udp.c 39.1 KB
Newer Older
L
Linus Torvalds 已提交
1 2
/*
 *	UDP over IPv6
3
 *	Linux INET6 implementation
L
Linus Torvalds 已提交
4 5
 *
 *	Authors:
6
 *	Pedro Roque		<roque@di.fc.ul.pt>
L
Linus Torvalds 已提交
7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34
 *
 *	Based on linux/ipv4/udp.c
 *
 *	Fixes:
 *	Hideaki YOSHIFUJI	:	sin6_scope_id support
 *	YOSHIFUJI Hideaki @USAGI and:	Support IPV6_V6ONLY socket option, which
 *	Alexey Kuznetsov		allow both IPv4 and IPv6 sockets to bind
 *					a single port at the same time.
 *      Kazunori MIYAZAWA @USAGI:       change process style to use ip6_append_data
 *      YOSHIFUJI Hideaki @USAGI:	convert /proc/net/udp6 to seq_file.
 *
 *	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.
 */

#include <linux/errno.h>
#include <linux/types.h>
#include <linux/socket.h>
#include <linux/sockios.h>
#include <linux/net.h>
#include <linux/in6.h>
#include <linux/netdevice.h>
#include <linux/if_arp.h>
#include <linux/ipv6.h>
#include <linux/icmpv6.h>
#include <linux/init.h>
35
#include <linux/module.h>
36
#include <linux/skbuff.h>
37
#include <linux/slab.h>
L
Linus Torvalds 已提交
38 39
#include <asm/uaccess.h>

40
#include <net/addrconf.h>
L
Linus Torvalds 已提交
41 42 43 44 45
#include <net/ndisc.h>
#include <net/protocol.h>
#include <net/transp_v6.h>
#include <net/ip6_route.h>
#include <net/raw.h>
46
#include <net/tcp_states.h>
L
Linus Torvalds 已提交
47 48
#include <net/ip6_checksum.h>
#include <net/xfrm.h>
49
#include <net/inet6_hashtables.h>
50
#include <net/busy_poll.h>
51
#include <net/sock_reuseport.h>
L
Linus Torvalds 已提交
52 53 54

#include <linux/proc_fs.h>
#include <linux/seq_file.h>
55
#include <trace/events/skb.h>
56
#include "udp_impl.h"
L
Linus Torvalds 已提交
57

58 59 60 61 62
static u32 udp6_ehashfn(const struct net *net,
			const struct in6_addr *laddr,
			const u16 lport,
			const struct in6_addr *faddr,
			const __be16 fport)
63
{
64 65 66 67 68 69 70 71 72 73 74 75 76
	static u32 udp6_ehash_secret __read_mostly;
	static u32 udp_ipv6_hash_secret __read_mostly;

	u32 lhash, fhash;

	net_get_random_once(&udp6_ehash_secret,
			    sizeof(udp6_ehash_secret));
	net_get_random_once(&udp_ipv6_hash_secret,
			    sizeof(udp_ipv6_hash_secret));

	lhash = (__force u32)laddr->s6_addr32[3];
	fhash = __ipv6_addr_jhash(faddr, udp_ipv6_hash_secret);

77
	return __inet6_ehashfn(lhash, lport, fhash, fport,
78
			       udp_ipv6_hash_secret + net_hash_mix(net));
79 80
}

81 82 83
static u32 udp6_portaddr_hash(const struct net *net,
			      const struct in6_addr *addr6,
			      unsigned int port)
84 85 86 87 88
{
	unsigned int hash, mix = net_hash_mix(net);

	if (ipv6_addr_any(addr6))
		hash = jhash_1word(0, mix);
B
Brian Haley 已提交
89
	else if (ipv6_addr_v4mapped(addr6))
90
		hash = jhash_1word((__force u32)addr6->s6_addr32[3], mix);
91
	else
92
		hash = jhash2((__force u32 *)addr6->s6_addr32, 4, mix);
93 94 95 96

	return hash ^ port;
}

97
int udp_v6_get_port(struct sock *sk, unsigned short snum)
L
Linus Torvalds 已提交
98
{
E
Eric Dumazet 已提交
99 100
	unsigned int hash2_nulladdr =
		udp6_portaddr_hash(sock_net(sk), &in6addr_any, snum);
101
	unsigned int hash2_partial =
102
		udp6_portaddr_hash(sock_net(sk), &sk->sk_v6_rcv_saddr, 0);
E
Eric Dumazet 已提交
103

104
	/* precompute partial secondary hash */
E
Eric Dumazet 已提交
105 106
	udp_sk(sk)->udp_portaddr_hash = hash2_partial;
	return udp_lib_get_port(sk, snum, ipv6_rcv_saddr_equal, hash2_nulladdr);
L
Linus Torvalds 已提交
107 108
}

E
Eric Dumazet 已提交
109 110 111
static void udp_v6_rehash(struct sock *sk)
{
	u16 new_hash = udp6_portaddr_hash(sock_net(sk),
112
					  &sk->sk_v6_rcv_saddr,
E
Eric Dumazet 已提交
113 114 115 116 117
					  inet_sk(sk)->inet_num);

	udp_lib_rehash(sk, new_hash);
}

118 119
static inline int compute_score(struct sock *sk, struct net *net,
				unsigned short hnum,
120 121
				const struct in6_addr *saddr, __be16 sport,
				const struct in6_addr *daddr, __be16 dport,
122 123
				int dif)
{
124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156
	int score;
	struct inet_sock *inet;

	if (!net_eq(sock_net(sk), net) ||
	    udp_sk(sk)->udp_port_hash != hnum ||
	    sk->sk_family != PF_INET6)
		return -1;

	score = 0;
	inet = inet_sk(sk);

	if (inet->inet_dport) {
		if (inet->inet_dport != sport)
			return -1;
		score++;
	}

	if (!ipv6_addr_any(&sk->sk_v6_rcv_saddr)) {
		if (!ipv6_addr_equal(&sk->sk_v6_rcv_saddr, daddr))
			return -1;
		score++;
	}

	if (!ipv6_addr_any(&sk->sk_v6_daddr)) {
		if (!ipv6_addr_equal(&sk->sk_v6_daddr, saddr))
			return -1;
		score++;
	}

	if (sk->sk_bound_dev_if) {
		if (sk->sk_bound_dev_if != dif)
			return -1;
		score++;
157
	}
158

159 160 161
	if (sk->sk_incoming_cpu == raw_smp_processor_id())
		score++;

162 163 164
	return score;
}

E
Eric Dumazet 已提交
165
static inline int compute_score2(struct sock *sk, struct net *net,
166 167 168
				 const struct in6_addr *saddr, __be16 sport,
				 const struct in6_addr *daddr,
				 unsigned short hnum, int dif)
E
Eric Dumazet 已提交
169
{
170 171
	int score;
	struct inet_sock *inet;
E
Eric Dumazet 已提交
172

173 174 175 176
	if (!net_eq(sock_net(sk), net) ||
	    udp_sk(sk)->udp_port_hash != hnum ||
	    sk->sk_family != PF_INET6)
		return -1;
E
Eric Dumazet 已提交
177

178 179 180 181 182 183 184 185 186 187 188 189 190 191
	if (!ipv6_addr_equal(&sk->sk_v6_rcv_saddr, daddr))
		return -1;

	score = 0;
	inet = inet_sk(sk);

	if (inet->inet_dport) {
		if (inet->inet_dport != sport)
			return -1;
		score++;
	}

	if (!ipv6_addr_any(&sk->sk_v6_daddr)) {
		if (!ipv6_addr_equal(&sk->sk_v6_daddr, saddr))
E
Eric Dumazet 已提交
192
			return -1;
193
		score++;
E
Eric Dumazet 已提交
194
	}
195 196 197 198 199 200 201

	if (sk->sk_bound_dev_if) {
		if (sk->sk_bound_dev_if != dif)
			return -1;
		score++;
	}

202 203 204
	if (sk->sk_incoming_cpu == raw_smp_processor_id())
		score++;

E
Eric Dumazet 已提交
205 206 207 208 209 210 211
	return score;
}

/* called with read_rcu_lock() */
static struct sock *udp6_lib_lookup2(struct net *net,
		const struct in6_addr *saddr, __be16 sport,
		const struct in6_addr *daddr, unsigned int hnum, int dif,
212 213
		struct udp_hslot *hslot2, unsigned int slot2,
		struct sk_buff *skb)
E
Eric Dumazet 已提交
214 215 216
{
	struct sock *sk, *result;
	struct hlist_nulls_node *node;
217
	int score, badness, matches = 0, reuseport = 0;
218
	bool select_ok = true;
219
	u32 hash = 0;
E
Eric Dumazet 已提交
220 221 222 223 224 225 226 227 228 229

begin:
	result = NULL;
	badness = -1;
	udp_portaddr_for_each_entry_rcu(sk, node, &hslot2->head) {
		score = compute_score2(sk, net, saddr, sport,
				      daddr, hnum, dif);
		if (score > badness) {
			result = sk;
			badness = score;
230 231
			reuseport = sk->sk_reuseport;
			if (reuseport) {
232 233
				hash = udp6_ehashfn(net, daddr, hnum,
						    saddr, sport);
234 235 236 237 238 239 240 241 242 243
				if (select_ok) {
					struct sock *sk2;

					sk2 = reuseport_select_sock(sk, hash, skb,
							sizeof(struct udphdr));
					if (sk2) {
						result = sk2;
						select_ok = false;
						goto found;
					}
244
				}
245
				matches = 1;
246
			}
247 248
		} else if (score == badness && reuseport) {
			matches++;
249
			if (reciprocal_scale(hash, matches) == 0)
250 251
				result = sk;
			hash = next_pseudo_random32(hash);
E
Eric Dumazet 已提交
252 253 254 255 256 257 258 259 260 261 262
		}
	}
	/*
	 * if the nulls value we got at the end of this lookup is
	 * not the expected one, we must restart lookup.
	 * We probably met an item that was moved to another chain.
	 */
	if (get_nulls_value(node) != slot2)
		goto begin;

	if (result) {
263
found:
E
Eric Dumazet 已提交
264
		if (unlikely(!atomic_inc_not_zero_hint(&result->sk_refcnt, 2)))
E
Eric Dumazet 已提交
265 266 267 268 269 270 271 272 273 274
			result = NULL;
		else if (unlikely(compute_score2(result, net, saddr, sport,
				  daddr, hnum, dif) < badness)) {
			sock_put(result);
			goto begin;
		}
	}
	return result;
}

275
struct sock *__udp6_lib_lookup(struct net *net,
276 277
				      const struct in6_addr *saddr, __be16 sport,
				      const struct in6_addr *daddr, __be16 dport,
278 279
				      int dif, struct udp_table *udptable,
				      struct sk_buff *skb)
L
Linus Torvalds 已提交
280
{
281
	struct sock *sk, *result;
282
	struct hlist_nulls_node *node;
L
Linus Torvalds 已提交
283
	unsigned short hnum = ntohs(dport);
E
Eric Dumazet 已提交
284 285
	unsigned int hash2, slot2, slot = udp_hashfn(net, hnum, udptable->mask);
	struct udp_hslot *hslot2, *hslot = &udptable->hash[slot];
286
	int score, badness, matches = 0, reuseport = 0;
287
	bool select_ok = true;
288
	u32 hash = 0;
289

290
	rcu_read_lock();
E
Eric Dumazet 已提交
291 292 293 294 295 296 297 298 299
	if (hslot->count > 10) {
		hash2 = udp6_portaddr_hash(net, daddr, hnum);
		slot2 = hash2 & udptable->mask;
		hslot2 = &udptable->hash2[slot2];
		if (hslot->count < hslot2->count)
			goto begin;

		result = udp6_lib_lookup2(net, saddr, sport,
					  daddr, hnum, dif,
300
					  hslot2, slot2, skb);
E
Eric Dumazet 已提交
301 302 303 304 305 306 307
		if (!result) {
			hash2 = udp6_portaddr_hash(net, &in6addr_any, hnum);
			slot2 = hash2 & udptable->mask;
			hslot2 = &udptable->hash2[slot2];
			if (hslot->count < hslot2->count)
				goto begin;

308 309
			result = udp6_lib_lookup2(net, saddr, sport,
						  &in6addr_any, hnum, dif,
310
						  hslot2, slot2, skb);
E
Eric Dumazet 已提交
311 312 313 314
		}
		rcu_read_unlock();
		return result;
	}
315 316 317
begin:
	result = NULL;
	badness = -1;
318
	sk_nulls_for_each_rcu(sk, node, &hslot->head) {
319 320 321 322
		score = compute_score(sk, net, hnum, saddr, sport, daddr, dport, dif);
		if (score > badness) {
			result = sk;
			badness = score;
323 324
			reuseport = sk->sk_reuseport;
			if (reuseport) {
325 326
				hash = udp6_ehashfn(net, daddr, hnum,
						    saddr, sport);
327 328 329 330
				if (select_ok) {
					struct sock *sk2;

					sk2 = reuseport_select_sock(sk, hash, skb,
331
							sizeof(struct udphdr));
332 333 334 335 336
					if (sk2) {
						result = sk2;
						select_ok = false;
						goto found;
					}
337
				}
338 339 340 341
				matches = 1;
			}
		} else if (score == badness && reuseport) {
			matches++;
342
			if (reciprocal_scale(hash, matches) == 0)
343 344
				result = sk;
			hash = next_pseudo_random32(hash);
L
Linus Torvalds 已提交
345 346
		}
	}
347 348 349 350 351
	/*
	 * if the nulls value we got at the end of this lookup is
	 * not the expected one, we must restart lookup.
	 * We probably met an item that was moved to another chain.
	 */
E
Eric Dumazet 已提交
352
	if (get_nulls_value(node) != slot)
353 354
		goto begin;

355
	if (result) {
356
found:
E
Eric Dumazet 已提交
357
		if (unlikely(!atomic_inc_not_zero_hint(&result->sk_refcnt, 2)))
358 359 360 361 362 363 364 365
			result = NULL;
		else if (unlikely(compute_score(result, net, hnum, saddr, sport,
					daddr, dport, dif) < badness)) {
			sock_put(result);
			goto begin;
		}
	}
	rcu_read_unlock();
L
Linus Torvalds 已提交
366 367
	return result;
}
368
EXPORT_SYMBOL_GPL(__udp6_lib_lookup);
L
Linus Torvalds 已提交
369

370 371
static struct sock *__udp6_lib_lookup_skb(struct sk_buff *skb,
					  __be16 sport, __be16 dport,
372
					  struct udp_table *udptable)
373
{
374
	struct sock *sk;
375
	const struct ipv6hdr *iph = ipv6_hdr(skb);
376

377 378
	sk = skb_steal_sock(skb);
	if (unlikely(sk))
379
		return sk;
E
Eric Dumazet 已提交
380 381
	return __udp6_lib_lookup(dev_net(skb_dst(skb)->dev), &iph->saddr, sport,
				 &iph->daddr, dport, inet6_iif(skb),
382
				 udptable, skb);
383 384
}

385 386 387
struct sock *udp6_lib_lookup(struct net *net, const struct in6_addr *saddr, __be16 sport,
			     const struct in6_addr *daddr, __be16 dport, int dif)
{
388
	return __udp6_lib_lookup(net, saddr, sport, daddr, dport, dif, &udp_table, NULL);
389 390 391
}
EXPORT_SYMBOL_GPL(udp6_lib_lookup);

L
Linus Torvalds 已提交
392
/*
393 394
 *	This should be easy, if there is something there we
 *	return it, otherwise we block.
L
Linus Torvalds 已提交
395 396
 */

397
int udpv6_recvmsg(struct sock *sk, struct msghdr *msg, size_t len,
L
Linus Torvalds 已提交
398 399 400 401
		  int noblock, int flags, int *addr_len)
{
	struct ipv6_pinfo *np = inet6_sk(sk);
	struct inet_sock *inet = inet_sk(sk);
402
	struct sk_buff *skb;
403
	unsigned int ulen, copied;
404
	int peeked, off = 0;
405 406
	int err;
	int is_udplite = IS_UDPLITE(sk);
407
	bool checksum_valid = false;
408
	int is_udp4;
409
	bool slow;
L
Linus Torvalds 已提交
410 411

	if (flags & MSG_ERRQUEUE)
412
		return ipv6_recv_error(sk, msg, len, addr_len);
L
Linus Torvalds 已提交
413

414
	if (np->rxpmtu && np->rxopt.bits.rxpmtu)
415
		return ipv6_recv_rxpmtu(sk, msg, len, addr_len);
416

L
Linus Torvalds 已提交
417
try_again:
418
	skb = __skb_recv_datagram(sk, flags | (noblock ? MSG_DONTWAIT : 0),
419
				  &peeked, &off, &err);
L
Linus Torvalds 已提交
420 421 422
	if (!skb)
		goto out;

423
	ulen = skb->len - sizeof(struct udphdr);
424 425 426 427
	copied = len;
	if (copied > ulen)
		copied = ulen;
	else if (copied < ulen)
428
		msg->msg_flags |= MSG_TRUNC;
L
Linus Torvalds 已提交
429

430 431
	is_udp4 = (skb->protocol == htons(ETH_P_IP));

432
	/*
433 434 435
	 * If checksum is needed at all, try to do it while copying the
	 * data.  If the data is truncated, or if we only want a partial
	 * coverage checksum (UDP-Lite), do it before the copy.
436 437
	 */

438
	if (copied < ulen || UDP_SKB_CB(skb)->partial_cov) {
439 440
		checksum_valid = !udp_lib_checksum_complete(skb);
		if (!checksum_valid)
L
Linus Torvalds 已提交
441
			goto csum_copy_err;
442 443
	}

444
	if (checksum_valid || skb_csum_unnecessary(skb))
445 446
		err = skb_copy_datagram_msg(skb, sizeof(struct udphdr),
					    msg, copied);
447
	else {
448
		err = skb_copy_and_csum_datagram_msg(skb, sizeof(struct udphdr), msg);
L
Linus Torvalds 已提交
449 450 451
		if (err == -EINVAL)
			goto csum_copy_err;
	}
452 453
	if (unlikely(err)) {
		trace_kfree_skb(skb, udpv6_recvmsg);
454 455 456 457 458 459 460 461 462 463 464
		if (!peeked) {
			atomic_inc(&sk->sk_drops);
			if (is_udp4)
				UDP_INC_STATS_USER(sock_net(sk),
						   UDP_MIB_INERRORS,
						   is_udplite);
			else
				UDP6_INC_STATS_USER(sock_net(sk),
						    UDP_MIB_INERRORS,
						    is_udplite);
		}
L
Linus Torvalds 已提交
465
		goto out_free;
466
	}
467 468 469 470 471 472 473 474
	if (!peeked) {
		if (is_udp4)
			UDP_INC_STATS_USER(sock_net(sk),
					UDP_MIB_INDATAGRAMS, is_udplite);
		else
			UDP6_INC_STATS_USER(sock_net(sk),
					UDP_MIB_INDATAGRAMS, is_udplite);
	}
475

476
	sock_recv_ts_and_drops(msg, sk, skb);
L
Linus Torvalds 已提交
477 478 479

	/* Copy the address. */
	if (msg->msg_name) {
480
		DECLARE_SOCKADDR(struct sockaddr_in6 *, sin6, msg->msg_name);
L
Linus Torvalds 已提交
481
		sin6->sin6_family = AF_INET6;
482
		sin6->sin6_port = udp_hdr(skb)->source;
L
Linus Torvalds 已提交
483 484
		sin6->sin6_flowinfo = 0;

485
		if (is_udp4) {
B
Brian Haley 已提交
486 487
			ipv6_addr_set_v4mapped(ip_hdr(skb)->saddr,
					       &sin6->sin6_addr);
488 489
			sin6->sin6_scope_id = 0;
		} else {
A
Alexey Dobriyan 已提交
490
			sin6->sin6_addr = ipv6_hdr(skb)->saddr;
491 492
			sin6->sin6_scope_id =
				ipv6_iface_scope_id(&sin6->sin6_addr,
493
						    inet6_iif(skb));
L
Linus Torvalds 已提交
494
		}
495
		*addr_len = sizeof(*sin6);
L
Linus Torvalds 已提交
496
	}
497 498 499 500

	if (np->rxopt.all)
		ip6_datagram_recv_common_ctl(sk, msg, skb);

501
	if (is_udp4) {
L
Linus Torvalds 已提交
502 503 504 505
		if (inet->cmsg_flags)
			ip_cmsg_recv(msg, skb);
	} else {
		if (np->rxopt.all)
506
			ip6_datagram_recv_specific_ctl(sk, msg, skb);
507
	}
L
Linus Torvalds 已提交
508

509
	err = copied;
L
Linus Torvalds 已提交
510
	if (flags & MSG_TRUNC)
511
		err = ulen;
L
Linus Torvalds 已提交
512 513

out_free:
514
	skb_free_datagram_locked(sk, skb);
L
Linus Torvalds 已提交
515 516 517 518
out:
	return err;

csum_copy_err:
519
	slow = lock_sock_fast(sk);
520
	if (!skb_kill_datagram(sk, skb, flags)) {
521 522 523
		if (is_udp4) {
			UDP_INC_STATS_USER(sock_net(sk),
					UDP_MIB_CSUMERRORS, is_udplite);
524 525
			UDP_INC_STATS_USER(sock_net(sk),
					UDP_MIB_INERRORS, is_udplite);
526 527 528
		} else {
			UDP6_INC_STATS_USER(sock_net(sk),
					UDP_MIB_CSUMERRORS, is_udplite);
529 530
			UDP6_INC_STATS_USER(sock_net(sk),
					UDP_MIB_INERRORS, is_udplite);
531
		}
532
	}
533
	unlock_sock_fast(sk, slow);
L
Linus Torvalds 已提交
534

535 536
	/* starting over for a new packet, but check if we need to yield */
	cond_resched();
537
	msg->msg_flags &= ~MSG_TRUNC;
L
Linus Torvalds 已提交
538 539 540
	goto try_again;
}

541
void __udp6_lib_err(struct sk_buff *skb, struct inet6_skb_parm *opt,
542
		    u8 type, u8 code, int offset, __be32 info,
543
		    struct udp_table *udptable)
L
Linus Torvalds 已提交
544 545
{
	struct ipv6_pinfo *np;
546 547 548
	const struct ipv6hdr *hdr = (const struct ipv6hdr *)skb->data;
	const struct in6_addr *saddr = &hdr->saddr;
	const struct in6_addr *daddr = &hdr->daddr;
549
	struct udphdr *uh = (struct udphdr *)(skb->data+offset);
L
Linus Torvalds 已提交
550
	struct sock *sk;
551
	int harderr;
L
Linus Torvalds 已提交
552
	int err;
553
	struct net *net = dev_net(skb->dev);
L
Linus Torvalds 已提交
554

555
	sk = __udp6_lib_lookup(net, daddr, uh->dest, saddr, uh->source,
556
			       inet6_iif(skb), udptable, skb);
557
	if (!sk) {
558 559
		ICMP6_INC_STATS_BH(net, __in6_dev_get(skb->dev),
				   ICMP6_MIB_INERRORS);
L
Linus Torvalds 已提交
560
		return;
561
	}
L
Linus Torvalds 已提交
562

563 564 565
	harderr = icmpv6_err_convert(type, code, &err);
	np = inet6_sk(sk);

566 567 568
	if (type == ICMPV6_PKT_TOOBIG) {
		if (!ip6_sk_accept_pmtu(sk))
			goto out;
569
		ip6_sk_update_pmtu(skb, sk, info);
570 571
		if (np->pmtudisc != IPV6_PMTUDISC_DONT)
			harderr = 1;
572
	}
573
	if (type == NDISC_REDIRECT) {
574
		ip6_sk_redirect(skb, sk);
575 576
		goto out;
	}
577

578 579 580 581
	if (!np->recverr) {
		if (!harderr || sk->sk_state != TCP_ESTABLISHED)
			goto out;
	} else {
L
Linus Torvalds 已提交
582
		ipv6_icmp_error(sk, skb, err, uh->dest, ntohl(info), (u8 *)(uh+1));
583
	}
584

L
Linus Torvalds 已提交
585 586 587 588 589 590
	sk->sk_err = err;
	sk->sk_error_report(sk);
out:
	sock_put(sk);
}

591 592 593 594
static int __udpv6_queue_rcv_skb(struct sock *sk, struct sk_buff *skb)
{
	int rc;

595
	if (!ipv6_addr_any(&sk->sk_v6_daddr)) {
596
		sock_rps_save_rxhash(sk, skb);
597
		sk_mark_napi_id(sk, skb);
E
Eric Dumazet 已提交
598
		sk_incoming_cpu_update(sk);
599
	}
600 601 602 603 604 605 606 607 608 609 610 611 612 613 614 615

	rc = sock_queue_rcv_skb(sk, skb);
	if (rc < 0) {
		int is_udplite = IS_UDPLITE(sk);

		/* Note that an ENOMEM error is charged twice */
		if (rc == -ENOMEM)
			UDP6_INC_STATS_BH(sock_net(sk),
					UDP_MIB_RCVBUFERRORS, is_udplite);
		UDP6_INC_STATS_BH(sock_net(sk), UDP_MIB_INERRORS, is_udplite);
		kfree_skb(skb);
		return -1;
	}
	return 0;
}

616
static __inline__ void udpv6_err(struct sk_buff *skb,
617
				 struct inet6_skb_parm *opt, u8 type,
618
				 u8 code, int offset, __be32 info)
619
{
620
	__udp6_lib_err(skb, opt, type, code, offset, info, &udp_table);
621 622
}

623 624 625 626 627 628 629 630
static struct static_key udpv6_encap_needed __read_mostly;
void udpv6_encap_enable(void)
{
	if (!static_key_enabled(&udpv6_encap_needed))
		static_key_slow_inc(&udpv6_encap_needed);
}
EXPORT_SYMBOL(udpv6_encap_enable);

631
int udpv6_queue_rcv_skb(struct sock *sk, struct sk_buff *skb)
L
Linus Torvalds 已提交
632
{
633
	struct udp_sock *up = udp_sk(sk);
634
	int rc;
W
Wang Chen 已提交
635
	int is_udplite = IS_UDPLITE(sk);
636

637 638
	if (!xfrm6_policy_check(sk, XFRM_POLICY_IN, skb))
		goto drop;
L
Linus Torvalds 已提交
639

640 641 642 643 644 645 646 647 648 649 650 651 652 653 654 655
	if (static_key_false(&udpv6_encap_needed) && up->encap_type) {
		int (*encap_rcv)(struct sock *sk, struct sk_buff *skb);

		/*
		 * This is an encapsulation socket so pass the skb to
		 * the socket's udp_encap_rcv() hook. Otherwise, just
		 * fall through and pass this up the UDP socket.
		 * up->encap_rcv() returns the following value:
		 * =0 if skb was successfully passed to the encap
		 *    handler or was discarded by it.
		 * >0 if skb should be passed on to UDP.
		 * <0 if skb should be resubmitted as proto -N
		 */

		/* if we're overly short, let UDP handle it */
		encap_rcv = ACCESS_ONCE(up->encap_rcv);
656
		if (skb->len > sizeof(struct udphdr) && encap_rcv) {
657 658
			int ret;

659 660 661 662
			/* Verify checksum before giving to encap */
			if (udp_lib_checksum_complete(skb))
				goto csum_error;

663 664 665 666 667 668 669 670 671 672 673 674
			ret = encap_rcv(sk, skb);
			if (ret <= 0) {
				UDP_INC_STATS_BH(sock_net(sk),
						 UDP_MIB_INDATAGRAMS,
						 is_udplite);
				return -ret;
			}
		}

		/* FALLTHROUGH -- it's a UDP Packet */
	}

675 676 677
	/*
	 * UDP-Lite specific tests, ignored on UDP sockets (see net/ipv4/udp.c).
	 */
W
Wang Chen 已提交
678
	if ((is_udplite & UDPLITE_RECV_CC)  &&  UDP_SKB_CB(skb)->partial_cov) {
679 680

		if (up->pcrlen == 0) {          /* full coverage was set  */
681 682
			net_dbg_ratelimited("UDPLITE6: partial coverage %d while full coverage %d requested\n",
					    UDP_SKB_CB(skb)->cscov, skb->len);
683 684 685
			goto drop;
		}
		if (UDP_SKB_CB(skb)->cscov  <  up->pcrlen) {
686 687
			net_dbg_ratelimited("UDPLITE6: coverage %d too small, need min %d\n",
					    UDP_SKB_CB(skb)->cscov, up->pcrlen);
688 689
			goto drop;
		}
L
Linus Torvalds 已提交
690 691
	}

692
	if (rcu_access_pointer(sk->sk_filter)) {
693
		if (udp_lib_checksum_complete(skb))
694
			goto csum_error;
695
	}
696

697
	if (sk_rcvqueues_full(sk, sk->sk_rcvbuf)) {
698 699
		UDP6_INC_STATS_BH(sock_net(sk),
				  UDP_MIB_RCVBUFERRORS, is_udplite);
700
		goto drop;
701
	}
702

703
	skb_dst_drop(skb);
704

705 706 707 708 709 710 711 712 713
	bh_lock_sock(sk);
	rc = 0;
	if (!sock_owned_by_user(sk))
		rc = __udpv6_queue_rcv_skb(sk, skb);
	else if (sk_add_backlog(sk, skb, sk->sk_rcvbuf)) {
		bh_unlock_sock(sk);
		goto drop;
	}
	bh_unlock_sock(sk);
714 715

	return rc;
716

717 718
csum_error:
	UDP6_INC_STATS_BH(sock_net(sk), UDP_MIB_CSUMERRORS, is_udplite);
719
drop:
720
	UDP6_INC_STATS_BH(sock_net(sk), UDP_MIB_INERRORS, is_udplite);
721
	atomic_inc(&sk->sk_drops);
722 723
	kfree_skb(skb);
	return -1;
L
Linus Torvalds 已提交
724 725
}

726 727 728 729
static bool __udp_v6_is_mcast_sock(struct net *net, struct sock *sk,
				   __be16 loc_port, const struct in6_addr *loc_addr,
				   __be16 rmt_port, const struct in6_addr *rmt_addr,
				   int dif, unsigned short hnum)
L
Linus Torvalds 已提交
730
{
731
	struct inet_sock *inet = inet_sk(sk);
L
Linus Torvalds 已提交
732

733 734 735 736 737 738 739 740
	if (!net_eq(sock_net(sk), net))
		return false;

	if (udp_sk(sk)->udp_port_hash != hnum ||
	    sk->sk_family != PF_INET6 ||
	    (inet->inet_dport && inet->inet_dport != rmt_port) ||
	    (!ipv6_addr_any(&sk->sk_v6_daddr) &&
		    !ipv6_addr_equal(&sk->sk_v6_daddr, rmt_addr)) ||
741 742 743
	    (sk->sk_bound_dev_if && sk->sk_bound_dev_if != dif) ||
	    (!ipv6_addr_any(&sk->sk_v6_rcv_saddr) &&
		    !ipv6_addr_equal(&sk->sk_v6_rcv_saddr, loc_addr)))
744 745 746 747
		return false;
	if (!inet6_mc_check(sk, loc_addr, rmt_addr))
		return false;
	return true;
L
Linus Torvalds 已提交
748 749
}

750 751 752
static void flush_stack(struct sock **stack, unsigned int count,
			struct sk_buff *skb, unsigned int final)
{
753
	struct sk_buff *skb1 = NULL;
754
	struct sock *sk;
755
	unsigned int i;
756 757

	for (i = 0; i < count; i++) {
758
		sk = stack[i];
759
		if (likely(!skb1))
760 761 762 763 764 765 766
			skb1 = (i == final) ? skb : skb_clone(skb, GFP_ATOMIC);
		if (!skb1) {
			atomic_inc(&sk->sk_drops);
			UDP6_INC_STATS_BH(sock_net(sk), UDP_MIB_RCVBUFERRORS,
					  IS_UDPLITE(sk));
			UDP6_INC_STATS_BH(sock_net(sk), UDP_MIB_INERRORS,
					  IS_UDPLITE(sk));
767
		}
768 769 770

		if (skb1 && udpv6_queue_rcv_skb(sk, skb1) <= 0)
			skb1 = NULL;
771
		sock_put(sk);
772
	}
773 774
	if (unlikely(skb1))
		kfree_skb(skb1);
775
}
776 777 778 779 780 781

static void udp6_csum_zero_error(struct sk_buff *skb)
{
	/* RFC 2460 section 8.1 says that we SHOULD log
	 * this error. Well, it is reasonable.
	 */
782 783 784
	net_dbg_ratelimited("IPv6: udp checksum is 0 for [%pI6c]:%u->[%pI6c]:%u\n",
			    &ipv6_hdr(skb)->saddr, ntohs(udp_hdr(skb)->source),
			    &ipv6_hdr(skb)->daddr, ntohs(udp_hdr(skb)->dest));
785 786
}

L
Linus Torvalds 已提交
787 788 789 790
/*
 * Note: called only from the BH handler context,
 * so we don't need to lock the hashes.
 */
791
static int __udp6_lib_mcast_deliver(struct net *net, struct sk_buff *skb,
792
		const struct in6_addr *saddr, const struct in6_addr *daddr,
793
		struct udp_table *udptable, int proto)
L
Linus Torvalds 已提交
794
{
795
	struct sock *sk, *stack[256 / sizeof(struct sock *)];
796
	const struct udphdr *uh = udp_hdr(skb);
797 798 799 800
	struct hlist_nulls_node *node;
	unsigned short hnum = ntohs(uh->dest);
	struct udp_hslot *hslot = udp_hashslot(udptable, net, hnum);
	int dif = inet6_iif(skb);
801 802
	unsigned int count = 0, offset = offsetof(typeof(*sk), sk_nulls_node);
	unsigned int hash2 = 0, hash2_any = 0, use_hash2 = (hslot->count > 10);
803
	bool inner_flushed = false;
804 805 806 807 808 809 810 811 812

	if (use_hash2) {
		hash2_any = udp6_portaddr_hash(net, &in6addr_any, hnum) &
			    udp_table.mask;
		hash2 = udp6_portaddr_hash(net, daddr, hnum) & udp_table.mask;
start_lookup:
		hslot = &udp_table.hash2[hash2];
		offset = offsetof(typeof(*sk), __sk_common.skc_portaddr_node);
	}
L
Linus Torvalds 已提交
813

814
	spin_lock(&hslot->lock);
815
	sk_nulls_for_each_entry_offset(sk, node, &hslot->head, offset) {
816 817 818 819 820 821 822 823 824 825
		if (__udp_v6_is_mcast_sock(net, sk,
					   uh->dest, daddr,
					   uh->source, saddr,
					   dif, hnum) &&
		    /* If zero checksum and no_check is not on for
		     * the socket then skip it.
		     */
		    (uh->check || udp_sk(sk)->no_check6_rx)) {
			if (unlikely(count == ARRAY_SIZE(stack))) {
				flush_stack(stack, count, skb, ~0);
826
				inner_flushed = true;
827 828
				count = 0;
			}
829
			stack[count++] = sk;
830
			sock_hold(sk);
H
Hideo Aoki 已提交
831
		}
L
Linus Torvalds 已提交
832
	}
833

834
	spin_unlock(&hslot->lock);
835

836 837 838 839 840 841
	/* Also lookup *:port if we are using hash2 and haven't done so yet. */
	if (use_hash2 && hash2 != hash2_any) {
		hash2 = hash2_any;
		goto start_lookup;
	}

842 843 844
	if (count) {
		flush_stack(stack, count, skb, count - 1);
	} else {
845
		if (!inner_flushed)
846 847
			UDP6_INC_STATS_BH(net, UDP_MIB_IGNOREDMULTI,
					  proto == IPPROTO_UDPLITE);
848
		consume_skb(skb);
849
	}
850
	return 0;
L
Linus Torvalds 已提交
851 852
}

853
int __udp6_lib_rcv(struct sk_buff *skb, struct udp_table *udptable,
854
		   int proto)
L
Linus Torvalds 已提交
855
{
856
	struct net *net = dev_net(skb->dev);
L
Linus Torvalds 已提交
857
	struct sock *sk;
858
	struct udphdr *uh;
859
	const struct in6_addr *saddr, *daddr;
L
Linus Torvalds 已提交
860 861 862
	u32 ulen = 0;

	if (!pskb_may_pull(skb, sizeof(struct udphdr)))
863
		goto discard;
L
Linus Torvalds 已提交
864

865 866
	saddr = &ipv6_hdr(skb)->saddr;
	daddr = &ipv6_hdr(skb)->daddr;
867
	uh = udp_hdr(skb);
L
Linus Torvalds 已提交
868 869

	ulen = ntohs(uh->len);
870 871
	if (ulen > skb->len)
		goto short_packet;
L
Linus Torvalds 已提交
872

873 874
	if (proto == IPPROTO_UDP) {
		/* UDP validates ulen. */
L
Linus Torvalds 已提交
875

876 877 878
		/* Check for jumbo payload */
		if (ulen == 0)
			ulen = skb->len;
L
Linus Torvalds 已提交
879

880 881
		if (ulen < sizeof(*uh))
			goto short_packet;
L
Linus Torvalds 已提交
882

883 884 885
		if (ulen < skb->len) {
			if (pskb_trim_rcsum(skb, ulen))
				goto short_packet;
886 887
			saddr = &ipv6_hdr(skb)->saddr;
			daddr = &ipv6_hdr(skb)->daddr;
888
			uh = udp_hdr(skb);
889 890
		}
	}
L
Linus Torvalds 已提交
891

892
	if (udp6_csum_init(skb, uh, proto))
893
		goto csum_error;
894

895 896
	/*
	 *	Multicast receive code
L
Linus Torvalds 已提交
897
	 */
898
	if (ipv6_addr_is_multicast(daddr))
899
		return __udp6_lib_mcast_deliver(net, skb,
900
				saddr, daddr, udptable, proto);
L
Linus Torvalds 已提交
901 902 903

	/* Unicast */

904
	/*
L
Linus Torvalds 已提交
905 906 907
	 * check socket cache ... must talk to Alan about his plans
	 * for sock caches... i'll skip this for now.
	 */
908
	sk = __udp6_lib_lookup_skb(skb, uh->source, uh->dest, udptable);
909
	if (sk) {
910 911
		int ret;

912
		if (!uh->check && !udp_sk(sk)->no_check6_rx) {
913
			sock_put(sk);
914 915 916 917
			udp6_csum_zero_error(skb);
			goto csum_error;
		}

918
		if (inet_get_convert_csum(sk) && uh->check && !IS_UDPLITE(sk))
919 920 921
			skb_checksum_try_convert(skb, IPPROTO_UDP, uh->check,
						 ip6_compute_pseudo);

922
		ret = udpv6_queue_rcv_skb(sk, skb);
923
		sock_put(sk);
L
Linus Torvalds 已提交
924

925
		/* a return value > 0 means to resubmit the input */
926
		if (ret > 0)
927
			return ret;
L
Linus Torvalds 已提交
928

S
Stephen Hemminger 已提交
929
		return 0;
L
Linus Torvalds 已提交
930
	}
931

932 933 934 935 936
	if (!uh->check) {
		udp6_csum_zero_error(skb);
		goto csum_error;
	}

937
	if (!xfrm6_policy_check(NULL, XFRM_POLICY_IN, skb))
938
		goto discard;
939 940

	if (udp_lib_checksum_complete(skb))
941
		goto csum_error;
942 943 944 945 946

	UDP6_INC_STATS_BH(net, UDP_MIB_NOPORTS, proto == IPPROTO_UDPLITE);
	icmpv6_send(skb, ICMPV6_DEST_UNREACH, ICMPV6_PORT_UNREACH, 0);

	kfree_skb(skb);
S
Stephen Hemminger 已提交
947
	return 0;
L
Linus Torvalds 已提交
948

949
short_packet:
950 951 952 953 954
	net_dbg_ratelimited("UDP%sv6: short packet: From [%pI6c]:%u %d/%d to [%pI6c]:%u\n",
			    proto == IPPROTO_UDPLITE ? "-Lite" : "",
			    saddr, ntohs(uh->source),
			    ulen, skb->len,
			    daddr, ntohs(uh->dest));
955 956 957
	goto discard;
csum_error:
	UDP6_INC_STATS_BH(net, UDP_MIB_CSUMERRORS, proto == IPPROTO_UDPLITE);
L
Linus Torvalds 已提交
958
discard:
959
	UDP6_INC_STATS_BH(net, UDP_MIB_INERRORS, proto == IPPROTO_UDPLITE);
L
Linus Torvalds 已提交
960
	kfree_skb(skb);
S
Stephen Hemminger 已提交
961
	return 0;
L
Linus Torvalds 已提交
962
}
963

964
static __inline__ int udpv6_rcv(struct sk_buff *skb)
965
{
966
	return __udp6_lib_rcv(skb, &udp_table, IPPROTO_UDP);
967 968
}

L
Linus Torvalds 已提交
969 970 971 972 973 974 975
/*
 * Throw away all pending data and cancel the corking. Socket is locked.
 */
static void udp_v6_flush_pending_frames(struct sock *sk)
{
	struct udp_sock *up = udp_sk(sk);

976 977 978
	if (up->pending == AF_INET)
		udp_flush_pending_frames(sk);
	else if (up->pending) {
L
Linus Torvalds 已提交
979 980 981
		up->len = 0;
		up->pending = 0;
		ip6_flush_pending_frames(sk);
982
	}
L
Linus Torvalds 已提交
983 984
}

985
/**
986 987 988 989
 *	udp6_hwcsum_outgoing  -  handle outgoing HW checksumming
 *	@sk:	socket we are sending on
 *	@skb:	sk_buff containing the filled-in UDP header
 *		(checksum field must be zeroed out)
990 991 992 993 994 995 996
 */
static void udp6_hwcsum_outgoing(struct sock *sk, struct sk_buff *skb,
				 const struct in6_addr *saddr,
				 const struct in6_addr *daddr, int len)
{
	unsigned int offset;
	struct udphdr *uh = udp_hdr(skb);
V
Vlad Yasevich 已提交
997
	struct sk_buff *frags = skb_shinfo(skb)->frag_list;
998 999
	__wsum csum = 0;

V
Vlad Yasevich 已提交
1000
	if (!frags) {
1001 1002 1003 1004 1005 1006 1007 1008 1009 1010 1011 1012 1013 1014 1015
		/* Only one fragment on the socket.  */
		skb->csum_start = skb_transport_header(skb) - skb->head;
		skb->csum_offset = offsetof(struct udphdr, check);
		uh->check = ~csum_ipv6_magic(saddr, daddr, len, IPPROTO_UDP, 0);
	} else {
		/*
		 * HW-checksum won't work as there are two or more
		 * fragments on the socket so that all csums of sk_buffs
		 * should be together
		 */
		offset = skb_transport_offset(skb);
		skb->csum = skb_checksum(skb, offset, skb->len - offset, 0);

		skb->ip_summed = CHECKSUM_NONE;

V
Vlad Yasevich 已提交
1016 1017 1018
		do {
			csum = csum_add(csum, frags->csum);
		} while ((frags = frags->next));
1019 1020 1021 1022 1023 1024 1025 1026

		uh->check = csum_ipv6_magic(saddr, daddr, len, IPPROTO_UDP,
					    csum);
		if (uh->check == 0)
			uh->check = CSUM_MANGLED_0;
	}
}

L
Linus Torvalds 已提交
1027 1028 1029 1030
/*
 *	Sending
 */

V
Vlad Yasevich 已提交
1031
static int udp_v6_send_skb(struct sk_buff *skb, struct flowi6 *fl6)
L
Linus Torvalds 已提交
1032
{
V
Vlad Yasevich 已提交
1033
	struct sock *sk = skb->sk;
L
Linus Torvalds 已提交
1034 1035
	struct udphdr *uh;
	int err = 0;
W
Wang Chen 已提交
1036
	int is_udplite = IS_UDPLITE(sk);
1037
	__wsum csum = 0;
V
Vlad Yasevich 已提交
1038 1039
	int offset = skb_transport_offset(skb);
	int len = skb->len - offset;
L
Linus Torvalds 已提交
1040 1041 1042 1043

	/*
	 * Create a UDP header
	 */
1044
	uh = udp_hdr(skb);
1045 1046
	uh->source = fl6->fl6_sport;
	uh->dest = fl6->fl6_dport;
V
Vlad Yasevich 已提交
1047
	uh->len = htons(len);
L
Linus Torvalds 已提交
1048 1049
	uh->check = 0;

W
Wang Chen 已提交
1050
	if (is_udplite)
V
Vlad Yasevich 已提交
1051 1052
		csum = udplite_csum(skb);
	else if (udp_sk(sk)->no_check6_tx) {   /* UDP csum disabled */
1053 1054 1055
		skb->ip_summed = CHECKSUM_NONE;
		goto send;
	} else if (skb->ip_summed == CHECKSUM_PARTIAL) { /* UDP hardware csum */
V
Vlad Yasevich 已提交
1056
		udp6_hwcsum_outgoing(sk, skb, &fl6->saddr, &fl6->daddr, len);
1057 1058
		goto send;
	} else
V
Vlad Yasevich 已提交
1059
		csum = udp_csum(skb);
L
Linus Torvalds 已提交
1060

1061
	/* add protocol-dependent pseudo-header */
1062
	uh->check = csum_ipv6_magic(&fl6->saddr, &fl6->daddr,
V
Vlad Yasevich 已提交
1063
				    len, fl6->flowi6_proto, csum);
L
Linus Torvalds 已提交
1064
	if (uh->check == 0)
1065
		uh->check = CSUM_MANGLED_0;
L
Linus Torvalds 已提交
1066

1067
send:
V
Vlad Yasevich 已提交
1068
	err = ip6_send_skb(skb);
E
Eric Dumazet 已提交
1069 1070 1071 1072 1073 1074 1075 1076 1077
	if (err) {
		if (err == -ENOBUFS && !inet6_sk(sk)->recverr) {
			UDP6_INC_STATS_USER(sock_net(sk),
					    UDP_MIB_SNDBUFERRORS, is_udplite);
			err = 0;
		}
	} else
		UDP6_INC_STATS_USER(sock_net(sk),
				    UDP_MIB_OUTDATAGRAMS, is_udplite);
V
Vlad Yasevich 已提交
1078 1079 1080 1081 1082 1083 1084 1085 1086 1087 1088 1089 1090 1091 1092 1093 1094 1095 1096 1097 1098 1099 1100 1101
	return err;
}

static int udp_v6_push_pending_frames(struct sock *sk)
{
	struct sk_buff *skb;
	struct udp_sock  *up = udp_sk(sk);
	struct flowi6 fl6;
	int err = 0;

	if (up->pending == AF_INET)
		return udp_push_pending_frames(sk);

	/* ip6_finish_skb will release the cork, so make a copy of
	 * fl6 here.
	 */
	fl6 = inet_sk(sk)->cork.fl.u.ip6;

	skb = ip6_finish_skb(sk);
	if (!skb)
		goto out;

	err = udp_v6_send_skb(skb, &fl6);

L
Linus Torvalds 已提交
1102 1103 1104 1105 1106 1107
out:
	up->len = 0;
	up->pending = 0;
	return err;
}

1108
int udpv6_sendmsg(struct sock *sk, struct msghdr *msg, size_t len)
L
Linus Torvalds 已提交
1109 1110 1111 1112 1113
{
	struct ipv6_txoptions opt_space;
	struct udp_sock *up = udp_sk(sk);
	struct inet_sock *inet = inet_sk(sk);
	struct ipv6_pinfo *np = inet6_sk(sk);
1114
	DECLARE_SOCKADDR(struct sockaddr_in6 *, sin6, msg->msg_name);
1115
	struct in6_addr *daddr, *final_p, final;
L
Linus Torvalds 已提交
1116
	struct ipv6_txoptions *opt = NULL;
1117
	struct ipv6_txoptions *opt_to_free = NULL;
L
Linus Torvalds 已提交
1118
	struct ip6_flowlabel *flowlabel = NULL;
1119
	struct flowi6 fl6;
L
Linus Torvalds 已提交
1120 1121 1122 1123
	struct dst_entry *dst;
	int addr_len = msg->msg_namelen;
	int ulen = len;
	int hlimit = -1;
1124
	int tclass = -1;
1125
	int dontfrag = -1;
L
Linus Torvalds 已提交
1126 1127
	int corkreq = up->corkflag || msg->msg_flags&MSG_MORE;
	int err;
1128
	int connected = 0;
W
Wang Chen 已提交
1129
	int is_udplite = IS_UDPLITE(sk);
1130
	int (*getfrag)(void *, char *, int, int, int, struct sk_buff *);
L
Linus Torvalds 已提交
1131 1132 1133 1134 1135 1136 1137 1138 1139 1140 1141 1142 1143 1144 1145 1146 1147 1148 1149 1150 1151 1152 1153 1154 1155

	/* destination address check */
	if (sin6) {
		if (addr_len < offsetof(struct sockaddr, sa_data))
			return -EINVAL;

		switch (sin6->sin6_family) {
		case AF_INET6:
			if (addr_len < SIN6_LEN_RFC2133)
				return -EINVAL;
			daddr = &sin6->sin6_addr;
			break;
		case AF_INET:
			goto do_udp_sendmsg;
		case AF_UNSPEC:
			msg->msg_name = sin6 = NULL;
			msg->msg_namelen = addr_len = 0;
			daddr = NULL;
			break;
		default:
			return -EINVAL;
		}
	} else if (!up->pending) {
		if (sk->sk_state != TCP_ESTABLISHED)
			return -EDESTADDRREQ;
1156
		daddr = &sk->sk_v6_daddr;
1157
	} else
L
Linus Torvalds 已提交
1158 1159 1160
		daddr = NULL;

	if (daddr) {
B
Brian Haley 已提交
1161
		if (ipv6_addr_v4mapped(daddr)) {
L
Linus Torvalds 已提交
1162 1163
			struct sockaddr_in sin;
			sin.sin_family = AF_INET;
E
Eric Dumazet 已提交
1164
			sin.sin_port = sin6 ? sin6->sin6_port : inet->inet_dport;
L
Linus Torvalds 已提交
1165 1166 1167 1168 1169 1170
			sin.sin_addr.s_addr = daddr->s6_addr32[3];
			msg->msg_name = &sin;
			msg->msg_namelen = sizeof(sin);
do_udp_sendmsg:
			if (__ipv6_only_sock(sk))
				return -ENETUNREACH;
1171
			return udp_sendmsg(sk, msg, len);
L
Linus Torvalds 已提交
1172 1173 1174 1175
		}
	}

	if (up->pending == AF_INET)
1176
		return udp_sendmsg(sk, msg, len);
L
Linus Torvalds 已提交
1177 1178

	/* Rough check on arithmetic overflow,
1179
	   better check is made in ip6_append_data().
L
Linus Torvalds 已提交
1180 1181 1182
	   */
	if (len > INT_MAX - sizeof(struct udphdr))
		return -EMSGSIZE;
1183

1184
	getfrag  =  is_udplite ?  udplite_getfrag : ip_generic_getfrag;
L
Linus Torvalds 已提交
1185 1186 1187 1188 1189 1190 1191 1192 1193 1194 1195 1196 1197 1198 1199 1200 1201 1202
	if (up->pending) {
		/*
		 * There are pending frames.
		 * The socket lock must be held while it's corked.
		 */
		lock_sock(sk);
		if (likely(up->pending)) {
			if (unlikely(up->pending != AF_INET6)) {
				release_sock(sk);
				return -EAFNOSUPPORT;
			}
			dst = NULL;
			goto do_append_data;
		}
		release_sock(sk);
	}
	ulen += sizeof(struct udphdr);

1203
	memset(&fl6, 0, sizeof(fl6));
L
Linus Torvalds 已提交
1204 1205 1206 1207 1208

	if (sin6) {
		if (sin6->sin6_port == 0)
			return -EINVAL;

1209
		fl6.fl6_dport = sin6->sin6_port;
L
Linus Torvalds 已提交
1210 1211 1212
		daddr = &sin6->sin6_addr;

		if (np->sndflow) {
1213 1214 1215
			fl6.flowlabel = sin6->sin6_flowinfo&IPV6_FLOWINFO_MASK;
			if (fl6.flowlabel&IPV6_FLOWLABEL_MASK) {
				flowlabel = fl6_sock_lookup(sk, fl6.flowlabel);
1216
				if (!flowlabel)
L
Linus Torvalds 已提交
1217 1218 1219 1220 1221 1222 1223 1224 1225
					return -EINVAL;
			}
		}

		/*
		 * Otherwise it will be difficult to maintain
		 * sk->sk_dst_cache.
		 */
		if (sk->sk_state == TCP_ESTABLISHED &&
1226 1227
		    ipv6_addr_equal(daddr, &sk->sk_v6_daddr))
			daddr = &sk->sk_v6_daddr;
L
Linus Torvalds 已提交
1228 1229 1230

		if (addr_len >= sizeof(struct sockaddr_in6) &&
		    sin6->sin6_scope_id &&
1231
		    __ipv6_addr_needs_scope_id(__ipv6_addr_type(daddr)))
1232
			fl6.flowi6_oif = sin6->sin6_scope_id;
L
Linus Torvalds 已提交
1233 1234 1235 1236
	} else {
		if (sk->sk_state != TCP_ESTABLISHED)
			return -EDESTADDRREQ;

1237
		fl6.fl6_dport = inet->inet_dport;
1238
		daddr = &sk->sk_v6_daddr;
1239
		fl6.flowlabel = np->flow_label;
1240
		connected = 1;
L
Linus Torvalds 已提交
1241 1242
	}

1243 1244
	if (!fl6.flowi6_oif)
		fl6.flowi6_oif = sk->sk_bound_dev_if;
L
Linus Torvalds 已提交
1245

1246 1247
	if (!fl6.flowi6_oif)
		fl6.flowi6_oif = np->sticky_pktinfo.ipi6_ifindex;
1248

1249
	fl6.flowi6_mark = sk->sk_mark;
B
Brian Haley 已提交
1250

L
Linus Torvalds 已提交
1251 1252 1253 1254 1255
	if (msg->msg_controllen) {
		opt = &opt_space;
		memset(opt, 0, sizeof(struct ipv6_txoptions));
		opt->tot_len = sizeof(*opt);

1256 1257
		err = ip6_datagram_send_ctl(sock_net(sk), sk, msg, &fl6, opt,
					    &hlimit, &tclass, &dontfrag);
L
Linus Torvalds 已提交
1258 1259 1260 1261
		if (err < 0) {
			fl6_sock_release(flowlabel);
			return err;
		}
1262 1263
		if ((fl6.flowlabel&IPV6_FLOWLABEL_MASK) && !flowlabel) {
			flowlabel = fl6_sock_lookup(sk, fl6.flowlabel);
1264
			if (!flowlabel)
L
Linus Torvalds 已提交
1265 1266 1267 1268
				return -EINVAL;
		}
		if (!(opt->opt_nflen|opt->opt_flen))
			opt = NULL;
1269
		connected = 0;
L
Linus Torvalds 已提交
1270
	}
1271 1272 1273 1274
	if (!opt) {
		opt = txopt_get(np);
		opt_to_free = opt;
	}
1275 1276 1277
	if (flowlabel)
		opt = fl6_merge_options(&opt_space, flowlabel, opt);
	opt = ipv6_fixup_options(&opt_space, opt);
L
Linus Torvalds 已提交
1278

1279
	fl6.flowi6_proto = sk->sk_protocol;
1280
	if (!ipv6_addr_any(daddr))
A
Alexey Dobriyan 已提交
1281
		fl6.daddr = *daddr;
1282
	else
1283 1284
		fl6.daddr.s6_addr[15] = 0x1; /* :: means loopback (BSD'ism) */
	if (ipv6_addr_any(&fl6.saddr) && !ipv6_addr_any(&np->saddr))
A
Alexey Dobriyan 已提交
1285
		fl6.saddr = np->saddr;
1286
	fl6.fl6_sport = inet->inet_sport;
1287

1288
	final_p = fl6_update_dst(&fl6, opt, &final);
1289
	if (final_p)
1290
		connected = 0;
L
Linus Torvalds 已提交
1291

1292 1293
	if (!fl6.flowi6_oif && ipv6_addr_is_multicast(&fl6.daddr)) {
		fl6.flowi6_oif = np->mcast_oif;
1294
		connected = 0;
1295 1296
	} else if (!fl6.flowi6_oif)
		fl6.flowi6_oif = np->ucast_oif;
L
Linus Torvalds 已提交
1297

1298
	security_sk_classify_flow(sk, flowi6_to_flowi(&fl6));
V
Venkat Yekkirala 已提交
1299

1300
	dst = ip6_sk_dst_lookup_flow(sk, &fl6, final_p);
1301 1302 1303
	if (IS_ERR(dst)) {
		err = PTR_ERR(dst);
		dst = NULL;
L
Linus Torvalds 已提交
1304
		goto out;
1305
	}
L
Linus Torvalds 已提交
1306

1307 1308
	if (hlimit < 0)
		hlimit = ip6_sk_dst_hoplimit(np, &fl6, dst);
L
Linus Torvalds 已提交
1309

G
Gerrit Renker 已提交
1310
	if (tclass < 0)
1311 1312
		tclass = np->tclass;

L
Linus Torvalds 已提交
1313 1314 1315 1316
	if (msg->msg_flags&MSG_CONFIRM)
		goto do_confirm;
back_from_confirm:

1317 1318 1319 1320 1321 1322 1323 1324 1325 1326 1327 1328 1329 1330
	/* Lockless fast path for the non-corking case */
	if (!corkreq) {
		struct sk_buff *skb;

		skb = ip6_make_skb(sk, getfrag, msg, ulen,
				   sizeof(struct udphdr), hlimit, tclass, opt,
				   &fl6, (struct rt6_info *)dst,
				   msg->msg_flags, dontfrag);
		err = PTR_ERR(skb);
		if (!IS_ERR_OR_NULL(skb))
			err = udp_v6_send_skb(skb, &fl6);
		goto release_dst;
	}

L
Linus Torvalds 已提交
1331 1332 1333 1334 1335 1336
	lock_sock(sk);
	if (unlikely(up->pending)) {
		/* The socket is already corked while preparing it. */
		/* ... which is an evident application bug. --ANK */
		release_sock(sk);

1337
		net_dbg_ratelimited("udp cork app bug 2\n");
L
Linus Torvalds 已提交
1338 1339 1340 1341 1342 1343 1344
		err = -EINVAL;
		goto out;
	}

	up->pending = AF_INET6;

do_append_data:
1345 1346
	if (dontfrag < 0)
		dontfrag = np->dontfrag;
L
Linus Torvalds 已提交
1347
	up->len += ulen;
1348
	err = ip6_append_data(sk, getfrag, msg, ulen,
1349
		sizeof(struct udphdr), hlimit, tclass, opt, &fl6,
1350
		(struct rt6_info *)dst,
1351
		corkreq ? msg->msg_flags|MSG_MORE : msg->msg_flags, dontfrag);
L
Linus Torvalds 已提交
1352 1353 1354
	if (err)
		udp_v6_flush_pending_frames(sk);
	else if (!corkreq)
1355
		err = udp_v6_push_pending_frames(sk);
H
Herbert Xu 已提交
1356 1357
	else if (unlikely(skb_queue_empty(&sk->sk_write_queue)))
		up->pending = 0;
L
Linus Torvalds 已提交
1358

1359 1360 1361 1362 1363
	if (err > 0)
		err = np->recverr ? net_xmit_errno(err) : 0;
	release_sock(sk);

release_dst:
1364 1365 1366
	if (dst) {
		if (connected) {
			ip6_dst_store(sk, dst,
1367 1368
				      ipv6_addr_equal(&fl6.daddr, &sk->sk_v6_daddr) ?
				      &sk->sk_v6_daddr : NULL,
1369
#ifdef CONFIG_IPV6_SUBTREES
1370
				      ipv6_addr_equal(&fl6.saddr, &np->saddr) ?
1371 1372 1373
				      &np->saddr :
#endif
				      NULL);
1374 1375 1376
		} else {
			dst_release(dst);
		}
1377
		dst = NULL;
1378 1379
	}

L
Linus Torvalds 已提交
1380
out:
1381
	dst_release(dst);
L
Linus Torvalds 已提交
1382
	fl6_sock_release(flowlabel);
1383
	txopt_put(opt_to_free);
1384
	if (!err)
L
Linus Torvalds 已提交
1385
		return len;
1386 1387 1388 1389 1390 1391 1392 1393
	/*
	 * ENOBUFS = no kernel mem, SOCK_NOSPACE = no sndbuf space.  Reporting
	 * ENOBUFS might not be good (it's not tunable per se), but otherwise
	 * we don't have a good statistic (IpOutDiscards but it can be too many
	 * things).  We could add another new stat but at least for now that
	 * seems like overkill.
	 */
	if (err == -ENOBUFS || test_bit(SOCK_NOSPACE, &sk->sk_socket->flags)) {
1394 1395
		UDP6_INC_STATS_USER(sock_net(sk),
				UDP_MIB_SNDBUFERRORS, is_udplite);
1396
	}
L
Linus Torvalds 已提交
1397 1398 1399 1400 1401 1402 1403 1404 1405 1406
	return err;

do_confirm:
	dst_confirm(dst);
	if (!(msg->msg_flags&MSG_PROBE) || len)
		goto back_from_confirm;
	err = 0;
	goto out;
}

1407
void udpv6_destroy_sock(struct sock *sk)
L
Linus Torvalds 已提交
1408
{
T
Tom Parkin 已提交
1409
	struct udp_sock *up = udp_sk(sk);
L
Linus Torvalds 已提交
1410 1411 1412 1413
	lock_sock(sk);
	udp_v6_flush_pending_frames(sk);
	release_sock(sk);

T
Tom Parkin 已提交
1414 1415 1416 1417 1418 1419 1420
	if (static_key_false(&udpv6_encap_needed) && up->encap_type) {
		void (*encap_destroy)(struct sock *sk);
		encap_destroy = ACCESS_ONCE(up->encap_destroy);
		if (encap_destroy)
			encap_destroy(sk);
	}

L
Linus Torvalds 已提交
1421 1422 1423 1424 1425 1426
	inet6_destroy_sock(sk);
}

/*
 *	Socket option code for UDP
 */
1427
int udpv6_setsockopt(struct sock *sk, int level, int optname,
1428
		     char __user *optval, unsigned int optlen)
1429
{
1430
	if (level == SOL_UDP  ||  level == SOL_UDPLITE)
1431 1432
		return udp_lib_setsockopt(sk, level, optname, optval, optlen,
					  udp_v6_push_pending_frames);
1433
	return ipv6_setsockopt(sk, level, optname, optval, optlen);
1434 1435 1436
}

#ifdef CONFIG_COMPAT
1437
int compat_udpv6_setsockopt(struct sock *sk, int level, int optname,
1438
			    char __user *optval, unsigned int optlen)
1439
{
1440
	if (level == SOL_UDP  ||  level == SOL_UDPLITE)
1441 1442
		return udp_lib_setsockopt(sk, level, optname, optval, optlen,
					  udp_v6_push_pending_frames);
1443
	return compat_ipv6_setsockopt(sk, level, optname, optval, optlen);
1444 1445 1446
}
#endif

1447 1448
int udpv6_getsockopt(struct sock *sk, int level, int optname,
		     char __user *optval, int __user *optlen)
1449
{
1450
	if (level == SOL_UDP  ||  level == SOL_UDPLITE)
1451
		return udp_lib_getsockopt(sk, level, optname, optval, optlen);
1452
	return ipv6_getsockopt(sk, level, optname, optval, optlen);
1453 1454 1455
}

#ifdef CONFIG_COMPAT
1456 1457
int compat_udpv6_getsockopt(struct sock *sk, int level, int optname,
			    char __user *optval, int __user *optlen)
1458
{
1459
	if (level == SOL_UDP  ||  level == SOL_UDPLITE)
1460
		return udp_lib_getsockopt(sk, level, optname, optval, optlen);
1461
	return compat_ipv6_getsockopt(sk, level, optname, optval, optlen);
1462 1463 1464
}
#endif

1465
static const struct inet6_protocol udpv6_protocol = {
L
Linus Torvalds 已提交
1466 1467 1468 1469 1470 1471 1472
	.handler	=	udpv6_rcv,
	.err_handler	=	udpv6_err,
	.flags		=	INET6_PROTO_NOPOLICY|INET6_PROTO_FINAL,
};

/* ------------------------------------------------------------------------ */
#ifdef CONFIG_PROC_FS
1473
int udp6_seq_show(struct seq_file *seq, void *v)
L
Linus Torvalds 已提交
1474
{
1475 1476 1477 1478 1479 1480 1481 1482 1483
	if (v == SEQ_START_TOKEN) {
		seq_puts(seq, IPV6_SEQ_DGRAM_HEADER);
	} else {
		int bucket = ((struct udp_iter_state *)seq->private)->bucket;
		struct inet_sock *inet = inet_sk(v);
		__u16 srcp = ntohs(inet->inet_sport);
		__u16 destp = ntohs(inet->inet_dport);
		ip6_dgram_sock_seq_show(seq, v, srcp, destp, bucket);
	}
L
Linus Torvalds 已提交
1484 1485 1486
	return 0;
}

1487 1488 1489 1490 1491 1492 1493 1494
static const struct file_operations udp6_afinfo_seq_fops = {
	.owner    = THIS_MODULE,
	.open     = udp_seq_open,
	.read     = seq_read,
	.llseek   = seq_lseek,
	.release  = seq_release_net
};

L
Linus Torvalds 已提交
1495 1496 1497
static struct udp_seq_afinfo udp6_seq_afinfo = {
	.name		= "udp6",
	.family		= AF_INET6,
1498
	.udp_table	= &udp_table,
1499
	.seq_fops	= &udp6_afinfo_seq_fops,
1500 1501 1502
	.seq_ops	= {
		.show		= udp6_seq_show,
	},
L
Linus Torvalds 已提交
1503 1504
};

1505
int __net_init udp6_proc_init(struct net *net)
L
Linus Torvalds 已提交
1506
{
1507
	return udp_proc_register(net, &udp6_seq_afinfo);
L
Linus Torvalds 已提交
1508 1509
}

I
Ian Morris 已提交
1510 1511
void udp6_proc_exit(struct net *net)
{
1512
	udp_proc_unregister(net, &udp6_seq_afinfo);
L
Linus Torvalds 已提交
1513 1514 1515
}
#endif /* CONFIG_PROC_FS */

E
Eric Dumazet 已提交
1516 1517 1518 1519 1520 1521 1522 1523 1524 1525 1526
void udp_v6_clear_sk(struct sock *sk, int size)
{
	struct inet_sock *inet = inet_sk(sk);

	/* we do not want to clear pinet6 field, because of RCU lookups */
	sk_prot_clear_portaddr_nulls(sk, offsetof(struct inet_sock, pinet6));

	size -= offsetof(struct inet_sock, pinet6) + sizeof(inet->pinet6);
	memset(&inet->pinet6 + 1, 0, size);
}

L
Linus Torvalds 已提交
1527 1528 1529
/* ------------------------------------------------------------------------ */

struct proto udpv6_prot = {
1530 1531
	.name		   = "UDPv6",
	.owner		   = THIS_MODULE,
1532
	.close		   = udp_lib_close,
1533 1534 1535 1536 1537 1538 1539 1540
	.connect	   = ip6_datagram_connect,
	.disconnect	   = udp_disconnect,
	.ioctl		   = udp_ioctl,
	.destroy	   = udpv6_destroy_sock,
	.setsockopt	   = udpv6_setsockopt,
	.getsockopt	   = udpv6_getsockopt,
	.sendmsg	   = udpv6_sendmsg,
	.recvmsg	   = udpv6_recvmsg,
1541
	.backlog_rcv	   = __udpv6_queue_rcv_skb,
1542 1543
	.hash		   = udp_lib_hash,
	.unhash		   = udp_lib_unhash,
E
Eric Dumazet 已提交
1544
	.rehash		   = udp_v6_rehash,
1545
	.get_port	   = udp_v6_get_port,
H
Hideo Aoki 已提交
1546 1547 1548 1549
	.memory_allocated  = &udp_memory_allocated,
	.sysctl_mem	   = sysctl_udp_mem,
	.sysctl_wmem	   = &sysctl_udp_wmem_min,
	.sysctl_rmem	   = &sysctl_udp_rmem_min,
1550
	.obj_size	   = sizeof(struct udp6_sock),
1551
	.slab_flags	   = SLAB_DESTROY_BY_RCU,
1552
	.h.udp_table	   = &udp_table,
1553
#ifdef CONFIG_COMPAT
1554 1555
	.compat_setsockopt = compat_udpv6_setsockopt,
	.compat_getsockopt = compat_udpv6_getsockopt,
1556
#endif
E
Eric Dumazet 已提交
1557
	.clear_sk	   = udp_v6_clear_sk,
L
Linus Torvalds 已提交
1558 1559 1560 1561 1562 1563 1564 1565 1566 1567
};

static struct inet_protosw udpv6_protosw = {
	.type =      SOCK_DGRAM,
	.protocol =  IPPROTO_UDP,
	.prot =      &udpv6_prot,
	.ops =       &inet6_dgram_ops,
	.flags =     INET_PROTOSW_PERMANENT,
};

1568
int __init udpv6_init(void)
L
Linus Torvalds 已提交
1569
{
1570 1571
	int ret;

1572 1573
	ret = inet6_add_protocol(&udpv6_protocol, IPPROTO_UDP);
	if (ret)
1574
		goto out;
1575

1576 1577 1578 1579 1580 1581 1582 1583 1584 1585 1586
	ret = inet6_register_protosw(&udpv6_protosw);
	if (ret)
		goto out_udpv6_protocol;
out:
	return ret;

out_udpv6_protocol:
	inet6_del_protocol(&udpv6_protocol, IPPROTO_UDP);
	goto out;
}

1587
void udpv6_exit(void)
1588 1589 1590
{
	inet6_unregister_protosw(&udpv6_protosw);
	inet6_del_protocol(&udpv6_protocol, IPPROTO_UDP);
L
Linus Torvalds 已提交
1591
}