seq, v);
4357	++*pos;
4358	return ifa;
4359}
4360
4361static void if6_seq_stop(struct seq_file *seq, void *v)
4362	__releases(rcu)
4363{
4364	rcu_read_unlock();
4365}
4366
4367static int if6_seq_show(struct seq_file *seq, void *v)
4368{
4369	struct inet6_ifaddr *ifp = (struct inet6_ifaddr *)v;
4370	seq_printf(seq, "%pi6 %02x %02x %02x %02x %8s\n",
4371		   &ifp->addr,
4372		   ifp->idev->dev->ifindex,
4373		   ifp->prefix_len,
4374		   ifp->scope,
4375		   (u8) ifp->flags,
4376		   ifp->idev->dev->name);
4377	return 0;
4378}
4379
4380static const struct seq_operations if6_seq_ops = {
4381	.start	= if6_seq_start,
4382	.next	= if6_seq_next,
4383	.show	= if6_seq_show,
4384	.stop	= if6_seq_stop,
4385};
4386
4387static int __net_init if6_proc_net_init(struct net *net)
4388{
4389	if (!proc_create_net("if_inet6", 0444, net->proc_net, &if6_seq_ops,
4390			sizeof(struct if6_iter_state)))
4391		return -ENOMEM;
4392	return 0;
4393}
4394
4395static void __net_exit if6_proc_net_exit(struct net *net)
4396{
4397	remove_proc_entry("if_inet6", net->proc_net);
4398}
4399
4400static struct pernet_operations if6_proc_net_ops = {
4401	.init = if6_proc_net_init,
4402	.exit = if6_proc_net_exit,
4403};
4404
4405int __init if6_proc_init(void)
4406{
4407	return register_pernet_subsys(&if6_proc_net_ops);
4408}
4409
4410void if6_proc_exit(void)
4411{
4412	unregister_pernet_subsys(&if6_proc_net_ops);
4413}
4414#endif	/* CONFIG_PROC_FS */
4415
4416#if IS_ENABLED(CONFIG_IPV6_MIP6)
4417/* Check if address is a home address configured on any interface. */
4418int ipv6_chk_home_addr(struct net *net, const struct in6_addr *addr)
4419{
4420	unsigned int hash = inet6_addr_hash(net, addr);
4421	struct inet6_ifaddr *ifp = NULL;
4422	int ret = 0;
4423
4424	rcu_read_lock();
4425	hlist_for_each_entry_rcu(ifp, &inet6_addr_lst[hash], addr_lst) {
4426		if (!net_eq(dev_net(ifp->idev->dev), net))
4427			continue;
4428		if (ipv6_addr_equal(&ifp->addr, addr) &&
4429		    (ifp->flags & IFA_F_HOMEADDRESS)) {
4430			ret = 1;
4431			break;
4432		}
4433	}
4434	rcu_read_unlock();
4435	return ret;
4436}
4437#endif
4438
4439/* RFC6554 has some algorithm to avoid loops in segment routing by
4440 * checking if the segments contains any of a local interface address.
4441 *
4442 * Quote:
4443 *
4444 * To detect loops in the SRH, a router MUST determine if the SRH
4445 * includes multiple addresses assigned to any interface on that router.
4446 * If such addresses appear more than once and are separated by at least
4447 * one address not assigned to that router.
4448 */
4449int ipv6_chk_rpl_srh_loop(struct net *net, const struct in6_addr *segs,
4450			  unsigned char nsegs)
4451{
4452	const struct in6_addr *addr;
4453	int i, ret = 0, found = 0;
4454	struct inet6_ifaddr *ifp;
4455	bool separated = false;
4456	unsigned int hash;
4457	bool hash_found;
4458
4459	rcu_read_lock();
4460	for (i = 0; i < nsegs; i++) {
4461		addr = &segs[i];
4462		hash = inet6_addr_hash(net, addr);
4463
4464		hash_found = false;
4465		hlist_for_each_entry_rcu(ifp, &inet6_addr_lst[hash], addr_lst) {
4466			if (!net_eq(dev_net(ifp->idev->dev), net))
4467				continue;
4468
4469			if (ipv6_addr_equal(&ifp->addr, addr)) {
4470				hash_found = true;
4471				break;
4472			}
4473		}
4474
4475		if (hash_found) {
4476			if (found > 1 && separated) {
4477				ret = 1;
4478				break;
4479			}
4480
4481			separated = false;
4482			found++;
4483		} else {
4484			separated = true;
4485		}
4486	}
4487	rcu_read_unlock();
4488
4489	return ret;
4490}
4491
4492/*
4493 *	Periodic address status verification
4494 */
4495
4496static void addrconf_verify_rtnl(void)
4497{
4498	unsigned long now, next, next_sec, next_sched;
4499	struct inet6_ifaddr *ifp;
4500	int i;
4501
4502	ASSERT_RTNL();
4503
4504	rcu_read_lock_bh();
4505	now = jiffies;
4506	next = round_jiffies_up(now + ADDR_CHECK_FREQUENCY);
4507
4508	cancel_delayed_work(&addr_chk_work);
4509
4510	for (i = 0; i < IN6_ADDR_HSIZE; i++) {
4511restart:
4512		hlist_for_each_entry_rcu_bh(ifp, &inet6_addr_lst[i], addr_lst) {
4513			unsigned long age;
4514
4515			/* When setting preferred_lft to a value not zero or
4516			 * infinity, while valid_lft is infinity
4517			 * IFA_F_PERMANENT has a non-infinity life time.
4518			 */
4519			if ((ifp->flags & IFA_F_PERMANENT) &&
4520			    (ifp->prefered_lft == INFINITY_LIFE_TIME))
4521				continue;
4522
4523			spin_lock(&ifp->lock);
4524			/* We try to batch several events at once. */
4525			age = (now - ifp->tstamp + ADDRCONF_TIMER_FUZZ_MINUS) / HZ;
4526
4527			if (ifp->valid_lft != INFINITY_LIFE_TIME &&
4528			    age >= ifp->valid_lft) {
4529				spin_unlock(&ifp->lock);
4530				in6_ifa_hold(ifp);
4531				ipv6_del_addr(ifp);
4532				goto restart;
4533			} else if (ifp->prefered_lft == INFINITY_LIFE_TIME) {
4534				spin_unlock(&ifp->lock);
4535				continue;
4536			} else if (age >= ifp->prefered_lft) {
4537				/* jiffies - ifp->tstamp > age >= ifp->prefered_lft */
4538				int deprecate = 0;
4539
4540				if (!(ifp->flags&IFA_F_DEPRECATED)) {
4541					deprecate = 1;
4542					ifp->flags |= IFA_F_DEPRECATED;
4543				}
4544
4545				if ((ifp->valid_lft != INFINITY_LIFE_TIME) &&
4546				    (time_before(ifp->tstamp + ifp->valid_lft * HZ, next)))
4547					next = ifp->tstamp + ifp->valid_lft * HZ;
4548
4549				spin_unlock(&ifp->lock);
4550
4551				if (deprecate) {
4552					in6_ifa_hold(ifp);
4553
4554					ipv6_ifa_notify(0, ifp);
4555					in6_ifa_put(ifp);
4556					goto restart;
4557				}
4558			} else if ((ifp->flags&IFA_F_TEMPORARY) &&
4559				   !(ifp->flags&IFA_F_TENTATIVE)) {
4560				unsigned long regen_advance = ifp->idev->cnf.regen_max_retry *
4561					ifp->idev->cnf.dad_transmits *
4562					max(NEIGH_VAR(ifp->idev->nd_parms, RETRANS_TIME), HZ/100) / HZ;
4563
4564				if (age >= ifp->prefered_lft - regen_advance) {
4565					struct inet6_ifaddr *ifpub = ifp->ifpub;
4566					if (time_before(ifp->tstamp + ifp->prefered_lft * HZ, next))
4567						next = ifp->tstamp + ifp->prefered_lft * HZ;
4568					if (!ifp->regen_count && ifpub) {
4569						ifp->regen_count++;
4570						in6_ifa_hold(ifp);
4571						in6_ifa_hold(ifpub);
4572						spin_unlock(&ifp->lock);
4573
4574						spin_lock(&ifpub->lock);
4575						ifpub->regen_count = 0;
4576						spin_unlock(&ifpub->lock);
4577						rcu_read_unlock_bh();
4578						ipv6_create_tempaddr(ifpub, true);
4579						in6_ifa_put(ifpub);
4580						in6_ifa_put(ifp);
4581						rcu_read_lock_bh();
4582						goto restart;
4583					}
4584				} else if (time_before(ifp->tstamp + ifp->prefered_lft * HZ - regen_advance * HZ, next))
4585					next = ifp->tstamp + ifp->prefered_lft * HZ - regen_advance * HZ;
4586				spin_unlock(&ifp->lock);
4587			} else {
4588				/* ifp->prefered_lft <= ifp->valid_lft */
4589				if (time_before(ifp->tstamp + ifp->prefered_lft * HZ, next))
4590					next = ifp->tstamp + ifp->prefered_lft * HZ;
4591				spin_unlock(&ifp->lock);
4592			}
4593		}
4594	}
4595
4596	next_sec = round_jiffies_up(next);
4597	next_sched = next;
4598
4599	/* If rounded timeout is accurate enough, accept it. */
4600	if (time_before(next_sec, next + ADDRCONF_TIMER_FUZZ))
4601		next_sched = next_sec;
4602
4603	/* And minimum interval is ADDRCONF_TIMER_FUZZ_MAX. */
4604	if (time_before(next_sched, jiffies + ADDRCONF_TIMER_FUZZ_MAX))
4605		next_sched = jiffies + ADDRCONF_TIMER_FUZZ_MAX;
4606
4607	pr_debug("now = %lu, schedule = %lu, rounded schedule = %lu => %lu\n",
4608		 now, next, next_sec, next_sched);
4609	mod_delayed_work(addrconf_wq, &addr_chk_work, next_sched - now);
4610	rcu_read_unlock_bh();
4611}
4612
4613static void addrconf_verify_work(struct work_struct *w)
4614{
4615	rtnl_lock();
4616	addrconf_verify_rtnl();
4617	rtnl_unlock();
4618}
4619
4620static void addrconf_verify(void)
4621{
4622	mod_delayed_work(addrconf_wq, &addr_chk_work, 0);
4623}
4624
4625static struct in6_addr *extract_addr(struct nlattr *addr, struct nlattr *local,
4626				     struct in6_addr **peer_pfx)
4627{
4628	struct in6_addr *pfx = NULL;
4629
4630	*peer_pfx = NULL;
4631
4632	if (addr)
4633		pfx = nla_data(addr);
4634
4635	if (local) {
4636		if (pfx && nla_memcmp(local, pfx, sizeof(*pfx)))
4637			*peer_pfx = pfx;
4638		pfx = nla_data(local);
4639	}
4640
4641	return pfx;
4642}
4643
4644static const struct nla_policy ifa_ipv6_policy[IFA_MAX+1] = {
4645	[IFA_ADDRESS]		= { .len = sizeof(struct in6_addr) },
4646	[IFA_LOCAL]		= { .len = sizeof(struct in6_addr) },
4647	[IFA_CACHEINFO]		= { .len = sizeof(struct ifa_cacheinfo) },
4648	[IFA_FLAGS]		= { .len = sizeof(u32) },
4649	[IFA_RT_PRIORITY]	= { .len = sizeof(u32) },
4650	[IFA_TARGET_NETNSID]	= { .type = NLA_S32 },
4651};
4652
4653static int
4654inet6_rtm_deladdr(struct sk_buff *skb, struct nlmsghdr *nlh,
4655		  struct netlink_ext_ack *extack)
4656{
4657	struct net *net = sock_net(skb->sk);
4658	struct ifaddrmsg *ifm;
4659	struct nlattr *tb[IFA_MAX+1];
4660	struct in6_addr *pfx, *peer_pfx;
4661	u32 ifa_flags;
4662	int err;
4663
4664	err = nlmsg_parse_deprecated(nlh, sizeof(*ifm), tb, IFA_MAX,
4665				     ifa_ipv6_policy, extack);
4666	if (err < 0)
4667		return err;
4668
4669	ifm = nlmsg_data(nlh);
4670	pfx = extract_addr(tb[IFA_ADDRESS], tb[IFA_LOCAL], &peer_pfx);
4671	if (!pfx)
4672		return -EINVAL;
4673
4674	ifa_flags = tb[IFA_FLAGS] ? nla_get_u32(tb[IFA_FLAGS]) : ifm->ifa_flags;
4675
4676	/* We ignore other flags so far. */
4677	ifa_flags &= IFA_F_MANAGETEMPADDR;
4678
4679	return inet6_addr_del(net, ifm->ifa_index, ifa_flags, pfx,
4680			      ifm->ifa_prefixlen);
4681}
4682
4683static int modify_prefix_route(struct inet6_ifaddr *ifp,
4684			       unsigned long expires, u32 flags,
4685			       bool modify_peer)
4686{
4687	struct fib6_info *f6i;
4688	u32 prio;
4689
4690	f6i = addrconf_get_prefix_route(modify_peer ? &ifp->peer_addr : &ifp->addr,
4691					ifp->prefix_len,
4692					ifp->idev->dev, 0, RTF_DEFAULT, true);
4693	if (!f6i)
4694		return -ENOENT;
4695
4696	prio = ifp->rt_priority ? : IP6_RT_PRIO_ADDRCONF;
4697	if (f6i->fib6_metric != prio) {
4698		/* delete old one */
4699		ip6_del_rt(dev_net(ifp->idev->dev), f6i, false);
4700
4701		/* add new one */
4702		addrconf_prefix_route(modify_peer ? &ifp->peer_addr : &ifp->addr,
4703				      ifp->prefix_len,
4704				      ifp->rt_priority, ifp->idev->dev,
4705				      expires, flags, GFP_KERNEL);
4706	} else {
4707		if (!expires)
4708			fib6_clean_expires(f6i);
4709		else
4710			fib6_set_expires(f6i, expires);
4711
4712		fib6_info_release(f6i);
4713	}
4714
4715	return 0;
4716}
4717
4718static int inet6_addr_modify(struct inet6_ifaddr *ifp, struct ifa6_config *cfg)
4719{
4720	u32 flags;
4721	clock_t expires;
4722	unsigned long timeout;
4723	bool was_managetempaddr;
4724	bool had_prefixroute;
4725	bool new_peer = false;
4726
4727	ASSERT_RTNL();
4728
4729	if (!cfg->valid_lft || cfg->preferred_lft > cfg->valid_lft)
4730		return -EINVAL;
4731
4732	if (cfg->ifa_flags & IFA_F_MANAGETEMPADDR &&
4733	    (ifp->flags & IFA_F_TEMPORARY || ifp->prefix_len != 64))
4734		return -EINVAL;
4735
4736	if (!(ifp->flags & IFA_F_TENTATIVE) || ifp->flags & IFA_F_DADFAILED)
4737		cfg->ifa_flags &= ~IFA_F_OPTIMISTIC;
4738
4739	timeout = addrconf_timeout_fixup(cfg->valid_lft, HZ);
4740	if (addrconf_finite_timeout(timeout)) {
4741		expires = jiffies_to_clock_t(timeout * HZ);
4742		cfg->valid_lft = timeout;
4743		flags = RTF_EXPIRES;
4744	} else {
4745		expires = 0;
4746		flags = 0;
4747		cfg->ifa_flags |= IFA_F_PERMANENT;
4748	}
4749
4750	timeout = addrconf_timeout_fixup(cfg->preferred_lft, HZ);
4751	if (addrconf_finite_timeout(timeout)) {
4752		if (timeout == 0)
4753			cfg->ifa_flags |= IFA_F_DEPRECATED;
4754		cfg->preferred_lft = timeout;
4755	}
4756
4757	if (cfg->peer_pfx &&
4758	    memcmp(&ifp->peer_addr, cfg->peer_pfx, sizeof(struct in6_addr))) {
4759		if (!ipv6_addr_any(&ifp->peer_addr))
4760			cleanup_prefix_route(ifp, expires, true, true);
4761		new_peer = true;
4762	}
4763
4764	spin_lock_bh(&ifp->lock);
4765	was_managetempaddr = ifp->flags & IFA_F_MANAGETEMPADDR;
4766	had_prefixroute = ifp->flags & IFA_F_PERMANENT &&
4767			  !(ifp->flags & IFA_F_NOPREFIXROUTE);
4768	ifp->flags &= ~(IFA_F_DEPRECATED | IFA_F_PERMANENT | IFA_F_NODAD |
4769			IFA_F_HOMEADDRESS | IFA_F_MANAGETEMPADDR |
4770			IFA_F_NOPREFIXROUTE);
4771	ifp->flags |= cfg->ifa_flags;
4772	ifp->tstamp = jiffies;
4773	ifp->valid_lft = cfg->valid_lft;
4774	ifp->prefered_lft = cfg->preferred_lft;
4775
4776	if (cfg->rt_priority && cfg->rt_priority != ifp->rt_priority)
4777		ifp->rt_priority = cfg->rt_priority;
4778
4779	if (new_peer)
4780		ifp->peer_addr = *cfg->peer_pfx;
4781
4782	spin_unlock_bh(&ifp->lock);
4783	if (!(ifp->flags&IFA_F_TENTATIVE))
4784		ipv6_ifa_notify(0, ifp);
4785
4786	if (!(cfg->ifa_flags & IFA_F_NOPREFIXROUTE)) {
4787		int rc = -ENOENT;
4788
4789		if (had_prefixroute)
4790			rc = modify_prefix_route(ifp, expires, flags, false);
4791
4792		/* prefix route could have been deleted; if so restore it */
4793		if (rc == -ENOENT) {
4794			addrconf_prefix_route(&ifp->addr, ifp->prefix_len,
4795					      ifp->rt_priority, ifp->idev->dev,
4796					      expires, flags, GFP_KERNEL);
4797		}
4798
4799		if (had_prefixroute && !ipv6_addr_any(&ifp->peer_addr))
4800			rc = modify_prefix_route(ifp, expires, flags, true);
4801
4802		if (rc == -ENOENT && !ipv6_addr_any(&ifp->peer_addr)) {
4803			addrconf_prefix_route(&ifp->peer_addr, ifp->prefix_len,
4804					      ifp->rt_priority, ifp->idev->dev,
4805					      expires, flags, GFP_KERNEL);
4806		}
4807	} else if (had_prefixroute) {
4808		enum cleanup_prefix_rt_t action;
4809		unsigned long rt_expires;
4810
4811		write_lock_bh(&ifp->idev->lock);
4812		action = check_cleanup_prefix_route(ifp, &rt_expires);
4813		write_unlock_bh(&ifp->idev->lock);
4814
4815		if (action != CLEANUP_PREFIX_RT_NOP) {
4816			cleanup_prefix_route(ifp, rt_expires,
4817				action == CLEANUP_PREFIX_RT_DEL, false);
4818		}
4819	}
4820
4821	if (was_managetempaddr || ifp->flags & IFA_F_MANAGETEMPADDR) {
4822		if (was_managetempaddr &&
4823		    !(ifp->flags & IFA_F_MANAGETEMPADDR)) {
4824			cfg->valid_lft = 0;
4825			cfg->preferred_lft = 0;
4826		}
4827		manage_tempaddrs(ifp->idev, ifp, cfg->valid_lft,
4828				 cfg->preferred_lft, !was_managetempaddr,
4829				 jiffies);
4830	}
4831
4832	addrconf_verify_rtnl();
4833
4834	return 0;
4835}
4836
4837static int
4838inet6_rtm_newaddr(struct sk_buff *skb, struct nlmsghdr *nlh,
4839		  struct netlink_ext_ack *extack)
4840{
4841	struct net *net = sock_net(skb->sk);
4842	struct ifaddrmsg *ifm;
4843	struct nlattr *tb[IFA_MAX+1];
4844	struct in6_addr *peer_pfx;
4845	struct inet6_ifaddr *ifa;
4846	struct net_device *dev;
4847	struct inet6_dev *idev;
4848	struct ifa6_config cfg;
4849	int err;
4850
4851	err = nlmsg_parse_deprecated(nlh, sizeof(*ifm), tb, IFA_MAX,
4852				     ifa_ipv6_policy, extack);
4853	if (err < 0)
4854		return err;
4855
4856	memset(&cfg, 0, sizeof(cfg));
4857
4858	ifm = nlmsg_data(nlh);
4859	cfg.pfx = extract_addr(tb[IFA_ADDRESS], tb[IFA_LOCAL], &peer_pfx);
4860	if (!cfg.pfx)
4861		return -EINVAL;
4862
4863	cfg.peer_pfx = peer_pfx;
4864	cfg.plen = ifm->ifa_prefixlen;
4865	if (tb[IFA_RT_PRIORITY])
4866		cfg.rt_priority = nla_get_u32(tb[IFA_RT_PRIORITY]);
4867
4868	cfg.valid_lft = INFINITY_LIFE_TIME;
4869	cfg.preferred_lft = INFINITY_LIFE_TIME;
4870
4871	if (tb[IFA_CACHEINFO]) {
4872		struct ifa_cacheinfo *ci;
4873
4874		ci = nla_data(tb[IFA_CACHEINFO]);
4875		cfg.valid_lft = ci->ifa_valid;
4876		cfg.preferred_lft = ci->ifa_prefered;
4877	}
4878
4879	dev =  __dev_get_by_index(net, ifm->ifa_index);
4880	if (!dev)
4881		return -ENODEV;
4882
4883	if (tb[IFA_FLAGS])
4884		cfg.ifa_flags = nla_get_u32(tb[IFA_FLAGS]);
4885	else
4886		cfg.ifa_flags = ifm->ifa_flags;
4887
4888	/* We ignore other flags so far. */
4889	cfg.ifa_flags &= IFA_F_NODAD | IFA_F_HOMEADDRESS |
4890			 IFA_F_MANAGETEMPADDR | IFA_F_NOPREFIXROUTE |
4891			 IFA_F_MCAUTOJOIN | IFA_F_OPTIMISTIC;
4892
4893	idev = ipv6_find_idev(dev);
4894	if (IS_ERR(idev))
4895		return PTR_ERR(idev);
4896
4897	if (!ipv6_allow_optimistic_dad(net, idev))
4898		cfg.ifa_flags &= ~IFA_F_OPTIMISTIC;
4899
4900	if (cfg.ifa_flags & IFA_F_NODAD &&
4901	    cfg.ifa_flags & IFA_F_OPTIMISTIC) {
4902		NL_SET_ERR_MSG(extack, "IFA_F_NODAD and IFA_F_OPTIMISTIC are mutually exclusive");
4903		return -EINVAL;
4904	}
4905
4906	ifa = ipv6_get_ifaddr(net, cfg.pfx, dev, 1);
4907	if (!ifa) {
4908		/*
4909		 * It would be best to check for !NLM_F_CREATE here but
4910		 * userspace already relies on not having to provide this.
4911		 */
4912		return inet6_addr_add(net, ifm->ifa_index, &cfg, extack);
4913	}
4914
4915	if (nlh->nlmsg_flags & NLM_F_EXCL ||
4916	    !(nlh->nlmsg_flags & NLM_F_REPLACE))
4917		err = -EEXIST;
4918	else
4919		err = inet6_addr_modify(ifa, &cfg);
4920
4921	in6_ifa_put(ifa);
4922
4923	return err;
4924}
4925
4926static void put_ifaddrmsg(struct nlmsghdr *nlh, u8 prefixlen, u32 flags,
4927			  u8 scope, int ifindex)
4928{
4929	struct ifaddrmsg *ifm;
4930
4931	ifm = nlmsg_data(nlh);
4932	ifm->ifa_family = AF_INET6;
4933	ifm->ifa_prefixlen = prefixlen;
4934	ifm->ifa_flags = flags;
4935	ifm->ifa_scope = scope;
4936	ifm->ifa_index = ifindex;
4937}
4938
4939static int put_cacheinfo(struct sk_buff *skb, unsigned long cstamp,
4940			 unsigned long tstamp, u32 preferred, u32 valid)
4941{
4942	struct ifa_cacheinfo ci;
4943
4944	ci.cstamp = cstamp_delta(cstamp);
4945	ci.tstamp = cstamp_delta(tstamp);
4946	ci.ifa_prefered = preferred;
4947	ci.ifa_valid = valid;
4948
4949	return nla_put(skb, IFA_CACHEINFO, sizeof(ci), &ci);
4950}
4951
4952static inline int rt_scope(int ifa_scope)
4953{
4954	if (ifa_scope & IFA_HOST)
4955		return RT_SCOPE_HOST;
4956	else if (ifa_scope & IFA_LINK)
4957		return RT_SCOPE_LINK;
4958	else if (ifa_scope & IFA_SITE)
4959		return RT_SCOPE_SITE;
4960	else
4961		return RT_SCOPE_UNIVERSE;
4962}
4963
4964static inline int inet6_ifaddr_msgsize(void)
4965{
4966	return NLMSG_ALIGN(sizeof(struct ifaddrmsg))
4967	       + nla_total_size(16) /* IFA_LOCAL */
4968	       + nla_total_size(16) /* IFA_ADDRESS */
4969	       + nla_total_size(sizeof(struct ifa_cacheinfo))
4970	       + nla_total_size(4)  /* IFA_FLAGS */
4971	       + nla_total_size(4)  /* IFA_RT_PRIORITY */;
4972}
4973
4974enum addr_type_t {
4975	UNICAST_ADDR,
4976	MULTICAST_ADDR,
4977	ANYCAST_ADDR,
4978};
4979
4980struct inet6_fill_args {
4981	u32 portid;
4982	u32 seq;
4983	int event;
4984	unsigned int flags;
4985	int netnsid;
4986	int ifindex;
4987	enum addr_type_t type;
4988};
4989
4990static int inet6_fill_ifaddr(struct sk_buff *skb, struct inet6_ifaddr *ifa,
4991			     struct inet6_fill_args *args)
4992{
4993	struct nlmsghdr  *nlh;
4994	u32 preferred, valid;
4995
4996	nlh = nlmsg_put(skb, args->portid, args->seq, args->event,
4997			sizeof(struct ifaddrmsg), args->flags);
4998	if (!nlh)
4999		return -EMSGSIZE;
5000
5001	put_ifaddrmsg(nlh, ifa->prefix_len, ifa->flags, rt_scope(ifa->scope),
5002		      ifa->idev->dev->ifindex);
5003
5004	if (args->netnsid >= 0 &&
5005	    nla_put_s32(skb, IFA_TARGET_NETNSID, args->netnsid))
5006		goto error;
5007
5008	spin_lock_bh(&ifa->lock);
5009	if (!((ifa->flags&IFA_F_PERMANENT) &&
5010	      (ifa->prefered_lft == INFINITY_LIFE_TIME))) {
5011		preferred = ifa->prefered_lft;
5012		valid = ifa->valid_lft;
5013		if (preferred != INFINITY_LIFE_TIME) {
5014			long tval = (jiffies - ifa->tstamp)/HZ;
5015			if (preferred > tval)
5016				preferred -= tval;
5017			else
5018				preferred = 0;
5019			if (valid != INFINITY_LIFE_TIME) {
5020				if (valid > tval)
5021					valid -= tval;
5022				else
5023					valid = 0;
5024			}
5025		}
5026	} else {
5027		preferred = INFINITY_LIFE_TIME;
5028		valid = INFINITY_LIFE_TIME;
5029	}
5030	spin_unlock_bh(&ifa->lock);
5031
5032	if (!ipv6_addr_any(&ifa->peer_addr)) {
5033		if (nla_put_in6_addr(skb, IFA_LOCAL, &ifa->addr) < 0 ||
5034		    nla_put_in6_addr(skb, IFA_ADDRESS, &ifa->peer_addr) < 0)
5035			goto error;
5036	} else
5037		if (nla_put_in6_addr(skb, IFA_ADDRESS, &ifa->addr) < 0)
5038			goto error;
5039
5040	if (ifa->rt_priority &&
5041	    nla_put_u32(skb, IFA_RT_PRIORITY, ifa->rt_priority))
5042		goto error;
5043
5044	if (put_cacheinfo(skb, ifa->cstamp, ifa->tstamp, preferred, valid) < 0)
5045		goto error;
5046
5047	if (nla_put_u32(skb, IFA_FLAGS, ifa->flags) < 0)
5048		goto error;
5049
5050	nlmsg_end(skb, nlh);
5051	return 0;
5052
5053error:
5054	nlmsg_cancel(skb, nlh);
5055	return -EMSGSIZE;
5056}
5057
5058static int inet6_fill_ifmcaddr(struct sk_buff *skb, struct ifmcaddr6 *ifmca,
5059			       struct inet6_fill_args *args)
5060{
5061	struct nlmsghdr  *nlh;
5062	u8 scope = RT_SCOPE_UNIVERSE;
5063	int ifindex = ifmca->idev->dev->ifindex;
5064
5065	if (ipv6_addr_scope(&ifmca->mca_addr) & IFA_SITE)
5066		scope = RT_SCOPE_SITE;
5067
5068	nlh = nlmsg_put(skb, args->portid, args->seq, args->event,
5069			sizeof(struct ifaddrmsg), args->flags);
5070	if (!nlh)
5071		return -EMSGSIZE;
5072
5073	if (args->netnsid >= 0 &&
5074	    nla_put_s32(skb, IFA_TARGET_NETNSID, args->netnsid)) {
5075		nlmsg_cancel(skb, nlh);
5076		return -EMSGSIZE;
5077	}
5078
5079	put_ifaddrmsg(nlh, 128, IFA_F_PERMANENT, scope, ifindex);
5080	if (nla_put_in6_addr(skb, IFA_MULTICAST, &ifmca->mca_addr) < 0 ||
5081	    put_cacheinfo(skb, ifmca->mca_cstamp, ifmca->mca_tstamp,
5082			  INFINITY_LIFE_TIME, INFINITY_LIFE_TIME) < 0) {
5083		nlmsg_cancel(skb, nlh);
5084		return -EMSGSIZE;
5085	}
5086
5087	nlmsg_end(skb, nlh);
5088	return 0;
5089}
5090
5091static int inet6_fill_ifacaddr(struct sk_buff *skb, struct ifacaddr6 *ifaca,
5092			       struct inet6_fill_args *args)
5093{
5094	struct net_device *dev = fib6_info_nh_dev(ifaca->aca_rt);
5095	int ifindex = dev ? dev->ifindex : 1;
5096	struct nlmsghdr  *nlh;
5097	u8 scope = RT_SCOPE_UNIVERSE;
5098
5099	if (ipv6_addr_scope(&ifaca->aca_addr) & IFA_SITE)
5100		scope = RT_SCOPE_SITE;
5101
5102	nlh = nlmsg_put(skb, args->portid, args->seq, args->event,
5103			sizeof(struct ifaddrmsg), args->flags);
5104	if (!nlh)
5105		return -EMSGSIZE;
5106
5107	if (args->netnsid >= 0 &&
5108	    nla_put_s32(skb, IFA_TARGET_NETNSID, args->netnsid)) {
5109		nlmsg_cancel(skb, nlh);
5110		return -EMSGSIZE;
5111	}
5112
5113	put_ifaddrmsg(nlh, 128, IFA_F_PERMANENT, scope, ifindex);
5114	if (nla_put_in6_addr(skb, IFA_ANYCAST, &ifaca->aca_addr) < 0 ||
5115	    put_cacheinfo(skb, ifaca->aca_cstamp, ifaca->aca_tstamp,
5116			  INFINITY_LIFE_TIME, INFINITY_LIFE_TIME) < 0) {
5117		nlmsg_cancel(skb, nlh);
5118		return -EMSGSIZE;
5119	}
5120
5121	nlmsg_end(skb, nlh);
5122	return 0;
5123}
5124
5125/* called with rcu_read_lock() */
5126static int in6_dump_addrs(struct inet6_dev *idev, struct sk_buff *skb,
5127			  struct netlink_callback *cb, int s_ip_idx,
5128			  struct inet6_fill_args *fillargs)
5129{
5130	struct ifmcaddr6 *ifmca;
5131	struct ifacaddr6 *ifaca;
5132	int ip_idx = 0;
5133	int err = 1;
5134
5135	read_lock_bh(&idev->lock);
5136	switch (fillargs->type) {
5137	case UNICAST_ADDR: {
5138		struct inet6_ifaddr *ifa;
5139		fillargs->event = RTM_NEWADDR;
5140
5141		/* unicast address incl. temp addr */
5142		list_for_each_entry(ifa, &idev->addr_list, if_list) {
5143			if (ip_idx < s_ip_idx)
5144				goto next;
5145			err = inet6_fill_ifaddr(skb, ifa, fillargs);
5146			if (err < 0)
5147				break;
5148			nl_dump_check_consistent(cb, nlmsg_hdr(skb));
5149next:
5150			ip_idx++;
5151		}
5152		break;
5153	}
5154	case MULTICAST_ADDR:
5155		fillargs->event = RTM_GETMULTICAST;
5156
5157		/* multicast address */
5158		for (ifmca = idev->mc_list; ifmca;
5159		     ifmca = ifmca->next, ip_idx++) {
5160			if (ip_idx < s_ip_idx)
5161				continue;
5162			err = inet6_fill_ifmcaddr(skb, ifmca, fillargs);
5163			if (err < 0)
5164				break;
5165		}
5166		break;
5167	case ANYCAST_ADDR:
5168		fillargs->event = RTM_GETANYCAST;
5169		/* anycast address */
5170		for (ifaca = idev->ac_list; ifaca;
5171		     ifaca = ifaca->aca_next, ip_idx++) {
5172			if (ip_idx < s_ip_idx)
5173				continue;
5174			err = inet6_fill_ifacaddr(skb, ifaca, fillargs);
5175			if (err < 0)
5176				break;
5177		}
5178		break;
5179	default:
5180		break;
5181	}
5182	read_unlock_bh(&idev->lock);
5183	cb->args[2] = ip_idx;
5184	return err;
5185}
5186
5187static int inet6_valid_dump_ifaddr_req(const struct nlmsghdr *nlh,
5188				       struct inet6_fill_args *fillargs,
5189				       struct net **tgt_net, struct sock *sk,
5190				       struct netlink_callback *cb)
5191{
5192	struct netlink_ext_ack *extack = cb->extack;
5193	struct nlattr *tb[IFA_MAX+1];
5194	struct ifaddrmsg *ifm;
5195	int err, i;
5196
5197	if (nlh->nlmsg_len < nlmsg_msg_size(sizeof(*ifm))) {
5198		NL_SET_ERR_MSG_MOD(extack, "Invalid header for address dump request");
5199		return -EINVAL;
5200	}
5201
5202	ifm = nlmsg_data(nlh);
5203	if (ifm->ifa_prefixlen || ifm->ifa_flags || ifm->ifa_scope) {
5204		NL_SET_ERR_MSG_MOD(extack, "Invalid values in header for address dump request");
5205		return -EINVAL;
5206	}
5207
5208	fillargs->ifindex = ifm->ifa_index;
5209	if (fillargs->ifindex) {
5210		cb->answer_flags |= NLM_F_DUMP_FILTERED;
5211		fillargs->flags |= NLM_F_DUMP_FILTERED;
5212	}
5213
5214	err = nlmsg_parse_deprecated_strict(nlh, sizeof(*ifm), tb, IFA_MAX,
5215					    ifa_ipv6_policy, extack);
5216	if (err < 0)
5217		return err;
5218
5219	for (i = 0; i <= IFA_MAX; ++i) {
5220		if (!tb[i])
5221			continue;
5222
5223		if (i == IFA_TARGET_NETNSID) {
5224			struct net *net;
5225
5226			fillargs->netnsid = nla_get_s32(tb[i]);
5227			net = rtnl_get_net_ns_capable(sk, fillargs->netnsid);
5228			if (IS_ERR(net)) {
5229				fillargs->netnsid = -1;
5230				NL_SET_ERR_MSG_MOD(extack, "Invalid target network namespace id");
5231				return PTR_ERR(net);
5232			}
5233			*tgt_net = net;
5234		} else {
5235			NL_SET_ERR_MSG_MOD(extack, "Unsupported attribute in dump request");
5236			return -EINVAL;
5237		}
5238	}
5239
5240	return 0;
5241}
5242
5243static int inet6_dump_addr(struct sk_buff *skb, struct netlink_callback *cb,
5244			   enum addr_type_t type)
5245{
5246	const struct nlmsghdr *nlh = cb->nlh;
5247	struct inet6_fill_args fillargs = {
5248		.portid = NETLINK_CB(cb->skb).portid,
5249		.seq = cb->nlh->nlmsg_seq,
5250		.flags = NLM_F_MULTI,
5251		.netnsid = -1,
5252		.type = type,
5253	};
5254	struct net *net = sock_net(skb->sk);
5255	struct net *tgt_net = net;
5256	int idx, s_idx, s_ip_idx;
5257	int h, s_h;
5258	struct net_device *dev;
5259	struct inet6_dev *idev;
5260	struct hlist_head *head;
5261	int err = 0;
5262
5263	s_h = cb->args[0];
5264	s_idx = idx = cb->args[1];
5265	s_ip_idx = cb->args[2];
5266
5267	if (cb->strict_check) {
5268		err = inet6_valid_dump_ifaddr_req(nlh, &fillargs, &tgt_net,
5269						  skb->sk, cb);
5270		if (err < 0)
5271			goto put_tgt_net;
5272
5273		err = 0;
5274		if (fillargs.ifindex) {
5275			dev = __dev_get_by_index(tgt_net, fillargs.ifindex);
5276			if (!dev) {
5277				err = -ENODEV;
5278				goto put_tgt_net;
5279			}
5280			idev = __in6_dev_get(dev);
5281			if (idev) {
5282				err = in6_dump_addrs(idev, skb, cb, s_ip_idx,
5283						     &fillargs);
5284				if (err > 0)
5285					err = 0;
5286			}
5287			goto put_tgt_net;
5288		}
5289	}
5290
5291	rcu_read_lock();
5292	cb->seq = atomic_read(&tgt_net->ipv6.dev_addr_genid) ^ tgt_net->dev_base_seq;
5293	for (h = s_h; h < NETDEV_HASHENTRIES; h++, s_idx = 0) {
5294		idx = 0;
5295		head = &tgt_net->dev_index_head[h];
5296		hlist_for_each_entry_rcu(dev, head, index_hlist) {
5297			if (idx < s_idx)
5298				goto cont;
5299			if (h > s_h || idx > s_idx)
5300				s_ip_idx = 0;
5301			idev = __in6_dev_get(dev);
5302			if (!idev)
5303				goto cont;
5304
5305			if (in6_dump_addrs(idev, skb, cb, s_ip_idx,
5306					   &fillargs) < 0)
5307				goto done;
5308cont:
5309			idx++;
5310		}
5311	}
5312done:
5313	rcu_read_unlock();
5314	cb->args[0] = h;
5315	cb->args[1] = idx;
5316put_tgt_net:
5317	if (fillargs.netnsid >= 0)
5318		put_net(tgt_net);
5319
5320	return skb->len ? : err;
5321}
5322
5323static int inet6_dump_ifaddr(struct sk_buff *skb, struct netlink_callback *cb)
5324{
5325	enum addr_type_t type = UNICAST_ADDR;
5326
5327	return inet6_dump_addr(skb, cb, type);
5328}
5329
5330static int inet6_dump_ifmcaddr(struct sk_buff *skb, struct netlink_callback *cb)
5331{
5332	enum addr_type_t type = MULTICAST_ADDR;
5333
5334	return inet6_dump_addr(skb, cb, type);
5335}
5336
5337
5338static int inet6_dump_ifacaddr(struct sk_buff *skb, struct netlink_callback *cb)
5339{
5340	enum addr_type_t type = ANYCAST_ADDR;
5341
5342	return inet6_dump_addr(skb, cb, type);
5343}
5344
5345static int inet6_rtm_valid_getaddr_req(struct sk_buff *skb,
5346				       const struct nlmsghdr *nlh,
5347				       struct nlattr **tb,
5348				       struct netlink_ext_ack *extack)
5349{
5350	struct ifaddrmsg *ifm;
5351	int i, err;
5352
5353	if (nlh->nlmsg_len < nlmsg_msg_size(sizeof(*ifm))) {
5354		NL_SET_ERR_MSG_MOD(extack, "Invalid header for get address request");
5355		return -EINVAL;
5356	}
5357
5358	if (!netlink_strict_get_check(skb))
5359		return nlmsg_parse_deprecated(nlh, sizeof(*ifm), tb, IFA_MAX,
5360					      ifa_ipv6_policy, extack);
5361
5362	ifm = nlmsg_data(nlh);
5363	if (ifm->ifa_prefixlen || ifm->ifa_flags || ifm->ifa_scope) {
5364		NL_SET_ERR_MSG_MOD(extack, "Invalid values in header for get address request");
5365		return -EINVAL;
5366	}
5367
5368	err = nlmsg_parse_deprecated_strict(nlh, sizeof(*ifm), tb, IFA_MAX,
5369					    ifa_ipv6_policy, extack);
5370	if (err)
5371		return err;
5372
5373	for (i = 0; i <= IFA_MAX; i++) {
5374		if (!tb[i])
5375			continue;
5376
5377		switch (i) {
5378		case IFA_TARGET_NETNSID:
5379		case IFA_ADDRESS:
5380		case IFA_LOCAL:
5381			break;
5382		default:
5383			NL_SET_ERR_MSG_MOD(extack, "Unsupported attribute in get address request");
5384			return -EINVAL;
5385		}
5386	}
5387
5388	return 0;
5389}
5390
5391static int inet6_rtm_getaddr(struct sk_buff *in_skb, struct nlmsghdr *nlh,
5392			     struct netlink_ext_ack *extack)
5393{
5394	struct net *net = sock_net(in_skb->sk);
5395	struct inet6_fill_args fillargs = {
5396		.portid = NETLINK_CB(in_skb).portid,
5397		.seq = nlh->nlmsg_seq,
5398		.event = RTM_NEWADDR,
5399		.flags = 0,
5400		.netnsid = -1,
5401	};
5402	struct net *tgt_net = net;
5403	struct ifaddrmsg *ifm;
5404	struct nlattr *tb[IFA_MAX+1];
5405	struct in6_addr *addr = NULL, *peer;
5406	struct net_device *dev = NULL;
5407	struct inet6_ifaddr *ifa;
5408	struct sk_buff *skb;
5409	int err;
5410
5411	err = inet6_rtm_valid_getaddr_req(in_skb, nlh, tb, extack);
5412	if (err < 0)
5413		return err;
5414
5415	if (tb[IFA_TARGET_NETNSID]) {
5416		fillargs.netnsid = nla_get_s32(tb[IFA_TARGET_NETNSID]);
5417
5418		tgt_net = rtnl_get_net_ns_capable(NETLINK_CB(in_skb).sk,
5419						  fillargs.netnsid);
5420		if (IS_ERR(tgt_net))
5421			return PTR_ERR(tgt_net);
5422	}
5423
5424	addr = extract_addr(tb[IFA_ADDRESS], tb[IFA_LOCAL], &peer);
5425	if (!addr) {
5426		err = -EINVAL;
5427		goto errout;
5428	}
5429	ifm = nlmsg_data(nlh);
5430	if (ifm->ifa_index)
5431		dev = dev_get_by_index(tgt_net, ifm->ifa_index);
5432
5433	ifa = ipv6_get_ifaddr(tgt_net, addr, dev, 1);
5434	if (!ifa) {
5435		err = -EADDRNOTAVAIL;
5436		goto errout;
5437	}
5438
5439	skb = nlmsg_new(inet6_ifaddr_msgsize(), GFP_KERNEL);
5440	if (!skb) {
5441		err = -ENOBUFS;
5442		goto errout_ifa;
5443	}
5444
5445	err = inet6_fill_ifaddr(skb, ifa, &fillargs);
5446	if (err < 0) {
5447		/* -EMSGSIZE implies BUG in inet6_ifaddr_msgsize() */
5448		WARN_ON(err == -EMSGSIZE);
5449		kfree_skb(skb);
5450		goto errout_ifa;
5451	}
5452	err = rtnl_unicast(skb, tgt_net, NETLINK_CB(in_skb).portid);
5453errout_ifa:
5454	in6_ifa_put(ifa);
5455errout:
5456	if (dev)
5457		dev_put(dev);
5458	if (fillargs.netnsid >= 0)
5459		put_net(tgt_net);
5460
5461	return err;
5462}
5463
5464static void inet6_ifa_notify(int event, struct inet6_ifaddr *ifa)
5465{
5466	struct sk_buff *skb;
5467	struct net *net = dev_net(ifa->idev->dev);
5468	struct inet6_fill_args fillargs = {
5469		.portid = 0,
5470		.seq = 0,
5471		.event = event,
5472		.flags = 0,
5473		.netnsid = -1,
5474	};
5475	int err = -ENOBUFS;
5476
5477	skb = nlmsg_new(inet6_ifaddr_msgsize(), GFP_ATOMIC);
5478	if (!skb)
5479		goto errout;
5480
5481	err = inet6_fill_ifaddr(skb, ifa, &fillargs);
5482	if (err < 0) {
5483		/* -EMSGSIZE implies BUG in inet6_ifaddr_msgsize() */
5484		WARN_ON(err == -EMSGSIZE);
5485		kfree_skb(skb);
5486		goto errout;
5487	}
5488	rtnl_notify(skb, net, 0, RTNLGRP_IPV6_IFADDR, NULL, GFP_ATOMIC);
5489	return;
5490errout:
5491	if (err < 0)
5492		rtnl_set_sk_err(net, RTNLGRP_IPV6_IFADDR, err);
5493}
5494
5495static inline void ipv6_store_devconf(struct ipv6_devconf *cnf,
5496				__s32 *array, int bytes)
5497{
5498	BUG_ON(bytes < (DEVCONF_MAX * 4));
5499
5500	memset(array, 0, bytes);
5501	array[DEVCONF_FORWARDING] = cnf->forwarding;
5502	array[DEVCONF_HOPLIMIT] = cnf->hop_limit;
5503	array[DEVCONF_MTU6] = cnf->mtu6;
5504	array[DEVCONF_ACCEPT_RA] = cnf->accept_ra;
5505	array[DEVCONF_ACCEPT_REDIRECTS] = cnf->accept_redirects;
5506	array[DEVCONF_AUTOCONF] = cnf->autoconf;
5507	array[DEVCONF_DAD_TRANSMITS] = cnf->dad_transmits;
5508	array[DEVCONF_RTR_SOLICITS] = cnf->rtr_solicits;
5509	array[DEVCONF_RTR_SOLICIT_INTERVAL] =
5510		jiffies_to_msecs(cnf->rtr_solicit_interval);
5511	array[DEVCONF_RTR_SOLICIT_MAX_INTERVAL] =
5512		jiffies_to_msecs(cnf->rtr_solicit_max_interval);
5513	array[DEVCONF_RTR_SOLICIT_DELAY] =
5514		jiffies_to_msecs(cnf->rtr_solicit_delay);
5515	array[DEVCONF_FORCE_MLD_VERSION] = cnf->force_mld_version;
5516	array[DEVCONF_MLDV1_UNSOLICITED_REPORT_INTERVAL] =
5517		jiffies_to_msecs(cnf->mldv1_unsolicited_report_interval);
5518	array[DEVCONF_MLDV2_UNSOLICITED_REPORT_INTERVAL] =
5519		jiffies_to_msecs(cnf->mldv2_unsolicited_report_interval);
5520	array[DEVCONF_USE_TEMPADDR] = cnf->use_tempaddr;
5521	array[DEVCONF_TEMP_VALID_LFT] = cnf->temp_valid_lft;
5522	array[DEVCONF_TEMP_PREFERED_LFT] = cnf->temp_prefered_lft;
5523	array[DEVCONF_REGEN_MAX_RETRY] = cnf->regen_max_retry;
5524	array[DEVCONF_MAX_DESYNC_FACTOR] = cnf->max_desync_factor;
5525	array[DEVCONF_MAX_ADDRESSES] = cnf->max_addresses;
5526	array[DEVCONF_ACCEPT_RA_DEFRTR] = cnf->accept_ra_defrtr;
5527	array[DEVCONF_ACCEPT_RA_MIN_HOP_LIMIT] = cnf->accept_ra_min_hop_limit;
5528	array[DEVCONF_ACCEPT_RA_PINFO] = cnf->accept_ra_pinfo;
5529#ifdef CONFIG_IPV6_ROUTER_PREF
5530	array[DEVCONF_ACCEPT_RA_RTR_PREF] = cnf->accept_ra_rtr_pref;
5531	array[DEVCONF_RTR_PROBE_INTERVAL] =
5532		jiffies_to_msecs(cnf->rtr_probe_interval);
5533#ifdef CONFIG_IPV6_ROUTE_INFO
5534	array[DEVCONF_ACCEPT_RA_RT_INFO_MIN_PLEN] = cnf->accept_ra_rt_info_min_plen;
5535	array[DEVCONF_ACCEPT_RA_RT_INFO_MAX_PLEN] = cnf->accept_ra_rt_info_max_plen;
5536#endif
5537#endif
5538	array[DEVCONF_PROXY_NDP] = cnf->proxy_ndp;
5539	array[DEVCONF_ACCEPT_SOURCE_ROUTE] = cnf->accept_source_route;
5540#ifdef CONFIG_IPV6_OPTIMISTIC_DAD
5541	array[DEVCONF_OPTIMISTIC_DAD] = cnf->optimistic_dad;
5542	array[DEVCONF_USE_OPTIMISTIC] = cnf->use_optimistic;
5543#endif
5544#ifdef CONFIG_IPV6_MROUTE
5545	array[DEVCONF_MC_FORWARDING] = atomic_read(&cnf->mc_forwarding);
5546#endif
5547	array[DEVCONF_DISABLE_IPV6] = cnf->disable_ipv6;
5548	array[DEVCONF_ACCEPT_DAD] = cnf->accept_dad;
5549	array[DEVCONF_FORCE_TLLAO] = cnf->force_tllao;
5550	array[DEVCONF_NDISC_NOTIFY] = cnf->ndisc_notify;
5551	array[DEVCONF_SUPPRESS_FRAG_NDISC] = cnf->suppress_frag_ndisc;
5552	array[DEVCONF_ACCEPT_RA_FROM_LOCAL] = cnf->accept_ra_from_local;
5553	array[DEVCONF_ACCEPT_RA_MTU] = cnf->accept_ra_mtu;
5554	array[DEVCONF_IGNORE_ROUTES_WITH_LINKDOWN] = cnf->ignore_routes_with_linkdown;
5555	/* we omit DEVCONF_STABLE_SECRET for now */
5556	array[DEVCONF_USE_OIF_ADDRS_ONLY] = cnf->use_oif_addrs_only;
5557	array[DEVCONF_DROP_UNICAST_IN_L2_MULTICAST] = cnf->drop_unicast_in_l2_multicast;
5558	array[DEVCONF_DROP_UNSOLICITED_NA] = cnf->drop_unsolicited_na;
5559	array[DEVCONF_KEEP_ADDR_ON_DOWN] = cnf->keep_addr_on_down;
5560	array[DEVCONF_SEG6_ENABLED] = cnf->seg6_enabled;
5561#ifdef CONFIG_IPV6_SEG6_HMAC
5562	array[DEVCONF_SEG6_REQUIRE_HMAC] = cnf->seg6_require_hmac;
5563#endif
5564	array[DEVCONF_ENHANCED_DAD] = cnf->enhanced_dad;
5565	array[DEVCONF_ADDR_GEN_MODE] = cnf->addr_gen_mode;
5566	array[DEVCONF_DISABLE_POLICY] = cnf->disable_policy;
5567	array[DEVCONF_NDISC_TCLASS] = cnf->ndisc_tclass;
5568	array[DEVCONF_RPL_SEG_ENABLED] = cnf->rpl_seg_enabled;
5569	array[DEVCONF_ACCEPT_RA_MIN_LFT] = cnf->accept_ra_min_lft;
5570}
5571
5572static inline size_t inet6_ifla6_size(void)
5573{
5574	return nla_total_size(4) /* IFLA_INET6_FLAGS */
5575	     + nla_total_size(sizeof(struct ifla_cacheinfo))
5576	     + nla_total_size(DEVCONF_MAX * 4) /* IFLA_INET6_CONF */
5577	     + nla_total_size(IPSTATS_MIB_MAX * 8) /* IFLA_INET6_STATS */
5578	     + nla_total_size(ICMP6_MIB_MAX * 8) /* IFLA_INET6_ICMP6STATS */
5579	     + nla_total_size(sizeof(struct in6_addr)) /* IFLA_INET6_TOKEN */
5580	     + nla_total_size(1) /* IFLA_INET6_ADDR_GEN_MODE */
5581	     + 0;
5582}
5583
5584static inline size_t inet6_if_nlmsg_size(void)
5585{
5586	return NLMSG_ALIGN(sizeof(struct ifinfomsg))
5587	       + nla_total_size(IFNAMSIZ) /* IFLA_IFNAME */
5588	       + nla_total_size(MAX_ADDR_LEN) /* IFLA_ADDRESS */
5589	       + nla_total_size(4) /* IFLA_MTU */
5590	       + nla_total_size(4) /* IFLA_LINK */
5591	       + nla_total_size(1) /* IFLA_OPERSTATE */
5592	       + nla_total_size(inet6_ifla6_size()); /* IFLA_PROTINFO */
5593}
5594
5595static inline void __snmp6_fill_statsdev(u64 *stats, atomic_long_t *mib,
5596					int bytes)
5597{
5598	int i;
5599	int pad = bytes - sizeof(u64) * ICMP6_MIB_MAX;
5600	BUG_ON(pad < 0);
5601
5602	/* Use put_unaligned() because stats may not be aligned for u64. */
5603	put_unaligned(ICMP6_MIB_MAX, &stats[0]);
5604	for (i = 1; i < ICMP6_MIB_MAX; i++)
5605		put_unaligned(atomic_long_read(&mib[i]), &stats[i]);
5606
5607	memset(&stats[ICMP6_MIB_MAX], 0, pad);
5608}
5609
5610static inline void __snmp6_fill_stats64(u64 *stats, void __percpu *mib,
5611					int bytes, size_t syncpoff)
5612{
5613	int i, c;
5614	u64 buff[IPSTATS_MIB_MAX];
5615	int pad = bytes - sizeof(u64) * IPSTATS_MIB_MAX;
5616
5617	BUG_ON(pad < 0);
5618
5619	memset(buff, 0, sizeof(buff));
5620	buff[0] = IPSTATS_MIB_MAX;
5621
5622	for_each_possible_cpu(c) {
5623		for (i = 1; i < IPSTATS_MIB_MAX; i++)
5624			buff[i] += snmp_get_cpu_field64(mib, c, i, syncpoff);
5625	}
5626
5627	memcpy(stats, buff, IPSTATS_MIB_MAX * sizeof(u64));
5628	memset(&stats[IPSTATS_MIB_MAX], 0, pad);
5629}
5630
5631static void snmp6_fill_stats(u64 *stats, struct inet6_dev *idev, int attrtype,
5632			     int bytes)
5633{
5634	switch (attrtype) {
5635	case IFLA_INET6_STATS:
5636		__snmp6_fill_stats64(stats, idev->stats.ipv6, bytes,
5637				     offsetof(struct ipstats_mib, syncp));
5638		break;
5639	case IFLA_INET6_ICMP6STATS:
5640		__snmp6_fill_statsdev(stats, idev->stats.icmpv6dev->mibs, bytes);
5641		break;
5642	}
5643}
5644
5645static int inet6_fill_ifla6_attrs(struct sk_buff *skb, struct inet6_dev *idev,
5646				  u32 ext_filter_mask)
5647{
5648	struct nlattr *nla;
5649	struct ifla_cacheinfo ci;
5650
5651	if (nla_put_u32(skb, IFLA_INET6_FLAGS, idev->if_flags))
5652		goto nla_put_failure;
5653	ci.max_reasm_len = IPV6_MAXPLEN;
5654	ci.tstamp = cstamp_delta(idev->tstamp);
5655	ci.reachable_time = jiffies_to_msecs(idev->nd_parms->reachable_time);
5656	ci.retrans_time = jiffies_to_msecs(NEIGH_VAR(idev->nd_parms, RETRANS_TIME));
5657	if (nla_put(skb, IFLA_INET6_CACHEINFO, sizeof(ci), &ci))
5658		goto nla_put_failure;
5659	nla = nla_reserve(skb, IFLA_INET6_CONF, DEVCONF_MAX * sizeof(s32));
5660	if (!nla)
5661		goto nla_put_failure;
5662	ipv6_store_devconf(&idev->cnf, nla_data(nla), nla_len(nla));
5663
5664	/* XXX - MC not implemented */
5665
5666	if (ext_filter_mask & RTEXT_FILTER_SKIP_STATS)
5667		return 0;
5668
5669	nla = nla_reserve(skb, IFLA_INET6_STATS, IPSTATS_MIB_MAX * sizeof(u64));
5670	if (!nla)
5671		goto nla_put_failure;
5672	snmp6_fill_stats(nla_data(nla), idev, IFLA_INET6_STATS, nla_len(nla));
5673
5674	nla = nla_reserve(skb, IFLA_INET6_ICMP6STATS, ICMP6_MIB_MAX * sizeof(u64));
5675	if (!nla)
5676		goto nla_put_failure;
5677	snmp6_fill_stats(nla_data(nla), idev, IFLA_INET6_ICMP6STATS, nla_len(nla));
5678
5679	nla = nla_reserve(skb, IFLA_INET6_TOKEN, sizeof(struct in6_addr));
5680	if (!nla)
5681		goto nla_put_failure;
5682	read_lock_bh(&idev->lock);
5683	memcpy(nla_data(nla), idev->token.s6_addr, nla_len(nla));
5684	read_unlock_bh(&idev->lock);
5685
5686	if (nla_put_u8(skb, IFLA_INET6_ADDR_GEN_MODE, idev->cnf.addr_gen_mode))
5687		goto nla_put_failure;
5688
5689	return 0;
5690
5691nla_put_failure:
5692	return -EMSGSIZE;
5693}
5694
5695static size_t inet6_get_link_af_size(const struct net_device *dev,
5696				     u32 ext_filter_mask)
5697{
5698	if (!__in6_dev_get(dev))
5699		return 0;
5700
5701	return inet6_ifla6_size();
5702}
5703
5704static int inet6_fill_link_af(struct sk_buff *skb, const struct net_device *dev,
5705			      u32 ext_filter_mask)
5706{
5707	struct inet6_dev *idev = __in6_dev_get(dev);
5708
5709	if (!idev)
5710		return -ENODATA;
5711
5712	if (inet6_fill_ifla6_attrs(skb, idev, ext_filter_mask) < 0)
5713		return -EMSGSIZE;
5714
5715	return 0;
5716}
5717
5718static int inet6_set_iftoken(struct inet6_dev *idev, struct in6_addr *token)
5719{
5720	struct inet6_ifaddr *ifp;
5721	struct net_device *dev = idev->dev;
5722	bool clear_token, update_rs = false;
5723	struct in6_addr ll_addr;
5724
5725	ASSERT_RTNL();
5726
5727	if (!token)
5728		return -EINVAL;
5729	if (dev->flags & (IFF_LOOPBACK | IFF_NOARP))
5730		return -EINVAL;
5731	if (!ipv6_accept_ra(idev))
5732		return -EINVAL;
5733	if (idev->cnf.rtr_solicits == 0)
5734		return -EINVAL;
5735
5736	write_lock_bh(&idev->lock);
5737
5738	BUILD_BUG_ON(sizeof(token->s6_addr) != 16);
5739	memcpy(idev->token.s6_addr + 8, token->s6_addr + 8, 8);
5740
5741	write_unlock_bh(&idev->lock);
5742
5743	clear_token = ipv6_addr_any(token);
5744	if (clear_token)
5745		goto update_lft;
5746
5747	if (!idev->dead && (idev->if_flags & IF_READY) &&
5748	    !ipv6_get_lladdr(dev, &ll_addr, IFA_F_TENTATIVE |
5749			     IFA_F_OPTIMISTIC)) {
5750		/* If we're not ready, then normal ifup will take care
5751		 * of this. Otherwise, we need to request our rs here.
5752		 */
5753		ndisc_send_rs(dev, &ll_addr, &in6addr_linklocal_allrouters);
5754		update_rs = true;
5755	}
5756
5757update_lft:
5758	write_lock_bh(&idev->lock);
5759
5760	if (update_rs) {
5761		idev->if_flags |= IF_RS_SENT;
5762		idev->rs_interval = rfc3315_s14_backoff_init(
5763			idev->cnf.rtr_solicit_interval);
5764		idev->rs_probes = 1;
5765		addrconf_mod_rs_timer(idev, idev->rs_interval);
5766	}
5767
5768	/* Well, that's kinda nasty ... */
5769	list_for_each_entry(ifp, &idev->addr_list, if_list) {
5770		spin_lock(&ifp->lock);
5771		if (ifp->tokenized) {
5772			ifp->valid_lft = 0;
5773			ifp->prefered_lft = 0;
5774		}
5775		spin_unlock(&ifp->lock);
5776	}
5777
5778	write_unlock_bh(&idev->lock);
5779	inet6_ifinfo_notify(RTM_NEWLINK, idev);
5780	addrconf_verify_rtnl();
5781	return 0;
5782}
5783
5784static const struct nla_policy inet6_af_policy[IFLA_INET6_MAX + 1] = {
5785	[IFLA_INET6_ADDR_GEN_MODE]	= { .type = NLA_U8 },
5786	[IFLA_INET6_TOKEN]		= { .len = sizeof(struct in6_addr) },
5787};
5788
5789static int check_addr_gen_mode(int mode)
5790{
5791	if (mode != IN6_ADDR_GEN_MODE_EUI64 &&
5792	    mode != IN6_ADDR_GEN_MODE_NONE &&
5793	    mode != IN6_ADDR_GEN_MODE_STABLE_PRIVACY &&
5794	    mode != IN6_ADDR_GEN_MODE_RANDOM)
5795		return -EINVAL;
5796	return 1;
5797}
5798
5799static int check_stable_privacy(struct inet6_dev *idev, struct net *net,
5800				int mode)
5801{
5802	if (mode == IN6_ADDR_GEN_MODE_STABLE_PRIVACY &&
5803	    !idev->cnf.stable_secret.initialized &&
5804	    !net->ipv6.devconf_dflt->stable_secret.initialized)
5805		return -EINVAL;
5806	return 1;
5807}
5808
5809static int inet6_validate_link_af(const struct net_device *dev,
5810				  const struct nlattr *nla)
5811{
5812	struct nlattr *tb[IFLA_INET6_MAX + 1];
5813	struct inet6_dev *idev = NULL;
5814	int err;
5815
5816	if (dev) {
5817		idev = __in6_dev_get(dev);
5818		if (!idev)
5819			return -EAFNOSUPPORT;
5820	}
5821
5822	err = nla_parse_nested_deprecated(tb, IFLA_INET6_MAX, nla,
5823					  inet6_af_policy, NULL);
5824	if (err)
5825		return err;
5826
5827	if (!tb[IFLA_INET6_TOKEN] && !tb[IFLA_INET6_ADDR_GEN_MODE])
5828		return -EINVAL;
5829
5830	if (tb[IFLA_INET6_ADDR_GEN_MODE]) {
5831		u8 mode = nla_get_u8(tb[IFLA_INET6_ADDR_GEN_MODE]);
5832
5833		if (check_addr_gen_mode(mode) < 0)
5834			return -EINVAL;
5835		if (dev && check_stable_privacy(idev, dev_net(dev), mode) < 0)
5836			return -EINVAL;
5837	}
5838
5839	return 0;
5840}
5841
5842static int inet6_set_link_af(struct net_device *dev, const struct nlattr *nla)
5843{
5844	struct inet6_dev *idev = __in6_dev_get(dev);
5845	struct nlattr *tb[IFLA_INET6_MAX + 1];
5846	int err;
5847
5848	if (!idev)
5849		return -EAFNOSUPPORT;
5850
5851	if (nla_parse_nested_deprecated(tb, IFLA_INET6_MAX, nla, NULL, NULL) < 0)
5852		return -EINVAL;
5853
5854	if (tb[IFLA_INET6_TOKEN]) {
5855		err = inet6_set_iftoken(idev, nla_data(tb[IFLA_INET6_TOKEN]));
5856		if (err)
5857			return err;
5858	}
5859
5860	if (tb[IFLA_INET6_ADDR_GEN_MODE]) {
5861		u8 mode = nla_get_u8(tb[IFLA_INET6_ADDR_GEN_MODE]);
5862
5863		idev->cnf.addr_gen_mode = mode;
5864	}
5865
5866	return 0;
5867}
5868
5869static int inet6_fill_ifinfo(struct sk_buff *skb, struct inet6_dev *idev,
5870			     u32 portid, u32 seq, int event, unsigned int flags)
5871{
5872	struct net_device *dev = idev->dev;
5873	struct ifinfomsg *hdr;
5874	struct nlmsghdr *nlh;
5875	void *protoinfo;
5876
5877	nlh = nlmsg_put(skb, portid, seq, event, sizeof(*hdr), flags);
5878	if (!nlh)
5879		return -EMSGSIZE;
5880
5881	hdr = nlmsg_data(nlh);
5882	hdr->ifi_family = AF_INET6;
5883	hdr->__ifi_pad = 0;
5884	hdr->ifi_type = dev->type;
5885	hdr->ifi_index = dev->ifindex;
5886	hdr->ifi_flags = dev_get_flags(dev);
5887	hdr->ifi_change = 0;
5888
5889	if (nla_put_string(skb, IFLA_IFNAME, dev->name) ||
5890	    (dev->addr_len &&
5891	     nla_put(skb, IFLA_ADDRESS, dev->addr_len, dev->dev_addr)) ||
5892	    nla_put_u32(skb, IFLA_MTU, dev->mtu) ||
5893	    (dev->ifindex != dev_get_iflink(dev) &&
5894	     nla_put_u32(skb, IFLA_LINK, dev_get_iflink(dev))) ||
5895	    nla_put_u8(skb, IFLA_OPERSTATE,
5896		       netif_running(dev) ? dev->operstate : IF_OPER_DOWN))
5897		goto nla_put_failure;
5898	protoinfo = nla_nest_start_noflag(skb, IFLA_PROTINFO);
5899	if (!protoinfo)
5900		goto nla_put_failure;
5901
5902	if (inet6_fill_ifla6_attrs(skb, idev, 0) < 0)
5903		goto nla_put_failure;
5904
5905	nla_nest_end(skb, protoinfo);
5906	nlmsg_end(skb, nlh);
5907	return 0;
5908
5909nla_put_failure:
5910	nlmsg_cancel(skb, nlh);
5911	return -EMSGSIZE;
5912}
5913
5914static int inet6_valid_dump_ifinfo(const struct nlmsghdr *nlh,
5915				   struct netlink_ext_ack *extack)
5916{
5917	struct ifinfomsg *ifm;
5918
5919	if (nlh->nlmsg_len < nlmsg_msg_size(sizeof(*ifm))) {
5920		NL_SET_ERR_MSG_MOD(extack, "Invalid header for link dump request");
5921		return -EINVAL;
5922	}
5923
5924	if (nlmsg_attrlen(nlh, sizeof(*ifm))) {
5925		NL_SET_ERR_MSG_MOD(extack, "Invalid data after header");
5926		return -EINVAL;
5927	}
5928
5929	ifm = nlmsg_data(nlh);
5930	if (ifm->__ifi_pad || ifm->ifi_type || ifm->ifi_flags ||
5931	    ifm->ifi_change || ifm->ifi_index) {
5932		NL_SET_ERR_MSG_MOD(extack, "Invalid values in header for dump request");
5933		return -EINVAL;
5934	}
5935
5936	return 0;
5937}
5938
5939static int inet6_dump_ifinfo(struct sk_buff *skb, struct netlink_callback *cb)
5940{
5941	struct net *net = sock_net(skb->sk);
5942	int h, s_h;
5943	int idx = 0, s_idx;
5944	struct net_device *dev;
5945	struct inet6_dev *idev;
5946	struct hlist_head *head;
5947
5948	/* only requests using strict checking can pass data to
5949	 * influence the dump
5950	 */
5951	if (cb->strict_check) {
5952		int err = inet6_valid_dump_ifinfo(cb->nlh, cb->extack);
5953
5954		if (err < 0)
5955			return err;
5956	}
5957
5958	s_h = cb->args[0];
5959	s_idx = cb->args[1];
5960
5961	rcu_read_lock();
5962	for (h = s_h; h < NETDEV_HASHENTRIES; h++, s_idx = 0) {
5963		idx = 0;
5964		head = &net->dev_index_head[h];
5965		hlist_for_each_entry_rcu(dev, head, index_hlist) {
5966			if (idx < s_idx)
5967				goto cont;
5968			idev = __in6_dev_get(dev);
5969			if (!idev)
5970				goto cont;
5971			if (inet6_fill_ifinfo(skb, idev,
5972					      NETLINK_CB(cb->skb).portid,
5973					      cb->nlh->nlmsg_seq,
5974					      RTM_NEWLINK, NLM_F_MULTI) < 0)
5975				goto out;
5976cont:
5977			idx++;
5978		}
5979	}
5980out:
5981	rcu_read_unlock();
5982	cb->args[1] = idx;
5983	cb->args[0] = h;
5984
5985	return skb->len;
5986}
5987
5988void inet6_ifinfo_notify(int event, struct inet6_dev *idev)
5989{
5990	struct sk_buff *skb;
5991	struct net *net = dev_net(idev->dev);
5992	int err = -ENOBUFS;
5993
5994	skb = nlmsg_new(inet6_if_nlmsg_size(), GFP_ATOMIC);
5995	if (!skb)
5996		goto errout;
5997
5998	err = inet6_fill_ifinfo(skb, idev, 0, 0, event, 0);
5999	if (err < 0) {
6000		/* -EMSGSIZE implies BUG in inet6_if_nlmsg_size() */
6001		WARN_ON(err == -EMSGSIZE);
6002		kfree_skb(skb);
6003		goto errout;
6004	}
6005	rtnl_notify(skb, net, 0, RTNLGRP_IPV6_IFINFO, NULL, GFP_ATOMIC);
6006	return;
6007errout:
6008	if (err < 0)
6009		rtnl_set_sk_err(net, RTNLGRP_IPV6_IFINFO, err);
6010}
6011
6012static inline size_t inet6_prefix_nlmsg_size(void)
6013{
6014	return NLMSG_ALIGN(sizeof(struct prefixmsg))
6015	       + nla_total_size(sizeof(struct in6_addr))
6016	       + nla_total_size(sizeof(struct prefix_cacheinfo));
6017}
6018
6019static int inet6_fill_prefix(struct sk_buff *skb, struct inet6_dev *idev,
6020			     struct prefix_info *pinfo, u32 portid, u32 seq,
6021			     int event, unsigned int flags)
6022{
6023	struct prefixmsg *pmsg;
6024	struct nlmsghdr *nlh;
6025	struct prefix_cacheinfo	ci;
6026
6027	nlh = nlmsg_put(skb, portid, seq, event, sizeof(*pmsg), flags);
6028	if (!nlh)
6029		return -EMSGSIZE;
6030
6031	pmsg = nlmsg_data(nlh);
6032	pmsg->prefix_family = AF_INET6;
6033	pmsg->prefix_pad1 = 0;
6034	pmsg->prefix_pad2 = 0;
6035	pmsg->prefix_ifindex = idev->dev->ifindex;
6036	pmsg->prefix_len = pinfo->prefix_len;
6037	pmsg->prefix_type = pinfo->type;
6038	pmsg->prefix_pad3 = 0;
6039	pmsg->prefix_flags = pinfo->flags;
6040
6041	if (nla_put(skb, PREFIX_ADDRESS, sizeof(pinfo->prefix), &pinfo->prefix))
6042		goto nla_put_failure;
6043	ci.preferred_time = ntohl(pinfo->prefered);
6044	ci.valid_time = ntohl(pinfo->valid);
6045	if (nla_put(skb, PREFIX_CACHEINFO, sizeof(ci), &ci))
6046		goto nla_put_failure;
6047	nlmsg_end(skb, nlh);
6048	return 0;
6049
6050nla_put_failure:
6051	nlmsg_cancel(skb, nlh);
6052	return -EMSGSIZE;
6053}
6054
6055static void inet6_prefix_notify(int event, struct inet6_dev *idev,
6056			 struct prefix_info *pinfo)
6057{
6058	struct sk_buff *skb;
6059	struct net *net = dev_net(idev->dev);
6060	int err = -ENOBUFS;
6061
6062	skb = nlmsg_new(inet6_prefix_nlmsg_size(), GFP_ATOMIC);
6063	if (!skb)
6064		goto errout;
6065
6066	err = inet6_fill_prefix(skb, idev, pinfo, 0, 0, event, 0);
6067	if (err < 0) {
6068		/* -EMSGSIZE implies BUG in inet6_prefix_nlmsg_size() */
6069		WARN_ON(err == -EMSGSIZE);
6070		kfree_skb(skb);
6071		goto errout;
6072	}
6073	rtnl_notify(skb, net, 0, RTNLGRP_IPV6_PREFIX, NULL, GFP_ATOMIC);
6074	return;
6075errout:
6076	if (err < 0)
6077		rtnl_set_sk_err(net, RTNLGRP_IPV6_PREFIX, err);
6078}
6079
6080static void __ipv6_ifa_notify(int event, struct inet6_ifaddr *ifp)
6081{
6082	struct net *net = dev_net(ifp->idev->dev);
6083
6084	if (event)
6085		ASSERT_RTNL();
6086
6087	inet6_ifa_notify(event ? : RTM_NEWADDR, ifp);
6088
6089	switch (event) {
6090	case RTM_NEWADDR:
6091		/*
6092		 * If the address was optimistic we inserted the route at the
6093		 * start of our DAD process, so we don't need to do it again.
6094		 * If the device was taken down in the middle of the DAD
6095		 * cycle there is a race where we could get here without a
6096		 * host route, so nothing to insert. That will be fixed when
6097		 * the device is brought up.
6098		 */
6099		if (ifp->rt && !rcu_access_pointer(ifp->rt->fib6_node)) {
6100			ip6_ins_rt(net, ifp->rt);
6101		} else if (!ifp->rt && (ifp->idev->dev->flags & IFF_UP)) {
6102			pr_warn("BUG: Address %pI6c on device %s is missing its host route.\n",
6103				&ifp->addr, ifp->idev->dev->name);
6104		}
6105
6106		if (ifp->idev->cnf.forwarding)
6107			addrconf_join_anycast(ifp);
6108		if (!ipv6_addr_any(&ifp->peer_addr))
6109			addrconf_prefix_route(&ifp->peer_addr, 128,
6110					      ifp->rt_priority, ifp->idev->dev,
6111					      0, 0, GFP_ATOMIC);
6112		break;
6113	case RTM_DELADDR:
6114		if (ifp->idev->cnf.forwarding)
6115			addrconf_leave_anycast(ifp);
6116		addrconf_leave_solict(ifp->idev, &ifp->addr);
6117		if (!ipv6_addr_any(&ifp->peer_addr)) {
6118			struct fib6_info *rt;
6119
6120			rt = addrconf_get_prefix_route(&ifp->peer_addr, 128,
6121						       ifp->idev->dev, 0, 0,
6122						       false);
6123			if (rt)
6124				ip6_del_rt(net, rt, false);
6125		}
6126		if (ifp->rt) {
6127			ip6_del_rt(net, ifp->rt, false);
6128			ifp->rt = NULL;
6129		}
6130		rt_genid_bump_ipv6(net);
6131		break;
6132	}
6133	atomic_inc(&net->ipv6.dev_addr_genid);
6134}
6135
6136static void ipv6_ifa_notify(int event, struct inet6_ifaddr *ifp)
6137{
6138	rcu_read_lock_bh();
6139	if (likely(ifp->idev->dead == 0))
6140		__ipv6_ifa_notify(event, ifp);
6141	rcu_read_unlock_bh();
6142}
6143
6144#ifdef CONFIG_SYSCTL
6145
6146static int addrconf_sysctl_forward(struct ctl_table *ctl, int write,
6147		void *buffer, size_t *lenp, loff_t *ppos)
6148{
6149	int *valp = ctl->data;
6150	int val = *valp;
6151	loff_t pos = *ppos;
6152	struct ctl_table lctl;
6153	int ret;
6154
6155	/*
6156	 * ctl->data points to idev->cnf.forwarding, we should
6157	 * not modify it until we get the rtnl lock.
6158	 */
6159	lctl = *ctl;
6160	lctl.data = &val;
6161
6162	ret = proc_dointvec(&lctl, write, buffer, lenp, ppos);
6163
6164	if (write)
6165		ret = addrconf_fixup_forwarding(ctl, valp, val);
6166	if (ret)
6167		*ppos = pos;
6168	return ret;
6169}
6170
6171static int addrconf_sysctl_mtu(struct ctl_table *ctl, int write,
6172		void *buffer, size_t *lenp, loff_t *ppos)
6173{
6174	struct inet6_dev *idev = ctl->extra1;
6175	int min_mtu = IPV6_MIN_MTU;
6176	struct ctl_table lctl;
6177
6178	lctl = *ctl;
6179	lctl.extra1 = &min_mtu;
6180	lctl.extra2 = idev ? &idev->dev->mtu : NULL;
6181
6182	return proc_dointvec_minmax(&lctl, write, buffer, lenp, ppos);
6183}
6184
6185static void dev_disable_change(struct inet6_dev *idev)
6186{
6187	struct netdev_notifier_info info;
6188
6189	if (!idev || !idev->dev)
6190		return;
6191
6192	netdev_notifier_info_init(&info, idev->dev);
6193	if (idev->cnf.disable_ipv6)
6194		addrconf_notify(NULL, NETDEV_DOWN, &info);
6195	else
6196		addrconf_notify(NULL, NETDEV_UP, &info);
6197}
6198
6199static void addrconf_disable_change(struct net *net, __s32 newf)
6200{
6201	struct net_device *dev;
6202	struct inet6_dev *idev;
6203
6204	for_each_netdev(net, dev) {
6205		idev = __in6_dev_get(dev);
6206		if (idev) {
6207			int changed = (!idev->cnf.disable_ipv6) ^ (!newf);
6208			idev->cnf.disable_ipv6 = newf;
6209			if (changed)
6210				dev_disable_change(idev);
6211		}
6212	}
6213}
6214
6215static int addrconf_disable_ipv6(struct ctl_table *table, int *p, int newf)
6216{
6217	struct net *net;
6218	int old;
6219
6220	if (!rtnl_trylock())
6221		return restart_syscall();
6222
6223	net = (struct net *)table->extra2;
6224	old = *p;
6225	*p = newf;
6226
6227	if (p == &net->ipv6.devconf_dflt->disable_ipv6) {
6228		rtnl_unlock();
6229		return 0;
6230	}
6231
6232	if (p == &net->ipv6.devconf_all->disable_ipv6) {
6233		net->ipv6.devconf_dflt->disable_ipv6 = newf;
6234		addrconf_disable_change(net, newf);
6235	} else if ((!newf) ^ (!old))
6236		dev_disable_change((struct inet6_dev *)table->extra1);
6237
6238	rtnl_unlock();
6239	return 0;
6240}
6241
6242static int addrconf_sysctl_disable(struct ctl_table *ctl, int write,
6243		void *buffer, size_t *lenp, loff_t *ppos)
6244{
6245	int *valp = ctl->data;
6246	int val = *valp;
6247	loff_t pos = *ppos;
6248	struct ctl_table lctl;
6249	int ret;
6250
6251	/*
6252	 * ctl->data points to idev->cnf.disable_ipv6, we should
6253	 * not modify it until we get the rtnl lock.
6254	 */
6255	lctl = *ctl;
6256	lctl.data = &val;
6257
6258	ret = proc_dointvec(&lctl, write, buffer, lenp, ppos);
6259
6260	if (write)
6261		ret = addrconf_disable_ipv6(ctl, valp, val);
6262	if (ret)
6263		*ppos = pos;
6264	return ret;
6265}
6266
6267static int addrconf_sysctl_proxy_ndp(struct ctl_table *ctl, int write,
6268		void *buffer, size_t *lenp, loff_t *ppos)
6269{
6270	int *valp = ctl->data;
6271	int ret;
6272	int old, new;
6273
6274	old = *valp;
6275	ret = proc_dointvec(ctl, write, buffer, lenp, ppos);
6276	new = *valp;
6277
6278	if (write && old != new) {
6279		struct net *net = ctl->extra2;
6280
6281		if (!rtnl_trylock())
6282			return restart_syscall();
6283
6284		if (valp == &net->ipv6.devconf_dflt->proxy_ndp)
6285			inet6_netconf_notify_devconf(net, RTM_NEWNETCONF,
6286						     NETCONFA_PROXY_NEIGH,
6287						     NETCONFA_IFINDEX_DEFAULT,
6288						     net->ipv6.devconf_dflt);
6289		else if (valp == &net->ipv6.devconf_all->proxy_ndp)
6290			inet6_netconf_notify_devconf(net, RTM_NEWNETCONF,
6291						     NETCONFA_PROXY_NEIGH,
6292						     NETCONFA_IFINDEX_ALL,
6293						     net->ipv6.devconf_all);
6294		else {
6295			struct inet6_dev *idev = ctl->extra1;
6296
6297			inet6_netconf_notify_devconf(net, RTM_NEWNETCONF,
6298						     NETCONFA_PROXY_NEIGH,
6299						     idev->dev->ifindex,
6300						     &idev->cnf);
6301		}
6302		rtnl_unlock();
6303	}
6304
6305	return ret;
6306}
6307
6308static int addrconf_sysctl_addr_gen_mode(struct ctl_table *ctl, int write,
6309					 void *buffer, size_t *lenp,
6310					 loff_t *ppos)
6311{
6312	int ret = 0;
6313	u32 new_val;
6314	struct inet6_dev *idev = (struct inet6_dev *)ctl->extra1;
6315	struct net *net = (struct net *)ctl->extra2;
6316	struct ctl_table tmp = {
6317		.data = &new_val,
6318		.maxlen = sizeof(new_val),
6319		.mode = ctl->mode,
6320	};
6321
6322	if (!rtnl_trylock())
6323		return restart_syscall();
6324
6325	new_val = *((u32 *)ctl->data);
6326
6327	ret = proc_douintvec(&tmp, write, buffer, lenp, ppos);
6328	if (ret != 0)
6329		goto out;
6330
6331	if (write) {
6332		if (check_addr_gen_mode(new_val) < 0) {
6333			ret = -EINVAL;
6334			goto out;
6335		}
6336
6337		if (idev) {
6338			if (check_stable_privacy(idev, net, new_val) < 0) {
6339				ret = -EINVAL;
6340				goto out;
6341			}
6342
6343			if (idev->cnf.addr_gen_mode != new_val) {
6344				idev->cnf.addr_gen_mode = new_val;
6345				addrconf_dev_config(idev->dev);
6346			}
6347		} else if (&net->ipv6.devconf_all->addr_gen_mode == ctl->data) {
6348			struct net_device *dev;
6349
6350			net->ipv6.devconf_dflt->addr_gen_mode = new_val;
6351			for_each_netdev(net, dev) {
6352				idev = __in6_dev_get(dev);
6353				if (idev &&
6354				    idev->cnf.addr_gen_mode != new_val) {
6355					idev->cnf.addr_gen_mode = new_val;
6356					addrconf_dev_config(idev->dev);
6357				}
6358			}
6359		}
6360
6361		*((u32 *)ctl->data) = new_val;
6362	}
6363
6364out:
6365	rtnl_unlock();
6366
6367	return ret;
6368}
6369
6370static int addrconf_sysctl_stable_secret(struct ctl_table *ctl, int write,
6371					 void *buffer, size_t *lenp,
6372					 loff_t *ppos)
6373{
6374	int err;
6375	struct in6_addr addr;
6376	char str[IPV6_MAX_STRLEN];
6377	struct ctl_table lctl = *ctl;
6378	struct net *net = ctl->extra2;
6379	struct ipv6_stable_secret *secret = ctl->data;
6380
6381	if (&net->ipv6.devconf_all->stable_secret == ctl->data)
6382		return -EIO;
6383
6384	lctl.maxlen = IPV6_MAX_STRLEN;
6385	lctl.data = str;
6386
6387	if (!rtnl_trylock())
6388		return restart_syscall();
6389
6390	if (!write && !secret->initialized) {
6391		err = -EIO;
6392		goto out;
6393	}
6394
6395	err = snprintf(str, sizeof(str), "%pI6", &secret->secret);
6396	if (err >= sizeof(str)) {
6397		err = -EIO;
6398		goto out;
6399	}
6400
6401	err = proc_dostring(&lctl, write, buffer, lenp, ppos);
6402	if (err || !write)
6403		goto out;
6404
6405	if (in6_pton(str, -1, addr.in6_u.u6_addr8, -1, NULL) != 1) {
6406		err = -EIO;
6407		goto out;
6408	}
6409
6410	secret->initialized = true;
6411	secret->secret = addr;
6412
6413	if (&net->ipv6.devconf_dflt->stable_secret == ctl->data) {
6414		struct net_device *dev;
6415
6416		for_each_netdev(net, dev) {
6417			struct inet6_dev *idev = __in6_dev_get(dev);
6418
6419			if (idev) {
6420				idev->cnf.addr_gen_mode =
6421					IN6_ADDR_GEN_MODE_STABLE_PRIVACY;
6422			}
6423		}
6424	} else {
6425		struct inet6_dev *idev = ctl->extra1;
6426
6427		idev->cnf.addr_gen_mode = IN6_ADDR_GEN_MODE_STABLE_PRIVACY;
6428	}
6429
6430out:
6431	rtnl_unlock();
6432
6433	return err;
6434}
6435
6436static
6437int addrconf_sysctl_ignore_routes_with_linkdown(struct ctl_table *ctl,
6438						int write, void *buffer,
6439						size_t *lenp,
6440						loff_t *ppos)
6441{
6442	int *valp = ctl->data;
6443	int val = *valp;
6444	loff_t pos = *ppos;
6445	struct ctl_table lctl;
6446	int ret;
6447
6448	/* ctl->data points to idev->cnf.ignore_routes_when_linkdown
6449	 * we should not modify it until we get the rtnl lock.
6450	 */
6451	lctl = *ctl;
6452	lctl.data = &val;
6453
6454	ret = proc_dointvec(&lctl, write, buffer, lenp, ppos);
6455
6456	if (write)
6457		ret = addrconf_fixup_linkdown(ctl, valp, val);
6458	if (ret)
6459		*ppos = pos;
6460	return ret;
6461}
6462
6463static
6464void addrconf_set_nopolicy(struct rt6_info *rt, int action)
6465{
6466	if (rt) {
6467		if (action)
6468			rt->dst.flags |= DST_NOPOLICY;
6469		else
6470			rt->dst.flags &= ~DST_NOPOLICY;
6471	}
6472}
6473
6474static
6475void addrconf_disable_policy_idev(struct inet6_dev *idev, int val)
6476{
6477	struct inet6_ifaddr *ifa;
6478
6479	read_lock_bh(&idev->lock);
6480	list_for_each_entry(ifa, &idev->addr_list, if_list) {
6481		spin_lock(&ifa->lock);
6482		if (ifa->rt) {
6483			/* host routes only use builtin fib6_nh */
6484			struct fib6_nh *nh = ifa->rt->fib6_nh;
6485			int cpu;
6486
6487			rcu_read_lock();
6488			ifa->rt->dst_nopolicy = val ? true : false;
6489			if (nh->rt6i_pcpu) {
6490				for_each_possible_cpu(cpu) {
6491					struct rt6_info **rtp;
6492
6493					rtp = per_cpu_ptr(nh->rt6i_pcpu, cpu);
6494					addrconf_set_nopolicy(*rtp, val);
6495				}
6496			}
6497			rcu_read_unlock();
6498		}
6499		spin_unlock(&ifa->lock);
6500	}
6501	read_unlock_bh(&idev->lock);
6502}
6503
6504static
6505int addrconf_disable_policy(struct ctl_table *ctl, int *valp, int val)
6506{
6507	struct inet6_dev *idev;
6508	struct net *net;
6509
6510	if (!rtnl_trylock())
6511		return restart_syscall();
6512
6513	*valp = val;
6514
6515	net = (struct net *)ctl->extra2;
6516	if (valp == &net->ipv6.devconf_dflt->disable_policy) {
6517		rtnl_unlock();
6518		return 0;
6519	}
6520
6521	if (valp == &net->ipv6.devconf_all->disable_policy)  {
6522		struct net_device *dev;
6523
6524		for_each_netdev(net, dev) {
6525			idev = __in6_dev_get(dev);
6526			if (idev)
6527				addrconf_disable_policy_idev(idev, val);
6528		}
6529	} else {
6530		idev = (struct inet6_dev *)ctl->extra1;
6531		addrconf_disable_policy_idev(idev, val);
6532	}
6533
6534	rtnl_unlock();
6535	return 0;
6536}
6537
6538static int addrconf_sysctl_disable_policy(struct ctl_table *ctl, int write,
6539				   void *buffer, size_t *lenp, loff_t *ppos)
6540{
6541	int *valp = ctl->data;
6542	int val = *valp;
6543	loff_t pos = *ppos;
6544	struct ctl_table lctl;
6545	int ret;
6546
6547	lctl = *ctl;
6548	lctl.data = &val;
6549	ret = proc_dointvec(&lctl, write, buffer, lenp, ppos);
6550
6551	if (write && (*valp != val))
6552		ret = addrconf_disable_policy(ctl, valp, val);
6553
6554	if (ret)
6555		*ppos = pos;
6556
6557	return ret;
6558}
6559
6560static int minus_one = -1;
6561static const int two_five_five = 255;
6562
6563static const struct ctl_table addrconf_sysctl[] = {
6564	{
6565		.procname	= "forwarding",
6566		.data		= &ipv6_devconf.forwarding,
6567		.maxlen		= sizeof(int),
6568		.mode		= 0644,
6569		.proc_handler	= addrconf_sysctl_forward,
6570	},
6571	{
6572		.procname	= "hop_limit",
6573		.data		= &ipv6_devconf.hop_limit,
6574		.maxlen		= sizeof(int),
6575		.mode		= 0644,
6576		.proc_handler	= proc_dointvec_minmax,
6577		.extra1		= (void *)SYSCTL_ONE,
6578		.extra2		= (void *)&two_five_five,
6579	},
6580	{
6581		.procname	= "mtu",
6582		.data		= &ipv6_devconf.mtu6,
6583		.maxlen		= sizeof(int),
6584		.mode		= 0644,
6585		.proc_handler	= addrconf_sysctl_mtu,
6586	},
6587	{
6588		.procname	= "accept_ra",
6589		.data		= &ipv6_devconf.accept_ra,
6590		.maxlen		= sizeof(int),
6591		.mode		= 0644,
6592		.proc_handler	= proc_dointvec,
6593	},
6594	{
6595		.procname	= "accept_redirects",
6596		.data		= &ipv6_devconf.accept_redirects,
6597		.maxlen		= sizeof(int),
6598		.mode		= 0644,
6599		.proc_handler	= proc_dointvec,
6600	},
6601	{
6602		.procname	= "autoconf",
6603		.data		= &ipv6_devconf.autoconf,
6604		.maxlen		= sizeof(int),
6605		.mode		= 0644,
6606		.proc_handler	= proc_dointvec,
6607	},
6608	{
6609		.procname	= "dad_transmits",
6610		.data		= &ipv6_devconf.dad_transmits,
6611		.maxlen		= sizeof(int),
6612		.mode		= 0644,
6613		.proc_handler	= proc_dointvec,
6614	},
6615	{
6616		.procname	= "router_solicitations",
6617		.data		= &ipv6_devconf.rtr_solicits,
6618		.maxlen		= sizeof(int),
6619		.mode		= 0644,
6620		.proc_handler	= proc_dointvec_minmax,
6621		.extra1		= &minus_one,
6622	},
6623	{
6624		.procname	= "router_solicitation_interval",
6625		.data		= &ipv6_devconf.rtr_solicit_interval,
6626		.maxlen		= sizeof(int),
6627		.mode		= 0644,
6628		.proc_handler	= proc_dointvec_jiffies,
6629	},
6630	{
6631		.procname	= "router_solicitation_max_interval",
6632		.data		= &ipv6_devconf.rtr_solicit_max_interval,
6633		.maxlen		= sizeof(int),
6634		.mode		= 0644,
6635		.proc_handler	= proc_dointvec_jiffies,
6636	},
6637	{
6638		.procname	= "router_solicitation_delay",
6639		.data		= &ipv6_devconf.rtr_solicit_delay,
6640		.maxlen		= sizeof(int),
6641		.mode		= 0644,
6642		.proc_handler	= proc_dointvec_jiffies,
6643	},
6644	{
6645		.procname	= "force_mld_version",
6646		.data		= &ipv6_devconf.force_mld_version,
6647		.maxlen		= sizeof(int),
6648		.mode		= 0644,
6649		.proc_handler	= proc_dointvec,
6650	},
6651	{
6652		.procname	= "mldv1_unsolicited_report_interval",
6653		.data		=
6654			&ipv6_devconf.mldv1_unsolicited_report_interval,
6655		.maxlen		= sizeof(int),
6656		.mode		= 0644,
6657		.proc_handler	= proc_dointvec_ms_jiffies,
6658	},
6659	{
6660		.procname	= "mldv2_unsolicited_report_interval",
6661		.data		=
6662			&ipv6_devconf.mldv2_unsolicited_report_interval,
6663		.maxlen		= sizeof(int),
6664		.mode		= 0644,
6665		.proc_handler	= proc_dointvec_ms_jiffies,
6666	},
6667	{
6668		.procname	= "use_tempaddr",
6669		.data		= &ipv6_devconf.use_tempaddr,
6670		.maxlen		= sizeof(int),
6671		.mode		= 0644,
6672		.proc_handler	= proc_dointvec,
6673	},
6674	{
6675		.procname	= "temp_valid_lft",
6676		.data		= &ipv6_devconf.temp_valid_lft,
6677		.maxlen		= sizeof(int),
6678		.mode		= 0644,
6679		.proc_handler	= proc_dointvec,
6680	},
6681	{
6682		.procname	= "temp_prefered_lft",
6683		.data		= &ipv6_devconf.temp_prefered_lft,
6684		.maxlen		= sizeof(int),
6685		.mode		= 0644,
6686		.proc_handler	= proc_dointvec,
6687	},
6688	{
6689		.procname	= "regen_max_retry",
6690		.data		= &ipv6_devconf.regen_max_retry,
6691		.maxlen		= sizeof(int),
6692		.mode		= 0644,
6693		.proc_handler	= proc_dointvec,
6694	},
6695	{
6696		.procname	= "max_desync_factor",
6697		.data		= &ipv6_devconf.max_desync_factor,
6698		.maxlen		= sizeof(int),
6699		.mode		= 0644,
6700		.proc_handler	= proc_dointvec,
6701	},
6702	{
6703		.procname	= "max_addresses",
6704		.data		= &ipv6_devconf.max_addresses,
6705		.maxlen		= sizeof(int),
6706		.mode		= 0644,
6707		.proc_handler	= proc_dointvec,
6708	},
6709	{
6710		.procname	= "accept_ra_defrtr",
6711		.data		= &ipv6_devconf.accept_ra_defrtr,
6712		.maxlen		= sizeof(int),
6713		.mode		= 0644,
6714		.proc_handler	= proc_dointvec,
6715	},
6716	{
6717		.procname	= "accept_ra_min_hop_limit",
6718		.data		= &ipv6_devconf.accept_ra_min_hop_limit,
6719		.maxlen		= sizeof(int),
6720		.mode		= 0644,
6721		.proc_handler	= proc_dointvec,
6722	},
6723	{
6724		.procname	= "accept_ra_min_lft",
6725		.data		= &ipv6_devconf.accept_ra_min_lft,
6726		.maxlen		= sizeof(int),
6727		.mode		= 0644,
6728		.proc_handler	= proc_dointvec,
6729	},
6730	{
6731		.procname	= "accept_ra_pinfo",
6732		.data		= &ipv6_devconf.accept_ra_pinfo,
6733		.maxlen		= sizeof(int),
6734		.mode		= 0644,
6735		.proc_handler	= proc_dointvec,
6736	},
6737#ifdef CONFIG_IPV6_ROUTER_PREF
6738	{
6739		.procname	= "accept_ra_rtr_pref",
6740		.data		= &ipv6_devconf.accept_ra_rtr_pref,
6741		.maxlen		= sizeof(int),
6742		.mode		= 0644,
6743		.proc_handler	= proc_dointvec,
6744	},
6745	{
6746		.procname	= "router_probe_interval",
6747		.data		= &ipv6_devconf.rtr_probe_interval,
6748		.maxlen		= sizeof(int),
6749		.mode		= 0644,
6750		.proc_handler	= proc_dointvec_jiffies,
6751	},
6752#ifdef CONFIG_IPV6_ROUTE_INFO
6753	{
6754		.procname	= "accept_ra_rt_info_min_plen",
6755		.data		= &ipv6_devconf.accept_ra_rt_info_min_plen,
6756		.maxlen		= sizeof(int),
6757		.mode		= 0644,
6758		.proc_handler	= proc_dointvec,
6759	},
6760	{
6761		.procname	= "accept_ra_rt_info_max_plen",
6762		.data		= &ipv6_devconf.accept_ra_rt_info_max_plen,
6763		.maxlen		= sizeof(int),
6764		.mode		= 0644,
6765		.proc_handler	= proc_dointvec,
6766	},
6767#endif
6768#endif
6769	{
6770		.procname	= "proxy_ndp",
6771		.data		= &ipv6_devconf.proxy_ndp,
6772		.maxlen		= sizeof(int),
6773		.mode		= 0644,
6774		.proc_handler	= addrconf_sysctl_proxy_ndp,
6775	},
6776	{
6777		.procname	= "accept_source_route",
6778		.data		= &ipv6_devconf.accept_source_route,
6779		.maxlen		= sizeof(int),
6780		.mode		= 0644,
6781		.proc_handler	= proc_dointvec,
6782	},
6783#ifdef CONFIG_IPV6_OPTIMISTIC_DAD
6784	{
6785		.procname	= "optimistic_dad",
6786		.data		= &ipv6_devconf.optimistic_dad,
6787		.maxlen		= sizeof(int),
6788		.mode		= 0644,
6789		.proc_handler   = proc_dointvec,
6790	},
6791	{
6792		.procname	= "use_optimistic",
6793		.data		= &ipv6_devconf.use_optimistic,
6794		.maxlen		= sizeof(int),
6795		.mode		= 0644,
6796		.proc_handler	= proc_dointvec,
6797	},
6798#endif
6799#ifdef CONFIG_IPV6_MROUTE
6800	{
6801		.procname	= "mc_forwarding",
6802		.data		= &ipv6_devconf.mc_forwarding,
6803		.maxlen		= sizeof(int),
6804		.mode		= 0444,
6805		.proc_handler	= proc_dointvec,
6806	},
6807#endif
6808	{
6809		.procname	= "disable_ipv6",
6810		.data		= &ipv6_devconf.disable_ipv6,
6811		.maxlen		= sizeof(int),
6812		.mode		= 0644,
6813		.proc_handler	= addrconf_sysctl_disable,
6814	},
6815	{
6816		.procname	= "accept_dad",
6817		.data		= &ipv6_devconf.accept_dad,
6818		.maxlen		= sizeof(int),
6819		.mode		= 0644,
6820		.proc_handler	= proc_dointvec,
6821	},
6822	{
6823		.procname	= "force_tllao",
6824		.data		= &ipv6_devconf.force_tllao,
6825		.maxlen		= sizeof(int),
6826		.mode		= 0644,
6827		.proc_handler	= proc_dointvec
6828	},
6829	{
6830		.procname	= "ndisc_notify",
6831		.data		= &ipv6_devconf.ndisc_notify,
6832		.maxlen		= sizeof(int),
6833		.mode		= 0644,
6834		.proc_handler	= proc_dointvec
6835	},
6836	{
6837		.procname	= "suppress_frag_ndisc",
6838		.data		= &ipv6_devconf.suppress_frag_ndisc,
6839		.maxlen		= sizeof(int),
6840		.mode		= 0644,
6841		.proc_handler	= proc_dointvec
6842	},
6843	{
6844		.procname	= "accept_ra_from_local",
6845		.data		= &ipv6_devconf.accept_ra_from_local,
6846		.maxlen		= sizeof(int),
6847		.mode		= 0644,
6848		.proc_handler	= proc_dointvec,
6849	},
6850	{
6851		.procname	= "accept_ra_mtu",
6852		.data		= &ipv6_devconf.accept_ra_mtu,
6853		.maxlen		= sizeof(int),
6854		.mode		= 0644,
6855		.proc_handler	= proc_dointvec,
6856	},
6857	{
6858		.procname	= "stable_secret",
6859		.data		= &ipv6_devconf.stable_secret,
6860		.maxlen		= IPV6_MAX_STRLEN,
6861		.mode		= 0600,
6862		.proc_handler	= addrconf_sysctl_stable_secret,
6863	},
6864	{
6865		.procname	= "use_oif_addrs_only",
6866		.data		= &ipv6_devconf.use_oif_addrs_only,
6867		.maxlen		= sizeof(int),
6868		.mode		= 0644,
6869		.proc_handler	= proc_dointvec,
6870	},
6871	{
6872		.procname	= "ignore_routes_with_linkdown",
6873		.data		= &ipv6_devconf.ignore_routes_with_linkdown,
6874		.maxlen		= sizeof(int),
6875		.mode		= 0644,
6876		.proc_handler	= addrconf_sysctl_ignore_routes_with_linkdown,
6877	},
6878	{
6879		.procname	= "drop_unicast_in_l2_multicast",
6880		.data		= &ipv6_devconf.drop_unicast_in_l2_multicast,
6881		.maxlen		= sizeof(int),
6882		.mode		= 0644,
6883		.proc_handler	= proc_dointvec,
6884	},
6885	{
6886		.procname	= "drop_unsolicited_na",
6887		.data		= &ipv6_devconf.drop_unsolicited_na,
6888		.maxlen		= sizeof(int),
6889		.mode		= 0644,
6890		.proc_handler	= proc_dointvec,
6891	},
6892	{
6893		.procname	= "keep_addr_on_down",
6894		.data		= &ipv6_devconf.keep_addr_on_down,
6895		.maxlen		= sizeof(int),
6896		.mode		= 0644,
6897		.proc_handler	= proc_dointvec,
6898
6899	},
6900	{
6901		.procname	= "seg6_enabled",
6902		.data		= &ipv6_devconf.seg6_enabled,
6903		.maxlen		= sizeof(int),
6904		.mode		= 0644,
6905		.proc_handler	= proc_dointvec,
6906	},
6907#ifdef CONFIG_IPV6_SEG6_HMAC
6908	{
6909		.procname	= "seg6_require_hmac",
6910		.data		= &ipv6_devconf.seg6_require_hmac,
6911		.maxlen		= sizeof(int),
6912		.mode		= 0644,
6913		.proc_handler	= proc_dointvec,
6914	},
6915#endif
6916	{
6917		.procname       = "enhanced_dad",
6918		.data           = &ipv6_devconf.enhanced_dad,
6919		.maxlen         = sizeof(int),
6920		.mode           = 0644,
6921		.proc_handler   = proc_dointvec,
6922	},
6923	{
6924		.procname		= "addr_gen_mode",
6925		.data			= &ipv6_devconf.addr_gen_mode,
6926		.maxlen			= sizeof(int),
6927		.mode			= 0644,
6928		.proc_handler	= addrconf_sysctl_addr_gen_mode,
6929	},
6930	{
6931		.procname       = "disable_policy",
6932		.data           = &ipv6_devconf.disable_policy,
6933		.maxlen         = sizeof(int),
6934		.mode           = 0644,
6935		.proc_handler   = addrconf_sysctl_disable_policy,
6936	},
6937	{
6938		.procname	= "ndisc_tclass",
6939		.data		= &ipv6_devconf.ndisc_tclass,
6940		.maxlen		= sizeof(int),
6941		.mode		= 0644,
6942		.proc_handler	= proc_dointvec_minmax,
6943		.extra1		= (void *)SYSCTL_ZERO,
6944		.extra2		= (void *)&two_five_five,
6945	},
6946	{
6947		.procname	= "rpl_seg_enabled",
6948		.data		= &ipv6_devconf.rpl_seg_enabled,
6949		.maxlen		= sizeof(int),
6950		.mode		= 0644,
6951		.proc_handler	= proc_dointvec,
6952	},
6953	{
6954		/* sentinel */
6955	}
6956};
6957
6958static int __addrconf_sysctl_register(struct net *net, char *dev_name,
6959		struct inet6_dev *idev, struct ipv6_devconf *p)
6960{
6961	int i, ifindex;
6962	struct ctl_table *table;
6963	char path[sizeof("net/ipv6/conf/") + IFNAMSIZ];
6964
6965	table = kmemdup(addrconf_sysctl, sizeof(addrconf_sysctl), GFP_KERNEL);
6966	if (!table)
6967		goto out;
6968
6969	for (i = 0; table[i].data; i++) {
6970		table[i].data += (char *)p - (char *)&ipv6_devconf;
6971		/* If one of these is already set, then it is not safe to
6972		 * overwrite either of them: this makes proc_dointvec_minmax
6973		 * usable.
6974		 */
6975		if (!table[i].extra1 && !table[i].extra2) {
6976			table[i].extra1 = idev; /* embedded; no ref */
6977			table[i].extra2 = net;
6978		}
6979	}
6980
6981	snprintf(path, sizeof(path), "net/ipv6/conf/%s", dev_name);
6982
6983	p->sysctl_header = register_net_sysctl(net, path, table);
6984	if (!p->sysctl_header)
6985		goto free;
6986
6987	if (!strcmp(dev_name, "all"))
6988		ifindex = NETCONFA_IFINDEX_ALL;
6989	else if (!strcmp(dev_name, "default"))
6990		ifindex = NETCONFA_IFINDEX_DEFAULT;
6991	else
6992		ifindex = idev->dev->ifindex;
6993	inet6_netconf_notify_devconf(net, RTM_NEWNETCONF, NETCONFA_ALL,
6994				     ifindex, p);
6995	return 0;
6996
6997free:
6998	kfree(table);
6999out:
7000	return -ENOBUFS;
7001}
7002
7003static void __addrconf_sysctl_unregister(struct net *net,
7004					 struct ipv6_devconf *p, int ifindex)
7005{
7006	struct ctl_table *table;
7007
7008	if (!p->sysctl_header)
7009		return;
7010
7011	table = p->sysctl_header->ctl_table_arg;
7012	unregister_net_sysctl_table(p->sysctl_header);
7013	p->sysctl_header = NULL;
7014	kfree(table);
7015
7016	inet6_netconf_notify_devconf(net, RTM_DELNETCONF, 0, ifindex, NULL);
7017}
7018
7019static int addrconf_sysctl_register(struct inet6_dev *idev)
7020{
7021	int err;
7022
7023	if (!sysctl_dev_name_is_allowed(idev->dev->name))
7024		return -EINVAL;
7025
7026	err = neigh_sysctl_register(idev->dev, idev->nd_parms,
7027				    &ndisc_ifinfo_sysctl_change);
7028	if (err)
7029		return err;
7030	err = __addrconf_sysctl_register(dev_net(idev->dev), idev->dev->name,
7031					 idev, &idev->cnf);
7032	if (err)
7033		neigh_sysctl_unregister(idev->nd_parms);
7034
7035	return err;
7036}
7037
7038static void addrconf_sysctl_unregister(struct inet6_dev *idev)
7039{
7040	__addrconf_sysctl_unregister(dev_net(idev->dev), &idev->cnf,
7041				     idev->dev->ifindex);
7042	neigh_sysctl_unregister(idev->nd_parms);
7043}
7044
7045
7046#endif
7047
7048static int __net_init addrconf_init_net(struct net *net)
7049{
7050	int err = -ENOMEM;
7051	struct ipv6_devconf *all, *dflt;
7052
7053	all = kmemdup(&ipv6_devconf, sizeof(ipv6_devconf), GFP_KERNEL);
7054	if (!all)
7055		goto err_alloc_all;
7056
7057	dflt = kmemdup(&ipv6_devconf_dflt, sizeof(ipv6_devconf_dflt), GFP_KERNEL);
7058	if (!dflt)
7059		goto err_alloc_dflt;
7060
7061	if (!net_eq(net, &init_net)) {
7062		switch (net_inherit_devconf()) {
7063		case 1:  /* copy from init_net */
7064			memcpy(all, init_net.ipv6.devconf_all,
7065			       sizeof(ipv6_devconf));
7066			memcpy(dflt, init_net.ipv6.devconf_dflt,
7067			       sizeof(ipv6_devconf_dflt));
7068			break;
7069		case 3: /* copy from the current netns */
7070			memcpy(all, current->nsproxy->net_ns->ipv6.devconf_all,
7071			       sizeof(ipv6_devconf));
7072			memcpy(dflt,
7073			       current->nsproxy->net_ns->ipv6.devconf_dflt,
7074			       sizeof(ipv6_devconf_dflt));
7075			break;
7076		case 0:
7077		case 2:
7078			/* use compiled values */
7079			break;
7080		}
7081	}
7082
7083	/* these will be inherited by all namespaces */
7084	dflt->autoconf = ipv6_defaults.autoconf;
7085	dflt->disable_ipv6 = ipv6_defaults.disable_ipv6;
7086
7087	dflt->stable_secret.initialized = false;
7088	all->stable_secret.initialized = false;
7089
7090	net->ipv6.devconf_all = all;
7091	net->ipv6.devconf_dflt = dflt;
7092
7093#ifdef CONFIG_SYSCTL
7094	err = __addrconf_sysctl_register(net, "all", NULL, all);
7095	if (err < 0)
7096		goto err_reg_all;
7097
7098	err = __addrconf_sysctl_register(net, "default", NULL, dflt);
7099	if (err < 0)
7100		goto err_reg_dflt;
7101#endif
7102	return 0;
7103
7104#ifdef CONFIG_SYSCTL
7105err_reg_dflt:
7106	__addrconf_sysctl_unregister(net, all, NETCONFA_IFINDEX_ALL);
7107err_reg_all:
7108	kfree(dflt);
7109#endif
7110err_alloc_dflt:
7111	kfree(all);
7112err_alloc_all:
7113	return err;
7114}
7115
7116static void __net_exit addrconf_exit_net(struct net *net)
7117{
7118#ifdef CONFIG_SYSCTL
7119	__addrconf_sysctl_unregister(net, net->ipv6.devconf_dflt,
7120				     NETCONFA_IFINDEX_DEFAULT);
7121	__addrconf_sysctl_unregister(net, net->ipv6.devconf_all,
7122				     NETCONFA_IFINDEX_ALL);
7123#endif
7124	kfree(net->ipv6.devconf_dflt);
7125	kfree(net->ipv6.devconf_all);
7126}
7127
7128static struct pernet_operations addrconf_ops = {
7129	.init = addrconf_init_net,
7130	.exit = addrconf_exit_net,
7131};
7132
7133static struct rtnl_af_ops inet6_ops __read_mostly = {
7134	.family		  = AF_INET6,
7135	.fill_link_af	  = inet6_fill_link_af,
7136	.get_link_af_size = inet6_get_link_af_size,
7137	.validate_link_af = inet6_validate_link_af,
7138	.set_link_af	  = inet6_set_link_af,
7139};
7140
7141/*
7142 *	Init / cleanup code
7143 */
7144
7145int __init addrconf_init(void)
7146{
7147	struct inet6_dev *idev;
7148	int i, err;
7149
7150	err = ipv6_addr_label_init();
7151	if (err < 0) {
7152		pr_crit("%s: cannot initialize default policy table: %d\n",
7153			__func__, err);
7154		goto out;
7155	}
7156
7157	err = register_pernet_subsys(&addrconf_ops);
7158	if (err < 0)
7159		goto out_addrlabel;
7160
7161	addrconf_wq = create_workqueue("ipv6_addrconf");
7162	if (!addrconf_wq) {
7163		err = -ENOMEM;
7164		goto out_nowq;
7165	}
7166
7167	/* The addrconf netdev notifier requires that loopback_dev
7168	 * has it's ipv6 private information allocated and setup
7169	 * before it can bring up and give link-local addresses
7170	 * to other devices which are up.
7171	 *
7172	 * Unfortunately, loopback_dev is not necessarily the first
7173	 * entry in the global dev_base list of net devices.  In fact,
7174	 * it is likely to be the very last entry on that list.
7175	 * So this causes the notifier registry below to try and
7176	 * give link-local addresses to all devices besides loopback_dev
7177	 * first, then loopback_dev, which cases all the non-loopback_dev
7178	 * devices to fail to get a link-local address.
7179	 *
7180	 * So, as a temporary fix, allocate the ipv6 structure for
7181	 * loopback_dev first by hand.
7182	 * Longer term, all of the dependencies ipv6 has upon the loopback
7183	 * device and it being up should be removed.
7184	 */
7185	rtnl_lock();
7186	idev = ipv6_add_dev(init_net.loopback_dev);
7187	rtnl_unlock();
7188	if (IS_ERR(idev)) {
7189		err = PTR_ERR(idev);
7190		goto errlo;
7191	}
7192
7193	ip6_route_init_special_entries();
7194
7195	for (i = 0; i < IN6_ADDR_HSIZE; i++)
7196		INIT_HLIST_HEAD(&inet6_addr_lst[i]);
7197
7198	register_netdevice_notifier(&ipv6_dev_notf);
7199
7200	addrconf_verify();
7201
7202	rtnl_af_register(&inet6_ops);
7203
7204	err = rtnl_register_module(THIS_MODULE, PF_INET6, RTM_GETLINK,
7205				   NULL, inet6_dump_ifinfo, 0);
7206	if (err < 0)
7207		goto errout;
7208
7209	err = rtnl_register_module(THIS_MODULE, PF_INET6, RTM_NEWADDR,
7210				   inet6_rtm_newaddr, NULL, 0);
7211	if (err < 0)
7212		goto errout;
7213	err = rtnl_register_module(THIS_MODULE, PF_INET6, RTM_DELADDR,
7214				   inet6_rtm_deladdr, NULL, 0);
7215	if (err < 0)
7216		goto errout;
7217	err = rtnl_register_module(THIS_MODULE, PF_INET6, RTM_GETADDR,
7218				   inet6_rtm_getaddr, inet6_dump_ifaddr,
7219				   RTNL_FLAG_DOIT_UNLOCKED);
7220	if (err < 0)
7221		goto errout;
7222	err = rtnl_register_module(THIS_MODULE, PF_INET6, RTM_GETMULTICAST,
7223				   NULL, inet6_dump_ifmcaddr, 0);
7224	if (err < 0)
7225		goto errout;
7226	err = rtnl_register_module(THIS_MODULE, PF_INET6, RTM_GETANYCAST,
7227				   NULL, inet6_dump_ifacaddr, 0);
7228	if (err < 0)
7229		goto errout;
7230	err = rtnl_register_module(THIS_MODULE, PF_INET6, RTM_GETNETCONF,
7231				   inet6_netconf_get_devconf,
7232				   inet6_netconf_dump_devconf,
7233				   RTNL_FLAG_DOIT_UNLOCKED);
7234	if (err < 0)
7235		goto errout;
7236	err = ipv6_addr_label_rtnl_register();
7237	if (err < 0)
7238		goto errout;
7239
7240	return 0;
7241errout:
7242	rtnl_unregister_all(PF_INET6);
7243	rtnl_af_unregister(&inet6_ops);
7244	unregister_netdevice_notifier(&ipv6_dev_notf);
7245errlo:
7246	destroy_workqueue(addrconf_wq);
7247out_nowq:
7248	unregister_pernet_subsys(&addrconf_ops);
7249out_addrlabel:
7250	ipv6_addr_label_cleanup();
7251out:
7252	return err;
7253}
7254
7255void addrconf_cleanup(void)
7256{
7257	struct net_device *dev;
7258	int i;
7259
7260	unregister_netdevice_notifier(&ipv6_dev_notf);
7261	unregister_pernet_subsys(&addrconf_ops);
7262	ipv6_addr_label_cleanup();
7263
7264	rtnl_af_unregister(&inet6_ops);
7265
7266	rtnl_lock();
7267
7268	/* clean dev list */
7269	for_each_netdev(&init_net, dev) {
7270		if (__in6_dev_get(dev) == NULL)
7271			continue;
7272		addrconf_ifdown(dev, true);
7273	}
7274	addrconf_ifdown(init_net.loopback_dev, true);
7275
7276	/*
7277	 *	Check hash table.
7278	 */
7279	spin_lock_bh(&addrconf_hash_lock);
7280	for (i = 0; i < IN6_ADDR_HSIZE; i++)
7281		WARN_ON(!hlist_empty(&inet6_addr_lst[i]));
7282	spin_unlock_bh(&addrconf_hash_lock);
7283	cancel_delayed_work(&addr_chk_work);
7284	rtnl_unlock();
7285
7286	destroy_workqueue(addrconf_wq);
7287}
7288