udp.c 38.6 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 40 41 42 43 44
#include <asm/uaccess.h>

#include <net/ndisc.h>
#include <net/protocol.h>
#include <net/transp_v6.h>
#include <net/ip6_route.h>
#include <net/raw.h>
45
#include <net/tcp_states.h>
L
Linus Torvalds 已提交
46 47
#include <net/ip6_checksum.h>
#include <net/xfrm.h>
48
#include <net/inet6_hashtables.h>
49
#include <net/busy_poll.h>
L
Linus Torvalds 已提交
50 51 52

#include <linux/proc_fs.h>
#include <linux/seq_file.h>
53
#include <trace/events/skb.h>
54
#include "udp_impl.h"
L
Linus Torvalds 已提交
55

56 57 58 59 60 61
static unsigned int udp6_ehashfn(struct net *net,
				  const struct in6_addr *laddr,
				  const u16 lport,
				  const struct in6_addr *faddr,
				  const __be16 fport)
{
62 63 64 65 66 67 68 69 70 71 72 73 74
	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);

75
	return __inet6_ehashfn(lhash, lport, fhash, fport,
76
			       udp_ipv6_hash_secret + net_hash_mix(net));
77 78
}

79 80 81 82
int ipv6_rcv_saddr_equal(const struct sock *sk, const struct sock *sk2)
{
	const struct in6_addr *sk2_rcv_saddr6 = inet6_rcv_saddr(sk2);
	int sk2_ipv6only = inet_v6_ipv6only(sk2);
83
	int addr_type = ipv6_addr_type(&sk->sk_v6_rcv_saddr);
84 85 86 87
	int addr_type2 = sk2_rcv_saddr6 ? ipv6_addr_type(sk2_rcv_saddr6) : IPV6_ADDR_MAPPED;

	/* if both are mapped, treat as IPv4 */
	if (addr_type == IPV6_ADDR_MAPPED && addr_type2 == IPV6_ADDR_MAPPED)
88
		return (!sk2_ipv6only &&
E
Eric Dumazet 已提交
89 90
			(!sk->sk_rcv_saddr || !sk2->sk_rcv_saddr ||
			  sk->sk_rcv_saddr == sk2->sk_rcv_saddr));
91 92 93 94 95 96

	if (addr_type2 == IPV6_ADDR_ANY &&
	    !(sk2_ipv6only && addr_type == IPV6_ADDR_MAPPED))
		return 1;

	if (addr_type == IPV6_ADDR_ANY &&
97
	    !(ipv6_only_sock(sk) && addr_type2 == IPV6_ADDR_MAPPED))
98 99 100
		return 1;

	if (sk2_rcv_saddr6 &&
101
	    ipv6_addr_equal(&sk->sk_v6_rcv_saddr, sk2_rcv_saddr6))
102 103 104 105 106
		return 1;

	return 0;
}

107 108 109 110 111 112 113 114
static unsigned int udp6_portaddr_hash(struct net *net,
				       const struct in6_addr *addr6,
				       unsigned int port)
{
	unsigned int hash, mix = net_hash_mix(net);

	if (ipv6_addr_any(addr6))
		hash = jhash_1word(0, mix);
B
Brian Haley 已提交
115
	else if (ipv6_addr_v4mapped(addr6))
116
		hash = jhash_1word((__force u32)addr6->s6_addr32[3], mix);
117
	else
118
		hash = jhash2((__force u32 *)addr6->s6_addr32, 4, mix);
119 120 121 122 123

	return hash ^ port;
}


124
int udp_v6_get_port(struct sock *sk, unsigned short snum)
L
Linus Torvalds 已提交
125
{
E
Eric Dumazet 已提交
126 127
	unsigned int hash2_nulladdr =
		udp6_portaddr_hash(sock_net(sk), &in6addr_any, snum);
128
	unsigned int hash2_partial =
129
		udp6_portaddr_hash(sock_net(sk), &sk->sk_v6_rcv_saddr, 0);
E
Eric Dumazet 已提交
130

131
	/* precompute partial secondary hash */
E
Eric Dumazet 已提交
132 133
	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
134 135
}

E
Eric Dumazet 已提交
136 137 138
static void udp_v6_rehash(struct sock *sk)
{
	u16 new_hash = udp6_portaddr_hash(sock_net(sk),
139
					  &sk->sk_v6_rcv_saddr,
E
Eric Dumazet 已提交
140 141 142 143 144
					  inet_sk(sk)->inet_num);

	udp_lib_rehash(sk, new_hash);
}

145 146
static inline int compute_score(struct sock *sk, struct net *net,
				unsigned short hnum,
147 148
				const struct in6_addr *saddr, __be16 sport,
				const struct in6_addr *daddr, __be16 dport,
149 150
				int dif)
{
151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183
	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++;
184
	}
185

186 187 188
	return score;
}

E
Eric Dumazet 已提交
189 190
#define SCORE2_MAX (1 + 1 + 1)
static inline int compute_score2(struct sock *sk, struct net *net,
191 192 193
				 const struct in6_addr *saddr, __be16 sport,
				 const struct in6_addr *daddr,
				 unsigned short hnum, int dif)
E
Eric Dumazet 已提交
194
{
195 196
	int score;
	struct inet_sock *inet;
E
Eric Dumazet 已提交
197

198 199 200 201
	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
202

203 204 205 206 207 208 209 210 211 212 213 214 215 216
	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
217
			return -1;
218
		score++;
E
Eric Dumazet 已提交
219
	}
220 221 222 223 224 225 226

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

E
Eric Dumazet 已提交
227 228 229 230 231 232 233 234 235 236 237
	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,
		struct udp_hslot *hslot2, unsigned int slot2)
{
	struct sock *sk, *result;
	struct hlist_nulls_node *node;
238 239
	int score, badness, matches = 0, reuseport = 0;
	u32 hash = 0;
E
Eric Dumazet 已提交
240 241 242 243 244 245 246 247 248 249

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;
250 251
			reuseport = sk->sk_reuseport;
			if (reuseport) {
252 253
				hash = udp6_ehashfn(net, daddr, hnum,
						    saddr, sport);
254 255
				matches = 1;
			} else if (score == SCORE2_MAX)
E
Eric Dumazet 已提交
256
				goto exact_match;
257 258
		} else if (score == badness && reuseport) {
			matches++;
259
			if (reciprocal_scale(hash, matches) == 0)
260 261
				result = sk;
			hash = next_pseudo_random32(hash);
E
Eric Dumazet 已提交
262 263 264 265 266 267 268 269 270 271 272 273
		}
	}
	/*
	 * 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) {
exact_match:
E
Eric Dumazet 已提交
274
		if (unlikely(!atomic_inc_not_zero_hint(&result->sk_refcnt, 2)))
E
Eric Dumazet 已提交
275 276 277 278 279 280 281 282 283 284
			result = NULL;
		else if (unlikely(compute_score2(result, net, saddr, sport,
				  daddr, hnum, dif) < badness)) {
			sock_put(result);
			goto begin;
		}
	}
	return result;
}

285
struct sock *__udp6_lib_lookup(struct net *net,
286 287
				      const struct in6_addr *saddr, __be16 sport,
				      const struct in6_addr *daddr, __be16 dport,
288
				      int dif, struct udp_table *udptable)
L
Linus Torvalds 已提交
289
{
290
	struct sock *sk, *result;
291
	struct hlist_nulls_node *node;
L
Linus Torvalds 已提交
292
	unsigned short hnum = ntohs(dport);
E
Eric Dumazet 已提交
293 294
	unsigned int hash2, slot2, slot = udp_hashfn(net, hnum, udptable->mask);
	struct udp_hslot *hslot2, *hslot = &udptable->hash[slot];
295 296
	int score, badness, matches = 0, reuseport = 0;
	u32 hash = 0;
297

298
	rcu_read_lock();
E
Eric Dumazet 已提交
299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315
	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,
					  hslot2, slot2);
		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;

316 317
			result = udp6_lib_lookup2(net, saddr, sport,
						  &in6addr_any, hnum, dif,
E
Eric Dumazet 已提交
318 319 320 321 322
						  hslot2, slot2);
		}
		rcu_read_unlock();
		return result;
	}
323 324 325
begin:
	result = NULL;
	badness = -1;
326
	sk_nulls_for_each_rcu(sk, node, &hslot->head) {
327 328 329 330
		score = compute_score(sk, net, hnum, saddr, sport, daddr, dport, dif);
		if (score > badness) {
			result = sk;
			badness = score;
331 332
			reuseport = sk->sk_reuseport;
			if (reuseport) {
333 334
				hash = udp6_ehashfn(net, daddr, hnum,
						    saddr, sport);
335 336 337 338
				matches = 1;
			}
		} else if (score == badness && reuseport) {
			matches++;
339
			if (reciprocal_scale(hash, matches) == 0)
340 341
				result = sk;
			hash = next_pseudo_random32(hash);
L
Linus Torvalds 已提交
342 343
		}
	}
344 345 346 347 348
	/*
	 * 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
349
	if (get_nulls_value(node) != slot)
350 351
		goto begin;

352
	if (result) {
E
Eric Dumazet 已提交
353
		if (unlikely(!atomic_inc_not_zero_hint(&result->sk_refcnt, 2)))
354 355 356 357 358 359 360 361
			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
362 363
	return result;
}
364
EXPORT_SYMBOL_GPL(__udp6_lib_lookup);
L
Linus Torvalds 已提交
365

366 367
static struct sock *__udp6_lib_lookup_skb(struct sk_buff *skb,
					  __be16 sport, __be16 dport,
368
					  struct udp_table *udptable)
369
{
370
	struct sock *sk;
371
	const struct ipv6hdr *iph = ipv6_hdr(skb);
372

373 374
	sk = skb_steal_sock(skb);
	if (unlikely(sk))
375
		return sk;
E
Eric Dumazet 已提交
376 377 378
	return __udp6_lib_lookup(dev_net(skb_dst(skb)->dev), &iph->saddr, sport,
				 &iph->daddr, dport, inet6_iif(skb),
				 udptable);
379 380
}

381 382 383 384 385 386 387 388
struct sock *udp6_lib_lookup(struct net *net, const struct in6_addr *saddr, __be16 sport,
			     const struct in6_addr *daddr, __be16 dport, int dif)
{
	return __udp6_lib_lookup(net, saddr, sport, daddr, dport, dif, &udp_table);
}
EXPORT_SYMBOL_GPL(udp6_lib_lookup);


L
Linus Torvalds 已提交
389
/*
390 391
 *	This should be easy, if there is something there we
 *	return it, otherwise we block.
L
Linus Torvalds 已提交
392 393
 */

394
int udpv6_recvmsg(struct kiocb *iocb, struct sock *sk,
L
Linus Torvalds 已提交
395 396 397 398 399
		  struct msghdr *msg, size_t len,
		  int noblock, int flags, int *addr_len)
{
	struct ipv6_pinfo *np = inet6_sk(sk);
	struct inet_sock *inet = inet_sk(sk);
400
	struct sk_buff *skb;
401
	unsigned int ulen, copied;
402
	int peeked, off = 0;
403 404
	int err;
	int is_udplite = IS_UDPLITE(sk);
405
	int is_udp4;
406
	bool slow;
L
Linus Torvalds 已提交
407 408

	if (flags & MSG_ERRQUEUE)
409
		return ipv6_recv_error(sk, msg, len, addr_len);
L
Linus Torvalds 已提交
410

411
	if (np->rxpmtu && np->rxopt.bits.rxpmtu)
412
		return ipv6_recv_rxpmtu(sk, msg, len, addr_len);
413

L
Linus Torvalds 已提交
414
try_again:
415
	skb = __skb_recv_datagram(sk, flags | (noblock ? MSG_DONTWAIT : 0),
416
				  &peeked, &off, &err);
L
Linus Torvalds 已提交
417 418 419
	if (!skb)
		goto out;

420
	ulen = skb->len - sizeof(struct udphdr);
421 422 423 424
	copied = len;
	if (copied > ulen)
		copied = ulen;
	else if (copied < ulen)
425
		msg->msg_flags |= MSG_TRUNC;
L
Linus Torvalds 已提交
426

427 428
	is_udp4 = (skb->protocol == htons(ETH_P_IP));

429
	/*
430 431 432
	 * 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.
433 434
	 */

435
	if (copied < ulen || UDP_SKB_CB(skb)->partial_cov) {
436
		if (udp_lib_checksum_complete(skb))
L
Linus Torvalds 已提交
437
			goto csum_copy_err;
438 439
	}

440
	if (skb_csum_unnecessary(skb))
441 442
		err = skb_copy_datagram_msg(skb, sizeof(struct udphdr),
					    msg, copied);
443
	else {
444
		err = skb_copy_and_csum_datagram_msg(skb, sizeof(struct udphdr), msg);
L
Linus Torvalds 已提交
445 446 447
		if (err == -EINVAL)
			goto csum_copy_err;
	}
448 449
	if (unlikely(err)) {
		trace_kfree_skb(skb, udpv6_recvmsg);
450 451 452 453 454 455 456 457 458 459 460
		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
461
		goto out_free;
462
	}
463 464 465 466 467 468 469 470
	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);
	}
471

472
	sock_recv_ts_and_drops(msg, sk, skb);
L
Linus Torvalds 已提交
473 474 475

	/* Copy the address. */
	if (msg->msg_name) {
476
		DECLARE_SOCKADDR(struct sockaddr_in6 *, sin6, msg->msg_name);
L
Linus Torvalds 已提交
477
		sin6->sin6_family = AF_INET6;
478
		sin6->sin6_port = udp_hdr(skb)->source;
L
Linus Torvalds 已提交
479 480
		sin6->sin6_flowinfo = 0;

481
		if (is_udp4) {
B
Brian Haley 已提交
482 483
			ipv6_addr_set_v4mapped(ip_hdr(skb)->saddr,
					       &sin6->sin6_addr);
484 485
			sin6->sin6_scope_id = 0;
		} else {
A
Alexey Dobriyan 已提交
486
			sin6->sin6_addr = ipv6_hdr(skb)->saddr;
487 488
			sin6->sin6_scope_id =
				ipv6_iface_scope_id(&sin6->sin6_addr,
489
						    inet6_iif(skb));
L
Linus Torvalds 已提交
490
		}
491
		*addr_len = sizeof(*sin6);
L
Linus Torvalds 已提交
492
	}
493 494 495 496

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

497
	if (is_udp4) {
L
Linus Torvalds 已提交
498 499 500 501
		if (inet->cmsg_flags)
			ip_cmsg_recv(msg, skb);
	} else {
		if (np->rxopt.all)
502
			ip6_datagram_recv_specific_ctl(sk, msg, skb);
503
	}
L
Linus Torvalds 已提交
504

505
	err = copied;
L
Linus Torvalds 已提交
506
	if (flags & MSG_TRUNC)
507
		err = ulen;
L
Linus Torvalds 已提交
508 509

out_free:
510
	skb_free_datagram_locked(sk, skb);
L
Linus Torvalds 已提交
511 512 513 514
out:
	return err;

csum_copy_err:
515
	slow = lock_sock_fast(sk);
516
	if (!skb_kill_datagram(sk, skb, flags)) {
517 518 519
		if (is_udp4) {
			UDP_INC_STATS_USER(sock_net(sk),
					UDP_MIB_CSUMERRORS, is_udplite);
520 521
			UDP_INC_STATS_USER(sock_net(sk),
					UDP_MIB_INERRORS, is_udplite);
522 523 524
		} else {
			UDP6_INC_STATS_USER(sock_net(sk),
					UDP_MIB_CSUMERRORS, is_udplite);
525 526
			UDP6_INC_STATS_USER(sock_net(sk),
					UDP_MIB_INERRORS, is_udplite);
527
		}
528
	}
529
	unlock_sock_fast(sk, slow);
L
Linus Torvalds 已提交
530

531
	if (noblock)
L
Linus Torvalds 已提交
532
		return -EAGAIN;
533 534 535

	/* starting over for a new packet */
	msg->msg_flags &= ~MSG_TRUNC;
L
Linus Torvalds 已提交
536 537 538
	goto try_again;
}

539
void __udp6_lib_err(struct sk_buff *skb, struct inet6_skb_parm *opt,
540
		    u8 type, u8 code, int offset, __be32 info,
541
		    struct udp_table *udptable)
L
Linus Torvalds 已提交
542 543
{
	struct ipv6_pinfo *np;
544 545 546
	const struct ipv6hdr *hdr = (const struct ipv6hdr *)skb->data;
	const struct in6_addr *saddr = &hdr->saddr;
	const struct in6_addr *daddr = &hdr->daddr;
547
	struct udphdr *uh = (struct udphdr *)(skb->data+offset);
L
Linus Torvalds 已提交
548 549
	struct sock *sk;
	int err;
550
	struct net *net = dev_net(skb->dev);
L
Linus Torvalds 已提交
551

552
	sk = __udp6_lib_lookup(net, daddr, uh->dest,
553
			       saddr, uh->source, inet6_iif(skb), udptable);
554 555 556
	if (sk == NULL) {
		ICMP6_INC_STATS_BH(net, __in6_dev_get(skb->dev),
				   ICMP6_MIB_INERRORS);
L
Linus Torvalds 已提交
557
		return;
558
	}
L
Linus Torvalds 已提交
559

560 561 562
	if (type == ICMPV6_PKT_TOOBIG) {
		if (!ip6_sk_accept_pmtu(sk))
			goto out;
563
		ip6_sk_update_pmtu(skb, sk, info);
564
	}
565
	if (type == NDISC_REDIRECT) {
566
		ip6_sk_redirect(skb, sk);
567 568
		goto out;
	}
569

L
Linus Torvalds 已提交
570 571 572 573 574 575 576 577
	np = inet6_sk(sk);

	if (!icmpv6_err_convert(type, code, &err) && !np->recverr)
		goto out;

	if (sk->sk_state != TCP_ESTABLISHED && !np->recverr)
		goto out;

578
	if (np->recverr)
L
Linus Torvalds 已提交
579
		ipv6_icmp_error(sk, skb, err, uh->dest, ntohl(info), (u8 *)(uh+1));
580

L
Linus Torvalds 已提交
581 582 583 584 585 586
	sk->sk_err = err;
	sk->sk_error_report(sk);
out:
	sock_put(sk);
}

587 588 589 590
static int __udpv6_queue_rcv_skb(struct sock *sk, struct sk_buff *skb)
{
	int rc;

591
	if (!ipv6_addr_any(&sk->sk_v6_daddr)) {
592
		sock_rps_save_rxhash(sk, skb);
593
		sk_mark_napi_id(sk, skb);
E
Eric Dumazet 已提交
594
		sk_incoming_cpu_update(sk);
595
	}
596 597 598 599 600 601 602 603 604 605 606 607 608 609 610 611

	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;
}

612
static __inline__ void udpv6_err(struct sk_buff *skb,
613
				 struct inet6_skb_parm *opt, u8 type,
614
				 u8 code, int offset, __be32 info)
615
{
616
	__udp6_lib_err(skb, opt, type, code, offset, info, &udp_table);
617 618
}

619 620 621 622 623 624 625 626
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);

627
int udpv6_queue_rcv_skb(struct sock *sk, struct sk_buff *skb)
L
Linus Torvalds 已提交
628
{
629
	struct udp_sock *up = udp_sk(sk);
630
	int rc;
W
Wang Chen 已提交
631
	int is_udplite = IS_UDPLITE(sk);
632

633 634
	if (!xfrm6_policy_check(sk, XFRM_POLICY_IN, skb))
		goto drop;
L
Linus Torvalds 已提交
635

636 637 638 639 640 641 642 643 644 645 646 647 648 649 650 651 652 653 654
	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);
		if (skb->len > sizeof(struct udphdr) && encap_rcv != NULL) {
			int ret;

655 656 657 658
			/* Verify checksum before giving to encap */
			if (udp_lib_checksum_complete(skb))
				goto csum_error;

659 660 661 662 663 664 665 666 667 668 669 670
			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 */
	}

671 672 673
	/*
	 * UDP-Lite specific tests, ignored on UDP sockets (see net/ipv4/udp.c).
	 */
W
Wang Chen 已提交
674
	if ((is_udplite & UDPLITE_RECV_CC)  &&  UDP_SKB_CB(skb)->partial_cov) {
675 676

		if (up->pcrlen == 0) {          /* full coverage was set  */
677 678
			net_dbg_ratelimited("UDPLITE6: partial coverage %d while full coverage %d requested\n",
					    UDP_SKB_CB(skb)->cscov, skb->len);
679 680 681
			goto drop;
		}
		if (UDP_SKB_CB(skb)->cscov  <  up->pcrlen) {
682 683
			net_dbg_ratelimited("UDPLITE6: coverage %d too small, need min %d\n",
					    UDP_SKB_CB(skb)->cscov, up->pcrlen);
684 685
			goto drop;
		}
L
Linus Torvalds 已提交
686 687
	}

688
	if (rcu_access_pointer(sk->sk_filter)) {
689
		if (udp_lib_checksum_complete(skb))
690
			goto csum_error;
691
	}
692

693
	if (sk_rcvqueues_full(sk, sk->sk_rcvbuf)) {
694 695
		UDP6_INC_STATS_BH(sock_net(sk),
				  UDP_MIB_RCVBUFERRORS, is_udplite);
696
		goto drop;
697
	}
698

699
	skb_dst_drop(skb);
700

701 702 703 704 705 706 707 708 709
	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);
710 711

	return rc;
712

713 714
csum_error:
	UDP6_INC_STATS_BH(sock_net(sk), UDP_MIB_CSUMERRORS, is_udplite);
715
drop:
716
	UDP6_INC_STATS_BH(sock_net(sk), UDP_MIB_INERRORS, is_udplite);
717
	atomic_inc(&sk->sk_drops);
718 719
	kfree_skb(skb);
	return -1;
L
Linus Torvalds 已提交
720 721
}

722 723 724 725
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
726
{
727
	struct inet_sock *inet = inet_sk(sk);
L
Linus Torvalds 已提交
728

729 730 731 732 733 734 735 736 737 738 739 740 741
	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)) ||
	    (sk->sk_bound_dev_if && sk->sk_bound_dev_if != dif))
		return false;
	if (!inet6_mc_check(sk, loc_addr, rmt_addr))
		return false;
	return true;
L
Linus Torvalds 已提交
742 743
}

744 745 746
static void flush_stack(struct sock **stack, unsigned int count,
			struct sk_buff *skb, unsigned int final)
{
747
	struct sk_buff *skb1 = NULL;
748
	struct sock *sk;
749
	unsigned int i;
750 751

	for (i = 0; i < count; i++) {
752
		sk = stack[i];
753 754 755 756 757 758 759 760
		if (likely(skb1 == NULL))
			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));
761
		}
762 763 764

		if (skb1 && udpv6_queue_rcv_skb(sk, skb1) <= 0)
			skb1 = NULL;
765
		sock_put(sk);
766
	}
767 768
	if (unlikely(skb1))
		kfree_skb(skb1);
769
}
770 771 772 773 774 775

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.
	 */
776 777 778
	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));
779 780
}

L
Linus Torvalds 已提交
781 782 783 784
/*
 * Note: called only from the BH handler context,
 * so we don't need to lock the hashes.
 */
785
static int __udp6_lib_mcast_deliver(struct net *net, struct sk_buff *skb,
786
		const struct in6_addr *saddr, const struct in6_addr *daddr,
787
		struct udp_table *udptable, int proto)
L
Linus Torvalds 已提交
788
{
789
	struct sock *sk, *stack[256 / sizeof(struct sock *)];
790
	const struct udphdr *uh = udp_hdr(skb);
791 792 793 794
	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);
795 796
	unsigned int count = 0, offset = offsetof(typeof(*sk), sk_nulls_node);
	unsigned int hash2 = 0, hash2_any = 0, use_hash2 = (hslot->count > 10);
797
	bool inner_flushed = false;
798 799 800 801 802 803 804 805 806

	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
807

808
	spin_lock(&hslot->lock);
809
	sk_nulls_for_each_entry_offset(sk, node, &hslot->head, offset) {
810 811 812 813 814 815 816 817 818 819
		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);
820
				inner_flushed = true;
821 822
				count = 0;
			}
823
			stack[count++] = sk;
824
			sock_hold(sk);
H
Hideo Aoki 已提交
825
		}
L
Linus Torvalds 已提交
826
	}
827

828
	spin_unlock(&hslot->lock);
829

830 831 832 833 834 835
	/* 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;
	}

836 837 838
	if (count) {
		flush_stack(stack, count, skb, count - 1);
	} else {
839 840 841 842
		if (!inner_flushed)
			UDP_INC_STATS_BH(net, UDP_MIB_IGNOREDMULTI,
					 proto == IPPROTO_UDPLITE);
		consume_skb(skb);
843
	}
844
	return 0;
L
Linus Torvalds 已提交
845 846
}

847
int __udp6_lib_rcv(struct sk_buff *skb, struct udp_table *udptable,
848
		   int proto)
L
Linus Torvalds 已提交
849
{
850
	struct net *net = dev_net(skb->dev);
L
Linus Torvalds 已提交
851
	struct sock *sk;
852
	struct udphdr *uh;
853
	const struct in6_addr *saddr, *daddr;
L
Linus Torvalds 已提交
854 855 856
	u32 ulen = 0;

	if (!pskb_may_pull(skb, sizeof(struct udphdr)))
857
		goto discard;
L
Linus Torvalds 已提交
858

859 860
	saddr = &ipv6_hdr(skb)->saddr;
	daddr = &ipv6_hdr(skb)->daddr;
861
	uh = udp_hdr(skb);
L
Linus Torvalds 已提交
862 863

	ulen = ntohs(uh->len);
864 865
	if (ulen > skb->len)
		goto short_packet;
L
Linus Torvalds 已提交
866

867 868
	if (proto == IPPROTO_UDP) {
		/* UDP validates ulen. */
L
Linus Torvalds 已提交
869

870 871 872
		/* Check for jumbo payload */
		if (ulen == 0)
			ulen = skb->len;
L
Linus Torvalds 已提交
873

874 875
		if (ulen < sizeof(*uh))
			goto short_packet;
L
Linus Torvalds 已提交
876

877 878 879
		if (ulen < skb->len) {
			if (pskb_trim_rcsum(skb, ulen))
				goto short_packet;
880 881
			saddr = &ipv6_hdr(skb)->saddr;
			daddr = &ipv6_hdr(skb)->daddr;
882
			uh = udp_hdr(skb);
883 884
		}
	}
L
Linus Torvalds 已提交
885

886
	if (udp6_csum_init(skb, uh, proto))
887
		goto csum_error;
888

889 890
	/*
	 *	Multicast receive code
L
Linus Torvalds 已提交
891
	 */
892
	if (ipv6_addr_is_multicast(daddr))
893
		return __udp6_lib_mcast_deliver(net, skb,
894
				saddr, daddr, udptable, proto);
L
Linus Torvalds 已提交
895 896 897

	/* Unicast */

898
	/*
L
Linus Torvalds 已提交
899 900 901
	 * check socket cache ... must talk to Alan about his plans
	 * for sock caches... i'll skip this for now.
	 */
902
	sk = __udp6_lib_lookup_skb(skb, uh->source, uh->dest, udptable);
903
	if (sk != NULL) {
904 905
		int ret;

906
		if (!uh->check && !udp_sk(sk)->no_check6_rx) {
907
			sock_put(sk);
908 909 910 911
			udp6_csum_zero_error(skb);
			goto csum_error;
		}

912 913 914 915
		if (udp_sk(sk)->convert_csum && uh->check && !IS_UDPLITE(sk))
			skb_checksum_try_convert(skb, IPPROTO_UDP, uh->check,
						 ip6_compute_pseudo);

916
		ret = udpv6_queue_rcv_skb(sk, skb);
917
		sock_put(sk);
L
Linus Torvalds 已提交
918

919 920 921 922 923
		/* a return value > 0 means to resubmit the input, but
		 * it wants the return to be -protocol, or 0
		 */
		if (ret > 0)
			return -ret;
L
Linus Torvalds 已提交
924

S
Stephen Hemminger 已提交
925
		return 0;
L
Linus Torvalds 已提交
926
	}
927

928 929 930 931 932
	if (!uh->check) {
		udp6_csum_zero_error(skb);
		goto csum_error;
	}

933
	if (!xfrm6_policy_check(NULL, XFRM_POLICY_IN, skb))
934
		goto discard;
935 936

	if (udp_lib_checksum_complete(skb))
937
		goto csum_error;
938 939 940 941 942

	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
943
	return 0;
L
Linus Torvalds 已提交
944

945
short_packet:
946 947 948 949 950
	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));
951 952 953
	goto discard;
csum_error:
	UDP6_INC_STATS_BH(net, UDP_MIB_CSUMERRORS, proto == IPPROTO_UDPLITE);
L
Linus Torvalds 已提交
954
discard:
955
	UDP6_INC_STATS_BH(net, UDP_MIB_INERRORS, proto == IPPROTO_UDPLITE);
L
Linus Torvalds 已提交
956
	kfree_skb(skb);
S
Stephen Hemminger 已提交
957
	return 0;
L
Linus Torvalds 已提交
958
}
959

960
static __inline__ int udpv6_rcv(struct sk_buff *skb)
961
{
962
	return __udp6_lib_rcv(skb, &udp_table, IPPROTO_UDP);
963 964
}

L
Linus Torvalds 已提交
965 966 967 968 969 970 971
/*
 * 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);

972 973 974
	if (up->pending == AF_INET)
		udp_flush_pending_frames(sk);
	else if (up->pending) {
L
Linus Torvalds 已提交
975 976 977
		up->len = 0;
		up->pending = 0;
		ip6_flush_pending_frames(sk);
978
	}
L
Linus Torvalds 已提交
979 980
}

981
/**
982 983 984 985
 *	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)
986 987 988 989 990 991 992 993 994 995 996 997 998 999 1000 1001 1002 1003 1004 1005 1006 1007 1008 1009 1010 1011 1012 1013 1014 1015 1016 1017 1018 1019 1020 1021
 */
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);
	__wsum csum = 0;

	if (skb_queue_len(&sk->sk_write_queue) == 1) {
		/* 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;

		skb_queue_walk(&sk->sk_write_queue, skb) {
			csum = csum_add(csum, skb->csum);
		}

		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
1022 1023 1024 1025
/*
 *	Sending
 */

1026
static int udp_v6_push_pending_frames(struct sock *sk)
L
Linus Torvalds 已提交
1027 1028 1029
{
	struct sk_buff *skb;
	struct udphdr *uh;
1030
	struct udp_sock  *up = udp_sk(sk);
L
Linus Torvalds 已提交
1031
	struct inet_sock *inet = inet_sk(sk);
1032
	struct flowi6 *fl6;
L
Linus Torvalds 已提交
1033
	int err = 0;
W
Wang Chen 已提交
1034
	int is_udplite = IS_UDPLITE(sk);
1035
	__wsum csum = 0;
L
Linus Torvalds 已提交
1036

1037 1038 1039 1040 1041
	if (up->pending == AF_INET)
		return udp_push_pending_frames(sk);

	fl6 = &inet->cork.fl.u.ip6;

L
Linus Torvalds 已提交
1042
	/* Grab the skbuff where UDP header space exists. */
1043 1044
	skb = skb_peek(&sk->sk_write_queue);
	if (skb == NULL)
L
Linus Torvalds 已提交
1045 1046 1047 1048 1049
		goto out;

	/*
	 * Create a UDP header
	 */
1050
	uh = udp_hdr(skb);
1051 1052
	uh->source = fl6->fl6_sport;
	uh->dest = fl6->fl6_dport;
L
Linus Torvalds 已提交
1053 1054 1055
	uh->len = htons(up->len);
	uh->check = 0;

W
Wang Chen 已提交
1056
	if (is_udplite)
1057
		csum = udplite_csum_outgoing(sk, skb);
1058
	else if (up->no_check6_tx) {   /* UDP csum disabled */
1059 1060 1061
		skb->ip_summed = CHECKSUM_NONE;
		goto send;
	} else if (skb->ip_summed == CHECKSUM_PARTIAL) { /* UDP hardware csum */
1062
		udp6_hwcsum_outgoing(sk, skb, &fl6->saddr, &fl6->daddr,
1063 1064 1065
				     up->len);
		goto send;
	} else
1066
		csum = udp_csum_outgoing(sk, skb);
L
Linus Torvalds 已提交
1067

1068
	/* add protocol-dependent pseudo-header */
1069 1070
	uh->check = csum_ipv6_magic(&fl6->saddr, &fl6->daddr,
				    up->len, fl6->flowi6_proto, csum);
L
Linus Torvalds 已提交
1071
	if (uh->check == 0)
1072
		uh->check = CSUM_MANGLED_0;
L
Linus Torvalds 已提交
1073

1074
send:
L
Linus Torvalds 已提交
1075
	err = ip6_push_pending_frames(sk);
E
Eric Dumazet 已提交
1076 1077 1078 1079 1080 1081 1082 1083 1084
	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);
L
Linus Torvalds 已提交
1085 1086 1087 1088 1089 1090
out:
	up->len = 0;
	up->pending = 0;
	return err;
}

1091
int udpv6_sendmsg(struct kiocb *iocb, struct sock *sk,
L
Linus Torvalds 已提交
1092 1093 1094 1095 1096 1097
		  struct msghdr *msg, size_t len)
{
	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);
1098
	DECLARE_SOCKADDR(struct sockaddr_in6 *, sin6, msg->msg_name);
1099
	struct in6_addr *daddr, *final_p, final;
L
Linus Torvalds 已提交
1100 1101
	struct ipv6_txoptions *opt = NULL;
	struct ip6_flowlabel *flowlabel = NULL;
1102
	struct flowi6 fl6;
L
Linus Torvalds 已提交
1103 1104 1105 1106
	struct dst_entry *dst;
	int addr_len = msg->msg_namelen;
	int ulen = len;
	int hlimit = -1;
1107
	int tclass = -1;
1108
	int dontfrag = -1;
L
Linus Torvalds 已提交
1109 1110
	int corkreq = up->corkflag || msg->msg_flags&MSG_MORE;
	int err;
1111
	int connected = 0;
W
Wang Chen 已提交
1112
	int is_udplite = IS_UDPLITE(sk);
1113
	int (*getfrag)(void *, char *, int, int, int, struct sk_buff *);
L
Linus Torvalds 已提交
1114 1115 1116 1117 1118 1119 1120 1121 1122 1123 1124 1125 1126 1127 1128 1129 1130 1131 1132 1133 1134 1135 1136 1137 1138

	/* 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;
1139
		daddr = &sk->sk_v6_daddr;
1140
	} else
L
Linus Torvalds 已提交
1141 1142 1143
		daddr = NULL;

	if (daddr) {
B
Brian Haley 已提交
1144
		if (ipv6_addr_v4mapped(daddr)) {
L
Linus Torvalds 已提交
1145 1146
			struct sockaddr_in sin;
			sin.sin_family = AF_INET;
E
Eric Dumazet 已提交
1147
			sin.sin_port = sin6 ? sin6->sin6_port : inet->inet_dport;
L
Linus Torvalds 已提交
1148 1149 1150 1151 1152 1153 1154 1155 1156 1157 1158 1159 1160 1161
			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;
			return udp_sendmsg(iocb, sk, msg, len);
		}
	}

	if (up->pending == AF_INET)
		return udp_sendmsg(iocb, sk, msg, len);

	/* Rough check on arithmetic overflow,
1162
	   better check is made in ip6_append_data().
L
Linus Torvalds 已提交
1163 1164 1165
	   */
	if (len > INT_MAX - sizeof(struct udphdr))
		return -EMSGSIZE;
1166

L
Linus Torvalds 已提交
1167 1168 1169 1170 1171 1172 1173 1174 1175 1176 1177 1178 1179 1180 1181 1182 1183 1184
	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);

1185
	memset(&fl6, 0, sizeof(fl6));
L
Linus Torvalds 已提交
1186 1187 1188 1189 1190

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

1191
		fl6.fl6_dport = sin6->sin6_port;
L
Linus Torvalds 已提交
1192 1193 1194
		daddr = &sin6->sin6_addr;

		if (np->sndflow) {
1195 1196 1197
			fl6.flowlabel = sin6->sin6_flowinfo&IPV6_FLOWINFO_MASK;
			if (fl6.flowlabel&IPV6_FLOWLABEL_MASK) {
				flowlabel = fl6_sock_lookup(sk, fl6.flowlabel);
L
Linus Torvalds 已提交
1198 1199 1200 1201 1202 1203 1204 1205 1206 1207
				if (flowlabel == NULL)
					return -EINVAL;
			}
		}

		/*
		 * Otherwise it will be difficult to maintain
		 * sk->sk_dst_cache.
		 */
		if (sk->sk_state == TCP_ESTABLISHED &&
1208 1209
		    ipv6_addr_equal(daddr, &sk->sk_v6_daddr))
			daddr = &sk->sk_v6_daddr;
L
Linus Torvalds 已提交
1210 1211 1212

		if (addr_len >= sizeof(struct sockaddr_in6) &&
		    sin6->sin6_scope_id &&
1213
		    __ipv6_addr_needs_scope_id(__ipv6_addr_type(daddr)))
1214
			fl6.flowi6_oif = sin6->sin6_scope_id;
L
Linus Torvalds 已提交
1215 1216 1217 1218
	} else {
		if (sk->sk_state != TCP_ESTABLISHED)
			return -EDESTADDRREQ;

1219
		fl6.fl6_dport = inet->inet_dport;
1220
		daddr = &sk->sk_v6_daddr;
1221
		fl6.flowlabel = np->flow_label;
1222
		connected = 1;
L
Linus Torvalds 已提交
1223 1224
	}

1225 1226
	if (!fl6.flowi6_oif)
		fl6.flowi6_oif = sk->sk_bound_dev_if;
L
Linus Torvalds 已提交
1227

1228 1229
	if (!fl6.flowi6_oif)
		fl6.flowi6_oif = np->sticky_pktinfo.ipi6_ifindex;
1230

1231
	fl6.flowi6_mark = sk->sk_mark;
B
Brian Haley 已提交
1232

L
Linus Torvalds 已提交
1233 1234 1235 1236 1237
	if (msg->msg_controllen) {
		opt = &opt_space;
		memset(opt, 0, sizeof(struct ipv6_txoptions));
		opt->tot_len = sizeof(*opt);

1238 1239
		err = ip6_datagram_send_ctl(sock_net(sk), sk, msg, &fl6, opt,
					    &hlimit, &tclass, &dontfrag);
L
Linus Torvalds 已提交
1240 1241 1242 1243
		if (err < 0) {
			fl6_sock_release(flowlabel);
			return err;
		}
1244 1245
		if ((fl6.flowlabel&IPV6_FLOWLABEL_MASK) && !flowlabel) {
			flowlabel = fl6_sock_lookup(sk, fl6.flowlabel);
L
Linus Torvalds 已提交
1246 1247 1248 1249 1250
			if (flowlabel == NULL)
				return -EINVAL;
		}
		if (!(opt->opt_nflen|opt->opt_flen))
			opt = NULL;
1251
		connected = 0;
L
Linus Torvalds 已提交
1252 1253 1254
	}
	if (opt == NULL)
		opt = np->opt;
1255 1256 1257
	if (flowlabel)
		opt = fl6_merge_options(&opt_space, flowlabel, opt);
	opt = ipv6_fixup_options(&opt_space, opt);
L
Linus Torvalds 已提交
1258

1259
	fl6.flowi6_proto = sk->sk_protocol;
1260
	if (!ipv6_addr_any(daddr))
A
Alexey Dobriyan 已提交
1261
		fl6.daddr = *daddr;
1262
	else
1263 1264
		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
1265
		fl6.saddr = np->saddr;
1266
	fl6.fl6_sport = inet->inet_sport;
1267

1268
	final_p = fl6_update_dst(&fl6, opt, &final);
1269
	if (final_p)
1270
		connected = 0;
L
Linus Torvalds 已提交
1271

1272 1273
	if (!fl6.flowi6_oif && ipv6_addr_is_multicast(&fl6.daddr)) {
		fl6.flowi6_oif = np->mcast_oif;
1274
		connected = 0;
1275 1276
	} else if (!fl6.flowi6_oif)
		fl6.flowi6_oif = np->ucast_oif;
L
Linus Torvalds 已提交
1277

1278
	security_sk_classify_flow(sk, flowi6_to_flowi(&fl6));
V
Venkat Yekkirala 已提交
1279

1280
	dst = ip6_sk_dst_lookup_flow(sk, &fl6, final_p);
1281 1282 1283
	if (IS_ERR(dst)) {
		err = PTR_ERR(dst);
		dst = NULL;
L
Linus Torvalds 已提交
1284
		goto out;
1285
	}
L
Linus Torvalds 已提交
1286

1287 1288
	if (hlimit < 0)
		hlimit = ip6_sk_dst_hoplimit(np, &fl6, dst);
L
Linus Torvalds 已提交
1289

G
Gerrit Renker 已提交
1290
	if (tclass < 0)
1291 1292
		tclass = np->tclass;

L
Linus Torvalds 已提交
1293 1294 1295 1296 1297 1298 1299 1300 1301 1302
	if (msg->msg_flags&MSG_CONFIRM)
		goto do_confirm;
back_from_confirm:

	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);

1303
		net_dbg_ratelimited("udp cork app bug 2\n");
L
Linus Torvalds 已提交
1304 1305 1306 1307 1308 1309 1310
		err = -EINVAL;
		goto out;
	}

	up->pending = AF_INET6;

do_append_data:
1311 1312
	if (dontfrag < 0)
		dontfrag = np->dontfrag;
L
Linus Torvalds 已提交
1313
	up->len += ulen;
1314 1315
	getfrag  =  is_udplite ?  udplite_getfrag : ip_generic_getfrag;
	err = ip6_append_data(sk, getfrag, msg->msg_iov, ulen,
1316
		sizeof(struct udphdr), hlimit, tclass, opt, &fl6,
1317
		(struct rt6_info *)dst,
1318
		corkreq ? msg->msg_flags|MSG_MORE : msg->msg_flags, dontfrag);
L
Linus Torvalds 已提交
1319 1320 1321
	if (err)
		udp_v6_flush_pending_frames(sk);
	else if (!corkreq)
1322
		err = udp_v6_push_pending_frames(sk);
H
Herbert Xu 已提交
1323 1324
	else if (unlikely(skb_queue_empty(&sk->sk_write_queue)))
		up->pending = 0;
L
Linus Torvalds 已提交
1325

1326 1327 1328
	if (dst) {
		if (connected) {
			ip6_dst_store(sk, dst,
1329 1330
				      ipv6_addr_equal(&fl6.daddr, &sk->sk_v6_daddr) ?
				      &sk->sk_v6_daddr : NULL,
1331
#ifdef CONFIG_IPV6_SUBTREES
1332
				      ipv6_addr_equal(&fl6.saddr, &np->saddr) ?
1333 1334 1335
				      &np->saddr :
#endif
				      NULL);
1336 1337 1338
		} else {
			dst_release(dst);
		}
1339
		dst = NULL;
1340 1341
	}

L
Linus Torvalds 已提交
1342 1343 1344 1345
	if (err > 0)
		err = np->recverr ? net_xmit_errno(err) : 0;
	release_sock(sk);
out:
1346
	dst_release(dst);
L
Linus Torvalds 已提交
1347
	fl6_sock_release(flowlabel);
1348
	if (!err)
L
Linus Torvalds 已提交
1349
		return len;
1350 1351 1352 1353 1354 1355 1356 1357
	/*
	 * 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)) {
1358 1359
		UDP6_INC_STATS_USER(sock_net(sk),
				UDP_MIB_SNDBUFERRORS, is_udplite);
1360
	}
L
Linus Torvalds 已提交
1361 1362 1363 1364 1365 1366 1367 1368 1369 1370
	return err;

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

1371
void udpv6_destroy_sock(struct sock *sk)
L
Linus Torvalds 已提交
1372
{
T
Tom Parkin 已提交
1373
	struct udp_sock *up = udp_sk(sk);
L
Linus Torvalds 已提交
1374 1375 1376 1377
	lock_sock(sk);
	udp_v6_flush_pending_frames(sk);
	release_sock(sk);

T
Tom Parkin 已提交
1378 1379 1380 1381 1382 1383 1384
	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
1385 1386 1387 1388 1389 1390
	inet6_destroy_sock(sk);
}

/*
 *	Socket option code for UDP
 */
1391
int udpv6_setsockopt(struct sock *sk, int level, int optname,
1392
		     char __user *optval, unsigned int optlen)
1393
{
1394
	if (level == SOL_UDP  ||  level == SOL_UDPLITE)
1395 1396
		return udp_lib_setsockopt(sk, level, optname, optval, optlen,
					  udp_v6_push_pending_frames);
1397
	return ipv6_setsockopt(sk, level, optname, optval, optlen);
1398 1399 1400
}

#ifdef CONFIG_COMPAT
1401
int compat_udpv6_setsockopt(struct sock *sk, int level, int optname,
1402
			    char __user *optval, unsigned int optlen)
1403
{
1404
	if (level == SOL_UDP  ||  level == SOL_UDPLITE)
1405 1406
		return udp_lib_setsockopt(sk, level, optname, optval, optlen,
					  udp_v6_push_pending_frames);
1407
	return compat_ipv6_setsockopt(sk, level, optname, optval, optlen);
1408 1409 1410
}
#endif

1411 1412
int udpv6_getsockopt(struct sock *sk, int level, int optname,
		     char __user *optval, int __user *optlen)
1413
{
1414
	if (level == SOL_UDP  ||  level == SOL_UDPLITE)
1415
		return udp_lib_getsockopt(sk, level, optname, optval, optlen);
1416
	return ipv6_getsockopt(sk, level, optname, optval, optlen);
1417 1418 1419
}

#ifdef CONFIG_COMPAT
1420 1421
int compat_udpv6_getsockopt(struct sock *sk, int level, int optname,
			    char __user *optval, int __user *optlen)
1422
{
1423
	if (level == SOL_UDP  ||  level == SOL_UDPLITE)
1424
		return udp_lib_getsockopt(sk, level, optname, optval, optlen);
1425
	return compat_ipv6_getsockopt(sk, level, optname, optval, optlen);
1426 1427 1428
}
#endif

1429
static const struct inet6_protocol udpv6_protocol = {
L
Linus Torvalds 已提交
1430 1431 1432 1433 1434 1435 1436
	.handler	=	udpv6_rcv,
	.err_handler	=	udpv6_err,
	.flags		=	INET6_PROTO_NOPOLICY|INET6_PROTO_FINAL,
};

/* ------------------------------------------------------------------------ */
#ifdef CONFIG_PROC_FS
1437
int udp6_seq_show(struct seq_file *seq, void *v)
L
Linus Torvalds 已提交
1438
{
1439 1440 1441 1442 1443 1444 1445 1446 1447
	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
1448 1449 1450
	return 0;
}

1451 1452 1453 1454 1455 1456 1457 1458
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
1459 1460 1461
static struct udp_seq_afinfo udp6_seq_afinfo = {
	.name		= "udp6",
	.family		= AF_INET6,
1462
	.udp_table	= &udp_table,
1463
	.seq_fops	= &udp6_afinfo_seq_fops,
1464 1465 1466
	.seq_ops	= {
		.show		= udp6_seq_show,
	},
L
Linus Torvalds 已提交
1467 1468
};

1469
int __net_init udp6_proc_init(struct net *net)
L
Linus Torvalds 已提交
1470
{
1471
	return udp_proc_register(net, &udp6_seq_afinfo);
L
Linus Torvalds 已提交
1472 1473
}

1474 1475
void udp6_proc_exit(struct net *net) {
	udp_proc_unregister(net, &udp6_seq_afinfo);
L
Linus Torvalds 已提交
1476 1477 1478
}
#endif /* CONFIG_PROC_FS */

E
Eric Dumazet 已提交
1479 1480 1481 1482 1483 1484 1485 1486 1487 1488 1489
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
1490 1491 1492
/* ------------------------------------------------------------------------ */

struct proto udpv6_prot = {
1493 1494
	.name		   = "UDPv6",
	.owner		   = THIS_MODULE,
1495
	.close		   = udp_lib_close,
1496 1497 1498 1499 1500 1501 1502 1503
	.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,
1504
	.backlog_rcv	   = __udpv6_queue_rcv_skb,
1505 1506
	.hash		   = udp_lib_hash,
	.unhash		   = udp_lib_unhash,
E
Eric Dumazet 已提交
1507
	.rehash		   = udp_v6_rehash,
1508
	.get_port	   = udp_v6_get_port,
H
Hideo Aoki 已提交
1509 1510 1511 1512
	.memory_allocated  = &udp_memory_allocated,
	.sysctl_mem	   = sysctl_udp_mem,
	.sysctl_wmem	   = &sysctl_udp_wmem_min,
	.sysctl_rmem	   = &sysctl_udp_rmem_min,
1513
	.obj_size	   = sizeof(struct udp6_sock),
1514
	.slab_flags	   = SLAB_DESTROY_BY_RCU,
1515
	.h.udp_table	   = &udp_table,
1516
#ifdef CONFIG_COMPAT
1517 1518
	.compat_setsockopt = compat_udpv6_setsockopt,
	.compat_getsockopt = compat_udpv6_getsockopt,
1519
#endif
E
Eric Dumazet 已提交
1520
	.clear_sk	   = udp_v6_clear_sk,
L
Linus Torvalds 已提交
1521 1522 1523 1524 1525 1526 1527 1528 1529 1530 1531
};

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


1532
int __init udpv6_init(void)
L
Linus Torvalds 已提交
1533
{
1534 1535
	int ret;

1536 1537
	ret = inet6_add_protocol(&udpv6_protocol, IPPROTO_UDP);
	if (ret)
1538
		goto out;
1539

1540 1541 1542 1543 1544 1545 1546 1547 1548 1549 1550
	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;
}

1551
void udpv6_exit(void)
1552 1553 1554
{
	inet6_unregister_protosw(&udpv6_protosw);
	inet6_del_protocol(&udpv6_protocol, IPPROTO_UDP);
L
Linus Torvalds 已提交
1555
}