1// SPDX-License-Identifier: GPL-2.0-or-later
   2/*
   3 *	Linux INET6 implementation
   4 *	FIB front-end.
   5 *
   6 *	Authors:
   7 *	Pedro Roque		<roque@di.fc.ul.pt>
   8 */
   9
  10/*	Changes:
  11 *
  12 *	YOSHIFUJI Hideaki @USAGI
  13 *		reworked default router selection.
  14 *		- respect outgoing interface
  15 *		- select from (probably) reachable routers (i.e.
  16 *		routers in REACHABLE, STALE, DELAY or PROBE states).
  17 *		- always select the same router if it is (probably)
  18 *		reachable.  otherwise, round-robin the list.
  19 *	Ville Nuorvala
  20 *		Fixed routing subtrees.
  21 */
  22
  23#define pr_fmt(fmt) "IPv6: " fmt
  24
  25#include <linux/capability.h>
  26#include <linux/errno.h>
  27#include <linux/export.h>
  28#include <linux/types.h>
  29#include <linux/times.h>
  30#include <linux/socket.h>
  31#include <linux/sockios.h>
  32#include <linux/net.h>
  33#include <linux/route.h>
  34#include <linux/netdevice.h>
  35#include <linux/in6.h>
  36#include <linux/mroute6.h>
  37#include <linux/init.h>
  38#include <linux/if_arp.h>
  39#include <linux/proc_fs.h>
  40#include <linux/seq_file.h>
  41#include <linux/nsproxy.h>
  42#include <linux/slab.h>
  43#include <linux/jhash.h>
 
  44#include <net/net_namespace.h>
  45#include <net/snmp.h>
  46#include <net/ipv6.h>
  47#include <net/ip6_fib.h>
  48#include <net/ip6_route.h>
  49#include <net/ndisc.h>
  50#include <net/addrconf.h>
  51#include <net/tcp.h>
  52#include <linux/rtnetlink.h>
  53#include <net/dst.h>
  54#include <net/dst_metadata.h>
  55#include <net/xfrm.h>
  56#include <net/netevent.h>
  57#include <net/netlink.h>
  58#include <net/rtnh.h>
  59#include <net/lwtunnel.h>
  60#include <net/ip_tunnels.h>
  61#include <net/l3mdev.h>
  62#include <net/ip.h>
  63#include <linux/uaccess.h>
  64#include <linux/btf_ids.h>
  65
  66#ifdef CONFIG_SYSCTL
  67#include <linux/sysctl.h>
  68#endif
  69
  70static int ip6_rt_type_to_error(u8 fib6_type);
  71
  72#define CREATE_TRACE_POINTS
  73#include <trace/events/fib6.h>
  74EXPORT_TRACEPOINT_SYMBOL_GPL(fib6_table_lookup);
  75#undef CREATE_TRACE_POINTS
  76
  77enum rt6_nud_state {
  78	RT6_NUD_FAIL_HARD = -3,
  79	RT6_NUD_FAIL_PROBE = -2,
  80	RT6_NUD_FAIL_DO_RR = -1,
  81	RT6_NUD_SUCCEED = 1
  82};
  83
  84static struct dst_entry	*ip6_dst_check(struct dst_entry *dst, u32 cookie);
 
  85static unsigned int	 ip6_default_advmss(const struct dst_entry *dst);
  86static unsigned int	 ip6_mtu(const struct dst_entry *dst);
  87static struct dst_entry *ip6_negative_advice(struct dst_entry *);
 
 
  88static void		ip6_dst_destroy(struct dst_entry *);
  89static void		ip6_dst_ifdown(struct dst_entry *,
  90				       struct net_device *dev, int how);
  91static int		 ip6_dst_gc(struct dst_ops *ops);
  92
  93static int		ip6_pkt_discard(struct sk_buff *skb);
  94static int		ip6_pkt_discard_out(struct net *net, struct sock *sk, struct sk_buff *skb);
  95static int		ip6_pkt_prohibit(struct sk_buff *skb);
  96static int		ip6_pkt_prohibit_out(struct net *net, struct sock *sk, struct sk_buff *skb);
  97static void		ip6_link_failure(struct sk_buff *skb);
  98static void		ip6_rt_update_pmtu(struct dst_entry *dst, struct sock *sk,
  99					   struct sk_buff *skb, u32 mtu,
 100					   bool confirm_neigh);
 101static void		rt6_do_redirect(struct dst_entry *dst, struct sock *sk,
 102					struct sk_buff *skb);
 103static int rt6_score_route(const struct fib6_nh *nh, u32 fib6_flags, int oif,
 104			   int strict);
 105static size_t rt6_nlmsg_size(struct fib6_info *f6i);
 106static int rt6_fill_node(struct net *net, struct sk_buff *skb,
 107			 struct fib6_info *rt, struct dst_entry *dst,
 108			 struct in6_addr *dest, struct in6_addr *src,
 109			 int iif, int type, u32 portid, u32 seq,
 110			 unsigned int flags);
 111static struct rt6_info *rt6_find_cached_rt(const struct fib6_result *res,
 112					   const struct in6_addr *daddr,
 113					   const struct in6_addr *saddr);
 114
 115#ifdef CONFIG_IPV6_ROUTE_INFO
 116static struct fib6_info *rt6_add_route_info(struct net *net,
 117					   const struct in6_addr *prefix, int prefixlen,
 118					   const struct in6_addr *gwaddr,
 119					   struct net_device *dev,
 120					   unsigned int pref);
 121static struct fib6_info *rt6_get_route_info(struct net *net,
 122					   const struct in6_addr *prefix, int prefixlen,
 123					   const struct in6_addr *gwaddr,
 124					   struct net_device *dev);
 125#endif
 126
 127struct uncached_list {
 128	spinlock_t		lock;
 129	struct list_head	head;
 
 130};
 131
 132static DEFINE_PER_CPU_ALIGNED(struct uncached_list, rt6_uncached_list);
 133
 134void rt6_uncached_list_add(struct rt6_info *rt)
 135{
 136	struct uncached_list *ul = raw_cpu_ptr(&rt6_uncached_list);
 137
 138	rt->rt6i_uncached_list = ul;
 139
 140	spin_lock_bh(&ul->lock);
 141	list_add_tail(&rt->rt6i_uncached, &ul->head);
 142	spin_unlock_bh(&ul->lock);
 143}
 144
 145void rt6_uncached_list_del(struct rt6_info *rt)
 146{
 147	if (!list_empty(&rt->rt6i_uncached)) {
 148		struct uncached_list *ul = rt->rt6i_uncached_list;
 149		struct net *net = dev_net(rt->dst.dev);
 150
 151		spin_lock_bh(&ul->lock);
 152		list_del(&rt->rt6i_uncached);
 153		atomic_dec(&net->ipv6.rt6_stats->fib_rt_uncache);
 154		spin_unlock_bh(&ul->lock);
 155	}
 156}
 157
 158static void rt6_uncached_list_flush_dev(struct net *net, struct net_device *dev)
 159{
 160	struct net_device *loopback_dev = net->loopback_dev;
 161	int cpu;
 162
 163	if (dev == loopback_dev)
 164		return;
 165
 166	for_each_possible_cpu(cpu) {
 167		struct uncached_list *ul = per_cpu_ptr(&rt6_uncached_list, cpu);
 168		struct rt6_info *rt;
 
 
 
 169
 170		spin_lock_bh(&ul->lock);
 171		list_for_each_entry(rt, &ul->head, rt6i_uncached) {
 172			struct inet6_dev *rt_idev = rt->rt6i_idev;
 173			struct net_device *rt_dev = rt->dst.dev;
 
 174
 175			if (rt_idev->dev == dev) {
 176				rt->rt6i_idev = in6_dev_get(loopback_dev);
 177				in6_dev_put(rt_idev);
 
 178			}
 179
 180			if (rt_dev == dev) {
 181				rt->dst.dev = blackhole_netdev;
 182				dev_hold(rt->dst.dev);
 183				dev_put(rt_dev);
 
 
 184			}
 
 
 
 185		}
 186		spin_unlock_bh(&ul->lock);
 187	}
 188}
 189
 190static inline const void *choose_neigh_daddr(const struct in6_addr *p,
 191					     struct sk_buff *skb,
 192					     const void *daddr)
 193{
 194	if (!ipv6_addr_any(p))
 195		return (const void *) p;
 196	else if (skb)
 197		return &ipv6_hdr(skb)->daddr;
 198	return daddr;
 199}
 200
 201struct neighbour *ip6_neigh_lookup(const struct in6_addr *gw,
 202				   struct net_device *dev,
 203				   struct sk_buff *skb,
 204				   const void *daddr)
 205{
 206	struct neighbour *n;
 207
 208	daddr = choose_neigh_daddr(gw, skb, daddr);
 209	n = __ipv6_neigh_lookup(dev, daddr);
 210	if (n)
 211		return n;
 212
 213	n = neigh_create(&nd_tbl, daddr, dev);
 214	return IS_ERR(n) ? NULL : n;
 215}
 216
 217static struct neighbour *ip6_dst_neigh_lookup(const struct dst_entry *dst,
 218					      struct sk_buff *skb,
 219					      const void *daddr)
 220{
 221	const struct rt6_info *rt = container_of(dst, struct rt6_info, dst);
 222
 223	return ip6_neigh_lookup(rt6_nexthop(rt, &in6addr_any),
 224				dst->dev, skb, daddr);
 225}
 226
 227static void ip6_confirm_neigh(const struct dst_entry *dst, const void *daddr)
 228{
 
 229	struct net_device *dev = dst->dev;
 230	struct rt6_info *rt = (struct rt6_info *)dst;
 231
 232	daddr = choose_neigh_daddr(rt6_nexthop(rt, &in6addr_any), NULL, daddr);
 233	if (!daddr)
 234		return;
 235	if (dev->flags & (IFF_NOARP | IFF_LOOPBACK))
 236		return;
 237	if (ipv6_addr_is_multicast((const struct in6_addr *)daddr))
 238		return;
 239	__ipv6_confirm_neigh(dev, daddr);
 240}
 241
 242static struct dst_ops ip6_dst_ops_template = {
 243	.family			=	AF_INET6,
 244	.gc			=	ip6_dst_gc,
 245	.gc_thresh		=	1024,
 246	.check			=	ip6_dst_check,
 247	.default_advmss		=	ip6_default_advmss,
 248	.mtu			=	ip6_mtu,
 249	.cow_metrics		=	dst_cow_metrics_generic,
 250	.destroy		=	ip6_dst_destroy,
 251	.ifdown			=	ip6_dst_ifdown,
 252	.negative_advice	=	ip6_negative_advice,
 253	.link_failure		=	ip6_link_failure,
 254	.update_pmtu		=	ip6_rt_update_pmtu,
 255	.redirect		=	rt6_do_redirect,
 256	.local_out		=	__ip6_local_out,
 257	.neigh_lookup		=	ip6_dst_neigh_lookup,
 258	.confirm_neigh		=	ip6_confirm_neigh,
 259};
 260
 261static unsigned int ip6_blackhole_mtu(const struct dst_entry *dst)
 262{
 263	unsigned int mtu = dst_metric_raw(dst, RTAX_MTU);
 264
 265	return mtu ? : dst->dev->mtu;
 266}
 267
 268static void ip6_rt_blackhole_update_pmtu(struct dst_entry *dst, struct sock *sk,
 269					 struct sk_buff *skb, u32 mtu,
 270					 bool confirm_neigh)
 271{
 272}
 273
 274static void ip6_rt_blackhole_redirect(struct dst_entry *dst, struct sock *sk,
 275				      struct sk_buff *skb)
 276{
 277}
 278
 279static struct dst_ops ip6_dst_blackhole_ops = {
 280	.family			=	AF_INET6,
 281	.destroy		=	ip6_dst_destroy,
 282	.check			=	ip6_dst_check,
 283	.mtu			=	ip6_blackhole_mtu,
 284	.default_advmss		=	ip6_default_advmss,
 285	.update_pmtu		=	ip6_rt_blackhole_update_pmtu,
 286	.redirect		=	ip6_rt_blackhole_redirect,
 287	.cow_metrics		=	dst_cow_metrics_generic,
 288	.neigh_lookup		=	ip6_dst_neigh_lookup,
 289};
 290
 291static const u32 ip6_template_metrics[RTAX_MAX] = {
 292	[RTAX_HOPLIMIT - 1] = 0,
 293};
 294
 295static const struct fib6_info fib6_null_entry_template = {
 296	.fib6_flags	= (RTF_REJECT | RTF_NONEXTHOP),
 297	.fib6_protocol  = RTPROT_KERNEL,
 298	.fib6_metric	= ~(u32)0,
 299	.fib6_ref	= REFCOUNT_INIT(1),
 300	.fib6_type	= RTN_UNREACHABLE,
 301	.fib6_metrics	= (struct dst_metrics *)&dst_default_metrics,
 302};
 303
 304static const struct rt6_info ip6_null_entry_template = {
 305	.dst = {
 306		.__refcnt	= ATOMIC_INIT(1),
 307		.__use		= 1,
 308		.obsolete	= DST_OBSOLETE_FORCE_CHK,
 309		.error		= -ENETUNREACH,
 310		.input		= ip6_pkt_discard,
 311		.output		= ip6_pkt_discard_out,
 312	},
 313	.rt6i_flags	= (RTF_REJECT | RTF_NONEXTHOP),
 314};
 315
 316#ifdef CONFIG_IPV6_MULTIPLE_TABLES
 317
 318static const struct rt6_info ip6_prohibit_entry_template = {
 319	.dst = {
 320		.__refcnt	= ATOMIC_INIT(1),
 321		.__use		= 1,
 322		.obsolete	= DST_OBSOLETE_FORCE_CHK,
 323		.error		= -EACCES,
 324		.input		= ip6_pkt_prohibit,
 325		.output		= ip6_pkt_prohibit_out,
 326	},
 327	.rt6i_flags	= (RTF_REJECT | RTF_NONEXTHOP),
 328};
 329
 330static const struct rt6_info ip6_blk_hole_entry_template = {
 331	.dst = {
 332		.__refcnt	= ATOMIC_INIT(1),
 333		.__use		= 1,
 334		.obsolete	= DST_OBSOLETE_FORCE_CHK,
 335		.error		= -EINVAL,
 336		.input		= dst_discard,
 337		.output		= dst_discard_out,
 338	},
 339	.rt6i_flags	= (RTF_REJECT | RTF_NONEXTHOP),
 340};
 341
 342#endif
 343
 344static void rt6_info_init(struct rt6_info *rt)
 345{
 346	struct dst_entry *dst = &rt->dst;
 347
 348	memset(dst + 1, 0, sizeof(*rt) - sizeof(*dst));
 349	INIT_LIST_HEAD(&rt->rt6i_uncached);
 350}
 351
 352/* allocate dst with ip6_dst_ops */
 353struct rt6_info *ip6_dst_alloc(struct net *net, struct net_device *dev,
 354			       int flags)
 355{
 356	struct rt6_info *rt = dst_alloc(&net->ipv6.ip6_dst_ops, dev,
 357					1, DST_OBSOLETE_FORCE_CHK, flags);
 358
 359	if (rt) {
 360		rt6_info_init(rt);
 361		atomic_inc(&net->ipv6.rt6_stats->fib_rt_alloc);
 362	}
 363
 364	return rt;
 365}
 366EXPORT_SYMBOL(ip6_dst_alloc);
 367
 368static void ip6_dst_destroy(struct dst_entry *dst)
 369{
 370	struct rt6_info *rt = (struct rt6_info *)dst;
 371	struct fib6_info *from;
 372	struct inet6_dev *idev;
 373
 374	ip_dst_metrics_put(dst);
 375	rt6_uncached_list_del(rt);
 376
 377	idev = rt->rt6i_idev;
 378	if (idev) {
 379		rt->rt6i_idev = NULL;
 380		in6_dev_put(idev);
 381	}
 382
 383	from = xchg((__force struct fib6_info **)&rt->from, NULL);
 384	fib6_info_release(from);
 385}
 386
 387static void ip6_dst_ifdown(struct dst_entry *dst, struct net_device *dev,
 388			   int how)
 389{
 390	struct rt6_info *rt = (struct rt6_info *)dst;
 391	struct inet6_dev *idev = rt->rt6i_idev;
 392	struct net_device *loopback_dev =
 393		dev_net(dev)->loopback_dev;
 394
 395	if (idev && idev->dev != loopback_dev) {
 396		struct inet6_dev *loopback_idev = in6_dev_get(loopback_dev);
 397		if (loopback_idev) {
 398			rt->rt6i_idev = loopback_idev;
 
 399			in6_dev_put(idev);
 400		}
 401	}
 402}
 403
 404static bool __rt6_check_expired(const struct rt6_info *rt)
 405{
 406	if (rt->rt6i_flags & RTF_EXPIRES)
 407		return time_after(jiffies, rt->dst.expires);
 408	else
 409		return false;
 410}
 411
 412static bool rt6_check_expired(const struct rt6_info *rt)
 413{
 414	struct fib6_info *from;
 415
 416	from = rcu_dereference(rt->from);
 417
 418	if (rt->rt6i_flags & RTF_EXPIRES) {
 419		if (time_after(jiffies, rt->dst.expires))
 420			return true;
 421	} else if (from) {
 422		return rt->dst.obsolete != DST_OBSOLETE_FORCE_CHK ||
 423			fib6_check_expired(from);
 424	}
 425	return false;
 426}
 427
 428void fib6_select_path(const struct net *net, struct fib6_result *res,
 429		      struct flowi6 *fl6, int oif, bool have_oif_match,
 430		      const struct sk_buff *skb, int strict)
 431{
 432	struct fib6_info *sibling, *next_sibling;
 433	struct fib6_info *match = res->f6i;
 434
 435	if (!match->nh && (!match->fib6_nsiblings || have_oif_match))
 436		goto out;
 437
 438	if (match->nh && have_oif_match && res->nh)
 439		return;
 440
 
 
 
 441	/* We might have already computed the hash for ICMPv6 errors. In such
 442	 * case it will always be non-zero. Otherwise now is the time to do it.
 443	 */
 444	if (!fl6->mp_hash &&
 445	    (!match->nh || nexthop_is_multipath(match->nh)))
 446		fl6->mp_hash = rt6_multipath_hash(net, fl6, skb, NULL);
 447
 448	if (unlikely(match->nh)) {
 449		nexthop_path_fib6_result(res, fl6->mp_hash);
 450		return;
 451	}
 452
 453	if (fl6->mp_hash <= atomic_read(&match->fib6_nh->fib_nh_upper_bound))
 454		goto out;
 455
 456	list_for_each_entry_safe(sibling, next_sibling, &match->fib6_siblings,
 457				 fib6_siblings) {
 458		const struct fib6_nh *nh = sibling->fib6_nh;
 459		int nh_upper_bound;
 460
 461		nh_upper_bound = atomic_read(&nh->fib_nh_upper_bound);
 462		if (fl6->mp_hash > nh_upper_bound)
 463			continue;
 464		if (rt6_score_route(nh, sibling->fib6_flags, oif, strict) < 0)
 465			break;
 466		match = sibling;
 467		break;
 468	}
 469
 470out:
 471	res->f6i = match;
 472	res->nh = match->fib6_nh;
 473}
 474
 475/*
 476 *	Route lookup. rcu_read_lock() should be held.
 477 */
 478
 479static bool __rt6_device_match(struct net *net, const struct fib6_nh *nh,
 480			       const struct in6_addr *saddr, int oif, int flags)
 481{
 482	const struct net_device *dev;
 483
 484	if (nh->fib_nh_flags & RTNH_F_DEAD)
 485		return false;
 486
 487	dev = nh->fib_nh_dev;
 488	if (oif) {
 489		if (dev->ifindex == oif)
 490			return true;
 491	} else {
 492		if (ipv6_chk_addr(net, saddr, dev,
 493				  flags & RT6_LOOKUP_F_IFACE))
 494			return true;
 495	}
 496
 497	return false;
 498}
 499
 500struct fib6_nh_dm_arg {
 501	struct net		*net;
 502	const struct in6_addr	*saddr;
 503	int			oif;
 504	int			flags;
 505	struct fib6_nh		*nh;
 506};
 507
 508static int __rt6_nh_dev_match(struct fib6_nh *nh, void *_arg)
 509{
 510	struct fib6_nh_dm_arg *arg = _arg;
 511
 512	arg->nh = nh;
 513	return __rt6_device_match(arg->net, nh, arg->saddr, arg->oif,
 514				  arg->flags);
 515}
 516
 517/* returns fib6_nh from nexthop or NULL */
 518static struct fib6_nh *rt6_nh_dev_match(struct net *net, struct nexthop *nh,
 519					struct fib6_result *res,
 520					const struct in6_addr *saddr,
 521					int oif, int flags)
 522{
 523	struct fib6_nh_dm_arg arg = {
 524		.net   = net,
 525		.saddr = saddr,
 526		.oif   = oif,
 527		.flags = flags,
 528	};
 529
 530	if (nexthop_is_blackhole(nh))
 531		return NULL;
 532
 533	if (nexthop_for_each_fib6_nh(nh, __rt6_nh_dev_match, &arg))
 534		return arg.nh;
 535
 536	return NULL;
 537}
 538
 539static void rt6_device_match(struct net *net, struct fib6_result *res,
 540			     const struct in6_addr *saddr, int oif, int flags)
 541{
 542	struct fib6_info *f6i = res->f6i;
 543	struct fib6_info *spf6i;
 544	struct fib6_nh *nh;
 545
 546	if (!oif && ipv6_addr_any(saddr)) {
 547		if (unlikely(f6i->nh)) {
 548			nh = nexthop_fib6_nh(f6i->nh);
 549			if (nexthop_is_blackhole(f6i->nh))
 550				goto out_blackhole;
 551		} else {
 552			nh = f6i->fib6_nh;
 553		}
 554		if (!(nh->fib_nh_flags & RTNH_F_DEAD))
 555			goto out;
 556	}
 557
 558	for (spf6i = f6i; spf6i; spf6i = rcu_dereference(spf6i->fib6_next)) {
 559		bool matched = false;
 560
 561		if (unlikely(spf6i->nh)) {
 562			nh = rt6_nh_dev_match(net, spf6i->nh, res, saddr,
 563					      oif, flags);
 564			if (nh)
 565				matched = true;
 566		} else {
 567			nh = spf6i->fib6_nh;
 568			if (__rt6_device_match(net, nh, saddr, oif, flags))
 569				matched = true;
 570		}
 571		if (matched) {
 572			res->f6i = spf6i;
 573			goto out;
 574		}
 575	}
 576
 577	if (oif && flags & RT6_LOOKUP_F_IFACE) {
 578		res->f6i = net->ipv6.fib6_null_entry;
 579		nh = res->f6i->fib6_nh;
 580		goto out;
 581	}
 582
 583	if (unlikely(f6i->nh)) {
 584		nh = nexthop_fib6_nh(f6i->nh);
 585		if (nexthop_is_blackhole(f6i->nh))
 586			goto out_blackhole;
 587	} else {
 588		nh = f6i->fib6_nh;
 589	}
 590
 591	if (nh->fib_nh_flags & RTNH_F_DEAD) {
 592		res->f6i = net->ipv6.fib6_null_entry;
 593		nh = res->f6i->fib6_nh;
 594	}
 595out:
 596	res->nh = nh;
 597	res->fib6_type = res->f6i->fib6_type;
 598	res->fib6_flags = res->f6i->fib6_flags;
 599	return;
 600
 601out_blackhole:
 602	res->fib6_flags |= RTF_REJECT;
 603	res->fib6_type = RTN_BLACKHOLE;
 604	res->nh = nh;
 605}
 606
 607#ifdef CONFIG_IPV6_ROUTER_PREF
 608struct __rt6_probe_work {
 609	struct work_struct work;
 610	struct in6_addr target;
 611	struct net_device *dev;
 
 612};
 613
 614static void rt6_probe_deferred(struct work_struct *w)
 615{
 616	struct in6_addr mcaddr;
 617	struct __rt6_probe_work *work =
 618		container_of(w, struct __rt6_probe_work, work);
 619
 620	addrconf_addr_solict_mult(&work->target, &mcaddr);
 621	ndisc_send_ns(work->dev, &work->target, &mcaddr, NULL, 0);
 622	dev_put(work->dev);
 623	kfree(work);
 624}
 625
 626static void rt6_probe(struct fib6_nh *fib6_nh)
 627{
 628	struct __rt6_probe_work *work = NULL;
 629	const struct in6_addr *nh_gw;
 630	unsigned long last_probe;
 631	struct neighbour *neigh;
 632	struct net_device *dev;
 633	struct inet6_dev *idev;
 634
 635	/*
 636	 * Okay, this does not seem to be appropriate
 637	 * for now, however, we need to check if it
 638	 * is really so; aka Router Reachability Probing.
 639	 *
 640	 * Router Reachability Probe MUST be rate-limited
 641	 * to no more than one per minute.
 642	 */
 643	if (!fib6_nh->fib_nh_gw_family)
 644		return;
 645
 646	nh_gw = &fib6_nh->fib_nh_gw6;
 647	dev = fib6_nh->fib_nh_dev;
 648	rcu_read_lock_bh();
 649	last_probe = READ_ONCE(fib6_nh->last_probe);
 650	idev = __in6_dev_get(dev);
 651	neigh = __ipv6_neigh_lookup_noref(dev, nh_gw);
 652	if (neigh) {
 653		if (neigh->nud_state & NUD_VALID)
 654			goto out;
 655
 656		write_lock(&neigh->lock);
 657		if (!(neigh->nud_state & NUD_VALID) &&
 658		    time_after(jiffies,
 659			       neigh->updated + idev->cnf.rtr_probe_interval)) {
 
 660			work = kmalloc(sizeof(*work), GFP_ATOMIC);
 661			if (work)
 662				__neigh_set_probe_once(neigh);
 663		}
 664		write_unlock(&neigh->lock);
 665	} else if (time_after(jiffies, last_probe +
 666				       idev->cnf.rtr_probe_interval)) {
 667		work = kmalloc(sizeof(*work), GFP_ATOMIC);
 668	}
 669
 670	if (!work || cmpxchg(&fib6_nh->last_probe,
 671			     last_probe, jiffies) != last_probe) {
 672		kfree(work);
 673	} else {
 674		INIT_WORK(&work->work, rt6_probe_deferred);
 675		work->target = *nh_gw;
 676		dev_hold(dev);
 677		work->dev = dev;
 678		schedule_work(&work->work);
 679	}
 680
 681out:
 682	rcu_read_unlock_bh();
 683}
 684#else
 685static inline void rt6_probe(struct fib6_nh *fib6_nh)
 686{
 687}
 688#endif
 689
 690/*
 691 * Default Router Selection (RFC 2461 6.3.6)
 692 */
 693static enum rt6_nud_state rt6_check_neigh(const struct fib6_nh *fib6_nh)
 694{
 695	enum rt6_nud_state ret = RT6_NUD_FAIL_HARD;
 696	struct neighbour *neigh;
 697
 698	rcu_read_lock_bh();
 699	neigh = __ipv6_neigh_lookup_noref(fib6_nh->fib_nh_dev,
 700					  &fib6_nh->fib_nh_gw6);
 701	if (neigh) {
 702		read_lock(&neigh->lock);
 703		if (neigh->nud_state & NUD_VALID)
 
 704			ret = RT6_NUD_SUCCEED;
 705#ifdef CONFIG_IPV6_ROUTER_PREF
 706		else if (!(neigh->nud_state & NUD_FAILED))
 707			ret = RT6_NUD_SUCCEED;
 708		else
 709			ret = RT6_NUD_FAIL_PROBE;
 710#endif
 711		read_unlock(&neigh->lock);
 712	} else {
 713		ret = IS_ENABLED(CONFIG_IPV6_ROUTER_PREF) ?
 714		      RT6_NUD_SUCCEED : RT6_NUD_FAIL_DO_RR;
 715	}
 716	rcu_read_unlock_bh();
 717
 718	return ret;
 719}
 720
 721static int rt6_score_route(const struct fib6_nh *nh, u32 fib6_flags, int oif,
 722			   int strict)
 723{
 724	int m = 0;
 725
 726	if (!oif || nh->fib_nh_dev->ifindex == oif)
 727		m = 2;
 728
 729	if (!m && (strict & RT6_LOOKUP_F_IFACE))
 730		return RT6_NUD_FAIL_HARD;
 731#ifdef CONFIG_IPV6_ROUTER_PREF
 732	m |= IPV6_DECODE_PREF(IPV6_EXTRACT_PREF(fib6_flags)) << 2;
 733#endif
 734	if ((strict & RT6_LOOKUP_F_REACHABLE) &&
 735	    !(fib6_flags & RTF_NONEXTHOP) && nh->fib_nh_gw_family) {
 736		int n = rt6_check_neigh(nh);
 737		if (n < 0)
 738			return n;
 739	}
 740	return m;
 741}
 742
 743static bool find_match(struct fib6_nh *nh, u32 fib6_flags,
 744		       int oif, int strict, int *mpri, bool *do_rr)
 745{
 746	bool match_do_rr = false;
 747	bool rc = false;
 748	int m;
 749
 750	if (nh->fib_nh_flags & RTNH_F_DEAD)
 751		goto out;
 752
 753	if (ip6_ignore_linkdown(nh->fib_nh_dev) &&
 754	    nh->fib_nh_flags & RTNH_F_LINKDOWN &&
 755	    !(strict & RT6_LOOKUP_F_IGNORE_LINKSTATE))
 756		goto out;
 757
 758	m = rt6_score_route(nh, fib6_flags, oif, strict);
 759	if (m == RT6_NUD_FAIL_DO_RR) {
 760		match_do_rr = true;
 761		m = 0; /* lowest valid score */
 762	} else if (m == RT6_NUD_FAIL_HARD) {
 763		goto out;
 764	}
 765
 766	if (strict & RT6_LOOKUP_F_REACHABLE)
 767		rt6_probe(nh);
 768
 769	/* note that m can be RT6_NUD_FAIL_PROBE at this point */
 770	if (m > *mpri) {
 771		*do_rr = match_do_rr;
 772		*mpri = m;
 773		rc = true;
 774	}
 775out:
 776	return rc;
 777}
 778
 779struct fib6_nh_frl_arg {
 780	u32		flags;
 781	int		oif;
 782	int		strict;
 783	int		*mpri;
 784	bool		*do_rr;
 785	struct fib6_nh	*nh;
 786};
 787
 788static int rt6_nh_find_match(struct fib6_nh *nh, void *_arg)
 789{
 790	struct fib6_nh_frl_arg *arg = _arg;
 791
 792	arg->nh = nh;
 793	return find_match(nh, arg->flags, arg->oif, arg->strict,
 794			  arg->mpri, arg->do_rr);
 795}
 796
 797static void __find_rr_leaf(struct fib6_info *f6i_start,
 798			   struct fib6_info *nomatch, u32 metric,
 799			   struct fib6_result *res, struct fib6_info **cont,
 800			   int oif, int strict, bool *do_rr, int *mpri)
 801{
 802	struct fib6_info *f6i;
 803
 804	for (f6i = f6i_start;
 805	     f6i && f6i != nomatch;
 806	     f6i = rcu_dereference(f6i->fib6_next)) {
 807		bool matched = false;
 808		struct fib6_nh *nh;
 809
 810		if (cont && f6i->fib6_metric != metric) {
 811			*cont = f6i;
 812			return;
 813		}
 814
 815		if (fib6_check_expired(f6i))
 816			continue;
 817
 818		if (unlikely(f6i->nh)) {
 819			struct fib6_nh_frl_arg arg = {
 820				.flags  = f6i->fib6_flags,
 821				.oif    = oif,
 822				.strict = strict,
 823				.mpri   = mpri,
 824				.do_rr  = do_rr
 825			};
 826
 827			if (nexthop_is_blackhole(f6i->nh)) {
 828				res->fib6_flags = RTF_REJECT;
 829				res->fib6_type = RTN_BLACKHOLE;
 830				res->f6i = f6i;
 831				res->nh = nexthop_fib6_nh(f6i->nh);
 832				return;
 833			}
 834			if (nexthop_for_each_fib6_nh(f6i->nh, rt6_nh_find_match,
 835						     &arg)) {
 836				matched = true;
 837				nh = arg.nh;
 838			}
 839		} else {
 840			nh = f6i->fib6_nh;
 841			if (find_match(nh, f6i->fib6_flags, oif, strict,
 842				       mpri, do_rr))
 843				matched = true;
 844		}
 845		if (matched) {
 846			res->f6i = f6i;
 847			res->nh = nh;
 848			res->fib6_flags = f6i->fib6_flags;
 849			res->fib6_type = f6i->fib6_type;
 850		}
 851	}
 852}
 853
 854static void find_rr_leaf(struct fib6_node *fn, struct fib6_info *leaf,
 855			 struct fib6_info *rr_head, int oif, int strict,
 856			 bool *do_rr, struct fib6_result *res)
 857{
 858	u32 metric = rr_head->fib6_metric;
 859	struct fib6_info *cont = NULL;
 860	int mpri = -1;
 861
 862	__find_rr_leaf(rr_head, NULL, metric, res, &cont,
 863		       oif, strict, do_rr, &mpri);
 864
 865	__find_rr_leaf(leaf, rr_head, metric, res, &cont,
 866		       oif, strict, do_rr, &mpri);
 867
 868	if (res->f6i || !cont)
 869		return;
 870
 871	__find_rr_leaf(cont, NULL, metric, res, NULL,
 872		       oif, strict, do_rr, &mpri);
 873}
 874
 875static void rt6_select(struct net *net, struct fib6_node *fn, int oif,
 876		       struct fib6_result *res, int strict)
 877{
 878	struct fib6_info *leaf = rcu_dereference(fn->leaf);
 879	struct fib6_info *rt0;
 880	bool do_rr = false;
 881	int key_plen;
 882
 883	/* make sure this function or its helpers sets f6i */
 884	res->f6i = NULL;
 885
 886	if (!leaf || leaf == net->ipv6.fib6_null_entry)
 887		goto out;
 888
 889	rt0 = rcu_dereference(fn->rr_ptr);
 890	if (!rt0)
 891		rt0 = leaf;
 892
 893	/* Double check to make sure fn is not an intermediate node
 894	 * and fn->leaf does not points to its child's leaf
 895	 * (This might happen if all routes under fn are deleted from
 896	 * the tree and fib6_repair_tree() is called on the node.)
 897	 */
 898	key_plen = rt0->fib6_dst.plen;
 899#ifdef CONFIG_IPV6_SUBTREES
 900	if (rt0->fib6_src.plen)
 901		key_plen = rt0->fib6_src.plen;
 902#endif
 903	if (fn->fn_bit != key_plen)
 904		goto out;
 905
 906	find_rr_leaf(fn, leaf, rt0, oif, strict, &do_rr, res);
 907	if (do_rr) {
 908		struct fib6_info *next = rcu_dereference(rt0->fib6_next);
 909
 910		/* no entries matched; do round-robin */
 911		if (!next || next->fib6_metric != rt0->fib6_metric)
 912			next = leaf;
 913
 914		if (next != rt0) {
 915			spin_lock_bh(&leaf->fib6_table->tb6_lock);
 916			/* make sure next is not being deleted from the tree */
 917			if (next->fib6_node)
 918				rcu_assign_pointer(fn->rr_ptr, next);
 919			spin_unlock_bh(&leaf->fib6_table->tb6_lock);
 920		}
 921	}
 922
 923out:
 924	if (!res->f6i) {
 925		res->f6i = net->ipv6.fib6_null_entry;
 926		res->nh = res->f6i->fib6_nh;
 927		res->fib6_flags = res->f6i->fib6_flags;
 928		res->fib6_type = res->f6i->fib6_type;
 929	}
 930}
 931
 932static bool rt6_is_gw_or_nonexthop(const struct fib6_result *res)
 933{
 934	return (res->f6i->fib6_flags & RTF_NONEXTHOP) ||
 935	       res->nh->fib_nh_gw_family;
 936}
 937
 938#ifdef CONFIG_IPV6_ROUTE_INFO
 939int rt6_route_rcv(struct net_device *dev, u8 *opt, int len,
 940		  const struct in6_addr *gwaddr)
 941{
 942	struct net *net = dev_net(dev);
 943	struct route_info *rinfo = (struct route_info *) opt;
 944	struct in6_addr prefix_buf, *prefix;
 
 945	unsigned int pref;
 946	unsigned long lifetime;
 947	struct fib6_info *rt;
 948
 949	if (len < sizeof(struct route_info)) {
 950		return -EINVAL;
 951	}
 952
 953	/* Sanity check for prefix_len and length */
 954	if (rinfo->length > 3) {
 955		return -EINVAL;
 956	} else if (rinfo->prefix_len > 128) {
 957		return -EINVAL;
 958	} else if (rinfo->prefix_len > 64) {
 959		if (rinfo->length < 2) {
 960			return -EINVAL;
 961		}
 962	} else if (rinfo->prefix_len > 0) {
 963		if (rinfo->length < 1) {
 964			return -EINVAL;
 965		}
 966	}
 967
 968	pref = rinfo->route_pref;
 969	if (pref == ICMPV6_ROUTER_PREF_INVALID)
 970		return -EINVAL;
 971
 972	lifetime = addrconf_timeout_fixup(ntohl(rinfo->lifetime), HZ);
 973
 974	if (rinfo->length == 3)
 975		prefix = (struct in6_addr *)rinfo->prefix;
 976	else {
 977		/* this function is safe */
 978		ipv6_addr_prefix(&prefix_buf,
 979				 (struct in6_addr *)rinfo->prefix,
 980				 rinfo->prefix_len);
 981		prefix = &prefix_buf;
 982	}
 983
 984	if (rinfo->prefix_len == 0)
 985		rt = rt6_get_dflt_router(net, gwaddr, dev);
 986	else
 987		rt = rt6_get_route_info(net, prefix, rinfo->prefix_len,
 988					gwaddr, dev);
 989
 990	if (rt && !lifetime) {
 991		ip6_del_rt(net, rt, false);
 992		rt = NULL;
 993	}
 994
 995	if (!rt && lifetime)
 996		rt = rt6_add_route_info(net, prefix, rinfo->prefix_len, gwaddr,
 997					dev, pref);
 998	else if (rt)
 999		rt->fib6_flags = RTF_ROUTEINFO |
1000				 (rt->fib6_flags & ~RTF_PREF_MASK) | RTF_PREF(pref);
1001
1002	if (rt) {
1003		if (!addrconf_finite_timeout(lifetime))
 
 
 
1004			fib6_clean_expires(rt);
1005		else
 
1006			fib6_set_expires(rt, jiffies + HZ * lifetime);
 
 
 
 
1007
1008		fib6_info_release(rt);
1009	}
1010	return 0;
1011}
1012#endif
1013
1014/*
1015 *	Misc support functions
1016 */
1017
1018/* called with rcu_lock held */
1019static struct net_device *ip6_rt_get_dev_rcu(const struct fib6_result *res)
1020{
1021	struct net_device *dev = res->nh->fib_nh_dev;
1022
1023	if (res->fib6_flags & (RTF_LOCAL | RTF_ANYCAST)) {
1024		/* for copies of local routes, dst->dev needs to be the
1025		 * device if it is a master device, the master device if
1026		 * device is enslaved, and the loopback as the default
1027		 */
1028		if (netif_is_l3_slave(dev) &&
1029		    !rt6_need_strict(&res->f6i->fib6_dst.addr))
1030			dev = l3mdev_master_dev_rcu(dev);
1031		else if (!netif_is_l3_master(dev))
1032			dev = dev_net(dev)->loopback_dev;
1033		/* last case is netif_is_l3_master(dev) is true in which
1034		 * case we want dev returned to be dev
1035		 */
1036	}
1037
1038	return dev;
1039}
1040
1041static const int fib6_prop[RTN_MAX + 1] = {
1042	[RTN_UNSPEC]	= 0,
1043	[RTN_UNICAST]	= 0,
1044	[RTN_LOCAL]	= 0,
1045	[RTN_BROADCAST]	= 0,
1046	[RTN_ANYCAST]	= 0,
1047	[RTN_MULTICAST]	= 0,
1048	[RTN_BLACKHOLE]	= -EINVAL,
1049	[RTN_UNREACHABLE] = -EHOSTUNREACH,
1050	[RTN_PROHIBIT]	= -EACCES,
1051	[RTN_THROW]	= -EAGAIN,
1052	[RTN_NAT]	= -EINVAL,
1053	[RTN_XRESOLVE]	= -EINVAL,
1054};
1055
1056static int ip6_rt_type_to_error(u8 fib6_type)
1057{
1058	return fib6_prop[fib6_type];
1059}
1060
1061static unsigned short fib6_info_dst_flags(struct fib6_info *rt)
1062{
1063	unsigned short flags = 0;
1064
1065	if (rt->dst_nocount)
1066		flags |= DST_NOCOUNT;
1067	if (rt->dst_nopolicy)
1068		flags |= DST_NOPOLICY;
1069
1070	return flags;
1071}
1072
1073static void ip6_rt_init_dst_reject(struct rt6_info *rt, u8 fib6_type)
1074{
1075	rt->dst.error = ip6_rt_type_to_error(fib6_type);
1076
1077	switch (fib6_type) {
1078	case RTN_BLACKHOLE:
1079		rt->dst.output = dst_discard_out;
1080		rt->dst.input = dst_discard;
1081		break;
1082	case RTN_PROHIBIT:
1083		rt->dst.output = ip6_pkt_prohibit_out;
1084		rt->dst.input = ip6_pkt_prohibit;
1085		break;
1086	case RTN_THROW:
1087	case RTN_UNREACHABLE:
1088	default:
1089		rt->dst.output = ip6_pkt_discard_out;
1090		rt->dst.input = ip6_pkt_discard;
1091		break;
1092	}
1093}
1094
1095static void ip6_rt_init_dst(struct rt6_info *rt, const struct fib6_result *res)
1096{
1097	struct fib6_info *f6i = res->f6i;
1098
1099	if (res->fib6_flags & RTF_REJECT) {
1100		ip6_rt_init_dst_reject(rt, res->fib6_type);
1101		return;
1102	}
1103
1104	rt->dst.error = 0;
1105	rt->dst.output = ip6_output;
1106
1107	if (res->fib6_type == RTN_LOCAL || res->fib6_type == RTN_ANYCAST) {
1108		rt->dst.input = ip6_input;
1109	} else if (ipv6_addr_type(&f6i->fib6_dst.addr) & IPV6_ADDR_MULTICAST) {
1110		rt->dst.input = ip6_mc_input;
1111	} else {
1112		rt->dst.input = ip6_forward;
1113	}
1114
1115	if (res->nh->fib_nh_lws) {
1116		rt->dst.lwtstate = lwtstate_get(res->nh->fib_nh_lws);
1117		lwtunnel_set_redirect(&rt->dst);
1118	}
1119
1120	rt->dst.lastuse = jiffies;
1121}
1122
1123/* Caller must already hold reference to @from */
1124static void rt6_set_from(struct rt6_info *rt, struct fib6_info *from)
1125{
1126	rt->rt6i_flags &= ~RTF_EXPIRES;
1127	rcu_assign_pointer(rt->from, from);
1128	ip_dst_init_metrics(&rt->dst, from->fib6_metrics);
1129}
1130
1131/* Caller must already hold reference to f6i in result */
1132static void ip6_rt_copy_init(struct rt6_info *rt, const struct fib6_result *res)
1133{
1134	const struct fib6_nh *nh = res->nh;
1135	const struct net_device *dev = nh->fib_nh_dev;
1136	struct fib6_info *f6i = res->f6i;
1137
1138	ip6_rt_init_dst(rt, res);
1139
1140	rt->rt6i_dst = f6i->fib6_dst;
1141	rt->rt6i_idev = dev ? in6_dev_get(dev) : NULL;
1142	rt->rt6i_flags = res->fib6_flags;
1143	if (nh->fib_nh_gw_family) {
1144		rt->rt6i_gateway = nh->fib_nh_gw6;
1145		rt->rt6i_flags |= RTF_GATEWAY;
1146	}
1147	rt6_set_from(rt, f6i);
1148#ifdef CONFIG_IPV6_SUBTREES
1149	rt->rt6i_src = f6i->fib6_src;
1150#endif
1151}
1152
1153static struct fib6_node* fib6_backtrack(struct fib6_node *fn,
1154					struct in6_addr *saddr)
1155{
1156	struct fib6_node *pn, *sn;
1157	while (1) {
1158		if (fn->fn_flags & RTN_TL_ROOT)
1159			return NULL;
1160		pn = rcu_dereference(fn->parent);
1161		sn = FIB6_SUBTREE(pn);
1162		if (sn && sn != fn)
1163			fn = fib6_node_lookup(sn, NULL, saddr);
1164		else
1165			fn = pn;
1166		if (fn->fn_flags & RTN_RTINFO)
1167			return fn;
1168	}
1169}
1170
1171static bool ip6_hold_safe(struct net *net, struct rt6_info **prt)
1172{
1173	struct rt6_info *rt = *prt;
1174
1175	if (dst_hold_safe(&rt->dst))
1176		return true;
1177	if (net) {
1178		rt = net->ipv6.ip6_null_entry;
1179		dst_hold(&rt->dst);
1180	} else {
1181		rt = NULL;
1182	}
1183	*prt = rt;
1184	return false;
1185}
1186
1187/* called with rcu_lock held */
1188static struct rt6_info *ip6_create_rt_rcu(const struct fib6_result *res)
1189{
1190	struct net_device *dev = res->nh->fib_nh_dev;
1191	struct fib6_info *f6i = res->f6i;
1192	unsigned short flags;
1193	struct rt6_info *nrt;
1194
1195	if (!fib6_info_hold_safe(f6i))
1196		goto fallback;
1197
1198	flags = fib6_info_dst_flags(f6i);
1199	nrt = ip6_dst_alloc(dev_net(dev), dev, flags);
1200	if (!nrt) {
1201		fib6_info_release(f6i);
1202		goto fallback;
1203	}
1204
1205	ip6_rt_copy_init(nrt, res);
1206	return nrt;
1207
1208fallback:
1209	nrt = dev_net(dev)->ipv6.ip6_null_entry;
1210	dst_hold(&nrt->dst);
1211	return nrt;
1212}
1213
1214INDIRECT_CALLABLE_SCOPE struct rt6_info *ip6_pol_route_lookup(struct net *net,
1215					     struct fib6_table *table,
1216					     struct flowi6 *fl6,
1217					     const struct sk_buff *skb,
1218					     int flags)
1219{
1220	struct fib6_result res = {};
1221	struct fib6_node *fn;
1222	struct rt6_info *rt;
1223
1224	if (fl6->flowi6_flags & FLOWI_FLAG_SKIP_NH_OIF)
1225		flags &= ~RT6_LOOKUP_F_IFACE;
1226
1227	rcu_read_lock();
1228	fn = fib6_node_lookup(&table->tb6_root, &fl6->daddr, &fl6->saddr);
1229restart:
1230	res.f6i = rcu_dereference(fn->leaf);
1231	if (!res.f6i)
1232		res.f6i = net->ipv6.fib6_null_entry;
1233	else
1234		rt6_device_match(net, &res, &fl6->saddr, fl6->flowi6_oif,
1235				 flags);
1236
1237	if (res.f6i == net->ipv6.fib6_null_entry) {
1238		fn = fib6_backtrack(fn, &fl6->saddr);
1239		if (fn)
1240			goto restart;
1241
1242		rt = net->ipv6.ip6_null_entry;
1243		dst_hold(&rt->dst);
1244		goto out;
1245	} else if (res.fib6_flags & RTF_REJECT) {
1246		goto do_create;
1247	}
1248
1249	fib6_select_path(net, &res, fl6, fl6->flowi6_oif,
1250			 fl6->flowi6_oif != 0, skb, flags);
1251
1252	/* Search through exception table */
1253	rt = rt6_find_cached_rt(&res, &fl6->daddr, &fl6->saddr);
1254	if (rt) {
1255		if (ip6_hold_safe(net, &rt))
1256			dst_use_noref(&rt->dst, jiffies);
1257	} else {
1258do_create:
1259		rt = ip6_create_rt_rcu(&res);
1260	}
1261
1262out:
1263	trace_fib6_table_lookup(net, &res, table, fl6);
1264
1265	rcu_read_unlock();
1266
1267	return rt;
1268}
1269
1270struct dst_entry *ip6_route_lookup(struct net *net, struct flowi6 *fl6,
1271				   const struct sk_buff *skb, int flags)
1272{
1273	return fib6_rule_lookup(net, fl6, skb, flags, ip6_pol_route_lookup);
1274}
1275EXPORT_SYMBOL_GPL(ip6_route_lookup);
1276
1277struct rt6_info *rt6_lookup(struct net *net, const struct in6_addr *daddr,
1278			    const struct in6_addr *saddr, int oif,
1279			    const struct sk_buff *skb, int strict)
1280{
1281	struct flowi6 fl6 = {
1282		.flowi6_oif = oif,
1283		.daddr = *daddr,
1284	};
1285	struct dst_entry *dst;
1286	int flags = strict ? RT6_LOOKUP_F_IFACE : 0;
1287
1288	if (saddr) {
1289		memcpy(&fl6.saddr, saddr, sizeof(*saddr));
1290		flags |= RT6_LOOKUP_F_HAS_SADDR;
1291	}
1292
1293	dst = fib6_rule_lookup(net, &fl6, skb, flags, ip6_pol_route_lookup);
1294	if (dst->error == 0)
1295		return (struct rt6_info *) dst;
1296
1297	dst_release(dst);
1298
1299	return NULL;
1300}
1301EXPORT_SYMBOL(rt6_lookup);
1302
1303/* ip6_ins_rt is called with FREE table->tb6_lock.
1304 * It takes new route entry, the addition fails by any reason the
1305 * route is released.
1306 * Caller must hold dst before calling it.
1307 */
1308
1309static int __ip6_ins_rt(struct fib6_info *rt, struct nl_info *info,
1310			struct netlink_ext_ack *extack)
1311{
1312	int err;
1313	struct fib6_table *table;
1314
1315	table = rt->fib6_table;
1316	spin_lock_bh(&table->tb6_lock);
1317	err = fib6_add(&table->tb6_root, rt, info, extack);
1318	spin_unlock_bh(&table->tb6_lock);
1319
1320	return err;
1321}
1322
1323int ip6_ins_rt(struct net *net, struct fib6_info *rt)
1324{
1325	struct nl_info info = {	.nl_net = net, };
1326
1327	return __ip6_ins_rt(rt, &info, NULL);
1328}
1329
1330static struct rt6_info *ip6_rt_cache_alloc(const struct fib6_result *res,
1331					   const struct in6_addr *daddr,
1332					   const struct in6_addr *saddr)
1333{
1334	struct fib6_info *f6i = res->f6i;
1335	struct net_device *dev;
1336	struct rt6_info *rt;
1337
1338	/*
1339	 *	Clone the route.
1340	 */
1341
1342	if (!fib6_info_hold_safe(f6i))
1343		return NULL;
1344
1345	dev = ip6_rt_get_dev_rcu(res);
1346	rt = ip6_dst_alloc(dev_net(dev), dev, 0);
1347	if (!rt) {
1348		fib6_info_release(f6i);
1349		return NULL;
1350	}
1351
1352	ip6_rt_copy_init(rt, res);
1353	rt->rt6i_flags |= RTF_CACHE;
1354	rt->rt6i_dst.addr = *daddr;
1355	rt->rt6i_dst.plen = 128;
1356
1357	if (!rt6_is_gw_or_nonexthop(res)) {
1358		if (f6i->fib6_dst.plen != 128 &&
1359		    ipv6_addr_equal(&f6i->fib6_dst.addr, daddr))
1360			rt->rt6i_flags |= RTF_ANYCAST;
1361#ifdef CONFIG_IPV6_SUBTREES
1362		if (rt->rt6i_src.plen && saddr) {
1363			rt->rt6i_src.addr = *saddr;
1364			rt->rt6i_src.plen = 128;
1365		}
1366#endif
1367	}
1368
1369	return rt;
1370}
1371
1372static struct rt6_info *ip6_rt_pcpu_alloc(const struct fib6_result *res)
1373{
1374	struct fib6_info *f6i = res->f6i;
1375	unsigned short flags = fib6_info_dst_flags(f6i);
1376	struct net_device *dev;
1377	struct rt6_info *pcpu_rt;
1378
1379	if (!fib6_info_hold_safe(f6i))
1380		return NULL;
1381
1382	rcu_read_lock();
1383	dev = ip6_rt_get_dev_rcu(res);
1384	pcpu_rt = ip6_dst_alloc(dev_net(dev), dev, flags | DST_NOCOUNT);
1385	rcu_read_unlock();
1386	if (!pcpu_rt) {
1387		fib6_info_release(f6i);
1388		return NULL;
1389	}
1390	ip6_rt_copy_init(pcpu_rt, res);
1391	pcpu_rt->rt6i_flags |= RTF_PCPU;
1392
1393	if (f6i->nh)
1394		pcpu_rt->sernum = rt_genid_ipv6(dev_net(dev));
1395
1396	return pcpu_rt;
1397}
1398
1399static bool rt6_is_valid(const struct rt6_info *rt6)
1400{
1401	return rt6->sernum == rt_genid_ipv6(dev_net(rt6->dst.dev));
1402}
1403
1404/* It should be called with rcu_read_lock() acquired */
1405static struct rt6_info *rt6_get_pcpu_route(const struct fib6_result *res)
1406{
1407	struct rt6_info *pcpu_rt;
1408
1409	pcpu_rt = this_cpu_read(*res->nh->rt6i_pcpu);
1410
1411	if (pcpu_rt && pcpu_rt->sernum && !rt6_is_valid(pcpu_rt)) {
1412		struct rt6_info *prev, **p;
1413
1414		p = this_cpu_ptr(res->nh->rt6i_pcpu);
1415		prev = xchg(p, NULL);
1416		if (prev) {
1417			dst_dev_put(&prev->dst);
1418			dst_release(&prev->dst);
1419		}
1420
1421		pcpu_rt = NULL;
1422	}
1423
1424	return pcpu_rt;
1425}
1426
1427static struct rt6_info *rt6_make_pcpu_route(struct net *net,
1428					    const struct fib6_result *res)
1429{
1430	struct rt6_info *pcpu_rt, *prev, **p;
1431
1432	pcpu_rt = ip6_rt_pcpu_alloc(res);
1433	if (!pcpu_rt)
1434		return NULL;
1435
1436	p = this_cpu_ptr(res->nh->rt6i_pcpu);
1437	prev = cmpxchg(p, NULL, pcpu_rt);
1438	BUG_ON(prev);
1439
1440	if (res->f6i->fib6_destroying) {
1441		struct fib6_info *from;
1442
1443		from = xchg((__force struct fib6_info **)&pcpu_rt->from, NULL);
1444		fib6_info_release(from);
1445	}
1446
1447	return pcpu_rt;
1448}
1449
1450/* exception hash table implementation
1451 */
1452static DEFINE_SPINLOCK(rt6_exception_lock);
1453
1454/* Remove rt6_ex from hash table and free the memory
1455 * Caller must hold rt6_exception_lock
1456 */
1457static void rt6_remove_exception(struct rt6_exception_bucket *bucket,
1458				 struct rt6_exception *rt6_ex)
1459{
1460	struct fib6_info *from;
1461	struct net *net;
1462
1463	if (!bucket || !rt6_ex)
1464		return;
1465
1466	net = dev_net(rt6_ex->rt6i->dst.dev);
1467	net->ipv6.rt6_stats->fib_rt_cache--;
1468
1469	/* purge completely the exception to allow releasing the held resources:
1470	 * some [sk] cache may keep the dst around for unlimited time
1471	 */
1472	from = xchg((__force struct fib6_info **)&rt6_ex->rt6i->from, NULL);
1473	fib6_info_release(from);
1474	dst_dev_put(&rt6_ex->rt6i->dst);
1475
1476	hlist_del_rcu(&rt6_ex->hlist);
1477	dst_release(&rt6_ex->rt6i->dst);
1478	kfree_rcu(rt6_ex, rcu);
1479	WARN_ON_ONCE(!bucket->depth);
1480	bucket->depth--;
1481}
1482
1483/* Remove oldest rt6_ex in bucket and free the memory
1484 * Caller must hold rt6_exception_lock
1485 */
1486static void rt6_exception_remove_oldest(struct rt6_exception_bucket *bucket)
1487{
1488	struct rt6_exception *rt6_ex, *oldest = NULL;
1489
1490	if (!bucket)
1491		return;
1492
1493	hlist_for_each_entry(rt6_ex, &bucket->chain, hlist) {
1494		if (!oldest || time_before(rt6_ex->stamp, oldest->stamp))
1495			oldest = rt6_ex;
1496	}
1497	rt6_remove_exception(bucket, oldest);
1498}
1499
1500static u32 rt6_exception_hash(const struct in6_addr *dst,
1501			      const struct in6_addr *src)
1502{
1503	static u32 seed __read_mostly;
1504	u32 val;
 
 
 
 
 
 
1505
1506	net_get_random_once(&seed, sizeof(seed));
1507	val = jhash2((const u32 *)dst, sizeof(*dst)/sizeof(u32), seed);
1508
1509#ifdef CONFIG_IPV6_SUBTREES
1510	if (src)
1511		val = jhash2((const u32 *)src, sizeof(*src)/sizeof(u32), val);
1512#endif
1513	return hash_32(val, FIB6_EXCEPTION_BUCKET_SIZE_SHIFT);
 
 
1514}
1515
1516/* Helper function to find the cached rt in the hash table
1517 * and update bucket pointer to point to the bucket for this
1518 * (daddr, saddr) pair
1519 * Caller must hold rt6_exception_lock
1520 */
1521static struct rt6_exception *
1522__rt6_find_exception_spinlock(struct rt6_exception_bucket **bucket,
1523			      const struct in6_addr *daddr,
1524			      const struct in6_addr *saddr)
1525{
1526	struct rt6_exception *rt6_ex;
1527	u32 hval;
1528
1529	if (!(*bucket) || !daddr)
1530		return NULL;
1531
1532	hval = rt6_exception_hash(daddr, saddr);
1533	*bucket += hval;
1534
1535	hlist_for_each_entry(rt6_ex, &(*bucket)->chain, hlist) {
1536		struct rt6_info *rt6 = rt6_ex->rt6i;
1537		bool matched = ipv6_addr_equal(daddr, &rt6->rt6i_dst.addr);
1538
1539#ifdef CONFIG_IPV6_SUBTREES
1540		if (matched && saddr)
1541			matched = ipv6_addr_equal(saddr, &rt6->rt6i_src.addr);
1542#endif
1543		if (matched)
1544			return rt6_ex;
1545	}
1546	return NULL;
1547}
1548
1549/* Helper function to find the cached rt in the hash table
1550 * and update bucket pointer to point to the bucket for this
1551 * (daddr, saddr) pair
1552 * Caller must hold rcu_read_lock()
1553 */
1554static struct rt6_exception *
1555__rt6_find_exception_rcu(struct rt6_exception_bucket **bucket,
1556			 const struct in6_addr *daddr,
1557			 const struct in6_addr *saddr)
1558{
1559	struct rt6_exception *rt6_ex;
1560	u32 hval;
1561
1562	WARN_ON_ONCE(!rcu_read_lock_held());
1563
1564	if (!(*bucket) || !daddr)
1565		return NULL;
1566
1567	hval = rt6_exception_hash(daddr, saddr);
1568	*bucket += hval;
1569
1570	hlist_for_each_entry_rcu(rt6_ex, &(*bucket)->chain, hlist) {
1571		struct rt6_info *rt6 = rt6_ex->rt6i;
1572		bool matched = ipv6_addr_equal(daddr, &rt6->rt6i_dst.addr);
1573
1574#ifdef CONFIG_IPV6_SUBTREES
1575		if (matched && saddr)
1576			matched = ipv6_addr_equal(saddr, &rt6->rt6i_src.addr);
1577#endif
1578		if (matched)
1579			return rt6_ex;
1580	}
1581	return NULL;
1582}
1583
1584static unsigned int fib6_mtu(const struct fib6_result *res)
1585{
1586	const struct fib6_nh *nh = res->nh;
1587	unsigned int mtu;
1588
1589	if (res->f6i->fib6_pmtu) {
1590		mtu = res->f6i->fib6_pmtu;
1591	} else {
1592		struct net_device *dev = nh->fib_nh_dev;
1593		struct inet6_dev *idev;
1594
1595		rcu_read_lock();
1596		idev = __in6_dev_get(dev);
1597		mtu = idev->cnf.mtu6;
1598		rcu_read_unlock();
1599	}
1600
1601	mtu = min_t(unsigned int, mtu, IP6_MAX_MTU);
1602
1603	return mtu - lwtunnel_headroom(nh->fib_nh_lws, mtu);
1604}
1605
1606#define FIB6_EXCEPTION_BUCKET_FLUSHED  0x1UL
1607
1608/* used when the flushed bit is not relevant, only access to the bucket
1609 * (ie., all bucket users except rt6_insert_exception);
1610 *
1611 * called under rcu lock; sometimes called with rt6_exception_lock held
1612 */
1613static
1614struct rt6_exception_bucket *fib6_nh_get_excptn_bucket(const struct fib6_nh *nh,
1615						       spinlock_t *lock)
1616{
1617	struct rt6_exception_bucket *bucket;
1618
1619	if (lock)
1620		bucket = rcu_dereference_protected(nh->rt6i_exception_bucket,
1621						   lockdep_is_held(lock));
1622	else
1623		bucket = rcu_dereference(nh->rt6i_exception_bucket);
1624
1625	/* remove bucket flushed bit if set */
1626	if (bucket) {
1627		unsigned long p = (unsigned long)bucket;
1628
1629		p &= ~FIB6_EXCEPTION_BUCKET_FLUSHED;
1630		bucket = (struct rt6_exception_bucket *)p;
1631	}
1632
1633	return bucket;
1634}
1635
1636static bool fib6_nh_excptn_bucket_flushed(struct rt6_exception_bucket *bucket)
1637{
1638	unsigned long p = (unsigned long)bucket;
1639
1640	return !!(p & FIB6_EXCEPTION_BUCKET_FLUSHED);
1641}
1642
1643/* called with rt6_exception_lock held */
1644static void fib6_nh_excptn_bucket_set_flushed(struct fib6_nh *nh,
1645					      spinlock_t *lock)
1646{
1647	struct rt6_exception_bucket *bucket;
1648	unsigned long p;
1649
1650	bucket = rcu_dereference_protected(nh->rt6i_exception_bucket,
1651					   lockdep_is_held(lock));
1652
1653	p = (unsigned long)bucket;
1654	p |= FIB6_EXCEPTION_BUCKET_FLUSHED;
1655	bucket = (struct rt6_exception_bucket *)p;
1656	rcu_assign_pointer(nh->rt6i_exception_bucket, bucket);
1657}
1658
1659static int rt6_insert_exception(struct rt6_info *nrt,
1660				const struct fib6_result *res)
1661{
1662	struct net *net = dev_net(nrt->dst.dev);
1663	struct rt6_exception_bucket *bucket;
1664	struct fib6_info *f6i = res->f6i;
1665	struct in6_addr *src_key = NULL;
1666	struct rt6_exception *rt6_ex;
1667	struct fib6_nh *nh = res->nh;
 
1668	int err = 0;
1669
1670	spin_lock_bh(&rt6_exception_lock);
1671
1672	bucket = rcu_dereference_protected(nh->rt6i_exception_bucket,
1673					  lockdep_is_held(&rt6_exception_lock));
1674	if (!bucket) {
1675		bucket = kcalloc(FIB6_EXCEPTION_BUCKET_SIZE, sizeof(*bucket),
1676				 GFP_ATOMIC);
1677		if (!bucket) {
1678			err = -ENOMEM;
1679			goto out;
1680		}
1681		rcu_assign_pointer(nh->rt6i_exception_bucket, bucket);
1682	} else if (fib6_nh_excptn_bucket_flushed(bucket)) {
1683		err = -EINVAL;
1684		goto out;
1685	}
1686
1687#ifdef CONFIG_IPV6_SUBTREES
1688	/* fib6_src.plen != 0 indicates f6i is in subtree
1689	 * and exception table is indexed by a hash of
1690	 * both fib6_dst and fib6_src.
1691	 * Otherwise, the exception table is indexed by
1692	 * a hash of only fib6_dst.
1693	 */
1694	if (f6i->fib6_src.plen)
1695		src_key = &nrt->rt6i_src.addr;
1696#endif
1697	/* rt6_mtu_change() might lower mtu on f6i.
1698	 * Only insert this exception route if its mtu
1699	 * is less than f6i's mtu value.
1700	 */
1701	if (dst_metric_raw(&nrt->dst, RTAX_MTU) >= fib6_mtu(res)) {
1702		err = -EINVAL;
1703		goto out;
1704	}
1705
1706	rt6_ex = __rt6_find_exception_spinlock(&bucket, &nrt->rt6i_dst.addr,
1707					       src_key);
1708	if (rt6_ex)
1709		rt6_remove_exception(bucket, rt6_ex);
1710
1711	rt6_ex = kzalloc(sizeof(*rt6_ex), GFP_ATOMIC);
1712	if (!rt6_ex) {
1713		err = -ENOMEM;
1714		goto out;
1715	}
1716	rt6_ex->rt6i = nrt;
1717	rt6_ex->stamp = jiffies;
1718	hlist_add_head_rcu(&rt6_ex->hlist, &bucket->chain);
1719	bucket->depth++;
1720	net->ipv6.rt6_stats->fib_rt_cache++;
1721
1722	if (bucket->depth > FIB6_MAX_DEPTH)
 
 
1723		rt6_exception_remove_oldest(bucket);
1724
1725out:
1726	spin_unlock_bh(&rt6_exception_lock);
1727
1728	/* Update fn->fn_sernum to invalidate all cached dst */
1729	if (!err) {
1730		spin_lock_bh(&f6i->fib6_table->tb6_lock);
1731		fib6_update_sernum(net, f6i);
1732		spin_unlock_bh(&f6i->fib6_table->tb6_lock);
1733		fib6_force_start_gc(net);
1734	}
1735
1736	return err;
1737}
1738
1739static void fib6_nh_flush_exceptions(struct fib6_nh *nh, struct fib6_info *from)
1740{
1741	struct rt6_exception_bucket *bucket;
1742	struct rt6_exception *rt6_ex;
1743	struct hlist_node *tmp;
1744	int i;
1745
1746	spin_lock_bh(&rt6_exception_lock);
1747
1748	bucket = fib6_nh_get_excptn_bucket(nh, &rt6_exception_lock);
1749	if (!bucket)
1750		goto out;
1751
1752	/* Prevent rt6_insert_exception() to recreate the bucket list */
1753	if (!from)
1754		fib6_nh_excptn_bucket_set_flushed(nh, &rt6_exception_lock);
1755
1756	for (i = 0; i < FIB6_EXCEPTION_BUCKET_SIZE; i++) {
1757		hlist_for_each_entry_safe(rt6_ex, tmp, &bucket->chain, hlist) {
1758			if (!from ||
1759			    rcu_access_pointer(rt6_ex->rt6i->from) == from)
1760				rt6_remove_exception(bucket, rt6_ex);
1761		}
1762		WARN_ON_ONCE(!from && bucket->depth);
1763		bucket++;
1764	}
1765out:
1766	spin_unlock_bh(&rt6_exception_lock);
1767}
1768
1769static int rt6_nh_flush_exceptions(struct fib6_nh *nh, void *arg)
1770{
1771	struct fib6_info *f6i = arg;
1772
1773	fib6_nh_flush_exceptions(nh, f6i);
1774
1775	return 0;
1776}
1777
1778void rt6_flush_exceptions(struct fib6_info *f6i)
1779{
1780	if (f6i->nh)
1781		nexthop_for_each_fib6_nh(f6i->nh, rt6_nh_flush_exceptions,
1782					 f6i);
1783	else
1784		fib6_nh_flush_exceptions(f6i->fib6_nh, f6i);
1785}
1786
1787/* Find cached rt in the hash table inside passed in rt
1788 * Caller has to hold rcu_read_lock()
1789 */
1790static struct rt6_info *rt6_find_cached_rt(const struct fib6_result *res,
1791					   const struct in6_addr *daddr,
1792					   const struct in6_addr *saddr)
1793{
1794	const struct in6_addr *src_key = NULL;
1795	struct rt6_exception_bucket *bucket;
1796	struct rt6_exception *rt6_ex;
1797	struct rt6_info *ret = NULL;
1798
1799#ifdef CONFIG_IPV6_SUBTREES
1800	/* fib6i_src.plen != 0 indicates f6i is in subtree
1801	 * and exception table is indexed by a hash of
1802	 * both fib6_dst and fib6_src.
1803	 * However, the src addr used to create the hash
1804	 * might not be exactly the passed in saddr which
1805	 * is a /128 addr from the flow.
1806	 * So we need to use f6i->fib6_src to redo lookup
1807	 * if the passed in saddr does not find anything.
1808	 * (See the logic in ip6_rt_cache_alloc() on how
1809	 * rt->rt6i_src is updated.)
1810	 */
1811	if (res->f6i->fib6_src.plen)
1812		src_key = saddr;
1813find_ex:
1814#endif
1815	bucket = fib6_nh_get_excptn_bucket(res->nh, NULL);
1816	rt6_ex = __rt6_find_exception_rcu(&bucket, daddr, src_key);
1817
1818	if (rt6_ex && !rt6_check_expired(rt6_ex->rt6i))
1819		ret = rt6_ex->rt6i;
1820
1821#ifdef CONFIG_IPV6_SUBTREES
1822	/* Use fib6_src as src_key and redo lookup */
1823	if (!ret && src_key && src_key != &res->f6i->fib6_src.addr) {
1824		src_key = &res->f6i->fib6_src.addr;
1825		goto find_ex;
1826	}
1827#endif
1828
1829	return ret;
1830}
1831
1832/* Remove the passed in cached rt from the hash table that contains it */
1833static int fib6_nh_remove_exception(const struct fib6_nh *nh, int plen,
1834				    const struct rt6_info *rt)
1835{
1836	const struct in6_addr *src_key = NULL;
1837	struct rt6_exception_bucket *bucket;
1838	struct rt6_exception *rt6_ex;
1839	int err;
1840
1841	if (!rcu_access_pointer(nh->rt6i_exception_bucket))
1842		return -ENOENT;
1843
1844	spin_lock_bh(&rt6_exception_lock);
1845	bucket = fib6_nh_get_excptn_bucket(nh, &rt6_exception_lock);
1846
1847#ifdef CONFIG_IPV6_SUBTREES
1848	/* rt6i_src.plen != 0 indicates 'from' is in subtree
1849	 * and exception table is indexed by a hash of
1850	 * both rt6i_dst and rt6i_src.
1851	 * Otherwise, the exception table is indexed by
1852	 * a hash of only rt6i_dst.
1853	 */
1854	if (plen)
1855		src_key = &rt->rt6i_src.addr;
1856#endif
1857	rt6_ex = __rt6_find_exception_spinlock(&bucket,
1858					       &rt->rt6i_dst.addr,
1859					       src_key);
1860	if (rt6_ex) {
1861		rt6_remove_exception(bucket, rt6_ex);
1862		err = 0;
1863	} else {
1864		err = -ENOENT;
1865	}
1866
1867	spin_unlock_bh(&rt6_exception_lock);
1868	return err;
1869}
1870
1871struct fib6_nh_excptn_arg {
1872	struct rt6_info	*rt;
1873	int		plen;
1874};
1875
1876static int rt6_nh_remove_exception_rt(struct fib6_nh *nh, void *_arg)
1877{
1878	struct fib6_nh_excptn_arg *arg = _arg;
1879	int err;
1880
1881	err = fib6_nh_remove_exception(nh, arg->plen, arg->rt);
1882	if (err == 0)
1883		return 1;
1884
1885	return 0;
1886}
1887
1888static int rt6_remove_exception_rt(struct rt6_info *rt)
1889{
1890	struct fib6_info *from;
1891
1892	from = rcu_dereference(rt->from);
1893	if (!from || !(rt->rt6i_flags & RTF_CACHE))
1894		return -EINVAL;
1895
1896	if (from->nh) {
1897		struct fib6_nh_excptn_arg arg = {
1898			.rt = rt,
1899			.plen = from->fib6_src.plen
1900		};
1901		int rc;
1902
1903		/* rc = 1 means an entry was found */
1904		rc = nexthop_for_each_fib6_nh(from->nh,
1905					      rt6_nh_remove_exception_rt,
1906					      &arg);
1907		return rc ? 0 : -ENOENT;
1908	}
1909
1910	return fib6_nh_remove_exception(from->fib6_nh,
1911					from->fib6_src.plen, rt);
1912}
1913
1914/* Find rt6_ex which contains the passed in rt cache and
1915 * refresh its stamp
1916 */
1917static void fib6_nh_update_exception(const struct fib6_nh *nh, int plen,
1918				     const struct rt6_info *rt)
1919{
1920	const struct in6_addr *src_key = NULL;
1921	struct rt6_exception_bucket *bucket;
1922	struct rt6_exception *rt6_ex;
1923
1924	bucket = fib6_nh_get_excptn_bucket(nh, NULL);
1925#ifdef CONFIG_IPV6_SUBTREES
1926	/* rt6i_src.plen != 0 indicates 'from' is in subtree
1927	 * and exception table is indexed by a hash of
1928	 * both rt6i_dst and rt6i_src.
1929	 * Otherwise, the exception table is indexed by
1930	 * a hash of only rt6i_dst.
1931	 */
1932	if (plen)
1933		src_key = &rt->rt6i_src.addr;
1934#endif
1935	rt6_ex = __rt6_find_exception_rcu(&bucket, &rt->rt6i_dst.addr, src_key);
1936	if (rt6_ex)
1937		rt6_ex->stamp = jiffies;
1938}
1939
1940struct fib6_nh_match_arg {
1941	const struct net_device *dev;
1942	const struct in6_addr	*gw;
1943	struct fib6_nh		*match;
1944};
1945
1946/* determine if fib6_nh has given device and gateway */
1947static int fib6_nh_find_match(struct fib6_nh *nh, void *_arg)
1948{
1949	struct fib6_nh_match_arg *arg = _arg;
1950
1951	if (arg->dev != nh->fib_nh_dev ||
1952	    (arg->gw && !nh->fib_nh_gw_family) ||
1953	    (!arg->gw && nh->fib_nh_gw_family) ||
1954	    (arg->gw && !ipv6_addr_equal(arg->gw, &nh->fib_nh_gw6)))
1955		return 0;
1956
1957	arg->match = nh;
1958
1959	/* found a match, break the loop */
1960	return 1;
1961}
1962
1963static void rt6_update_exception_stamp_rt(struct rt6_info *rt)
1964{
1965	struct fib6_info *from;
1966	struct fib6_nh *fib6_nh;
1967
1968	rcu_read_lock();
1969
1970	from = rcu_dereference(rt->from);
1971	if (!from || !(rt->rt6i_flags & RTF_CACHE))
1972		goto unlock;
1973
1974	if (from->nh) {
1975		struct fib6_nh_match_arg arg = {
1976			.dev = rt->dst.dev,
1977			.gw = &rt->rt6i_gateway,
1978		};
1979
1980		nexthop_for_each_fib6_nh(from->nh, fib6_nh_find_match, &arg);
1981
1982		if (!arg.match)
1983			goto unlock;
1984		fib6_nh = arg.match;
1985	} else {
1986		fib6_nh = from->fib6_nh;
1987	}
1988	fib6_nh_update_exception(fib6_nh, from->fib6_src.plen, rt);
1989unlock:
1990	rcu_read_unlock();
1991}
1992
1993static bool rt6_mtu_change_route_allowed(struct inet6_dev *idev,
1994					 struct rt6_info *rt, int mtu)
1995{
1996	/* If the new MTU is lower than the route PMTU, this new MTU will be the
1997	 * lowest MTU in the path: always allow updating the route PMTU to
1998	 * reflect PMTU decreases.
1999	 *
2000	 * If the new MTU is higher, and the route PMTU is equal to the local
2001	 * MTU, this means the old MTU is the lowest in the path, so allow
2002	 * updating it: if other nodes now have lower MTUs, PMTU discovery will
2003	 * handle this.
2004	 */
2005
2006	if (dst_mtu(&rt->dst) >= mtu)
2007		return true;
2008
2009	if (dst_mtu(&rt->dst) == idev->cnf.mtu6)
2010		return true;
2011
2012	return false;
2013}
2014
2015static void rt6_exceptions_update_pmtu(struct inet6_dev *idev,
2016				       const struct fib6_nh *nh, int mtu)
2017{
2018	struct rt6_exception_bucket *bucket;
2019	struct rt6_exception *rt6_ex;
2020	int i;
2021
2022	bucket = fib6_nh_get_excptn_bucket(nh, &rt6_exception_lock);
2023	if (!bucket)
2024		return;
2025
2026	for (i = 0; i < FIB6_EXCEPTION_BUCKET_SIZE; i++) {
2027		hlist_for_each_entry(rt6_ex, &bucket->chain, hlist) {
2028			struct rt6_info *entry = rt6_ex->rt6i;
2029
2030			/* For RTF_CACHE with rt6i_pmtu == 0 (i.e. a redirected
2031			 * route), the metrics of its rt->from have already
2032			 * been updated.
2033			 */
2034			if (dst_metric_raw(&entry->dst, RTAX_MTU) &&
2035			    rt6_mtu_change_route_allowed(idev, entry, mtu))
2036				dst_metric_set(&entry->dst, RTAX_MTU, mtu);
2037		}
2038		bucket++;
2039	}
2040}
2041
2042#define RTF_CACHE_GATEWAY	(RTF_GATEWAY | RTF_CACHE)
2043
2044static void fib6_nh_exceptions_clean_tohost(const struct fib6_nh *nh,
2045					    const struct in6_addr *gateway)
2046{
2047	struct rt6_exception_bucket *bucket;
2048	struct rt6_exception *rt6_ex;
2049	struct hlist_node *tmp;
2050	int i;
2051
2052	if (!rcu_access_pointer(nh->rt6i_exception_bucket))
2053		return;
2054
2055	spin_lock_bh(&rt6_exception_lock);
2056	bucket = fib6_nh_get_excptn_bucket(nh, &rt6_exception_lock);
2057	if (bucket) {
2058		for (i = 0; i < FIB6_EXCEPTION_BUCKET_SIZE; i++) {
2059			hlist_for_each_entry_safe(rt6_ex, tmp,
2060						  &bucket->chain, hlist) {
2061				struct rt6_info *entry = rt6_ex->rt6i;
2062
2063				if ((entry->rt6i_flags & RTF_CACHE_GATEWAY) ==
2064				    RTF_CACHE_GATEWAY &&
2065				    ipv6_addr_equal(gateway,
2066						    &entry->rt6i_gateway)) {
2067					rt6_remove_exception(bucket, rt6_ex);
2068				}
2069			}
2070			bucket++;
2071		}
2072	}
2073
2074	spin_unlock_bh(&rt6_exception_lock);
2075}
2076
2077static void rt6_age_examine_exception(struct rt6_exception_bucket *bucket,
2078				      struct rt6_exception *rt6_ex,
2079				      struct fib6_gc_args *gc_args,
2080				      unsigned long now)
2081{
2082	struct rt6_info *rt = rt6_ex->rt6i;
2083
2084	/* we are pruning and obsoleting aged-out and non gateway exceptions
2085	 * even if others have still references to them, so that on next
2086	 * dst_check() such references can be dropped.
2087	 * EXPIRES exceptions - e.g. pmtu-generated ones are pruned when
2088	 * expired, independently from their aging, as per RFC 8201 section 4
2089	 */
2090	if (!(rt->rt6i_flags & RTF_EXPIRES)) {
2091		if (time_after_eq(now, rt->dst.lastuse + gc_args->timeout)) {
2092			RT6_TRACE("aging clone %p\n", rt);
2093			rt6_remove_exception(bucket, rt6_ex);
2094			return;
2095		}
2096	} else if (time_after(jiffies, rt->dst.expires)) {
2097		RT6_TRACE("purging expired route %p\n", rt);
2098		rt6_remove_exception(bucket, rt6_ex);
2099		return;
2100	}
2101
2102	if (rt->rt6i_flags & RTF_GATEWAY) {
2103		struct neighbour *neigh;
2104		__u8 neigh_flags = 0;
2105
2106		neigh = __ipv6_neigh_lookup_noref(rt->dst.dev, &rt->rt6i_gateway);
2107		if (neigh)
2108			neigh_flags = neigh->flags;
2109
2110		if (!(neigh_flags & NTF_ROUTER)) {
2111			RT6_TRACE("purging route %p via non-router but gateway\n",
2112				  rt);
2113			rt6_remove_exception(bucket, rt6_ex);
2114			return;
2115		}
2116	}
2117
2118	gc_args->more++;
2119}
2120
2121static void fib6_nh_age_exceptions(const struct fib6_nh *nh,
2122				   struct fib6_gc_args *gc_args,
2123				   unsigned long now)
2124{
2125	struct rt6_exception_bucket *bucket;
2126	struct rt6_exception *rt6_ex;
2127	struct hlist_node *tmp;
2128	int i;
2129
2130	if (!rcu_access_pointer(nh->rt6i_exception_bucket))
2131		return;
2132
2133	rcu_read_lock_bh();
2134	spin_lock(&rt6_exception_lock);
2135	bucket = fib6_nh_get_excptn_bucket(nh, &rt6_exception_lock);
2136	if (bucket) {
2137		for (i = 0; i < FIB6_EXCEPTION_BUCKET_SIZE; i++) {
2138			hlist_for_each_entry_safe(rt6_ex, tmp,
2139						  &bucket->chain, hlist) {
2140				rt6_age_examine_exception(bucket, rt6_ex,
2141							  gc_args, now);
2142			}
2143			bucket++;
2144		}
2145	}
2146	spin_unlock(&rt6_exception_lock);
2147	rcu_read_unlock_bh();
2148}
2149
2150struct fib6_nh_age_excptn_arg {
2151	struct fib6_gc_args	*gc_args;
2152	unsigned long		now;
2153};
2154
2155static int rt6_nh_age_exceptions(struct fib6_nh *nh, void *_arg)
2156{
2157	struct fib6_nh_age_excptn_arg *arg = _arg;
2158
2159	fib6_nh_age_exceptions(nh, arg->gc_args, arg->now);
2160	return 0;
2161}
2162
2163void rt6_age_exceptions(struct fib6_info *f6i,
2164			struct fib6_gc_args *gc_args,
2165			unsigned long now)
2166{
2167	if (f6i->nh) {
2168		struct fib6_nh_age_excptn_arg arg = {
2169			.gc_args = gc_args,
2170			.now = now
2171		};
2172
2173		nexthop_for_each_fib6_nh(f6i->nh, rt6_nh_age_exceptions,
2174					 &arg);
2175	} else {
2176		fib6_nh_age_exceptions(f6i->fib6_nh, gc_args, now);
2177	}
2178}
2179
2180/* must be called with rcu lock held */
2181int fib6_table_lookup(struct net *net, struct fib6_table *table, int oif,
2182		      struct flowi6 *fl6, struct fib6_result *res, int strict)
2183{
2184	struct fib6_node *fn, *saved_fn;
2185
2186	fn = fib6_node_lookup(&table->tb6_root, &fl6->daddr, &fl6->saddr);
2187	saved_fn = fn;
2188
2189	if (fl6->flowi6_flags & FLOWI_FLAG_SKIP_NH_OIF)
2190		oif = 0;
2191
2192redo_rt6_select:
2193	rt6_select(net, fn, oif, res, strict);
2194	if (res->f6i == net->ipv6.fib6_null_entry) {
2195		fn = fib6_backtrack(fn, &fl6->saddr);
2196		if (fn)
2197			goto redo_rt6_select;
2198		else if (strict & RT6_LOOKUP_F_REACHABLE) {
2199			/* also consider unreachable route */
2200			strict &= ~RT6_LOOKUP_F_REACHABLE;
2201			fn = saved_fn;
2202			goto redo_rt6_select;
2203		}
2204	}
2205
2206	trace_fib6_table_lookup(net, res, table, fl6);
2207
2208	return 0;
2209}
2210
2211struct rt6_info *ip6_pol_route(struct net *net, struct fib6_table *table,
2212			       int oif, struct flowi6 *fl6,
2213			       const struct sk_buff *skb, int flags)
2214{
2215	struct fib6_result res = {};
2216	struct rt6_info *rt = NULL;
2217	int strict = 0;
2218
2219	WARN_ON_ONCE((flags & RT6_LOOKUP_F_DST_NOREF) &&
2220		     !rcu_read_lock_held());
2221
2222	strict |= flags & RT6_LOOKUP_F_IFACE;
2223	strict |= flags & RT6_LOOKUP_F_IGNORE_LINKSTATE;
2224	if (net->ipv6.devconf_all->forwarding == 0)
2225		strict |= RT6_LOOKUP_F_REACHABLE;
2226
2227	rcu_read_lock();
2228
2229	fib6_table_lookup(net, table, oif, fl6, &res, strict);
2230	if (res.f6i == net->ipv6.fib6_null_entry)
2231		goto out;
2232
2233	fib6_select_path(net, &res, fl6, oif, false, skb, strict);
2234
2235	/*Search through exception table */
2236	rt = rt6_find_cached_rt(&res, &fl6->daddr, &fl6->saddr);
2237	if (rt) {
2238		goto out;
2239	} else if (unlikely((fl6->flowi6_flags & FLOWI_FLAG_KNOWN_NH) &&
2240			    !res.nh->fib_nh_gw_family)) {
2241		/* Create a RTF_CACHE clone which will not be
2242		 * owned by the fib6 tree.  It is for the special case where
2243		 * the daddr in the skb during the neighbor look-up is different
2244		 * from the fl6->daddr used to look-up route here.
2245		 */
2246		rt = ip6_rt_cache_alloc(&res, &fl6->daddr, NULL);
2247
2248		if (rt) {
2249			/* 1 refcnt is taken during ip6_rt_cache_alloc().
2250			 * As rt6_uncached_list_add() does not consume refcnt,
2251			 * this refcnt is always returned to the caller even
2252			 * if caller sets RT6_LOOKUP_F_DST_NOREF flag.
2253			 */
2254			rt6_uncached_list_add(rt);
2255			atomic_inc(&net->ipv6.rt6_stats->fib_rt_uncache);
2256			rcu_read_unlock();
2257
2258			return rt;
2259		}
2260	} else {
2261		/* Get a percpu copy */
2262		local_bh_disable();
2263		rt = rt6_get_pcpu_route(&res);
2264
2265		if (!rt)
2266			rt = rt6_make_pcpu_route(net, &res);
2267
2268		local_bh_enable();
2269	}
2270out:
2271	if (!rt)
2272		rt = net->ipv6.ip6_null_entry;
2273	if (!(flags & RT6_LOOKUP_F_DST_NOREF))
2274		ip6_hold_safe(net, &rt);
2275	rcu_read_unlock();
2276
2277	return rt;
2278}
2279EXPORT_SYMBOL_GPL(ip6_pol_route);
2280
2281INDIRECT_CALLABLE_SCOPE struct rt6_info *ip6_pol_route_input(struct net *net,
2282					    struct fib6_table *table,
2283					    struct flowi6 *fl6,
2284					    const struct sk_buff *skb,
2285					    int flags)
2286{
2287	return ip6_pol_route(net, table, fl6->flowi6_iif, fl6, skb, flags);
2288}
2289
2290struct dst_entry *ip6_route_input_lookup(struct net *net,
2291					 struct net_device *dev,
2292					 struct flowi6 *fl6,
2293					 const struct sk_buff *skb,
2294					 int flags)
2295{
2296	if (rt6_need_strict(&fl6->daddr) && dev->type != ARPHRD_PIMREG)
2297		flags |= RT6_LOOKUP_F_IFACE;
2298
2299	return fib6_rule_lookup(net, fl6, skb, flags, ip6_pol_route_input);
2300}
2301EXPORT_SYMBOL_GPL(ip6_route_input_lookup);
2302
2303static void ip6_multipath_l3_keys(const struct sk_buff *skb,
2304				  struct flow_keys *keys,
2305				  struct flow_keys *flkeys)
2306{
2307	const struct ipv6hdr *outer_iph = ipv6_hdr(skb);
2308	const struct ipv6hdr *key_iph = outer_iph;
2309	struct flow_keys *_flkeys = flkeys;
2310	const struct ipv6hdr *inner_iph;
2311	const struct icmp6hdr *icmph;
2312	struct ipv6hdr _inner_iph;
2313	struct icmp6hdr _icmph;
2314
2315	if (likely(outer_iph->nexthdr != IPPROTO_ICMPV6))
2316		goto out;
2317
2318	icmph = skb_header_pointer(skb, skb_transport_offset(skb),
2319				   sizeof(_icmph), &_icmph);
2320	if (!icmph)
2321		goto out;
2322
2323	if (!icmpv6_is_err(icmph->icmp6_type))
2324		goto out;
2325
2326	inner_iph = skb_header_pointer(skb,
2327				       skb_transport_offset(skb) + sizeof(*icmph),
2328				       sizeof(_inner_iph), &_inner_iph);
2329	if (!inner_iph)
2330		goto out;
2331
2332	key_iph = inner_iph;
2333	_flkeys = NULL;
2334out:
2335	if (_flkeys) {
2336		keys->addrs.v6addrs.src = _flkeys->addrs.v6addrs.src;
2337		keys->addrs.v6addrs.dst = _flkeys->addrs.v6addrs.dst;
2338		keys->tags.flow_label = _flkeys->tags.flow_label;
2339		keys->basic.ip_proto = _flkeys->basic.ip_proto;
2340	} else {
2341		keys->addrs.v6addrs.src = key_iph->saddr;
2342		keys->addrs.v6addrs.dst = key_iph->daddr;
2343		keys->tags.flow_label = ip6_flowlabel(key_iph);
2344		keys->basic.ip_proto = key_iph->nexthdr;
2345	}
2346}
2347
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
2348/* if skb is set it will be used and fl6 can be NULL */
2349u32 rt6_multipath_hash(const struct net *net, const struct flowi6 *fl6,
2350		       const struct sk_buff *skb, struct flow_keys *flkeys)
2351{
2352	struct flow_keys hash_keys;
2353	u32 mhash;
2354
2355	switch (ip6_multipath_hash_policy(net)) {
2356	case 0:
2357		memset(&hash_keys, 0, sizeof(hash_keys));
2358		hash_keys.control.addr_type = FLOW_DISSECTOR_KEY_IPV6_ADDRS;
2359		if (skb) {
2360			ip6_multipath_l3_keys(skb, &hash_keys, flkeys);
2361		} else {
2362			hash_keys.addrs.v6addrs.src = fl6->saddr;
2363			hash_keys.addrs.v6addrs.dst = fl6->daddr;
2364			hash_keys.tags.flow_label = (__force u32)flowi6_get_flowlabel(fl6);
2365			hash_keys.basic.ip_proto = fl6->flowi6_proto;
2366		}
 
2367		break;
2368	case 1:
2369		if (skb) {
2370			unsigned int flag = FLOW_DISSECTOR_F_STOP_AT_ENCAP;
2371			struct flow_keys keys;
2372
2373			/* short-circuit if we already have L4 hash present */
2374			if (skb->l4_hash)
2375				return skb_get_hash_raw(skb) >> 1;
2376
2377			memset(&hash_keys, 0, sizeof(hash_keys));
2378
2379                        if (!flkeys) {
2380				skb_flow_dissect_flow_keys(skb, &keys, flag);
2381				flkeys = &keys;
2382			}
2383			hash_keys.control.addr_type = FLOW_DISSECTOR_KEY_IPV6_ADDRS;
2384			hash_keys.addrs.v6addrs.src = flkeys->addrs.v6addrs.src;
2385			hash_keys.addrs.v6addrs.dst = flkeys->addrs.v6addrs.dst;
2386			hash_keys.ports.src = flkeys->ports.src;
2387			hash_keys.ports.dst = flkeys->ports.dst;
2388			hash_keys.basic.ip_proto = flkeys->basic.ip_proto;
2389		} else {
2390			memset(&hash_keys, 0, sizeof(hash_keys));
2391			hash_keys.control.addr_type = FLOW_DISSECTOR_KEY_IPV6_ADDRS;
2392			hash_keys.addrs.v6addrs.src = fl6->saddr;
2393			hash_keys.addrs.v6addrs.dst = fl6->daddr;
2394			hash_keys.ports.src = fl6->fl6_sport;
2395			hash_keys.ports.dst = fl6->fl6_dport;
2396			hash_keys.basic.ip_proto = fl6->flowi6_proto;
2397		}
 
2398		break;
2399	case 2:
2400		memset(&hash_keys, 0, sizeof(hash_keys));
2401		hash_keys.control.addr_type = FLOW_DISSECTOR_KEY_IPV6_ADDRS;
2402		if (skb) {
2403			struct flow_keys keys;
2404
2405			if (!flkeys) {
2406				skb_flow_dissect_flow_keys(skb, &keys, 0);
2407				flkeys = &keys;
2408			}
2409
2410			/* Inner can be v4 or v6 */
2411			if (flkeys->control.addr_type == FLOW_DISSECTOR_KEY_IPV4_ADDRS) {
2412				hash_keys.control.addr_type = FLOW_DISSECTOR_KEY_IPV4_ADDRS;
2413				hash_keys.addrs.v4addrs.src = flkeys->addrs.v4addrs.src;
2414				hash_keys.addrs.v4addrs.dst = flkeys->addrs.v4addrs.dst;
2415			} else if (flkeys->control.addr_type == FLOW_DISSECTOR_KEY_IPV6_ADDRS) {
2416				hash_keys.control.addr_type = FLOW_DISSECTOR_KEY_IPV6_ADDRS;
2417				hash_keys.addrs.v6addrs.src = flkeys->addrs.v6addrs.src;
2418				hash_keys.addrs.v6addrs.dst = flkeys->addrs.v6addrs.dst;
2419				hash_keys.tags.flow_label = flkeys->tags.flow_label;
2420				hash_keys.basic.ip_proto = flkeys->basic.ip_proto;
2421			} else {
2422				/* Same as case 0 */
2423				hash_keys.control.addr_type = FLOW_DISSECTOR_KEY_IPV6_ADDRS;
2424				ip6_multipath_l3_keys(skb, &hash_keys, flkeys);
2425			}
2426		} else {
2427			/* Same as case 0 */
2428			hash_keys.control.addr_type = FLOW_DISSECTOR_KEY_IPV6_ADDRS;
2429			hash_keys.addrs.v6addrs.src = fl6->saddr;
2430			hash_keys.addrs.v6addrs.dst = fl6->daddr;
2431			hash_keys.tags.flow_label = (__force u32)flowi6_get_flowlabel(fl6);
2432			hash_keys.basic.ip_proto = fl6->flowi6_proto;
2433		}
 
 
 
 
 
 
 
2434		break;
2435	}
2436	mhash = flow_hash_from_keys(&hash_keys);
2437
2438	return mhash >> 1;
2439}
2440
2441/* Called with rcu held */
2442void ip6_route_input(struct sk_buff *skb)
2443{
2444	const struct ipv6hdr *iph = ipv6_hdr(skb);
2445	struct net *net = dev_net(skb->dev);
2446	int flags = RT6_LOOKUP_F_HAS_SADDR | RT6_LOOKUP_F_DST_NOREF;
2447	struct ip_tunnel_info *tun_info;
2448	struct flowi6 fl6 = {
2449		.flowi6_iif = skb->dev->ifindex,
2450		.daddr = iph->daddr,
2451		.saddr = iph->saddr,
2452		.flowlabel = ip6_flowinfo(iph),
2453		.flowi6_mark = skb->mark,
2454		.flowi6_proto = iph->nexthdr,
2455	};
2456	struct flow_keys *flkeys = NULL, _flkeys;
2457
2458	tun_info = skb_tunnel_info(skb);
2459	if (tun_info && !(tun_info->mode & IP_TUNNEL_INFO_TX))
2460		fl6.flowi6_tun_key.tun_id = tun_info->key.tun_id;
2461
2462	if (fib6_rules_early_flow_dissect(net, skb, &fl6, &_flkeys))
2463		flkeys = &_flkeys;
2464
2465	if (unlikely(fl6.flowi6_proto == IPPROTO_ICMPV6))
2466		fl6.mp_hash = rt6_multipath_hash(net, &fl6, skb, flkeys);
2467	skb_dst_drop(skb);
2468	skb_dst_set_noref(skb, ip6_route_input_lookup(net, skb->dev,
2469						      &fl6, skb, flags));
2470}
2471
2472INDIRECT_CALLABLE_SCOPE struct rt6_info *ip6_pol_route_output(struct net *net,
2473					     struct fib6_table *table,
2474					     struct flowi6 *fl6,
2475					     const struct sk_buff *skb,
2476					     int flags)
2477{
2478	return ip6_pol_route(net, table, fl6->flowi6_oif, fl6, skb, flags);
2479}
2480
2481struct dst_entry *ip6_route_output_flags_noref(struct net *net,
2482					       const struct sock *sk,
2483					       struct flowi6 *fl6, int flags)
 
2484{
2485	bool any_src;
2486
2487	if (ipv6_addr_type(&fl6->daddr) &
2488	    (IPV6_ADDR_MULTICAST | IPV6_ADDR_LINKLOCAL)) {
2489		struct dst_entry *dst;
2490
2491		/* This function does not take refcnt on the dst */
2492		dst = l3mdev_link_scope_lookup(net, fl6);
2493		if (dst)
2494			return dst;
2495	}
2496
2497	fl6->flowi6_iif = LOOPBACK_IFINDEX;
2498
2499	flags |= RT6_LOOKUP_F_DST_NOREF;
2500	any_src = ipv6_addr_any(&fl6->saddr);
2501	if ((sk && sk->sk_bound_dev_if) || rt6_need_strict(&fl6->daddr) ||
2502	    (fl6->flowi6_oif && any_src))
2503		flags |= RT6_LOOKUP_F_IFACE;
2504
2505	if (!any_src)
2506		flags |= RT6_LOOKUP_F_HAS_SADDR;
2507	else if (sk)
2508		flags |= rt6_srcprefs2flags(inet6_sk(sk)->srcprefs);
2509
2510	return fib6_rule_lookup(net, fl6, NULL, flags, ip6_pol_route_output);
2511}
2512EXPORT_SYMBOL_GPL(ip6_route_output_flags_noref);
2513
2514struct dst_entry *ip6_route_output_flags(struct net *net,
2515					 const struct sock *sk,
2516					 struct flowi6 *fl6,
2517					 int flags)
2518{
2519        struct dst_entry *dst;
2520        struct rt6_info *rt6;
2521
2522        rcu_read_lock();
2523        dst = ip6_route_output_flags_noref(net, sk, fl6, flags);
2524        rt6 = (struct rt6_info *)dst;
2525        /* For dst cached in uncached_list, refcnt is already taken. */
2526        if (list_empty(&rt6->rt6i_uncached) && !dst_hold_safe(dst)) {
2527                dst = &net->ipv6.ip6_null_entry->dst;
2528                dst_hold(dst);
2529        }
2530        rcu_read_unlock();
2531
2532        return dst;
2533}
2534EXPORT_SYMBOL_GPL(ip6_route_output_flags);
2535
2536struct dst_entry *ip6_blackhole_route(struct net *net, struct dst_entry *dst_orig)
2537{
2538	struct rt6_info *rt, *ort = (struct rt6_info *) dst_orig;
2539	struct net_device *loopback_dev = net->loopback_dev;
2540	struct dst_entry *new = NULL;
2541
2542	rt = dst_alloc(&ip6_dst_blackhole_ops, loopback_dev, 1,
2543		       DST_OBSOLETE_DEAD, 0);
2544	if (rt) {
2545		rt6_info_init(rt);
2546		atomic_inc(&net->ipv6.rt6_stats->fib_rt_alloc);
2547
2548		new = &rt->dst;
2549		new->__use = 1;
2550		new->input = dst_discard;
2551		new->output = dst_discard_out;
2552
2553		dst_copy_metrics(new, &ort->dst);
2554
2555		rt->rt6i_idev = in6_dev_get(loopback_dev);
2556		rt->rt6i_gateway = ort->rt6i_gateway;
2557		rt->rt6i_flags = ort->rt6i_flags & ~RTF_PCPU;
2558
2559		memcpy(&rt->rt6i_dst, &ort->rt6i_dst, sizeof(struct rt6key));
2560#ifdef CONFIG_IPV6_SUBTREES
2561		memcpy(&rt->rt6i_src, &ort->rt6i_src, sizeof(struct rt6key));
2562#endif
2563	}
2564
2565	dst_release(dst_orig);
2566	return new ? new : ERR_PTR(-ENOMEM);
2567}
2568
2569/*
2570 *	Destination cache support functions
2571 */
2572
2573static bool fib6_check(struct fib6_info *f6i, u32 cookie)
2574{
2575	u32 rt_cookie = 0;
2576
2577	if (!fib6_get_cookie_safe(f6i, &rt_cookie) || rt_cookie != cookie)
2578		return false;
2579
2580	if (fib6_check_expired(f6i))
2581		return false;
2582
2583	return true;
2584}
2585
2586static struct dst_entry *rt6_check(struct rt6_info *rt,
2587				   struct fib6_info *from,
2588				   u32 cookie)
2589{
2590	u32 rt_cookie = 0;
2591
2592	if (!from || !fib6_get_cookie_safe(from, &rt_cookie) ||
2593	    rt_cookie != cookie)
2594		return NULL;
2595
2596	if (rt6_check_expired(rt))
2597		return NULL;
2598
2599	return &rt->dst;
2600}
2601
2602static struct dst_entry *rt6_dst_from_check(struct rt6_info *rt,
2603					    struct fib6_info *from,
2604					    u32 cookie)
2605{
2606	if (!__rt6_check_expired(rt) &&
2607	    rt->dst.obsolete == DST_OBSOLETE_FORCE_CHK &&
2608	    fib6_check(from, cookie))
2609		return &rt->dst;
2610	else
2611		return NULL;
2612}
2613
2614static struct dst_entry *ip6_dst_check(struct dst_entry *dst, u32 cookie)
 
2615{
2616	struct dst_entry *dst_ret;
2617	struct fib6_info *from;
2618	struct rt6_info *rt;
2619
2620	rt = container_of(dst, struct rt6_info, dst);
2621
2622	if (rt->sernum)
2623		return rt6_is_valid(rt) ? dst : NULL;
2624
2625	rcu_read_lock();
2626
2627	/* All IPV6 dsts are created with ->obsolete set to the value
2628	 * DST_OBSOLETE_FORCE_CHK which forces validation calls down
2629	 * into this function always.
2630	 */
2631
2632	from = rcu_dereference(rt->from);
2633
2634	if (from && (rt->rt6i_flags & RTF_PCPU ||
2635	    unlikely(!list_empty(&rt->rt6i_uncached))))
2636		dst_ret = rt6_dst_from_check(rt, from, cookie);
2637	else
2638		dst_ret = rt6_check(rt, from, cookie);
2639
2640	rcu_read_unlock();
2641
2642	return dst_ret;
2643}
 
2644
2645static struct dst_entry *ip6_negative_advice(struct dst_entry *dst)
 
2646{
2647	struct rt6_info *rt = (struct rt6_info *) dst;
2648
2649	if (rt) {
2650		if (rt->rt6i_flags & RTF_CACHE) {
2651			rcu_read_lock();
2652			if (rt6_check_expired(rt)) {
2653				rt6_remove_exception_rt(rt);
2654				dst = NULL;
2655			}
2656			rcu_read_unlock();
2657		} else {
2658			dst_release(dst);
2659			dst = NULL;
2660		}
 
 
2661	}
2662	return dst;
2663}
2664
2665static void ip6_link_failure(struct sk_buff *skb)
2666{
2667	struct rt6_info *rt;
2668
2669	icmpv6_send(skb, ICMPV6_DEST_UNREACH, ICMPV6_ADDR_UNREACH, 0);
2670
2671	rt = (struct rt6_info *) skb_dst(skb);
2672	if (rt) {
2673		rcu_read_lock();
2674		if (rt->rt6i_flags & RTF_CACHE) {
2675			rt6_remove_exception_rt(rt);
2676		} else {
2677			struct fib6_info *from;
2678			struct fib6_node *fn;
2679
2680			from = rcu_dereference(rt->from);
2681			if (from) {
2682				fn = rcu_dereference(from->fib6_node);
2683				if (fn && (rt->rt6i_flags & RTF_DEFAULT))
2684					fn->fn_sernum = -1;
2685			}
2686		}
2687		rcu_read_unlock();
2688	}
2689}
2690
2691static void rt6_update_expires(struct rt6_info *rt0, int timeout)
2692{
2693	if (!(rt0->rt6i_flags & RTF_EXPIRES)) {
2694		struct fib6_info *from;
2695
2696		rcu_read_lock();
2697		from = rcu_dereference(rt0->from);
2698		if (from)
2699			rt0->dst.expires = from->expires;
2700		rcu_read_unlock();
2701	}
2702
2703	dst_set_expires(&rt0->dst, timeout);
2704	rt0->rt6i_flags |= RTF_EXPIRES;
2705}
2706
2707static void rt6_do_update_pmtu(struct rt6_info *rt, u32 mtu)
2708{
2709	struct net *net = dev_net(rt->dst.dev);
2710
2711	dst_metric_set(&rt->dst, RTAX_MTU, mtu);
2712	rt->rt6i_flags |= RTF_MODIFIED;
2713	rt6_update_expires(rt, net->ipv6.sysctl.ip6_rt_mtu_expires);
2714}
2715
2716static bool rt6_cache_allowed_for_pmtu(const struct rt6_info *rt)
2717{
2718	return !(rt->rt6i_flags & RTF_CACHE) &&
2719		(rt->rt6i_flags & RTF_PCPU || rcu_access_pointer(rt->from));
2720}
2721
2722static void __ip6_rt_update_pmtu(struct dst_entry *dst, const struct sock *sk,
2723				 const struct ipv6hdr *iph, u32 mtu,
2724				 bool confirm_neigh)
2725{
2726	const struct in6_addr *daddr, *saddr;
2727	struct rt6_info *rt6 = (struct rt6_info *)dst;
2728
2729	/* Note: do *NOT* check dst_metric_locked(dst, RTAX_MTU)
2730	 * IPv6 pmtu discovery isn't optional, so 'mtu lock' cannot disable it.
2731	 * [see also comment in rt6_mtu_change_route()]
2732	 */
2733
2734	if (iph) {
2735		daddr = &iph->daddr;
2736		saddr = &iph->saddr;
2737	} else if (sk) {
2738		daddr = &sk->sk_v6_daddr;
2739		saddr = &inet6_sk(sk)->saddr;
2740	} else {
2741		daddr = NULL;
2742		saddr = NULL;
2743	}
2744
2745	if (confirm_neigh)
2746		dst_confirm_neigh(dst, daddr);
2747
2748	mtu = max_t(u32, mtu, IPV6_MIN_MTU);
 
2749	if (mtu >= dst_mtu(dst))
2750		return;
2751
2752	if (!rt6_cache_allowed_for_pmtu(rt6)) {
2753		rt6_do_update_pmtu(rt6, mtu);
2754		/* update rt6_ex->stamp for cache */
2755		if (rt6->rt6i_flags & RTF_CACHE)
2756			rt6_update_exception_stamp_rt(rt6);
2757	} else if (daddr) {
2758		struct fib6_result res = {};
2759		struct rt6_info *nrt6;
2760
2761		rcu_read_lock();
2762		res.f6i = rcu_dereference(rt6->from);
2763		if (!res.f6i)
2764			goto out_unlock;
2765
2766		res.fib6_flags = res.f6i->fib6_flags;
2767		res.fib6_type = res.f6i->fib6_type;
2768
2769		if (res.f6i->nh) {
2770			struct fib6_nh_match_arg arg = {
2771				.dev = dst->dev,
2772				.gw = &rt6->rt6i_gateway,
2773			};
2774
2775			nexthop_for_each_fib6_nh(res.f6i->nh,
2776						 fib6_nh_find_match, &arg);
2777
2778			/* fib6_info uses a nexthop that does not have fib6_nh
2779			 * using the dst->dev + gw. Should be impossible.
2780			 */
2781			if (!arg.match)
2782				goto out_unlock;
2783
2784			res.nh = arg.match;
2785		} else {
2786			res.nh = res.f6i->fib6_nh;
2787		}
2788
2789		nrt6 = ip6_rt_cache_alloc(&res, daddr, saddr);
2790		if (nrt6) {
2791			rt6_do_update_pmtu(nrt6, mtu);
2792			if (rt6_insert_exception(nrt6, &res))
2793				dst_release_immediate(&nrt6->dst);
2794		}
2795out_unlock:
2796		rcu_read_unlock();
2797	}
2798}
2799
2800static void ip6_rt_update_pmtu(struct dst_entry *dst, struct sock *sk,
2801			       struct sk_buff *skb, u32 mtu,
2802			       bool confirm_neigh)
2803{
2804	__ip6_rt_update_pmtu(dst, sk, skb ? ipv6_hdr(skb) : NULL, mtu,
2805			     confirm_neigh);
2806}
2807
2808void ip6_update_pmtu(struct sk_buff *skb, struct net *net, __be32 mtu,
2809		     int oif, u32 mark, kuid_t uid)
2810{
2811	const struct ipv6hdr *iph = (struct ipv6hdr *) skb->data;
2812	struct dst_entry *dst;
2813	struct flowi6 fl6 = {
2814		.flowi6_oif = oif,
2815		.flowi6_mark = mark ? mark : IP6_REPLY_MARK(net, skb->mark),
2816		.daddr = iph->daddr,
2817		.saddr = iph->saddr,
2818		.flowlabel = ip6_flowinfo(iph),
2819		.flowi6_uid = uid,
2820	};
2821
2822	dst = ip6_route_output(net, NULL, &fl6);
2823	if (!dst->error)
2824		__ip6_rt_update_pmtu(dst, NULL, iph, ntohl(mtu), true);
2825	dst_release(dst);
2826}
2827EXPORT_SYMBOL_GPL(ip6_update_pmtu);
2828
2829void ip6_sk_update_pmtu(struct sk_buff *skb, struct sock *sk, __be32 mtu)
2830{
2831	int oif = sk->sk_bound_dev_if;
2832	struct dst_entry *dst;
2833
2834	if (!oif && skb->dev)
2835		oif = l3mdev_master_ifindex(skb->dev);
2836
2837	ip6_update_pmtu(skb, sock_net(sk), mtu, oif, sk->sk_mark, sk->sk_uid);
 
2838
2839	dst = __sk_dst_get(sk);
2840	if (!dst || !dst->obsolete ||
2841	    dst->ops->check(dst, inet6_sk(sk)->dst_cookie))
2842		return;
2843
2844	bh_lock_sock(sk);
2845	if (!sock_owned_by_user(sk) && !ipv6_addr_v4mapped(&sk->sk_v6_daddr))
2846		ip6_datagram_dst_update(sk, false);
2847	bh_unlock_sock(sk);
2848}
2849EXPORT_SYMBOL_GPL(ip6_sk_update_pmtu);
2850
2851void ip6_sk_dst_store_flow(struct sock *sk, struct dst_entry *dst,
2852			   const struct flowi6 *fl6)
2853{
2854#ifdef CONFIG_IPV6_SUBTREES
2855	struct ipv6_pinfo *np = inet6_sk(sk);
2856#endif
2857
2858	ip6_dst_store(sk, dst,
2859		      ipv6_addr_equal(&fl6->daddr, &sk->sk_v6_daddr) ?
2860		      &sk->sk_v6_daddr : NULL,
2861#ifdef CONFIG_IPV6_SUBTREES
2862		      ipv6_addr_equal(&fl6->saddr, &np->saddr) ?
2863		      &np->saddr :
2864#endif
2865		      NULL);
2866}
2867
2868static bool ip6_redirect_nh_match(const struct fib6_result *res,
2869				  struct flowi6 *fl6,
2870				  const struct in6_addr *gw,
2871				  struct rt6_info **ret)
2872{
2873	const struct fib6_nh *nh = res->nh;
2874
2875	if (nh->fib_nh_flags & RTNH_F_DEAD || !nh->fib_nh_gw_family ||
2876	    fl6->flowi6_oif != nh->fib_nh_dev->ifindex)
2877		return false;
2878
2879	/* rt_cache's gateway might be different from its 'parent'
2880	 * in the case of an ip redirect.
2881	 * So we keep searching in the exception table if the gateway
2882	 * is different.
2883	 */
2884	if (!ipv6_addr_equal(gw, &nh->fib_nh_gw6)) {
2885		struct rt6_info *rt_cache;
2886
2887		rt_cache = rt6_find_cached_rt(res, &fl6->daddr, &fl6->saddr);
2888		if (rt_cache &&
2889		    ipv6_addr_equal(gw, &rt_cache->rt6i_gateway)) {
2890			*ret = rt_cache;
2891			return true;
2892		}
2893		return false;
2894	}
2895	return true;
2896}
2897
2898struct fib6_nh_rd_arg {
2899	struct fib6_result	*res;
2900	struct flowi6		*fl6;
2901	const struct in6_addr	*gw;
2902	struct rt6_info		**ret;
2903};
2904
2905static int fib6_nh_redirect_match(struct fib6_nh *nh, void *_arg)
2906{
2907	struct fib6_nh_rd_arg *arg = _arg;
2908
2909	arg->res->nh = nh;
2910	return ip6_redirect_nh_match(arg->res, arg->fl6, arg->gw, arg->ret);
2911}
2912
2913/* Handle redirects */
2914struct ip6rd_flowi {
2915	struct flowi6 fl6;
2916	struct in6_addr gateway;
2917};
2918
2919INDIRECT_CALLABLE_SCOPE struct rt6_info *__ip6_route_redirect(struct net *net,
2920					     struct fib6_table *table,
2921					     struct flowi6 *fl6,
2922					     const struct sk_buff *skb,
2923					     int flags)
2924{
2925	struct ip6rd_flowi *rdfl = (struct ip6rd_flowi *)fl6;
2926	struct rt6_info *ret = NULL;
2927	struct fib6_result res = {};
2928	struct fib6_nh_rd_arg arg = {
2929		.res = &res,
2930		.fl6 = fl6,
2931		.gw  = &rdfl->gateway,
2932		.ret = &ret
2933	};
2934	struct fib6_info *rt;
2935	struct fib6_node *fn;
2936
2937	/* l3mdev_update_flow overrides oif if the device is enslaved; in
2938	 * this case we must match on the real ingress device, so reset it
2939	 */
2940	if (fl6->flowi6_flags & FLOWI_FLAG_SKIP_NH_OIF)
2941		fl6->flowi6_oif = skb->dev->ifindex;
2942
2943	/* Get the "current" route for this destination and
2944	 * check if the redirect has come from appropriate router.
2945	 *
2946	 * RFC 4861 specifies that redirects should only be
2947	 * accepted if they come from the nexthop to the target.
2948	 * Due to the way the routes are chosen, this notion
2949	 * is a bit fuzzy and one might need to check all possible
2950	 * routes.
2951	 */
2952
2953	rcu_read_lock();
2954	fn = fib6_node_lookup(&table->tb6_root, &fl6->daddr, &fl6->saddr);
2955restart:
2956	for_each_fib6_node_rt_rcu(fn) {
2957		res.f6i = rt;
2958		if (fib6_check_expired(rt))
2959			continue;
2960		if (rt->fib6_flags & RTF_REJECT)
2961			break;
2962		if (unlikely(rt->nh)) {
2963			if (nexthop_is_blackhole(rt->nh))
2964				continue;
2965			/* on match, res->nh is filled in and potentially ret */
2966			if (nexthop_for_each_fib6_nh(rt->nh,
2967						     fib6_nh_redirect_match,
2968						     &arg))
2969				goto out;
2970		} else {
2971			res.nh = rt->fib6_nh;
2972			if (ip6_redirect_nh_match(&res, fl6, &rdfl->gateway,
2973						  &ret))
2974				goto out;
2975		}
2976	}
2977
2978	if (!rt)
2979		rt = net->ipv6.fib6_null_entry;
2980	else if (rt->fib6_flags & RTF_REJECT) {
2981		ret = net->ipv6.ip6_null_entry;
2982		goto out;
2983	}
2984
2985	if (rt == net->ipv6.fib6_null_entry) {
2986		fn = fib6_backtrack(fn, &fl6->saddr);
2987		if (fn)
2988			goto restart;
2989	}
2990
2991	res.f6i = rt;
2992	res.nh = rt->fib6_nh;
2993out:
2994	if (ret) {
2995		ip6_hold_safe(net, &ret);
2996	} else {
2997		res.fib6_flags = res.f6i->fib6_flags;
2998		res.fib6_type = res.f6i->fib6_type;
2999		ret = ip6_create_rt_rcu(&res);
3000	}
3001
3002	rcu_read_unlock();
3003
3004	trace_fib6_table_lookup(net, &res, table, fl6);
3005	return ret;
3006};
3007
3008static struct dst_entry *ip6_route_redirect(struct net *net,
3009					    const struct flowi6 *fl6,
3010					    const struct sk_buff *skb,
3011					    const struct in6_addr *gateway)
3012{
3013	int flags = RT6_LOOKUP_F_HAS_SADDR;
3014	struct ip6rd_flowi rdfl;
3015
3016	rdfl.fl6 = *fl6;
3017	rdfl.gateway = *gateway;
3018
3019	return fib6_rule_lookup(net, &rdfl.fl6, skb,
3020				flags, __ip6_route_redirect);
3021}
3022
3023void ip6_redirect(struct sk_buff *skb, struct net *net, int oif, u32 mark,
3024		  kuid_t uid)
3025{
3026	const struct ipv6hdr *iph = (struct ipv6hdr *) skb->data;
3027	struct dst_entry *dst;
3028	struct flowi6 fl6 = {
3029		.flowi6_iif = LOOPBACK_IFINDEX,
3030		.flowi6_oif = oif,
3031		.flowi6_mark = mark,
3032		.daddr = iph->daddr,
3033		.saddr = iph->saddr,
3034		.flowlabel = ip6_flowinfo(iph),
3035		.flowi6_uid = uid,
3036	};
3037
3038	dst = ip6_route_redirect(net, &fl6, skb, &ipv6_hdr(skb)->saddr);
3039	rt6_do_redirect(dst, NULL, skb);
3040	dst_release(dst);
3041}
3042EXPORT_SYMBOL_GPL(ip6_redirect);
3043
3044void ip6_redirect_no_header(struct sk_buff *skb, struct net *net, int oif)
3045{
3046	const struct ipv6hdr *iph = ipv6_hdr(skb);
3047	const struct rd_msg *msg = (struct rd_msg *)icmp6_hdr(skb);
3048	struct dst_entry *dst;
3049	struct flowi6 fl6 = {
3050		.flowi6_iif = LOOPBACK_IFINDEX,
3051		.flowi6_oif = oif,
3052		.daddr = msg->dest,
3053		.saddr = iph->daddr,
3054		.flowi6_uid = sock_net_uid(net, NULL),
3055	};
3056
3057	dst = ip6_route_redirect(net, &fl6, skb, &iph->saddr);
3058	rt6_do_redirect(dst, NULL, skb);
3059	dst_release(dst);
3060}
3061
3062void ip6_sk_redirect(struct sk_buff *skb, struct sock *sk)
3063{
3064	ip6_redirect(skb, sock_net(sk), sk->sk_bound_dev_if, sk->sk_mark,
3065		     sk->sk_uid);
3066}
3067EXPORT_SYMBOL_GPL(ip6_sk_redirect);
3068
3069static unsigned int ip6_default_advmss(const struct dst_entry *dst)
3070{
3071	struct net_device *dev = dst->dev;
3072	unsigned int mtu = dst_mtu(dst);
3073	struct net *net = dev_net(dev);
3074
3075	mtu -= sizeof(struct ipv6hdr) + sizeof(struct tcphdr);
3076
3077	if (mtu < net->ipv6.sysctl.ip6_rt_min_advmss)
3078		mtu = net->ipv6.sysctl.ip6_rt_min_advmss;
3079
3080	/*
3081	 * Maximal non-jumbo IPv6 payload is IPV6_MAXPLEN and
3082	 * corresponding MSS is IPV6_MAXPLEN - tcp_header_size.
3083	 * IPV6_MAXPLEN is also valid and means: "any MSS,
3084	 * rely only on pmtu discovery"
3085	 */
3086	if (mtu > IPV6_MAXPLEN - sizeof(struct tcphdr))
3087		mtu = IPV6_MAXPLEN;
3088	return mtu;
3089}
3090
3091static unsigned int ip6_mtu(const struct dst_entry *dst)
3092{
3093	struct inet6_dev *idev;
3094	unsigned int mtu;
3095
3096	mtu = dst_metric_raw(dst, RTAX_MTU);
3097	if (mtu)
3098		goto out;
3099
3100	mtu = IPV6_MIN_MTU;
3101
3102	rcu_read_lock();
3103	idev = __in6_dev_get(dst->dev);
3104	if (idev)
3105		mtu = idev->cnf.mtu6;
3106	rcu_read_unlock();
3107
3108out:
3109	mtu = min_t(unsigned int, mtu, IP6_MAX_MTU);
3110
3111	return mtu - lwtunnel_headroom(dst->lwtstate, mtu);
3112}
 
3113
3114/* MTU selection:
3115 * 1. mtu on route is locked - use it
3116 * 2. mtu from nexthop exception
3117 * 3. mtu from egress device
3118 *
3119 * based on ip6_dst_mtu_forward and exception logic of
3120 * rt6_find_cached_rt; called with rcu_read_lock
3121 */
3122u32 ip6_mtu_from_fib6(const struct fib6_result *res,
3123		      const struct in6_addr *daddr,
3124		      const struct in6_addr *saddr)
3125{
3126	const struct fib6_nh *nh = res->nh;
3127	struct fib6_info *f6i = res->f6i;
3128	struct inet6_dev *idev;
3129	struct rt6_info *rt;
3130	u32 mtu = 0;
3131
3132	if (unlikely(fib6_metric_locked(f6i, RTAX_MTU))) {
3133		mtu = f6i->fib6_pmtu;
3134		if (mtu)
3135			goto out;
3136	}
3137
3138	rt = rt6_find_cached_rt(res, daddr, saddr);
3139	if (unlikely(rt)) {
3140		mtu = dst_metric_raw(&rt->dst, RTAX_MTU);
3141	} else {
3142		struct net_device *dev = nh->fib_nh_dev;
3143
3144		mtu = IPV6_MIN_MTU;
3145		idev = __in6_dev_get(dev);
3146		if (idev && idev->cnf.mtu6 > mtu)
3147			mtu = idev->cnf.mtu6;
3148	}
3149
3150	mtu = min_t(unsigned int, mtu, IP6_MAX_MTU);
3151out:
3152	return mtu - lwtunnel_headroom(nh->fib_nh_lws, mtu);
3153}
3154
3155struct dst_entry *icmp6_dst_alloc(struct net_device *dev,
3156				  struct flowi6 *fl6)
3157{
3158	struct dst_entry *dst;
3159	struct rt6_info *rt;
3160	struct inet6_dev *idev = in6_dev_get(dev);
3161	struct net *net = dev_net(dev);
3162
3163	if (unlikely(!idev))
3164		return ERR_PTR(-ENODEV);
3165
3166	rt = ip6_dst_alloc(net, dev, 0);
3167	if (unlikely(!rt)) {
3168		in6_dev_put(idev);
3169		dst = ERR_PTR(-ENOMEM);
3170		goto out;
3171	}
3172
3173	rt->dst.input = ip6_input;
3174	rt->dst.output  = ip6_output;
3175	rt->rt6i_gateway  = fl6->daddr;
3176	rt->rt6i_dst.addr = fl6->daddr;
3177	rt->rt6i_dst.plen = 128;
3178	rt->rt6i_idev     = idev;
3179	dst_metric_set(&rt->dst, RTAX_HOPLIMIT, 0);
3180
3181	/* Add this dst into uncached_list so that rt6_disable_ip() can
3182	 * do proper release of the net_device
3183	 */
3184	rt6_uncached_list_add(rt);
3185	atomic_inc(&net->ipv6.rt6_stats->fib_rt_uncache);
3186
3187	dst = xfrm_lookup(net, &rt->dst, flowi6_to_flowi(fl6), NULL, 0);
3188
3189out:
3190	return dst;
3191}
3192
3193static int ip6_dst_gc(struct dst_ops *ops)
3194{
3195	struct net *net = container_of(ops, struct net, ipv6.ip6_dst_ops);
3196	int rt_min_interval = net->ipv6.sysctl.ip6_rt_gc_min_interval;
3197	int rt_max_size = net->ipv6.sysctl.ip6_rt_max_size;
3198	int rt_elasticity = net->ipv6.sysctl.ip6_rt_gc_elasticity;
3199	int rt_gc_timeout = net->ipv6.sysctl.ip6_rt_gc_timeout;
3200	unsigned long rt_last_gc = net->ipv6.ip6_rt_last_gc;
 
3201	int entries;
3202
3203	entries = dst_entries_get_fast(ops);
3204	if (entries > rt_max_size)
3205		entries = dst_entries_get_slow(ops);
3206
3207	if (time_after(rt_last_gc + rt_min_interval, jiffies) &&
3208	    entries <= rt_max_size)
3209		goto out;
3210
3211	net->ipv6.ip6_rt_gc_expire++;
3212	fib6_run_gc(net->ipv6.ip6_rt_gc_expire, net, true);
3213	entries = dst_entries_get_slow(ops);
3214	if (entries < ops->gc_thresh)
3215		net->ipv6.ip6_rt_gc_expire = rt_gc_timeout>>1;
3216out:
3217	net->ipv6.ip6_rt_gc_expire -= net->ipv6.ip6_rt_gc_expire>>rt_elasticity;
3218	return entries > rt_max_size;
3219}
3220
3221static int ip6_nh_lookup_table(struct net *net, struct fib6_config *cfg,
3222			       const struct in6_addr *gw_addr, u32 tbid,
3223			       int flags, struct fib6_result *res)
3224{
3225	struct flowi6 fl6 = {
3226		.flowi6_oif = cfg->fc_ifindex,
3227		.daddr = *gw_addr,
3228		.saddr = cfg->fc_prefsrc,
3229	};
3230	struct fib6_table *table;
3231	int err;
3232
3233	table = fib6_get_table(net, tbid);
3234	if (!table)
3235		return -EINVAL;
3236
3237	if (!ipv6_addr_any(&cfg->fc_prefsrc))
3238		flags |= RT6_LOOKUP_F_HAS_SADDR;
3239
3240	flags |= RT6_LOOKUP_F_IGNORE_LINKSTATE;
3241
3242	err = fib6_table_lookup(net, table, cfg->fc_ifindex, &fl6, res, flags);
3243	if (!err && res->f6i != net->ipv6.fib6_null_entry)
3244		fib6_select_path(net, res, &fl6, cfg->fc_ifindex,
3245				 cfg->fc_ifindex != 0, NULL, flags);
3246
3247	return err;
3248}
3249
3250static int ip6_route_check_nh_onlink(struct net *net,
3251				     struct fib6_config *cfg,
3252				     const struct net_device *dev,
3253				     struct netlink_ext_ack *extack)
3254{
3255	u32 tbid = l3mdev_fib_table_rcu(dev) ? : RT_TABLE_MAIN;
3256	const struct in6_addr *gw_addr = &cfg->fc_gateway;
3257	struct fib6_result res = {};
3258	int err;
3259
3260	err = ip6_nh_lookup_table(net, cfg, gw_addr, tbid, 0, &res);
3261	if (!err && !(res.fib6_flags & RTF_REJECT) &&
3262	    /* ignore match if it is the default route */
3263	    !ipv6_addr_any(&res.f6i->fib6_dst.addr) &&
3264	    (res.fib6_type != RTN_UNICAST || dev != res.nh->fib_nh_dev)) {
3265		NL_SET_ERR_MSG(extack,
3266			       "Nexthop has invalid gateway or device mismatch");
3267		err = -EINVAL;
3268	}
3269
3270	return err;
3271}
3272
3273static int ip6_route_check_nh(struct net *net,
3274			      struct fib6_config *cfg,
3275			      struct net_device **_dev,
 
3276			      struct inet6_dev **idev)
3277{
3278	const struct in6_addr *gw_addr = &cfg->fc_gateway;
3279	struct net_device *dev = _dev ? *_dev : NULL;
3280	int flags = RT6_LOOKUP_F_IFACE;
3281	struct fib6_result res = {};
3282	int err = -EHOSTUNREACH;
3283
3284	if (cfg->fc_table) {
3285		err = ip6_nh_lookup_table(net, cfg, gw_addr,
3286					  cfg->fc_table, flags, &res);
3287		/* gw_addr can not require a gateway or resolve to a reject
3288		 * route. If a device is given, it must match the result.
3289		 */
3290		if (err || res.fib6_flags & RTF_REJECT ||
3291		    res.nh->fib_nh_gw_family ||
3292		    (dev && dev != res.nh->fib_nh_dev))
3293			err = -EHOSTUNREACH;
3294	}
3295
3296	if (err < 0) {
3297		struct flowi6 fl6 = {
3298			.flowi6_oif = cfg->fc_ifindex,
3299			.daddr = *gw_addr,
3300		};
3301
3302		err = fib6_lookup(net, cfg->fc_ifindex, &fl6, &res, flags);
3303		if (err || res.fib6_flags & RTF_REJECT ||
3304		    res.nh->fib_nh_gw_family)
3305			err = -EHOSTUNREACH;
3306
3307		if (err)
3308			return err;
3309
3310		fib6_select_path(net, &res, &fl6, cfg->fc_ifindex,
3311				 cfg->fc_ifindex != 0, NULL, flags);
3312	}
3313
3314	err = 0;
3315	if (dev) {
3316		if (dev != res.nh->fib_nh_dev)
3317			err = -EHOSTUNREACH;
3318	} else {
3319		*_dev = dev = res.nh->fib_nh_dev;
3320		dev_hold(dev);
3321		*idev = in6_dev_get(dev);
3322	}
3323
3324	return err;
3325}
3326
3327static int ip6_validate_gw(struct net *net, struct fib6_config *cfg,
3328			   struct net_device **_dev, struct inet6_dev **idev,
 
 
3329			   struct netlink_ext_ack *extack)
3330{
3331	const struct in6_addr *gw_addr = &cfg->fc_gateway;
3332	int gwa_type = ipv6_addr_type(gw_addr);
3333	bool skip_dev = gwa_type & IPV6_ADDR_LINKLOCAL ? false : true;
3334	const struct net_device *dev = *_dev;
3335	bool need_addr_check = !dev;
3336	int err = -EINVAL;
3337
3338	/* if gw_addr is local we will fail to detect this in case
3339	 * address is still TENTATIVE (DAD in progress). rt6_lookup()
3340	 * will return already-added prefix route via interface that
3341	 * prefix route was assigned to, which might be non-loopback.
3342	 */
3343	if (dev &&
3344	    ipv6_chk_addr_and_flags(net, gw_addr, dev, skip_dev, 0, 0)) {
3345		NL_SET_ERR_MSG(extack, "Gateway can not be a local address");
3346		goto out;
3347	}
3348
3349	if (gwa_type != (IPV6_ADDR_LINKLOCAL | IPV6_ADDR_UNICAST)) {
3350		/* IPv6 strictly inhibits using not link-local
3351		 * addresses as nexthop address.
3352		 * Otherwise, router will not able to send redirects.
3353		 * It is very good, but in some (rare!) circumstances
3354		 * (SIT, PtP, NBMA NOARP links) it is handy to allow
3355		 * some exceptions. --ANK
3356		 * We allow IPv4-mapped nexthops to support RFC4798-type
3357		 * addressing
3358		 */
3359		if (!(gwa_type & (IPV6_ADDR_UNICAST | IPV6_ADDR_MAPPED))) {
3360			NL_SET_ERR_MSG(extack, "Invalid gateway address");
3361			goto out;
3362		}
3363
3364		rcu_read_lock();
3365
3366		if (cfg->fc_flags & RTNH_F_ONLINK)
3367			err = ip6_route_check_nh_onlink(net, cfg, dev, extack);
3368		else
3369			err = ip6_route_check_nh(net, cfg, _dev, idev);
 
3370
3371		rcu_read_unlock();
3372
3373		if (err)
3374			goto out;
3375	}
3376
3377	/* reload in case device was changed */
3378	dev = *_dev;
3379
3380	err = -EINVAL;
3381	if (!dev) {
3382		NL_SET_ERR_MSG(extack, "Egress device not specified");
3383		goto out;
3384	} else if (dev->flags & IFF_LOOPBACK) {
3385		NL_SET_ERR_MSG(extack,
3386			       "Egress device can not be loopback device for this route");
3387		goto out;
3388	}
3389
3390	/* if we did not check gw_addr above, do so now that the
3391	 * egress device has been resolved.
3392	 */
3393	if (need_addr_check &&
3394	    ipv6_chk_addr_and_flags(net, gw_addr, dev, skip_dev, 0, 0)) {
3395		NL_SET_ERR_MSG(extack, "Gateway can not be a local address");
3396		goto out;
3397	}
3398
3399	err = 0;
3400out:
3401	return err;
3402}
3403
3404static bool fib6_is_reject(u32 flags, struct net_device *dev, int addr_type)
3405{
3406	if ((flags & RTF_REJECT) ||
3407	    (dev && (dev->flags & IFF_LOOPBACK) &&
3408	     !(addr_type & IPV6_ADDR_LOOPBACK) &&
3409	     !(flags & (RTF_ANYCAST | RTF_LOCAL))))
3410		return true;
3411
3412	return false;
3413}
3414
3415int fib6_nh_init(struct net *net, struct fib6_nh *fib6_nh,
3416		 struct fib6_config *cfg, gfp_t gfp_flags,
3417		 struct netlink_ext_ack *extack)
3418{
 
3419	struct net_device *dev = NULL;
3420	struct inet6_dev *idev = NULL;
3421	int addr_type;
3422	int err;
3423
3424	fib6_nh->fib_nh_family = AF_INET6;
3425#ifdef CONFIG_IPV6_ROUTER_PREF
3426	fib6_nh->last_probe = jiffies;
3427#endif
3428	if (cfg->fc_is_fdb) {
3429		fib6_nh->fib_nh_gw6 = cfg->fc_gateway;
3430		fib6_nh->fib_nh_gw_family = AF_INET6;
3431		return 0;
3432	}
3433
3434	err = -ENODEV;
3435	if (cfg->fc_ifindex) {
3436		dev = dev_get_by_index(net, cfg->fc_ifindex);
 
3437		if (!dev)
3438			goto out;
3439		idev = in6_dev_get(dev);
3440		if (!idev)
3441			goto out;
3442	}
3443
3444	if (cfg->fc_flags & RTNH_F_ONLINK) {
3445		if (!dev) {
3446			NL_SET_ERR_MSG(extack,
3447				       "Nexthop device required for onlink");
3448			goto out;
3449		}
3450
3451		if (!(dev->flags & IFF_UP)) {
3452			NL_SET_ERR_MSG(extack, "Nexthop device is not up");
3453			err = -ENETDOWN;
3454			goto out;
3455		}
3456
3457		fib6_nh->fib_nh_flags |= RTNH_F_ONLINK;
3458	}
3459
3460	fib6_nh->fib_nh_weight = 1;
3461
3462	/* We cannot add true routes via loopback here,
3463	 * they would result in kernel looping; promote them to reject routes
3464	 */
3465	addr_type = ipv6_addr_type(&cfg->fc_dst);
3466	if (fib6_is_reject(cfg->fc_flags, dev, addr_type)) {
3467		/* hold loopback dev/idev if we haven't done so. */
3468		if (dev != net->loopback_dev) {
3469			if (dev) {
3470				dev_put(dev);
3471				in6_dev_put(idev);
3472			}
3473			dev = net->loopback_dev;
3474			dev_hold(dev);
3475			idev = in6_dev_get(dev);
3476			if (!idev) {
3477				err = -ENODEV;
3478				goto out;
3479			}
3480		}
3481		goto pcpu_alloc;
3482	}
3483
3484	if (cfg->fc_flags & RTF_GATEWAY) {
3485		err = ip6_validate_gw(net, cfg, &dev, &idev, extack);
 
3486		if (err)
3487			goto out;
3488
3489		fib6_nh->fib_nh_gw6 = cfg->fc_gateway;
3490		fib6_nh->fib_nh_gw_family = AF_INET6;
3491	}
3492
3493	err = -ENODEV;
3494	if (!dev)
3495		goto out;
3496
3497	if (idev->cnf.disable_ipv6) {
3498		NL_SET_ERR_MSG(extack, "IPv6 is disabled on nexthop device");
3499		err = -EACCES;
3500		goto out;
3501	}
3502
3503	if (!(dev->flags & IFF_UP) && !cfg->fc_ignore_dev_down) {
3504		NL_SET_ERR_MSG(extack, "Nexthop device is not up");
3505		err = -ENETDOWN;
3506		goto out;
3507	}
3508
3509	if (!(cfg->fc_flags & (RTF_LOCAL | RTF_ANYCAST)) &&
3510	    !netif_carrier_ok(dev))
3511		fib6_nh->fib_nh_flags |= RTNH_F_LINKDOWN;
3512
3513	err = fib_nh_common_init(net, &fib6_nh->nh_common, cfg->fc_encap,
3514				 cfg->fc_encap_type, cfg, gfp_flags, extack);
3515	if (err)
3516		goto out;
3517
3518pcpu_alloc:
3519	fib6_nh->rt6i_pcpu = alloc_percpu_gfp(struct rt6_info *, gfp_flags);
3520	if (!fib6_nh->rt6i_pcpu) {
3521		err = -ENOMEM;
3522		goto out;
3523	}
3524
3525	fib6_nh->fib_nh_dev = dev;
3526	fib6_nh->fib_nh_oif = dev->ifindex;
3527	err = 0;
3528out:
3529	if (idev)
3530		in6_dev_put(idev);
3531
3532	if (err) {
3533		lwtstate_put(fib6_nh->fib_nh_lws);
3534		fib6_nh->fib_nh_lws = NULL;
3535		if (dev)
3536			dev_put(dev);
3537	}
3538
3539	return err;
3540}
3541
3542void fib6_nh_release(struct fib6_nh *fib6_nh)
3543{
3544	struct rt6_exception_bucket *bucket;
3545
3546	rcu_read_lock();
3547
3548	fib6_nh_flush_exceptions(fib6_nh, NULL);
3549	bucket = fib6_nh_get_excptn_bucket(fib6_nh, NULL);
3550	if (bucket) {
3551		rcu_assign_pointer(fib6_nh->rt6i_exception_bucket, NULL);
3552		kfree(bucket);
3553	}
3554
3555	rcu_read_unlock();
3556
3557	if (fib6_nh->rt6i_pcpu) {
3558		int cpu;
3559
3560		for_each_possible_cpu(cpu) {
3561			struct rt6_info **ppcpu_rt;
3562			struct rt6_info *pcpu_rt;
3563
3564			ppcpu_rt = per_cpu_ptr(fib6_nh->rt6i_pcpu, cpu);
3565			pcpu_rt = *ppcpu_rt;
3566			if (pcpu_rt) {
3567				dst_dev_put(&pcpu_rt->dst);
3568				dst_release(&pcpu_rt->dst);
3569				*ppcpu_rt = NULL;
3570			}
3571		}
3572
3573		free_percpu(fib6_nh->rt6i_pcpu);
3574	}
 
3575
3576	fib_nh_common_release(&fib6_nh->nh_common);
 
 
 
 
 
 
 
 
 
 
 
 
3577}
3578
3579static struct fib6_info *ip6_route_info_create(struct fib6_config *cfg,
3580					      gfp_t gfp_flags,
3581					      struct netlink_ext_ack *extack)
3582{
3583	struct net *net = cfg->fc_nlinfo.nl_net;
3584	struct fib6_info *rt = NULL;
3585	struct nexthop *nh = NULL;
3586	struct fib6_table *table;
3587	struct fib6_nh *fib6_nh;
3588	int err = -EINVAL;
3589	int addr_type;
3590
3591	/* RTF_PCPU is an internal flag; can not be set by userspace */
3592	if (cfg->fc_flags & RTF_PCPU) {
3593		NL_SET_ERR_MSG(extack, "Userspace can not set RTF_PCPU");
3594		goto out;
3595	}
3596
3597	/* RTF_CACHE is an internal flag; can not be set by userspace */
3598	if (cfg->fc_flags & RTF_CACHE) {
3599		NL_SET_ERR_MSG(extack, "Userspace can not set RTF_CACHE");
3600		goto out;
3601	}
3602
3603	if (cfg->fc_type > RTN_MAX) {
3604		NL_SET_ERR_MSG(extack, "Invalid route type");
3605		goto out;
3606	}
3607
3608	if (cfg->fc_dst_len > 128) {
3609		NL_SET_ERR_MSG(extack, "Invalid prefix length");
3610		goto out;
3611	}
3612	if (cfg->fc_src_len > 128) {
3613		NL_SET_ERR_MSG(extack, "Invalid source address length");
3614		goto out;
3615	}
3616#ifndef CONFIG_IPV6_SUBTREES
3617	if (cfg->fc_src_len) {
3618		NL_SET_ERR_MSG(extack,
3619			       "Specifying source address requires IPV6_SUBTREES to be enabled");
3620		goto out;
3621	}
3622#endif
3623	if (cfg->fc_nh_id) {
3624		nh = nexthop_find_by_id(net, cfg->fc_nh_id);
3625		if (!nh) {
3626			NL_SET_ERR_MSG(extack, "Nexthop id does not exist");
3627			goto out;
3628		}
3629		err = fib6_check_nexthop(nh, cfg, extack);
3630		if (err)
3631			goto out;
3632	}
3633
3634	err = -ENOBUFS;
3635	if (cfg->fc_nlinfo.nlh &&
3636	    !(cfg->fc_nlinfo.nlh->nlmsg_flags & NLM_F_CREATE)) {
3637		table = fib6_get_table(net, cfg->fc_table);
3638		if (!table) {
3639			pr_warn("NLM_F_CREATE should be specified when creating new route\n");
3640			table = fib6_new_table(net, cfg->fc_table);
3641		}
3642	} else {
3643		table = fib6_new_table(net, cfg->fc_table);
3644	}
3645
3646	if (!table)
3647		goto out;
3648
3649	err = -ENOMEM;
3650	rt = fib6_info_alloc(gfp_flags, !nh);
3651	if (!rt)
3652		goto out;
3653
3654	rt->fib6_metrics = ip_fib_metrics_init(net, cfg->fc_mx, cfg->fc_mx_len,
3655					       extack);
3656	if (IS_ERR(rt->fib6_metrics)) {
3657		err = PTR_ERR(rt->fib6_metrics);
3658		/* Do not leave garbage there. */
3659		rt->fib6_metrics = (struct dst_metrics *)&dst_default_metrics;
3660		goto out;
3661	}
3662
3663	if (cfg->fc_flags & RTF_ADDRCONF)
3664		rt->dst_nocount = true;
3665
3666	if (cfg->fc_flags & RTF_EXPIRES)
3667		fib6_set_expires(rt, jiffies +
3668				clock_t_to_jiffies(cfg->fc_expires));
3669	else
3670		fib6_clean_expires(rt);
3671
3672	if (cfg->fc_protocol == RTPROT_UNSPEC)
3673		cfg->fc_protocol = RTPROT_BOOT;
3674	rt->fib6_protocol = cfg->fc_protocol;
3675
3676	rt->fib6_table = table;
3677	rt->fib6_metric = cfg->fc_metric;
3678	rt->fib6_type = cfg->fc_type ? : RTN_UNICAST;
3679	rt->fib6_flags = cfg->fc_flags & ~RTF_GATEWAY;
3680
3681	ipv6_addr_prefix(&rt->fib6_dst.addr, &cfg->fc_dst, cfg->fc_dst_len);
3682	rt->fib6_dst.plen = cfg->fc_dst_len;
3683
3684#ifdef CONFIG_IPV6_SUBTREES
3685	ipv6_addr_prefix(&rt->fib6_src.addr, &cfg->fc_src, cfg->fc_src_len);
3686	rt->fib6_src.plen = cfg->fc_src_len;
3687#endif
3688	if (nh) {
3689		if (rt->fib6_src.plen) {
3690			NL_SET_ERR_MSG(extack, "Nexthops can not be used with source routing");
3691			goto out;
3692		}
3693		if (!nexthop_get(nh)) {
3694			NL_SET_ERR_MSG(extack, "Nexthop has been deleted");
3695			goto out;
3696		}
3697		rt->nh = nh;
3698		fib6_nh = nexthop_fib6_nh(rt->nh);
3699	} else {
3700		err = fib6_nh_init(net, rt->fib6_nh, cfg, gfp_flags, extack);
3701		if (err)
3702			goto out;
3703
3704		fib6_nh = rt->fib6_nh;
3705
3706		/* We cannot add true routes via loopback here, they would
3707		 * result in kernel looping; promote them to reject routes
3708		 */
3709		addr_type = ipv6_addr_type(&cfg->fc_dst);
3710		if (fib6_is_reject(cfg->fc_flags, rt->fib6_nh->fib_nh_dev,
3711				   addr_type))
3712			rt->fib6_flags = RTF_REJECT | RTF_NONEXTHOP;
3713	}
3714
3715	if (!ipv6_addr_any(&cfg->fc_prefsrc)) {
3716		struct net_device *dev = fib6_nh->fib_nh_dev;
3717
3718		if (!ipv6_chk_addr(net, &cfg->fc_prefsrc, dev, 0)) {
3719			NL_SET_ERR_MSG(extack, "Invalid source address");
3720			err = -EINVAL;
3721			goto out;
3722		}
3723		rt->fib6_prefsrc.addr = cfg->fc_prefsrc;
3724		rt->fib6_prefsrc.plen = 128;
3725	} else
3726		rt->fib6_prefsrc.plen = 0;
3727
3728	return rt;
3729out:
3730	fib6_info_release(rt);
3731	return ERR_PTR(err);
 
 
 
 
3732}
3733
3734int ip6_route_add(struct fib6_config *cfg, gfp_t gfp_flags,
3735		  struct netlink_ext_ack *extack)
3736{
3737	struct fib6_info *rt;
3738	int err;
3739
3740	rt = ip6_route_info_create(cfg, gfp_flags, extack);
3741	if (IS_ERR(rt))
3742		return PTR_ERR(rt);
3743
3744	err = __ip6_ins_rt(rt, &cfg->fc_nlinfo, extack);
3745	fib6_info_release(rt);
3746
3747	return err;
3748}
3749
3750static int __ip6_del_rt(struct fib6_info *rt, struct nl_info *info)
3751{
3752	struct net *net = info->nl_net;
3753	struct fib6_table *table;
3754	int err;
3755
3756	if (rt == net->ipv6.fib6_null_entry) {
3757		err = -ENOENT;
3758		goto out;
3759	}
3760
3761	table = rt->fib6_table;
3762	spin_lock_bh(&table->tb6_lock);
3763	err = fib6_del(rt, info);
3764	spin_unlock_bh(&table->tb6_lock);
3765
3766out:
3767	fib6_info_release(rt);
3768	return err;
3769}
3770
3771int ip6_del_rt(struct net *net, struct fib6_info *rt, bool skip_notify)
3772{
3773	struct nl_info info = {
3774		.nl_net = net,
3775		.skip_notify = skip_notify
3776	};
3777
3778	return __ip6_del_rt(rt, &info);
3779}
3780
3781static int __ip6_del_rt_siblings(struct fib6_info *rt, struct fib6_config *cfg)
3782{
3783	struct nl_info *info = &cfg->fc_nlinfo;
3784	struct net *net = info->nl_net;
3785	struct sk_buff *skb = NULL;
3786	struct fib6_table *table;
3787	int err = -ENOENT;
3788
3789	if (rt == net->ipv6.fib6_null_entry)
3790		goto out_put;
3791	table = rt->fib6_table;
3792	spin_lock_bh(&table->tb6_lock);
3793
3794	if (rt->fib6_nsiblings && cfg->fc_delete_all_nh) {
3795		struct fib6_info *sibling, *next_sibling;
3796		struct fib6_node *fn;
3797
3798		/* prefer to send a single notification with all hops */
3799		skb = nlmsg_new(rt6_nlmsg_size(rt), gfp_any());
3800		if (skb) {
3801			u32 seq = info->nlh ? info->nlh->nlmsg_seq : 0;
3802
3803			if (rt6_fill_node(net, skb, rt, NULL,
3804					  NULL, NULL, 0, RTM_DELROUTE,
3805					  info->portid, seq, 0) < 0) {
3806				kfree_skb(skb);
3807				skb = NULL;
3808			} else
3809				info->skip_notify = 1;
3810		}
3811
3812		/* 'rt' points to the first sibling route. If it is not the
3813		 * leaf, then we do not need to send a notification. Otherwise,
3814		 * we need to check if the last sibling has a next route or not
3815		 * and emit a replace or delete notification, respectively.
3816		 */
3817		info->skip_notify_kernel = 1;
3818		fn = rcu_dereference_protected(rt->fib6_node,
3819					    lockdep_is_held(&table->tb6_lock));
3820		if (rcu_access_pointer(fn->leaf) == rt) {
3821			struct fib6_info *last_sibling, *replace_rt;
3822
3823			last_sibling = list_last_entry(&rt->fib6_siblings,
3824						       struct fib6_info,
3825						       fib6_siblings);
3826			replace_rt = rcu_dereference_protected(
3827					    last_sibling->fib6_next,
3828					    lockdep_is_held(&table->tb6_lock));
3829			if (replace_rt)
3830				call_fib6_entry_notifiers_replace(net,
3831								  replace_rt);
3832			else
3833				call_fib6_multipath_entry_notifiers(net,
3834						       FIB_EVENT_ENTRY_DEL,
3835						       rt, rt->fib6_nsiblings,
3836						       NULL);
3837		}
3838		list_for_each_entry_safe(sibling, next_sibling,
3839					 &rt->fib6_siblings,
3840					 fib6_siblings) {
3841			err = fib6_del(sibling, info);
3842			if (err)
3843				goto out_unlock;
3844		}
3845	}
3846
3847	err = fib6_del(rt, info);
3848out_unlock:
3849	spin_unlock_bh(&table->tb6_lock);
3850out_put:
3851	fib6_info_release(rt);
3852
3853	if (skb) {
3854		rtnl_notify(skb, net, info->portid, RTNLGRP_IPV6_ROUTE,
3855			    info->nlh, gfp_any());
3856	}
3857	return err;
3858}
3859
3860static int __ip6_del_cached_rt(struct rt6_info *rt, struct fib6_config *cfg)
3861{
3862	int rc = -ESRCH;
3863
3864	if (cfg->fc_ifindex && rt->dst.dev->ifindex != cfg->fc_ifindex)
3865		goto out;
3866
3867	if (cfg->fc_flags & RTF_GATEWAY &&
3868	    !ipv6_addr_equal(&cfg->fc_gateway, &rt->rt6i_gateway))
3869		goto out;
3870
3871	rc = rt6_remove_exception_rt(rt);
3872out:
3873	return rc;
3874}
3875
3876static int ip6_del_cached_rt(struct fib6_config *cfg, struct fib6_info *rt,
3877			     struct fib6_nh *nh)
3878{
3879	struct fib6_result res = {
3880		.f6i = rt,
3881		.nh = nh,
3882	};
3883	struct rt6_info *rt_cache;
3884
3885	rt_cache = rt6_find_cached_rt(&res, &cfg->fc_dst, &cfg->fc_src);
3886	if (rt_cache)
3887		return __ip6_del_cached_rt(rt_cache, cfg);
3888
3889	return 0;
3890}
3891
3892struct fib6_nh_del_cached_rt_arg {
3893	struct fib6_config *cfg;
3894	struct fib6_info *f6i;
3895};
3896
3897static int fib6_nh_del_cached_rt(struct fib6_nh *nh, void *_arg)
3898{
3899	struct fib6_nh_del_cached_rt_arg *arg = _arg;
3900	int rc;
3901
3902	rc = ip6_del_cached_rt(arg->cfg, arg->f6i, nh);
3903	return rc != -ESRCH ? rc : 0;
3904}
3905
3906static int ip6_del_cached_rt_nh(struct fib6_config *cfg, struct fib6_info *f6i)
3907{
3908	struct fib6_nh_del_cached_rt_arg arg = {
3909		.cfg = cfg,
3910		.f6i = f6i
3911	};
3912
3913	return nexthop_for_each_fib6_nh(f6i->nh, fib6_nh_del_cached_rt, &arg);
3914}
3915
3916static int ip6_route_del(struct fib6_config *cfg,
3917			 struct netlink_ext_ack *extack)
3918{
3919	struct fib6_table *table;
3920	struct fib6_info *rt;
3921	struct fib6_node *fn;
3922	int err = -ESRCH;
3923
3924	table = fib6_get_table(cfg->fc_nlinfo.nl_net, cfg->fc_table);
3925	if (!table) {
3926		NL_SET_ERR_MSG(extack, "FIB table does not exist");
3927		return err;
3928	}
3929
3930	rcu_read_lock();
3931
3932	fn = fib6_locate(&table->tb6_root,
3933			 &cfg->fc_dst, cfg->fc_dst_len,
3934			 &cfg->fc_src, cfg->fc_src_len,
3935			 !(cfg->fc_flags & RTF_CACHE));
3936
3937	if (fn) {
3938		for_each_fib6_node_rt_rcu(fn) {
3939			struct fib6_nh *nh;
3940
3941			if (rt->nh && cfg->fc_nh_id &&
3942			    rt->nh->id != cfg->fc_nh_id)
3943				continue;
3944
3945			if (cfg->fc_flags & RTF_CACHE) {
3946				int rc = 0;
3947
3948				if (rt->nh) {
3949					rc = ip6_del_cached_rt_nh(cfg, rt);
3950				} else if (cfg->fc_nh_id) {
3951					continue;
3952				} else {
3953					nh = rt->fib6_nh;
3954					rc = ip6_del_cached_rt(cfg, rt, nh);
3955				}
3956				if (rc != -ESRCH) {
3957					rcu_read_unlock();
3958					return rc;
3959				}
3960				continue;
3961			}
3962
3963			if (cfg->fc_metric && cfg->fc_metric != rt->fib6_metric)
3964				continue;
3965			if (cfg->fc_protocol &&
3966			    cfg->fc_protocol != rt->fib6_protocol)
3967				continue;
3968
3969			if (rt->nh) {
3970				if (!fib6_info_hold_safe(rt))
3971					continue;
3972				rcu_read_unlock();
3973
3974				return __ip6_del_rt(rt, &cfg->fc_nlinfo);
3975			}
3976			if (cfg->fc_nh_id)
3977				continue;
3978
3979			nh = rt->fib6_nh;
3980			if (cfg->fc_ifindex &&
3981			    (!nh->fib_nh_dev ||
3982			     nh->fib_nh_dev->ifindex != cfg->fc_ifindex))
3983				continue;
3984			if (cfg->fc_flags & RTF_GATEWAY &&
3985			    !ipv6_addr_equal(&cfg->fc_gateway, &nh->fib_nh_gw6))
3986				continue;
3987			if (!fib6_info_hold_safe(rt))
3988				continue;
3989			rcu_read_unlock();
3990
3991			/* if gateway was specified only delete the one hop */
3992			if (cfg->fc_flags & RTF_GATEWAY)
3993				return __ip6_del_rt(rt, &cfg->fc_nlinfo);
3994
3995			return __ip6_del_rt_siblings(rt, cfg);
3996		}
3997	}
3998	rcu_read_unlock();
3999
4000	return err;
4001}
4002
4003static void rt6_do_redirect(struct dst_entry *dst, struct sock *sk, struct sk_buff *skb)
4004{
4005	struct netevent_redirect netevent;
4006	struct rt6_info *rt, *nrt = NULL;
4007	struct fib6_result res = {};
4008	struct ndisc_options ndopts;
4009	struct inet6_dev *in6_dev;
4010	struct neighbour *neigh;
4011	struct rd_msg *msg;
4012	int optlen, on_link;
4013	u8 *lladdr;
4014
4015	optlen = skb_tail_pointer(skb) - skb_transport_header(skb);
4016	optlen -= sizeof(*msg);
4017
4018	if (optlen < 0) {
4019		net_dbg_ratelimited("rt6_do_redirect: packet too short\n");
4020		return;
4021	}
4022
4023	msg = (struct rd_msg *)icmp6_hdr(skb);
4024
4025	if (ipv6_addr_is_multicast(&msg->dest)) {
4026		net_dbg_ratelimited("rt6_do_redirect: destination address is multicast\n");
4027		return;
4028	}
4029
4030	on_link = 0;
4031	if (ipv6_addr_equal(&msg->dest, &msg->target)) {
4032		on_link = 1;
4033	} else if (ipv6_addr_type(&msg->target) !=
4034		   (IPV6_ADDR_UNICAST|IPV6_ADDR_LINKLOCAL)) {
4035		net_dbg_ratelimited("rt6_do_redirect: target address is not link-local unicast\n");
4036		return;
4037	}
4038
4039	in6_dev = __in6_dev_get(skb->dev);
4040	if (!in6_dev)
4041		return;
4042	if (in6_dev->cnf.forwarding || !in6_dev->cnf.accept_redirects)
 
4043		return;
4044
4045	/* RFC2461 8.1:
4046	 *	The IP source address of the Redirect MUST be the same as the current
4047	 *	first-hop router for the specified ICMP Destination Address.
4048	 */
4049
4050	if (!ndisc_parse_options(skb->dev, msg->opt, optlen, &ndopts)) {
4051		net_dbg_ratelimited("rt6_redirect: invalid ND options\n");
4052		return;
4053	}
4054
4055	lladdr = NULL;
4056	if (ndopts.nd_opts_tgt_lladdr) {
4057		lladdr = ndisc_opt_addr_data(ndopts.nd_opts_tgt_lladdr,
4058					     skb->dev);
4059		if (!lladdr) {
4060			net_dbg_ratelimited("rt6_redirect: invalid link-layer address length\n");
4061			return;
4062		}
4063	}
4064
4065	rt = (struct rt6_info *) dst;
4066	if (rt->rt6i_flags & RTF_REJECT) {
4067		net_dbg_ratelimited("rt6_redirect: source isn't a valid nexthop for redirect target\n");
4068		return;
4069	}
4070
4071	/* Redirect received -> path was valid.
4072	 * Look, redirects are sent only in response to data packets,
4073	 * so that this nexthop apparently is reachable. --ANK
4074	 */
4075	dst_confirm_neigh(&rt->dst, &ipv6_hdr(skb)->saddr);
4076
4077	neigh = __neigh_lookup(&nd_tbl, &msg->target, skb->dev, 1);
4078	if (!neigh)
4079		return;
4080
4081	/*
4082	 *	We have finally decided to accept it.
4083	 */
4084
4085	ndisc_update(skb->dev, neigh, lladdr, NUD_STALE,
4086		     NEIGH_UPDATE_F_WEAK_OVERRIDE|
4087		     NEIGH_UPDATE_F_OVERRIDE|
4088		     (on_link ? 0 : (NEIGH_UPDATE_F_OVERRIDE_ISROUTER|
4089				     NEIGH_UPDATE_F_ISROUTER)),
4090		     NDISC_REDIRECT, &ndopts);
4091
4092	rcu_read_lock();
4093	res.f6i = rcu_dereference(rt->from);
4094	if (!res.f6i)
4095		goto out;
4096
4097	if (res.f6i->nh) {
4098		struct fib6_nh_match_arg arg = {
4099			.dev = dst->dev,
4100			.gw = &rt->rt6i_gateway,
4101		};
4102
4103		nexthop_for_each_fib6_nh(res.f6i->nh,
4104					 fib6_nh_find_match, &arg);
4105
4106		/* fib6_info uses a nexthop that does not have fib6_nh
4107		 * using the dst->dev. Should be impossible
4108		 */
4109		if (!arg.match)
4110			goto out;
4111		res.nh = arg.match;
4112	} else {
4113		res.nh = res.f6i->fib6_nh;
4114	}
4115
4116	res.fib6_flags = res.f6i->fib6_flags;
4117	res.fib6_type = res.f6i->fib6_type;
4118	nrt = ip6_rt_cache_alloc(&res, &msg->dest, NULL);
4119	if (!nrt)
4120		goto out;
4121
4122	nrt->rt6i_flags = RTF_GATEWAY|RTF_UP|RTF_DYNAMIC|RTF_CACHE;
4123	if (on_link)
4124		nrt->rt6i_flags &= ~RTF_GATEWAY;
4125
4126	nrt->rt6i_gateway = *(struct in6_addr *)neigh->primary_key;
4127
4128	/* rt6_insert_exception() will take care of duplicated exceptions */
4129	if (rt6_insert_exception(nrt, &res)) {
4130		dst_release_immediate(&nrt->dst);
4131		goto out;
4132	}
4133
4134	netevent.old = &rt->dst;
4135	netevent.new = &nrt->dst;
4136	netevent.daddr = &msg->dest;
4137	netevent.neigh = neigh;
4138	call_netevent_notifiers(NETEVENT_REDIRECT, &netevent);
4139
4140out:
4141	rcu_read_unlock();
4142	neigh_release(neigh);
4143}
4144
4145#ifdef CONFIG_IPV6_ROUTE_INFO
4146static struct fib6_info *rt6_get_route_info(struct net *net,
4147					   const struct in6_addr *prefix, int prefixlen,
4148					   const struct in6_addr *gwaddr,
4149					   struct net_device *dev)
4150{
4151	u32 tb_id = l3mdev_fib_table(dev) ? : RT6_TABLE_INFO;
4152	int ifindex = dev->ifindex;
4153	struct fib6_node *fn;
4154	struct fib6_info *rt = NULL;
4155	struct fib6_table *table;
4156
4157	table = fib6_get_table(net, tb_id);
4158	if (!table)
4159		return NULL;
4160
4161	rcu_read_lock();
4162	fn = fib6_locate(&table->tb6_root, prefix, prefixlen, NULL, 0, true);
4163	if (!fn)
4164		goto out;
4165
4166	for_each_fib6_node_rt_rcu(fn) {
4167		/* these routes do not use nexthops */
4168		if (rt->nh)
4169			continue;
4170		if (rt->fib6_nh->fib_nh_dev->ifindex != ifindex)
4171			continue;
4172		if (!(rt->fib6_flags & RTF_ROUTEINFO) ||
4173		    !rt->fib6_nh->fib_nh_gw_family)
4174			continue;
4175		if (!ipv6_addr_equal(&rt->fib6_nh->fib_nh_gw6, gwaddr))
4176			continue;
4177		if (!fib6_info_hold_safe(rt))
4178			continue;
4179		break;
4180	}
4181out:
4182	rcu_read_unlock();
4183	return rt;
4184}
4185
4186static struct fib6_info *rt6_add_route_info(struct net *net,
4187					   const struct in6_addr *prefix, int prefixlen,
4188					   const struct in6_addr *gwaddr,
4189					   struct net_device *dev,
4190					   unsigned int pref)
4191{
4192	struct fib6_config cfg = {
4193		.fc_metric	= IP6_RT_PRIO_USER,
4194		.fc_ifindex	= dev->ifindex,
4195		.fc_dst_len	= prefixlen,
4196		.fc_flags	= RTF_GATEWAY | RTF_ADDRCONF | RTF_ROUTEINFO |
4197				  RTF_UP | RTF_PREF(pref),
4198		.fc_protocol = RTPROT_RA,
4199		.fc_type = RTN_UNICAST,
4200		.fc_nlinfo.portid = 0,
4201		.fc_nlinfo.nlh = NULL,
4202		.fc_nlinfo.nl_net = net,
4203	};
4204
4205	cfg.fc_table = l3mdev_fib_table(dev) ? : RT6_TABLE_INFO;
4206	cfg.fc_dst = *prefix;
4207	cfg.fc_gateway = *gwaddr;
4208
4209	/* We should treat it as a default route if prefix length is 0. */
4210	if (!prefixlen)
4211		cfg.fc_flags |= RTF_DEFAULT;
4212
4213	ip6_route_add(&cfg, GFP_ATOMIC, NULL);
4214
4215	return rt6_get_route_info(net, prefix, prefixlen, gwaddr, dev);
4216}
4217#endif
4218
4219struct fib6_info *rt6_get_dflt_router(struct net *net,
4220				     const struct in6_addr *addr,
4221				     struct net_device *dev)
4222{
4223	u32 tb_id = l3mdev_fib_table(dev) ? : RT6_TABLE_DFLT;
4224	struct fib6_info *rt;
4225	struct fib6_table *table;
4226
4227	table = fib6_get_table(net, tb_id);
4228	if (!table)
4229		return NULL;
4230
4231	rcu_read_lock();
4232	for_each_fib6_node_rt_rcu(&table->tb6_root) {
4233		struct fib6_nh *nh;
4234
4235		/* RA routes do not use nexthops */
4236		if (rt->nh)
4237			continue;
4238
4239		nh = rt->fib6_nh;
4240		if (dev == nh->fib_nh_dev &&
4241		    ((rt->fib6_flags & (RTF_ADDRCONF | RTF_DEFAULT)) == (RTF_ADDRCONF | RTF_DEFAULT)) &&
4242		    ipv6_addr_equal(&nh->fib_nh_gw6, addr))
4243			break;
4244	}
4245	if (rt && !fib6_info_hold_safe(rt))
4246		rt = NULL;
4247	rcu_read_unlock();
4248	return rt;
4249}
4250
4251struct fib6_info *rt6_add_dflt_router(struct net *net,
4252				     const struct in6_addr *gwaddr,
4253				     struct net_device *dev,
4254				     unsigned int pref)
 
 
4255{
4256	struct fib6_config cfg = {
4257		.fc_table	= l3mdev_fib_table(dev) ? : RT6_TABLE_DFLT,
4258		.fc_metric	= IP6_RT_PRIO_USER,
4259		.fc_ifindex	= dev->ifindex,
4260		.fc_flags	= RTF_GATEWAY | RTF_ADDRCONF | RTF_DEFAULT |
4261				  RTF_UP | RTF_EXPIRES | RTF_PREF(pref),
4262		.fc_protocol = RTPROT_RA,
4263		.fc_type = RTN_UNICAST,
4264		.fc_nlinfo.portid = 0,
4265		.fc_nlinfo.nlh = NULL,
4266		.fc_nlinfo.nl_net = net,
 
4267	};
4268
4269	cfg.fc_gateway = *gwaddr;
4270
4271	if (!ip6_route_add(&cfg, GFP_ATOMIC, NULL)) {
4272		struct fib6_table *table;
4273
4274		table = fib6_get_table(dev_net(dev), cfg.fc_table);
4275		if (table)
4276			table->flags |= RT6_TABLE_HAS_DFLT_ROUTER;
4277	}
4278
4279	return rt6_get_dflt_router(net, gwaddr, dev);
4280}
4281
4282static void __rt6_purge_dflt_routers(struct net *net,
4283				     struct fib6_table *table)
4284{
4285	struct fib6_info *rt;
4286
4287restart:
4288	rcu_read_lock();
4289	for_each_fib6_node_rt_rcu(&table->tb6_root) {
4290		struct net_device *dev = fib6_info_nh_dev(rt);
4291		struct inet6_dev *idev = dev ? __in6_dev_get(dev) : NULL;
4292
4293		if (rt->fib6_flags & (RTF_DEFAULT | RTF_ADDRCONF) &&
4294		    (!idev || idev->cnf.accept_ra != 2) &&
4295		    fib6_info_hold_safe(rt)) {
4296			rcu_read_unlock();
4297			ip6_del_rt(net, rt, false);
4298			goto restart;
4299		}
4300	}
4301	rcu_read_unlock();
4302
4303	table->flags &= ~RT6_TABLE_HAS_DFLT_ROUTER;
4304}
4305
4306void rt6_purge_dflt_routers(struct net *net)
4307{
4308	struct fib6_table *table;
4309	struct hlist_head *head;
4310	unsigned int h;
4311
4312	rcu_read_lock();
4313
4314	for (h = 0; h < FIB6_TABLE_HASHSZ; h++) {
4315		head = &net->ipv6.fib_table_hash[h];
4316		hlist_for_each_entry_rcu(table, head, tb6_hlist) {
4317			if (table->flags & RT6_TABLE_HAS_DFLT_ROUTER)
4318				__rt6_purge_dflt_routers(net, table);
4319		}
4320	}
4321
4322	rcu_read_unlock();
4323}
4324
4325static void rtmsg_to_fib6_config(struct net *net,
4326				 struct in6_rtmsg *rtmsg,
4327				 struct fib6_config *cfg)
4328{
4329	*cfg = (struct fib6_config){
4330		.fc_table = l3mdev_fib_table_by_index(net, rtmsg->rtmsg_ifindex) ?
4331			 : RT6_TABLE_MAIN,
4332		.fc_ifindex = rtmsg->rtmsg_ifindex,
4333		.fc_metric = rtmsg->rtmsg_metric ? : IP6_RT_PRIO_USER,
4334		.fc_expires = rtmsg->rtmsg_info,
4335		.fc_dst_len = rtmsg->rtmsg_dst_len,
4336		.fc_src_len = rtmsg->rtmsg_src_len,
4337		.fc_flags = rtmsg->rtmsg_flags,
4338		.fc_type = rtmsg->rtmsg_type,
4339
4340		.fc_nlinfo.nl_net = net,
4341
4342		.fc_dst = rtmsg->rtmsg_dst,
4343		.fc_src = rtmsg->rtmsg_src,
4344		.fc_gateway = rtmsg->rtmsg_gateway,
4345	};
4346}
4347
4348int ipv6_route_ioctl(struct net *net, unsigned int cmd, struct in6_rtmsg *rtmsg)
4349{
4350	struct fib6_config cfg;
4351	int err;
4352
4353	if (cmd != SIOCADDRT && cmd != SIOCDELRT)
4354		return -EINVAL;
4355	if (!ns_capable(net->user_ns, CAP_NET_ADMIN))
4356		return -EPERM;
4357
4358	rtmsg_to_fib6_config(net, rtmsg, &cfg);
4359
4360	rtnl_lock();
4361	switch (cmd) {
4362	case SIOCADDRT:
4363		err = ip6_route_add(&cfg, GFP_KERNEL, NULL);
4364		break;
4365	case SIOCDELRT:
4366		err = ip6_route_del(&cfg, NULL);
4367		break;
4368	}
4369	rtnl_unlock();
4370	return err;
4371}
4372
4373/*
4374 *	Drop the packet on the floor
4375 */
4376
4377static int ip6_pkt_drop(struct sk_buff *skb, u8 code, int ipstats_mib_noroutes)
4378{
4379	struct dst_entry *dst = skb_dst(skb);
4380	struct net *net = dev_net(dst->dev);
4381	struct inet6_dev *idev;
 
4382	int type;
4383
4384	if (netif_is_l3_master(skb->dev) &&
4385	    dst->dev == net->loopback_dev)
4386		idev = __in6_dev_get_safely(dev_get_by_index_rcu(net, IP6CB(skb)->iif));
4387	else
4388		idev = ip6_dst_idev(dst);
4389
4390	switch (ipstats_mib_noroutes) {
4391	case IPSTATS_MIB_INNOROUTES:
4392		type = ipv6_addr_type(&ipv6_hdr(skb)->daddr);
4393		if (type == IPV6_ADDR_ANY) {
 
4394			IP6_INC_STATS(net, idev, IPSTATS_MIB_INADDRERRORS);
4395			break;
4396		}
 
4397		fallthrough;
4398	case IPSTATS_MIB_OUTNOROUTES:
 
4399		IP6_INC_STATS(net, idev, ipstats_mib_noroutes);
4400		break;
4401	}
4402
4403	/* Start over by dropping the dst for l3mdev case */
4404	if (netif_is_l3_master(skb->dev))
4405		skb_dst_drop(skb);
4406
4407	icmpv6_send(skb, ICMPV6_DEST_UNREACH, code, 0);
4408	kfree_skb(skb);
4409	return 0;
4410}
4411
4412static int ip6_pkt_discard(struct sk_buff *skb)
4413{
4414	return ip6_pkt_drop(skb, ICMPV6_NOROUTE, IPSTATS_MIB_INNOROUTES);
4415}
4416
4417static int ip6_pkt_discard_out(struct net *net, struct sock *sk, struct sk_buff *skb)
4418{
4419	skb->dev = skb_dst(skb)->dev;
4420	return ip6_pkt_drop(skb, ICMPV6_NOROUTE, IPSTATS_MIB_OUTNOROUTES);
4421}
4422
4423static int ip6_pkt_prohibit(struct sk_buff *skb)
4424{
4425	return ip6_pkt_drop(skb, ICMPV6_ADM_PROHIBITED, IPSTATS_MIB_INNOROUTES);
4426}
4427
4428static int ip6_pkt_prohibit_out(struct net *net, struct sock *sk, struct sk_buff *skb)
4429{
4430	skb->dev = skb_dst(skb)->dev;
4431	return ip6_pkt_drop(skb, ICMPV6_ADM_PROHIBITED, IPSTATS_MIB_OUTNOROUTES);
4432}
4433
4434/*
4435 *	Allocate a dst for local (unicast / anycast) address.
4436 */
4437
4438struct fib6_info *addrconf_f6i_alloc(struct net *net,
4439				     struct inet6_dev *idev,
4440				     const struct in6_addr *addr,
4441				     bool anycast, gfp_t gfp_flags)
 
4442{
4443	struct fib6_config cfg = {
4444		.fc_table = l3mdev_fib_table(idev->dev) ? : RT6_TABLE_LOCAL,
4445		.fc_ifindex = idev->dev->ifindex,
4446		.fc_flags = RTF_UP | RTF_NONEXTHOP,
4447		.fc_dst = *addr,
4448		.fc_dst_len = 128,
4449		.fc_protocol = RTPROT_KERNEL,
4450		.fc_nlinfo.nl_net = net,
4451		.fc_ignore_dev_down = true,
4452	};
4453	struct fib6_info *f6i;
4454
4455	if (anycast) {
4456		cfg.fc_type = RTN_ANYCAST;
4457		cfg.fc_flags |= RTF_ANYCAST;
4458	} else {
4459		cfg.fc_type = RTN_LOCAL;
4460		cfg.fc_flags |= RTF_LOCAL;
4461	}
4462
4463	f6i = ip6_route_info_create(&cfg, gfp_flags, NULL);
4464	if (!IS_ERR(f6i))
4465		f6i->dst_nocount = true;
 
 
 
 
 
 
 
4466	return f6i;
4467}
4468
4469/* remove deleted ip from prefsrc entries */
4470struct arg_dev_net_ip {
4471	struct net_device *dev;
4472	struct net *net;
4473	struct in6_addr *addr;
4474};
4475
4476static int fib6_remove_prefsrc(struct fib6_info *rt, void *arg)
4477{
4478	struct net_device *dev = ((struct arg_dev_net_ip *)arg)->dev;
4479	struct net *net = ((struct arg_dev_net_ip *)arg)->net;
4480	struct in6_addr *addr = ((struct arg_dev_net_ip *)arg)->addr;
4481
4482	if (!rt->nh &&
4483	    ((void *)rt->fib6_nh->fib_nh_dev == dev || !dev) &&
4484	    rt != net->ipv6.fib6_null_entry &&
4485	    ipv6_addr_equal(addr, &rt->fib6_prefsrc.addr)) {
 
4486		spin_lock_bh(&rt6_exception_lock);
4487		/* remove prefsrc entry */
4488		rt->fib6_prefsrc.plen = 0;
4489		spin_unlock_bh(&rt6_exception_lock);
4490	}
4491	return 0;
4492}
4493
4494void rt6_remove_prefsrc(struct inet6_ifaddr *ifp)
4495{
4496	struct net *net = dev_net(ifp->idev->dev);
4497	struct arg_dev_net_ip adni = {
4498		.dev = ifp->idev->dev,
4499		.net = net,
4500		.addr = &ifp->addr,
4501	};
4502	fib6_clean_all(net, fib6_remove_prefsrc, &adni);
4503}
4504
4505#define RTF_RA_ROUTER		(RTF_ADDRCONF | RTF_DEFAULT)
4506
4507/* Remove routers and update dst entries when gateway turn into host. */
4508static int fib6_clean_tohost(struct fib6_info *rt, void *arg)
4509{
4510	struct in6_addr *gateway = (struct in6_addr *)arg;
4511	struct fib6_nh *nh;
4512
4513	/* RA routes do not use nexthops */
4514	if (rt->nh)
4515		return 0;
4516
4517	nh = rt->fib6_nh;
4518	if (((rt->fib6_flags & RTF_RA_ROUTER) == RTF_RA_ROUTER) &&
4519	    nh->fib_nh_gw_family && ipv6_addr_equal(gateway, &nh->fib_nh_gw6))
4520		return -1;
4521
4522	/* Further clean up cached routes in exception table.
4523	 * This is needed because cached route may have a different
4524	 * gateway than its 'parent' in the case of an ip redirect.
4525	 */
4526	fib6_nh_exceptions_clean_tohost(nh, gateway);
4527
4528	return 0;
4529}
4530
4531void rt6_clean_tohost(struct net *net, struct in6_addr *gateway)
4532{
4533	fib6_clean_all(net, fib6_clean_tohost, gateway);
4534}
4535
4536struct arg_netdev_event {
4537	const struct net_device *dev;
4538	union {
4539		unsigned char nh_flags;
4540		unsigned long event;
4541	};
4542};
4543
4544static struct fib6_info *rt6_multipath_first_sibling(const struct fib6_info *rt)
4545{
4546	struct fib6_info *iter;
4547	struct fib6_node *fn;
4548
4549	fn = rcu_dereference_protected(rt->fib6_node,
4550			lockdep_is_held(&rt->fib6_table->tb6_lock));
4551	iter = rcu_dereference_protected(fn->leaf,
4552			lockdep_is_held(&rt->fib6_table->tb6_lock));
4553	while (iter) {
4554		if (iter->fib6_metric == rt->fib6_metric &&
4555		    rt6_qualify_for_ecmp(iter))
4556			return iter;
4557		iter = rcu_dereference_protected(iter->fib6_next,
4558				lockdep_is_held(&rt->fib6_table->tb6_lock));
4559	}
4560
4561	return NULL;
4562}
4563
4564/* only called for fib entries with builtin fib6_nh */
4565static bool rt6_is_dead(const struct fib6_info *rt)
4566{
4567	if (rt->fib6_nh->fib_nh_flags & RTNH_F_DEAD ||
4568	    (rt->fib6_nh->fib_nh_flags & RTNH_F_LINKDOWN &&
4569	     ip6_ignore_linkdown(rt->fib6_nh->fib_nh_dev)))
4570		return true;
4571
4572	return false;
4573}
4574
4575static int rt6_multipath_total_weight(const struct fib6_info *rt)
4576{
4577	struct fib6_info *iter;
4578	int total = 0;
4579
4580	if (!rt6_is_dead(rt))
4581		total += rt->fib6_nh->fib_nh_weight;
4582
4583	list_for_each_entry(iter, &rt->fib6_siblings, fib6_siblings) {
4584		if (!rt6_is_dead(iter))
4585			total += iter->fib6_nh->fib_nh_weight;
4586	}
4587
4588	return total;
4589}
4590
4591static void rt6_upper_bound_set(struct fib6_info *rt, int *weight, int total)
4592{
4593	int upper_bound = -1;
4594
4595	if (!rt6_is_dead(rt)) {
4596		*weight += rt->fib6_nh->fib_nh_weight;
4597		upper_bound = DIV_ROUND_CLOSEST_ULL((u64) (*weight) << 31,
4598						    total) - 1;
4599	}
4600	atomic_set(&rt->fib6_nh->fib_nh_upper_bound, upper_bound);
4601}
4602
4603static void rt6_multipath_upper_bound_set(struct fib6_info *rt, int total)
4604{
4605	struct fib6_info *iter;
4606	int weight = 0;
4607
4608	rt6_upper_bound_set(rt, &weight, total);
4609
4610	list_for_each_entry(iter, &rt->fib6_siblings, fib6_siblings)
4611		rt6_upper_bound_set(iter, &weight, total);
4612}
4613
4614void rt6_multipath_rebalance(struct fib6_info *rt)
4615{
4616	struct fib6_info *first;
4617	int total;
4618
4619	/* In case the entire multipath route was marked for flushing,
4620	 * then there is no need to rebalance upon the removal of every
4621	 * sibling route.
4622	 */
4623	if (!rt->fib6_nsiblings || rt->should_flush)
4624		return;
4625
4626	/* During lookup routes are evaluated in order, so we need to
4627	 * make sure upper bounds are assigned from the first sibling
4628	 * onwards.
4629	 */
4630	first = rt6_multipath_first_sibling(rt);
4631	if (WARN_ON_ONCE(!first))
4632		return;
4633
4634	total = rt6_multipath_total_weight(first);
4635	rt6_multipath_upper_bound_set(first, total);
4636}
4637
4638static int fib6_ifup(struct fib6_info *rt, void *p_arg)
4639{
4640	const struct arg_netdev_event *arg = p_arg;
4641	struct net *net = dev_net(arg->dev);
4642
4643	if (rt != net->ipv6.fib6_null_entry && !rt->nh &&
4644	    rt->fib6_nh->fib_nh_dev == arg->dev) {
4645		rt->fib6_nh->fib_nh_flags &= ~arg->nh_flags;
4646		fib6_update_sernum_upto_root(net, rt);
4647		rt6_multipath_rebalance(rt);
4648	}
4649
4650	return 0;
4651}
4652
4653void rt6_sync_up(struct net_device *dev, unsigned char nh_flags)
4654{
4655	struct arg_netdev_event arg = {
4656		.dev = dev,
4657		{
4658			.nh_flags = nh_flags,
4659		},
4660	};
4661
4662	if (nh_flags & RTNH_F_DEAD && netif_carrier_ok(dev))
4663		arg.nh_flags |= RTNH_F_LINKDOWN;
4664
4665	fib6_clean_all(dev_net(dev), fib6_ifup, &arg);
4666}
4667
4668/* only called for fib entries with inline fib6_nh */
4669static bool rt6_multipath_uses_dev(const struct fib6_info *rt,
4670				   const struct net_device *dev)
4671{
4672	struct fib6_info *iter;
4673
4674	if (rt->fib6_nh->fib_nh_dev == dev)
4675		return true;
4676	list_for_each_entry(iter, &rt->fib6_siblings, fib6_siblings)
4677		if (iter->fib6_nh->fib_nh_dev == dev)
4678			return true;
4679
4680	return false;
4681}
4682
4683static void rt6_multipath_flush(struct fib6_info *rt)
4684{
4685	struct fib6_info *iter;
4686
4687	rt->should_flush = 1;
4688	list_for_each_entry(iter, &rt->fib6_siblings, fib6_siblings)
4689		iter->should_flush = 1;
4690}
4691
4692static unsigned int rt6_multipath_dead_count(const struct fib6_info *rt,
4693					     const struct net_device *down_dev)
4694{
4695	struct fib6_info *iter;
4696	unsigned int dead = 0;
4697
4698	if (rt->fib6_nh->fib_nh_dev == down_dev ||
4699	    rt->fib6_nh->fib_nh_flags & RTNH_F_DEAD)
4700		dead++;
4701	list_for_each_entry(iter, &rt->fib6_siblings, fib6_siblings)
4702		if (iter->fib6_nh->fib_nh_dev == down_dev ||
4703		    iter->fib6_nh->fib_nh_flags & RTNH_F_DEAD)
4704			dead++;
4705
4706	return dead;
4707}
4708
4709static void rt6_multipath_nh_flags_set(struct fib6_info *rt,
4710				       const struct net_device *dev,
4711				       unsigned char nh_flags)
4712{
4713	struct fib6_info *iter;
4714
4715	if (rt->fib6_nh->fib_nh_dev == dev)
4716		rt->fib6_nh->fib_nh_flags |= nh_flags;
4717	list_for_each_entry(iter, &rt->fib6_siblings, fib6_siblings)
4718		if (iter->fib6_nh->fib_nh_dev == dev)
4719			iter->fib6_nh->fib_nh_flags |= nh_flags;
4720}
4721
4722/* called with write lock held for table with rt */
4723static int fib6_ifdown(struct fib6_info *rt, void *p_arg)
4724{
4725	const struct arg_netdev_event *arg = p_arg;
4726	const struct net_device *dev = arg->dev;
4727	struct net *net = dev_net(dev);
4728
4729	if (rt == net->ipv6.fib6_null_entry || rt->nh)
4730		return 0;
4731
4732	switch (arg->event) {
4733	case NETDEV_UNREGISTER:
4734		return rt->fib6_nh->fib_nh_dev == dev ? -1 : 0;
4735	case NETDEV_DOWN:
4736		if (rt->should_flush)
4737			return -1;
4738		if (!rt->fib6_nsiblings)
4739			return rt->fib6_nh->fib_nh_dev == dev ? -1 : 0;
4740		if (rt6_multipath_uses_dev(rt, dev)) {
4741			unsigned int count;
4742
4743			count = rt6_multipath_dead_count(rt, dev);
4744			if (rt->fib6_nsiblings + 1 == count) {
4745				rt6_multipath_flush(rt);
4746				return -1;
4747			}
4748			rt6_multipath_nh_flags_set(rt, dev, RTNH_F_DEAD |
4749						   RTNH_F_LINKDOWN);
4750			fib6_update_sernum(net, rt);
4751			rt6_multipath_rebalance(rt);
4752		}
4753		return -2;
4754	case NETDEV_CHANGE:
4755		if (rt->fib6_nh->fib_nh_dev != dev ||
4756		    rt->fib6_flags & (RTF_LOCAL | RTF_ANYCAST))
4757			break;
4758		rt->fib6_nh->fib_nh_flags |= RTNH_F_LINKDOWN;
4759		rt6_multipath_rebalance(rt);
4760		break;
4761	}
4762
4763	return 0;
4764}
4765
4766void rt6_sync_down_dev(struct net_device *dev, unsigned long event)
4767{
4768	struct arg_netdev_event arg = {
4769		.dev = dev,
4770		{
4771			.event = event,
4772		},
4773	};
4774	struct net *net = dev_net(dev);
4775
4776	if (net->ipv6.sysctl.skip_notify_on_dev_down)
4777		fib6_clean_all_skip_notify(net, fib6_ifdown, &arg);
4778	else
4779		fib6_clean_all(net, fib6_ifdown, &arg);
4780}
4781
4782void rt6_disable_ip(struct net_device *dev, unsigned long event)
4783{
4784	rt6_sync_down_dev(dev, event);
4785	rt6_uncached_list_flush_dev(dev_net(dev), dev);
4786	neigh_ifdown(&nd_tbl, dev);
4787}
4788
4789struct rt6_mtu_change_arg {
4790	struct net_device *dev;
4791	unsigned int mtu;
4792	struct fib6_info *f6i;
4793};
4794
4795static int fib6_nh_mtu_change(struct fib6_nh *nh, void *_arg)
4796{
4797	struct rt6_mtu_change_arg *arg = (struct rt6_mtu_change_arg *)_arg;
4798	struct fib6_info *f6i = arg->f6i;
4799
4800	/* For administrative MTU increase, there is no way to discover
4801	 * IPv6 PMTU increase, so PMTU increase should be updated here.
4802	 * Since RFC 1981 doesn't include administrative MTU increase
4803	 * update PMTU increase is a MUST. (i.e. jumbo frame)
4804	 */
4805	if (nh->fib_nh_dev == arg->dev) {
4806		struct inet6_dev *idev = __in6_dev_get(arg->dev);
4807		u32 mtu = f6i->fib6_pmtu;
4808
4809		if (mtu >= arg->mtu ||
4810		    (mtu < arg->mtu && mtu == idev->cnf.mtu6))
4811			fib6_metric_set(f6i, RTAX_MTU, arg->mtu);
4812
4813		spin_lock_bh(&rt6_exception_lock);
4814		rt6_exceptions_update_pmtu(idev, nh, arg->mtu);
4815		spin_unlock_bh(&rt6_exception_lock);
4816	}
4817
4818	return 0;
4819}
4820
4821static int rt6_mtu_change_route(struct fib6_info *f6i, void *p_arg)
4822{
4823	struct rt6_mtu_change_arg *arg = (struct rt6_mtu_change_arg *) p_arg;
4824	struct inet6_dev *idev;
4825
4826	/* In IPv6 pmtu discovery is not optional,
4827	   so that RTAX_MTU lock cannot disable it.
4828	   We still use this lock to block changes
4829	   caused by addrconf/ndisc.
4830	*/
4831
4832	idev = __in6_dev_get(arg->dev);
4833	if (!idev)
4834		return 0;
4835
4836	if (fib6_metric_locked(f6i, RTAX_MTU))
4837		return 0;
4838
4839	arg->f6i = f6i;
4840	if (f6i->nh) {
4841		/* fib6_nh_mtu_change only returns 0, so this is safe */
4842		return nexthop_for_each_fib6_nh(f6i->nh, fib6_nh_mtu_change,
4843						arg);
4844	}
4845
4846	return fib6_nh_mtu_change(f6i->fib6_nh, arg);
4847}
4848
4849void rt6_mtu_change(struct net_device *dev, unsigned int mtu)
4850{
4851	struct rt6_mtu_change_arg arg = {
4852		.dev = dev,
4853		.mtu = mtu,
4854	};
4855
4856	fib6_clean_all(dev_net(dev), rt6_mtu_change_route, &arg);
4857}
4858
4859static const struct nla_policy rtm_ipv6_policy[RTA_MAX+1] = {
4860	[RTA_UNSPEC]		= { .strict_start_type = RTA_DPORT + 1 },
4861	[RTA_GATEWAY]           = { .len = sizeof(struct in6_addr) },
4862	[RTA_PREFSRC]		= { .len = sizeof(struct in6_addr) },
4863	[RTA_OIF]               = { .type = NLA_U32 },
4864	[RTA_IIF]		= { .type = NLA_U32 },
4865	[RTA_PRIORITY]          = { .type = NLA_U32 },
4866	[RTA_METRICS]           = { .type = NLA_NESTED },
4867	[RTA_MULTIPATH]		= { .len = sizeof(struct rtnexthop) },
4868	[RTA_PREF]              = { .type = NLA_U8 },
4869	[RTA_ENCAP_TYPE]	= { .type = NLA_U16 },
4870	[RTA_ENCAP]		= { .type = NLA_NESTED },
4871	[RTA_EXPIRES]		= { .type = NLA_U32 },
4872	[RTA_UID]		= { .type = NLA_U32 },
4873	[RTA_MARK]		= { .type = NLA_U32 },
4874	[RTA_TABLE]		= { .type = NLA_U32 },
4875	[RTA_IP_PROTO]		= { .type = NLA_U8 },
4876	[RTA_SPORT]		= { .type = NLA_U16 },
4877	[RTA_DPORT]		= { .type = NLA_U16 },
4878	[RTA_NH_ID]		= { .type = NLA_U32 },
4879};
4880
4881static int rtm_to_fib6_config(struct sk_buff *skb, struct nlmsghdr *nlh,
4882			      struct fib6_config *cfg,
4883			      struct netlink_ext_ack *extack)
4884{
4885	struct rtmsg *rtm;
4886	struct nlattr *tb[RTA_MAX+1];
4887	unsigned int pref;
4888	int err;
4889
4890	err = nlmsg_parse_deprecated(nlh, sizeof(*rtm), tb, RTA_MAX,
4891				     rtm_ipv6_policy, extack);
4892	if (err < 0)
4893		goto errout;
4894
4895	err = -EINVAL;
4896	rtm = nlmsg_data(nlh);
4897
 
 
 
 
 
 
4898	*cfg = (struct fib6_config){
4899		.fc_table = rtm->rtm_table,
4900		.fc_dst_len = rtm->rtm_dst_len,
4901		.fc_src_len = rtm->rtm_src_len,
4902		.fc_flags = RTF_UP,
4903		.fc_protocol = rtm->rtm_protocol,
4904		.fc_type = rtm->rtm_type,
4905
4906		.fc_nlinfo.portid = NETLINK_CB(skb).portid,
4907		.fc_nlinfo.nlh = nlh,
4908		.fc_nlinfo.nl_net = sock_net(skb->sk),
4909	};
4910
4911	if (rtm->rtm_type == RTN_UNREACHABLE ||
4912	    rtm->rtm_type == RTN_BLACKHOLE ||
4913	    rtm->rtm_type == RTN_PROHIBIT ||
4914	    rtm->rtm_type == RTN_THROW)
4915		cfg->fc_flags |= RTF_REJECT;
4916
4917	if (rtm->rtm_type == RTN_LOCAL)
4918		cfg->fc_flags |= RTF_LOCAL;
4919
4920	if (rtm->rtm_flags & RTM_F_CLONED)
4921		cfg->fc_flags |= RTF_CACHE;
4922
4923	cfg->fc_flags |= (rtm->rtm_flags & RTNH_F_ONLINK);
4924
4925	if (tb[RTA_NH_ID]) {
4926		if (tb[RTA_GATEWAY]   || tb[RTA_OIF] ||
4927		    tb[RTA_MULTIPATH] || tb[RTA_ENCAP]) {
4928			NL_SET_ERR_MSG(extack,
4929				       "Nexthop specification and nexthop id are mutually exclusive");
4930			goto errout;
4931		}
4932		cfg->fc_nh_id = nla_get_u32(tb[RTA_NH_ID]);
4933	}
4934
4935	if (tb[RTA_GATEWAY]) {
4936		cfg->fc_gateway = nla_get_in6_addr(tb[RTA_GATEWAY]);
4937		cfg->fc_flags |= RTF_GATEWAY;
4938	}
4939	if (tb[RTA_VIA]) {
4940		NL_SET_ERR_MSG(extack, "IPv6 does not support RTA_VIA attribute");
4941		goto errout;
4942	}
4943
4944	if (tb[RTA_DST]) {
4945		int plen = (rtm->rtm_dst_len + 7) >> 3;
4946
4947		if (nla_len(tb[RTA_DST]) < plen)
4948			goto errout;
4949
4950		nla_memcpy(&cfg->fc_dst, tb[RTA_DST], plen);
4951	}
4952
4953	if (tb[RTA_SRC]) {
4954		int plen = (rtm->rtm_src_len + 7) >> 3;
4955
4956		if (nla_len(tb[RTA_SRC]) < plen)
4957			goto errout;
4958
4959		nla_memcpy(&cfg->fc_src, tb[RTA_SRC], plen);
4960	}
4961
4962	if (tb[RTA_PREFSRC])
4963		cfg->fc_prefsrc = nla_get_in6_addr(tb[RTA_PREFSRC]);
4964
4965	if (tb[RTA_OIF])
4966		cfg->fc_ifindex = nla_get_u32(tb[RTA_OIF]);
4967
4968	if (tb[RTA_PRIORITY])
4969		cfg->fc_metric = nla_get_u32(tb[RTA_PRIORITY]);
4970
4971	if (tb[RTA_METRICS]) {
4972		cfg->fc_mx = nla_data(tb[RTA_METRICS]);
4973		cfg->fc_mx_len = nla_len(tb[RTA_METRICS]);
4974	}
4975
4976	if (tb[RTA_TABLE])
4977		cfg->fc_table = nla_get_u32(tb[RTA_TABLE]);
4978
4979	if (tb[RTA_MULTIPATH]) {
4980		cfg->fc_mp = nla_data(tb[RTA_MULTIPATH]);
4981		cfg->fc_mp_len = nla_len(tb[RTA_MULTIPATH]);
4982
4983		err = lwtunnel_valid_encap_type_attr(cfg->fc_mp,
4984						     cfg->fc_mp_len, extack);
4985		if (err < 0)
4986			goto errout;
4987	}
4988
4989	if (tb[RTA_PREF]) {
4990		pref = nla_get_u8(tb[RTA_PREF]);
4991		if (pref != ICMPV6_ROUTER_PREF_LOW &&
4992		    pref != ICMPV6_ROUTER_PREF_HIGH)
4993			pref = ICMPV6_ROUTER_PREF_MEDIUM;
4994		cfg->fc_flags |= RTF_PREF(pref);
4995	}
4996
4997	if (tb[RTA_ENCAP])
4998		cfg->fc_encap = tb[RTA_ENCAP];
4999
5000	if (tb[RTA_ENCAP_TYPE]) {
5001		cfg->fc_encap_type = nla_get_u16(tb[RTA_ENCAP_TYPE]);
5002
5003		err = lwtunnel_valid_encap_type(cfg->fc_encap_type, extack);
5004		if (err < 0)
5005			goto errout;
5006	}
5007
5008	if (tb[RTA_EXPIRES]) {
5009		unsigned long timeout = addrconf_timeout_fixup(nla_get_u32(tb[RTA_EXPIRES]), HZ);
5010
5011		if (addrconf_finite_timeout(timeout)) {
5012			cfg->fc_expires = jiffies_to_clock_t(timeout * HZ);
5013			cfg->fc_flags |= RTF_EXPIRES;
5014		}
5015	}
5016
5017	err = 0;
5018errout:
5019	return err;
5020}
5021
5022struct rt6_nh {
5023	struct fib6_info *fib6_info;
5024	struct fib6_config r_cfg;
5025	struct list_head next;
5026};
5027
5028static int ip6_route_info_append(struct net *net,
5029				 struct list_head *rt6_nh_list,
5030				 struct fib6_info *rt,
5031				 struct fib6_config *r_cfg)
5032{
5033	struct rt6_nh *nh;
5034	int err = -EEXIST;
5035
5036	list_for_each_entry(nh, rt6_nh_list, next) {
5037		/* check if fib6_info already exists */
5038		if (rt6_duplicate_nexthop(nh->fib6_info, rt))
5039			return err;
5040	}
5041
5042	nh = kzalloc(sizeof(*nh), GFP_KERNEL);
5043	if (!nh)
5044		return -ENOMEM;
5045	nh->fib6_info = rt;
5046	memcpy(&nh->r_cfg, r_cfg, sizeof(*r_cfg));
5047	list_add_tail(&nh->next, rt6_nh_list);
5048
5049	return 0;
5050}
5051
5052static void ip6_route_mpath_notify(struct fib6_info *rt,
5053				   struct fib6_info *rt_last,
5054				   struct nl_info *info,
5055				   __u16 nlflags)
5056{
5057	/* if this is an APPEND route, then rt points to the first route
5058	 * inserted and rt_last points to last route inserted. Userspace
5059	 * wants a consistent dump of the route which starts at the first
5060	 * nexthop. Since sibling routes are always added at the end of
5061	 * the list, find the first sibling of the last route appended
5062	 */
5063	if ((nlflags & NLM_F_APPEND) && rt_last && rt_last->fib6_nsiblings) {
5064		rt = list_first_entry(&rt_last->fib6_siblings,
5065				      struct fib6_info,
5066				      fib6_siblings);
5067	}
5068
5069	if (rt)
5070		inet6_rt_notify(RTM_NEWROUTE, rt, info, nlflags);
5071}
5072
5073static bool ip6_route_mpath_should_notify(const struct fib6_info *rt)
5074{
5075	bool rt_can_ecmp = rt6_qualify_for_ecmp(rt);
5076	bool should_notify = false;
5077	struct fib6_info *leaf;
5078	struct fib6_node *fn;
5079
5080	rcu_read_lock();
5081	fn = rcu_dereference(rt->fib6_node);
5082	if (!fn)
5083		goto out;
5084
5085	leaf = rcu_dereference(fn->leaf);
5086	if (!leaf)
5087		goto out;
5088
5089	if (rt == leaf ||
5090	    (rt_can_ecmp && rt->fib6_metric == leaf->fib6_metric &&
5091	     rt6_qualify_for_ecmp(leaf)))
5092		should_notify = true;
5093out:
5094	rcu_read_unlock();
5095
5096	return should_notify;
5097}
5098
 
 
 
 
 
 
 
 
 
 
 
 
 
5099static int ip6_route_multipath_add(struct fib6_config *cfg,
5100				   struct netlink_ext_ack *extack)
5101{
5102	struct fib6_info *rt_notif = NULL, *rt_last = NULL;
5103	struct nl_info *info = &cfg->fc_nlinfo;
5104	struct fib6_config r_cfg;
5105	struct rtnexthop *rtnh;
5106	struct fib6_info *rt;
5107	struct rt6_nh *err_nh;
5108	struct rt6_nh *nh, *nh_safe;
5109	__u16 nlflags;
5110	int remaining;
5111	int attrlen;
5112	int err = 1;
5113	int nhn = 0;
5114	int replace = (cfg->fc_nlinfo.nlh &&
5115		       (cfg->fc_nlinfo.nlh->nlmsg_flags & NLM_F_REPLACE));
5116	LIST_HEAD(rt6_nh_list);
5117
5118	nlflags = replace ? NLM_F_REPLACE : NLM_F_CREATE;
5119	if (info->nlh && info->nlh->nlmsg_flags & NLM_F_APPEND)
5120		nlflags |= NLM_F_APPEND;
5121
5122	remaining = cfg->fc_mp_len;
5123	rtnh = (struct rtnexthop *)cfg->fc_mp;
5124
5125	/* Parse a Multipath Entry and build a list (rt6_nh_list) of
5126	 * fib6_info structs per nexthop
5127	 */
5128	while (rtnh_ok(rtnh, remaining)) {
5129		memcpy(&r_cfg, cfg, sizeof(*cfg));
5130		if (rtnh->rtnh_ifindex)
5131			r_cfg.fc_ifindex = rtnh->rtnh_ifindex;
5132
5133		attrlen = rtnh_attrlen(rtnh);
5134		if (attrlen > 0) {
5135			struct nlattr *nla, *attrs = rtnh_attrs(rtnh);
5136
5137			nla = nla_find(attrs, attrlen, RTA_GATEWAY);
5138			if (nla) {
5139				r_cfg.fc_gateway = nla_get_in6_addr(nla);
 
 
 
 
5140				r_cfg.fc_flags |= RTF_GATEWAY;
5141			}
5142			r_cfg.fc_encap = nla_find(attrs, attrlen, RTA_ENCAP);
 
 
 
 
5143			nla = nla_find(attrs, attrlen, RTA_ENCAP_TYPE);
5144			if (nla)
5145				r_cfg.fc_encap_type = nla_get_u16(nla);
5146		}
5147
5148		r_cfg.fc_flags |= (rtnh->rtnh_flags & RTNH_F_ONLINK);
5149		rt = ip6_route_info_create(&r_cfg, GFP_KERNEL, extack);
5150		if (IS_ERR(rt)) {
5151			err = PTR_ERR(rt);
5152			rt = NULL;
5153			goto cleanup;
5154		}
5155		if (!rt6_qualify_for_ecmp(rt)) {
5156			err = -EINVAL;
5157			NL_SET_ERR_MSG(extack,
5158				       "Device only routes can not be added for IPv6 using the multipath API.");
5159			fib6_info_release(rt);
5160			goto cleanup;
5161		}
5162
5163		rt->fib6_nh->fib_nh_weight = rtnh->rtnh_hops + 1;
5164
5165		err = ip6_route_info_append(info->nl_net, &rt6_nh_list,
5166					    rt, &r_cfg);
5167		if (err) {
5168			fib6_info_release(rt);
5169			goto cleanup;
5170		}
5171
5172		rtnh = rtnh_next(rtnh, &remaining);
5173	}
5174
5175	if (list_empty(&rt6_nh_list)) {
5176		NL_SET_ERR_MSG(extack,
5177			       "Invalid nexthop configuration - no valid nexthops");
5178		return -EINVAL;
5179	}
5180
5181	/* for add and replace send one notification with all nexthops.
5182	 * Skip the notification in fib6_add_rt2node and send one with
5183	 * the full route when done
5184	 */
5185	info->skip_notify = 1;
5186
5187	/* For add and replace, send one notification with all nexthops. For
5188	 * append, send one notification with all appended nexthops.
5189	 */
5190	info->skip_notify_kernel = 1;
5191
5192	err_nh = NULL;
5193	list_for_each_entry(nh, &rt6_nh_list, next) {
5194		err = __ip6_ins_rt(nh->fib6_info, info, extack);
5195		fib6_info_release(nh->fib6_info);
5196
5197		if (!err) {
5198			/* save reference to last route successfully inserted */
5199			rt_last = nh->fib6_info;
5200
5201			/* save reference to first route for notification */
5202			if (!rt_notif)
5203				rt_notif = nh->fib6_info;
5204		}
5205
5206		/* nh->fib6_info is used or freed at this point, reset to NULL*/
5207		nh->fib6_info = NULL;
5208		if (err) {
5209			if (replace && nhn)
5210				NL_SET_ERR_MSG_MOD(extack,
5211						   "multipath route replace failed (check consistency of installed routes)");
5212			err_nh = nh;
5213			goto add_errout;
5214		}
 
 
 
 
 
 
5215
5216		/* Because each route is added like a single route we remove
5217		 * these flags after the first nexthop: if there is a collision,
5218		 * we have already failed to add the first nexthop:
5219		 * fib6_add_rt2node() has rejected it; when replacing, old
5220		 * nexthops have been replaced by first new, the rest should
5221		 * be added to it.
5222		 */
5223		cfg->fc_nlinfo.nlh->nlmsg_flags &= ~(NLM_F_EXCL |
5224						     NLM_F_REPLACE);
5225		cfg->fc_nlinfo.nlh->nlmsg_flags |= NLM_F_CREATE;
 
 
5226		nhn++;
5227	}
5228
5229	/* An in-kernel notification should only be sent in case the new
5230	 * multipath route is added as the first route in the node, or if
5231	 * it was appended to it. We pass 'rt_notif' since it is the first
5232	 * sibling and might allow us to skip some checks in the replace case.
5233	 */
5234	if (ip6_route_mpath_should_notify(rt_notif)) {
5235		enum fib_event_type fib_event;
5236
5237		if (rt_notif->fib6_nsiblings != nhn - 1)
5238			fib_event = FIB_EVENT_ENTRY_APPEND;
5239		else
5240			fib_event = FIB_EVENT_ENTRY_REPLACE;
5241
5242		err = call_fib6_multipath_entry_notifiers(info->nl_net,
5243							  fib_event, rt_notif,
5244							  nhn - 1, extack);
5245		if (err) {
5246			/* Delete all the siblings that were just added */
5247			err_nh = NULL;
5248			goto add_errout;
5249		}
5250	}
5251
5252	/* success ... tell user about new route */
5253	ip6_route_mpath_notify(rt_notif, rt_last, info, nlflags);
5254	goto cleanup;
5255
5256add_errout:
5257	/* send notification for routes that were added so that
5258	 * the delete notifications sent by ip6_route_del are
5259	 * coherent
5260	 */
5261	if (rt_notif)
5262		ip6_route_mpath_notify(rt_notif, rt_last, info, nlflags);
5263
5264	/* Delete routes that were already added */
5265	list_for_each_entry(nh, &rt6_nh_list, next) {
5266		if (err_nh == nh)
5267			break;
5268		ip6_route_del(&nh->r_cfg, extack);
5269	}
5270
5271cleanup:
5272	list_for_each_entry_safe(nh, nh_safe, &rt6_nh_list, next) {
5273		if (nh->fib6_info)
5274			fib6_info_release(nh->fib6_info);
5275		list_del(&nh->next);
5276		kfree(nh);
5277	}
5278
5279	return err;
5280}
5281
5282static int ip6_route_multipath_del(struct fib6_config *cfg,
5283				   struct netlink_ext_ack *extack)
5284{
5285	struct fib6_config r_cfg;
5286	struct rtnexthop *rtnh;
 
5287	int remaining;
5288	int attrlen;
5289	int err = 1, last_err = 0;
5290
5291	remaining = cfg->fc_mp_len;
5292	rtnh = (struct rtnexthop *)cfg->fc_mp;
5293
5294	/* Parse a Multipath Entry */
5295	while (rtnh_ok(rtnh, remaining)) {
5296		memcpy(&r_cfg, cfg, sizeof(*cfg));
5297		if (rtnh->rtnh_ifindex)
5298			r_cfg.fc_ifindex = rtnh->rtnh_ifindex;
5299
5300		attrlen = rtnh_attrlen(rtnh);
5301		if (attrlen > 0) {
5302			struct nlattr *nla, *attrs = rtnh_attrs(rtnh);
5303
5304			nla = nla_find(attrs, attrlen, RTA_GATEWAY);
5305			if (nla) {
5306				nla_memcpy(&r_cfg.fc_gateway, nla, 16);
 
 
 
 
 
 
5307				r_cfg.fc_flags |= RTF_GATEWAY;
5308			}
5309		}
5310		err = ip6_route_del(&r_cfg, extack);
5311		if (err)
5312			last_err = err;
5313
 
5314		rtnh = rtnh_next(rtnh, &remaining);
5315	}
5316
5317	return last_err;
5318}
5319
5320static int inet6_rtm_delroute(struct sk_buff *skb, struct nlmsghdr *nlh,
5321			      struct netlink_ext_ack *extack)
5322{
5323	struct fib6_config cfg;
5324	int err;
5325
5326	err = rtm_to_fib6_config(skb, nlh, &cfg, extack);
5327	if (err < 0)
5328		return err;
5329
5330	if (cfg.fc_nh_id &&
5331	    !nexthop_find_by_id(sock_net(skb->sk), cfg.fc_nh_id)) {
5332		NL_SET_ERR_MSG(extack, "Nexthop id does not exist");
5333		return -EINVAL;
5334	}
5335
5336	if (cfg.fc_mp)
5337		return ip6_route_multipath_del(&cfg, extack);
5338	else {
5339		cfg.fc_delete_all_nh = 1;
5340		return ip6_route_del(&cfg, extack);
5341	}
5342}
5343
5344static int inet6_rtm_newroute(struct sk_buff *skb, struct nlmsghdr *nlh,
5345			      struct netlink_ext_ack *extack)
5346{
5347	struct fib6_config cfg;
5348	int err;
5349
5350	err = rtm_to_fib6_config(skb, nlh, &cfg, extack);
5351	if (err < 0)
5352		return err;
5353
5354	if (cfg.fc_metric == 0)
5355		cfg.fc_metric = IP6_RT_PRIO_USER;
5356
5357	if (cfg.fc_mp)
5358		return ip6_route_multipath_add(&cfg, extack);
5359	else
5360		return ip6_route_add(&cfg, GFP_KERNEL, extack);
5361}
5362
5363/* add the overhead of this fib6_nh to nexthop_len */
5364static int rt6_nh_nlmsg_size(struct fib6_nh *nh, void *arg)
5365{
5366	int *nexthop_len = arg;
5367
5368	*nexthop_len += nla_total_size(0)	 /* RTA_MULTIPATH */
5369		     + NLA_ALIGN(sizeof(struct rtnexthop))
5370		     + nla_total_size(16); /* RTA_GATEWAY */
5371
5372	if (nh->fib_nh_lws) {
5373		/* RTA_ENCAP_TYPE */
5374		*nexthop_len += lwtunnel_get_encap_size(nh->fib_nh_lws);
5375		/* RTA_ENCAP */
5376		*nexthop_len += nla_total_size(2);
5377	}
5378
5379	return 0;
5380}
5381
5382static size_t rt6_nlmsg_size(struct fib6_info *f6i)
5383{
5384	int nexthop_len;
5385
5386	if (f6i->nh) {
5387		nexthop_len = nla_total_size(4); /* RTA_NH_ID */
5388		nexthop_for_each_fib6_nh(f6i->nh, rt6_nh_nlmsg_size,
5389					 &nexthop_len);
5390	} else {
 
5391		struct fib6_nh *nh = f6i->fib6_nh;
5392
5393		nexthop_len = 0;
5394		if (f6i->fib6_nsiblings) {
5395			nexthop_len = nla_total_size(0)	 /* RTA_MULTIPATH */
5396				    + NLA_ALIGN(sizeof(struct rtnexthop))
5397				    + nla_total_size(16) /* RTA_GATEWAY */
5398				    + lwtunnel_get_encap_size(nh->fib_nh_lws);
5399
5400			nexthop_len *= f6i->fib6_nsiblings;
 
 
 
5401		}
5402		nexthop_len += lwtunnel_get_encap_size(nh->fib_nh_lws);
5403	}
5404
5405	return NLMSG_ALIGN(sizeof(struct rtmsg))
5406	       + nla_total_size(16) /* RTA_SRC */
5407	       + nla_total_size(16) /* RTA_DST */
5408	       + nla_total_size(16) /* RTA_GATEWAY */
5409	       + nla_total_size(16) /* RTA_PREFSRC */
5410	       + nla_total_size(4) /* RTA_TABLE */
5411	       + nla_total_size(4) /* RTA_IIF */
5412	       + nla_total_size(4) /* RTA_OIF */
5413	       + nla_total_size(4) /* RTA_PRIORITY */
5414	       + RTAX_MAX * nla_total_size(4) /* RTA_METRICS */
5415	       + nla_total_size(sizeof(struct rta_cacheinfo))
5416	       + nla_total_size(TCP_CA_NAME_MAX) /* RTAX_CC_ALGO */
5417	       + nla_total_size(1) /* RTA_PREF */
5418	       + nexthop_len;
5419}
5420
5421static int rt6_fill_node_nexthop(struct sk_buff *skb, struct nexthop *nh,
5422				 unsigned char *flags)
5423{
5424	if (nexthop_is_multipath(nh)) {
5425		struct nlattr *mp;
5426
5427		mp = nla_nest_start_noflag(skb, RTA_MULTIPATH);
5428		if (!mp)
5429			goto nla_put_failure;
5430
5431		if (nexthop_mpath_fill_node(skb, nh, AF_INET6))
5432			goto nla_put_failure;
5433
5434		nla_nest_end(skb, mp);
5435	} else {
5436		struct fib6_nh *fib6_nh;
5437
5438		fib6_nh = nexthop_fib6_nh(nh);
5439		if (fib_nexthop_info(skb, &fib6_nh->nh_common, AF_INET6,
5440				     flags, false) < 0)
5441			goto nla_put_failure;
5442	}
5443
5444	return 0;
5445
5446nla_put_failure:
5447	return -EMSGSIZE;
5448}
5449
5450static int rt6_fill_node(struct net *net, struct sk_buff *skb,
5451			 struct fib6_info *rt, struct dst_entry *dst,
5452			 struct in6_addr *dest, struct in6_addr *src,
5453			 int iif, int type, u32 portid, u32 seq,
5454			 unsigned int flags)
5455{
5456	struct rt6_info *rt6 = (struct rt6_info *)dst;
5457	struct rt6key *rt6_dst, *rt6_src;
5458	u32 *pmetrics, table, rt6_flags;
5459	unsigned char nh_flags = 0;
5460	struct nlmsghdr *nlh;
5461	struct rtmsg *rtm;
5462	long expires = 0;
5463
5464	nlh = nlmsg_put(skb, portid, seq, type, sizeof(*rtm), flags);
5465	if (!nlh)
5466		return -EMSGSIZE;
5467
5468	if (rt6) {
5469		rt6_dst = &rt6->rt6i_dst;
5470		rt6_src = &rt6->rt6i_src;
5471		rt6_flags = rt6->rt6i_flags;
5472	} else {
5473		rt6_dst = &rt->fib6_dst;
5474		rt6_src = &rt->fib6_src;
5475		rt6_flags = rt->fib6_flags;
5476	}
5477
5478	rtm = nlmsg_data(nlh);
5479	rtm->rtm_family = AF_INET6;
5480	rtm->rtm_dst_len = rt6_dst->plen;
5481	rtm->rtm_src_len = rt6_src->plen;
5482	rtm->rtm_tos = 0;
5483	if (rt->fib6_table)
5484		table = rt->fib6_table->tb6_id;
5485	else
5486		table = RT6_TABLE_UNSPEC;
5487	rtm->rtm_table = table < 256 ? table : RT_TABLE_COMPAT;
5488	if (nla_put_u32(skb, RTA_TABLE, table))
5489		goto nla_put_failure;
5490
5491	rtm->rtm_type = rt->fib6_type;
5492	rtm->rtm_flags = 0;
5493	rtm->rtm_scope = RT_SCOPE_UNIVERSE;
5494	rtm->rtm_protocol = rt->fib6_protocol;
5495
5496	if (rt6_flags & RTF_CACHE)
5497		rtm->rtm_flags |= RTM_F_CLONED;
5498
5499	if (dest) {
5500		if (nla_put_in6_addr(skb, RTA_DST, dest))
5501			goto nla_put_failure;
5502		rtm->rtm_dst_len = 128;
5503	} else if (rtm->rtm_dst_len)
5504		if (nla_put_in6_addr(skb, RTA_DST, &rt6_dst->addr))
5505			goto nla_put_failure;
5506#ifdef CONFIG_IPV6_SUBTREES
5507	if (src) {
5508		if (nla_put_in6_addr(skb, RTA_SRC, src))
5509			goto nla_put_failure;
5510		rtm->rtm_src_len = 128;
5511	} else if (rtm->rtm_src_len &&
5512		   nla_put_in6_addr(skb, RTA_SRC, &rt6_src->addr))
5513		goto nla_put_failure;
5514#endif
5515	if (iif) {
5516#ifdef CONFIG_IPV6_MROUTE
5517		if (ipv6_addr_is_multicast(&rt6_dst->addr)) {
5518			int err = ip6mr_get_route(net, skb, rtm, portid);
5519
5520			if (err == 0)
5521				return 0;
5522			if (err < 0)
5523				goto nla_put_failure;
5524		} else
5525#endif
5526			if (nla_put_u32(skb, RTA_IIF, iif))
5527				goto nla_put_failure;
5528	} else if (dest) {
5529		struct in6_addr saddr_buf;
5530		if (ip6_route_get_saddr(net, rt, dest, 0, &saddr_buf) == 0 &&
5531		    nla_put_in6_addr(skb, RTA_PREFSRC, &saddr_buf))
5532			goto nla_put_failure;
5533	}
5534
5535	if (rt->fib6_prefsrc.plen) {
5536		struct in6_addr saddr_buf;
5537		saddr_buf = rt->fib6_prefsrc.addr;
5538		if (nla_put_in6_addr(skb, RTA_PREFSRC, &saddr_buf))
5539			goto nla_put_failure;
5540	}
5541
5542	pmetrics = dst ? dst_metrics_ptr(dst) : rt->fib6_metrics->metrics;
5543	if (rtnetlink_put_metrics(skb, pmetrics) < 0)
5544		goto nla_put_failure;
5545
5546	if (nla_put_u32(skb, RTA_PRIORITY, rt->fib6_metric))
5547		goto nla_put_failure;
5548
5549	/* For multipath routes, walk the siblings list and add
5550	 * each as a nexthop within RTA_MULTIPATH.
5551	 */
5552	if (rt6) {
5553		if (rt6_flags & RTF_GATEWAY &&
5554		    nla_put_in6_addr(skb, RTA_GATEWAY, &rt6->rt6i_gateway))
5555			goto nla_put_failure;
5556
5557		if (dst->dev && nla_put_u32(skb, RTA_OIF, dst->dev->ifindex))
5558			goto nla_put_failure;
 
 
 
 
5559	} else if (rt->fib6_nsiblings) {
5560		struct fib6_info *sibling, *next_sibling;
5561		struct nlattr *mp;
5562
5563		mp = nla_nest_start_noflag(skb, RTA_MULTIPATH);
5564		if (!mp)
5565			goto nla_put_failure;
5566
5567		if (fib_add_nexthop(skb, &rt->fib6_nh->nh_common,
5568				    rt->fib6_nh->fib_nh_weight, AF_INET6) < 0)
 
5569			goto nla_put_failure;
5570
5571		list_for_each_entry_safe(sibling, next_sibling,
5572					 &rt->fib6_siblings, fib6_siblings) {
5573			if (fib_add_nexthop(skb, &sibling->fib6_nh->nh_common,
5574					    sibling->fib6_nh->fib_nh_weight,
5575					    AF_INET6) < 0)
5576				goto nla_put_failure;
5577		}
5578
5579		nla_nest_end(skb, mp);
5580	} else if (rt->nh) {
5581		if (nla_put_u32(skb, RTA_NH_ID, rt->nh->id))
5582			goto nla_put_failure;
5583
5584		if (nexthop_is_blackhole(rt->nh))
5585			rtm->rtm_type = RTN_BLACKHOLE;
5586
5587		if (net->ipv4.sysctl_nexthop_compat_mode &&
5588		    rt6_fill_node_nexthop(skb, rt->nh, &nh_flags) < 0)
5589			goto nla_put_failure;
5590
5591		rtm->rtm_flags |= nh_flags;
5592	} else {
5593		if (fib_nexthop_info(skb, &rt->fib6_nh->nh_common, AF_INET6,
5594				     &nh_flags, false) < 0)
5595			goto nla_put_failure;
5596
5597		rtm->rtm_flags |= nh_flags;
5598	}
5599
5600	if (rt6_flags & RTF_EXPIRES) {
5601		expires = dst ? dst->expires : rt->expires;
5602		expires -= jiffies;
5603	}
5604
5605	if (!dst) {
5606		if (rt->offload)
5607			rtm->rtm_flags |= RTM_F_OFFLOAD;
5608		if (rt->trap)
5609			rtm->rtm_flags |= RTM_F_TRAP;
 
 
5610	}
5611
5612	if (rtnl_put_cacheinfo(skb, dst, 0, expires, dst ? dst->error : 0) < 0)
5613		goto nla_put_failure;
5614
5615	if (nla_put_u8(skb, RTA_PREF, IPV6_EXTRACT_PREF(rt6_flags)))
5616		goto nla_put_failure;
5617
5618
5619	nlmsg_end(skb, nlh);
5620	return 0;
5621
5622nla_put_failure:
5623	nlmsg_cancel(skb, nlh);
5624	return -EMSGSIZE;
5625}
5626
5627static int fib6_info_nh_uses_dev(struct fib6_nh *nh, void *arg)
5628{
5629	const struct net_device *dev = arg;
5630
5631	if (nh->fib_nh_dev == dev)
5632		return 1;
5633
5634	return 0;
5635}
5636
5637static bool fib6_info_uses_dev(const struct fib6_info *f6i,
5638			       const struct net_device *dev)
5639{
5640	if (f6i->nh) {
5641		struct net_device *_dev = (struct net_device *)dev;
5642
5643		return !!nexthop_for_each_fib6_nh(f6i->nh,
5644						  fib6_info_nh_uses_dev,
5645						  _dev);
5646	}
5647
5648	if (f6i->fib6_nh->fib_nh_dev == dev)
5649		return true;
5650
5651	if (f6i->fib6_nsiblings) {
5652		struct fib6_info *sibling, *next_sibling;
5653
5654		list_for_each_entry_safe(sibling, next_sibling,
5655					 &f6i->fib6_siblings, fib6_siblings) {
5656			if (sibling->fib6_nh->fib_nh_dev == dev)
5657				return true;
5658		}
5659	}
5660
5661	return false;
5662}
5663
5664struct fib6_nh_exception_dump_walker {
5665	struct rt6_rtnl_dump_arg *dump;
5666	struct fib6_info *rt;
5667	unsigned int flags;
5668	unsigned int skip;
5669	unsigned int count;
5670};
5671
5672static int rt6_nh_dump_exceptions(struct fib6_nh *nh, void *arg)
5673{
5674	struct fib6_nh_exception_dump_walker *w = arg;
5675	struct rt6_rtnl_dump_arg *dump = w->dump;
5676	struct rt6_exception_bucket *bucket;
5677	struct rt6_exception *rt6_ex;
5678	int i, err;
5679
5680	bucket = fib6_nh_get_excptn_bucket(nh, NULL);
5681	if (!bucket)
5682		return 0;
5683
5684	for (i = 0; i < FIB6_EXCEPTION_BUCKET_SIZE; i++) {
5685		hlist_for_each_entry(rt6_ex, &bucket->chain, hlist) {
5686			if (w->skip) {
5687				w->skip--;
5688				continue;
5689			}
5690
5691			/* Expiration of entries doesn't bump sernum, insertion
5692			 * does. Removal is triggered by insertion, so we can
5693			 * rely on the fact that if entries change between two
5694			 * partial dumps, this node is scanned again completely,
5695			 * see rt6_insert_exception() and fib6_dump_table().
5696			 *
5697			 * Count expired entries we go through as handled
5698			 * entries that we'll skip next time, in case of partial
5699			 * node dump. Otherwise, if entries expire meanwhile,
5700			 * we'll skip the wrong amount.
5701			 */
5702			if (rt6_check_expired(rt6_ex->rt6i)) {
5703				w->count++;
5704				continue;
5705			}
5706
5707			err = rt6_fill_node(dump->net, dump->skb, w->rt,
5708					    &rt6_ex->rt6i->dst, NULL, NULL, 0,
5709					    RTM_NEWROUTE,
5710					    NETLINK_CB(dump->cb->skb).portid,
5711					    dump->cb->nlh->nlmsg_seq, w->flags);
5712			if (err)
5713				return err;
5714
5715			w->count++;
5716		}
5717		bucket++;
5718	}
5719
5720	return 0;
5721}
5722
5723/* Return -1 if done with node, number of handled routes on partial dump */
5724int rt6_dump_route(struct fib6_info *rt, void *p_arg, unsigned int skip)
5725{
5726	struct rt6_rtnl_dump_arg *arg = (struct rt6_rtnl_dump_arg *) p_arg;
5727	struct fib_dump_filter *filter = &arg->filter;
5728	unsigned int flags = NLM_F_MULTI;
5729	struct net *net = arg->net;
5730	int count = 0;
5731
5732	if (rt == net->ipv6.fib6_null_entry)
5733		return -1;
5734
5735	if ((filter->flags & RTM_F_PREFIX) &&
5736	    !(rt->fib6_flags & RTF_PREFIX_RT)) {
5737		/* success since this is not a prefix route */
5738		return -1;
5739	}
5740	if (filter->filter_set &&
5741	    ((filter->rt_type  && rt->fib6_type != filter->rt_type) ||
5742	     (filter->dev      && !fib6_info_uses_dev(rt, filter->dev)) ||
5743	     (filter->protocol && rt->fib6_protocol != filter->protocol))) {
5744		return -1;
5745	}
5746
5747	if (filter->filter_set ||
5748	    !filter->dump_routes || !filter->dump_exceptions) {
5749		flags |= NLM_F_DUMP_FILTERED;
5750	}
5751
5752	if (filter->dump_routes) {
5753		if (skip) {
5754			skip--;
5755		} else {
5756			if (rt6_fill_node(net, arg->skb, rt, NULL, NULL, NULL,
5757					  0, RTM_NEWROUTE,
5758					  NETLINK_CB(arg->cb->skb).portid,
5759					  arg->cb->nlh->nlmsg_seq, flags)) {
5760				return 0;
5761			}
5762			count++;
5763		}
5764	}
5765
5766	if (filter->dump_exceptions) {
5767		struct fib6_nh_exception_dump_walker w = { .dump = arg,
5768							   .rt = rt,
5769							   .flags = flags,
5770							   .skip = skip,
5771							   .count = 0 };
5772		int err;
5773
5774		rcu_read_lock();
5775		if (rt->nh) {
5776			err = nexthop_for_each_fib6_nh(rt->nh,
5777						       rt6_nh_dump_exceptions,
5778						       &w);
5779		} else {
5780			err = rt6_nh_dump_exceptions(rt->fib6_nh, &w);
5781		}
5782		rcu_read_unlock();
5783
5784		if (err)
5785			return count += w.count;
5786	}
5787
5788	return -1;
5789}
5790
5791static int inet6_rtm_valid_getroute_req(struct sk_buff *skb,
5792					const struct nlmsghdr *nlh,
5793					struct nlattr **tb,
5794					struct netlink_ext_ack *extack)
5795{
5796	struct rtmsg *rtm;
5797	int i, err;
5798
5799	if (nlh->nlmsg_len < nlmsg_msg_size(sizeof(*rtm))) {
5800		NL_SET_ERR_MSG_MOD(extack,
5801				   "Invalid header for get route request");
5802		return -EINVAL;
5803	}
5804
5805	if (!netlink_strict_get_check(skb))
5806		return nlmsg_parse_deprecated(nlh, sizeof(*rtm), tb, RTA_MAX,
5807					      rtm_ipv6_policy, extack);
5808
5809	rtm = nlmsg_data(nlh);
5810	if ((rtm->rtm_src_len && rtm->rtm_src_len != 128) ||
5811	    (rtm->rtm_dst_len && rtm->rtm_dst_len != 128) ||
5812	    rtm->rtm_table || rtm->rtm_protocol || rtm->rtm_scope ||
5813	    rtm->rtm_type) {
5814		NL_SET_ERR_MSG_MOD(extack, "Invalid values in header for get route request");
5815		return -EINVAL;
5816	}
5817	if (rtm->rtm_flags & ~RTM_F_FIB_MATCH) {
5818		NL_SET_ERR_MSG_MOD(extack,
5819				   "Invalid flags for get route request");
5820		return -EINVAL;
5821	}
5822
5823	err = nlmsg_parse_deprecated_strict(nlh, sizeof(*rtm), tb, RTA_MAX,
5824					    rtm_ipv6_policy, extack);
5825	if (err)
5826		return err;
5827
5828	if ((tb[RTA_SRC] && !rtm->rtm_src_len) ||
5829	    (tb[RTA_DST] && !rtm->rtm_dst_len)) {
5830		NL_SET_ERR_MSG_MOD(extack, "rtm_src_len and rtm_dst_len must be 128 for IPv6");
5831		return -EINVAL;
5832	}
5833
5834	for (i = 0; i <= RTA_MAX; i++) {
5835		if (!tb[i])
5836			continue;
5837
5838		switch (i) {
5839		case RTA_SRC:
5840		case RTA_DST:
5841		case RTA_IIF:
5842		case RTA_OIF:
5843		case RTA_MARK:
5844		case RTA_UID:
5845		case RTA_SPORT:
5846		case RTA_DPORT:
5847		case RTA_IP_PROTO:
5848			break;
5849		default:
5850			NL_SET_ERR_MSG_MOD(extack, "Unsupported attribute in get route request");
5851			return -EINVAL;
5852		}
5853	}
5854
5855	return 0;
5856}
5857
5858static int inet6_rtm_getroute(struct sk_buff *in_skb, struct nlmsghdr *nlh,
5859			      struct netlink_ext_ack *extack)
5860{
5861	struct net *net = sock_net(in_skb->sk);
5862	struct nlattr *tb[RTA_MAX+1];
5863	int err, iif = 0, oif = 0;
5864	struct fib6_info *from;
5865	struct dst_entry *dst;
5866	struct rt6_info *rt;
5867	struct sk_buff *skb;
5868	struct rtmsg *rtm;
5869	struct flowi6 fl6 = {};
5870	bool fibmatch;
5871
5872	err = inet6_rtm_valid_getroute_req(in_skb, nlh, tb, extack);
5873	if (err < 0)
5874		goto errout;
5875
5876	err = -EINVAL;
5877	rtm = nlmsg_data(nlh);
5878	fl6.flowlabel = ip6_make_flowinfo(rtm->rtm_tos, 0);
5879	fibmatch = !!(rtm->rtm_flags & RTM_F_FIB_MATCH);
5880
5881	if (tb[RTA_SRC]) {
5882		if (nla_len(tb[RTA_SRC]) < sizeof(struct in6_addr))
5883			goto errout;
5884
5885		fl6.saddr = *(struct in6_addr *)nla_data(tb[RTA_SRC]);
5886	}
5887
5888	if (tb[RTA_DST]) {
5889		if (nla_len(tb[RTA_DST]) < sizeof(struct in6_addr))
5890			goto errout;
5891
5892		fl6.daddr = *(struct in6_addr *)nla_data(tb[RTA_DST]);
5893	}
5894
5895	if (tb[RTA_IIF])
5896		iif = nla_get_u32(tb[RTA_IIF]);
5897
5898	if (tb[RTA_OIF])
5899		oif = nla_get_u32(tb[RTA_OIF]);
5900
5901	if (tb[RTA_MARK])
5902		fl6.flowi6_mark = nla_get_u32(tb[RTA_MARK]);
5903
5904	if (tb[RTA_UID])
5905		fl6.flowi6_uid = make_kuid(current_user_ns(),
5906					   nla_get_u32(tb[RTA_UID]));
5907	else
5908		fl6.flowi6_uid = iif ? INVALID_UID : current_uid();
5909
5910	if (tb[RTA_SPORT])
5911		fl6.fl6_sport = nla_get_be16(tb[RTA_SPORT]);
5912
5913	if (tb[RTA_DPORT])
5914		fl6.fl6_dport = nla_get_be16(tb[RTA_DPORT]);
5915
5916	if (tb[RTA_IP_PROTO]) {
5917		err = rtm_getroute_parse_ip_proto(tb[RTA_IP_PROTO],
5918						  &fl6.flowi6_proto, AF_INET6,
5919						  extack);
5920		if (err)
5921			goto errout;
5922	}
5923
5924	if (iif) {
5925		struct net_device *dev;
5926		int flags = 0;
5927
5928		rcu_read_lock();
5929
5930		dev = dev_get_by_index_rcu(net, iif);
5931		if (!dev) {
5932			rcu_read_unlock();
5933			err = -ENODEV;
5934			goto errout;
5935		}
5936
5937		fl6.flowi6_iif = iif;
5938
5939		if (!ipv6_addr_any(&fl6.saddr))
5940			flags |= RT6_LOOKUP_F_HAS_SADDR;
5941
5942		dst = ip6_route_input_lookup(net, dev, &fl6, NULL, flags);
5943
5944		rcu_read_unlock();
5945	} else {
5946		fl6.flowi6_oif = oif;
5947
5948		dst = ip6_route_output(net, NULL, &fl6);
5949	}
5950
5951
5952	rt = container_of(dst, struct rt6_info, dst);
5953	if (rt->dst.error) {
5954		err = rt->dst.error;
5955		ip6_rt_put(rt);
5956		goto errout;
5957	}
5958
5959	if (rt == net->ipv6.ip6_null_entry) {
5960		err = rt->dst.error;
5961		ip6_rt_put(rt);
5962		goto errout;
5963	}
5964
5965	skb = alloc_skb(NLMSG_GOODSIZE, GFP_KERNEL);
5966	if (!skb) {
5967		ip6_rt_put(rt);
5968		err = -ENOBUFS;
5969		goto errout;
5970	}
5971
5972	skb_dst_set(skb, &rt->dst);
5973
5974	rcu_read_lock();
5975	from = rcu_dereference(rt->from);
5976	if (from) {
5977		if (fibmatch)
5978			err = rt6_fill_node(net, skb, from, NULL, NULL, NULL,
5979					    iif, RTM_NEWROUTE,
5980					    NETLINK_CB(in_skb).portid,
5981					    nlh->nlmsg_seq, 0);
5982		else
5983			err = rt6_fill_node(net, skb, from, dst, &fl6.daddr,
5984					    &fl6.saddr, iif, RTM_NEWROUTE,
5985					    NETLINK_CB(in_skb).portid,
5986					    nlh->nlmsg_seq, 0);
5987	} else {
5988		err = -ENETUNREACH;
5989	}
5990	rcu_read_unlock();
5991
5992	if (err < 0) {
5993		kfree_skb(skb);
5994		goto errout;
5995	}
5996
5997	err = rtnl_unicast(skb, net, NETLINK_CB(in_skb).portid);
5998errout:
5999	return err;
6000}
6001
6002void inet6_rt_notify(int event, struct fib6_info *rt, struct nl_info *info,
6003		     unsigned int nlm_flags)
6004{
6005	struct sk_buff *skb;
6006	struct net *net = info->nl_net;
6007	u32 seq;
6008	int err;
6009
6010	err = -ENOBUFS;
6011	seq = info->nlh ? info->nlh->nlmsg_seq : 0;
6012
6013	skb = nlmsg_new(rt6_nlmsg_size(rt), gfp_any());
6014	if (!skb)
6015		goto errout;
6016
6017	err = rt6_fill_node(net, skb, rt, NULL, NULL, NULL, 0,
6018			    event, info->portid, seq, nlm_flags);
6019	if (err < 0) {
6020		/* -EMSGSIZE implies BUG in rt6_nlmsg_size() */
6021		WARN_ON(err == -EMSGSIZE);
6022		kfree_skb(skb);
6023		goto errout;
6024	}
6025	rtnl_notify(skb, net, info->portid, RTNLGRP_IPV6_ROUTE,
6026		    info->nlh, gfp_any());
6027	return;
6028errout:
6029	if (err < 0)
6030		rtnl_set_sk_err(net, RTNLGRP_IPV6_ROUTE, err);
6031}
6032
6033void fib6_rt_update(struct net *net, struct fib6_info *rt,
6034		    struct nl_info *info)
6035{
6036	u32 seq = info->nlh ? info->nlh->nlmsg_seq : 0;
6037	struct sk_buff *skb;
6038	int err = -ENOBUFS;
6039
6040	/* call_fib6_entry_notifiers will be removed when in-kernel notifier
6041	 * is implemented and supported for nexthop objects
6042	 */
6043	call_fib6_entry_notifiers(net, FIB_EVENT_ENTRY_REPLACE, rt, NULL);
6044
6045	skb = nlmsg_new(rt6_nlmsg_size(rt), gfp_any());
6046	if (!skb)
6047		goto errout;
6048
6049	err = rt6_fill_node(net, skb, rt, NULL, NULL, NULL, 0,
6050			    RTM_NEWROUTE, info->portid, seq, NLM_F_REPLACE);
6051	if (err < 0) {
6052		/* -EMSGSIZE implies BUG in rt6_nlmsg_size() */
6053		WARN_ON(err == -EMSGSIZE);
6054		kfree_skb(skb);
6055		goto errout;
6056	}
6057	rtnl_notify(skb, net, info->portid, RTNLGRP_IPV6_ROUTE,
6058		    info->nlh, gfp_any());
6059	return;
6060errout:
6061	if (err < 0)
6062		rtnl_set_sk_err(net, RTNLGRP_IPV6_ROUTE, err);
6063}
6064
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
6065static int ip6_route_dev_notify(struct notifier_block *this,
6066				unsigned long event, void *ptr)
6067{
6068	struct net_device *dev = netdev_notifier_info_to_dev(ptr);
6069	struct net *net = dev_net(dev);
6070
6071	if (!(dev->flags & IFF_LOOPBACK))
6072		return NOTIFY_OK;
6073
6074	if (event == NETDEV_REGISTER) {
6075		net->ipv6.fib6_null_entry->fib6_nh->fib_nh_dev = dev;
6076		net->ipv6.ip6_null_entry->dst.dev = dev;
6077		net->ipv6.ip6_null_entry->rt6i_idev = in6_dev_get(dev);
6078#ifdef CONFIG_IPV6_MULTIPLE_TABLES
6079		net->ipv6.ip6_prohibit_entry->dst.dev = dev;
6080		net->ipv6.ip6_prohibit_entry->rt6i_idev = in6_dev_get(dev);
6081		net->ipv6.ip6_blk_hole_entry->dst.dev = dev;
6082		net->ipv6.ip6_blk_hole_entry->rt6i_idev = in6_dev_get(dev);
6083#endif
6084	 } else if (event == NETDEV_UNREGISTER &&
6085		    dev->reg_state != NETREG_UNREGISTERED) {
6086		/* NETDEV_UNREGISTER could be fired for multiple times by
6087		 * netdev_wait_allrefs(). Make sure we only call this once.
6088		 */
6089		in6_dev_put_clear(&net->ipv6.ip6_null_entry->rt6i_idev);
6090#ifdef CONFIG_IPV6_MULTIPLE_TABLES
6091		in6_dev_put_clear(&net->ipv6.ip6_prohibit_entry->rt6i_idev);
6092		in6_dev_put_clear(&net->ipv6.ip6_blk_hole_entry->rt6i_idev);
6093#endif
6094	}
6095
6096	return NOTIFY_OK;
6097}
6098
6099/*
6100 *	/proc
6101 */
6102
6103#ifdef CONFIG_PROC_FS
6104static int rt6_stats_seq_show(struct seq_file *seq, void *v)
6105{
6106	struct net *net = (struct net *)seq->private;
6107	seq_printf(seq, "%04x %04x %04x %04x %04x %04x %04x\n",
6108		   net->ipv6.rt6_stats->fib_nodes,
6109		   net->ipv6.rt6_stats->fib_route_nodes,
6110		   atomic_read(&net->ipv6.rt6_stats->fib_rt_alloc),
6111		   net->ipv6.rt6_stats->fib_rt_entries,
6112		   net->ipv6.rt6_stats->fib_rt_cache,
6113		   dst_entries_get_slow(&net->ipv6.ip6_dst_ops),
6114		   net->ipv6.rt6_stats->fib_discarded_routes);
6115
6116	return 0;
6117}
6118#endif	/* CONFIG_PROC_FS */
6119
6120#ifdef CONFIG_SYSCTL
6121
6122static int ipv6_sysctl_rtcache_flush(struct ctl_table *ctl, int write,
6123			      void *buffer, size_t *lenp, loff_t *ppos)
6124{
6125	struct net *net;
6126	int delay;
6127	int ret;
6128	if (!write)
6129		return -EINVAL;
6130
6131	net = (struct net *)ctl->extra1;
6132	delay = net->ipv6.sysctl.flush_delay;
6133	ret = proc_dointvec(ctl, write, buffer, lenp, ppos);
6134	if (ret)
6135		return ret;
6136
6137	fib6_run_gc(delay <= 0 ? 0 : (unsigned long)delay, net, delay > 0);
6138	return 0;
6139}
6140
6141static struct ctl_table ipv6_route_table_template[] = {
6142	{
6143		.procname	=	"flush",
6144		.data		=	&init_net.ipv6.sysctl.flush_delay,
6145		.maxlen		=	sizeof(int),
6146		.mode		=	0200,
6147		.proc_handler	=	ipv6_sysctl_rtcache_flush
6148	},
6149	{
6150		.procname	=	"gc_thresh",
6151		.data		=	&ip6_dst_ops_template.gc_thresh,
6152		.maxlen		=	sizeof(int),
6153		.mode		=	0644,
6154		.proc_handler	=	proc_dointvec,
6155	},
6156	{
6157		.procname	=	"max_size",
6158		.data		=	&init_net.ipv6.sysctl.ip6_rt_max_size,
6159		.maxlen		=	sizeof(int),
6160		.mode		=	0644,
6161		.proc_handler	=	proc_dointvec,
6162	},
6163	{
6164		.procname	=	"gc_min_interval",
6165		.data		=	&init_net.ipv6.sysctl.ip6_rt_gc_min_interval,
6166		.maxlen		=	sizeof(int),
6167		.mode		=	0644,
6168		.proc_handler	=	proc_dointvec_jiffies,
6169	},
6170	{
6171		.procname	=	"gc_timeout",
6172		.data		=	&init_net.ipv6.sysctl.ip6_rt_gc_timeout,
6173		.maxlen		=	sizeof(int),
6174		.mode		=	0644,
6175		.proc_handler	=	proc_dointvec_jiffies,
6176	},
6177	{
6178		.procname	=	"gc_interval",
6179		.data		=	&init_net.ipv6.sysctl.ip6_rt_gc_interval,
6180		.maxlen		=	sizeof(int),
6181		.mode		=	0644,
6182		.proc_handler	=	proc_dointvec_jiffies,
6183	},
6184	{
6185		.procname	=	"gc_elasticity",
6186		.data		=	&init_net.ipv6.sysctl.ip6_rt_gc_elasticity,
6187		.maxlen		=	sizeof(int),
6188		.mode		=	0644,
6189		.proc_handler	=	proc_dointvec,
6190	},
6191	{
6192		.procname	=	"mtu_expires",
6193		.data		=	&init_net.ipv6.sysctl.ip6_rt_mtu_expires,
6194		.maxlen		=	sizeof(int),
6195		.mode		=	0644,
6196		.proc_handler	=	proc_dointvec_jiffies,
6197	},
6198	{
6199		.procname	=	"min_adv_mss",
6200		.data		=	&init_net.ipv6.sysctl.ip6_rt_min_advmss,
6201		.maxlen		=	sizeof(int),
6202		.mode		=	0644,
6203		.proc_handler	=	proc_dointvec,
6204	},
6205	{
6206		.procname	=	"gc_min_interval_ms",
6207		.data		=	&init_net.ipv6.sysctl.ip6_rt_gc_min_interval,
6208		.maxlen		=	sizeof(int),
6209		.mode		=	0644,
6210		.proc_handler	=	proc_dointvec_ms_jiffies,
6211	},
6212	{
6213		.procname	=	"skip_notify_on_dev_down",
6214		.data		=	&init_net.ipv6.sysctl.skip_notify_on_dev_down,
6215		.maxlen		=	sizeof(int),
6216		.mode		=	0644,
6217		.proc_handler	=	proc_dointvec_minmax,
6218		.extra1		=	SYSCTL_ZERO,
6219		.extra2		=	SYSCTL_ONE,
6220	},
6221	{ }
6222};
6223
6224struct ctl_table * __net_init ipv6_route_sysctl_init(struct net *net)
6225{
6226	struct ctl_table *table;
6227
6228	table = kmemdup(ipv6_route_table_template,
6229			sizeof(ipv6_route_table_template),
6230			GFP_KERNEL);
6231
6232	if (table) {
6233		table[0].data = &net->ipv6.sysctl.flush_delay;
6234		table[0].extra1 = net;
6235		table[1].data = &net->ipv6.ip6_dst_ops.gc_thresh;
6236		table[2].data = &net->ipv6.sysctl.ip6_rt_max_size;
 
6237		table[3].data = &net->ipv6.sysctl.ip6_rt_gc_min_interval;
6238		table[4].data = &net->ipv6.sysctl.ip6_rt_gc_timeout;
6239		table[5].data = &net->ipv6.sysctl.ip6_rt_gc_interval;
6240		table[6].data = &net->ipv6.sysctl.ip6_rt_gc_elasticity;
6241		table[7].data = &net->ipv6.sysctl.ip6_rt_mtu_expires;
6242		table[8].data = &net->ipv6.sysctl.ip6_rt_min_advmss;
6243		table[9].data = &net->ipv6.sysctl.ip6_rt_gc_min_interval;
6244		table[10].data = &net->ipv6.sysctl.skip_notify_on_dev_down;
6245
6246		/* Don't export sysctls to unprivileged users */
6247		if (net->user_ns != &init_user_ns)
6248			table[0].procname = NULL;
6249	}
6250
6251	return table;
6252}
 
 
 
 
 
 
 
 
 
6253#endif
6254
6255static int __net_init ip6_route_net_init(struct net *net)
6256{
6257	int ret = -ENOMEM;
6258
6259	memcpy(&net->ipv6.ip6_dst_ops, &ip6_dst_ops_template,
6260	       sizeof(net->ipv6.ip6_dst_ops));
6261
6262	if (dst_entries_init(&net->ipv6.ip6_dst_ops) < 0)
6263		goto out_ip6_dst_ops;
6264
6265	net->ipv6.fib6_null_entry = fib6_info_alloc(GFP_KERNEL, true);
6266	if (!net->ipv6.fib6_null_entry)
6267		goto out_ip6_dst_entries;
6268	memcpy(net->ipv6.fib6_null_entry, &fib6_null_entry_template,
6269	       sizeof(*net->ipv6.fib6_null_entry));
6270
6271	net->ipv6.ip6_null_entry = kmemdup(&ip6_null_entry_template,
6272					   sizeof(*net->ipv6.ip6_null_entry),
6273					   GFP_KERNEL);
6274	if (!net->ipv6.ip6_null_entry)
6275		goto out_fib6_null_entry;
6276	net->ipv6.ip6_null_entry->dst.ops = &net->ipv6.ip6_dst_ops;
6277	dst_init_metrics(&net->ipv6.ip6_null_entry->dst,
6278			 ip6_template_metrics, true);
6279	INIT_LIST_HEAD(&net->ipv6.ip6_null_entry->rt6i_uncached);
6280
6281#ifdef CONFIG_IPV6_MULTIPLE_TABLES
6282	net->ipv6.fib6_has_custom_rules = false;
6283	net->ipv6.ip6_prohibit_entry = kmemdup(&ip6_prohibit_entry_template,
6284					       sizeof(*net->ipv6.ip6_prohibit_entry),
6285					       GFP_KERNEL);
6286	if (!net->ipv6.ip6_prohibit_entry)
6287		goto out_ip6_null_entry;
6288	net->ipv6.ip6_prohibit_entry->dst.ops = &net->ipv6.ip6_dst_ops;
6289	dst_init_metrics(&net->ipv6.ip6_prohibit_entry->dst,
6290			 ip6_template_metrics, true);
6291	INIT_LIST_HEAD(&net->ipv6.ip6_prohibit_entry->rt6i_uncached);
6292
6293	net->ipv6.ip6_blk_hole_entry = kmemdup(&ip6_blk_hole_entry_template,
6294					       sizeof(*net->ipv6.ip6_blk_hole_entry),
6295					       GFP_KERNEL);
6296	if (!net->ipv6.ip6_blk_hole_entry)
6297		goto out_ip6_prohibit_entry;
6298	net->ipv6.ip6_blk_hole_entry->dst.ops = &net->ipv6.ip6_dst_ops;
6299	dst_init_metrics(&net->ipv6.ip6_blk_hole_entry->dst,
6300			 ip6_template_metrics, true);
6301	INIT_LIST_HEAD(&net->ipv6.ip6_blk_hole_entry->rt6i_uncached);
6302#ifdef CONFIG_IPV6_SUBTREES
6303	net->ipv6.fib6_routes_require_src = 0;
6304#endif
6305#endif
6306
6307	net->ipv6.sysctl.flush_delay = 0;
6308	net->ipv6.sysctl.ip6_rt_max_size = 4096;
6309	net->ipv6.sysctl.ip6_rt_gc_min_interval = HZ / 2;
6310	net->ipv6.sysctl.ip6_rt_gc_timeout = 60*HZ;
6311	net->ipv6.sysctl.ip6_rt_gc_interval = 30*HZ;
6312	net->ipv6.sysctl.ip6_rt_gc_elasticity = 9;
6313	net->ipv6.sysctl.ip6_rt_mtu_expires = 10*60*HZ;
6314	net->ipv6.sysctl.ip6_rt_min_advmss = IPV6_MIN_MTU - 20 - 40;
6315	net->ipv6.sysctl.skip_notify_on_dev_down = 0;
6316
6317	net->ipv6.ip6_rt_gc_expire = 30*HZ;
6318
6319	ret = 0;
6320out:
6321	return ret;
6322
6323#ifdef CONFIG_IPV6_MULTIPLE_TABLES
6324out_ip6_prohibit_entry:
6325	kfree(net->ipv6.ip6_prohibit_entry);
6326out_ip6_null_entry:
6327	kfree(net->ipv6.ip6_null_entry);
6328#endif
6329out_fib6_null_entry:
6330	kfree(net->ipv6.fib6_null_entry);
6331out_ip6_dst_entries:
6332	dst_entries_destroy(&net->ipv6.ip6_dst_ops);
6333out_ip6_dst_ops:
6334	goto out;
6335}
6336
6337static void __net_exit ip6_route_net_exit(struct net *net)
6338{
6339	kfree(net->ipv6.fib6_null_entry);
6340	kfree(net->ipv6.ip6_null_entry);
6341#ifdef CONFIG_IPV6_MULTIPLE_TABLES
6342	kfree(net->ipv6.ip6_prohibit_entry);
6343	kfree(net->ipv6.ip6_blk_hole_entry);
6344#endif
6345	dst_entries_destroy(&net->ipv6.ip6_dst_ops);
6346}
6347
6348static int __net_init ip6_route_net_init_late(struct net *net)
6349{
6350#ifdef CONFIG_PROC_FS
6351	proc_create_net("ipv6_route", 0, net->proc_net, &ipv6_route_seq_ops,
6352			sizeof(struct ipv6_route_iter));
6353	proc_create_net_single("rt6_stats", 0444, net->proc_net,
6354			rt6_stats_seq_show, NULL);
 
 
 
 
 
 
6355#endif
6356	return 0;
6357}
6358
6359static void __net_exit ip6_route_net_exit_late(struct net *net)
6360{
6361#ifdef CONFIG_PROC_FS
6362	remove_proc_entry("ipv6_route", net->proc_net);
6363	remove_proc_entry("rt6_stats", net->proc_net);
6364#endif
6365}
6366
6367static struct pernet_operations ip6_route_net_ops = {
6368	.init = ip6_route_net_init,
6369	.exit = ip6_route_net_exit,
6370};
6371
6372static int __net_init ipv6_inetpeer_init(struct net *net)
6373{
6374	struct inet_peer_base *bp = kmalloc(sizeof(*bp), GFP_KERNEL);
6375
6376	if (!bp)
6377		return -ENOMEM;
6378	inet_peer_base_init(bp);
6379	net->ipv6.peers = bp;
6380	return 0;
6381}
6382
6383static void __net_exit ipv6_inetpeer_exit(struct net *net)
6384{
6385	struct inet_peer_base *bp = net->ipv6.peers;
6386
6387	net->ipv6.peers = NULL;
6388	inetpeer_invalidate_tree(bp);
6389	kfree(bp);
6390}
6391
6392static struct pernet_operations ipv6_inetpeer_ops = {
6393	.init	=	ipv6_inetpeer_init,
6394	.exit	=	ipv6_inetpeer_exit,
6395};
6396
6397static struct pernet_operations ip6_route_net_late_ops = {
6398	.init = ip6_route_net_init_late,
6399	.exit = ip6_route_net_exit_late,
6400};
6401
6402static struct notifier_block ip6_route_dev_notifier = {
6403	.notifier_call = ip6_route_dev_notify,
6404	.priority = ADDRCONF_NOTIFY_PRIORITY - 10,
6405};
6406
6407void __init ip6_route_init_special_entries(void)
6408{
6409	/* Registering of the loopback is done before this portion of code,
6410	 * the loopback reference in rt6_info will not be taken, do it
6411	 * manually for init_net */
6412	init_net.ipv6.fib6_null_entry->fib6_nh->fib_nh_dev = init_net.loopback_dev;
6413	init_net.ipv6.ip6_null_entry->dst.dev = init_net.loopback_dev;
6414	init_net.ipv6.ip6_null_entry->rt6i_idev = in6_dev_get(init_net.loopback_dev);
6415  #ifdef CONFIG_IPV6_MULTIPLE_TABLES
6416	init_net.ipv6.ip6_prohibit_entry->dst.dev = init_net.loopback_dev;
6417	init_net.ipv6.ip6_prohibit_entry->rt6i_idev = in6_dev_get(init_net.loopback_dev);
6418	init_net.ipv6.ip6_blk_hole_entry->dst.dev = init_net.loopback_dev;
6419	init_net.ipv6.ip6_blk_hole_entry->rt6i_idev = in6_dev_get(init_net.loopback_dev);
6420  #endif
6421}
6422
6423#if IS_BUILTIN(CONFIG_IPV6)
6424#if defined(CONFIG_BPF_SYSCALL) && defined(CONFIG_PROC_FS)
6425DEFINE_BPF_ITER_FUNC(ipv6_route, struct bpf_iter_meta *meta, struct fib6_info *rt)
6426
6427BTF_ID_LIST(btf_fib6_info_id)
6428BTF_ID(struct, fib6_info)
6429
6430static const struct bpf_iter_seq_info ipv6_route_seq_info = {
6431	.seq_ops		= &ipv6_route_seq_ops,
6432	.init_seq_private	= bpf_iter_init_seq_net,
6433	.fini_seq_private	= bpf_iter_fini_seq_net,
6434	.seq_priv_size		= sizeof(struct ipv6_route_iter),
6435};
6436
6437static struct bpf_iter_reg ipv6_route_reg_info = {
6438	.target			= "ipv6_route",
6439	.ctx_arg_info_size	= 1,
6440	.ctx_arg_info		= {
6441		{ offsetof(struct bpf_iter__ipv6_route, rt),
6442		  PTR_TO_BTF_ID_OR_NULL },
6443	},
6444	.seq_info		= &ipv6_route_seq_info,
6445};
6446
6447static int __init bpf_iter_register(void)
6448{
6449	ipv6_route_reg_info.ctx_arg_info[0].btf_id = *btf_fib6_info_id;
6450	return bpf_iter_reg_target(&ipv6_route_reg_info);
6451}
6452
6453static void bpf_iter_unregister(void)
6454{
6455	bpf_iter_unreg_target(&ipv6_route_reg_info);
6456}
6457#endif
6458#endif
6459
6460int __init ip6_route_init(void)
6461{
6462	int ret;
6463	int cpu;
6464
6465	ret = -ENOMEM;
6466	ip6_dst_ops_template.kmem_cachep =
6467		kmem_cache_create("ip6_dst_cache", sizeof(struct rt6_info), 0,
6468				  SLAB_HWCACHE_ALIGN, NULL);
6469	if (!ip6_dst_ops_template.kmem_cachep)
6470		goto out;
6471
6472	ret = dst_entries_init(&ip6_dst_blackhole_ops);
6473	if (ret)
6474		goto out_kmem_cache;
6475
6476	ret = register_pernet_subsys(&ipv6_inetpeer_ops);
6477	if (ret)
6478		goto out_dst_entries;
6479
6480	ret = register_pernet_subsys(&ip6_route_net_ops);
6481	if (ret)
6482		goto out_register_inetpeer;
6483
6484	ip6_dst_blackhole_ops.kmem_cachep = ip6_dst_ops_template.kmem_cachep;
6485
6486	ret = fib6_init();
6487	if (ret)
6488		goto out_register_subsys;
6489
6490	ret = xfrm6_init();
6491	if (ret)
6492		goto out_fib6_init;
6493
6494	ret = fib6_rules_init();
6495	if (ret)
6496		goto xfrm6_init;
6497
6498	ret = register_pernet_subsys(&ip6_route_net_late_ops);
6499	if (ret)
6500		goto fib6_rules_init;
6501
6502	ret = rtnl_register_module(THIS_MODULE, PF_INET6, RTM_NEWROUTE,
6503				   inet6_rtm_newroute, NULL, 0);
6504	if (ret < 0)
6505		goto out_register_late_subsys;
6506
6507	ret = rtnl_register_module(THIS_MODULE, PF_INET6, RTM_DELROUTE,
6508				   inet6_rtm_delroute, NULL, 0);
6509	if (ret < 0)
6510		goto out_register_late_subsys;
6511
6512	ret = rtnl_register_module(THIS_MODULE, PF_INET6, RTM_GETROUTE,
6513				   inet6_rtm_getroute, NULL,
6514				   RTNL_FLAG_DOIT_UNLOCKED);
6515	if (ret < 0)
6516		goto out_register_late_subsys;
6517
6518	ret = register_netdevice_notifier(&ip6_route_dev_notifier);
6519	if (ret)
6520		goto out_register_late_subsys;
6521
6522#if IS_BUILTIN(CONFIG_IPV6)
6523#if defined(CONFIG_BPF_SYSCALL) && defined(CONFIG_PROC_FS)
6524	ret = bpf_iter_register();
6525	if (ret)
6526		goto out_register_late_subsys;
6527#endif
6528#endif
6529
6530	for_each_possible_cpu(cpu) {
6531		struct uncached_list *ul = per_cpu_ptr(&rt6_uncached_list, cpu);
6532
6533		INIT_LIST_HEAD(&ul->head);
 
6534		spin_lock_init(&ul->lock);
6535	}
6536
6537out:
6538	return ret;
6539
6540out_register_late_subsys:
6541	rtnl_unregister_all(PF_INET6);
6542	unregister_pernet_subsys(&ip6_route_net_late_ops);
6543fib6_rules_init:
6544	fib6_rules_cleanup();
6545xfrm6_init:
6546	xfrm6_fini();
6547out_fib6_init:
6548	fib6_gc_cleanup();
6549out_register_subsys:
6550	unregister_pernet_subsys(&ip6_route_net_ops);
6551out_register_inetpeer:
6552	unregister_pernet_subsys(&ipv6_inetpeer_ops);
6553out_dst_entries:
6554	dst_entries_destroy(&ip6_dst_blackhole_ops);
6555out_kmem_cache:
6556	kmem_cache_destroy(ip6_dst_ops_template.kmem_cachep);
6557	goto out;
6558}
6559
6560void ip6_route_cleanup(void)
6561{
6562#if IS_BUILTIN(CONFIG_IPV6)
6563#if defined(CONFIG_BPF_SYSCALL) && defined(CONFIG_PROC_FS)
6564	bpf_iter_unregister();
6565#endif
6566#endif
6567	unregister_netdevice_notifier(&ip6_route_dev_notifier);
6568	unregister_pernet_subsys(&ip6_route_net_late_ops);
6569	fib6_rules_cleanup();
6570	xfrm6_fini();
6571	fib6_gc_cleanup();
6572	unregister_pernet_subsys(&ipv6_inetpeer_ops);
6573	unregister_pernet_subsys(&ip6_route_net_ops);
6574	dst_entries_destroy(&ip6_dst_blackhole_ops);
6575	kmem_cache_destroy(ip6_dst_ops_template.kmem_cachep);
6576}