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 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
static u32 udp6_ehashfn(const struct net *net,
			const struct in6_addr *laddr,
			const u16 lport,
			const struct in6_addr *faddr,
			const __be16 fport)
61
{
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
static u32 udp6_portaddr_hash(const struct net *net,
			      const struct in6_addr *addr6,
			      unsigned int port)
110 111 112 113 114
{
	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 sock *sk, struct msghdr *msg, size_t len,
L
Linus Torvalds 已提交
395 396 397 398
		  int noblock, int flags, int *addr_len)
{
	struct ipv6_pinfo *np = inet6_sk(sk);
	struct inet_sock *inet = inet_sk(sk);
399
	struct sk_buff *skb;
400
	unsigned int ulen, copied;
401
	int peeked, off = 0;
402 403
	int err;
	int is_udplite = IS_UDPLITE(sk);
404
	int is_udp4;
405
	bool slow;
L
Linus Torvalds 已提交
406 407

	if (flags & MSG_ERRQUEUE)
408
		return ipv6_recv_error(sk, msg, len, addr_len);
L
Linus Torvalds 已提交
409

410
	if (np->rxpmtu && np->rxopt.bits.rxpmtu)
411
		return ipv6_recv_rxpmtu(sk, msg, len, addr_len);
412

L
Linus Torvalds 已提交
413
try_again:
414
	skb = __skb_recv_datagram(sk, flags | (noblock ? MSG_DONTWAIT : 0),
415
				  &peeked, &off, &err);
L
Linus Torvalds 已提交
416 417 418
	if (!skb)
		goto out;

419
	ulen = skb->len - sizeof(struct udphdr);
420 421 422 423
	copied = len;
	if (copied > ulen)
		copied = ulen;
	else if (copied < ulen)
424
		msg->msg_flags |= MSG_TRUNC;
L
Linus Torvalds 已提交
425

426 427
	is_udp4 = (skb->protocol == htons(ETH_P_IP));

428
	/*
429 430 431
	 * 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.
432 433
	 */

434
	if (copied < ulen || UDP_SKB_CB(skb)->partial_cov) {
435
		if (udp_lib_checksum_complete(skb))
L
Linus Torvalds 已提交
436
			goto csum_copy_err;
437 438
	}

439
	if (skb_csum_unnecessary(skb))
440 441
		err = skb_copy_datagram_msg(skb, sizeof(struct udphdr),
					    msg, copied);
442
	else {
443
		err = skb_copy_and_csum_datagram_msg(skb, sizeof(struct udphdr), msg);
L
Linus Torvalds 已提交
444 445 446
		if (err == -EINVAL)
			goto csum_copy_err;
	}
447 448
	if (unlikely(err)) {
		trace_kfree_skb(skb, udpv6_recvmsg);
449 450 451 452 453 454 455 456 457 458 459
		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
460
		goto out_free;
461
	}
462 463 464 465 466 467 468 469
	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);
	}
470

471
	sock_recv_ts_and_drops(msg, sk, skb);
L
Linus Torvalds 已提交
472 473 474

	/* Copy the address. */
	if (msg->msg_name) {
475
		DECLARE_SOCKADDR(struct sockaddr_in6 *, sin6, msg->msg_name);
L
Linus Torvalds 已提交
476
		sin6->sin6_family = AF_INET6;
477
		sin6->sin6_port = udp_hdr(skb)->source;
L
Linus Torvalds 已提交
478 479
		sin6->sin6_flowinfo = 0;

480
		if (is_udp4) {
B
Brian Haley 已提交
481 482
			ipv6_addr_set_v4mapped(ip_hdr(skb)->saddr,
					       &sin6->sin6_addr);
483 484
			sin6->sin6_scope_id = 0;
		} else {
A
Alexey Dobriyan 已提交
485
			sin6->sin6_addr = ipv6_hdr(skb)->saddr;
486 487
			sin6->sin6_scope_id =
				ipv6_iface_scope_id(&sin6->sin6_addr,
488
						    inet6_iif(skb));
L
Linus Torvalds 已提交
489
		}
490
		*addr_len = sizeof(*sin6);
L
Linus Torvalds 已提交
491
	}
492 493 494 495

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

496
	if (is_udp4) {
L
Linus Torvalds 已提交
497 498 499 500
		if (inet->cmsg_flags)
			ip_cmsg_recv(msg, skb);
	} else {
		if (np->rxopt.all)
501
			ip6_datagram_recv_specific_ctl(sk, msg, skb);
502
	}
L
Linus Torvalds 已提交
503

504
	err = copied;
L
Linus Torvalds 已提交
505
	if (flags & MSG_TRUNC)
506
		err = ulen;
L
Linus Torvalds 已提交
507 508

out_free:
509
	skb_free_datagram_locked(sk, skb);
L
Linus Torvalds 已提交
510 511 512 513
out:
	return err;

csum_copy_err:
514
	slow = lock_sock_fast(sk);
515
	if (!skb_kill_datagram(sk, skb, flags)) {
516 517 518
		if (is_udp4) {
			UDP_INC_STATS_USER(sock_net(sk),
					UDP_MIB_CSUMERRORS, is_udplite);
519 520
			UDP_INC_STATS_USER(sock_net(sk),
					UDP_MIB_INERRORS, is_udplite);
521 522 523
		} else {
			UDP6_INC_STATS_USER(sock_net(sk),
					UDP_MIB_CSUMERRORS, is_udplite);
524 525
			UDP6_INC_STATS_USER(sock_net(sk),
					UDP_MIB_INERRORS, is_udplite);
526
		}
527
	}
528
	unlock_sock_fast(sk, slow);
L
Linus Torvalds 已提交
529

530
	if (noblock)
L
Linus Torvalds 已提交
531
		return -EAGAIN;
532 533 534

	/* starting over for a new packet */
	msg->msg_flags &= ~MSG_TRUNC;
L
Linus Torvalds 已提交
535 536 537
	goto try_again;
}

538
void __udp6_lib_err(struct sk_buff *skb, struct inet6_skb_parm *opt,
539
		    u8 type, u8 code, int offset, __be32 info,
540
		    struct udp_table *udptable)
L
Linus Torvalds 已提交
541 542
{
	struct ipv6_pinfo *np;
543 544 545
	const struct ipv6hdr *hdr = (const struct ipv6hdr *)skb->data;
	const struct in6_addr *saddr = &hdr->saddr;
	const struct in6_addr *daddr = &hdr->daddr;
546
	struct udphdr *uh = (struct udphdr *)(skb->data+offset);
L
Linus Torvalds 已提交
547 548
	struct sock *sk;
	int err;
549
	struct net *net = dev_net(skb->dev);
L
Linus Torvalds 已提交
550

551
	sk = __udp6_lib_lookup(net, daddr, uh->dest,
552
			       saddr, uh->source, inet6_iif(skb), udptable);
553
	if (!sk) {
554 555
		ICMP6_INC_STATS_BH(net, __in6_dev_get(skb->dev),
				   ICMP6_MIB_INERRORS);
L
Linus Torvalds 已提交
556
		return;
557
	}
L
Linus Torvalds 已提交
558

559 560 561
	if (type == ICMPV6_PKT_TOOBIG) {
		if (!ip6_sk_accept_pmtu(sk))
			goto out;
562
		ip6_sk_update_pmtu(skb, sk, info);
563
	}
564
	if (type == NDISC_REDIRECT) {
565
		ip6_sk_redirect(skb, sk);
566 567
		goto out;
	}
568

L
Linus Torvalds 已提交
569 570 571 572 573 574 575 576
	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;

577
	if (np->recverr)
L
Linus Torvalds 已提交
578
		ipv6_icmp_error(sk, skb, err, uh->dest, ntohl(info), (u8 *)(uh+1));
579

L
Linus Torvalds 已提交
580 581 582 583 584 585
	sk->sk_err = err;
	sk->sk_error_report(sk);
out:
	sock_put(sk);
}

586 587 588 589
static int __udpv6_queue_rcv_skb(struct sock *sk, struct sk_buff *skb)
{
	int rc;

590
	if (!ipv6_addr_any(&sk->sk_v6_daddr)) {
591
		sock_rps_save_rxhash(sk, skb);
592
		sk_mark_napi_id(sk, skb);
E
Eric Dumazet 已提交
593
		sk_incoming_cpu_update(sk);
594
	}
595 596 597 598 599 600 601 602 603 604 605 606 607 608 609 610

	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;
}

611
static __inline__ void udpv6_err(struct sk_buff *skb,
612
				 struct inet6_skb_parm *opt, u8 type,
613
				 u8 code, int offset, __be32 info)
614
{
615
	__udp6_lib_err(skb, opt, type, code, offset, info, &udp_table);
616 617
}

618 619 620 621 622 623 624 625
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);

626
int udpv6_queue_rcv_skb(struct sock *sk, struct sk_buff *skb)
L
Linus Torvalds 已提交
627
{
628
	struct udp_sock *up = udp_sk(sk);
629
	int rc;
W
Wang Chen 已提交
630
	int is_udplite = IS_UDPLITE(sk);
631

632 633
	if (!xfrm6_policy_check(sk, XFRM_POLICY_IN, skb))
		goto drop;
L
Linus Torvalds 已提交
634

635 636 637 638 639 640 641 642 643 644 645 646 647 648 649 650 651 652 653
	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;

654 655 656 657
			/* Verify checksum before giving to encap */
			if (udp_lib_checksum_complete(skb))
				goto csum_error;

658 659 660 661 662 663 664 665 666 667 668 669
			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 */
	}

670 671 672
	/*
	 * UDP-Lite specific tests, ignored on UDP sockets (see net/ipv4/udp.c).
	 */
W
Wang Chen 已提交
673
	if ((is_udplite & UDPLITE_RECV_CC)  &&  UDP_SKB_CB(skb)->partial_cov) {
674 675

		if (up->pcrlen == 0) {          /* full coverage was set  */
676 677
			net_dbg_ratelimited("UDPLITE6: partial coverage %d while full coverage %d requested\n",
					    UDP_SKB_CB(skb)->cscov, skb->len);
678 679 680
			goto drop;
		}
		if (UDP_SKB_CB(skb)->cscov  <  up->pcrlen) {
681 682
			net_dbg_ratelimited("UDPLITE6: coverage %d too small, need min %d\n",
					    UDP_SKB_CB(skb)->cscov, up->pcrlen);
683 684
			goto drop;
		}
L
Linus Torvalds 已提交
685 686
	}

687
	if (rcu_access_pointer(sk->sk_filter)) {
688
		if (udp_lib_checksum_complete(skb))
689
			goto csum_error;
690
	}
691

692
	if (sk_rcvqueues_full(sk, sk->sk_rcvbuf)) {
693 694
		UDP6_INC_STATS_BH(sock_net(sk),
				  UDP_MIB_RCVBUFERRORS, is_udplite);
695
		goto drop;
696
	}
697

698
	skb_dst_drop(skb);
699

700 701 702 703 704 705 706 707 708
	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);
709 710

	return rc;
711

712 713
csum_error:
	UDP6_INC_STATS_BH(sock_net(sk), UDP_MIB_CSUMERRORS, is_udplite);
714
drop:
715
	UDP6_INC_STATS_BH(sock_net(sk), UDP_MIB_INERRORS, is_udplite);
716
	atomic_inc(&sk->sk_drops);
717 718
	kfree_skb(skb);
	return -1;
L
Linus Torvalds 已提交
719 720
}

721 722 723 724
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
725
{
726
	struct inet_sock *inet = inet_sk(sk);
L
Linus Torvalds 已提交
727

728 729 730 731 732 733 734 735 736 737 738 739 740
	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)) ||
	    (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
741 742
}

743 744 745
static void flush_stack(struct sock **stack, unsigned int count,
			struct sk_buff *skb, unsigned int final)
{
746
	struct sk_buff *skb1 = NULL;
747
	struct sock *sk;
748
	unsigned int i;
749 750

	for (i = 0; i < count; i++) {
751
		sk = stack[i];
752
		if (likely(!skb1))
753 754 755 756 757 758 759
			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));
760
		}
761 762 763

		if (skb1 && udpv6_queue_rcv_skb(sk, skb1) <= 0)
			skb1 = NULL;
764
		sock_put(sk);
765
	}
766 767
	if (unlikely(skb1))
		kfree_skb(skb1);
768
}
769 770 771 772 773 774

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.
	 */
775 776 777
	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));
778 779
}

L
Linus Torvalds 已提交
780 781 782 783
/*
 * Note: called only from the BH handler context,
 * so we don't need to lock the hashes.
 */
784
static int __udp6_lib_mcast_deliver(struct net *net, struct sk_buff *skb,
785
		const struct in6_addr *saddr, const struct in6_addr *daddr,
786
		struct udp_table *udptable, int proto)
L
Linus Torvalds 已提交
787
{
788
	struct sock *sk, *stack[256 / sizeof(struct sock *)];
789
	const struct udphdr *uh = udp_hdr(skb);
790 791 792 793
	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);
794 795
	unsigned int count = 0, offset = offsetof(typeof(*sk), sk_nulls_node);
	unsigned int hash2 = 0, hash2_any = 0, use_hash2 = (hslot->count > 10);
796
	bool inner_flushed = false;
797 798 799 800 801 802 803 804 805

	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
806

807
	spin_lock(&hslot->lock);
808
	sk_nulls_for_each_entry_offset(sk, node, &hslot->head, offset) {
809 810 811 812 813 814 815 816 817 818
		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);
819
				inner_flushed = true;
820 821
				count = 0;
			}
822
			stack[count++] = sk;
823
			sock_hold(sk);
H
Hideo Aoki 已提交
824
		}
L
Linus Torvalds 已提交
825
	}
826

827
	spin_unlock(&hslot->lock);
828

829 830 831 832 833 834
	/* 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;
	}

835 836 837
	if (count) {
		flush_stack(stack, count, skb, count - 1);
	} else {
838 839 840 841
		if (!inner_flushed)
			UDP_INC_STATS_BH(net, UDP_MIB_IGNOREDMULTI,
					 proto == IPPROTO_UDPLITE);
		consume_skb(skb);
842
	}
843
	return 0;
L
Linus Torvalds 已提交
844 845
}

846
int __udp6_lib_rcv(struct sk_buff *skb, struct udp_table *udptable,
847
		   int proto)
L
Linus Torvalds 已提交
848
{
849
	struct net *net = dev_net(skb->dev);
L
Linus Torvalds 已提交
850
	struct sock *sk;
851
	struct udphdr *uh;
852
	const struct in6_addr *saddr, *daddr;
L
Linus Torvalds 已提交
853 854 855
	u32 ulen = 0;

	if (!pskb_may_pull(skb, sizeof(struct udphdr)))
856
		goto discard;
L
Linus Torvalds 已提交
857

858 859
	saddr = &ipv6_hdr(skb)->saddr;
	daddr = &ipv6_hdr(skb)->daddr;
860
	uh = udp_hdr(skb);
L
Linus Torvalds 已提交
861 862

	ulen = ntohs(uh->len);
863 864
	if (ulen > skb->len)
		goto short_packet;
L
Linus Torvalds 已提交
865

866 867
	if (proto == IPPROTO_UDP) {
		/* UDP validates ulen. */
L
Linus Torvalds 已提交
868

869 870 871
		/* Check for jumbo payload */
		if (ulen == 0)
			ulen = skb->len;
L
Linus Torvalds 已提交
872

873 874
		if (ulen < sizeof(*uh))
			goto short_packet;
L
Linus Torvalds 已提交
875

876 877 878
		if (ulen < skb->len) {
			if (pskb_trim_rcsum(skb, ulen))
				goto short_packet;
879 880
			saddr = &ipv6_hdr(skb)->saddr;
			daddr = &ipv6_hdr(skb)->daddr;
881
			uh = udp_hdr(skb);
882 883
		}
	}
L
Linus Torvalds 已提交
884

885
	if (udp6_csum_init(skb, uh, proto))
886
		goto csum_error;
887

888 889
	/*
	 *	Multicast receive code
L
Linus Torvalds 已提交
890
	 */
891
	if (ipv6_addr_is_multicast(daddr))
892
		return __udp6_lib_mcast_deliver(net, skb,
893
				saddr, daddr, udptable, proto);
L
Linus Torvalds 已提交
894 895 896

	/* Unicast */

897
	/*
L
Linus Torvalds 已提交
898 899 900
	 * check socket cache ... must talk to Alan about his plans
	 * for sock caches... i'll skip this for now.
	 */
901
	sk = __udp6_lib_lookup_skb(skb, uh->source, uh->dest, udptable);
902
	if (sk != NULL) {
903 904
		int ret;

905
		if (!uh->check && !udp_sk(sk)->no_check6_rx) {
906
			sock_put(sk);
907 908 909 910
			udp6_csum_zero_error(skb);
			goto csum_error;
		}

911
		if (inet_get_convert_csum(sk) && uh->check && !IS_UDPLITE(sk))
912 913 914
			skb_checksum_try_convert(skb, IPPROTO_UDP, uh->check,
						 ip6_compute_pseudo);

915
		ret = udpv6_queue_rcv_skb(sk, skb);
916
		sock_put(sk);
L
Linus Torvalds 已提交
917

918 919 920 921 922
		/* 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
923

S
Stephen Hemminger 已提交
924
		return 0;
L
Linus Torvalds 已提交
925
	}
926

927 928 929 930 931
	if (!uh->check) {
		udp6_csum_zero_error(skb);
		goto csum_error;
	}

932
	if (!xfrm6_policy_check(NULL, XFRM_POLICY_IN, skb))
933
		goto discard;
934 935

	if (udp_lib_checksum_complete(skb))
936
		goto csum_error;
937 938 939 940 941

	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
942
	return 0;
L
Linus Torvalds 已提交
943

944
short_packet:
945 946 947 948 949
	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));
950 951 952
	goto discard;
csum_error:
	UDP6_INC_STATS_BH(net, UDP_MIB_CSUMERRORS, proto == IPPROTO_UDPLITE);
L
Linus Torvalds 已提交
953
discard:
954
	UDP6_INC_STATS_BH(net, UDP_MIB_INERRORS, proto == IPPROTO_UDPLITE);
L
Linus Torvalds 已提交
955
	kfree_skb(skb);
S
Stephen Hemminger 已提交
956
	return 0;
L
Linus Torvalds 已提交
957
}
958

959
static __inline__ int udpv6_rcv(struct sk_buff *skb)
960
{
961
	return __udp6_lib_rcv(skb, &udp_table, IPPROTO_UDP);
962 963
}

L
Linus Torvalds 已提交
964 965 966 967 968 969 970
/*
 * 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);

971 972 973
	if (up->pending == AF_INET)
		udp_flush_pending_frames(sk);
	else if (up->pending) {
L
Linus Torvalds 已提交
974 975 976
		up->len = 0;
		up->pending = 0;
		ip6_flush_pending_frames(sk);
977
	}
L
Linus Torvalds 已提交
978 979
}

980
/**
981 982 983 984
 *	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)
985 986 987 988 989 990 991
 */
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
992
	struct sk_buff *frags = skb_shinfo(skb)->frag_list;
993 994
	__wsum csum = 0;

V
Vlad Yasevich 已提交
995
	if (!frags) {
996 997 998 999 1000 1001 1002 1003 1004 1005 1006 1007 1008 1009 1010
		/* 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
1011 1012 1013
		do {
			csum = csum_add(csum, frags->csum);
		} while ((frags = frags->next));
1014 1015 1016 1017 1018 1019 1020 1021

		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
 */

V
Vlad Yasevich 已提交
1026
static int udp_v6_send_skb(struct sk_buff *skb, struct flowi6 *fl6)
L
Linus Torvalds 已提交
1027
{
V
Vlad Yasevich 已提交
1028
	struct sock *sk = skb->sk;
L
Linus Torvalds 已提交
1029 1030
	struct udphdr *uh;
	int err = 0;
W
Wang Chen 已提交
1031
	int is_udplite = IS_UDPLITE(sk);
1032
	__wsum csum = 0;
V
Vlad Yasevich 已提交
1033 1034
	int offset = skb_transport_offset(skb);
	int len = skb->len - offset;
L
Linus Torvalds 已提交
1035 1036 1037 1038

	/*
	 * Create a UDP header
	 */
1039
	uh = udp_hdr(skb);
1040 1041
	uh->source = fl6->fl6_sport;
	uh->dest = fl6->fl6_dport;
V
Vlad Yasevich 已提交
1042
	uh->len = htons(len);
L
Linus Torvalds 已提交
1043 1044
	uh->check = 0;

W
Wang Chen 已提交
1045
	if (is_udplite)
V
Vlad Yasevich 已提交
1046 1047
		csum = udplite_csum(skb);
	else if (udp_sk(sk)->no_check6_tx) {   /* UDP csum disabled */
1048 1049 1050
		skb->ip_summed = CHECKSUM_NONE;
		goto send;
	} else if (skb->ip_summed == CHECKSUM_PARTIAL) { /* UDP hardware csum */
V
Vlad Yasevich 已提交
1051
		udp6_hwcsum_outgoing(sk, skb, &fl6->saddr, &fl6->daddr, len);
1052 1053
		goto send;
	} else
V
Vlad Yasevich 已提交
1054
		csum = udp_csum(skb);
L
Linus Torvalds 已提交
1055

1056
	/* add protocol-dependent pseudo-header */
1057
	uh->check = csum_ipv6_magic(&fl6->saddr, &fl6->daddr,
V
Vlad Yasevich 已提交
1058
				    len, fl6->flowi6_proto, csum);
L
Linus Torvalds 已提交
1059
	if (uh->check == 0)
1060
		uh->check = CSUM_MANGLED_0;
L
Linus Torvalds 已提交
1061

1062
send:
V
Vlad Yasevich 已提交
1063
	err = ip6_send_skb(skb);
E
Eric Dumazet 已提交
1064 1065 1066 1067 1068 1069 1070 1071 1072
	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
1073 1074 1075 1076 1077 1078 1079 1080 1081 1082 1083 1084 1085 1086 1087 1088 1089 1090 1091 1092 1093 1094 1095 1096
	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
1097 1098 1099 1100 1101 1102
out:
	up->len = 0;
	up->pending = 0;
	return err;
}

1103
int udpv6_sendmsg(struct sock *sk, struct msghdr *msg, size_t len)
L
Linus Torvalds 已提交
1104 1105 1106 1107 1108
{
	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);
1109
	DECLARE_SOCKADDR(struct sockaddr_in6 *, sin6, msg->msg_name);
1110
	struct in6_addr *daddr, *final_p, final;
L
Linus Torvalds 已提交
1111 1112
	struct ipv6_txoptions *opt = NULL;
	struct ip6_flowlabel *flowlabel = NULL;
1113
	struct flowi6 fl6;
L
Linus Torvalds 已提交
1114 1115 1116 1117
	struct dst_entry *dst;
	int addr_len = msg->msg_namelen;
	int ulen = len;
	int hlimit = -1;
1118
	int tclass = -1;
1119
	int dontfrag = -1;
L
Linus Torvalds 已提交
1120 1121
	int corkreq = up->corkflag || msg->msg_flags&MSG_MORE;
	int err;
1122
	int connected = 0;
W
Wang Chen 已提交
1123
	int is_udplite = IS_UDPLITE(sk);
1124
	int (*getfrag)(void *, char *, int, int, int, struct sk_buff *);
L
Linus Torvalds 已提交
1125 1126 1127 1128 1129 1130 1131 1132 1133 1134 1135 1136 1137 1138 1139 1140 1141 1142 1143 1144 1145 1146 1147 1148 1149

	/* 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;
1150
		daddr = &sk->sk_v6_daddr;
1151
	} else
L
Linus Torvalds 已提交
1152 1153 1154
		daddr = NULL;

	if (daddr) {
B
Brian Haley 已提交
1155
		if (ipv6_addr_v4mapped(daddr)) {
L
Linus Torvalds 已提交
1156 1157
			struct sockaddr_in sin;
			sin.sin_family = AF_INET;
E
Eric Dumazet 已提交
1158
			sin.sin_port = sin6 ? sin6->sin6_port : inet->inet_dport;
L
Linus Torvalds 已提交
1159 1160 1161 1162 1163 1164
			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;
1165
			return udp_sendmsg(sk, msg, len);
L
Linus Torvalds 已提交
1166 1167 1168 1169
		}
	}

	if (up->pending == AF_INET)
1170
		return udp_sendmsg(sk, msg, len);
L
Linus Torvalds 已提交
1171 1172

	/* Rough check on arithmetic overflow,
1173
	   better check is made in ip6_append_data().
L
Linus Torvalds 已提交
1174 1175 1176
	   */
	if (len > INT_MAX - sizeof(struct udphdr))
		return -EMSGSIZE;
1177

1178
	getfrag  =  is_udplite ?  udplite_getfrag : ip_generic_getfrag;
L
Linus Torvalds 已提交
1179 1180 1181 1182 1183 1184 1185 1186 1187 1188 1189 1190 1191 1192 1193 1194 1195 1196
	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);

1197
	memset(&fl6, 0, sizeof(fl6));
L
Linus Torvalds 已提交
1198 1199 1200 1201 1202

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

1203
		fl6.fl6_dport = sin6->sin6_port;
L
Linus Torvalds 已提交
1204 1205 1206
		daddr = &sin6->sin6_addr;

		if (np->sndflow) {
1207 1208 1209
			fl6.flowlabel = sin6->sin6_flowinfo&IPV6_FLOWINFO_MASK;
			if (fl6.flowlabel&IPV6_FLOWLABEL_MASK) {
				flowlabel = fl6_sock_lookup(sk, fl6.flowlabel);
1210
				if (!flowlabel)
L
Linus Torvalds 已提交
1211 1212 1213 1214 1215 1216 1217 1218 1219
					return -EINVAL;
			}
		}

		/*
		 * Otherwise it will be difficult to maintain
		 * sk->sk_dst_cache.
		 */
		if (sk->sk_state == TCP_ESTABLISHED &&
1220 1221
		    ipv6_addr_equal(daddr, &sk->sk_v6_daddr))
			daddr = &sk->sk_v6_daddr;
L
Linus Torvalds 已提交
1222 1223 1224

		if (addr_len >= sizeof(struct sockaddr_in6) &&
		    sin6->sin6_scope_id &&
1225
		    __ipv6_addr_needs_scope_id(__ipv6_addr_type(daddr)))
1226
			fl6.flowi6_oif = sin6->sin6_scope_id;
L
Linus Torvalds 已提交
1227 1228 1229 1230
	} else {
		if (sk->sk_state != TCP_ESTABLISHED)
			return -EDESTADDRREQ;

1231
		fl6.fl6_dport = inet->inet_dport;
1232
		daddr = &sk->sk_v6_daddr;
1233
		fl6.flowlabel = np->flow_label;
1234
		connected = 1;
L
Linus Torvalds 已提交
1235 1236
	}

1237 1238
	if (!fl6.flowi6_oif)
		fl6.flowi6_oif = sk->sk_bound_dev_if;
L
Linus Torvalds 已提交
1239

1240 1241
	if (!fl6.flowi6_oif)
		fl6.flowi6_oif = np->sticky_pktinfo.ipi6_ifindex;
1242

1243
	fl6.flowi6_mark = sk->sk_mark;
B
Brian Haley 已提交
1244

L
Linus Torvalds 已提交
1245 1246 1247 1248 1249
	if (msg->msg_controllen) {
		opt = &opt_space;
		memset(opt, 0, sizeof(struct ipv6_txoptions));
		opt->tot_len = sizeof(*opt);

1250 1251
		err = ip6_datagram_send_ctl(sock_net(sk), sk, msg, &fl6, opt,
					    &hlimit, &tclass, &dontfrag);
L
Linus Torvalds 已提交
1252 1253 1254 1255
		if (err < 0) {
			fl6_sock_release(flowlabel);
			return err;
		}
1256 1257
		if ((fl6.flowlabel&IPV6_FLOWLABEL_MASK) && !flowlabel) {
			flowlabel = fl6_sock_lookup(sk, fl6.flowlabel);
1258
			if (!flowlabel)
L
Linus Torvalds 已提交
1259 1260 1261 1262
				return -EINVAL;
		}
		if (!(opt->opt_nflen|opt->opt_flen))
			opt = NULL;
1263
		connected = 0;
L
Linus Torvalds 已提交
1264
	}
1265
	if (!opt)
L
Linus Torvalds 已提交
1266
		opt = np->opt;
1267 1268 1269
	if (flowlabel)
		opt = fl6_merge_options(&opt_space, flowlabel, opt);
	opt = ipv6_fixup_options(&opt_space, opt);
L
Linus Torvalds 已提交
1270

1271
	fl6.flowi6_proto = sk->sk_protocol;
1272
	if (!ipv6_addr_any(daddr))
A
Alexey Dobriyan 已提交
1273
		fl6.daddr = *daddr;
1274
	else
1275 1276
		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
1277
		fl6.saddr = np->saddr;
1278
	fl6.fl6_sport = inet->inet_sport;
1279

1280
	final_p = fl6_update_dst(&fl6, opt, &final);
1281
	if (final_p)
1282
		connected = 0;
L
Linus Torvalds 已提交
1283

1284 1285
	if (!fl6.flowi6_oif && ipv6_addr_is_multicast(&fl6.daddr)) {
		fl6.flowi6_oif = np->mcast_oif;
1286
		connected = 0;
1287 1288
	} else if (!fl6.flowi6_oif)
		fl6.flowi6_oif = np->ucast_oif;
L
Linus Torvalds 已提交
1289

1290
	security_sk_classify_flow(sk, flowi6_to_flowi(&fl6));
V
Venkat Yekkirala 已提交
1291

1292
	dst = ip6_sk_dst_lookup_flow(sk, &fl6, final_p);
1293 1294 1295
	if (IS_ERR(dst)) {
		err = PTR_ERR(dst);
		dst = NULL;
L
Linus Torvalds 已提交
1296
		goto out;
1297
	}
L
Linus Torvalds 已提交
1298

1299 1300
	if (hlimit < 0)
		hlimit = ip6_sk_dst_hoplimit(np, &fl6, dst);
L
Linus Torvalds 已提交
1301

G
Gerrit Renker 已提交
1302
	if (tclass < 0)
1303 1304
		tclass = np->tclass;

L
Linus Torvalds 已提交
1305 1306 1307 1308
	if (msg->msg_flags&MSG_CONFIRM)
		goto do_confirm;
back_from_confirm:

1309 1310 1311 1312 1313 1314 1315 1316 1317 1318 1319 1320 1321 1322
	/* 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
1323 1324 1325 1326 1327 1328
	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);

1329
		net_dbg_ratelimited("udp cork app bug 2\n");
L
Linus Torvalds 已提交
1330 1331 1332 1333 1334 1335 1336
		err = -EINVAL;
		goto out;
	}

	up->pending = AF_INET6;

do_append_data:
1337 1338
	if (dontfrag < 0)
		dontfrag = np->dontfrag;
L
Linus Torvalds 已提交
1339
	up->len += ulen;
1340
	err = ip6_append_data(sk, getfrag, msg, ulen,
1341
		sizeof(struct udphdr), hlimit, tclass, opt, &fl6,
1342
		(struct rt6_info *)dst,
1343
		corkreq ? msg->msg_flags|MSG_MORE : msg->msg_flags, dontfrag);
L
Linus Torvalds 已提交
1344 1345 1346
	if (err)
		udp_v6_flush_pending_frames(sk);
	else if (!corkreq)
1347
		err = udp_v6_push_pending_frames(sk);
H
Herbert Xu 已提交
1348 1349
	else if (unlikely(skb_queue_empty(&sk->sk_write_queue)))
		up->pending = 0;
L
Linus Torvalds 已提交
1350

1351 1352 1353 1354 1355
	if (err > 0)
		err = np->recverr ? net_xmit_errno(err) : 0;
	release_sock(sk);

release_dst:
1356 1357 1358
	if (dst) {
		if (connected) {
			ip6_dst_store(sk, dst,
1359 1360
				      ipv6_addr_equal(&fl6.daddr, &sk->sk_v6_daddr) ?
				      &sk->sk_v6_daddr : NULL,
1361
#ifdef CONFIG_IPV6_SUBTREES
1362
				      ipv6_addr_equal(&fl6.saddr, &np->saddr) ?
1363 1364 1365
				      &np->saddr :
#endif
				      NULL);
1366 1367 1368
		} else {
			dst_release(dst);
		}
1369
		dst = NULL;
1370 1371
	}

L
Linus Torvalds 已提交
1372
out:
1373
	dst_release(dst);
L
Linus Torvalds 已提交
1374
	fl6_sock_release(flowlabel);
1375
	if (!err)
L
Linus Torvalds 已提交
1376
		return len;
1377 1378 1379 1380 1381 1382 1383 1384
	/*
	 * 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)) {
1385 1386
		UDP6_INC_STATS_USER(sock_net(sk),
				UDP_MIB_SNDBUFERRORS, is_udplite);
1387
	}
L
Linus Torvalds 已提交
1388 1389 1390 1391 1392 1393 1394 1395 1396 1397
	return err;

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

1398
void udpv6_destroy_sock(struct sock *sk)
L
Linus Torvalds 已提交
1399
{
T
Tom Parkin 已提交
1400
	struct udp_sock *up = udp_sk(sk);
L
Linus Torvalds 已提交
1401 1402 1403 1404
	lock_sock(sk);
	udp_v6_flush_pending_frames(sk);
	release_sock(sk);

T
Tom Parkin 已提交
1405 1406 1407 1408 1409 1410 1411
	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
1412 1413 1414 1415 1416 1417
	inet6_destroy_sock(sk);
}

/*
 *	Socket option code for UDP
 */
1418
int udpv6_setsockopt(struct sock *sk, int level, int optname,
1419
		     char __user *optval, unsigned int optlen)
1420
{
1421
	if (level == SOL_UDP  ||  level == SOL_UDPLITE)
1422 1423
		return udp_lib_setsockopt(sk, level, optname, optval, optlen,
					  udp_v6_push_pending_frames);
1424
	return ipv6_setsockopt(sk, level, optname, optval, optlen);
1425 1426 1427
}

#ifdef CONFIG_COMPAT
1428
int compat_udpv6_setsockopt(struct sock *sk, int level, int optname,
1429
			    char __user *optval, unsigned int optlen)
1430
{
1431
	if (level == SOL_UDP  ||  level == SOL_UDPLITE)
1432 1433
		return udp_lib_setsockopt(sk, level, optname, optval, optlen,
					  udp_v6_push_pending_frames);
1434
	return compat_ipv6_setsockopt(sk, level, optname, optval, optlen);
1435 1436 1437
}
#endif

1438 1439
int udpv6_getsockopt(struct sock *sk, int level, int optname,
		     char __user *optval, int __user *optlen)
1440
{
1441
	if (level == SOL_UDP  ||  level == SOL_UDPLITE)
1442
		return udp_lib_getsockopt(sk, level, optname, optval, optlen);
1443
	return ipv6_getsockopt(sk, level, optname, optval, optlen);
1444 1445 1446
}

#ifdef CONFIG_COMPAT
1447 1448
int compat_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 compat_ipv6_getsockopt(sk, level, optname, optval, optlen);
1453 1454 1455
}
#endif

1456
static const struct inet6_protocol udpv6_protocol = {
L
Linus Torvalds 已提交
1457 1458 1459 1460 1461 1462 1463
	.handler	=	udpv6_rcv,
	.err_handler	=	udpv6_err,
	.flags		=	INET6_PROTO_NOPOLICY|INET6_PROTO_FINAL,
};

/* ------------------------------------------------------------------------ */
#ifdef CONFIG_PROC_FS
1464
int udp6_seq_show(struct seq_file *seq, void *v)
L
Linus Torvalds 已提交
1465
{
1466 1467 1468 1469 1470 1471 1472 1473 1474
	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
1475 1476 1477
	return 0;
}

1478 1479 1480 1481 1482 1483 1484 1485
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
1486 1487 1488
static struct udp_seq_afinfo udp6_seq_afinfo = {
	.name		= "udp6",
	.family		= AF_INET6,
1489
	.udp_table	= &udp_table,
1490
	.seq_fops	= &udp6_afinfo_seq_fops,
1491 1492 1493
	.seq_ops	= {
		.show		= udp6_seq_show,
	},
L
Linus Torvalds 已提交
1494 1495
};

1496
int __net_init udp6_proc_init(struct net *net)
L
Linus Torvalds 已提交
1497
{
1498
	return udp_proc_register(net, &udp6_seq_afinfo);
L
Linus Torvalds 已提交
1499 1500
}

1501 1502
void udp6_proc_exit(struct net *net) {
	udp_proc_unregister(net, &udp6_seq_afinfo);
L
Linus Torvalds 已提交
1503 1504 1505
}
#endif /* CONFIG_PROC_FS */

E
Eric Dumazet 已提交
1506 1507 1508 1509 1510 1511 1512 1513 1514 1515 1516
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
1517 1518 1519
/* ------------------------------------------------------------------------ */

struct proto udpv6_prot = {
1520 1521
	.name		   = "UDPv6",
	.owner		   = THIS_MODULE,
1522
	.close		   = udp_lib_close,
1523 1524 1525 1526 1527 1528 1529 1530
	.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,
1531
	.backlog_rcv	   = __udpv6_queue_rcv_skb,
1532 1533
	.hash		   = udp_lib_hash,
	.unhash		   = udp_lib_unhash,
E
Eric Dumazet 已提交
1534
	.rehash		   = udp_v6_rehash,
1535
	.get_port	   = udp_v6_get_port,
H
Hideo Aoki 已提交
1536 1537 1538 1539
	.memory_allocated  = &udp_memory_allocated,
	.sysctl_mem	   = sysctl_udp_mem,
	.sysctl_wmem	   = &sysctl_udp_wmem_min,
	.sysctl_rmem	   = &sysctl_udp_rmem_min,
1540
	.obj_size	   = sizeof(struct udp6_sock),
1541
	.slab_flags	   = SLAB_DESTROY_BY_RCU,
1542
	.h.udp_table	   = &udp_table,
1543
#ifdef CONFIG_COMPAT
1544 1545
	.compat_setsockopt = compat_udpv6_setsockopt,
	.compat_getsockopt = compat_udpv6_getsockopt,
1546
#endif
E
Eric Dumazet 已提交
1547
	.clear_sk	   = udp_v6_clear_sk,
L
Linus Torvalds 已提交
1548 1549 1550 1551 1552 1553 1554 1555 1556 1557 1558
};

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


1559
int __init udpv6_init(void)
L
Linus Torvalds 已提交
1560
{
1561 1562
	int ret;

1563 1564
	ret = inet6_add_protocol(&udpv6_protocol, IPPROTO_UDP);
	if (ret)
1565
		goto out;
1566

1567 1568 1569 1570 1571 1572 1573 1574 1575 1576 1577
	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;
}

1578
void udpv6_exit(void)
1579 1580 1581
{
	inet6_unregister_protosw(&udpv6_protosw);
	inet6_del_protocol(&udpv6_protocol, IPPROTO_UDP);
L
Linus Torvalds 已提交
1582
}