route.c 65.7 KB
Newer Older
L
Linus Torvalds 已提交
1 2 3 4 5
/*
 *	Linux INET6 implementation
 *	FIB front-end.
 *
 *	Authors:
6
 *	Pedro Roque		<roque@di.fc.ul.pt>
L
Linus Torvalds 已提交
7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22
 *
 *	This program is free software; you can redistribute it and/or
 *      modify it under the terms of the GNU General Public License
 *      as published by the Free Software Foundation; either version
 *      2 of the License, or (at your option) any later version.
 */

/*	Changes:
 *
 *	YOSHIFUJI Hideaki @USAGI
 *		reworked default router selection.
 *		- respect outgoing interface
 *		- select from (probably) reachable routers (i.e.
 *		routers in REACHABLE, STALE, DELAY or PROBE states).
 *		- always select the same router if it is (probably)
 *		reachable.  otherwise, round-robin the list.
23 24
 *	Ville Nuorvala
 *		Fixed routing subtrees.
L
Linus Torvalds 已提交
25 26
 */

27
#include <linux/capability.h>
L
Linus Torvalds 已提交
28 29 30 31 32 33 34 35 36
#include <linux/errno.h>
#include <linux/types.h>
#include <linux/times.h>
#include <linux/socket.h>
#include <linux/sockios.h>
#include <linux/net.h>
#include <linux/route.h>
#include <linux/netdevice.h>
#include <linux/in6.h>
37
#include <linux/mroute6.h>
L
Linus Torvalds 已提交
38 39 40 41
#include <linux/init.h>
#include <linux/if_arp.h>
#include <linux/proc_fs.h>
#include <linux/seq_file.h>
42
#include <linux/nsproxy.h>
43
#include <net/net_namespace.h>
L
Linus Torvalds 已提交
44 45 46 47 48 49 50 51 52 53
#include <net/snmp.h>
#include <net/ipv6.h>
#include <net/ip6_fib.h>
#include <net/ip6_route.h>
#include <net/ndisc.h>
#include <net/addrconf.h>
#include <net/tcp.h>
#include <linux/rtnetlink.h>
#include <net/dst.h>
#include <net/xfrm.h>
54
#include <net/netevent.h>
55
#include <net/netlink.h>
L
Linus Torvalds 已提交
56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73

#include <asm/uaccess.h>

#ifdef CONFIG_SYSCTL
#include <linux/sysctl.h>
#endif

/* Set to 3 to get tracing. */
#define RT6_DEBUG 2

#if RT6_DEBUG >= 3
#define RDBG(x) printk x
#define RT6_TRACE(x...) printk(KERN_DEBUG x)
#else
#define RDBG(x)
#define RT6_TRACE(x...) do { ; } while (0)
#endif

74
#define CLONE_OFFLINK_ROUTE 0
L
Linus Torvalds 已提交
75 76 77 78 79 80 81

static struct rt6_info * ip6_rt_copy(struct rt6_info *ort);
static struct dst_entry	*ip6_dst_check(struct dst_entry *dst, u32 cookie);
static struct dst_entry *ip6_negative_advice(struct dst_entry *);
static void		ip6_dst_destroy(struct dst_entry *);
static void		ip6_dst_ifdown(struct dst_entry *,
				       struct net_device *dev, int how);
82
static int		 ip6_dst_gc(struct dst_ops *ops);
L
Linus Torvalds 已提交
83 84 85 86 87 88

static int		ip6_pkt_discard(struct sk_buff *skb);
static int		ip6_pkt_discard_out(struct sk_buff *skb);
static void		ip6_link_failure(struct sk_buff *skb);
static void		ip6_rt_update_pmtu(struct dst_entry *dst, u32 mtu);

89
#ifdef CONFIG_IPV6_ROUTE_INFO
90 91
static struct rt6_info *rt6_add_route_info(struct net *net,
					   struct in6_addr *prefix, int prefixlen,
92 93
					   struct in6_addr *gwaddr, int ifindex,
					   unsigned pref);
94 95
static struct rt6_info *rt6_get_route_info(struct net *net,
					   struct in6_addr *prefix, int prefixlen,
96 97 98
					   struct in6_addr *gwaddr, int ifindex);
#endif

99
static struct dst_ops ip6_dst_ops_template = {
L
Linus Torvalds 已提交
100 101 102 103 104 105 106 107 108 109
	.family			=	AF_INET6,
	.protocol		=	__constant_htons(ETH_P_IPV6),
	.gc			=	ip6_dst_gc,
	.gc_thresh		=	1024,
	.check			=	ip6_dst_check,
	.destroy		=	ip6_dst_destroy,
	.ifdown			=	ip6_dst_ifdown,
	.negative_advice	=	ip6_negative_advice,
	.link_failure		=	ip6_link_failure,
	.update_pmtu		=	ip6_rt_update_pmtu,
110
	.local_out		=	__ip6_local_out,
L
Linus Torvalds 已提交
111
	.entry_size		=	sizeof(struct rt6_info),
112
	.entries		=	ATOMIC_INIT(0),
L
Linus Torvalds 已提交
113 114
};

115 116 117 118 119 120 121 122 123 124 125
static void ip6_rt_blackhole_update_pmtu(struct dst_entry *dst, u32 mtu)
{
}

static struct dst_ops ip6_dst_blackhole_ops = {
	.family			=	AF_INET6,
	.protocol		=	__constant_htons(ETH_P_IPV6),
	.destroy		=	ip6_dst_destroy,
	.check			=	ip6_dst_check,
	.update_pmtu		=	ip6_rt_blackhole_update_pmtu,
	.entry_size		=	sizeof(struct rt6_info),
126
	.entries		=	ATOMIC_INIT(0),
127 128
};

129
static struct rt6_info ip6_null_entry_template = {
L
Linus Torvalds 已提交
130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145
	.u = {
		.dst = {
			.__refcnt	= ATOMIC_INIT(1),
			.__use		= 1,
			.obsolete	= -1,
			.error		= -ENETUNREACH,
			.metrics	= { [RTAX_HOPLIMIT - 1] = 255, },
			.input		= ip6_pkt_discard,
			.output		= ip6_pkt_discard_out,
		}
	},
	.rt6i_flags	= (RTF_REJECT | RTF_NONEXTHOP),
	.rt6i_metric	= ~(u32) 0,
	.rt6i_ref	= ATOMIC_INIT(1),
};

T
Thomas Graf 已提交
146 147
#ifdef CONFIG_IPV6_MULTIPLE_TABLES

148 149 150
static int ip6_pkt_prohibit(struct sk_buff *skb);
static int ip6_pkt_prohibit_out(struct sk_buff *skb);

151
static struct rt6_info ip6_prohibit_entry_template = {
T
Thomas Graf 已提交
152 153 154 155 156 157 158
	.u = {
		.dst = {
			.__refcnt	= ATOMIC_INIT(1),
			.__use		= 1,
			.obsolete	= -1,
			.error		= -EACCES,
			.metrics	= { [RTAX_HOPLIMIT - 1] = 255, },
159 160
			.input		= ip6_pkt_prohibit,
			.output		= ip6_pkt_prohibit_out,
T
Thomas Graf 已提交
161 162 163 164 165 166 167
		}
	},
	.rt6i_flags	= (RTF_REJECT | RTF_NONEXTHOP),
	.rt6i_metric	= ~(u32) 0,
	.rt6i_ref	= ATOMIC_INIT(1),
};

168
static struct rt6_info ip6_blk_hole_entry_template = {
T
Thomas Graf 已提交
169 170 171 172 173 174 175
	.u = {
		.dst = {
			.__refcnt	= ATOMIC_INIT(1),
			.__use		= 1,
			.obsolete	= -1,
			.error		= -EINVAL,
			.metrics	= { [RTAX_HOPLIMIT - 1] = 255, },
176 177
			.input		= dst_discard,
			.output		= dst_discard,
T
Thomas Graf 已提交
178 179 180 181 182 183 184 185 186
		}
	},
	.rt6i_flags	= (RTF_REJECT | RTF_NONEXTHOP),
	.rt6i_metric	= ~(u32) 0,
	.rt6i_ref	= ATOMIC_INIT(1),
};

#endif

L
Linus Torvalds 已提交
187
/* allocate dst with ip6_dst_ops */
188
static inline struct rt6_info *ip6_dst_alloc(struct dst_ops *ops)
L
Linus Torvalds 已提交
189
{
190
	return (struct rt6_info *)dst_alloc(ops);
L
Linus Torvalds 已提交
191 192 193 194 195 196 197 198 199 200
}

static void ip6_dst_destroy(struct dst_entry *dst)
{
	struct rt6_info *rt = (struct rt6_info *)dst;
	struct inet6_dev *idev = rt->rt6i_idev;

	if (idev != NULL) {
		rt->rt6i_idev = NULL;
		in6_dev_put(idev);
201
	}
L
Linus Torvalds 已提交
202 203 204 205 206 207 208
}

static void ip6_dst_ifdown(struct dst_entry *dst, struct net_device *dev,
			   int how)
{
	struct rt6_info *rt = (struct rt6_info *)dst;
	struct inet6_dev *idev = rt->rt6i_idev;
209
	struct net_device *loopback_dev =
210
		dev_net(dev)->loopback_dev;
L
Linus Torvalds 已提交
211

212 213 214
	if (dev != loopback_dev && idev != NULL && idev->dev == dev) {
		struct inet6_dev *loopback_idev =
			in6_dev_get(loopback_dev);
L
Linus Torvalds 已提交
215 216 217 218 219 220 221 222 223 224 225 226 227
		if (loopback_idev != NULL) {
			rt->rt6i_idev = loopback_idev;
			in6_dev_put(idev);
		}
	}
}

static __inline__ int rt6_check_expired(const struct rt6_info *rt)
{
	return (rt->rt6i_flags & RTF_EXPIRES &&
		time_after(jiffies, rt->rt6i_expires));
}

T
Thomas Graf 已提交
228 229 230
static inline int rt6_need_strict(struct in6_addr *daddr)
{
	return (ipv6_addr_type(daddr) &
231
		(IPV6_ADDR_MULTICAST | IPV6_ADDR_LINKLOCAL | IPV6_ADDR_LOOPBACK));
T
Thomas Graf 已提交
232 233
}

L
Linus Torvalds 已提交
234
/*
T
Thomas Graf 已提交
235
 *	Route lookup. Any table->tb6_lock is implied.
L
Linus Torvalds 已提交
236 237
 */

238 239
static inline struct rt6_info *rt6_device_match(struct net *net,
						    struct rt6_info *rt,
240
						    struct in6_addr *saddr,
L
Linus Torvalds 已提交
241
						    int oif,
242
						    int flags)
L
Linus Torvalds 已提交
243 244 245 246
{
	struct rt6_info *local = NULL;
	struct rt6_info *sprt;

247 248 249 250 251 252 253
	if (!oif && ipv6_addr_any(saddr))
		goto out;

	for (sprt = rt; sprt; sprt = sprt->u.dst.rt6_next) {
		struct net_device *dev = sprt->rt6i_dev;

		if (oif) {
L
Linus Torvalds 已提交
254 255 256 257 258
			if (dev->ifindex == oif)
				return sprt;
			if (dev->flags & IFF_LOOPBACK) {
				if (sprt->rt6i_idev == NULL ||
				    sprt->rt6i_idev->dev->ifindex != oif) {
259
					if (flags & RT6_LOOKUP_F_IFACE && oif)
L
Linus Torvalds 已提交
260
						continue;
261
					if (local && (!oif ||
L
Linus Torvalds 已提交
262 263 264 265 266
						      local->rt6i_idev->dev->ifindex == oif))
						continue;
				}
				local = sprt;
			}
267 268 269 270
		} else {
			if (ipv6_chk_addr(net, saddr, dev,
					  flags & RT6_LOOKUP_F_IFACE))
				return sprt;
L
Linus Torvalds 已提交
271
		}
272
	}
L
Linus Torvalds 已提交
273

274
	if (oif) {
L
Linus Torvalds 已提交
275 276 277
		if (local)
			return local;

278
		if (flags & RT6_LOOKUP_F_IFACE)
279
			return net->ipv6.ip6_null_entry;
L
Linus Torvalds 已提交
280
	}
281
out:
L
Linus Torvalds 已提交
282 283 284
	return rt;
}

285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300
#ifdef CONFIG_IPV6_ROUTER_PREF
static void rt6_probe(struct rt6_info *rt)
{
	struct neighbour *neigh = rt ? rt->rt6i_nexthop : NULL;
	/*
	 * Okay, this does not seem to be appropriate
	 * for now, however, we need to check if it
	 * is really so; aka Router Reachability Probing.
	 *
	 * Router Reachability Probe MUST be rate-limited
	 * to no more than one per minute.
	 */
	if (!neigh || (neigh->nud_state & NUD_VALID))
		return;
	read_lock_bh(&neigh->lock);
	if (!(neigh->nud_state & NUD_VALID) &&
301
	    time_after(jiffies, neigh->updated + rt->rt6i_idev->cnf.rtr_probe_interval)) {
302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320
		struct in6_addr mcaddr;
		struct in6_addr *target;

		neigh->updated = jiffies;
		read_unlock_bh(&neigh->lock);

		target = (struct in6_addr *)&neigh->primary_key;
		addrconf_addr_solict_mult(target, &mcaddr);
		ndisc_send_ns(rt->rt6i_dev, NULL, target, &mcaddr, NULL);
	} else
		read_unlock_bh(&neigh->lock);
}
#else
static inline void rt6_probe(struct rt6_info *rt)
{
	return;
}
#endif

L
Linus Torvalds 已提交
321
/*
322
 * Default Router Selection (RFC 2461 6.3.6)
L
Linus Torvalds 已提交
323
 */
D
Dave Jones 已提交
324
static inline int rt6_check_dev(struct rt6_info *rt, int oif)
325 326
{
	struct net_device *dev = rt->rt6i_dev;
327
	if (!oif || dev->ifindex == oif)
328
		return 2;
329 330 331 332
	if ((dev->flags & IFF_LOOPBACK) &&
	    rt->rt6i_idev && rt->rt6i_idev->dev->ifindex == oif)
		return 1;
	return 0;
333
}
L
Linus Torvalds 已提交
334

D
Dave Jones 已提交
335
static inline int rt6_check_neigh(struct rt6_info *rt)
L
Linus Torvalds 已提交
336
{
337
	struct neighbour *neigh = rt->rt6i_nexthop;
338
	int m;
339 340 341 342
	if (rt->rt6i_flags & RTF_NONEXTHOP ||
	    !(rt->rt6i_flags & RTF_GATEWAY))
		m = 1;
	else if (neigh) {
343 344
		read_lock_bh(&neigh->lock);
		if (neigh->nud_state & NUD_VALID)
345
			m = 2;
346 347 348 349 350
#ifdef CONFIG_IPV6_ROUTER_PREF
		else if (neigh->nud_state & NUD_FAILED)
			m = 0;
#endif
		else
351
			m = 1;
352
		read_unlock_bh(&neigh->lock);
353 354
	} else
		m = 0;
355
	return m;
L
Linus Torvalds 已提交
356 357
}

358 359
static int rt6_score_route(struct rt6_info *rt, int oif,
			   int strict)
L
Linus Torvalds 已提交
360
{
361
	int m, n;
362

363
	m = rt6_check_dev(rt, oif);
364
	if (!m && (strict & RT6_LOOKUP_F_IFACE))
365
		return -1;
366 367 368
#ifdef CONFIG_IPV6_ROUTER_PREF
	m |= IPV6_DECODE_PREF(IPV6_EXTRACT_PREF(rt->rt6i_flags)) << 2;
#endif
369
	n = rt6_check_neigh(rt);
370
	if (!n && (strict & RT6_LOOKUP_F_REACHABLE))
371 372 373 374
		return -1;
	return m;
}

375 376
static struct rt6_info *find_match(struct rt6_info *rt, int oif, int strict,
				   int *mpri, struct rt6_info *match)
377
{
378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404
	int m;

	if (rt6_check_expired(rt))
		goto out;

	m = rt6_score_route(rt, oif, strict);
	if (m < 0)
		goto out;

	if (m > *mpri) {
		if (strict & RT6_LOOKUP_F_REACHABLE)
			rt6_probe(match);
		*mpri = m;
		match = rt;
	} else if (strict & RT6_LOOKUP_F_REACHABLE) {
		rt6_probe(rt);
	}

out:
	return match;
}

static struct rt6_info *find_rr_leaf(struct fib6_node *fn,
				     struct rt6_info *rr_head,
				     u32 metric, int oif, int strict)
{
	struct rt6_info *rt, *match;
405
	int mpri = -1;
L
Linus Torvalds 已提交
406

407 408 409 410 411 412 413
	match = NULL;
	for (rt = rr_head; rt && rt->rt6i_metric == metric;
	     rt = rt->u.dst.rt6_next)
		match = find_match(rt, oif, strict, &mpri, match);
	for (rt = fn->leaf; rt && rt != rr_head && rt->rt6i_metric == metric;
	     rt = rt->u.dst.rt6_next)
		match = find_match(rt, oif, strict, &mpri, match);
L
Linus Torvalds 已提交
414

415 416
	return match;
}
L
Linus Torvalds 已提交
417

418 419 420
static struct rt6_info *rt6_select(struct fib6_node *fn, int oif, int strict)
{
	struct rt6_info *match, *rt0;
421
	struct net *net;
L
Linus Torvalds 已提交
422

423
	RT6_TRACE("%s(fn->leaf=%p, oif=%d)\n",
424
		  __func__, fn->leaf, oif);
425

426 427 428
	rt0 = fn->rr_ptr;
	if (!rt0)
		fn->rr_ptr = rt0 = fn->leaf;
L
Linus Torvalds 已提交
429

430
	match = find_rr_leaf(fn, rt0, rt0->rt6i_metric, oif, strict);
L
Linus Torvalds 已提交
431

432
	if (!match &&
433 434 435
	    (strict & RT6_LOOKUP_F_REACHABLE)) {
		struct rt6_info *next = rt0->u.dst.rt6_next;

436
		/* no entries matched; do round-robin */
437 438 439 440 441
		if (!next || next->rt6i_metric != rt0->rt6i_metric)
			next = fn->leaf;

		if (next != rt0)
			fn->rr_ptr = next;
L
Linus Torvalds 已提交
442 443
	}

444
	RT6_TRACE("%s() => %p\n",
445
		  __func__, match);
L
Linus Torvalds 已提交
446

447
	net = dev_net(rt0->rt6i_dev);
448
	return (match ? match : net->ipv6.ip6_null_entry);
L
Linus Torvalds 已提交
449 450
}

451 452 453 454
#ifdef CONFIG_IPV6_ROUTE_INFO
int rt6_route_rcv(struct net_device *dev, u8 *opt, int len,
		  struct in6_addr *gwaddr)
{
455
	struct net *net = dev_net(dev);
456 457 458
	struct route_info *rinfo = (struct route_info *) opt;
	struct in6_addr prefix_buf, *prefix;
	unsigned int pref;
459
	unsigned long lifetime;
460 461 462 463 464 465 466 467 468 469 470 471 472 473 474 475 476 477 478 479 480 481 482 483 484
	struct rt6_info *rt;

	if (len < sizeof(struct route_info)) {
		return -EINVAL;
	}

	/* Sanity check for prefix_len and length */
	if (rinfo->length > 3) {
		return -EINVAL;
	} else if (rinfo->prefix_len > 128) {
		return -EINVAL;
	} else if (rinfo->prefix_len > 64) {
		if (rinfo->length < 2) {
			return -EINVAL;
		}
	} else if (rinfo->prefix_len > 0) {
		if (rinfo->length < 1) {
			return -EINVAL;
		}
	}

	pref = rinfo->route_pref;
	if (pref == ICMPV6_ROUTER_PREF_INVALID)
		pref = ICMPV6_ROUTER_PREF_MEDIUM;

485
	lifetime = addrconf_timeout_fixup(ntohl(rinfo->lifetime), HZ);
486 487 488 489 490 491 492 493 494 495 496

	if (rinfo->length == 3)
		prefix = (struct in6_addr *)rinfo->prefix;
	else {
		/* this function is safe */
		ipv6_addr_prefix(&prefix_buf,
				 (struct in6_addr *)rinfo->prefix,
				 rinfo->prefix_len);
		prefix = &prefix_buf;
	}

497 498
	rt = rt6_get_route_info(net, prefix, rinfo->prefix_len, gwaddr,
				dev->ifindex);
499 500

	if (rt && !lifetime) {
501
		ip6_del_rt(rt);
502 503 504 505
		rt = NULL;
	}

	if (!rt && lifetime)
506
		rt = rt6_add_route_info(net, prefix, rinfo->prefix_len, gwaddr, dev->ifindex,
507 508 509 510 511 512
					pref);
	else if (rt)
		rt->rt6i_flags = RTF_ROUTEINFO |
				 (rt->rt6i_flags & ~RTF_PREF_MASK) | RTF_PREF(pref);

	if (rt) {
513
		if (!addrconf_finite_timeout(lifetime)) {
514 515 516 517 518 519 520 521 522 523 524
			rt->rt6i_flags &= ~RTF_EXPIRES;
		} else {
			rt->rt6i_expires = jiffies + HZ * lifetime;
			rt->rt6i_flags |= RTF_EXPIRES;
		}
		dst_release(&rt->u.dst);
	}
	return 0;
}
#endif

525
#define BACKTRACK(__net, saddr)			\
526
do { \
527
	if (rt == __net->ipv6.ip6_null_entry) {	\
528
		struct fib6_node *pn; \
V
Ville Nuorvala 已提交
529
		while (1) { \
530 531 532 533
			if (fn->fn_flags & RTN_TL_ROOT) \
				goto out; \
			pn = fn->parent; \
			if (FIB6_SUBTREE(pn) && FIB6_SUBTREE(pn) != fn) \
534
				fn = fib6_lookup(FIB6_SUBTREE(pn), NULL, saddr); \
535 536 537 538
			else \
				fn = pn; \
			if (fn->fn_flags & RTN_RTINFO) \
				goto restart; \
T
Thomas Graf 已提交
539 540
		} \
	} \
541
} while(0)
T
Thomas Graf 已提交
542

543 544
static struct rt6_info *ip6_pol_route_lookup(struct net *net,
					     struct fib6_table *table,
T
Thomas Graf 已提交
545
					     struct flowi *fl, int flags)
L
Linus Torvalds 已提交
546 547 548 549
{
	struct fib6_node *fn;
	struct rt6_info *rt;

T
Thomas Graf 已提交
550 551 552 553
	read_lock_bh(&table->tb6_lock);
	fn = fib6_lookup(&table->tb6_root, &fl->fl6_dst, &fl->fl6_src);
restart:
	rt = fn->leaf;
554
	rt = rt6_device_match(net, rt, &fl->fl6_src, fl->oif, flags);
555
	BACKTRACK(net, &fl->fl6_src);
T
Thomas Graf 已提交
556
out:
557
	dst_use(&rt->u.dst, jiffies);
T
Thomas Graf 已提交
558 559 560 561 562
	read_unlock_bh(&table->tb6_lock);
	return rt;

}

563 564
struct rt6_info *rt6_lookup(struct net *net, const struct in6_addr *daddr,
			    const struct in6_addr *saddr, int oif, int strict)
T
Thomas Graf 已提交
565 566 567 568 569 570 571 572 573 574
{
	struct flowi fl = {
		.oif = oif,
		.nl_u = {
			.ip6_u = {
				.daddr = *daddr,
			},
		},
	};
	struct dst_entry *dst;
575
	int flags = strict ? RT6_LOOKUP_F_IFACE : 0;
T
Thomas Graf 已提交
576

577 578 579 580 581
	if (saddr) {
		memcpy(&fl.fl6_src, saddr, sizeof(*saddr));
		flags |= RT6_LOOKUP_F_HAS_SADDR;
	}

582
	dst = fib6_rule_lookup(net, &fl, flags, ip6_pol_route_lookup);
T
Thomas Graf 已提交
583 584 585 586 587
	if (dst->error == 0)
		return (struct rt6_info *) dst;

	dst_release(dst);

L
Linus Torvalds 已提交
588 589 590
	return NULL;
}

591 592
EXPORT_SYMBOL(rt6_lookup);

T
Thomas Graf 已提交
593
/* ip6_ins_rt is called with FREE table->tb6_lock.
L
Linus Torvalds 已提交
594 595 596 597 598
   It takes new route entry, the addition fails by any reason the
   route is freed. In any case, if caller does not hold it, it may
   be destroyed.
 */

599
static int __ip6_ins_rt(struct rt6_info *rt, struct nl_info *info)
L
Linus Torvalds 已提交
600 601
{
	int err;
T
Thomas Graf 已提交
602
	struct fib6_table *table;
L
Linus Torvalds 已提交
603

T
Thomas Graf 已提交
604 605
	table = rt->rt6i_table;
	write_lock_bh(&table->tb6_lock);
606
	err = fib6_add(&table->tb6_root, rt, info);
T
Thomas Graf 已提交
607
	write_unlock_bh(&table->tb6_lock);
L
Linus Torvalds 已提交
608 609 610 611

	return err;
}

612 613
int ip6_ins_rt(struct rt6_info *rt)
{
614
	struct nl_info info = {
615
		.nl_net = dev_net(rt->rt6i_dev),
616
	};
617
	return __ip6_ins_rt(rt, &info);
618 619
}

620 621
static struct rt6_info *rt6_alloc_cow(struct rt6_info *ort, struct in6_addr *daddr,
				      struct in6_addr *saddr)
L
Linus Torvalds 已提交
622 623 624 625 626 627 628 629 630 631
{
	struct rt6_info *rt;

	/*
	 *	Clone the route.
	 */

	rt = ip6_rt_copy(ort);

	if (rt) {
632 633 634 635
		if (!(rt->rt6i_flags&RTF_GATEWAY)) {
			if (rt->rt6i_dst.plen != 128 &&
			    ipv6_addr_equal(&rt->rt6i_dst.addr, daddr))
				rt->rt6i_flags |= RTF_ANYCAST;
L
Linus Torvalds 已提交
636
			ipv6_addr_copy(&rt->rt6i_gateway, daddr);
637
		}
L
Linus Torvalds 已提交
638

639
		ipv6_addr_copy(&rt->rt6i_dst.addr, daddr);
L
Linus Torvalds 已提交
640 641 642 643 644 645 646 647 648 649 650 651 652
		rt->rt6i_dst.plen = 128;
		rt->rt6i_flags |= RTF_CACHE;
		rt->u.dst.flags |= DST_HOST;

#ifdef CONFIG_IPV6_SUBTREES
		if (rt->rt6i_src.plen && saddr) {
			ipv6_addr_copy(&rt->rt6i_src.addr, saddr);
			rt->rt6i_src.plen = 128;
		}
#endif

		rt->rt6i_nexthop = ndisc_get_neigh(rt->rt6i_dev, &rt->rt6i_gateway);

653
	}
L
Linus Torvalds 已提交
654

655 656
	return rt;
}
L
Linus Torvalds 已提交
657

658 659 660 661 662 663 664 665 666 667 668 669 670
static struct rt6_info *rt6_alloc_clone(struct rt6_info *ort, struct in6_addr *daddr)
{
	struct rt6_info *rt = ip6_rt_copy(ort);
	if (rt) {
		ipv6_addr_copy(&rt->rt6i_dst.addr, daddr);
		rt->rt6i_dst.plen = 128;
		rt->rt6i_flags |= RTF_CACHE;
		rt->u.dst.flags |= DST_HOST;
		rt->rt6i_nexthop = neigh_clone(ort->rt6i_nexthop);
	}
	return rt;
}

671 672
static struct rt6_info *ip6_pol_route(struct net *net, struct fib6_table *table, int oif,
				      struct flowi *fl, int flags)
L
Linus Torvalds 已提交
673 674
{
	struct fib6_node *fn;
675
	struct rt6_info *rt, *nrt;
T
Thomas Graf 已提交
676
	int strict = 0;
L
Linus Torvalds 已提交
677
	int attempts = 3;
678
	int err;
679
	int reachable = ipv6_devconf.forwarding ? 0 : RT6_LOOKUP_F_REACHABLE;
L
Linus Torvalds 已提交
680

681
	strict |= flags & RT6_LOOKUP_F_IFACE;
L
Linus Torvalds 已提交
682 683

relookup:
T
Thomas Graf 已提交
684
	read_lock_bh(&table->tb6_lock);
L
Linus Torvalds 已提交
685

686
restart_2:
T
Thomas Graf 已提交
687
	fn = fib6_lookup(&table->tb6_root, &fl->fl6_dst, &fl->fl6_src);
L
Linus Torvalds 已提交
688 689

restart:
690
	rt = rt6_select(fn, oif, strict | reachable);
691 692 693

	BACKTRACK(net, &fl->fl6_src);
	if (rt == net->ipv6.ip6_null_entry ||
694
	    rt->rt6i_flags & RTF_CACHE)
695
		goto out;
L
Linus Torvalds 已提交
696

697
	dst_hold(&rt->u.dst);
T
Thomas Graf 已提交
698
	read_unlock_bh(&table->tb6_lock);
699

700
	if (!rt->rt6i_nexthop && !(rt->rt6i_flags & RTF_NONEXTHOP))
T
Thomas Graf 已提交
701
		nrt = rt6_alloc_cow(rt, &fl->fl6_dst, &fl->fl6_src);
702 703
	else {
#if CLONE_OFFLINK_ROUTE
T
Thomas Graf 已提交
704
		nrt = rt6_alloc_clone(rt, &fl->fl6_dst);
705 706 707 708
#else
		goto out2;
#endif
	}
709

710
	dst_release(&rt->u.dst);
711
	rt = nrt ? : net->ipv6.ip6_null_entry;
L
Linus Torvalds 已提交
712

713 714
	dst_hold(&rt->u.dst);
	if (nrt) {
715
		err = ip6_ins_rt(nrt);
716
		if (!err)
L
Linus Torvalds 已提交
717 718 719
			goto out2;
	}

720 721 722 723
	if (--attempts <= 0)
		goto out2;

	/*
T
Thomas Graf 已提交
724
	 * Race condition! In the gap, when table->tb6_lock was
725 726 727 728 729 730
	 * released someone could insert this route.  Relookup.
	 */
	dst_release(&rt->u.dst);
	goto relookup;

out:
731 732 733 734
	if (reachable) {
		reachable = 0;
		goto restart_2;
	}
735
	dst_hold(&rt->u.dst);
T
Thomas Graf 已提交
736
	read_unlock_bh(&table->tb6_lock);
L
Linus Torvalds 已提交
737 738 739
out2:
	rt->u.dst.lastuse = jiffies;
	rt->u.dst.__use++;
T
Thomas Graf 已提交
740 741

	return rt;
L
Linus Torvalds 已提交
742 743
}

744
static struct rt6_info *ip6_pol_route_input(struct net *net, struct fib6_table *table,
745 746
					    struct flowi *fl, int flags)
{
747
	return ip6_pol_route(net, table, fl->iif, fl, flags);
748 749
}

T
Thomas Graf 已提交
750 751
void ip6_route_input(struct sk_buff *skb)
{
752
	struct ipv6hdr *iph = ipv6_hdr(skb);
753
	struct net *net = dev_net(skb->dev);
754
	int flags = RT6_LOOKUP_F_HAS_SADDR;
T
Thomas Graf 已提交
755 756 757 758 759 760
	struct flowi fl = {
		.iif = skb->dev->ifindex,
		.nl_u = {
			.ip6_u = {
				.daddr = iph->daddr,
				.saddr = iph->saddr,
A
Al Viro 已提交
761
				.flowlabel = (* (__be32 *) iph)&IPV6_FLOWINFO_MASK,
T
Thomas Graf 已提交
762 763
			},
		},
764
		.mark = skb->mark,
T
Thomas Graf 已提交
765 766
		.proto = iph->nexthdr,
	};
767 768 769

	if (rt6_need_strict(&iph->daddr))
		flags |= RT6_LOOKUP_F_IFACE;
T
Thomas Graf 已提交
770

771
	skb->dst = fib6_rule_lookup(net, &fl, flags, ip6_pol_route_input);
T
Thomas Graf 已提交
772 773
}

774
static struct rt6_info *ip6_pol_route_output(struct net *net, struct fib6_table *table,
T
Thomas Graf 已提交
775
					     struct flowi *fl, int flags)
L
Linus Torvalds 已提交
776
{
777
	return ip6_pol_route(net, table, fl->oif, fl, flags);
T
Thomas Graf 已提交
778 779
}

780 781
struct dst_entry * ip6_route_output(struct net *net, struct sock *sk,
				    struct flowi *fl)
T
Thomas Graf 已提交
782 783 784 785
{
	int flags = 0;

	if (rt6_need_strict(&fl->fl6_dst))
786
		flags |= RT6_LOOKUP_F_IFACE;
T
Thomas Graf 已提交
787

788 789
	if (!ipv6_addr_any(&fl->fl6_src))
		flags |= RT6_LOOKUP_F_HAS_SADDR;
790 791 792 793 794 795 796 797 798
	else if (sk) {
		unsigned int prefs = inet6_sk(sk)->srcprefs;
		if (prefs & IPV6_PREFER_SRC_TMP)
			flags |= RT6_LOOKUP_F_SRCPREF_TMP;
		if (prefs & IPV6_PREFER_SRC_PUBLIC)
			flags |= RT6_LOOKUP_F_SRCPREF_PUBLIC;
		if (prefs & IPV6_PREFER_SRC_COA)
			flags |= RT6_LOOKUP_F_SRCPREF_COA;
	}
799

800
	return fib6_rule_lookup(net, fl, flags, ip6_pol_route_output);
L
Linus Torvalds 已提交
801 802
}

803
EXPORT_SYMBOL(ip6_route_output);
L
Linus Torvalds 已提交
804

805 806 807 808 809 810 811 812 813 814 815 816
int ip6_dst_blackhole(struct sock *sk, struct dst_entry **dstp, struct flowi *fl)
{
	struct rt6_info *ort = (struct rt6_info *) *dstp;
	struct rt6_info *rt = (struct rt6_info *)
		dst_alloc(&ip6_dst_blackhole_ops);
	struct dst_entry *new = NULL;

	if (rt) {
		new = &rt->u.dst;

		atomic_set(&new->__refcnt, 1);
		new->__use = 1;
817 818
		new->input = dst_discard;
		new->output = dst_discard;
819 820 821 822 823 824 825 826 827 828 829 830 831 832 833 834 835 836 837 838 839 840 841 842 843 844 845 846

		memcpy(new->metrics, ort->u.dst.metrics, RTAX_MAX*sizeof(u32));
		new->dev = ort->u.dst.dev;
		if (new->dev)
			dev_hold(new->dev);
		rt->rt6i_idev = ort->rt6i_idev;
		if (rt->rt6i_idev)
			in6_dev_hold(rt->rt6i_idev);
		rt->rt6i_expires = 0;

		ipv6_addr_copy(&rt->rt6i_gateway, &ort->rt6i_gateway);
		rt->rt6i_flags = ort->rt6i_flags & ~RTF_EXPIRES;
		rt->rt6i_metric = 0;

		memcpy(&rt->rt6i_dst, &ort->rt6i_dst, sizeof(struct rt6key));
#ifdef CONFIG_IPV6_SUBTREES
		memcpy(&rt->rt6i_src, &ort->rt6i_src, sizeof(struct rt6key));
#endif

		dst_free(new);
	}

	dst_release(*dstp);
	*dstp = new;
	return (new ? 0 : -ENOMEM);
}
EXPORT_SYMBOL_GPL(ip6_dst_blackhole);

L
Linus Torvalds 已提交
847 848 849 850 851 852 853 854 855 856 857 858 859 860 861 862 863 864 865 866 867 868
/*
 *	Destination cache support functions
 */

static struct dst_entry *ip6_dst_check(struct dst_entry *dst, u32 cookie)
{
	struct rt6_info *rt;

	rt = (struct rt6_info *) dst;

	if (rt && rt->rt6i_node && (rt->rt6i_node->fn_sernum == cookie))
		return dst;

	return NULL;
}

static struct dst_entry *ip6_negative_advice(struct dst_entry *dst)
{
	struct rt6_info *rt = (struct rt6_info *) dst;

	if (rt) {
		if (rt->rt6i_flags & RTF_CACHE)
869
			ip6_del_rt(rt);
L
Linus Torvalds 已提交
870 871 872 873 874 875 876 877 878 879 880 881 882 883 884 885 886 887 888 889 890 891 892 893 894 895 896 897 898 899 900 901 902
		else
			dst_release(dst);
	}
	return NULL;
}

static void ip6_link_failure(struct sk_buff *skb)
{
	struct rt6_info *rt;

	icmpv6_send(skb, ICMPV6_DEST_UNREACH, ICMPV6_ADDR_UNREACH, 0, skb->dev);

	rt = (struct rt6_info *) skb->dst;
	if (rt) {
		if (rt->rt6i_flags&RTF_CACHE) {
			dst_set_expires(&rt->u.dst, 0);
			rt->rt6i_flags |= RTF_EXPIRES;
		} else if (rt->rt6i_node && (rt->rt6i_flags & RTF_DEFAULT))
			rt->rt6i_node->fn_sernum = -1;
	}
}

static void ip6_rt_update_pmtu(struct dst_entry *dst, u32 mtu)
{
	struct rt6_info *rt6 = (struct rt6_info*)dst;

	if (mtu < dst_mtu(dst) && rt6->rt6i_dst.plen == 128) {
		rt6->rt6i_flags |= RTF_MODIFIED;
		if (mtu < IPV6_MIN_MTU) {
			mtu = IPV6_MIN_MTU;
			dst->metrics[RTAX_FEATURES-1] |= RTAX_FEATURE_ALLFRAG;
		}
		dst->metrics[RTAX_MTU-1] = mtu;
903
		call_netevent_notifiers(NETEVENT_PMTU_UPDATE, dst);
L
Linus Torvalds 已提交
904 905 906 907 908
	}
}

static int ipv6_get_mtu(struct net_device *dev);

909
static inline unsigned int ipv6_advmss(struct net *net, unsigned int mtu)
L
Linus Torvalds 已提交
910 911 912
{
	mtu -= sizeof(struct ipv6hdr) + sizeof(struct tcphdr);

913 914
	if (mtu < net->ipv6.sysctl.ip6_rt_min_advmss)
		mtu = net->ipv6.sysctl.ip6_rt_min_advmss;
L
Linus Torvalds 已提交
915 916

	/*
917 918 919
	 * Maximal non-jumbo IPv6 payload is IPV6_MAXPLEN and
	 * corresponding MSS is IPV6_MAXPLEN - tcp_header_size.
	 * IPV6_MAXPLEN is also valid and means: "any MSS,
L
Linus Torvalds 已提交
920 921 922 923 924 925 926
	 * rely only on pmtu discovery"
	 */
	if (mtu > IPV6_MAXPLEN - sizeof(struct tcphdr))
		mtu = IPV6_MAXPLEN;
	return mtu;
}

927 928
static struct dst_entry *icmp6_dst_gc_list;
static DEFINE_SPINLOCK(icmp6_dst_lock);
929

930
struct dst_entry *icmp6_dst_alloc(struct net_device *dev,
L
Linus Torvalds 已提交
931
				  struct neighbour *neigh,
932
				  const struct in6_addr *addr)
L
Linus Torvalds 已提交
933 934 935
{
	struct rt6_info *rt;
	struct inet6_dev *idev = in6_dev_get(dev);
936
	struct net *net = dev_net(dev);
L
Linus Torvalds 已提交
937 938 939 940

	if (unlikely(idev == NULL))
		return NULL;

941
	rt = ip6_dst_alloc(net->ipv6.ip6_dst_ops);
L
Linus Torvalds 已提交
942 943 944 945 946 947 948 949 950 951 952 953 954 955 956 957 958
	if (unlikely(rt == NULL)) {
		in6_dev_put(idev);
		goto out;
	}

	dev_hold(dev);
	if (neigh)
		neigh_hold(neigh);
	else
		neigh = ndisc_get_neigh(dev, addr);

	rt->rt6i_dev	  = dev;
	rt->rt6i_idev     = idev;
	rt->rt6i_nexthop  = neigh;
	atomic_set(&rt->u.dst.__refcnt, 1);
	rt->u.dst.metrics[RTAX_HOPLIMIT-1] = 255;
	rt->u.dst.metrics[RTAX_MTU-1] = ipv6_get_mtu(rt->rt6i_dev);
959
	rt->u.dst.metrics[RTAX_ADVMSS-1] = ipv6_advmss(net, dst_mtu(&rt->u.dst));
960
	rt->u.dst.output  = ip6_output;
L
Linus Torvalds 已提交
961 962

#if 0	/* there's no chance to use these for ndisc */
963 964
	rt->u.dst.flags   = ipv6_addr_type(addr) & IPV6_ADDR_UNICAST
				? DST_HOST
L
Linus Torvalds 已提交
965 966 967 968 969
				: 0;
	ipv6_addr_copy(&rt->rt6i_dst.addr, addr);
	rt->rt6i_dst.plen = 128;
#endif

970 971 972 973
	spin_lock_bh(&icmp6_dst_lock);
	rt->u.dst.next = icmp6_dst_gc_list;
	icmp6_dst_gc_list = &rt->u.dst;
	spin_unlock_bh(&icmp6_dst_lock);
L
Linus Torvalds 已提交
974

975
	fib6_force_start_gc(net);
L
Linus Torvalds 已提交
976 977

out:
978
	return &rt->u.dst;
L
Linus Torvalds 已提交
979 980
}

981
int icmp6_dst_gc(int *more)
L
Linus Torvalds 已提交
982 983 984 985 986
{
	struct dst_entry *dst, *next, **pprev;
	int freed;

	next = NULL;
987
	freed = 0;
988

989 990
	spin_lock_bh(&icmp6_dst_lock);
	pprev = &icmp6_dst_gc_list;
991

L
Linus Torvalds 已提交
992 993 994 995 996 997 998 999 1000 1001 1002
	while ((dst = *pprev) != NULL) {
		if (!atomic_read(&dst->__refcnt)) {
			*pprev = dst->next;
			dst_free(dst);
			freed++;
		} else {
			pprev = &dst->next;
			(*more)++;
		}
	}

1003
	spin_unlock_bh(&icmp6_dst_lock);
1004

L
Linus Torvalds 已提交
1005 1006 1007
	return freed;
}

1008
static int ip6_dst_gc(struct dst_ops *ops)
L
Linus Torvalds 已提交
1009 1010
{
	unsigned long now = jiffies;
1011 1012 1013 1014 1015 1016 1017 1018 1019
	struct net *net = ops->dst_net;
	int rt_min_interval = net->ipv6.sysctl.ip6_rt_gc_min_interval;
	int rt_max_size = net->ipv6.sysctl.ip6_rt_max_size;
	int rt_elasticity = net->ipv6.sysctl.ip6_rt_gc_elasticity;
	int rt_gc_timeout = net->ipv6.sysctl.ip6_rt_gc_timeout;
	unsigned long rt_last_gc = net->ipv6.ip6_rt_last_gc;

	if (time_after(rt_last_gc + rt_min_interval, now) &&
	    atomic_read(&ops->entries) <= rt_max_size)
L
Linus Torvalds 已提交
1020 1021
		goto out;

1022 1023 1024
	net->ipv6.ip6_rt_gc_expire++;
	fib6_run_gc(net->ipv6.ip6_rt_gc_expire, net);
	net->ipv6.ip6_rt_last_gc = now;
1025 1026
	if (atomic_read(&ops->entries) < ops->gc_thresh)
		net->ipv6.ip6_rt_gc_expire = rt_gc_timeout>>1;
L
Linus Torvalds 已提交
1027
out:
1028 1029
	net->ipv6.ip6_rt_gc_expire -= net->ipv6.ip6_rt_gc_expire>>rt_elasticity;
	return (atomic_read(&ops->entries) > rt_max_size);
L
Linus Torvalds 已提交
1030 1031 1032 1033 1034 1035 1036 1037 1038 1039 1040 1041 1042 1043 1044 1045 1046 1047 1048 1049 1050
}

/* Clean host part of a prefix. Not necessary in radix tree,
   but results in cleaner routing tables.

   Remove it only when all the things will work!
 */

static int ipv6_get_mtu(struct net_device *dev)
{
	int mtu = IPV6_MIN_MTU;
	struct inet6_dev *idev;

	idev = in6_dev_get(dev);
	if (idev) {
		mtu = idev->cnf.mtu6;
		in6_dev_put(idev);
	}
	return mtu;
}

1051
int ip6_dst_hoplimit(struct dst_entry *dst)
L
Linus Torvalds 已提交
1052
{
1053 1054 1055 1056 1057 1058 1059 1060 1061
	int hoplimit = dst_metric(dst, RTAX_HOPLIMIT);
	if (hoplimit < 0) {
		struct net_device *dev = dst->dev;
		struct inet6_dev *idev = in6_dev_get(dev);
		if (idev) {
			hoplimit = idev->cnf.hop_limit;
			in6_dev_put(idev);
		} else
			hoplimit = ipv6_devconf.hop_limit;
L
Linus Torvalds 已提交
1062 1063 1064 1065 1066 1067 1068 1069
	}
	return hoplimit;
}

/*
 *
 */

1070
int ip6_route_add(struct fib6_config *cfg)
L
Linus Torvalds 已提交
1071 1072
{
	int err;
1073
	struct net *net = cfg->fc_nlinfo.nl_net;
L
Linus Torvalds 已提交
1074 1075 1076
	struct rt6_info *rt = NULL;
	struct net_device *dev = NULL;
	struct inet6_dev *idev = NULL;
T
Thomas Graf 已提交
1077
	struct fib6_table *table;
L
Linus Torvalds 已提交
1078 1079
	int addr_type;

1080
	if (cfg->fc_dst_len > 128 || cfg->fc_src_len > 128)
L
Linus Torvalds 已提交
1081 1082
		return -EINVAL;
#ifndef CONFIG_IPV6_SUBTREES
1083
	if (cfg->fc_src_len)
L
Linus Torvalds 已提交
1084 1085
		return -EINVAL;
#endif
1086
	if (cfg->fc_ifindex) {
L
Linus Torvalds 已提交
1087
		err = -ENODEV;
1088
		dev = dev_get_by_index(net, cfg->fc_ifindex);
L
Linus Torvalds 已提交
1089 1090 1091 1092 1093 1094 1095
		if (!dev)
			goto out;
		idev = in6_dev_get(dev);
		if (!idev)
			goto out;
	}

1096 1097
	if (cfg->fc_metric == 0)
		cfg->fc_metric = IP6_RT_PRIO_USER;
L
Linus Torvalds 已提交
1098

1099
	table = fib6_new_table(net, cfg->fc_table);
T
Thomas Graf 已提交
1100 1101 1102 1103 1104
	if (table == NULL) {
		err = -ENOBUFS;
		goto out;
	}

1105
	rt = ip6_dst_alloc(net->ipv6.ip6_dst_ops);
L
Linus Torvalds 已提交
1106 1107 1108 1109 1110 1111 1112

	if (rt == NULL) {
		err = -ENOMEM;
		goto out;
	}

	rt->u.dst.obsolete = -1;
1113 1114 1115
	rt->rt6i_expires = (cfg->fc_flags & RTF_EXPIRES) ?
				jiffies + clock_t_to_jiffies(cfg->fc_expires) :
				0;
L
Linus Torvalds 已提交
1116

1117 1118 1119 1120 1121
	if (cfg->fc_protocol == RTPROT_UNSPEC)
		cfg->fc_protocol = RTPROT_BOOT;
	rt->rt6i_protocol = cfg->fc_protocol;

	addr_type = ipv6_addr_type(&cfg->fc_dst);
L
Linus Torvalds 已提交
1122 1123 1124 1125 1126 1127 1128 1129

	if (addr_type & IPV6_ADDR_MULTICAST)
		rt->u.dst.input = ip6_mc_input;
	else
		rt->u.dst.input = ip6_forward;

	rt->u.dst.output = ip6_output;

1130 1131
	ipv6_addr_prefix(&rt->rt6i_dst.addr, &cfg->fc_dst, cfg->fc_dst_len);
	rt->rt6i_dst.plen = cfg->fc_dst_len;
L
Linus Torvalds 已提交
1132 1133 1134 1135
	if (rt->rt6i_dst.plen == 128)
	       rt->u.dst.flags = DST_HOST;

#ifdef CONFIG_IPV6_SUBTREES
1136 1137
	ipv6_addr_prefix(&rt->rt6i_src.addr, &cfg->fc_src, cfg->fc_src_len);
	rt->rt6i_src.plen = cfg->fc_src_len;
L
Linus Torvalds 已提交
1138 1139
#endif

1140
	rt->rt6i_metric = cfg->fc_metric;
L
Linus Torvalds 已提交
1141 1142 1143 1144

	/* We cannot add true routes via loopback here,
	   they would result in kernel looping; promote them to reject routes
	 */
1145
	if ((cfg->fc_flags & RTF_REJECT) ||
L
Linus Torvalds 已提交
1146 1147
	    (dev && (dev->flags&IFF_LOOPBACK) && !(addr_type&IPV6_ADDR_LOOPBACK))) {
		/* hold loopback dev/idev if we haven't done so. */
1148
		if (dev != net->loopback_dev) {
L
Linus Torvalds 已提交
1149 1150 1151 1152
			if (dev) {
				dev_put(dev);
				in6_dev_put(idev);
			}
1153
			dev = net->loopback_dev;
L
Linus Torvalds 已提交
1154 1155 1156 1157 1158 1159 1160 1161 1162 1163 1164 1165 1166 1167
			dev_hold(dev);
			idev = in6_dev_get(dev);
			if (!idev) {
				err = -ENODEV;
				goto out;
			}
		}
		rt->u.dst.output = ip6_pkt_discard_out;
		rt->u.dst.input = ip6_pkt_discard;
		rt->u.dst.error = -ENETUNREACH;
		rt->rt6i_flags = RTF_REJECT|RTF_NONEXTHOP;
		goto install_route;
	}

1168
	if (cfg->fc_flags & RTF_GATEWAY) {
L
Linus Torvalds 已提交
1169 1170 1171
		struct in6_addr *gw_addr;
		int gwa_type;

1172 1173
		gw_addr = &cfg->fc_gateway;
		ipv6_addr_copy(&rt->rt6i_gateway, gw_addr);
L
Linus Torvalds 已提交
1174 1175 1176 1177 1178 1179 1180 1181 1182 1183 1184 1185 1186 1187 1188 1189
		gwa_type = ipv6_addr_type(gw_addr);

		if (gwa_type != (IPV6_ADDR_LINKLOCAL|IPV6_ADDR_UNICAST)) {
			struct rt6_info *grt;

			/* IPv6 strictly inhibits using not link-local
			   addresses as nexthop address.
			   Otherwise, router will not able to send redirects.
			   It is very good, but in some (rare!) circumstances
			   (SIT, PtP, NBMA NOARP links) it is handy to allow
			   some exceptions. --ANK
			 */
			err = -EINVAL;
			if (!(gwa_type&IPV6_ADDR_UNICAST))
				goto out;

1190
			grt = rt6_lookup(net, gw_addr, NULL, cfg->fc_ifindex, 1);
L
Linus Torvalds 已提交
1191 1192 1193 1194 1195 1196 1197 1198 1199 1200 1201 1202 1203 1204 1205 1206 1207 1208 1209 1210 1211 1212 1213 1214 1215 1216 1217 1218 1219 1220 1221

			err = -EHOSTUNREACH;
			if (grt == NULL)
				goto out;
			if (dev) {
				if (dev != grt->rt6i_dev) {
					dst_release(&grt->u.dst);
					goto out;
				}
			} else {
				dev = grt->rt6i_dev;
				idev = grt->rt6i_idev;
				dev_hold(dev);
				in6_dev_hold(grt->rt6i_idev);
			}
			if (!(grt->rt6i_flags&RTF_GATEWAY))
				err = 0;
			dst_release(&grt->u.dst);

			if (err)
				goto out;
		}
		err = -EINVAL;
		if (dev == NULL || (dev->flags&IFF_LOOPBACK))
			goto out;
	}

	err = -ENODEV;
	if (dev == NULL)
		goto out;

1222
	if (cfg->fc_flags & (RTF_GATEWAY | RTF_NONEXTHOP)) {
L
Linus Torvalds 已提交
1223 1224 1225 1226 1227 1228 1229 1230
		rt->rt6i_nexthop = __neigh_lookup_errno(&nd_tbl, &rt->rt6i_gateway, dev);
		if (IS_ERR(rt->rt6i_nexthop)) {
			err = PTR_ERR(rt->rt6i_nexthop);
			rt->rt6i_nexthop = NULL;
			goto out;
		}
	}

1231
	rt->rt6i_flags = cfg->fc_flags;
L
Linus Torvalds 已提交
1232 1233

install_route:
1234 1235 1236 1237 1238
	if (cfg->fc_mx) {
		struct nlattr *nla;
		int remaining;

		nla_for_each_attr(nla, cfg->fc_mx, cfg->fc_mx_len, remaining) {
1239
			int type = nla_type(nla);
1240 1241 1242

			if (type) {
				if (type > RTAX_MAX) {
L
Linus Torvalds 已提交
1243 1244 1245
					err = -EINVAL;
					goto out;
				}
1246 1247

				rt->u.dst.metrics[type - 1] = nla_get_u32(nla);
L
Linus Torvalds 已提交
1248 1249 1250 1251
			}
		}
	}

1252
	if (dst_metric(&rt->u.dst, RTAX_HOPLIMIT) == 0)
L
Linus Torvalds 已提交
1253
		rt->u.dst.metrics[RTAX_HOPLIMIT-1] = -1;
1254
	if (!dst_metric(&rt->u.dst, RTAX_MTU))
L
Linus Torvalds 已提交
1255
		rt->u.dst.metrics[RTAX_MTU-1] = ipv6_get_mtu(dev);
1256
	if (!dst_metric(&rt->u.dst, RTAX_ADVMSS))
1257
		rt->u.dst.metrics[RTAX_ADVMSS-1] = ipv6_advmss(net, dst_mtu(&rt->u.dst));
L
Linus Torvalds 已提交
1258 1259
	rt->u.dst.dev = dev;
	rt->rt6i_idev = idev;
T
Thomas Graf 已提交
1260
	rt->rt6i_table = table;
1261

1262
	cfg->fc_nlinfo.nl_net = dev_net(dev);
1263

1264
	return __ip6_ins_rt(rt, &cfg->fc_nlinfo);
L
Linus Torvalds 已提交
1265 1266 1267 1268 1269 1270 1271

out:
	if (dev)
		dev_put(dev);
	if (idev)
		in6_dev_put(idev);
	if (rt)
1272
		dst_free(&rt->u.dst);
L
Linus Torvalds 已提交
1273 1274 1275
	return err;
}

1276
static int __ip6_del_rt(struct rt6_info *rt, struct nl_info *info)
L
Linus Torvalds 已提交
1277 1278
{
	int err;
T
Thomas Graf 已提交
1279
	struct fib6_table *table;
1280
	struct net *net = dev_net(rt->rt6i_dev);
L
Linus Torvalds 已提交
1281

1282
	if (rt == net->ipv6.ip6_null_entry)
1283 1284
		return -ENOENT;

T
Thomas Graf 已提交
1285 1286
	table = rt->rt6i_table;
	write_lock_bh(&table->tb6_lock);
L
Linus Torvalds 已提交
1287

1288
	err = fib6_del(rt, info);
L
Linus Torvalds 已提交
1289 1290
	dst_release(&rt->u.dst);

T
Thomas Graf 已提交
1291
	write_unlock_bh(&table->tb6_lock);
L
Linus Torvalds 已提交
1292 1293 1294 1295

	return err;
}

1296 1297
int ip6_del_rt(struct rt6_info *rt)
{
1298
	struct nl_info info = {
1299
		.nl_net = dev_net(rt->rt6i_dev),
1300
	};
1301
	return __ip6_del_rt(rt, &info);
1302 1303
}

1304
static int ip6_route_del(struct fib6_config *cfg)
L
Linus Torvalds 已提交
1305
{
T
Thomas Graf 已提交
1306
	struct fib6_table *table;
L
Linus Torvalds 已提交
1307 1308 1309 1310
	struct fib6_node *fn;
	struct rt6_info *rt;
	int err = -ESRCH;

1311
	table = fib6_get_table(cfg->fc_nlinfo.nl_net, cfg->fc_table);
T
Thomas Graf 已提交
1312 1313 1314 1315
	if (table == NULL)
		return err;

	read_lock_bh(&table->tb6_lock);
L
Linus Torvalds 已提交
1316

T
Thomas Graf 已提交
1317
	fn = fib6_locate(&table->tb6_root,
1318 1319
			 &cfg->fc_dst, cfg->fc_dst_len,
			 &cfg->fc_src, cfg->fc_src_len);
1320

L
Linus Torvalds 已提交
1321
	if (fn) {
1322
		for (rt = fn->leaf; rt; rt = rt->u.dst.rt6_next) {
1323
			if (cfg->fc_ifindex &&
L
Linus Torvalds 已提交
1324
			    (rt->rt6i_dev == NULL ||
1325
			     rt->rt6i_dev->ifindex != cfg->fc_ifindex))
L
Linus Torvalds 已提交
1326
				continue;
1327 1328
			if (cfg->fc_flags & RTF_GATEWAY &&
			    !ipv6_addr_equal(&cfg->fc_gateway, &rt->rt6i_gateway))
L
Linus Torvalds 已提交
1329
				continue;
1330
			if (cfg->fc_metric && cfg->fc_metric != rt->rt6i_metric)
L
Linus Torvalds 已提交
1331 1332
				continue;
			dst_hold(&rt->u.dst);
T
Thomas Graf 已提交
1333
			read_unlock_bh(&table->tb6_lock);
L
Linus Torvalds 已提交
1334

1335
			return __ip6_del_rt(rt, &cfg->fc_nlinfo);
L
Linus Torvalds 已提交
1336 1337
		}
	}
T
Thomas Graf 已提交
1338
	read_unlock_bh(&table->tb6_lock);
L
Linus Torvalds 已提交
1339 1340 1341 1342 1343 1344 1345

	return err;
}

/*
 *	Handle redirects
 */
1346 1347 1348 1349 1350
struct ip6rd_flowi {
	struct flowi fl;
	struct in6_addr gateway;
};

1351 1352
static struct rt6_info *__ip6_route_redirect(struct net *net,
					     struct fib6_table *table,
1353 1354
					     struct flowi *fl,
					     int flags)
L
Linus Torvalds 已提交
1355
{
1356 1357
	struct ip6rd_flowi *rdfl = (struct ip6rd_flowi *)fl;
	struct rt6_info *rt;
1358
	struct fib6_node *fn;
T
Thomas Graf 已提交
1359

L
Linus Torvalds 已提交
1360
	/*
1361 1362 1363 1364 1365 1366 1367 1368
	 * Get the "current" route for this destination and
	 * check if the redirect has come from approriate router.
	 *
	 * RFC 2461 specifies that redirects should only be
	 * accepted if they come from the nexthop to the target.
	 * Due to the way the routes are chosen, this notion
	 * is a bit fuzzy and one might need to check all possible
	 * routes.
L
Linus Torvalds 已提交
1369 1370
	 */

T
Thomas Graf 已提交
1371
	read_lock_bh(&table->tb6_lock);
1372
	fn = fib6_lookup(&table->tb6_root, &fl->fl6_dst, &fl->fl6_src);
1373
restart:
1374
	for (rt = fn->leaf; rt; rt = rt->u.dst.rt6_next) {
1375 1376 1377 1378 1379 1380 1381 1382 1383 1384 1385 1386
		/*
		 * Current route is on-link; redirect is always invalid.
		 *
		 * Seems, previous statement is not true. It could
		 * be node, which looks for us as on-link (f.e. proxy ndisc)
		 * But then router serving it might decide, that we should
		 * know truth 8)8) --ANK (980726).
		 */
		if (rt6_check_expired(rt))
			continue;
		if (!(rt->rt6i_flags & RTF_GATEWAY))
			continue;
1387
		if (fl->oif != rt->rt6i_dev->ifindex)
1388
			continue;
1389
		if (!ipv6_addr_equal(&rdfl->gateway, &rt->rt6i_gateway))
1390 1391 1392
			continue;
		break;
	}
1393

1394
	if (!rt)
1395 1396
		rt = net->ipv6.ip6_null_entry;
	BACKTRACK(net, &fl->fl6_src);
1397
out:
1398 1399
	dst_hold(&rt->u.dst);

T
Thomas Graf 已提交
1400
	read_unlock_bh(&table->tb6_lock);
1401

1402 1403 1404 1405 1406 1407 1408 1409
	return rt;
};

static struct rt6_info *ip6_route_redirect(struct in6_addr *dest,
					   struct in6_addr *src,
					   struct in6_addr *gateway,
					   struct net_device *dev)
{
1410
	int flags = RT6_LOOKUP_F_HAS_SADDR;
1411
	struct net *net = dev_net(dev);
1412 1413 1414 1415 1416 1417 1418 1419 1420 1421 1422 1423
	struct ip6rd_flowi rdfl = {
		.fl = {
			.oif = dev->ifindex,
			.nl_u = {
				.ip6_u = {
					.daddr = *dest,
					.saddr = *src,
				},
			},
		},
		.gateway = *gateway,
	};
1424 1425 1426

	if (rt6_need_strict(dest))
		flags |= RT6_LOOKUP_F_IFACE;
1427

1428
	return (struct rt6_info *)fib6_rule_lookup(net, (struct flowi *)&rdfl,
1429
						   flags, __ip6_route_redirect);
1430 1431 1432 1433 1434 1435 1436 1437
}

void rt6_redirect(struct in6_addr *dest, struct in6_addr *src,
		  struct in6_addr *saddr,
		  struct neighbour *neigh, u8 *lladdr, int on_link)
{
	struct rt6_info *rt, *nrt = NULL;
	struct netevent_redirect netevent;
1438
	struct net *net = dev_net(neigh->dev);
1439 1440 1441

	rt = ip6_route_redirect(dest, src, saddr, neigh->dev);

1442
	if (rt == net->ipv6.ip6_null_entry) {
L
Linus Torvalds 已提交
1443 1444 1445
		if (net_ratelimit())
			printk(KERN_DEBUG "rt6_redirect: source isn't a valid nexthop "
			       "for redirect target\n");
1446
		goto out;
L
Linus Torvalds 已提交
1447 1448 1449 1450 1451 1452
	}

	/*
	 *	We have finally decided to accept it.
	 */

1453
	neigh_update(neigh, lladdr, NUD_STALE,
L
Linus Torvalds 已提交
1454 1455 1456 1457 1458 1459 1460 1461 1462 1463 1464 1465 1466 1467 1468 1469 1470 1471 1472 1473 1474 1475 1476 1477 1478 1479 1480 1481 1482 1483 1484 1485 1486
		     NEIGH_UPDATE_F_WEAK_OVERRIDE|
		     NEIGH_UPDATE_F_OVERRIDE|
		     (on_link ? 0 : (NEIGH_UPDATE_F_OVERRIDE_ISROUTER|
				     NEIGH_UPDATE_F_ISROUTER))
		     );

	/*
	 * Redirect received -> path was valid.
	 * Look, redirects are sent only in response to data packets,
	 * so that this nexthop apparently is reachable. --ANK
	 */
	dst_confirm(&rt->u.dst);

	/* Duplicate redirect: silently ignore. */
	if (neigh == rt->u.dst.neighbour)
		goto out;

	nrt = ip6_rt_copy(rt);
	if (nrt == NULL)
		goto out;

	nrt->rt6i_flags = RTF_GATEWAY|RTF_UP|RTF_DYNAMIC|RTF_CACHE;
	if (on_link)
		nrt->rt6i_flags &= ~RTF_GATEWAY;

	ipv6_addr_copy(&nrt->rt6i_dst.addr, dest);
	nrt->rt6i_dst.plen = 128;
	nrt->u.dst.flags |= DST_HOST;

	ipv6_addr_copy(&nrt->rt6i_gateway, (struct in6_addr*)neigh->primary_key);
	nrt->rt6i_nexthop = neigh_clone(neigh);
	/* Reset pmtu, it may be better */
	nrt->u.dst.metrics[RTAX_MTU-1] = ipv6_get_mtu(neigh->dev);
1487
	nrt->u.dst.metrics[RTAX_ADVMSS-1] = ipv6_advmss(dev_net(neigh->dev),
1488
							dst_mtu(&nrt->u.dst));
L
Linus Torvalds 已提交
1489

1490
	if (ip6_ins_rt(nrt))
L
Linus Torvalds 已提交
1491 1492
		goto out;

1493 1494 1495 1496
	netevent.old = &rt->u.dst;
	netevent.new = &nrt->u.dst;
	call_netevent_notifiers(NETEVENT_REDIRECT, &netevent);

L
Linus Torvalds 已提交
1497
	if (rt->rt6i_flags&RTF_CACHE) {
1498
		ip6_del_rt(rt);
L
Linus Torvalds 已提交
1499 1500 1501 1502
		return;
	}

out:
1503
	dst_release(&rt->u.dst);
L
Linus Torvalds 已提交
1504 1505 1506 1507 1508 1509 1510 1511 1512 1513 1514 1515
	return;
}

/*
 *	Handle ICMP "packet too big" messages
 *	i.e. Path MTU discovery
 */

void rt6_pmtu_discovery(struct in6_addr *daddr, struct in6_addr *saddr,
			struct net_device *dev, u32 pmtu)
{
	struct rt6_info *rt, *nrt;
1516
	struct net *net = dev_net(dev);
L
Linus Torvalds 已提交
1517 1518
	int allfrag = 0;

1519
	rt = rt6_lookup(net, daddr, saddr, dev->ifindex, 0);
L
Linus Torvalds 已提交
1520 1521 1522 1523 1524 1525 1526 1527
	if (rt == NULL)
		return;

	if (pmtu >= dst_mtu(&rt->u.dst))
		goto out;

	if (pmtu < IPV6_MIN_MTU) {
		/*
1528
		 * According to RFC2460, PMTU is set to the IPv6 Minimum Link
L
Linus Torvalds 已提交
1529 1530 1531 1532 1533 1534 1535 1536 1537 1538 1539 1540 1541 1542 1543 1544 1545 1546 1547 1548 1549 1550 1551
		 * MTU (1280) and a fragment header should always be included
		 * after a node receiving Too Big message reporting PMTU is
		 * less than the IPv6 Minimum Link MTU.
		 */
		pmtu = IPV6_MIN_MTU;
		allfrag = 1;
	}

	/* New mtu received -> path was valid.
	   They are sent only in response to data packets,
	   so that this nexthop apparently is reachable. --ANK
	 */
	dst_confirm(&rt->u.dst);

	/* Host route. If it is static, it would be better
	   not to override it, but add new one, so that
	   when cache entry will expire old pmtu
	   would return automatically.
	 */
	if (rt->rt6i_flags & RTF_CACHE) {
		rt->u.dst.metrics[RTAX_MTU-1] = pmtu;
		if (allfrag)
			rt->u.dst.metrics[RTAX_FEATURES-1] |= RTAX_FEATURE_ALLFRAG;
1552
		dst_set_expires(&rt->u.dst, net->ipv6.sysctl.ip6_rt_mtu_expires);
L
Linus Torvalds 已提交
1553 1554 1555 1556 1557 1558 1559 1560 1561
		rt->rt6i_flags |= RTF_MODIFIED|RTF_EXPIRES;
		goto out;
	}

	/* Network route.
	   Two cases are possible:
	   1. It is connected route. Action: COW
	   2. It is gatewayed route or NONEXTHOP route. Action: clone it.
	 */
1562
	if (!rt->rt6i_nexthop && !(rt->rt6i_flags & RTF_NONEXTHOP))
1563
		nrt = rt6_alloc_cow(rt, daddr, saddr);
1564 1565
	else
		nrt = rt6_alloc_clone(rt, daddr);
1566

1567
	if (nrt) {
1568 1569 1570 1571 1572 1573 1574 1575 1576 1577
		nrt->u.dst.metrics[RTAX_MTU-1] = pmtu;
		if (allfrag)
			nrt->u.dst.metrics[RTAX_FEATURES-1] |= RTAX_FEATURE_ALLFRAG;

		/* According to RFC 1981, detecting PMTU increase shouldn't be
		 * happened within 5 mins, the recommended timer is 10 mins.
		 * Here this route expiration time is set to ip6_rt_mtu_expires
		 * which is 10 mins. After 10 mins the decreased pmtu is expired
		 * and detecting PMTU increase will be automatically happened.
		 */
1578
		dst_set_expires(&nrt->u.dst, net->ipv6.sysctl.ip6_rt_mtu_expires);
1579 1580
		nrt->rt6i_flags |= RTF_DYNAMIC|RTF_EXPIRES;

1581
		ip6_ins_rt(nrt);
L
Linus Torvalds 已提交
1582 1583 1584 1585 1586 1587 1588 1589 1590 1591 1592
	}
out:
	dst_release(&rt->u.dst);
}

/*
 *	Misc support functions
 */

static struct rt6_info * ip6_rt_copy(struct rt6_info *ort)
{
1593
	struct net *net = dev_net(ort->rt6i_dev);
1594
	struct rt6_info *rt = ip6_dst_alloc(net->ipv6.ip6_dst_ops);
L
Linus Torvalds 已提交
1595 1596 1597 1598 1599 1600

	if (rt) {
		rt->u.dst.input = ort->u.dst.input;
		rt->u.dst.output = ort->u.dst.output;

		memcpy(rt->u.dst.metrics, ort->u.dst.metrics, RTAX_MAX*sizeof(u32));
1601
		rt->u.dst.error = ort->u.dst.error;
L
Linus Torvalds 已提交
1602 1603 1604 1605 1606 1607 1608 1609 1610 1611 1612 1613 1614 1615 1616 1617 1618
		rt->u.dst.dev = ort->u.dst.dev;
		if (rt->u.dst.dev)
			dev_hold(rt->u.dst.dev);
		rt->rt6i_idev = ort->rt6i_idev;
		if (rt->rt6i_idev)
			in6_dev_hold(rt->rt6i_idev);
		rt->u.dst.lastuse = jiffies;
		rt->rt6i_expires = 0;

		ipv6_addr_copy(&rt->rt6i_gateway, &ort->rt6i_gateway);
		rt->rt6i_flags = ort->rt6i_flags & ~RTF_EXPIRES;
		rt->rt6i_metric = 0;

		memcpy(&rt->rt6i_dst, &ort->rt6i_dst, sizeof(struct rt6key));
#ifdef CONFIG_IPV6_SUBTREES
		memcpy(&rt->rt6i_src, &ort->rt6i_src, sizeof(struct rt6key));
#endif
T
Thomas Graf 已提交
1619
		rt->rt6i_table = ort->rt6i_table;
L
Linus Torvalds 已提交
1620 1621 1622 1623
	}
	return rt;
}

1624
#ifdef CONFIG_IPV6_ROUTE_INFO
1625 1626
static struct rt6_info *rt6_get_route_info(struct net *net,
					   struct in6_addr *prefix, int prefixlen,
1627 1628 1629 1630
					   struct in6_addr *gwaddr, int ifindex)
{
	struct fib6_node *fn;
	struct rt6_info *rt = NULL;
T
Thomas Graf 已提交
1631 1632
	struct fib6_table *table;

1633
	table = fib6_get_table(net, RT6_TABLE_INFO);
T
Thomas Graf 已提交
1634 1635
	if (table == NULL)
		return NULL;
1636

T
Thomas Graf 已提交
1637 1638
	write_lock_bh(&table->tb6_lock);
	fn = fib6_locate(&table->tb6_root, prefix ,prefixlen, NULL, 0);
1639 1640 1641
	if (!fn)
		goto out;

1642
	for (rt = fn->leaf; rt; rt = rt->u.dst.rt6_next) {
1643 1644 1645 1646 1647 1648 1649 1650 1651 1652
		if (rt->rt6i_dev->ifindex != ifindex)
			continue;
		if ((rt->rt6i_flags & (RTF_ROUTEINFO|RTF_GATEWAY)) != (RTF_ROUTEINFO|RTF_GATEWAY))
			continue;
		if (!ipv6_addr_equal(&rt->rt6i_gateway, gwaddr))
			continue;
		dst_hold(&rt->u.dst);
		break;
	}
out:
T
Thomas Graf 已提交
1653
	write_unlock_bh(&table->tb6_lock);
1654 1655 1656
	return rt;
}

1657 1658
static struct rt6_info *rt6_add_route_info(struct net *net,
					   struct in6_addr *prefix, int prefixlen,
1659 1660 1661
					   struct in6_addr *gwaddr, int ifindex,
					   unsigned pref)
{
1662 1663
	struct fib6_config cfg = {
		.fc_table	= RT6_TABLE_INFO,
1664
		.fc_metric	= IP6_RT_PRIO_USER,
1665 1666 1667 1668
		.fc_ifindex	= ifindex,
		.fc_dst_len	= prefixlen,
		.fc_flags	= RTF_GATEWAY | RTF_ADDRCONF | RTF_ROUTEINFO |
				  RTF_UP | RTF_PREF(pref),
1669 1670 1671
		.fc_nlinfo.pid = 0,
		.fc_nlinfo.nlh = NULL,
		.fc_nlinfo.nl_net = net,
1672 1673 1674 1675
	};

	ipv6_addr_copy(&cfg.fc_dst, prefix);
	ipv6_addr_copy(&cfg.fc_gateway, gwaddr);
1676

1677 1678
	/* We should treat it as a default route if prefix length is 0. */
	if (!prefixlen)
1679
		cfg.fc_flags |= RTF_DEFAULT;
1680

1681
	ip6_route_add(&cfg);
1682

1683
	return rt6_get_route_info(net, prefix, prefixlen, gwaddr, ifindex);
1684 1685 1686
}
#endif

L
Linus Torvalds 已提交
1687
struct rt6_info *rt6_get_dflt_router(struct in6_addr *addr, struct net_device *dev)
1688
{
L
Linus Torvalds 已提交
1689
	struct rt6_info *rt;
T
Thomas Graf 已提交
1690
	struct fib6_table *table;
L
Linus Torvalds 已提交
1691

1692
	table = fib6_get_table(dev_net(dev), RT6_TABLE_DFLT);
T
Thomas Graf 已提交
1693 1694
	if (table == NULL)
		return NULL;
L
Linus Torvalds 已提交
1695

T
Thomas Graf 已提交
1696
	write_lock_bh(&table->tb6_lock);
1697
	for (rt = table->tb6_root.leaf; rt; rt=rt->u.dst.rt6_next) {
L
Linus Torvalds 已提交
1698
		if (dev == rt->rt6i_dev &&
1699
		    ((rt->rt6i_flags & (RTF_ADDRCONF | RTF_DEFAULT)) == (RTF_ADDRCONF | RTF_DEFAULT)) &&
L
Linus Torvalds 已提交
1700 1701 1702 1703 1704
		    ipv6_addr_equal(&rt->rt6i_gateway, addr))
			break;
	}
	if (rt)
		dst_hold(&rt->u.dst);
T
Thomas Graf 已提交
1705
	write_unlock_bh(&table->tb6_lock);
L
Linus Torvalds 已提交
1706 1707 1708 1709
	return rt;
}

struct rt6_info *rt6_add_dflt_router(struct in6_addr *gwaddr,
1710 1711
				     struct net_device *dev,
				     unsigned int pref)
L
Linus Torvalds 已提交
1712
{
1713 1714
	struct fib6_config cfg = {
		.fc_table	= RT6_TABLE_DFLT,
1715
		.fc_metric	= IP6_RT_PRIO_USER,
1716 1717 1718
		.fc_ifindex	= dev->ifindex,
		.fc_flags	= RTF_GATEWAY | RTF_ADDRCONF | RTF_DEFAULT |
				  RTF_UP | RTF_EXPIRES | RTF_PREF(pref),
1719 1720
		.fc_nlinfo.pid = 0,
		.fc_nlinfo.nlh = NULL,
1721
		.fc_nlinfo.nl_net = dev_net(dev),
1722
	};
L
Linus Torvalds 已提交
1723

1724
	ipv6_addr_copy(&cfg.fc_gateway, gwaddr);
L
Linus Torvalds 已提交
1725

1726
	ip6_route_add(&cfg);
L
Linus Torvalds 已提交
1727 1728 1729 1730

	return rt6_get_dflt_router(gwaddr, dev);
}

1731
void rt6_purge_dflt_routers(struct net *net)
L
Linus Torvalds 已提交
1732 1733
{
	struct rt6_info *rt;
T
Thomas Graf 已提交
1734 1735 1736
	struct fib6_table *table;

	/* NOTE: Keep consistent with rt6_get_dflt_router */
1737
	table = fib6_get_table(net, RT6_TABLE_DFLT);
T
Thomas Graf 已提交
1738 1739
	if (table == NULL)
		return;
L
Linus Torvalds 已提交
1740 1741

restart:
T
Thomas Graf 已提交
1742
	read_lock_bh(&table->tb6_lock);
1743
	for (rt = table->tb6_root.leaf; rt; rt = rt->u.dst.rt6_next) {
L
Linus Torvalds 已提交
1744 1745
		if (rt->rt6i_flags & (RTF_DEFAULT | RTF_ADDRCONF)) {
			dst_hold(&rt->u.dst);
T
Thomas Graf 已提交
1746
			read_unlock_bh(&table->tb6_lock);
1747
			ip6_del_rt(rt);
L
Linus Torvalds 已提交
1748 1749 1750
			goto restart;
		}
	}
T
Thomas Graf 已提交
1751
	read_unlock_bh(&table->tb6_lock);
L
Linus Torvalds 已提交
1752 1753
}

1754 1755
static void rtmsg_to_fib6_config(struct net *net,
				 struct in6_rtmsg *rtmsg,
1756 1757 1758 1759 1760 1761 1762 1763 1764 1765 1766 1767
				 struct fib6_config *cfg)
{
	memset(cfg, 0, sizeof(*cfg));

	cfg->fc_table = RT6_TABLE_MAIN;
	cfg->fc_ifindex = rtmsg->rtmsg_ifindex;
	cfg->fc_metric = rtmsg->rtmsg_metric;
	cfg->fc_expires = rtmsg->rtmsg_info;
	cfg->fc_dst_len = rtmsg->rtmsg_dst_len;
	cfg->fc_src_len = rtmsg->rtmsg_src_len;
	cfg->fc_flags = rtmsg->rtmsg_flags;

1768
	cfg->fc_nlinfo.nl_net = net;
1769

1770 1771 1772 1773 1774
	ipv6_addr_copy(&cfg->fc_dst, &rtmsg->rtmsg_dst);
	ipv6_addr_copy(&cfg->fc_src, &rtmsg->rtmsg_src);
	ipv6_addr_copy(&cfg->fc_gateway, &rtmsg->rtmsg_gateway);
}

1775
int ipv6_route_ioctl(struct net *net, unsigned int cmd, void __user *arg)
L
Linus Torvalds 已提交
1776
{
1777
	struct fib6_config cfg;
L
Linus Torvalds 已提交
1778 1779 1780 1781 1782 1783 1784 1785 1786 1787 1788 1789
	struct in6_rtmsg rtmsg;
	int err;

	switch(cmd) {
	case SIOCADDRT:		/* Add a route */
	case SIOCDELRT:		/* Delete a route */
		if (!capable(CAP_NET_ADMIN))
			return -EPERM;
		err = copy_from_user(&rtmsg, arg,
				     sizeof(struct in6_rtmsg));
		if (err)
			return -EFAULT;
1790

1791
		rtmsg_to_fib6_config(net, &rtmsg, &cfg);
1792

L
Linus Torvalds 已提交
1793 1794 1795
		rtnl_lock();
		switch (cmd) {
		case SIOCADDRT:
1796
			err = ip6_route_add(&cfg);
L
Linus Torvalds 已提交
1797 1798
			break;
		case SIOCDELRT:
1799
			err = ip6_route_del(&cfg);
L
Linus Torvalds 已提交
1800 1801 1802 1803 1804 1805 1806
			break;
		default:
			err = -EINVAL;
		}
		rtnl_unlock();

		return err;
1807
	}
L
Linus Torvalds 已提交
1808 1809 1810 1811 1812 1813 1814 1815

	return -EINVAL;
}

/*
 *	Drop the packet on the floor
 */

I
Ilpo Järvinen 已提交
1816
static int ip6_pkt_drop(struct sk_buff *skb, int code, int ipstats_mib_noroutes)
L
Linus Torvalds 已提交
1817
{
1818 1819 1820
	int type;
	switch (ipstats_mib_noroutes) {
	case IPSTATS_MIB_INNOROUTES:
1821
		type = ipv6_addr_type(&ipv6_hdr(skb)->daddr);
1822 1823 1824 1825 1826 1827 1828 1829 1830
		if (type == IPV6_ADDR_ANY || type == IPV6_ADDR_RESERVED) {
			IP6_INC_STATS(ip6_dst_idev(skb->dst), IPSTATS_MIB_INADDRERRORS);
			break;
		}
		/* FALLTHROUGH */
	case IPSTATS_MIB_OUTNOROUTES:
		IP6_INC_STATS(ip6_dst_idev(skb->dst), ipstats_mib_noroutes);
		break;
	}
1831
	icmpv6_send(skb, ICMPV6_DEST_UNREACH, code, 0, skb->dev);
L
Linus Torvalds 已提交
1832 1833 1834 1835
	kfree_skb(skb);
	return 0;
}

1836 1837
static int ip6_pkt_discard(struct sk_buff *skb)
{
1838
	return ip6_pkt_drop(skb, ICMPV6_NOROUTE, IPSTATS_MIB_INNOROUTES);
1839 1840
}

1841
static int ip6_pkt_discard_out(struct sk_buff *skb)
L
Linus Torvalds 已提交
1842 1843
{
	skb->dev = skb->dst->dev;
1844
	return ip6_pkt_drop(skb, ICMPV6_NOROUTE, IPSTATS_MIB_OUTNOROUTES);
L
Linus Torvalds 已提交
1845 1846
}

1847 1848
#ifdef CONFIG_IPV6_MULTIPLE_TABLES

1849 1850
static int ip6_pkt_prohibit(struct sk_buff *skb)
{
1851
	return ip6_pkt_drop(skb, ICMPV6_ADM_PROHIBITED, IPSTATS_MIB_INNOROUTES);
1852 1853 1854 1855 1856
}

static int ip6_pkt_prohibit_out(struct sk_buff *skb)
{
	skb->dev = skb->dst->dev;
1857
	return ip6_pkt_drop(skb, ICMPV6_ADM_PROHIBITED, IPSTATS_MIB_OUTNOROUTES);
1858 1859
}

1860 1861
#endif

L
Linus Torvalds 已提交
1862 1863 1864 1865 1866 1867 1868 1869
/*
 *	Allocate a dst for local (unicast / anycast) address.
 */

struct rt6_info *addrconf_dst_alloc(struct inet6_dev *idev,
				    const struct in6_addr *addr,
				    int anycast)
{
1870
	struct net *net = dev_net(idev->dev);
1871
	struct rt6_info *rt = ip6_dst_alloc(net->ipv6.ip6_dst_ops);
L
Linus Torvalds 已提交
1872 1873 1874 1875

	if (rt == NULL)
		return ERR_PTR(-ENOMEM);

1876
	dev_hold(net->loopback_dev);
L
Linus Torvalds 已提交
1877 1878 1879 1880 1881
	in6_dev_hold(idev);

	rt->u.dst.flags = DST_HOST;
	rt->u.dst.input = ip6_input;
	rt->u.dst.output = ip6_output;
1882
	rt->rt6i_dev = net->loopback_dev;
L
Linus Torvalds 已提交
1883 1884
	rt->rt6i_idev = idev;
	rt->u.dst.metrics[RTAX_MTU-1] = ipv6_get_mtu(rt->rt6i_dev);
1885
	rt->u.dst.metrics[RTAX_ADVMSS-1] = ipv6_advmss(net, dst_mtu(&rt->u.dst));
L
Linus Torvalds 已提交
1886 1887 1888 1889
	rt->u.dst.metrics[RTAX_HOPLIMIT-1] = -1;
	rt->u.dst.obsolete = -1;

	rt->rt6i_flags = RTF_UP | RTF_NONEXTHOP;
1890 1891 1892
	if (anycast)
		rt->rt6i_flags |= RTF_ANYCAST;
	else
L
Linus Torvalds 已提交
1893 1894 1895
		rt->rt6i_flags |= RTF_LOCAL;
	rt->rt6i_nexthop = ndisc_get_neigh(rt->rt6i_dev, &rt->rt6i_gateway);
	if (rt->rt6i_nexthop == NULL) {
1896
		dst_free(&rt->u.dst);
L
Linus Torvalds 已提交
1897 1898 1899 1900 1901
		return ERR_PTR(-ENOMEM);
	}

	ipv6_addr_copy(&rt->rt6i_dst.addr, addr);
	rt->rt6i_dst.plen = 128;
1902
	rt->rt6i_table = fib6_get_table(net, RT6_TABLE_LOCAL);
L
Linus Torvalds 已提交
1903 1904 1905 1906 1907 1908

	atomic_set(&rt->u.dst.__refcnt, 1);

	return rt;
}

1909 1910 1911 1912 1913
struct arg_dev_net {
	struct net_device *dev;
	struct net *net;
};

L
Linus Torvalds 已提交
1914 1915
static int fib6_ifdown(struct rt6_info *rt, void *arg)
{
1916 1917 1918 1919 1920
	struct net_device *dev = ((struct arg_dev_net *)arg)->dev;
	struct net *net = ((struct arg_dev_net *)arg)->net;

	if (((void *)rt->rt6i_dev == dev || dev == NULL) &&
	    rt != net->ipv6.ip6_null_entry) {
L
Linus Torvalds 已提交
1921 1922 1923 1924 1925 1926
		RT6_TRACE("deleted by ifdown %p\n", rt);
		return -1;
	}
	return 0;
}

1927
void rt6_ifdown(struct net *net, struct net_device *dev)
L
Linus Torvalds 已提交
1928
{
1929 1930 1931 1932 1933 1934
	struct arg_dev_net adn = {
		.dev = dev,
		.net = net,
	};

	fib6_clean_all(net, fib6_ifdown, 0, &adn);
L
Linus Torvalds 已提交
1935 1936 1937 1938 1939 1940 1941 1942 1943 1944 1945 1946
}

struct rt6_mtu_change_arg
{
	struct net_device *dev;
	unsigned mtu;
};

static int rt6_mtu_change_route(struct rt6_info *rt, void *p_arg)
{
	struct rt6_mtu_change_arg *arg = (struct rt6_mtu_change_arg *) p_arg;
	struct inet6_dev *idev;
1947
	struct net *net = dev_net(arg->dev);
L
Linus Torvalds 已提交
1948 1949 1950 1951 1952 1953 1954 1955 1956 1957 1958 1959 1960 1961 1962 1963 1964 1965 1966 1967 1968 1969 1970 1971 1972 1973 1974

	/* In IPv6 pmtu discovery is not optional,
	   so that RTAX_MTU lock cannot disable it.
	   We still use this lock to block changes
	   caused by addrconf/ndisc.
	*/

	idev = __in6_dev_get(arg->dev);
	if (idev == NULL)
		return 0;

	/* For administrative MTU increase, there is no way to discover
	   IPv6 PMTU increase, so PMTU increase should be updated here.
	   Since RFC 1981 doesn't include administrative MTU increase
	   update PMTU increase is a MUST. (i.e. jumbo frame)
	 */
	/*
	   If new MTU is less than route PMTU, this new MTU will be the
	   lowest MTU in the path, update the route PMTU to reflect PMTU
	   decreases; if new MTU is greater than route PMTU, and the
	   old MTU is the lowest MTU in the path, update the route PMTU
	   to reflect the increase. In this case if the other nodes' MTU
	   also have the lowest MTU, TOO BIG MESSAGE will be lead to
	   PMTU discouvery.
	 */
	if (rt->rt6i_dev == arg->dev &&
	    !dst_metric_locked(&rt->u.dst, RTAX_MTU) &&
1975
	    (dst_mtu(&rt->u.dst) >= arg->mtu ||
1976
	     (dst_mtu(&rt->u.dst) < arg->mtu &&
1977
	      dst_mtu(&rt->u.dst) == idev->cnf.mtu6))) {
L
Linus Torvalds 已提交
1978
		rt->u.dst.metrics[RTAX_MTU-1] = arg->mtu;
1979
		rt->u.dst.metrics[RTAX_ADVMSS-1] = ipv6_advmss(net, arg->mtu);
1980
	}
L
Linus Torvalds 已提交
1981 1982 1983 1984 1985
	return 0;
}

void rt6_mtu_change(struct net_device *dev, unsigned mtu)
{
T
Thomas Graf 已提交
1986 1987 1988 1989
	struct rt6_mtu_change_arg arg = {
		.dev = dev,
		.mtu = mtu,
	};
L
Linus Torvalds 已提交
1990

1991
	fib6_clean_all(dev_net(dev), rt6_mtu_change_route, 0, &arg);
L
Linus Torvalds 已提交
1992 1993
}

1994
static const struct nla_policy rtm_ipv6_policy[RTA_MAX+1] = {
1995
	[RTA_GATEWAY]           = { .len = sizeof(struct in6_addr) },
1996
	[RTA_OIF]               = { .type = NLA_U32 },
1997
	[RTA_IIF]		= { .type = NLA_U32 },
1998 1999 2000 2001 2002 2003
	[RTA_PRIORITY]          = { .type = NLA_U32 },
	[RTA_METRICS]           = { .type = NLA_NESTED },
};

static int rtm_to_fib6_config(struct sk_buff *skb, struct nlmsghdr *nlh,
			      struct fib6_config *cfg)
L
Linus Torvalds 已提交
2004
{
2005 2006 2007
	struct rtmsg *rtm;
	struct nlattr *tb[RTA_MAX+1];
	int err;
L
Linus Torvalds 已提交
2008

2009 2010 2011
	err = nlmsg_parse(nlh, sizeof(*rtm), tb, RTA_MAX, rtm_ipv6_policy);
	if (err < 0)
		goto errout;
L
Linus Torvalds 已提交
2012

2013 2014 2015 2016 2017 2018 2019 2020 2021 2022 2023 2024 2025 2026 2027
	err = -EINVAL;
	rtm = nlmsg_data(nlh);
	memset(cfg, 0, sizeof(*cfg));

	cfg->fc_table = rtm->rtm_table;
	cfg->fc_dst_len = rtm->rtm_dst_len;
	cfg->fc_src_len = rtm->rtm_src_len;
	cfg->fc_flags = RTF_UP;
	cfg->fc_protocol = rtm->rtm_protocol;

	if (rtm->rtm_type == RTN_UNREACHABLE)
		cfg->fc_flags |= RTF_REJECT;

	cfg->fc_nlinfo.pid = NETLINK_CB(skb).pid;
	cfg->fc_nlinfo.nlh = nlh;
2028
	cfg->fc_nlinfo.nl_net = sock_net(skb->sk);
2029 2030 2031 2032

	if (tb[RTA_GATEWAY]) {
		nla_memcpy(&cfg->fc_gateway, tb[RTA_GATEWAY], 16);
		cfg->fc_flags |= RTF_GATEWAY;
L
Linus Torvalds 已提交
2033
	}
2034 2035 2036 2037 2038 2039 2040 2041

	if (tb[RTA_DST]) {
		int plen = (rtm->rtm_dst_len + 7) >> 3;

		if (nla_len(tb[RTA_DST]) < plen)
			goto errout;

		nla_memcpy(&cfg->fc_dst, tb[RTA_DST], plen);
L
Linus Torvalds 已提交
2042
	}
2043 2044 2045 2046 2047 2048 2049 2050

	if (tb[RTA_SRC]) {
		int plen = (rtm->rtm_src_len + 7) >> 3;

		if (nla_len(tb[RTA_SRC]) < plen)
			goto errout;

		nla_memcpy(&cfg->fc_src, tb[RTA_SRC], plen);
L
Linus Torvalds 已提交
2051
	}
2052 2053 2054 2055 2056 2057 2058 2059 2060 2061

	if (tb[RTA_OIF])
		cfg->fc_ifindex = nla_get_u32(tb[RTA_OIF]);

	if (tb[RTA_PRIORITY])
		cfg->fc_metric = nla_get_u32(tb[RTA_PRIORITY]);

	if (tb[RTA_METRICS]) {
		cfg->fc_mx = nla_data(tb[RTA_METRICS]);
		cfg->fc_mx_len = nla_len(tb[RTA_METRICS]);
L
Linus Torvalds 已提交
2062
	}
2063 2064 2065 2066 2067 2068 2069

	if (tb[RTA_TABLE])
		cfg->fc_table = nla_get_u32(tb[RTA_TABLE]);

	err = 0;
errout:
	return err;
L
Linus Torvalds 已提交
2070 2071
}

2072
static int inet6_rtm_delroute(struct sk_buff *skb, struct nlmsghdr* nlh, void *arg)
L
Linus Torvalds 已提交
2073
{
2074 2075
	struct fib6_config cfg;
	int err;
L
Linus Torvalds 已提交
2076

2077 2078 2079 2080 2081
	err = rtm_to_fib6_config(skb, nlh, &cfg);
	if (err < 0)
		return err;

	return ip6_route_del(&cfg);
L
Linus Torvalds 已提交
2082 2083
}

2084
static int inet6_rtm_newroute(struct sk_buff *skb, struct nlmsghdr* nlh, void *arg)
L
Linus Torvalds 已提交
2085
{
2086 2087
	struct fib6_config cfg;
	int err;
L
Linus Torvalds 已提交
2088

2089 2090 2091 2092 2093
	err = rtm_to_fib6_config(skb, nlh, &cfg);
	if (err < 0)
		return err;

	return ip6_route_add(&cfg);
L
Linus Torvalds 已提交
2094 2095
}

2096 2097 2098 2099 2100 2101 2102 2103 2104 2105 2106
static inline size_t rt6_nlmsg_size(void)
{
	return NLMSG_ALIGN(sizeof(struct rtmsg))
	       + nla_total_size(16) /* RTA_SRC */
	       + nla_total_size(16) /* RTA_DST */
	       + nla_total_size(16) /* RTA_GATEWAY */
	       + nla_total_size(16) /* RTA_PREFSRC */
	       + nla_total_size(4) /* RTA_TABLE */
	       + nla_total_size(4) /* RTA_IIF */
	       + nla_total_size(4) /* RTA_OIF */
	       + nla_total_size(4) /* RTA_PRIORITY */
2107
	       + RTAX_MAX * nla_total_size(4) /* RTA_METRICS */
2108 2109 2110
	       + nla_total_size(sizeof(struct rta_cacheinfo));
}

L
Linus Torvalds 已提交
2111
static int rt6_fill_node(struct sk_buff *skb, struct rt6_info *rt,
2112 2113
			 struct in6_addr *dst, struct in6_addr *src,
			 int iif, int type, u32 pid, u32 seq,
2114
			 int prefix, int nowait, unsigned int flags)
L
Linus Torvalds 已提交
2115 2116
{
	struct rtmsg *rtm;
2117
	struct nlmsghdr *nlh;
2118
	long expires;
2119
	u32 table;
L
Linus Torvalds 已提交
2120 2121 2122 2123 2124 2125 2126 2127

	if (prefix) {	/* user wants prefix routes only */
		if (!(rt->rt6i_flags & RTF_PREFIX_RT)) {
			/* success since this is not a prefix route */
			return 1;
		}
	}

2128 2129
	nlh = nlmsg_put(skb, pid, seq, type, sizeof(*rtm), flags);
	if (nlh == NULL)
2130
		return -EMSGSIZE;
2131 2132

	rtm = nlmsg_data(nlh);
L
Linus Torvalds 已提交
2133 2134 2135 2136
	rtm->rtm_family = AF_INET6;
	rtm->rtm_dst_len = rt->rt6i_dst.plen;
	rtm->rtm_src_len = rt->rt6i_src.plen;
	rtm->rtm_tos = 0;
T
Thomas Graf 已提交
2137
	if (rt->rt6i_table)
2138
		table = rt->rt6i_table->tb6_id;
T
Thomas Graf 已提交
2139
	else
2140 2141
		table = RT6_TABLE_UNSPEC;
	rtm->rtm_table = table;
2142
	NLA_PUT_U32(skb, RTA_TABLE, table);
L
Linus Torvalds 已提交
2143 2144 2145 2146 2147 2148 2149 2150 2151 2152 2153 2154 2155 2156 2157 2158 2159 2160 2161 2162
	if (rt->rt6i_flags&RTF_REJECT)
		rtm->rtm_type = RTN_UNREACHABLE;
	else if (rt->rt6i_dev && (rt->rt6i_dev->flags&IFF_LOOPBACK))
		rtm->rtm_type = RTN_LOCAL;
	else
		rtm->rtm_type = RTN_UNICAST;
	rtm->rtm_flags = 0;
	rtm->rtm_scope = RT_SCOPE_UNIVERSE;
	rtm->rtm_protocol = rt->rt6i_protocol;
	if (rt->rt6i_flags&RTF_DYNAMIC)
		rtm->rtm_protocol = RTPROT_REDIRECT;
	else if (rt->rt6i_flags & RTF_ADDRCONF)
		rtm->rtm_protocol = RTPROT_KERNEL;
	else if (rt->rt6i_flags&RTF_DEFAULT)
		rtm->rtm_protocol = RTPROT_RA;

	if (rt->rt6i_flags&RTF_CACHE)
		rtm->rtm_flags |= RTM_F_CLONED;

	if (dst) {
2163
		NLA_PUT(skb, RTA_DST, 16, dst);
2164
		rtm->rtm_dst_len = 128;
L
Linus Torvalds 已提交
2165
	} else if (rtm->rtm_dst_len)
2166
		NLA_PUT(skb, RTA_DST, 16, &rt->rt6i_dst.addr);
L
Linus Torvalds 已提交
2167 2168
#ifdef CONFIG_IPV6_SUBTREES
	if (src) {
2169
		NLA_PUT(skb, RTA_SRC, 16, src);
2170
		rtm->rtm_src_len = 128;
L
Linus Torvalds 已提交
2171
	} else if (rtm->rtm_src_len)
2172
		NLA_PUT(skb, RTA_SRC, 16, &rt->rt6i_src.addr);
L
Linus Torvalds 已提交
2173
#endif
2174 2175 2176 2177 2178 2179 2180 2181 2182 2183 2184 2185 2186 2187 2188 2189 2190 2191
	if (iif) {
#ifdef CONFIG_IPV6_MROUTE
		if (ipv6_addr_is_multicast(&rt->rt6i_dst.addr)) {
			int err = ip6mr_get_route(skb, rtm, nowait);
			if (err <= 0) {
				if (!nowait) {
					if (err == 0)
						return 0;
					goto nla_put_failure;
				} else {
					if (err == -EMSGSIZE)
						goto nla_put_failure;
				}
			}
		} else
#endif
			NLA_PUT_U32(skb, RTA_IIF, iif);
	} else if (dst) {
L
Linus Torvalds 已提交
2192
		struct in6_addr saddr_buf;
2193
		if (ipv6_dev_get_saddr(ip6_dst_idev(&rt->u.dst)->dev,
2194
				       dst, 0, &saddr_buf) == 0)
2195
			NLA_PUT(skb, RTA_PREFSRC, 16, &saddr_buf);
L
Linus Torvalds 已提交
2196
	}
2197

L
Linus Torvalds 已提交
2198
	if (rtnetlink_put_metrics(skb, rt->u.dst.metrics) < 0)
2199 2200
		goto nla_put_failure;

L
Linus Torvalds 已提交
2201
	if (rt->u.dst.neighbour)
2202 2203
		NLA_PUT(skb, RTA_GATEWAY, 16, &rt->u.dst.neighbour->primary_key);

L
Linus Torvalds 已提交
2204
	if (rt->u.dst.dev)
2205 2206 2207
		NLA_PUT_U32(skb, RTA_OIF, rt->rt6i_dev->ifindex);

	NLA_PUT_U32(skb, RTA_PRIORITY, rt->rt6i_metric);
2208

2209 2210 2211 2212 2213 2214
	if (!(rt->rt6i_flags & RTF_EXPIRES))
		expires = 0;
	else if (rt->rt6i_expires - jiffies < INT_MAX)
		expires = rt->rt6i_expires - jiffies;
	else
		expires = INT_MAX;
2215

2216 2217 2218
	if (rtnl_put_cacheinfo(skb, &rt->u.dst, 0, 0, 0,
			       expires, rt->u.dst.error) < 0)
		goto nla_put_failure;
2219 2220 2221 2222

	return nlmsg_end(skb, nlh);

nla_put_failure:
2223 2224
	nlmsg_cancel(skb, nlh);
	return -EMSGSIZE;
L
Linus Torvalds 已提交
2225 2226
}

2227
int rt6_dump_route(struct rt6_info *rt, void *p_arg)
L
Linus Torvalds 已提交
2228 2229 2230 2231
{
	struct rt6_rtnl_dump_arg *arg = (struct rt6_rtnl_dump_arg *) p_arg;
	int prefix;

2232 2233
	if (nlmsg_len(arg->cb->nlh) >= sizeof(struct rtmsg)) {
		struct rtmsg *rtm = nlmsg_data(arg->cb->nlh);
L
Linus Torvalds 已提交
2234 2235 2236 2237 2238 2239
		prefix = (rtm->rtm_flags & RTM_F_PREFIX) != 0;
	} else
		prefix = 0;

	return rt6_fill_node(arg->skb, rt, NULL, NULL, 0, RTM_NEWROUTE,
		     NETLINK_CB(arg->cb->skb).pid, arg->cb->nlh->nlmsg_seq,
2240
		     prefix, 0, NLM_F_MULTI);
L
Linus Torvalds 已提交
2241 2242
}

2243
static int inet6_rtm_getroute(struct sk_buff *in_skb, struct nlmsghdr* nlh, void *arg)
L
Linus Torvalds 已提交
2244
{
2245
	struct net *net = sock_net(in_skb->sk);
2246 2247
	struct nlattr *tb[RTA_MAX+1];
	struct rt6_info *rt;
L
Linus Torvalds 已提交
2248
	struct sk_buff *skb;
2249
	struct rtmsg *rtm;
L
Linus Torvalds 已提交
2250
	struct flowi fl;
2251
	int err, iif = 0;
L
Linus Torvalds 已提交
2252

2253 2254 2255
	err = nlmsg_parse(nlh, sizeof(*rtm), tb, RTA_MAX, rtm_ipv6_policy);
	if (err < 0)
		goto errout;
L
Linus Torvalds 已提交
2256

2257
	err = -EINVAL;
L
Linus Torvalds 已提交
2258 2259
	memset(&fl, 0, sizeof(fl));

2260 2261 2262 2263 2264 2265 2266 2267 2268 2269 2270 2271 2272 2273 2274 2275 2276 2277 2278
	if (tb[RTA_SRC]) {
		if (nla_len(tb[RTA_SRC]) < sizeof(struct in6_addr))
			goto errout;

		ipv6_addr_copy(&fl.fl6_src, nla_data(tb[RTA_SRC]));
	}

	if (tb[RTA_DST]) {
		if (nla_len(tb[RTA_DST]) < sizeof(struct in6_addr))
			goto errout;

		ipv6_addr_copy(&fl.fl6_dst, nla_data(tb[RTA_DST]));
	}

	if (tb[RTA_IIF])
		iif = nla_get_u32(tb[RTA_IIF]);

	if (tb[RTA_OIF])
		fl.oif = nla_get_u32(tb[RTA_OIF]);
L
Linus Torvalds 已提交
2279 2280 2281

	if (iif) {
		struct net_device *dev;
2282
		dev = __dev_get_by_index(net, iif);
L
Linus Torvalds 已提交
2283 2284
		if (!dev) {
			err = -ENODEV;
2285
			goto errout;
L
Linus Torvalds 已提交
2286 2287 2288
		}
	}

2289 2290 2291 2292 2293
	skb = alloc_skb(NLMSG_GOODSIZE, GFP_KERNEL);
	if (skb == NULL) {
		err = -ENOBUFS;
		goto errout;
	}
L
Linus Torvalds 已提交
2294

2295 2296 2297
	/* Reserve room for dummy headers, this skb can pass
	   through good chunk of routing engine.
	 */
2298
	skb_reset_mac_header(skb);
2299
	skb_reserve(skb, MAX_HEADER + sizeof(struct ipv6hdr));
L
Linus Torvalds 已提交
2300

2301
	rt = (struct rt6_info*) ip6_route_output(net, NULL, &fl);
L
Linus Torvalds 已提交
2302 2303
	skb->dst = &rt->u.dst;

2304
	err = rt6_fill_node(skb, rt, &fl.fl6_dst, &fl.fl6_src, iif,
L
Linus Torvalds 已提交
2305
			    RTM_NEWROUTE, NETLINK_CB(in_skb).pid,
2306
			    nlh->nlmsg_seq, 0, 0, 0);
L
Linus Torvalds 已提交
2307
	if (err < 0) {
2308 2309
		kfree_skb(skb);
		goto errout;
L
Linus Torvalds 已提交
2310 2311
	}

2312
	err = rtnl_unicast(skb, net, NETLINK_CB(in_skb).pid);
2313
errout:
L
Linus Torvalds 已提交
2314 2315 2316
	return err;
}

2317
void inet6_rt_notify(int event, struct rt6_info *rt, struct nl_info *info)
L
Linus Torvalds 已提交
2318 2319
{
	struct sk_buff *skb;
2320
	struct net *net = info->nl_net;
2321 2322 2323 2324 2325
	u32 seq;
	int err;

	err = -ENOBUFS;
	seq = info->nlh != NULL ? info->nlh->nlmsg_seq : 0;
2326

2327
	skb = nlmsg_new(rt6_nlmsg_size(), gfp_any());
2328 2329 2330
	if (skb == NULL)
		goto errout;

2331
	err = rt6_fill_node(skb, rt, NULL, NULL, 0,
2332
				event, info->pid, seq, 0, 0, 0);
2333 2334 2335 2336 2337 2338
	if (err < 0) {
		/* -EMSGSIZE implies BUG in rt6_nlmsg_size() */
		WARN_ON(err == -EMSGSIZE);
		kfree_skb(skb);
		goto errout;
	}
2339 2340
	err = rtnl_notify(skb, net, info->pid, RTNLGRP_IPV6_ROUTE,
			  info->nlh, gfp_any());
2341 2342
errout:
	if (err < 0)
2343
		rtnl_set_sk_err(net, RTNLGRP_IPV6_ROUTE, err);
L
Linus Torvalds 已提交
2344 2345
}

2346 2347 2348 2349
static int ip6_route_dev_notify(struct notifier_block *this,
				unsigned long event, void *data)
{
	struct net_device *dev = (struct net_device *)data;
2350
	struct net *net = dev_net(dev);
2351 2352 2353 2354 2355 2356 2357 2358 2359 2360 2361 2362 2363 2364 2365

	if (event == NETDEV_REGISTER && (dev->flags & IFF_LOOPBACK)) {
		net->ipv6.ip6_null_entry->u.dst.dev = dev;
		net->ipv6.ip6_null_entry->rt6i_idev = in6_dev_get(dev);
#ifdef CONFIG_IPV6_MULTIPLE_TABLES
		net->ipv6.ip6_prohibit_entry->u.dst.dev = dev;
		net->ipv6.ip6_prohibit_entry->rt6i_idev = in6_dev_get(dev);
		net->ipv6.ip6_blk_hole_entry->u.dst.dev = dev;
		net->ipv6.ip6_blk_hole_entry->rt6i_idev = in6_dev_get(dev);
#endif
	}

	return NOTIFY_OK;
}

L
Linus Torvalds 已提交
2366 2367 2368 2369 2370 2371 2372 2373 2374 2375 2376 2377 2378 2379 2380 2381 2382 2383 2384
/*
 *	/proc
 */

#ifdef CONFIG_PROC_FS

#define RT6_INFO_LEN (32 + 4 + 32 + 4 + 32 + 40 + 5 + 1)

struct rt6_proc_arg
{
	char *buffer;
	int offset;
	int length;
	int skip;
	int len;
};

static int rt6_info_route(struct rt6_info *rt, void *p_arg)
{
2385
	struct seq_file *m = p_arg;
L
Linus Torvalds 已提交
2386

2387 2388
	seq_printf(m, NIP6_SEQFMT " %02x ", NIP6(rt->rt6i_dst.addr),
		   rt->rt6i_dst.plen);
L
Linus Torvalds 已提交
2389 2390

#ifdef CONFIG_IPV6_SUBTREES
2391 2392
	seq_printf(m, NIP6_SEQFMT " %02x ", NIP6(rt->rt6i_src.addr),
		   rt->rt6i_src.plen);
L
Linus Torvalds 已提交
2393
#else
2394
	seq_puts(m, "00000000000000000000000000000000 00 ");
L
Linus Torvalds 已提交
2395 2396 2397
#endif

	if (rt->rt6i_nexthop) {
2398 2399
		seq_printf(m, NIP6_SEQFMT,
			   NIP6(*((struct in6_addr *)rt->rt6i_nexthop->primary_key)));
L
Linus Torvalds 已提交
2400
	} else {
2401
		seq_puts(m, "00000000000000000000000000000000");
L
Linus Torvalds 已提交
2402
	}
2403 2404 2405 2406
	seq_printf(m, " %08x %08x %08x %08x %8s\n",
		   rt->rt6i_metric, atomic_read(&rt->u.dst.__refcnt),
		   rt->u.dst.__use, rt->rt6i_flags,
		   rt->rt6i_dev ? rt->rt6i_dev->name : "");
L
Linus Torvalds 已提交
2407 2408 2409
	return 0;
}

2410
static int ipv6_route_show(struct seq_file *m, void *v)
L
Linus Torvalds 已提交
2411
{
2412 2413
	struct net *net = (struct net *)m->private;
	fib6_clean_all(net, rt6_info_route, 0, m);
2414 2415
	return 0;
}
L
Linus Torvalds 已提交
2416

2417 2418
static int ipv6_route_open(struct inode *inode, struct file *file)
{
2419
	return single_open_net(inode, file, ipv6_route_show);
2420 2421
}

2422 2423 2424 2425 2426
static const struct file_operations ipv6_route_proc_fops = {
	.owner		= THIS_MODULE,
	.open		= ipv6_route_open,
	.read		= seq_read,
	.llseek		= seq_lseek,
2427
	.release	= single_release_net,
2428 2429
};

L
Linus Torvalds 已提交
2430 2431
static int rt6_stats_seq_show(struct seq_file *seq, void *v)
{
2432
	struct net *net = (struct net *)seq->private;
L
Linus Torvalds 已提交
2433
	seq_printf(seq, "%04x %04x %04x %04x %04x %04x %04x\n",
2434 2435 2436 2437 2438
		   net->ipv6.rt6_stats->fib_nodes,
		   net->ipv6.rt6_stats->fib_route_nodes,
		   net->ipv6.rt6_stats->fib_rt_alloc,
		   net->ipv6.rt6_stats->fib_rt_entries,
		   net->ipv6.rt6_stats->fib_rt_cache,
2439
		   atomic_read(&net->ipv6.ip6_dst_ops->entries),
2440
		   net->ipv6.rt6_stats->fib_discarded_routes);
L
Linus Torvalds 已提交
2441 2442 2443 2444 2445 2446

	return 0;
}

static int rt6_stats_seq_open(struct inode *inode, struct file *file)
{
2447
	return single_open_net(inode, file, rt6_stats_seq_show);
2448 2449
}

2450
static const struct file_operations rt6_stats_seq_fops = {
L
Linus Torvalds 已提交
2451 2452 2453 2454
	.owner	 = THIS_MODULE,
	.open	 = rt6_stats_seq_open,
	.read	 = seq_read,
	.llseek	 = seq_lseek,
2455
	.release = single_release_net,
L
Linus Torvalds 已提交
2456 2457 2458 2459 2460 2461 2462 2463 2464
};
#endif	/* CONFIG_PROC_FS */

#ifdef CONFIG_SYSCTL

static
int ipv6_sysctl_rtcache_flush(ctl_table *ctl, int write, struct file * filp,
			      void __user *buffer, size_t *lenp, loff_t *ppos)
{
2465 2466
	struct net *net = current->nsproxy->net_ns;
	int delay = net->ipv6.sysctl.flush_delay;
L
Linus Torvalds 已提交
2467 2468
	if (write) {
		proc_dointvec(ctl, write, filp, buffer, lenp, ppos);
2469
		fib6_run_gc(delay <= 0 ? ~0UL : (unsigned long)delay, net);
L
Linus Torvalds 已提交
2470 2471 2472 2473 2474
		return 0;
	} else
		return -EINVAL;
}

2475
ctl_table ipv6_route_table_template[] = {
2476
	{
L
Linus Torvalds 已提交
2477
		.procname	=	"flush",
2478
		.data		=	&init_net.ipv6.sysctl.flush_delay,
L
Linus Torvalds 已提交
2479
		.maxlen		=	sizeof(int),
2480
		.mode		=	0200,
2481
		.proc_handler	=	&ipv6_sysctl_rtcache_flush
L
Linus Torvalds 已提交
2482 2483 2484 2485
	},
	{
		.ctl_name	=	NET_IPV6_ROUTE_GC_THRESH,
		.procname	=	"gc_thresh",
2486
		.data		=	&ip6_dst_ops_template.gc_thresh,
L
Linus Torvalds 已提交
2487 2488
		.maxlen		=	sizeof(int),
		.mode		=	0644,
2489
		.proc_handler	=	&proc_dointvec,
L
Linus Torvalds 已提交
2490 2491 2492 2493
	},
	{
		.ctl_name	=	NET_IPV6_ROUTE_MAX_SIZE,
		.procname	=	"max_size",
2494
		.data		=	&init_net.ipv6.sysctl.ip6_rt_max_size,
L
Linus Torvalds 已提交
2495 2496
		.maxlen		=	sizeof(int),
		.mode		=	0644,
2497
		.proc_handler	=	&proc_dointvec,
L
Linus Torvalds 已提交
2498 2499 2500 2501
	},
	{
		.ctl_name	=	NET_IPV6_ROUTE_GC_MIN_INTERVAL,
		.procname	=	"gc_min_interval",
2502
		.data		=	&init_net.ipv6.sysctl.ip6_rt_gc_min_interval,
L
Linus Torvalds 已提交
2503 2504
		.maxlen		=	sizeof(int),
		.mode		=	0644,
2505
		.proc_handler	=	&proc_dointvec_jiffies,
L
Linus Torvalds 已提交
2506 2507 2508 2509 2510
		.strategy	=	&sysctl_jiffies,
	},
	{
		.ctl_name	=	NET_IPV6_ROUTE_GC_TIMEOUT,
		.procname	=	"gc_timeout",
2511
		.data		=	&init_net.ipv6.sysctl.ip6_rt_gc_timeout,
L
Linus Torvalds 已提交
2512 2513
		.maxlen		=	sizeof(int),
		.mode		=	0644,
2514
		.proc_handler	=	&proc_dointvec_jiffies,
L
Linus Torvalds 已提交
2515 2516 2517 2518 2519
		.strategy	=	&sysctl_jiffies,
	},
	{
		.ctl_name	=	NET_IPV6_ROUTE_GC_INTERVAL,
		.procname	=	"gc_interval",
2520
		.data		=	&init_net.ipv6.sysctl.ip6_rt_gc_interval,
L
Linus Torvalds 已提交
2521 2522
		.maxlen		=	sizeof(int),
		.mode		=	0644,
2523
		.proc_handler	=	&proc_dointvec_jiffies,
L
Linus Torvalds 已提交
2524 2525 2526 2527 2528
		.strategy	=	&sysctl_jiffies,
	},
	{
		.ctl_name	=	NET_IPV6_ROUTE_GC_ELASTICITY,
		.procname	=	"gc_elasticity",
2529
		.data		=	&init_net.ipv6.sysctl.ip6_rt_gc_elasticity,
L
Linus Torvalds 已提交
2530 2531
		.maxlen		=	sizeof(int),
		.mode		=	0644,
2532
		.proc_handler	=	&proc_dointvec_jiffies,
L
Linus Torvalds 已提交
2533 2534 2535 2536 2537
		.strategy	=	&sysctl_jiffies,
	},
	{
		.ctl_name	=	NET_IPV6_ROUTE_MTU_EXPIRES,
		.procname	=	"mtu_expires",
2538
		.data		=	&init_net.ipv6.sysctl.ip6_rt_mtu_expires,
L
Linus Torvalds 已提交
2539 2540
		.maxlen		=	sizeof(int),
		.mode		=	0644,
2541
		.proc_handler	=	&proc_dointvec_jiffies,
L
Linus Torvalds 已提交
2542 2543 2544 2545 2546
		.strategy	=	&sysctl_jiffies,
	},
	{
		.ctl_name	=	NET_IPV6_ROUTE_MIN_ADVMSS,
		.procname	=	"min_adv_mss",
2547
		.data		=	&init_net.ipv6.sysctl.ip6_rt_min_advmss,
L
Linus Torvalds 已提交
2548 2549
		.maxlen		=	sizeof(int),
		.mode		=	0644,
2550
		.proc_handler	=	&proc_dointvec_jiffies,
L
Linus Torvalds 已提交
2551 2552 2553 2554 2555
		.strategy	=	&sysctl_jiffies,
	},
	{
		.ctl_name	=	NET_IPV6_ROUTE_GC_MIN_INTERVAL_MS,
		.procname	=	"gc_min_interval_ms",
2556
		.data		=	&init_net.ipv6.sysctl.ip6_rt_gc_min_interval,
L
Linus Torvalds 已提交
2557 2558
		.maxlen		=	sizeof(int),
		.mode		=	0644,
2559
		.proc_handler	=	&proc_dointvec_ms_jiffies,
L
Linus Torvalds 已提交
2560 2561 2562 2563 2564
		.strategy	=	&sysctl_ms_jiffies,
	},
	{ .ctl_name = 0 }
};

2565 2566 2567 2568 2569 2570 2571
struct ctl_table *ipv6_route_sysctl_init(struct net *net)
{
	struct ctl_table *table;

	table = kmemdup(ipv6_route_table_template,
			sizeof(ipv6_route_table_template),
			GFP_KERNEL);
2572 2573 2574

	if (table) {
		table[0].data = &net->ipv6.sysctl.flush_delay;
2575
		table[1].data = &net->ipv6.ip6_dst_ops->gc_thresh;
2576 2577 2578 2579 2580 2581 2582 2583 2584
		table[2].data = &net->ipv6.sysctl.ip6_rt_max_size;
		table[3].data = &net->ipv6.sysctl.ip6_rt_gc_min_interval;
		table[4].data = &net->ipv6.sysctl.ip6_rt_gc_timeout;
		table[5].data = &net->ipv6.sysctl.ip6_rt_gc_interval;
		table[6].data = &net->ipv6.sysctl.ip6_rt_gc_elasticity;
		table[7].data = &net->ipv6.sysctl.ip6_rt_mtu_expires;
		table[8].data = &net->ipv6.sysctl.ip6_rt_min_advmss;
	}

2585 2586
	return table;
}
L
Linus Torvalds 已提交
2587 2588
#endif

2589 2590
static int ip6_route_net_init(struct net *net)
{
2591
	int ret = -ENOMEM;
2592

2593 2594 2595 2596 2597
	net->ipv6.ip6_dst_ops = kmemdup(&ip6_dst_ops_template,
					sizeof(*net->ipv6.ip6_dst_ops),
					GFP_KERNEL);
	if (!net->ipv6.ip6_dst_ops)
		goto out;
2598
	net->ipv6.ip6_dst_ops->dst_net = hold_net(net);
2599

2600 2601 2602 2603
	net->ipv6.ip6_null_entry = kmemdup(&ip6_null_entry_template,
					   sizeof(*net->ipv6.ip6_null_entry),
					   GFP_KERNEL);
	if (!net->ipv6.ip6_null_entry)
2604
		goto out_ip6_dst_ops;
2605 2606
	net->ipv6.ip6_null_entry->u.dst.path =
		(struct dst_entry *)net->ipv6.ip6_null_entry;
2607
	net->ipv6.ip6_null_entry->u.dst.ops = net->ipv6.ip6_dst_ops;
2608 2609 2610 2611 2612 2613 2614 2615 2616 2617 2618

#ifdef CONFIG_IPV6_MULTIPLE_TABLES
	net->ipv6.ip6_prohibit_entry = kmemdup(&ip6_prohibit_entry_template,
					       sizeof(*net->ipv6.ip6_prohibit_entry),
					       GFP_KERNEL);
	if (!net->ipv6.ip6_prohibit_entry) {
		kfree(net->ipv6.ip6_null_entry);
		goto out;
	}
	net->ipv6.ip6_prohibit_entry->u.dst.path =
		(struct dst_entry *)net->ipv6.ip6_prohibit_entry;
2619
	net->ipv6.ip6_prohibit_entry->u.dst.ops = net->ipv6.ip6_dst_ops;
2620 2621 2622 2623 2624 2625 2626 2627 2628 2629 2630

	net->ipv6.ip6_blk_hole_entry = kmemdup(&ip6_blk_hole_entry_template,
					       sizeof(*net->ipv6.ip6_blk_hole_entry),
					       GFP_KERNEL);
	if (!net->ipv6.ip6_blk_hole_entry) {
		kfree(net->ipv6.ip6_null_entry);
		kfree(net->ipv6.ip6_prohibit_entry);
		goto out;
	}
	net->ipv6.ip6_blk_hole_entry->u.dst.path =
		(struct dst_entry *)net->ipv6.ip6_blk_hole_entry;
2631
	net->ipv6.ip6_blk_hole_entry->u.dst.ops = net->ipv6.ip6_dst_ops;
2632 2633
#endif

2634 2635 2636 2637
#ifdef CONFIG_PROC_FS
	proc_net_fops_create(net, "ipv6_route", 0, &ipv6_route_proc_fops);
	proc_net_fops_create(net, "rt6_stats", S_IRUGO, &rt6_stats_seq_fops);
#endif
2638 2639
	net->ipv6.ip6_rt_gc_expire = 30*HZ;

2640 2641 2642
	ret = 0;
out:
	return ret;
2643 2644

out_ip6_dst_ops:
2645
	release_net(net->ipv6.ip6_dst_ops->dst_net);
2646 2647
	kfree(net->ipv6.ip6_dst_ops);
	goto out;
2648 2649 2650 2651 2652 2653 2654 2655
}

static void ip6_route_net_exit(struct net *net)
{
#ifdef CONFIG_PROC_FS
	proc_net_remove(net, "ipv6_route");
	proc_net_remove(net, "rt6_stats");
#endif
2656 2657 2658 2659 2660
	kfree(net->ipv6.ip6_null_entry);
#ifdef CONFIG_IPV6_MULTIPLE_TABLES
	kfree(net->ipv6.ip6_prohibit_entry);
	kfree(net->ipv6.ip6_blk_hole_entry);
#endif
2661
	release_net(net->ipv6.ip6_dst_ops->dst_net);
2662
	kfree(net->ipv6.ip6_dst_ops);
2663 2664 2665 2666 2667 2668 2669
}

static struct pernet_operations ip6_route_net_ops = {
	.init = ip6_route_net_init,
	.exit = ip6_route_net_exit,
};

2670 2671 2672 2673 2674
static struct notifier_block ip6_route_dev_notifier = {
	.notifier_call = ip6_route_dev_notify,
	.priority = 0,
};

2675
int __init ip6_route_init(void)
L
Linus Torvalds 已提交
2676
{
2677 2678
	int ret;

2679 2680
	ret = -ENOMEM;
	ip6_dst_ops_template.kmem_cachep =
A
Alexey Dobriyan 已提交
2681
		kmem_cache_create("ip6_dst_cache", sizeof(struct rt6_info), 0,
2682
				  SLAB_HWCACHE_ALIGN, NULL);
2683
	if (!ip6_dst_ops_template.kmem_cachep)
2684
		goto out;;
2685

2686 2687
	ret = register_pernet_subsys(&ip6_route_net_ops);
	if (ret)
2688 2689
		goto out_kmem_cache;

2690 2691 2692 2693 2694 2695 2696 2697 2698 2699 2700
	/* Registering of the loopback is done before this portion of code,
	 * the loopback reference in rt6_info will not be taken, do it
	 * manually for init_net */
	init_net.ipv6.ip6_null_entry->u.dst.dev = init_net.loopback_dev;
	init_net.ipv6.ip6_null_entry->rt6i_idev = in6_dev_get(init_net.loopback_dev);
  #ifdef CONFIG_IPV6_MULTIPLE_TABLES
	init_net.ipv6.ip6_prohibit_entry->u.dst.dev = init_net.loopback_dev;
	init_net.ipv6.ip6_prohibit_entry->rt6i_idev = in6_dev_get(init_net.loopback_dev);
	init_net.ipv6.ip6_blk_hole_entry->u.dst.dev = init_net.loopback_dev;
	init_net.ipv6.ip6_blk_hole_entry->rt6i_idev = in6_dev_get(init_net.loopback_dev);
  #endif
2701 2702
	ret = fib6_init();
	if (ret)
2703
		goto out_register_subsys;
2704 2705 2706

	ret = xfrm6_init();
	if (ret)
2707
		goto out_fib6_init;
2708

2709 2710 2711
	ret = fib6_rules_init();
	if (ret)
		goto xfrm6_init;
2712

2713 2714 2715 2716 2717
	ret = -ENOBUFS;
	if (__rtnl_register(PF_INET6, RTM_NEWROUTE, inet6_rtm_newroute, NULL) ||
	    __rtnl_register(PF_INET6, RTM_DELROUTE, inet6_rtm_delroute, NULL) ||
	    __rtnl_register(PF_INET6, RTM_GETROUTE, inet6_rtm_getroute, NULL))
		goto fib6_rules_init;
2718

2719
	ret = register_netdevice_notifier(&ip6_route_dev_notifier);
2720 2721
	if (ret)
		goto fib6_rules_init;
2722

2723 2724 2725 2726 2727 2728 2729 2730 2731
out:
	return ret;

fib6_rules_init:
	fib6_rules_cleanup();
xfrm6_init:
	xfrm6_fini();
out_fib6_init:
	fib6_gc_cleanup();
2732 2733
out_register_subsys:
	unregister_pernet_subsys(&ip6_route_net_ops);
2734
out_kmem_cache:
2735
	kmem_cache_destroy(ip6_dst_ops_template.kmem_cachep);
2736
	goto out;
L
Linus Torvalds 已提交
2737 2738 2739 2740
}

void ip6_route_cleanup(void)
{
2741
	unregister_netdevice_notifier(&ip6_route_dev_notifier);
T
Thomas Graf 已提交
2742
	fib6_rules_cleanup();
L
Linus Torvalds 已提交
2743 2744
	xfrm6_fini();
	fib6_gc_cleanup();
2745
	unregister_pernet_subsys(&ip6_route_net_ops);
2746
	kmem_cache_destroy(ip6_dst_ops_template.kmem_cachep);
L
Linus Torvalds 已提交
2747
}