route.c 68.8 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 <linux/slab.h>
44
#include <net/net_namespace.h>
L
Linus Torvalds 已提交
45 46 47 48 49 50 51 52 53 54
#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>
55
#include <net/netevent.h>
56
#include <net/netlink.h>
L
Linus Torvalds 已提交
57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74

#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

75
#define CLONE_OFFLINK_ROUTE 0
L
Linus Torvalds 已提交
76 77 78 79 80 81 82

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);
83
static int		 ip6_dst_gc(struct dst_ops *ops);
L
Linus Torvalds 已提交
84 85 86 87 88 89

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);

90
#ifdef CONFIG_IPV6_ROUTE_INFO
91 92
static struct rt6_info *rt6_add_route_info(struct net *net,
					   struct in6_addr *prefix, int prefixlen,
93 94
					   struct in6_addr *gwaddr, int ifindex,
					   unsigned pref);
95 96
static struct rt6_info *rt6_get_route_info(struct net *net,
					   struct in6_addr *prefix, int prefixlen,
97 98 99
					   struct in6_addr *gwaddr, int ifindex);
#endif

100
static struct dst_ops ip6_dst_ops_template = {
L
Linus Torvalds 已提交
101
	.family			=	AF_INET6,
102
	.protocol		=	cpu_to_be16(ETH_P_IPV6),
L
Linus Torvalds 已提交
103 104 105 106 107 108 109 110
	.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,
111
	.local_out		=	__ip6_local_out,
L
Linus Torvalds 已提交
112 113
};

114 115 116 117 118 119
static void ip6_rt_blackhole_update_pmtu(struct dst_entry *dst, u32 mtu)
{
}

static struct dst_ops ip6_dst_blackhole_ops = {
	.family			=	AF_INET6,
120
	.protocol		=	cpu_to_be16(ETH_P_IPV6),
121 122 123 124 125
	.destroy		=	ip6_dst_destroy,
	.check			=	ip6_dst_check,
	.update_pmtu		=	ip6_rt_blackhole_update_pmtu,
};

126
static struct rt6_info ip6_null_entry_template = {
127 128 129 130 131 132 133
	.dst = {
		.__refcnt	= ATOMIC_INIT(1),
		.__use		= 1,
		.obsolete	= -1,
		.error		= -ENETUNREACH,
		.input		= ip6_pkt_discard,
		.output		= ip6_pkt_discard_out,
L
Linus Torvalds 已提交
134 135
	},
	.rt6i_flags	= (RTF_REJECT | RTF_NONEXTHOP),
136
	.rt6i_protocol  = RTPROT_KERNEL,
L
Linus Torvalds 已提交
137 138 139 140
	.rt6i_metric	= ~(u32) 0,
	.rt6i_ref	= ATOMIC_INIT(1),
};

T
Thomas Graf 已提交
141 142
#ifdef CONFIG_IPV6_MULTIPLE_TABLES

143 144 145
static int ip6_pkt_prohibit(struct sk_buff *skb);
static int ip6_pkt_prohibit_out(struct sk_buff *skb);

146
static struct rt6_info ip6_prohibit_entry_template = {
147 148 149 150 151 152 153
	.dst = {
		.__refcnt	= ATOMIC_INIT(1),
		.__use		= 1,
		.obsolete	= -1,
		.error		= -EACCES,
		.input		= ip6_pkt_prohibit,
		.output		= ip6_pkt_prohibit_out,
T
Thomas Graf 已提交
154 155
	},
	.rt6i_flags	= (RTF_REJECT | RTF_NONEXTHOP),
156
	.rt6i_protocol  = RTPROT_KERNEL,
T
Thomas Graf 已提交
157 158 159 160
	.rt6i_metric	= ~(u32) 0,
	.rt6i_ref	= ATOMIC_INIT(1),
};

161
static struct rt6_info ip6_blk_hole_entry_template = {
162 163 164 165 166 167 168
	.dst = {
		.__refcnt	= ATOMIC_INIT(1),
		.__use		= 1,
		.obsolete	= -1,
		.error		= -EINVAL,
		.input		= dst_discard,
		.output		= dst_discard,
T
Thomas Graf 已提交
169 170
	},
	.rt6i_flags	= (RTF_REJECT | RTF_NONEXTHOP),
171
	.rt6i_protocol  = RTPROT_KERNEL,
T
Thomas Graf 已提交
172 173 174 175 176 177
	.rt6i_metric	= ~(u32) 0,
	.rt6i_ref	= ATOMIC_INIT(1),
};

#endif

L
Linus Torvalds 已提交
178
/* allocate dst with ip6_dst_ops */
179
static inline struct rt6_info *ip6_dst_alloc(struct dst_ops *ops)
L
Linus Torvalds 已提交
180
{
181
	return (struct rt6_info *)dst_alloc(ops);
L
Linus Torvalds 已提交
182 183 184 185 186 187
}

static void ip6_dst_destroy(struct dst_entry *dst)
{
	struct rt6_info *rt = (struct rt6_info *)dst;
	struct inet6_dev *idev = rt->rt6i_idev;
188
	struct inet_peer *peer = rt->rt6i_peer;
L
Linus Torvalds 已提交
189 190 191 192

	if (idev != NULL) {
		rt->rt6i_idev = NULL;
		in6_dev_put(idev);
193
	}
194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210
	if (peer) {
		BUG_ON(!(rt->rt6i_flags & RTF_CACHE));
		rt->rt6i_peer = NULL;
		inet_putpeer(peer);
	}
}

void rt6_bind_peer(struct rt6_info *rt, int create)
{
	struct inet_peer *peer;

	if (WARN_ON(!(rt->rt6i_flags & RTF_CACHE)))
		return;

	peer = inet_getpeer_v6(&rt->rt6i_dst.addr, create);
	if (peer && cmpxchg(&rt->rt6i_peer, NULL, peer) != NULL)
		inet_putpeer(peer);
L
Linus Torvalds 已提交
211 212 213 214 215 216 217
}

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;
218
	struct net_device *loopback_dev =
219
		dev_net(dev)->loopback_dev;
L
Linus Torvalds 已提交
220

221 222 223
	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
224 225 226 227 228 229 230 231 232
		if (loopback_idev != NULL) {
			rt->rt6i_idev = loopback_idev;
			in6_dev_put(idev);
		}
	}
}

static __inline__ int rt6_check_expired(const struct rt6_info *rt)
{
E
Eric Dumazet 已提交
233 234
	return (rt->rt6i_flags & RTF_EXPIRES) &&
		time_after(jiffies, rt->rt6i_expires);
L
Linus Torvalds 已提交
235 236
}

T
Thomas Graf 已提交
237 238
static inline int rt6_need_strict(struct in6_addr *daddr)
{
E
Eric Dumazet 已提交
239 240
	return ipv6_addr_type(daddr) &
		(IPV6_ADDR_MULTICAST | IPV6_ADDR_LINKLOCAL | IPV6_ADDR_LOOPBACK);
T
Thomas Graf 已提交
241 242
}

L
Linus Torvalds 已提交
243
/*
T
Thomas Graf 已提交
244
 *	Route lookup. Any table->tb6_lock is implied.
L
Linus Torvalds 已提交
245 246
 */

247 248
static inline struct rt6_info *rt6_device_match(struct net *net,
						    struct rt6_info *rt,
249
						    struct in6_addr *saddr,
L
Linus Torvalds 已提交
250
						    int oif,
251
						    int flags)
L
Linus Torvalds 已提交
252 253 254 255
{
	struct rt6_info *local = NULL;
	struct rt6_info *sprt;

256 257 258
	if (!oif && ipv6_addr_any(saddr))
		goto out;

259
	for (sprt = rt; sprt; sprt = sprt->dst.rt6_next) {
260 261 262
		struct net_device *dev = sprt->rt6i_dev;

		if (oif) {
L
Linus Torvalds 已提交
263 264 265 266 267
			if (dev->ifindex == oif)
				return sprt;
			if (dev->flags & IFF_LOOPBACK) {
				if (sprt->rt6i_idev == NULL ||
				    sprt->rt6i_idev->dev->ifindex != oif) {
268
					if (flags & RT6_LOOKUP_F_IFACE && oif)
L
Linus Torvalds 已提交
269
						continue;
270
					if (local && (!oif ||
L
Linus Torvalds 已提交
271 272 273 274 275
						      local->rt6i_idev->dev->ifindex == oif))
						continue;
				}
				local = sprt;
			}
276 277 278 279
		} else {
			if (ipv6_chk_addr(net, saddr, dev,
					  flags & RT6_LOOKUP_F_IFACE))
				return sprt;
L
Linus Torvalds 已提交
280
		}
281
	}
L
Linus Torvalds 已提交
282

283
	if (oif) {
L
Linus Torvalds 已提交
284 285 286
		if (local)
			return local;

287
		if (flags & RT6_LOOKUP_F_IFACE)
288
			return net->ipv6.ip6_null_entry;
L
Linus Torvalds 已提交
289
	}
290
out:
L
Linus Torvalds 已提交
291 292 293
	return rt;
}

294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309
#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) &&
310
	    time_after(jiffies, neigh->updated + rt->rt6i_idev->cnf.rtr_probe_interval)) {
311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328
		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)
{
}
#endif

L
Linus Torvalds 已提交
329
/*
330
 * Default Router Selection (RFC 2461 6.3.6)
L
Linus Torvalds 已提交
331
 */
D
Dave Jones 已提交
332
static inline int rt6_check_dev(struct rt6_info *rt, int oif)
333 334
{
	struct net_device *dev = rt->rt6i_dev;
335
	if (!oif || dev->ifindex == oif)
336
		return 2;
337 338 339 340
	if ((dev->flags & IFF_LOOPBACK) &&
	    rt->rt6i_idev && rt->rt6i_idev->dev->ifindex == oif)
		return 1;
	return 0;
341
}
L
Linus Torvalds 已提交
342

D
Dave Jones 已提交
343
static inline int rt6_check_neigh(struct rt6_info *rt)
L
Linus Torvalds 已提交
344
{
345
	struct neighbour *neigh = rt->rt6i_nexthop;
346
	int m;
347 348 349 350
	if (rt->rt6i_flags & RTF_NONEXTHOP ||
	    !(rt->rt6i_flags & RTF_GATEWAY))
		m = 1;
	else if (neigh) {
351 352
		read_lock_bh(&neigh->lock);
		if (neigh->nud_state & NUD_VALID)
353
			m = 2;
354 355 356 357 358
#ifdef CONFIG_IPV6_ROUTER_PREF
		else if (neigh->nud_state & NUD_FAILED)
			m = 0;
#endif
		else
359
			m = 1;
360
		read_unlock_bh(&neigh->lock);
361 362
	} else
		m = 0;
363
	return m;
L
Linus Torvalds 已提交
364 365
}

366 367
static int rt6_score_route(struct rt6_info *rt, int oif,
			   int strict)
L
Linus Torvalds 已提交
368
{
369
	int m, n;
370

371
	m = rt6_check_dev(rt, oif);
372
	if (!m && (strict & RT6_LOOKUP_F_IFACE))
373
		return -1;
374 375 376
#ifdef CONFIG_IPV6_ROUTER_PREF
	m |= IPV6_DECODE_PREF(IPV6_EXTRACT_PREF(rt->rt6i_flags)) << 2;
#endif
377
	n = rt6_check_neigh(rt);
378
	if (!n && (strict & RT6_LOOKUP_F_REACHABLE))
379 380 381 382
		return -1;
	return m;
}

383 384
static struct rt6_info *find_match(struct rt6_info *rt, int oif, int strict,
				   int *mpri, struct rt6_info *match)
385
{
386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412
	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;
413
	int mpri = -1;
L
Linus Torvalds 已提交
414

415 416
	match = NULL;
	for (rt = rr_head; rt && rt->rt6i_metric == metric;
417
	     rt = rt->dst.rt6_next)
418 419
		match = find_match(rt, oif, strict, &mpri, match);
	for (rt = fn->leaf; rt && rt != rr_head && rt->rt6i_metric == metric;
420
	     rt = rt->dst.rt6_next)
421
		match = find_match(rt, oif, strict, &mpri, match);
L
Linus Torvalds 已提交
422

423 424
	return match;
}
L
Linus Torvalds 已提交
425

426 427 428
static struct rt6_info *rt6_select(struct fib6_node *fn, int oif, int strict)
{
	struct rt6_info *match, *rt0;
429
	struct net *net;
L
Linus Torvalds 已提交
430

431
	RT6_TRACE("%s(fn->leaf=%p, oif=%d)\n",
432
		  __func__, fn->leaf, oif);
433

434 435 436
	rt0 = fn->rr_ptr;
	if (!rt0)
		fn->rr_ptr = rt0 = fn->leaf;
L
Linus Torvalds 已提交
437

438
	match = find_rr_leaf(fn, rt0, rt0->rt6i_metric, oif, strict);
L
Linus Torvalds 已提交
439

440
	if (!match &&
441
	    (strict & RT6_LOOKUP_F_REACHABLE)) {
442
		struct rt6_info *next = rt0->dst.rt6_next;
443

444
		/* no entries matched; do round-robin */
445 446 447 448 449
		if (!next || next->rt6i_metric != rt0->rt6i_metric)
			next = fn->leaf;

		if (next != rt0)
			fn->rr_ptr = next;
L
Linus Torvalds 已提交
450 451
	}

452
	RT6_TRACE("%s() => %p\n",
453
		  __func__, match);
L
Linus Torvalds 已提交
454

455
	net = dev_net(rt0->rt6i_dev);
E
Eric Dumazet 已提交
456
	return match ? match : net->ipv6.ip6_null_entry;
L
Linus Torvalds 已提交
457 458
}

459 460 461 462
#ifdef CONFIG_IPV6_ROUTE_INFO
int rt6_route_rcv(struct net_device *dev, u8 *opt, int len,
		  struct in6_addr *gwaddr)
{
463
	struct net *net = dev_net(dev);
464 465 466
	struct route_info *rinfo = (struct route_info *) opt;
	struct in6_addr prefix_buf, *prefix;
	unsigned int pref;
467
	unsigned long lifetime;
468 469 470 471 472 473 474 475 476 477 478 479 480 481 482 483 484 485 486 487 488 489 490
	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)
491
		return -EINVAL;
492

493
	lifetime = addrconf_timeout_fixup(ntohl(rinfo->lifetime), HZ);
494 495 496 497 498 499 500 501 502 503 504

	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;
	}

505 506
	rt = rt6_get_route_info(net, prefix, rinfo->prefix_len, gwaddr,
				dev->ifindex);
507 508

	if (rt && !lifetime) {
509
		ip6_del_rt(rt);
510 511 512 513
		rt = NULL;
	}

	if (!rt && lifetime)
514
		rt = rt6_add_route_info(net, prefix, rinfo->prefix_len, gwaddr, dev->ifindex,
515 516 517 518 519 520
					pref);
	else if (rt)
		rt->rt6i_flags = RTF_ROUTEINFO |
				 (rt->rt6i_flags & ~RTF_PREF_MASK) | RTF_PREF(pref);

	if (rt) {
521
		if (!addrconf_finite_timeout(lifetime)) {
522 523 524 525 526
			rt->rt6i_flags &= ~RTF_EXPIRES;
		} else {
			rt->rt6i_expires = jiffies + HZ * lifetime;
			rt->rt6i_flags |= RTF_EXPIRES;
		}
527
		dst_release(&rt->dst);
528 529 530 531 532
	}
	return 0;
}
#endif

533
#define BACKTRACK(__net, saddr)			\
534
do { \
535
	if (rt == __net->ipv6.ip6_null_entry) {	\
536
		struct fib6_node *pn; \
V
Ville Nuorvala 已提交
537
		while (1) { \
538 539 540 541
			if (fn->fn_flags & RTN_TL_ROOT) \
				goto out; \
			pn = fn->parent; \
			if (FIB6_SUBTREE(pn) && FIB6_SUBTREE(pn) != fn) \
542
				fn = fib6_lookup(FIB6_SUBTREE(pn), NULL, saddr); \
543 544 545 546
			else \
				fn = pn; \
			if (fn->fn_flags & RTN_RTINFO) \
				goto restart; \
T
Thomas Graf 已提交
547 548
		} \
	} \
549
} while(0)
T
Thomas Graf 已提交
550

551 552
static struct rt6_info *ip6_pol_route_lookup(struct net *net,
					     struct fib6_table *table,
T
Thomas Graf 已提交
553
					     struct flowi *fl, int flags)
L
Linus Torvalds 已提交
554 555 556 557
{
	struct fib6_node *fn;
	struct rt6_info *rt;

T
Thomas Graf 已提交
558 559 560 561
	read_lock_bh(&table->tb6_lock);
	fn = fib6_lookup(&table->tb6_root, &fl->fl6_dst, &fl->fl6_src);
restart:
	rt = fn->leaf;
562
	rt = rt6_device_match(net, rt, &fl->fl6_src, fl->oif, flags);
563
	BACKTRACK(net, &fl->fl6_src);
T
Thomas Graf 已提交
564
out:
565
	dst_use(&rt->dst, jiffies);
T
Thomas Graf 已提交
566 567 568 569 570
	read_unlock_bh(&table->tb6_lock);
	return rt;

}

571 572
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
573 574 575
{
	struct flowi fl = {
		.oif = oif,
576
		.fl6_dst = *daddr,
T
Thomas Graf 已提交
577 578
	};
	struct dst_entry *dst;
579
	int flags = strict ? RT6_LOOKUP_F_IFACE : 0;
T
Thomas Graf 已提交
580

581 582 583 584 585
	if (saddr) {
		memcpy(&fl.fl6_src, saddr, sizeof(*saddr));
		flags |= RT6_LOOKUP_F_HAS_SADDR;
	}

586
	dst = fib6_rule_lookup(net, &fl, flags, ip6_pol_route_lookup);
T
Thomas Graf 已提交
587 588 589 590 591
	if (dst->error == 0)
		return (struct rt6_info *) dst;

	dst_release(dst);

L
Linus Torvalds 已提交
592 593 594
	return NULL;
}

595 596
EXPORT_SYMBOL(rt6_lookup);

T
Thomas Graf 已提交
597
/* ip6_ins_rt is called with FREE table->tb6_lock.
L
Linus Torvalds 已提交
598 599 600 601 602
   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.
 */

603
static int __ip6_ins_rt(struct rt6_info *rt, struct nl_info *info)
L
Linus Torvalds 已提交
604 605
{
	int err;
T
Thomas Graf 已提交
606
	struct fib6_table *table;
L
Linus Torvalds 已提交
607

T
Thomas Graf 已提交
608 609
	table = rt->rt6i_table;
	write_lock_bh(&table->tb6_lock);
610
	err = fib6_add(&table->tb6_root, rt, info);
T
Thomas Graf 已提交
611
	write_unlock_bh(&table->tb6_lock);
L
Linus Torvalds 已提交
612 613 614 615

	return err;
}

616 617
int ip6_ins_rt(struct rt6_info *rt)
{
618
	struct nl_info info = {
619
		.nl_net = dev_net(rt->rt6i_dev),
620
	};
621
	return __ip6_ins_rt(rt, &info);
622 623
}

624 625
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
626 627 628 629 630 631 632 633 634 635
{
	struct rt6_info *rt;

	/*
	 *	Clone the route.
	 */

	rt = ip6_rt_copy(ort);

	if (rt) {
636 637 638
		struct neighbour *neigh;
		int attempts = !in_softirq();

639 640 641 642
		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
643
			ipv6_addr_copy(&rt->rt6i_gateway, daddr);
644
		}
L
Linus Torvalds 已提交
645

646
		ipv6_addr_copy(&rt->rt6i_dst.addr, daddr);
L
Linus Torvalds 已提交
647 648
		rt->rt6i_dst.plen = 128;
		rt->rt6i_flags |= RTF_CACHE;
649
		rt->dst.flags |= DST_HOST;
L
Linus Torvalds 已提交
650 651 652 653 654 655 656 657

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

658 659 660 661 662 663 664 665 666 667 668 669 670
	retry:
		neigh = ndisc_get_neigh(rt->rt6i_dev, &rt->rt6i_gateway);
		if (IS_ERR(neigh)) {
			struct net *net = dev_net(rt->rt6i_dev);
			int saved_rt_min_interval =
				net->ipv6.sysctl.ip6_rt_gc_min_interval;
			int saved_rt_elasticity =
				net->ipv6.sysctl.ip6_rt_gc_elasticity;

			if (attempts-- > 0) {
				net->ipv6.sysctl.ip6_rt_gc_elasticity = 1;
				net->ipv6.sysctl.ip6_rt_gc_min_interval = 0;

671
				ip6_dst_gc(&net->ipv6.ip6_dst_ops);
672 673 674 675 676 677 678 679 680 681

				net->ipv6.sysctl.ip6_rt_gc_elasticity =
					saved_rt_elasticity;
				net->ipv6.sysctl.ip6_rt_gc_min_interval =
					saved_rt_min_interval;
				goto retry;
			}

			if (net_ratelimit())
				printk(KERN_WARNING
682
				       "ipv6: Neighbour table overflow.\n");
683
			dst_free(&rt->dst);
684 685 686
			return NULL;
		}
		rt->rt6i_nexthop = neigh;
L
Linus Torvalds 已提交
687

688
	}
L
Linus Torvalds 已提交
689

690 691
	return rt;
}
L
Linus Torvalds 已提交
692

693 694 695 696 697 698 699
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;
700
		rt->dst.flags |= DST_HOST;
701 702 703 704 705
		rt->rt6i_nexthop = neigh_clone(ort->rt6i_nexthop);
	}
	return rt;
}

706 707
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
708 709
{
	struct fib6_node *fn;
710
	struct rt6_info *rt, *nrt;
T
Thomas Graf 已提交
711
	int strict = 0;
L
Linus Torvalds 已提交
712
	int attempts = 3;
713
	int err;
714
	int reachable = net->ipv6.devconf_all->forwarding ? 0 : RT6_LOOKUP_F_REACHABLE;
L
Linus Torvalds 已提交
715

716
	strict |= flags & RT6_LOOKUP_F_IFACE;
L
Linus Torvalds 已提交
717 718

relookup:
T
Thomas Graf 已提交
719
	read_lock_bh(&table->tb6_lock);
L
Linus Torvalds 已提交
720

721
restart_2:
T
Thomas Graf 已提交
722
	fn = fib6_lookup(&table->tb6_root, &fl->fl6_dst, &fl->fl6_src);
L
Linus Torvalds 已提交
723 724

restart:
725
	rt = rt6_select(fn, oif, strict | reachable);
726 727 728

	BACKTRACK(net, &fl->fl6_src);
	if (rt == net->ipv6.ip6_null_entry ||
729
	    rt->rt6i_flags & RTF_CACHE)
730
		goto out;
L
Linus Torvalds 已提交
731

732
	dst_hold(&rt->dst);
T
Thomas Graf 已提交
733
	read_unlock_bh(&table->tb6_lock);
734

735
	if (!rt->rt6i_nexthop && !(rt->rt6i_flags & RTF_NONEXTHOP))
T
Thomas Graf 已提交
736
		nrt = rt6_alloc_cow(rt, &fl->fl6_dst, &fl->fl6_src);
737 738
	else {
#if CLONE_OFFLINK_ROUTE
T
Thomas Graf 已提交
739
		nrt = rt6_alloc_clone(rt, &fl->fl6_dst);
740 741 742 743
#else
		goto out2;
#endif
	}
744

745
	dst_release(&rt->dst);
746
	rt = nrt ? : net->ipv6.ip6_null_entry;
L
Linus Torvalds 已提交
747

748
	dst_hold(&rt->dst);
749
	if (nrt) {
750
		err = ip6_ins_rt(nrt);
751
		if (!err)
L
Linus Torvalds 已提交
752 753 754
			goto out2;
	}

755 756 757 758
	if (--attempts <= 0)
		goto out2;

	/*
T
Thomas Graf 已提交
759
	 * Race condition! In the gap, when table->tb6_lock was
760 761
	 * released someone could insert this route.  Relookup.
	 */
762
	dst_release(&rt->dst);
763 764 765
	goto relookup;

out:
766 767 768 769
	if (reachable) {
		reachable = 0;
		goto restart_2;
	}
770
	dst_hold(&rt->dst);
T
Thomas Graf 已提交
771
	read_unlock_bh(&table->tb6_lock);
L
Linus Torvalds 已提交
772
out2:
773 774
	rt->dst.lastuse = jiffies;
	rt->dst.__use++;
T
Thomas Graf 已提交
775 776

	return rt;
L
Linus Torvalds 已提交
777 778
}

779
static struct rt6_info *ip6_pol_route_input(struct net *net, struct fib6_table *table,
780 781
					    struct flowi *fl, int flags)
{
782
	return ip6_pol_route(net, table, fl->iif, fl, flags);
783 784
}

T
Thomas Graf 已提交
785 786
void ip6_route_input(struct sk_buff *skb)
{
787
	struct ipv6hdr *iph = ipv6_hdr(skb);
788
	struct net *net = dev_net(skb->dev);
789
	int flags = RT6_LOOKUP_F_HAS_SADDR;
T
Thomas Graf 已提交
790 791
	struct flowi fl = {
		.iif = skb->dev->ifindex,
792 793 794
		.fl6_dst = iph->daddr,
		.fl6_src = iph->saddr,
		.fl6_flowlabel = (* (__be32 *) iph)&IPV6_FLOWINFO_MASK,
795
		.mark = skb->mark,
T
Thomas Graf 已提交
796 797
		.proto = iph->nexthdr,
	};
798

T
Thomas Goff 已提交
799
	if (rt6_need_strict(&iph->daddr) && skb->dev->type != ARPHRD_PIMREG)
800
		flags |= RT6_LOOKUP_F_IFACE;
T
Thomas Graf 已提交
801

E
Eric Dumazet 已提交
802
	skb_dst_set(skb, fib6_rule_lookup(net, &fl, flags, ip6_pol_route_input));
T
Thomas Graf 已提交
803 804
}

805
static struct rt6_info *ip6_pol_route_output(struct net *net, struct fib6_table *table,
T
Thomas Graf 已提交
806
					     struct flowi *fl, int flags)
L
Linus Torvalds 已提交
807
{
808
	return ip6_pol_route(net, table, fl->oif, fl, flags);
T
Thomas Graf 已提交
809 810
}

811 812
struct dst_entry * ip6_route_output(struct net *net, struct sock *sk,
				    struct flowi *fl)
T
Thomas Graf 已提交
813 814 815
{
	int flags = 0;

B
Brian Haley 已提交
816
	if ((sk && sk->sk_bound_dev_if) || rt6_need_strict(&fl->fl6_dst))
817
		flags |= RT6_LOOKUP_F_IFACE;
T
Thomas Graf 已提交
818

819 820
	if (!ipv6_addr_any(&fl->fl6_src))
		flags |= RT6_LOOKUP_F_HAS_SADDR;
821 822
	else if (sk)
		flags |= rt6_srcprefs2flags(inet6_sk(sk)->srcprefs);
823

824
	return fib6_rule_lookup(net, fl, flags, ip6_pol_route_output);
L
Linus Torvalds 已提交
825 826
}

827
EXPORT_SYMBOL(ip6_route_output);
L
Linus Torvalds 已提交
828

829 830 831 832 833 834 835 836
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) {
837
		new = &rt->dst;
838 839 840

		atomic_set(&new->__refcnt, 1);
		new->__use = 1;
841 842
		new->input = dst_discard;
		new->output = dst_discard;
843

844
		dst_copy_metrics(new, &ort->dst);
845
		new->dev = ort->dst.dev;
846 847 848 849 850 851 852 853 854 855 856 857 858 859 860 861 862 863 864 865 866
		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;
E
Eric Dumazet 已提交
867
	return new ? 0 : -ENOMEM;
868 869 870
}
EXPORT_SYMBOL_GPL(ip6_dst_blackhole);

L
Linus Torvalds 已提交
871 872 873 874 875 876 877 878 879 880
/*
 *	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;

881
	if (rt->rt6i_node && (rt->rt6i_node->fn_sernum == cookie))
L
Linus Torvalds 已提交
882 883 884 885 886 887 888 889 890 891
		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) {
892 893 894 895 896 897
		if (rt->rt6i_flags & RTF_CACHE) {
			if (rt6_check_expired(rt)) {
				ip6_del_rt(rt);
				dst = NULL;
			}
		} else {
L
Linus Torvalds 已提交
898
			dst_release(dst);
899 900
			dst = NULL;
		}
L
Linus Torvalds 已提交
901
	}
902
	return dst;
L
Linus Torvalds 已提交
903 904 905 906 907 908
}

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

909
	icmpv6_send(skb, ICMPV6_DEST_UNREACH, ICMPV6_ADDR_UNREACH, 0);
L
Linus Torvalds 已提交
910

E
Eric Dumazet 已提交
911
	rt = (struct rt6_info *) skb_dst(skb);
L
Linus Torvalds 已提交
912 913
	if (rt) {
		if (rt->rt6i_flags&RTF_CACHE) {
914
			dst_set_expires(&rt->dst, 0);
L
Linus Torvalds 已提交
915 916 917 918 919 920 921 922 923 924 925 926 927
			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) {
928
			u32 features = dst_metric(dst, RTAX_FEATURES);
L
Linus Torvalds 已提交
929
			mtu = IPV6_MIN_MTU;
930 931
			features |= RTAX_FEATURE_ALLFRAG;
			dst_metric_set(dst, RTAX_FEATURES, features);
L
Linus Torvalds 已提交
932
		}
933
		dst_metric_set(dst, RTAX_MTU, mtu);
934
		call_netevent_notifiers(NETEVENT_PMTU_UPDATE, dst);
L
Linus Torvalds 已提交
935 936 937 938 939
	}
}

static int ipv6_get_mtu(struct net_device *dev);

940
static inline unsigned int ipv6_advmss(struct net *net, unsigned int mtu)
L
Linus Torvalds 已提交
941 942 943
{
	mtu -= sizeof(struct ipv6hdr) + sizeof(struct tcphdr);

944 945
	if (mtu < net->ipv6.sysctl.ip6_rt_min_advmss)
		mtu = net->ipv6.sysctl.ip6_rt_min_advmss;
L
Linus Torvalds 已提交
946 947

	/*
948 949 950
	 * 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
951 952 953 954 955 956 957
	 * rely only on pmtu discovery"
	 */
	if (mtu > IPV6_MAXPLEN - sizeof(struct tcphdr))
		mtu = IPV6_MAXPLEN;
	return mtu;
}

958 959
static struct dst_entry *icmp6_dst_gc_list;
static DEFINE_SPINLOCK(icmp6_dst_lock);
960

961
struct dst_entry *icmp6_dst_alloc(struct net_device *dev,
L
Linus Torvalds 已提交
962
				  struct neighbour *neigh,
963
				  const struct in6_addr *addr)
L
Linus Torvalds 已提交
964 965 966
{
	struct rt6_info *rt;
	struct inet6_dev *idev = in6_dev_get(dev);
967
	struct net *net = dev_net(dev);
L
Linus Torvalds 已提交
968 969 970 971

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

972
	rt = ip6_dst_alloc(&net->ipv6.ip6_dst_ops);
L
Linus Torvalds 已提交
973 974 975 976 977 978 979 980
	if (unlikely(rt == NULL)) {
		in6_dev_put(idev);
		goto out;
	}

	dev_hold(dev);
	if (neigh)
		neigh_hold(neigh);
981
	else {
L
Linus Torvalds 已提交
982
		neigh = ndisc_get_neigh(dev, addr);
983 984 985
		if (IS_ERR(neigh))
			neigh = NULL;
	}
L
Linus Torvalds 已提交
986 987 988 989

	rt->rt6i_dev	  = dev;
	rt->rt6i_idev     = idev;
	rt->rt6i_nexthop  = neigh;
990
	atomic_set(&rt->dst.__refcnt, 1);
991 992 993
	dst_metric_set(&rt->dst, RTAX_HOPLIMIT, 255);
	dst_metric_set(&rt->dst, RTAX_MTU, ipv6_get_mtu(rt->rt6i_dev));
	dst_metric_set(&rt->dst, RTAX_ADVMSS, ipv6_advmss(net, dst_mtu(&rt->dst)));
994
	rt->dst.output  = ip6_output;
L
Linus Torvalds 已提交
995 996

#if 0	/* there's no chance to use these for ndisc */
997
	rt->dst.flags   = ipv6_addr_type(addr) & IPV6_ADDR_UNICAST
998
				? DST_HOST
L
Linus Torvalds 已提交
999 1000 1001 1002 1003
				: 0;
	ipv6_addr_copy(&rt->rt6i_dst.addr, addr);
	rt->rt6i_dst.plen = 128;
#endif

1004
	spin_lock_bh(&icmp6_dst_lock);
1005 1006
	rt->dst.next = icmp6_dst_gc_list;
	icmp6_dst_gc_list = &rt->dst;
1007
	spin_unlock_bh(&icmp6_dst_lock);
L
Linus Torvalds 已提交
1008

1009
	fib6_force_start_gc(net);
L
Linus Torvalds 已提交
1010 1011

out:
1012
	return &rt->dst;
L
Linus Torvalds 已提交
1013 1014
}

1015
int icmp6_dst_gc(void)
L
Linus Torvalds 已提交
1016 1017
{
	struct dst_entry *dst, *next, **pprev;
1018
	int more = 0;
L
Linus Torvalds 已提交
1019 1020

	next = NULL;
1021

1022 1023
	spin_lock_bh(&icmp6_dst_lock);
	pprev = &icmp6_dst_gc_list;
1024

L
Linus Torvalds 已提交
1025 1026 1027 1028 1029 1030
	while ((dst = *pprev) != NULL) {
		if (!atomic_read(&dst->__refcnt)) {
			*pprev = dst->next;
			dst_free(dst);
		} else {
			pprev = &dst->next;
1031
			++more;
L
Linus Torvalds 已提交
1032 1033 1034
		}
	}

1035
	spin_unlock_bh(&icmp6_dst_lock);
1036

1037
	return more;
L
Linus Torvalds 已提交
1038 1039
}

1040 1041 1042 1043 1044 1045 1046 1047 1048 1049 1050 1051 1052 1053 1054 1055 1056 1057 1058
static void icmp6_clean_all(int (*func)(struct rt6_info *rt, void *arg),
			    void *arg)
{
	struct dst_entry *dst, **pprev;

	spin_lock_bh(&icmp6_dst_lock);
	pprev = &icmp6_dst_gc_list;
	while ((dst = *pprev) != NULL) {
		struct rt6_info *rt = (struct rt6_info *) dst;
		if (func(rt, arg)) {
			*pprev = dst->next;
			dst_free(dst);
		} else {
			pprev = &dst->next;
		}
	}
	spin_unlock_bh(&icmp6_dst_lock);
}

1059
static int ip6_dst_gc(struct dst_ops *ops)
L
Linus Torvalds 已提交
1060 1061
{
	unsigned long now = jiffies;
1062
	struct net *net = container_of(ops, struct net, ipv6.ip6_dst_ops);
1063 1064 1065 1066 1067
	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;
1068
	int entries;
1069

1070
	entries = dst_entries_get_fast(ops);
1071
	if (time_after(rt_last_gc + rt_min_interval, now) &&
1072
	    entries <= rt_max_size)
L
Linus Torvalds 已提交
1073 1074
		goto out;

1075 1076 1077
	net->ipv6.ip6_rt_gc_expire++;
	fib6_run_gc(net->ipv6.ip6_rt_gc_expire, net);
	net->ipv6.ip6_rt_last_gc = now;
1078 1079
	entries = dst_entries_get_slow(ops);
	if (entries < ops->gc_thresh)
1080
		net->ipv6.ip6_rt_gc_expire = rt_gc_timeout>>1;
L
Linus Torvalds 已提交
1081
out:
1082
	net->ipv6.ip6_rt_gc_expire -= net->ipv6.ip6_rt_gc_expire>>rt_elasticity;
1083
	return entries > rt_max_size;
L
Linus Torvalds 已提交
1084 1085 1086 1087 1088 1089 1090 1091 1092 1093 1094 1095 1096
}

/* 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;

1097 1098 1099
	rcu_read_lock();
	idev = __in6_dev_get(dev);
	if (idev)
L
Linus Torvalds 已提交
1100
		mtu = idev->cnf.mtu6;
1101
	rcu_read_unlock();
L
Linus Torvalds 已提交
1102 1103 1104
	return mtu;
}

1105
int ip6_dst_hoplimit(struct dst_entry *dst)
L
Linus Torvalds 已提交
1106
{
1107
	int hoplimit = dst_metric_raw(dst, RTAX_HOPLIMIT);
1108
	if (hoplimit == 0) {
1109
		struct net_device *dev = dst->dev;
1110 1111 1112 1113 1114
		struct inet6_dev *idev;

		rcu_read_lock();
		idev = __in6_dev_get(dev);
		if (idev)
1115
			hoplimit = idev->cnf.hop_limit;
1116
		else
1117
			hoplimit = dev_net(dev)->ipv6.devconf_all->hop_limit;
1118
		rcu_read_unlock();
L
Linus Torvalds 已提交
1119 1120 1121
	}
	return hoplimit;
}
1122
EXPORT_SYMBOL(ip6_dst_hoplimit);
L
Linus Torvalds 已提交
1123 1124 1125 1126 1127

/*
 *
 */

1128
int ip6_route_add(struct fib6_config *cfg)
L
Linus Torvalds 已提交
1129 1130
{
	int err;
1131
	struct net *net = cfg->fc_nlinfo.nl_net;
L
Linus Torvalds 已提交
1132 1133 1134
	struct rt6_info *rt = NULL;
	struct net_device *dev = NULL;
	struct inet6_dev *idev = NULL;
T
Thomas Graf 已提交
1135
	struct fib6_table *table;
L
Linus Torvalds 已提交
1136 1137
	int addr_type;

1138
	if (cfg->fc_dst_len > 128 || cfg->fc_src_len > 128)
L
Linus Torvalds 已提交
1139 1140
		return -EINVAL;
#ifndef CONFIG_IPV6_SUBTREES
1141
	if (cfg->fc_src_len)
L
Linus Torvalds 已提交
1142 1143
		return -EINVAL;
#endif
1144
	if (cfg->fc_ifindex) {
L
Linus Torvalds 已提交
1145
		err = -ENODEV;
1146
		dev = dev_get_by_index(net, cfg->fc_ifindex);
L
Linus Torvalds 已提交
1147 1148 1149 1150 1151 1152 1153
		if (!dev)
			goto out;
		idev = in6_dev_get(dev);
		if (!idev)
			goto out;
	}

1154 1155
	if (cfg->fc_metric == 0)
		cfg->fc_metric = IP6_RT_PRIO_USER;
L
Linus Torvalds 已提交
1156

1157
	table = fib6_new_table(net, cfg->fc_table);
T
Thomas Graf 已提交
1158 1159 1160 1161 1162
	if (table == NULL) {
		err = -ENOBUFS;
		goto out;
	}

1163
	rt = ip6_dst_alloc(&net->ipv6.ip6_dst_ops);
L
Linus Torvalds 已提交
1164 1165 1166 1167 1168 1169

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

1170
	rt->dst.obsolete = -1;
1171 1172 1173
	rt->rt6i_expires = (cfg->fc_flags & RTF_EXPIRES) ?
				jiffies + clock_t_to_jiffies(cfg->fc_expires) :
				0;
L
Linus Torvalds 已提交
1174

1175 1176 1177 1178 1179
	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
1180 1181

	if (addr_type & IPV6_ADDR_MULTICAST)
1182
		rt->dst.input = ip6_mc_input;
1183 1184
	else if (cfg->fc_flags & RTF_LOCAL)
		rt->dst.input = ip6_input;
L
Linus Torvalds 已提交
1185
	else
1186
		rt->dst.input = ip6_forward;
L
Linus Torvalds 已提交
1187

1188
	rt->dst.output = ip6_output;
L
Linus Torvalds 已提交
1189

1190 1191
	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
1192
	if (rt->rt6i_dst.plen == 128)
1193
	       rt->dst.flags = DST_HOST;
L
Linus Torvalds 已提交
1194 1195

#ifdef CONFIG_IPV6_SUBTREES
1196 1197
	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
1198 1199
#endif

1200
	rt->rt6i_metric = cfg->fc_metric;
L
Linus Torvalds 已提交
1201 1202 1203 1204

	/* We cannot add true routes via loopback here,
	   they would result in kernel looping; promote them to reject routes
	 */
1205
	if ((cfg->fc_flags & RTF_REJECT) ||
1206 1207
	    (dev && (dev->flags&IFF_LOOPBACK) && !(addr_type&IPV6_ADDR_LOOPBACK)
					      && !(cfg->fc_flags&RTF_LOCAL))) {
L
Linus Torvalds 已提交
1208
		/* hold loopback dev/idev if we haven't done so. */
1209
		if (dev != net->loopback_dev) {
L
Linus Torvalds 已提交
1210 1211 1212 1213
			if (dev) {
				dev_put(dev);
				in6_dev_put(idev);
			}
1214
			dev = net->loopback_dev;
L
Linus Torvalds 已提交
1215 1216 1217 1218 1219 1220 1221
			dev_hold(dev);
			idev = in6_dev_get(dev);
			if (!idev) {
				err = -ENODEV;
				goto out;
			}
		}
1222 1223 1224
		rt->dst.output = ip6_pkt_discard_out;
		rt->dst.input = ip6_pkt_discard;
		rt->dst.error = -ENETUNREACH;
L
Linus Torvalds 已提交
1225 1226 1227 1228
		rt->rt6i_flags = RTF_REJECT|RTF_NONEXTHOP;
		goto install_route;
	}

1229
	if (cfg->fc_flags & RTF_GATEWAY) {
L
Linus Torvalds 已提交
1230 1231 1232
		struct in6_addr *gw_addr;
		int gwa_type;

1233 1234
		gw_addr = &cfg->fc_gateway;
		ipv6_addr_copy(&rt->rt6i_gateway, gw_addr);
L
Linus Torvalds 已提交
1235 1236 1237 1238 1239 1240 1241 1242 1243 1244 1245 1246 1247 1248 1249 1250
		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;

1251
			grt = rt6_lookup(net, gw_addr, NULL, cfg->fc_ifindex, 1);
L
Linus Torvalds 已提交
1252 1253 1254 1255 1256 1257

			err = -EHOSTUNREACH;
			if (grt == NULL)
				goto out;
			if (dev) {
				if (dev != grt->rt6i_dev) {
1258
					dst_release(&grt->dst);
L
Linus Torvalds 已提交
1259 1260 1261 1262 1263 1264 1265 1266 1267 1268
					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;
1269
			dst_release(&grt->dst);
L
Linus Torvalds 已提交
1270 1271 1272 1273 1274 1275 1276 1277 1278 1279 1280 1281 1282

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

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

1283
	if (cfg->fc_flags & (RTF_GATEWAY | RTF_NONEXTHOP)) {
L
Linus Torvalds 已提交
1284 1285 1286 1287 1288 1289 1290 1291
		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;
		}
	}

1292
	rt->rt6i_flags = cfg->fc_flags;
L
Linus Torvalds 已提交
1293 1294

install_route:
1295 1296 1297 1298 1299
	if (cfg->fc_mx) {
		struct nlattr *nla;
		int remaining;

		nla_for_each_attr(nla, cfg->fc_mx, cfg->fc_mx_len, remaining) {
1300
			int type = nla_type(nla);
1301 1302 1303

			if (type) {
				if (type > RTAX_MAX) {
L
Linus Torvalds 已提交
1304 1305 1306
					err = -EINVAL;
					goto out;
				}
1307

1308
				dst_metric_set(&rt->dst, type, nla_get_u32(nla));
L
Linus Torvalds 已提交
1309 1310 1311 1312
			}
		}
	}

1313
	if (!dst_mtu(&rt->dst))
1314
		dst_metric_set(&rt->dst, RTAX_MTU, ipv6_get_mtu(dev));
1315
	if (!dst_metric(&rt->dst, RTAX_ADVMSS))
1316
		dst_metric_set(&rt->dst, RTAX_ADVMSS, ipv6_advmss(net, dst_mtu(&rt->dst)));
1317
	rt->dst.dev = dev;
L
Linus Torvalds 已提交
1318
	rt->rt6i_idev = idev;
T
Thomas Graf 已提交
1319
	rt->rt6i_table = table;
1320

1321
	cfg->fc_nlinfo.nl_net = dev_net(dev);
1322

1323
	return __ip6_ins_rt(rt, &cfg->fc_nlinfo);
L
Linus Torvalds 已提交
1324 1325 1326 1327 1328 1329 1330

out:
	if (dev)
		dev_put(dev);
	if (idev)
		in6_dev_put(idev);
	if (rt)
1331
		dst_free(&rt->dst);
L
Linus Torvalds 已提交
1332 1333 1334
	return err;
}

1335
static int __ip6_del_rt(struct rt6_info *rt, struct nl_info *info)
L
Linus Torvalds 已提交
1336 1337
{
	int err;
T
Thomas Graf 已提交
1338
	struct fib6_table *table;
1339
	struct net *net = dev_net(rt->rt6i_dev);
L
Linus Torvalds 已提交
1340

1341
	if (rt == net->ipv6.ip6_null_entry)
1342 1343
		return -ENOENT;

T
Thomas Graf 已提交
1344 1345
	table = rt->rt6i_table;
	write_lock_bh(&table->tb6_lock);
L
Linus Torvalds 已提交
1346

1347
	err = fib6_del(rt, info);
1348
	dst_release(&rt->dst);
L
Linus Torvalds 已提交
1349

T
Thomas Graf 已提交
1350
	write_unlock_bh(&table->tb6_lock);
L
Linus Torvalds 已提交
1351 1352 1353 1354

	return err;
}

1355 1356
int ip6_del_rt(struct rt6_info *rt)
{
1357
	struct nl_info info = {
1358
		.nl_net = dev_net(rt->rt6i_dev),
1359
	};
1360
	return __ip6_del_rt(rt, &info);
1361 1362
}

1363
static int ip6_route_del(struct fib6_config *cfg)
L
Linus Torvalds 已提交
1364
{
T
Thomas Graf 已提交
1365
	struct fib6_table *table;
L
Linus Torvalds 已提交
1366 1367 1368 1369
	struct fib6_node *fn;
	struct rt6_info *rt;
	int err = -ESRCH;

1370
	table = fib6_get_table(cfg->fc_nlinfo.nl_net, cfg->fc_table);
T
Thomas Graf 已提交
1371 1372 1373 1374
	if (table == NULL)
		return err;

	read_lock_bh(&table->tb6_lock);
L
Linus Torvalds 已提交
1375

T
Thomas Graf 已提交
1376
	fn = fib6_locate(&table->tb6_root,
1377 1378
			 &cfg->fc_dst, cfg->fc_dst_len,
			 &cfg->fc_src, cfg->fc_src_len);
1379

L
Linus Torvalds 已提交
1380
	if (fn) {
1381
		for (rt = fn->leaf; rt; rt = rt->dst.rt6_next) {
1382
			if (cfg->fc_ifindex &&
L
Linus Torvalds 已提交
1383
			    (rt->rt6i_dev == NULL ||
1384
			     rt->rt6i_dev->ifindex != cfg->fc_ifindex))
L
Linus Torvalds 已提交
1385
				continue;
1386 1387
			if (cfg->fc_flags & RTF_GATEWAY &&
			    !ipv6_addr_equal(&cfg->fc_gateway, &rt->rt6i_gateway))
L
Linus Torvalds 已提交
1388
				continue;
1389
			if (cfg->fc_metric && cfg->fc_metric != rt->rt6i_metric)
L
Linus Torvalds 已提交
1390
				continue;
1391
			dst_hold(&rt->dst);
T
Thomas Graf 已提交
1392
			read_unlock_bh(&table->tb6_lock);
L
Linus Torvalds 已提交
1393

1394
			return __ip6_del_rt(rt, &cfg->fc_nlinfo);
L
Linus Torvalds 已提交
1395 1396
		}
	}
T
Thomas Graf 已提交
1397
	read_unlock_bh(&table->tb6_lock);
L
Linus Torvalds 已提交
1398 1399 1400 1401 1402 1403 1404

	return err;
}

/*
 *	Handle redirects
 */
1405 1406 1407 1408 1409
struct ip6rd_flowi {
	struct flowi fl;
	struct in6_addr gateway;
};

1410 1411
static struct rt6_info *__ip6_route_redirect(struct net *net,
					     struct fib6_table *table,
1412 1413
					     struct flowi *fl,
					     int flags)
L
Linus Torvalds 已提交
1414
{
1415 1416
	struct ip6rd_flowi *rdfl = (struct ip6rd_flowi *)fl;
	struct rt6_info *rt;
1417
	struct fib6_node *fn;
T
Thomas Graf 已提交
1418

L
Linus Torvalds 已提交
1419
	/*
1420 1421 1422 1423 1424 1425 1426 1427
	 * 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
1428 1429
	 */

T
Thomas Graf 已提交
1430
	read_lock_bh(&table->tb6_lock);
1431
	fn = fib6_lookup(&table->tb6_root, &fl->fl6_dst, &fl->fl6_src);
1432
restart:
1433
	for (rt = fn->leaf; rt; rt = rt->dst.rt6_next) {
1434 1435 1436 1437 1438 1439 1440 1441 1442 1443 1444 1445
		/*
		 * 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;
1446
		if (fl->oif != rt->rt6i_dev->ifindex)
1447
			continue;
1448
		if (!ipv6_addr_equal(&rdfl->gateway, &rt->rt6i_gateway))
1449 1450 1451
			continue;
		break;
	}
1452

1453
	if (!rt)
1454 1455
		rt = net->ipv6.ip6_null_entry;
	BACKTRACK(net, &fl->fl6_src);
1456
out:
1457
	dst_hold(&rt->dst);
1458

T
Thomas Graf 已提交
1459
	read_unlock_bh(&table->tb6_lock);
1460

1461 1462 1463 1464 1465 1466 1467 1468
	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)
{
1469
	int flags = RT6_LOOKUP_F_HAS_SADDR;
1470
	struct net *net = dev_net(dev);
1471 1472 1473
	struct ip6rd_flowi rdfl = {
		.fl = {
			.oif = dev->ifindex,
1474 1475
			.fl6_dst = *dest,
			.fl6_src = *src,
1476 1477
		},
	};
1478

1479 1480
	ipv6_addr_copy(&rdfl.gateway, gateway);

1481 1482
	if (rt6_need_strict(dest))
		flags |= RT6_LOOKUP_F_IFACE;
1483

1484
	return (struct rt6_info *)fib6_rule_lookup(net, (struct flowi *)&rdfl,
1485
						   flags, __ip6_route_redirect);
1486 1487 1488 1489 1490 1491 1492 1493
}

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;
1494
	struct net *net = dev_net(neigh->dev);
1495 1496 1497

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

1498
	if (rt == net->ipv6.ip6_null_entry) {
L
Linus Torvalds 已提交
1499 1500 1501
		if (net_ratelimit())
			printk(KERN_DEBUG "rt6_redirect: source isn't a valid nexthop "
			       "for redirect target\n");
1502
		goto out;
L
Linus Torvalds 已提交
1503 1504 1505 1506 1507 1508
	}

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

1509
	neigh_update(neigh, lladdr, NUD_STALE,
L
Linus Torvalds 已提交
1510 1511 1512 1513 1514 1515 1516 1517 1518 1519 1520
		     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
	 */
1521
	dst_confirm(&rt->dst);
L
Linus Torvalds 已提交
1522 1523

	/* Duplicate redirect: silently ignore. */
1524
	if (neigh == rt->dst.neighbour)
L
Linus Torvalds 已提交
1525 1526 1527 1528 1529 1530 1531 1532 1533 1534 1535 1536
		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;
1537
	nrt->dst.flags |= DST_HOST;
L
Linus Torvalds 已提交
1538 1539 1540 1541

	ipv6_addr_copy(&nrt->rt6i_gateway, (struct in6_addr*)neigh->primary_key);
	nrt->rt6i_nexthop = neigh_clone(neigh);
	/* Reset pmtu, it may be better */
1542 1543 1544
	dst_metric_set(&nrt->dst, RTAX_MTU, ipv6_get_mtu(neigh->dev));
	dst_metric_set(&nrt->dst, RTAX_ADVMSS, ipv6_advmss(dev_net(neigh->dev),
							   dst_mtu(&nrt->dst)));
L
Linus Torvalds 已提交
1545

1546
	if (ip6_ins_rt(nrt))
L
Linus Torvalds 已提交
1547 1548
		goto out;

1549 1550
	netevent.old = &rt->dst;
	netevent.new = &nrt->dst;
1551 1552
	call_netevent_notifiers(NETEVENT_REDIRECT, &netevent);

L
Linus Torvalds 已提交
1553
	if (rt->rt6i_flags&RTF_CACHE) {
1554
		ip6_del_rt(rt);
L
Linus Torvalds 已提交
1555 1556 1557 1558
		return;
	}

out:
1559
	dst_release(&rt->dst);
L
Linus Torvalds 已提交
1560 1561 1562 1563 1564 1565 1566
}

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

1567 1568
static void rt6_do_pmtu_disc(struct in6_addr *daddr, struct in6_addr *saddr,
			     struct net *net, u32 pmtu, int ifindex)
L
Linus Torvalds 已提交
1569 1570 1571 1572
{
	struct rt6_info *rt, *nrt;
	int allfrag = 0;

1573
	rt = rt6_lookup(net, daddr, saddr, ifindex, 0);
L
Linus Torvalds 已提交
1574 1575 1576
	if (rt == NULL)
		return;

1577
	if (pmtu >= dst_mtu(&rt->dst))
L
Linus Torvalds 已提交
1578 1579 1580 1581
		goto out;

	if (pmtu < IPV6_MIN_MTU) {
		/*
1582
		 * According to RFC2460, PMTU is set to the IPv6 Minimum Link
L
Linus Torvalds 已提交
1583 1584 1585 1586 1587 1588 1589 1590 1591 1592 1593 1594
		 * 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
	 */
1595
	dst_confirm(&rt->dst);
L
Linus Torvalds 已提交
1596 1597 1598 1599 1600 1601 1602

	/* 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) {
1603 1604 1605 1606 1607 1608
		dst_metric_set(&rt->dst, RTAX_MTU, pmtu);
		if (allfrag) {
			u32 features = dst_metric(&rt->dst, RTAX_FEATURES);
			features |= RTAX_FEATURE_ALLFRAG;
			dst_metric_set(&rt->dst, RTAX_FEATURES, features);
		}
1609
		dst_set_expires(&rt->dst, net->ipv6.sysctl.ip6_rt_mtu_expires);
L
Linus Torvalds 已提交
1610 1611 1612 1613 1614 1615 1616 1617 1618
		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.
	 */
1619
	if (!rt->rt6i_nexthop && !(rt->rt6i_flags & RTF_NONEXTHOP))
1620
		nrt = rt6_alloc_cow(rt, daddr, saddr);
1621 1622
	else
		nrt = rt6_alloc_clone(rt, daddr);
1623

1624
	if (nrt) {
1625 1626 1627 1628 1629 1630
		dst_metric_set(&nrt->dst, RTAX_MTU, pmtu);
		if (allfrag) {
			u32 features = dst_metric(&nrt->dst, RTAX_FEATURES);
			features |= RTAX_FEATURE_ALLFRAG;
			dst_metric_set(&nrt->dst, RTAX_FEATURES, features);
		}
1631 1632 1633 1634 1635 1636 1637

		/* 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.
		 */
1638
		dst_set_expires(&nrt->dst, net->ipv6.sysctl.ip6_rt_mtu_expires);
1639 1640
		nrt->rt6i_flags |= RTF_DYNAMIC|RTF_EXPIRES;

1641
		ip6_ins_rt(nrt);
L
Linus Torvalds 已提交
1642 1643
	}
out:
1644
	dst_release(&rt->dst);
L
Linus Torvalds 已提交
1645 1646
}

1647 1648 1649 1650 1651 1652 1653 1654 1655 1656 1657 1658 1659 1660 1661 1662 1663 1664 1665 1666 1667
void rt6_pmtu_discovery(struct in6_addr *daddr, struct in6_addr *saddr,
			struct net_device *dev, u32 pmtu)
{
	struct net *net = dev_net(dev);

	/*
	 * RFC 1981 states that a node "MUST reduce the size of the packets it
	 * is sending along the path" that caused the Packet Too Big message.
	 * Since it's not possible in the general case to determine which
	 * interface was used to send the original packet, we update the MTU
	 * on the interface that will be used to send future packets. We also
	 * update the MTU on the interface that received the Packet Too Big in
	 * case the original packet was forced out that interface with
	 * SO_BINDTODEVICE or similar. This is the next best thing to the
	 * correct behaviour, which would be to update the MTU on all
	 * interfaces.
	 */
	rt6_do_pmtu_disc(daddr, saddr, net, pmtu, 0);
	rt6_do_pmtu_disc(daddr, saddr, net, pmtu, dev->ifindex);
}

L
Linus Torvalds 已提交
1668 1669 1670 1671 1672 1673
/*
 *	Misc support functions
 */

static struct rt6_info * ip6_rt_copy(struct rt6_info *ort)
{
1674
	struct net *net = dev_net(ort->rt6i_dev);
1675
	struct rt6_info *rt = ip6_dst_alloc(&net->ipv6.ip6_dst_ops);
L
Linus Torvalds 已提交
1676 1677

	if (rt) {
1678 1679 1680
		rt->dst.input = ort->dst.input;
		rt->dst.output = ort->dst.output;

1681
		dst_copy_metrics(&rt->dst, &ort->dst);
1682 1683 1684 1685
		rt->dst.error = ort->dst.error;
		rt->dst.dev = ort->dst.dev;
		if (rt->dst.dev)
			dev_hold(rt->dst.dev);
L
Linus Torvalds 已提交
1686 1687 1688
		rt->rt6i_idev = ort->rt6i_idev;
		if (rt->rt6i_idev)
			in6_dev_hold(rt->rt6i_idev);
1689
		rt->dst.lastuse = jiffies;
L
Linus Torvalds 已提交
1690 1691 1692 1693 1694 1695 1696 1697 1698 1699
		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
1700
		rt->rt6i_table = ort->rt6i_table;
L
Linus Torvalds 已提交
1701 1702 1703 1704
	}
	return rt;
}

1705
#ifdef CONFIG_IPV6_ROUTE_INFO
1706 1707
static struct rt6_info *rt6_get_route_info(struct net *net,
					   struct in6_addr *prefix, int prefixlen,
1708 1709 1710 1711
					   struct in6_addr *gwaddr, int ifindex)
{
	struct fib6_node *fn;
	struct rt6_info *rt = NULL;
T
Thomas Graf 已提交
1712 1713
	struct fib6_table *table;

1714
	table = fib6_get_table(net, RT6_TABLE_INFO);
T
Thomas Graf 已提交
1715 1716
	if (table == NULL)
		return NULL;
1717

T
Thomas Graf 已提交
1718 1719
	write_lock_bh(&table->tb6_lock);
	fn = fib6_locate(&table->tb6_root, prefix ,prefixlen, NULL, 0);
1720 1721 1722
	if (!fn)
		goto out;

1723
	for (rt = fn->leaf; rt; rt = rt->dst.rt6_next) {
1724 1725 1726 1727 1728 1729
		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;
1730
		dst_hold(&rt->dst);
1731 1732 1733
		break;
	}
out:
T
Thomas Graf 已提交
1734
	write_unlock_bh(&table->tb6_lock);
1735 1736 1737
	return rt;
}

1738 1739
static struct rt6_info *rt6_add_route_info(struct net *net,
					   struct in6_addr *prefix, int prefixlen,
1740 1741 1742
					   struct in6_addr *gwaddr, int ifindex,
					   unsigned pref)
{
1743 1744
	struct fib6_config cfg = {
		.fc_table	= RT6_TABLE_INFO,
1745
		.fc_metric	= IP6_RT_PRIO_USER,
1746 1747 1748 1749
		.fc_ifindex	= ifindex,
		.fc_dst_len	= prefixlen,
		.fc_flags	= RTF_GATEWAY | RTF_ADDRCONF | RTF_ROUTEINFO |
				  RTF_UP | RTF_PREF(pref),
1750 1751 1752
		.fc_nlinfo.pid = 0,
		.fc_nlinfo.nlh = NULL,
		.fc_nlinfo.nl_net = net,
1753 1754 1755 1756
	};

	ipv6_addr_copy(&cfg.fc_dst, prefix);
	ipv6_addr_copy(&cfg.fc_gateway, gwaddr);
1757

1758 1759
	/* We should treat it as a default route if prefix length is 0. */
	if (!prefixlen)
1760
		cfg.fc_flags |= RTF_DEFAULT;
1761

1762
	ip6_route_add(&cfg);
1763

1764
	return rt6_get_route_info(net, prefix, prefixlen, gwaddr, ifindex);
1765 1766 1767
}
#endif

L
Linus Torvalds 已提交
1768
struct rt6_info *rt6_get_dflt_router(struct in6_addr *addr, struct net_device *dev)
1769
{
L
Linus Torvalds 已提交
1770
	struct rt6_info *rt;
T
Thomas Graf 已提交
1771
	struct fib6_table *table;
L
Linus Torvalds 已提交
1772

1773
	table = fib6_get_table(dev_net(dev), RT6_TABLE_DFLT);
T
Thomas Graf 已提交
1774 1775
	if (table == NULL)
		return NULL;
L
Linus Torvalds 已提交
1776

T
Thomas Graf 已提交
1777
	write_lock_bh(&table->tb6_lock);
1778
	for (rt = table->tb6_root.leaf; rt; rt=rt->dst.rt6_next) {
L
Linus Torvalds 已提交
1779
		if (dev == rt->rt6i_dev &&
1780
		    ((rt->rt6i_flags & (RTF_ADDRCONF | RTF_DEFAULT)) == (RTF_ADDRCONF | RTF_DEFAULT)) &&
L
Linus Torvalds 已提交
1781 1782 1783 1784
		    ipv6_addr_equal(&rt->rt6i_gateway, addr))
			break;
	}
	if (rt)
1785
		dst_hold(&rt->dst);
T
Thomas Graf 已提交
1786
	write_unlock_bh(&table->tb6_lock);
L
Linus Torvalds 已提交
1787 1788 1789 1790
	return rt;
}

struct rt6_info *rt6_add_dflt_router(struct in6_addr *gwaddr,
1791 1792
				     struct net_device *dev,
				     unsigned int pref)
L
Linus Torvalds 已提交
1793
{
1794 1795
	struct fib6_config cfg = {
		.fc_table	= RT6_TABLE_DFLT,
1796
		.fc_metric	= IP6_RT_PRIO_USER,
1797 1798 1799
		.fc_ifindex	= dev->ifindex,
		.fc_flags	= RTF_GATEWAY | RTF_ADDRCONF | RTF_DEFAULT |
				  RTF_UP | RTF_EXPIRES | RTF_PREF(pref),
1800 1801
		.fc_nlinfo.pid = 0,
		.fc_nlinfo.nlh = NULL,
1802
		.fc_nlinfo.nl_net = dev_net(dev),
1803
	};
L
Linus Torvalds 已提交
1804

1805
	ipv6_addr_copy(&cfg.fc_gateway, gwaddr);
L
Linus Torvalds 已提交
1806

1807
	ip6_route_add(&cfg);
L
Linus Torvalds 已提交
1808 1809 1810 1811

	return rt6_get_dflt_router(gwaddr, dev);
}

1812
void rt6_purge_dflt_routers(struct net *net)
L
Linus Torvalds 已提交
1813 1814
{
	struct rt6_info *rt;
T
Thomas Graf 已提交
1815 1816 1817
	struct fib6_table *table;

	/* NOTE: Keep consistent with rt6_get_dflt_router */
1818
	table = fib6_get_table(net, RT6_TABLE_DFLT);
T
Thomas Graf 已提交
1819 1820
	if (table == NULL)
		return;
L
Linus Torvalds 已提交
1821 1822

restart:
T
Thomas Graf 已提交
1823
	read_lock_bh(&table->tb6_lock);
1824
	for (rt = table->tb6_root.leaf; rt; rt = rt->dst.rt6_next) {
L
Linus Torvalds 已提交
1825
		if (rt->rt6i_flags & (RTF_DEFAULT | RTF_ADDRCONF)) {
1826
			dst_hold(&rt->dst);
T
Thomas Graf 已提交
1827
			read_unlock_bh(&table->tb6_lock);
1828
			ip6_del_rt(rt);
L
Linus Torvalds 已提交
1829 1830 1831
			goto restart;
		}
	}
T
Thomas Graf 已提交
1832
	read_unlock_bh(&table->tb6_lock);
L
Linus Torvalds 已提交
1833 1834
}

1835 1836
static void rtmsg_to_fib6_config(struct net *net,
				 struct in6_rtmsg *rtmsg,
1837 1838 1839 1840 1841 1842 1843 1844 1845 1846 1847 1848
				 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;

1849
	cfg->fc_nlinfo.nl_net = net;
1850

1851 1852 1853 1854 1855
	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);
}

1856
int ipv6_route_ioctl(struct net *net, unsigned int cmd, void __user *arg)
L
Linus Torvalds 已提交
1857
{
1858
	struct fib6_config cfg;
L
Linus Torvalds 已提交
1859 1860 1861 1862 1863 1864 1865 1866 1867 1868 1869 1870
	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;
1871

1872
		rtmsg_to_fib6_config(net, &rtmsg, &cfg);
1873

L
Linus Torvalds 已提交
1874 1875 1876
		rtnl_lock();
		switch (cmd) {
		case SIOCADDRT:
1877
			err = ip6_route_add(&cfg);
L
Linus Torvalds 已提交
1878 1879
			break;
		case SIOCDELRT:
1880
			err = ip6_route_del(&cfg);
L
Linus Torvalds 已提交
1881 1882 1883 1884 1885 1886 1887
			break;
		default:
			err = -EINVAL;
		}
		rtnl_unlock();

		return err;
1888
	}
L
Linus Torvalds 已提交
1889 1890 1891 1892 1893 1894 1895 1896

	return -EINVAL;
}

/*
 *	Drop the packet on the floor
 */

1897
static int ip6_pkt_drop(struct sk_buff *skb, u8 code, int ipstats_mib_noroutes)
L
Linus Torvalds 已提交
1898
{
1899
	int type;
E
Eric Dumazet 已提交
1900
	struct dst_entry *dst = skb_dst(skb);
1901 1902
	switch (ipstats_mib_noroutes) {
	case IPSTATS_MIB_INNOROUTES:
1903
		type = ipv6_addr_type(&ipv6_hdr(skb)->daddr);
U
Ulrich Weber 已提交
1904
		if (type == IPV6_ADDR_ANY) {
1905 1906
			IP6_INC_STATS(dev_net(dst->dev), ip6_dst_idev(dst),
				      IPSTATS_MIB_INADDRERRORS);
1907 1908 1909 1910
			break;
		}
		/* FALLTHROUGH */
	case IPSTATS_MIB_OUTNOROUTES:
1911 1912
		IP6_INC_STATS(dev_net(dst->dev), ip6_dst_idev(dst),
			      ipstats_mib_noroutes);
1913 1914
		break;
	}
1915
	icmpv6_send(skb, ICMPV6_DEST_UNREACH, code, 0);
L
Linus Torvalds 已提交
1916 1917 1918 1919
	kfree_skb(skb);
	return 0;
}

1920 1921
static int ip6_pkt_discard(struct sk_buff *skb)
{
1922
	return ip6_pkt_drop(skb, ICMPV6_NOROUTE, IPSTATS_MIB_INNOROUTES);
1923 1924
}

1925
static int ip6_pkt_discard_out(struct sk_buff *skb)
L
Linus Torvalds 已提交
1926
{
E
Eric Dumazet 已提交
1927
	skb->dev = skb_dst(skb)->dev;
1928
	return ip6_pkt_drop(skb, ICMPV6_NOROUTE, IPSTATS_MIB_OUTNOROUTES);
L
Linus Torvalds 已提交
1929 1930
}

1931 1932
#ifdef CONFIG_IPV6_MULTIPLE_TABLES

1933 1934
static int ip6_pkt_prohibit(struct sk_buff *skb)
{
1935
	return ip6_pkt_drop(skb, ICMPV6_ADM_PROHIBITED, IPSTATS_MIB_INNOROUTES);
1936 1937 1938 1939
}

static int ip6_pkt_prohibit_out(struct sk_buff *skb)
{
E
Eric Dumazet 已提交
1940
	skb->dev = skb_dst(skb)->dev;
1941
	return ip6_pkt_drop(skb, ICMPV6_ADM_PROHIBITED, IPSTATS_MIB_OUTNOROUTES);
1942 1943
}

1944 1945
#endif

L
Linus Torvalds 已提交
1946 1947 1948 1949 1950 1951 1952 1953
/*
 *	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)
{
1954
	struct net *net = dev_net(idev->dev);
1955
	struct rt6_info *rt = ip6_dst_alloc(&net->ipv6.ip6_dst_ops);
1956
	struct neighbour *neigh;
L
Linus Torvalds 已提交
1957

1958 1959 1960 1961
	if (rt == NULL) {
		if (net_ratelimit())
			pr_warning("IPv6:  Maximum number of routes reached,"
				   " consider increasing route/max_size.\n");
L
Linus Torvalds 已提交
1962
		return ERR_PTR(-ENOMEM);
1963
	}
L
Linus Torvalds 已提交
1964

1965
	dev_hold(net->loopback_dev);
L
Linus Torvalds 已提交
1966 1967
	in6_dev_hold(idev);

1968 1969 1970
	rt->dst.flags = DST_HOST;
	rt->dst.input = ip6_input;
	rt->dst.output = ip6_output;
1971
	rt->rt6i_dev = net->loopback_dev;
L
Linus Torvalds 已提交
1972
	rt->rt6i_idev = idev;
1973 1974 1975
	dst_metric_set(&rt->dst, RTAX_MTU, ipv6_get_mtu(rt->rt6i_dev));
	dst_metric_set(&rt->dst, RTAX_ADVMSS, ipv6_advmss(net, dst_mtu(&rt->dst)));
	dst_metric_set(&rt->dst, RTAX_HOPLIMIT, -1);
1976
	rt->dst.obsolete = -1;
L
Linus Torvalds 已提交
1977 1978

	rt->rt6i_flags = RTF_UP | RTF_NONEXTHOP;
1979 1980 1981
	if (anycast)
		rt->rt6i_flags |= RTF_ANYCAST;
	else
L
Linus Torvalds 已提交
1982
		rt->rt6i_flags |= RTF_LOCAL;
1983 1984
	neigh = ndisc_get_neigh(rt->rt6i_dev, &rt->rt6i_gateway);
	if (IS_ERR(neigh)) {
1985
		dst_free(&rt->dst);
1986 1987 1988 1989 1990 1991 1992

		/* We are casting this because that is the return
		 * value type.  But an errno encoded pointer is the
		 * same regardless of the underlying pointer type,
		 * and that's what we are returning.  So this is OK.
		 */
		return (struct rt6_info *) neigh;
L
Linus Torvalds 已提交
1993
	}
1994
	rt->rt6i_nexthop = neigh;
L
Linus Torvalds 已提交
1995 1996 1997

	ipv6_addr_copy(&rt->rt6i_dst.addr, addr);
	rt->rt6i_dst.plen = 128;
1998
	rt->rt6i_table = fib6_get_table(net, RT6_TABLE_LOCAL);
L
Linus Torvalds 已提交
1999

2000
	atomic_set(&rt->dst.__refcnt, 1);
L
Linus Torvalds 已提交
2001 2002 2003 2004

	return rt;
}

2005 2006 2007 2008 2009
struct arg_dev_net {
	struct net_device *dev;
	struct net *net;
};

L
Linus Torvalds 已提交
2010 2011
static int fib6_ifdown(struct rt6_info *rt, void *arg)
{
2012 2013 2014 2015 2016
	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
2017 2018 2019 2020 2021 2022
		RT6_TRACE("deleted by ifdown %p\n", rt);
		return -1;
	}
	return 0;
}

2023
void rt6_ifdown(struct net *net, struct net_device *dev)
L
Linus Torvalds 已提交
2024
{
2025 2026 2027 2028 2029 2030
	struct arg_dev_net adn = {
		.dev = dev,
		.net = net,
	};

	fib6_clean_all(net, fib6_ifdown, 0, &adn);
2031
	icmp6_clean_all(fib6_ifdown, &adn);
L
Linus Torvalds 已提交
2032 2033 2034 2035 2036 2037 2038 2039 2040 2041 2042 2043
}

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;
2044
	struct net *net = dev_net(arg->dev);
L
Linus Torvalds 已提交
2045 2046 2047 2048 2049 2050 2051 2052 2053 2054 2055 2056 2057 2058 2059 2060 2061 2062 2063 2064 2065 2066 2067 2068 2069 2070

	/* 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 &&
2071 2072 2073 2074
	    !dst_metric_locked(&rt->dst, RTAX_MTU) &&
	    (dst_mtu(&rt->dst) >= arg->mtu ||
	     (dst_mtu(&rt->dst) < arg->mtu &&
	      dst_mtu(&rt->dst) == idev->cnf.mtu6))) {
2075 2076
		dst_metric_set(&rt->dst, RTAX_MTU, arg->mtu);
		dst_metric_set(&rt->dst, RTAX_ADVMSS, ipv6_advmss(net, arg->mtu));
2077
	}
L
Linus Torvalds 已提交
2078 2079 2080 2081 2082
	return 0;
}

void rt6_mtu_change(struct net_device *dev, unsigned mtu)
{
T
Thomas Graf 已提交
2083 2084 2085 2086
	struct rt6_mtu_change_arg arg = {
		.dev = dev,
		.mtu = mtu,
	};
L
Linus Torvalds 已提交
2087

2088
	fib6_clean_all(dev_net(dev), rt6_mtu_change_route, 0, &arg);
L
Linus Torvalds 已提交
2089 2090
}

2091
static const struct nla_policy rtm_ipv6_policy[RTA_MAX+1] = {
2092
	[RTA_GATEWAY]           = { .len = sizeof(struct in6_addr) },
2093
	[RTA_OIF]               = { .type = NLA_U32 },
2094
	[RTA_IIF]		= { .type = NLA_U32 },
2095 2096 2097 2098 2099 2100
	[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
2101
{
2102 2103 2104
	struct rtmsg *rtm;
	struct nlattr *tb[RTA_MAX+1];
	int err;
L
Linus Torvalds 已提交
2105

2106 2107 2108
	err = nlmsg_parse(nlh, sizeof(*rtm), tb, RTA_MAX, rtm_ipv6_policy);
	if (err < 0)
		goto errout;
L
Linus Torvalds 已提交
2109

2110 2111 2112 2113 2114 2115 2116 2117 2118 2119 2120 2121 2122
	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;

2123 2124 2125
	if (rtm->rtm_type == RTN_LOCAL)
		cfg->fc_flags |= RTF_LOCAL;

2126 2127
	cfg->fc_nlinfo.pid = NETLINK_CB(skb).pid;
	cfg->fc_nlinfo.nlh = nlh;
2128
	cfg->fc_nlinfo.nl_net = sock_net(skb->sk);
2129 2130 2131 2132

	if (tb[RTA_GATEWAY]) {
		nla_memcpy(&cfg->fc_gateway, tb[RTA_GATEWAY], 16);
		cfg->fc_flags |= RTF_GATEWAY;
L
Linus Torvalds 已提交
2133
	}
2134 2135 2136 2137 2138 2139 2140 2141

	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
2142
	}
2143 2144 2145 2146 2147 2148 2149 2150

	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
2151
	}
2152 2153 2154 2155 2156 2157 2158 2159 2160 2161

	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
2162
	}
2163 2164 2165 2166 2167 2168 2169

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

	err = 0;
errout:
	return err;
L
Linus Torvalds 已提交
2170 2171
}

2172
static int inet6_rtm_delroute(struct sk_buff *skb, struct nlmsghdr* nlh, void *arg)
L
Linus Torvalds 已提交
2173
{
2174 2175
	struct fib6_config cfg;
	int err;
L
Linus Torvalds 已提交
2176

2177 2178 2179 2180 2181
	err = rtm_to_fib6_config(skb, nlh, &cfg);
	if (err < 0)
		return err;

	return ip6_route_del(&cfg);
L
Linus Torvalds 已提交
2182 2183
}

2184
static int inet6_rtm_newroute(struct sk_buff *skb, struct nlmsghdr* nlh, void *arg)
L
Linus Torvalds 已提交
2185
{
2186 2187
	struct fib6_config cfg;
	int err;
L
Linus Torvalds 已提交
2188

2189 2190 2191 2192 2193
	err = rtm_to_fib6_config(skb, nlh, &cfg);
	if (err < 0)
		return err;

	return ip6_route_add(&cfg);
L
Linus Torvalds 已提交
2194 2195
}

2196 2197 2198 2199 2200 2201 2202 2203 2204 2205 2206
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 */
2207
	       + RTAX_MAX * nla_total_size(4) /* RTA_METRICS */
2208 2209 2210
	       + nla_total_size(sizeof(struct rta_cacheinfo));
}

2211 2212
static int rt6_fill_node(struct net *net,
			 struct sk_buff *skb, struct rt6_info *rt,
2213 2214
			 struct in6_addr *dst, struct in6_addr *src,
			 int iif, int type, u32 pid, u32 seq,
2215
			 int prefix, int nowait, unsigned int flags)
L
Linus Torvalds 已提交
2216 2217
{
	struct rtmsg *rtm;
2218
	struct nlmsghdr *nlh;
2219
	long expires;
2220
	u32 table;
L
Linus Torvalds 已提交
2221 2222 2223 2224 2225 2226 2227 2228

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

2229 2230
	nlh = nlmsg_put(skb, pid, seq, type, sizeof(*rtm), flags);
	if (nlh == NULL)
2231
		return -EMSGSIZE;
2232 2233

	rtm = nlmsg_data(nlh);
L
Linus Torvalds 已提交
2234 2235 2236 2237
	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
2238
	if (rt->rt6i_table)
2239
		table = rt->rt6i_table->tb6_id;
T
Thomas Graf 已提交
2240
	else
2241 2242
		table = RT6_TABLE_UNSPEC;
	rtm->rtm_table = table;
2243
	NLA_PUT_U32(skb, RTA_TABLE, table);
L
Linus Torvalds 已提交
2244 2245
	if (rt->rt6i_flags&RTF_REJECT)
		rtm->rtm_type = RTN_UNREACHABLE;
2246 2247
	else if (rt->rt6i_flags&RTF_LOCAL)
		rtm->rtm_type = RTN_LOCAL;
L
Linus Torvalds 已提交
2248 2249 2250 2251 2252 2253 2254 2255 2256 2257 2258 2259 2260 2261 2262 2263 2264 2265
	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) {
2266
		NLA_PUT(skb, RTA_DST, 16, dst);
2267
		rtm->rtm_dst_len = 128;
L
Linus Torvalds 已提交
2268
	} else if (rtm->rtm_dst_len)
2269
		NLA_PUT(skb, RTA_DST, 16, &rt->rt6i_dst.addr);
L
Linus Torvalds 已提交
2270 2271
#ifdef CONFIG_IPV6_SUBTREES
	if (src) {
2272
		NLA_PUT(skb, RTA_SRC, 16, src);
2273
		rtm->rtm_src_len = 128;
L
Linus Torvalds 已提交
2274
	} else if (rtm->rtm_src_len)
2275
		NLA_PUT(skb, RTA_SRC, 16, &rt->rt6i_src.addr);
L
Linus Torvalds 已提交
2276
#endif
2277 2278 2279
	if (iif) {
#ifdef CONFIG_IPV6_MROUTE
		if (ipv6_addr_is_multicast(&rt->rt6i_dst.addr)) {
2280
			int err = ip6mr_get_route(net, skb, rtm, nowait);
2281 2282 2283 2284 2285 2286 2287 2288 2289 2290 2291 2292 2293 2294
			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) {
2295
		struct inet6_dev *idev = ip6_dst_idev(&rt->dst);
L
Linus Torvalds 已提交
2296
		struct in6_addr saddr_buf;
2297
		if (ipv6_dev_get_saddr(net, idev ? idev->dev : NULL,
2298
				       dst, 0, &saddr_buf) == 0)
2299
			NLA_PUT(skb, RTA_PREFSRC, 16, &saddr_buf);
L
Linus Torvalds 已提交
2300
	}
2301

2302
	if (rtnetlink_put_metrics(skb, dst_metrics_ptr(&rt->dst)) < 0)
2303 2304
		goto nla_put_failure;

2305 2306
	if (rt->dst.neighbour)
		NLA_PUT(skb, RTA_GATEWAY, 16, &rt->dst.neighbour->primary_key);
2307

2308
	if (rt->dst.dev)
2309 2310 2311
		NLA_PUT_U32(skb, RTA_OIF, rt->rt6i_dev->ifindex);

	NLA_PUT_U32(skb, RTA_PRIORITY, rt->rt6i_metric);
2312

2313 2314 2315 2316 2317 2318
	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;
2319

2320 2321
	if (rtnl_put_cacheinfo(skb, &rt->dst, 0, 0, 0,
			       expires, rt->dst.error) < 0)
2322
		goto nla_put_failure;
2323 2324 2325 2326

	return nlmsg_end(skb, nlh);

nla_put_failure:
2327 2328
	nlmsg_cancel(skb, nlh);
	return -EMSGSIZE;
L
Linus Torvalds 已提交
2329 2330
}

2331
int rt6_dump_route(struct rt6_info *rt, void *p_arg)
L
Linus Torvalds 已提交
2332 2333 2334 2335
{
	struct rt6_rtnl_dump_arg *arg = (struct rt6_rtnl_dump_arg *) p_arg;
	int prefix;

2336 2337
	if (nlmsg_len(arg->cb->nlh) >= sizeof(struct rtmsg)) {
		struct rtmsg *rtm = nlmsg_data(arg->cb->nlh);
L
Linus Torvalds 已提交
2338 2339 2340 2341
		prefix = (rtm->rtm_flags & RTM_F_PREFIX) != 0;
	} else
		prefix = 0;

2342 2343
	return rt6_fill_node(arg->net,
		     arg->skb, rt, NULL, NULL, 0, RTM_NEWROUTE,
L
Linus Torvalds 已提交
2344
		     NETLINK_CB(arg->cb->skb).pid, arg->cb->nlh->nlmsg_seq,
2345
		     prefix, 0, NLM_F_MULTI);
L
Linus Torvalds 已提交
2346 2347
}

2348
static int inet6_rtm_getroute(struct sk_buff *in_skb, struct nlmsghdr* nlh, void *arg)
L
Linus Torvalds 已提交
2349
{
2350
	struct net *net = sock_net(in_skb->sk);
2351 2352
	struct nlattr *tb[RTA_MAX+1];
	struct rt6_info *rt;
L
Linus Torvalds 已提交
2353
	struct sk_buff *skb;
2354
	struct rtmsg *rtm;
L
Linus Torvalds 已提交
2355
	struct flowi fl;
2356
	int err, iif = 0;
L
Linus Torvalds 已提交
2357

2358 2359 2360
	err = nlmsg_parse(nlh, sizeof(*rtm), tb, RTA_MAX, rtm_ipv6_policy);
	if (err < 0)
		goto errout;
L
Linus Torvalds 已提交
2361

2362
	err = -EINVAL;
L
Linus Torvalds 已提交
2363 2364
	memset(&fl, 0, sizeof(fl));

2365 2366 2367 2368 2369 2370 2371 2372 2373 2374 2375 2376 2377 2378 2379 2380 2381 2382 2383
	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
2384 2385 2386

	if (iif) {
		struct net_device *dev;
2387
		dev = __dev_get_by_index(net, iif);
L
Linus Torvalds 已提交
2388 2389
		if (!dev) {
			err = -ENODEV;
2390
			goto errout;
L
Linus Torvalds 已提交
2391 2392 2393
		}
	}

2394 2395 2396 2397 2398
	skb = alloc_skb(NLMSG_GOODSIZE, GFP_KERNEL);
	if (skb == NULL) {
		err = -ENOBUFS;
		goto errout;
	}
L
Linus Torvalds 已提交
2399

2400 2401 2402
	/* Reserve room for dummy headers, this skb can pass
	   through good chunk of routing engine.
	 */
2403
	skb_reset_mac_header(skb);
2404
	skb_reserve(skb, MAX_HEADER + sizeof(struct ipv6hdr));
L
Linus Torvalds 已提交
2405

2406
	rt = (struct rt6_info*) ip6_route_output(net, NULL, &fl);
2407
	skb_dst_set(skb, &rt->dst);
L
Linus Torvalds 已提交
2408

2409
	err = rt6_fill_node(net, skb, rt, &fl.fl6_dst, &fl.fl6_src, iif,
L
Linus Torvalds 已提交
2410
			    RTM_NEWROUTE, NETLINK_CB(in_skb).pid,
2411
			    nlh->nlmsg_seq, 0, 0, 0);
L
Linus Torvalds 已提交
2412
	if (err < 0) {
2413 2414
		kfree_skb(skb);
		goto errout;
L
Linus Torvalds 已提交
2415 2416
	}

2417
	err = rtnl_unicast(skb, net, NETLINK_CB(in_skb).pid);
2418
errout:
L
Linus Torvalds 已提交
2419 2420 2421
	return err;
}

2422
void inet6_rt_notify(int event, struct rt6_info *rt, struct nl_info *info)
L
Linus Torvalds 已提交
2423 2424
{
	struct sk_buff *skb;
2425
	struct net *net = info->nl_net;
2426 2427 2428 2429 2430
	u32 seq;
	int err;

	err = -ENOBUFS;
	seq = info->nlh != NULL ? info->nlh->nlmsg_seq : 0;
2431

2432
	skb = nlmsg_new(rt6_nlmsg_size(), gfp_any());
2433 2434 2435
	if (skb == NULL)
		goto errout;

2436
	err = rt6_fill_node(net, skb, rt, NULL, NULL, 0,
2437
				event, info->pid, seq, 0, 0, 0);
2438 2439 2440 2441 2442 2443
	if (err < 0) {
		/* -EMSGSIZE implies BUG in rt6_nlmsg_size() */
		WARN_ON(err == -EMSGSIZE);
		kfree_skb(skb);
		goto errout;
	}
2444 2445 2446
	rtnl_notify(skb, net, info->pid, RTNLGRP_IPV6_ROUTE,
		    info->nlh, gfp_any());
	return;
2447 2448
errout:
	if (err < 0)
2449
		rtnl_set_sk_err(net, RTNLGRP_IPV6_ROUTE, err);
L
Linus Torvalds 已提交
2450 2451
}

2452 2453 2454 2455
static int ip6_route_dev_notify(struct notifier_block *this,
				unsigned long event, void *data)
{
	struct net_device *dev = (struct net_device *)data;
2456
	struct net *net = dev_net(dev);
2457 2458

	if (event == NETDEV_REGISTER && (dev->flags & IFF_LOOPBACK)) {
2459
		net->ipv6.ip6_null_entry->dst.dev = dev;
2460 2461
		net->ipv6.ip6_null_entry->rt6i_idev = in6_dev_get(dev);
#ifdef CONFIG_IPV6_MULTIPLE_TABLES
2462
		net->ipv6.ip6_prohibit_entry->dst.dev = dev;
2463
		net->ipv6.ip6_prohibit_entry->rt6i_idev = in6_dev_get(dev);
2464
		net->ipv6.ip6_blk_hole_entry->dst.dev = dev;
2465 2466 2467 2468 2469 2470 2471
		net->ipv6.ip6_blk_hole_entry->rt6i_idev = in6_dev_get(dev);
#endif
	}

	return NOTIFY_OK;
}

L
Linus Torvalds 已提交
2472 2473 2474 2475 2476 2477 2478 2479 2480 2481 2482 2483 2484 2485 2486 2487 2488
/*
 *	/proc
 */

#ifdef CONFIG_PROC_FS

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)
{
2489
	struct seq_file *m = p_arg;
L
Linus Torvalds 已提交
2490

2491
	seq_printf(m, "%pi6 %02x ", &rt->rt6i_dst.addr, rt->rt6i_dst.plen);
L
Linus Torvalds 已提交
2492 2493

#ifdef CONFIG_IPV6_SUBTREES
2494
	seq_printf(m, "%pi6 %02x ", &rt->rt6i_src.addr, rt->rt6i_src.plen);
L
Linus Torvalds 已提交
2495
#else
2496
	seq_puts(m, "00000000000000000000000000000000 00 ");
L
Linus Torvalds 已提交
2497 2498 2499
#endif

	if (rt->rt6i_nexthop) {
2500
		seq_printf(m, "%pi6", rt->rt6i_nexthop->primary_key);
L
Linus Torvalds 已提交
2501
	} else {
2502
		seq_puts(m, "00000000000000000000000000000000");
L
Linus Torvalds 已提交
2503
	}
2504
	seq_printf(m, " %08x %08x %08x %08x %8s\n",
2505 2506
		   rt->rt6i_metric, atomic_read(&rt->dst.__refcnt),
		   rt->dst.__use, rt->rt6i_flags,
2507
		   rt->rt6i_dev ? rt->rt6i_dev->name : "");
L
Linus Torvalds 已提交
2508 2509 2510
	return 0;
}

2511
static int ipv6_route_show(struct seq_file *m, void *v)
L
Linus Torvalds 已提交
2512
{
2513 2514
	struct net *net = (struct net *)m->private;
	fib6_clean_all(net, rt6_info_route, 0, m);
2515 2516
	return 0;
}
L
Linus Torvalds 已提交
2517

2518 2519
static int ipv6_route_open(struct inode *inode, struct file *file)
{
2520
	return single_open_net(inode, file, ipv6_route_show);
2521 2522
}

2523 2524 2525 2526 2527
static const struct file_operations ipv6_route_proc_fops = {
	.owner		= THIS_MODULE,
	.open		= ipv6_route_open,
	.read		= seq_read,
	.llseek		= seq_lseek,
2528
	.release	= single_release_net,
2529 2530
};

L
Linus Torvalds 已提交
2531 2532
static int rt6_stats_seq_show(struct seq_file *seq, void *v)
{
2533
	struct net *net = (struct net *)seq->private;
L
Linus Torvalds 已提交
2534
	seq_printf(seq, "%04x %04x %04x %04x %04x %04x %04x\n",
2535 2536 2537 2538 2539
		   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,
2540
		   dst_entries_get_slow(&net->ipv6.ip6_dst_ops),
2541
		   net->ipv6.rt6_stats->fib_discarded_routes);
L
Linus Torvalds 已提交
2542 2543 2544 2545 2546 2547

	return 0;
}

static int rt6_stats_seq_open(struct inode *inode, struct file *file)
{
2548
	return single_open_net(inode, file, rt6_stats_seq_show);
2549 2550
}

2551
static const struct file_operations rt6_stats_seq_fops = {
L
Linus Torvalds 已提交
2552 2553 2554 2555
	.owner	 = THIS_MODULE,
	.open	 = rt6_stats_seq_open,
	.read	 = seq_read,
	.llseek	 = seq_lseek,
2556
	.release = single_release_net,
L
Linus Torvalds 已提交
2557 2558 2559 2560 2561 2562
};
#endif	/* CONFIG_PROC_FS */

#ifdef CONFIG_SYSCTL

static
2563
int ipv6_sysctl_rtcache_flush(ctl_table *ctl, int write,
L
Linus Torvalds 已提交
2564 2565
			      void __user *buffer, size_t *lenp, loff_t *ppos)
{
2566 2567
	struct net *net = current->nsproxy->net_ns;
	int delay = net->ipv6.sysctl.flush_delay;
L
Linus Torvalds 已提交
2568
	if (write) {
2569
		proc_dointvec(ctl, write, buffer, lenp, ppos);
2570
		fib6_run_gc(delay <= 0 ? ~0UL : (unsigned long)delay, net);
L
Linus Torvalds 已提交
2571 2572 2573 2574 2575
		return 0;
	} else
		return -EINVAL;
}

2576
ctl_table ipv6_route_table_template[] = {
2577
	{
L
Linus Torvalds 已提交
2578
		.procname	=	"flush",
2579
		.data		=	&init_net.ipv6.sysctl.flush_delay,
L
Linus Torvalds 已提交
2580
		.maxlen		=	sizeof(int),
2581
		.mode		=	0200,
A
Alexey Dobriyan 已提交
2582
		.proc_handler	=	ipv6_sysctl_rtcache_flush
L
Linus Torvalds 已提交
2583 2584 2585
	},
	{
		.procname	=	"gc_thresh",
2586
		.data		=	&ip6_dst_ops_template.gc_thresh,
L
Linus Torvalds 已提交
2587 2588
		.maxlen		=	sizeof(int),
		.mode		=	0644,
A
Alexey Dobriyan 已提交
2589
		.proc_handler	=	proc_dointvec,
L
Linus Torvalds 已提交
2590 2591 2592
	},
	{
		.procname	=	"max_size",
2593
		.data		=	&init_net.ipv6.sysctl.ip6_rt_max_size,
L
Linus Torvalds 已提交
2594 2595
		.maxlen		=	sizeof(int),
		.mode		=	0644,
A
Alexey Dobriyan 已提交
2596
		.proc_handler	=	proc_dointvec,
L
Linus Torvalds 已提交
2597 2598 2599
	},
	{
		.procname	=	"gc_min_interval",
2600
		.data		=	&init_net.ipv6.sysctl.ip6_rt_gc_min_interval,
L
Linus Torvalds 已提交
2601 2602
		.maxlen		=	sizeof(int),
		.mode		=	0644,
A
Alexey Dobriyan 已提交
2603
		.proc_handler	=	proc_dointvec_jiffies,
L
Linus Torvalds 已提交
2604 2605 2606
	},
	{
		.procname	=	"gc_timeout",
2607
		.data		=	&init_net.ipv6.sysctl.ip6_rt_gc_timeout,
L
Linus Torvalds 已提交
2608 2609
		.maxlen		=	sizeof(int),
		.mode		=	0644,
A
Alexey Dobriyan 已提交
2610
		.proc_handler	=	proc_dointvec_jiffies,
L
Linus Torvalds 已提交
2611 2612 2613
	},
	{
		.procname	=	"gc_interval",
2614
		.data		=	&init_net.ipv6.sysctl.ip6_rt_gc_interval,
L
Linus Torvalds 已提交
2615 2616
		.maxlen		=	sizeof(int),
		.mode		=	0644,
A
Alexey Dobriyan 已提交
2617
		.proc_handler	=	proc_dointvec_jiffies,
L
Linus Torvalds 已提交
2618 2619 2620
	},
	{
		.procname	=	"gc_elasticity",
2621
		.data		=	&init_net.ipv6.sysctl.ip6_rt_gc_elasticity,
L
Linus Torvalds 已提交
2622 2623
		.maxlen		=	sizeof(int),
		.mode		=	0644,
2624
		.proc_handler	=	proc_dointvec,
L
Linus Torvalds 已提交
2625 2626 2627
	},
	{
		.procname	=	"mtu_expires",
2628
		.data		=	&init_net.ipv6.sysctl.ip6_rt_mtu_expires,
L
Linus Torvalds 已提交
2629 2630
		.maxlen		=	sizeof(int),
		.mode		=	0644,
A
Alexey Dobriyan 已提交
2631
		.proc_handler	=	proc_dointvec_jiffies,
L
Linus Torvalds 已提交
2632 2633 2634
	},
	{
		.procname	=	"min_adv_mss",
2635
		.data		=	&init_net.ipv6.sysctl.ip6_rt_min_advmss,
L
Linus Torvalds 已提交
2636 2637
		.maxlen		=	sizeof(int),
		.mode		=	0644,
2638
		.proc_handler	=	proc_dointvec,
L
Linus Torvalds 已提交
2639 2640 2641
	},
	{
		.procname	=	"gc_min_interval_ms",
2642
		.data		=	&init_net.ipv6.sysctl.ip6_rt_gc_min_interval,
L
Linus Torvalds 已提交
2643 2644
		.maxlen		=	sizeof(int),
		.mode		=	0644,
A
Alexey Dobriyan 已提交
2645
		.proc_handler	=	proc_dointvec_ms_jiffies,
L
Linus Torvalds 已提交
2646
	},
2647
	{ }
L
Linus Torvalds 已提交
2648 2649
};

2650
struct ctl_table * __net_init ipv6_route_sysctl_init(struct net *net)
2651 2652 2653 2654 2655 2656
{
	struct ctl_table *table;

	table = kmemdup(ipv6_route_table_template,
			sizeof(ipv6_route_table_template),
			GFP_KERNEL);
2657 2658 2659

	if (table) {
		table[0].data = &net->ipv6.sysctl.flush_delay;
2660
		table[1].data = &net->ipv6.ip6_dst_ops.gc_thresh;
2661 2662 2663 2664 2665 2666 2667
		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;
2668
		table[9].data = &net->ipv6.sysctl.ip6_rt_gc_min_interval;
2669 2670
	}

2671 2672
	return table;
}
L
Linus Torvalds 已提交
2673 2674
#endif

2675
static int __net_init ip6_route_net_init(struct net *net)
2676
{
2677
	int ret = -ENOMEM;
2678

2679 2680
	memcpy(&net->ipv6.ip6_dst_ops, &ip6_dst_ops_template,
	       sizeof(net->ipv6.ip6_dst_ops));
2681

2682 2683 2684
	if (dst_entries_init(&net->ipv6.ip6_dst_ops) < 0)
		goto out_ip6_dst_ops;

2685 2686 2687 2688
	net->ipv6.ip6_null_entry = kmemdup(&ip6_null_entry_template,
					   sizeof(*net->ipv6.ip6_null_entry),
					   GFP_KERNEL);
	if (!net->ipv6.ip6_null_entry)
2689
		goto out_ip6_dst_entries;
2690
	net->ipv6.ip6_null_entry->dst.path =
2691
		(struct dst_entry *)net->ipv6.ip6_null_entry;
2692
	net->ipv6.ip6_null_entry->dst.ops = &net->ipv6.ip6_dst_ops;
2693
	dst_metric_set(&net->ipv6.ip6_null_entry->dst, RTAX_HOPLIMIT, 255);
2694 2695 2696 2697 2698

#ifdef CONFIG_IPV6_MULTIPLE_TABLES
	net->ipv6.ip6_prohibit_entry = kmemdup(&ip6_prohibit_entry_template,
					       sizeof(*net->ipv6.ip6_prohibit_entry),
					       GFP_KERNEL);
2699 2700
	if (!net->ipv6.ip6_prohibit_entry)
		goto out_ip6_null_entry;
2701
	net->ipv6.ip6_prohibit_entry->dst.path =
2702
		(struct dst_entry *)net->ipv6.ip6_prohibit_entry;
2703
	net->ipv6.ip6_prohibit_entry->dst.ops = &net->ipv6.ip6_dst_ops;
2704
	dst_metric_set(&net->ipv6.ip6_prohibit_entry->dst, RTAX_HOPLIMIT, 255);
2705 2706 2707 2708

	net->ipv6.ip6_blk_hole_entry = kmemdup(&ip6_blk_hole_entry_template,
					       sizeof(*net->ipv6.ip6_blk_hole_entry),
					       GFP_KERNEL);
2709 2710
	if (!net->ipv6.ip6_blk_hole_entry)
		goto out_ip6_prohibit_entry;
2711
	net->ipv6.ip6_blk_hole_entry->dst.path =
2712
		(struct dst_entry *)net->ipv6.ip6_blk_hole_entry;
2713
	net->ipv6.ip6_blk_hole_entry->dst.ops = &net->ipv6.ip6_dst_ops;
2714
	dst_metric_set(&net->ipv6.ip6_blk_hole_entry->dst, RTAX_HOPLIMIT, 255);
2715 2716
#endif

2717 2718 2719 2720 2721 2722 2723 2724 2725
	net->ipv6.sysctl.flush_delay = 0;
	net->ipv6.sysctl.ip6_rt_max_size = 4096;
	net->ipv6.sysctl.ip6_rt_gc_min_interval = HZ / 2;
	net->ipv6.sysctl.ip6_rt_gc_timeout = 60*HZ;
	net->ipv6.sysctl.ip6_rt_gc_interval = 30*HZ;
	net->ipv6.sysctl.ip6_rt_gc_elasticity = 9;
	net->ipv6.sysctl.ip6_rt_mtu_expires = 10*60*HZ;
	net->ipv6.sysctl.ip6_rt_min_advmss = IPV6_MIN_MTU - 20 - 40;

2726 2727 2728 2729
#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
2730 2731
	net->ipv6.ip6_rt_gc_expire = 30*HZ;

2732 2733 2734
	ret = 0;
out:
	return ret;
2735

2736 2737 2738 2739 2740 2741
#ifdef CONFIG_IPV6_MULTIPLE_TABLES
out_ip6_prohibit_entry:
	kfree(net->ipv6.ip6_prohibit_entry);
out_ip6_null_entry:
	kfree(net->ipv6.ip6_null_entry);
#endif
2742 2743
out_ip6_dst_entries:
	dst_entries_destroy(&net->ipv6.ip6_dst_ops);
2744 2745
out_ip6_dst_ops:
	goto out;
2746 2747
}

2748
static void __net_exit ip6_route_net_exit(struct net *net)
2749 2750 2751 2752 2753
{
#ifdef CONFIG_PROC_FS
	proc_net_remove(net, "ipv6_route");
	proc_net_remove(net, "rt6_stats");
#endif
2754 2755 2756 2757 2758
	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
2759
	dst_entries_destroy(&net->ipv6.ip6_dst_ops);
2760 2761 2762 2763 2764 2765 2766
}

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

2767 2768 2769 2770 2771
static struct notifier_block ip6_route_dev_notifier = {
	.notifier_call = ip6_route_dev_notify,
	.priority = 0,
};

2772
int __init ip6_route_init(void)
L
Linus Torvalds 已提交
2773
{
2774 2775
	int ret;

2776 2777
	ret = -ENOMEM;
	ip6_dst_ops_template.kmem_cachep =
A
Alexey Dobriyan 已提交
2778
		kmem_cache_create("ip6_dst_cache", sizeof(struct rt6_info), 0,
2779
				  SLAB_HWCACHE_ALIGN, NULL);
2780
	if (!ip6_dst_ops_template.kmem_cachep)
2781
		goto out;
2782

2783
	ret = dst_entries_init(&ip6_dst_blackhole_ops);
2784
	if (ret)
2785 2786
		goto out_kmem_cache;

2787 2788 2789 2790
	ret = register_pernet_subsys(&ip6_route_net_ops);
	if (ret)
		goto out_dst_entries;

2791 2792
	ip6_dst_blackhole_ops.kmem_cachep = ip6_dst_ops_template.kmem_cachep;

2793 2794 2795
	/* 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 */
2796
	init_net.ipv6.ip6_null_entry->dst.dev = init_net.loopback_dev;
2797 2798
	init_net.ipv6.ip6_null_entry->rt6i_idev = in6_dev_get(init_net.loopback_dev);
  #ifdef CONFIG_IPV6_MULTIPLE_TABLES
2799
	init_net.ipv6.ip6_prohibit_entry->dst.dev = init_net.loopback_dev;
2800
	init_net.ipv6.ip6_prohibit_entry->rt6i_idev = in6_dev_get(init_net.loopback_dev);
2801
	init_net.ipv6.ip6_blk_hole_entry->dst.dev = init_net.loopback_dev;
2802 2803
	init_net.ipv6.ip6_blk_hole_entry->rt6i_idev = in6_dev_get(init_net.loopback_dev);
  #endif
2804 2805
	ret = fib6_init();
	if (ret)
2806
		goto out_register_subsys;
2807 2808 2809

	ret = xfrm6_init();
	if (ret)
2810
		goto out_fib6_init;
2811

2812 2813 2814
	ret = fib6_rules_init();
	if (ret)
		goto xfrm6_init;
2815

2816 2817 2818 2819 2820
	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;
2821

2822
	ret = register_netdevice_notifier(&ip6_route_dev_notifier);
2823 2824
	if (ret)
		goto fib6_rules_init;
2825

2826 2827 2828 2829 2830 2831 2832 2833 2834
out:
	return ret;

fib6_rules_init:
	fib6_rules_cleanup();
xfrm6_init:
	xfrm6_fini();
out_fib6_init:
	fib6_gc_cleanup();
2835 2836
out_register_subsys:
	unregister_pernet_subsys(&ip6_route_net_ops);
2837 2838
out_dst_entries:
	dst_entries_destroy(&ip6_dst_blackhole_ops);
2839
out_kmem_cache:
2840
	kmem_cache_destroy(ip6_dst_ops_template.kmem_cachep);
2841
	goto out;
L
Linus Torvalds 已提交
2842 2843 2844 2845
}

void ip6_route_cleanup(void)
{
2846
	unregister_netdevice_notifier(&ip6_route_dev_notifier);
T
Thomas Graf 已提交
2847
	fib6_rules_cleanup();
L
Linus Torvalds 已提交
2848 2849
	xfrm6_fini();
	fib6_gc_cleanup();
2850
	unregister_pernet_subsys(&ip6_route_net_ops);
2851
	dst_entries_destroy(&ip6_dst_blackhole_ops);
2852
	kmem_cache_destroy(ip6_dst_ops_template.kmem_cachep);
L
Linus Torvalds 已提交
2853
}