udp.c 38.4 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,
149 150 151 152
				int dif)
{
	int score = -1;

153
	if (net_eq(sock_net(sk), net) && udp_sk(sk)->udp_port_hash == hnum &&
154 155 156 157
			sk->sk_family == PF_INET6) {
		struct inet_sock *inet = inet_sk(sk);

		score = 0;
E
Eric Dumazet 已提交
158 159
		if (inet->inet_dport) {
			if (inet->inet_dport != sport)
160 161 162
				return -1;
			score++;
		}
163 164
		if (!ipv6_addr_any(&sk->sk_v6_rcv_saddr)) {
			if (!ipv6_addr_equal(&sk->sk_v6_rcv_saddr, daddr))
165 166 167
				return -1;
			score++;
		}
168 169
		if (!ipv6_addr_any(&sk->sk_v6_daddr)) {
			if (!ipv6_addr_equal(&sk->sk_v6_daddr, saddr))
170 171 172 173 174 175 176 177 178 179 180 181
				return -1;
			score++;
		}
		if (sk->sk_bound_dev_if) {
			if (sk->sk_bound_dev_if != dif)
				return -1;
			score++;
		}
	}
	return score;
}

E
Eric Dumazet 已提交
182 183 184 185 186 187 188 189 190 191 192 193
#define SCORE2_MAX (1 + 1 + 1)
static inline int compute_score2(struct sock *sk, struct net *net,
				const struct in6_addr *saddr, __be16 sport,
				const struct in6_addr *daddr, unsigned short hnum,
				int dif)
{
	int score = -1;

	if (net_eq(sock_net(sk), net) && udp_sk(sk)->udp_port_hash == hnum &&
			sk->sk_family == PF_INET6) {
		struct inet_sock *inet = inet_sk(sk);

194
		if (!ipv6_addr_equal(&sk->sk_v6_rcv_saddr, daddr))
E
Eric Dumazet 已提交
195 196 197 198 199 200 201
			return -1;
		score = 0;
		if (inet->inet_dport) {
			if (inet->inet_dport != sport)
				return -1;
			score++;
		}
202 203
		if (!ipv6_addr_any(&sk->sk_v6_daddr)) {
			if (!ipv6_addr_equal(&sk->sk_v6_daddr, saddr))
E
Eric Dumazet 已提交
204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224
				return -1;
			score++;
		}
		if (sk->sk_bound_dev_if) {
			if (sk->sk_bound_dev_if != dif)
				return -1;
			score++;
		}
	}
	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;
225 226
	int score, badness, matches = 0, reuseport = 0;
	u32 hash = 0;
E
Eric Dumazet 已提交
227 228 229 230 231 232 233 234 235 236

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;
237 238
			reuseport = sk->sk_reuseport;
			if (reuseport) {
239 240
				hash = udp6_ehashfn(net, daddr, hnum,
						    saddr, sport);
241 242
				matches = 1;
			} else if (score == SCORE2_MAX)
E
Eric Dumazet 已提交
243
				goto exact_match;
244 245
		} else if (score == badness && reuseport) {
			matches++;
246
			if (reciprocal_scale(hash, matches) == 0)
247 248
				result = sk;
			hash = next_pseudo_random32(hash);
E
Eric Dumazet 已提交
249 250 251 252 253 254 255 256 257 258 259 260
		}
	}
	/*
	 * 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
261
		if (unlikely(!atomic_inc_not_zero_hint(&result->sk_refcnt, 2)))
E
Eric Dumazet 已提交
262 263 264 265 266 267 268 269 270 271
			result = NULL;
		else if (unlikely(compute_score2(result, net, saddr, sport,
				  daddr, hnum, dif) < badness)) {
			sock_put(result);
			goto begin;
		}
	}
	return result;
}

272
struct sock *__udp6_lib_lookup(struct net *net,
273 274
				      const struct in6_addr *saddr, __be16 sport,
				      const struct in6_addr *daddr, __be16 dport,
275
				      int dif, struct udp_table *udptable)
L
Linus Torvalds 已提交
276
{
277
	struct sock *sk, *result;
278
	struct hlist_nulls_node *node;
L
Linus Torvalds 已提交
279
	unsigned short hnum = ntohs(dport);
E
Eric Dumazet 已提交
280 281
	unsigned int hash2, slot2, slot = udp_hashfn(net, hnum, udptable->mask);
	struct udp_hslot *hslot2, *hslot = &udptable->hash[slot];
282 283
	int score, badness, matches = 0, reuseport = 0;
	u32 hash = 0;
284

285
	rcu_read_lock();
E
Eric Dumazet 已提交
286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302
	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;

303 304
			result = udp6_lib_lookup2(net, saddr, sport,
						  &in6addr_any, hnum, dif,
E
Eric Dumazet 已提交
305 306 307 308 309
						  hslot2, slot2);
		}
		rcu_read_unlock();
		return result;
	}
310 311 312
begin:
	result = NULL;
	badness = -1;
313
	sk_nulls_for_each_rcu(sk, node, &hslot->head) {
314 315 316 317
		score = compute_score(sk, net, hnum, saddr, sport, daddr, dport, dif);
		if (score > badness) {
			result = sk;
			badness = score;
318 319
			reuseport = sk->sk_reuseport;
			if (reuseport) {
320 321
				hash = udp6_ehashfn(net, daddr, hnum,
						    saddr, sport);
322 323 324 325
				matches = 1;
			}
		} else if (score == badness && reuseport) {
			matches++;
326
			if (reciprocal_scale(hash, matches) == 0)
327 328
				result = sk;
			hash = next_pseudo_random32(hash);
L
Linus Torvalds 已提交
329 330
		}
	}
331 332 333 334 335
	/*
	 * 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
336
	if (get_nulls_value(node) != slot)
337 338
		goto begin;

339
	if (result) {
E
Eric Dumazet 已提交
340
		if (unlikely(!atomic_inc_not_zero_hint(&result->sk_refcnt, 2)))
341 342 343 344 345 346 347 348
			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
349 350
	return result;
}
351
EXPORT_SYMBOL_GPL(__udp6_lib_lookup);
L
Linus Torvalds 已提交
352

353 354
static struct sock *__udp6_lib_lookup_skb(struct sk_buff *skb,
					  __be16 sport, __be16 dport,
355
					  struct udp_table *udptable)
356
{
357
	struct sock *sk;
358
	const struct ipv6hdr *iph = ipv6_hdr(skb);
359

360 361
	if (unlikely(sk = skb_steal_sock(skb)))
		return sk;
E
Eric Dumazet 已提交
362 363 364
	return __udp6_lib_lookup(dev_net(skb_dst(skb)->dev), &iph->saddr, sport,
				 &iph->daddr, dport, inet6_iif(skb),
				 udptable);
365 366
}

367 368 369 370 371 372 373 374
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
375
/*
376 377
 *	This should be easy, if there is something there we
 *	return it, otherwise we block.
L
Linus Torvalds 已提交
378 379
 */

380
int udpv6_recvmsg(struct kiocb *iocb, struct sock *sk,
L
Linus Torvalds 已提交
381 382 383 384 385
		  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);
386
	struct sk_buff *skb;
387
	unsigned int ulen, copied;
388
	int peeked, off = 0;
389 390
	int err;
	int is_udplite = IS_UDPLITE(sk);
391
	int is_udp4;
392
	bool slow;
L
Linus Torvalds 已提交
393 394

	if (flags & MSG_ERRQUEUE)
395
		return ipv6_recv_error(sk, msg, len, addr_len);
L
Linus Torvalds 已提交
396

397
	if (np->rxpmtu && np->rxopt.bits.rxpmtu)
398
		return ipv6_recv_rxpmtu(sk, msg, len, addr_len);
399

L
Linus Torvalds 已提交
400
try_again:
401
	skb = __skb_recv_datagram(sk, flags | (noblock ? MSG_DONTWAIT : 0),
402
				  &peeked, &off, &err);
L
Linus Torvalds 已提交
403 404 405
	if (!skb)
		goto out;

406
	ulen = skb->len - sizeof(struct udphdr);
407 408 409 410
	copied = len;
	if (copied > ulen)
		copied = ulen;
	else if (copied < ulen)
411
		msg->msg_flags |= MSG_TRUNC;
L
Linus Torvalds 已提交
412

413 414
	is_udp4 = (skb->protocol == htons(ETH_P_IP));

415
	/*
416 417 418
	 * 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.
419 420
	 */

421
	if (copied < ulen || UDP_SKB_CB(skb)->partial_cov) {
422
		if (udp_lib_checksum_complete(skb))
L
Linus Torvalds 已提交
423
			goto csum_copy_err;
424 425
	}

426
	if (skb_csum_unnecessary(skb))
427
		err = skb_copy_datagram_iovec(skb, sizeof(struct udphdr),
428
					      msg->msg_iov, copied);
429
	else {
L
Linus Torvalds 已提交
430 431 432 433
		err = skb_copy_and_csum_datagram_iovec(skb, sizeof(struct udphdr), msg->msg_iov);
		if (err == -EINVAL)
			goto csum_copy_err;
	}
434 435
	if (unlikely(err)) {
		trace_kfree_skb(skb, udpv6_recvmsg);
436 437 438 439 440 441 442 443 444 445 446
		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
447
		goto out_free;
448
	}
449 450 451 452 453 454 455 456
	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);
	}
457

458
	sock_recv_ts_and_drops(msg, sk, skb);
L
Linus Torvalds 已提交
459 460 461

	/* Copy the address. */
	if (msg->msg_name) {
462
		DECLARE_SOCKADDR(struct sockaddr_in6 *, sin6, msg->msg_name);
L
Linus Torvalds 已提交
463
		sin6->sin6_family = AF_INET6;
464
		sin6->sin6_port = udp_hdr(skb)->source;
L
Linus Torvalds 已提交
465 466
		sin6->sin6_flowinfo = 0;

467
		if (is_udp4) {
B
Brian Haley 已提交
468 469
			ipv6_addr_set_v4mapped(ip_hdr(skb)->saddr,
					       &sin6->sin6_addr);
470 471
			sin6->sin6_scope_id = 0;
		} else {
A
Alexey Dobriyan 已提交
472
			sin6->sin6_addr = ipv6_hdr(skb)->saddr;
473 474
			sin6->sin6_scope_id =
				ipv6_iface_scope_id(&sin6->sin6_addr,
475
						    inet6_iif(skb));
L
Linus Torvalds 已提交
476
		}
477
		*addr_len = sizeof(*sin6);
L
Linus Torvalds 已提交
478
	}
479 480 481 482

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

483
	if (is_udp4) {
L
Linus Torvalds 已提交
484 485 486 487
		if (inet->cmsg_flags)
			ip_cmsg_recv(msg, skb);
	} else {
		if (np->rxopt.all)
488
			ip6_datagram_recv_specific_ctl(sk, msg, skb);
489
	}
L
Linus Torvalds 已提交
490

491
	err = copied;
L
Linus Torvalds 已提交
492
	if (flags & MSG_TRUNC)
493
		err = ulen;
L
Linus Torvalds 已提交
494 495

out_free:
496
	skb_free_datagram_locked(sk, skb);
L
Linus Torvalds 已提交
497 498 499 500
out:
	return err;

csum_copy_err:
501
	slow = lock_sock_fast(sk);
502
	if (!skb_kill_datagram(sk, skb, flags)) {
503 504 505
		if (is_udp4) {
			UDP_INC_STATS_USER(sock_net(sk),
					UDP_MIB_CSUMERRORS, is_udplite);
506 507
			UDP_INC_STATS_USER(sock_net(sk),
					UDP_MIB_INERRORS, is_udplite);
508 509 510
		} else {
			UDP6_INC_STATS_USER(sock_net(sk),
					UDP_MIB_CSUMERRORS, is_udplite);
511 512
			UDP6_INC_STATS_USER(sock_net(sk),
					UDP_MIB_INERRORS, is_udplite);
513
		}
514
	}
515
	unlock_sock_fast(sk, slow);
L
Linus Torvalds 已提交
516

517
	if (noblock)
L
Linus Torvalds 已提交
518
		return -EAGAIN;
519 520 521

	/* starting over for a new packet */
	msg->msg_flags &= ~MSG_TRUNC;
L
Linus Torvalds 已提交
522 523 524
	goto try_again;
}

525
void __udp6_lib_err(struct sk_buff *skb, struct inet6_skb_parm *opt,
526
		    u8 type, u8 code, int offset, __be32 info,
527
		    struct udp_table *udptable)
L
Linus Torvalds 已提交
528 529
{
	struct ipv6_pinfo *np;
530 531 532
	const struct ipv6hdr *hdr = (const struct ipv6hdr *)skb->data;
	const struct in6_addr *saddr = &hdr->saddr;
	const struct in6_addr *daddr = &hdr->daddr;
533
	struct udphdr *uh = (struct udphdr *)(skb->data+offset);
L
Linus Torvalds 已提交
534 535
	struct sock *sk;
	int err;
536
	struct net *net = dev_net(skb->dev);
L
Linus Torvalds 已提交
537

538
	sk = __udp6_lib_lookup(net, daddr, uh->dest,
539
			       saddr, uh->source, inet6_iif(skb), udptable);
540 541 542
	if (sk == NULL) {
		ICMP6_INC_STATS_BH(net, __in6_dev_get(skb->dev),
				   ICMP6_MIB_INERRORS);
L
Linus Torvalds 已提交
543
		return;
544
	}
L
Linus Torvalds 已提交
545

546 547 548
	if (type == ICMPV6_PKT_TOOBIG) {
		if (!ip6_sk_accept_pmtu(sk))
			goto out;
549
		ip6_sk_update_pmtu(skb, sk, info);
550
	}
551
	if (type == NDISC_REDIRECT) {
552
		ip6_sk_redirect(skb, sk);
553 554
		goto out;
	}
555

L
Linus Torvalds 已提交
556 557 558 559 560 561 562 563
	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;

564
	if (np->recverr)
L
Linus Torvalds 已提交
565
		ipv6_icmp_error(sk, skb, err, uh->dest, ntohl(info), (u8 *)(uh+1));
566

L
Linus Torvalds 已提交
567 568 569 570 571 572
	sk->sk_err = err;
	sk->sk_error_report(sk);
out:
	sock_put(sk);
}

573 574 575 576
static int __udpv6_queue_rcv_skb(struct sock *sk, struct sk_buff *skb)
{
	int rc;

577
	if (!ipv6_addr_any(&sk->sk_v6_daddr)) {
578
		sock_rps_save_rxhash(sk, skb);
579 580
		sk_mark_napi_id(sk, skb);
	}
581 582 583 584 585 586 587 588 589 590 591 592 593 594 595 596

	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;
}

597
static __inline__ void udpv6_err(struct sk_buff *skb,
598
				 struct inet6_skb_parm *opt, u8 type,
599
				 u8 code, int offset, __be32 info)
600
{
601
	__udp6_lib_err(skb, opt, type, code, offset, info, &udp_table);
602 603
}

604 605 606 607 608 609 610 611
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);

612
int udpv6_queue_rcv_skb(struct sock *sk, struct sk_buff *skb)
L
Linus Torvalds 已提交
613
{
614
	struct udp_sock *up = udp_sk(sk);
615
	int rc;
W
Wang Chen 已提交
616
	int is_udplite = IS_UDPLITE(sk);
617

618 619
	if (!xfrm6_policy_check(sk, XFRM_POLICY_IN, skb))
		goto drop;
L
Linus Torvalds 已提交
620

621 622 623 624 625 626 627 628 629 630 631 632 633 634 635 636 637 638 639
	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;

640 641 642 643
			/* Verify checksum before giving to encap */
			if (udp_lib_checksum_complete(skb))
				goto csum_error;

644 645 646 647 648 649 650 651 652 653 654 655
			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 */
	}

656 657 658
	/*
	 * UDP-Lite specific tests, ignored on UDP sockets (see net/ipv4/udp.c).
	 */
W
Wang Chen 已提交
659
	if ((is_udplite & UDPLITE_RECV_CC)  &&  UDP_SKB_CB(skb)->partial_cov) {
660 661 662 663 664 665 666 667 668 669 670 671 672

		if (up->pcrlen == 0) {          /* full coverage was set  */
			LIMIT_NETDEBUG(KERN_WARNING "UDPLITE6: partial coverage"
				" %d while full coverage %d requested\n",
				UDP_SKB_CB(skb)->cscov, skb->len);
			goto drop;
		}
		if (UDP_SKB_CB(skb)->cscov  <  up->pcrlen) {
			LIMIT_NETDEBUG(KERN_WARNING "UDPLITE6: coverage %d "
						    "too small, need min %d\n",
				       UDP_SKB_CB(skb)->cscov, up->pcrlen);
			goto drop;
		}
L
Linus Torvalds 已提交
673 674
	}

675
	if (rcu_access_pointer(sk->sk_filter)) {
676
		if (udp_lib_checksum_complete(skb))
677
			goto csum_error;
678
	}
679

680
	if (sk_rcvqueues_full(sk, sk->sk_rcvbuf)) {
681 682
		UDP6_INC_STATS_BH(sock_net(sk),
				  UDP_MIB_RCVBUFERRORS, is_udplite);
683
		goto drop;
684
	}
685

686
	skb_dst_drop(skb);
687

688 689 690 691 692 693 694 695 696
	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);
697 698

	return rc;
699

700 701
csum_error:
	UDP6_INC_STATS_BH(sock_net(sk), UDP_MIB_CSUMERRORS, is_udplite);
702
drop:
703
	UDP6_INC_STATS_BH(sock_net(sk), UDP_MIB_INERRORS, is_udplite);
704
	atomic_inc(&sk->sk_drops);
705 706
	kfree_skb(skb);
	return -1;
L
Linus Torvalds 已提交
707 708
}

709 710 711 712
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
713
{
714
	struct inet_sock *inet = inet_sk(sk);
L
Linus Torvalds 已提交
715

716 717 718 719 720 721 722 723 724 725 726 727 728
	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
729 730
}

731 732 733
static void flush_stack(struct sock **stack, unsigned int count,
			struct sk_buff *skb, unsigned int final)
{
734
	struct sk_buff *skb1 = NULL;
735
	struct sock *sk;
736
	unsigned int i;
737 738

	for (i = 0; i < count; i++) {
739
		sk = stack[i];
740 741 742 743 744 745 746 747
		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));
748
		}
749 750 751

		if (skb1 && udpv6_queue_rcv_skb(sk, skb1) <= 0)
			skb1 = NULL;
752
		sock_put(sk);
753
	}
754 755
	if (unlikely(skb1))
		kfree_skb(skb1);
756
}
757 758 759 760 761 762 763 764 765 766 767

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.
	 */
	LIMIT_NETDEBUG(KERN_INFO "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));
}

L
Linus Torvalds 已提交
768 769 770 771
/*
 * Note: called only from the BH handler context,
 * so we don't need to lock the hashes.
 */
772
static int __udp6_lib_mcast_deliver(struct net *net, struct sk_buff *skb,
773
		const struct in6_addr *saddr, const struct in6_addr *daddr,
774
		struct udp_table *udptable)
L
Linus Torvalds 已提交
775
{
776
	struct sock *sk, *stack[256 / sizeof(struct sock *)];
777
	const struct udphdr *uh = udp_hdr(skb);
778 779 780 781
	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);
782 783 784 785 786 787 788 789 790 791 792
	unsigned int count = 0, offset = offsetof(typeof(*sk), sk_nulls_node);
	unsigned int hash2 = 0, hash2_any = 0, use_hash2 = (hslot->count > 10);

	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
793

794
	spin_lock(&hslot->lock);
795
	sk_nulls_for_each_entry_offset(sk, node, &hslot->head, offset) {
796 797 798 799 800 801 802 803 804 805 806 807
		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);
				count = 0;
			}
808
			stack[count++] = sk;
809
			sock_hold(sk);
H
Hideo Aoki 已提交
810
		}
L
Linus Torvalds 已提交
811
	}
812

813
	spin_unlock(&hslot->lock);
814

815 816 817 818 819 820
	/* 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;
	}

821 822 823 824 825
	if (count) {
		flush_stack(stack, count, skb, count - 1);
	} else {
		kfree_skb(skb);
	}
826
	return 0;
L
Linus Torvalds 已提交
827 828
}

829
int __udp6_lib_rcv(struct sk_buff *skb, struct udp_table *udptable,
830
		   int proto)
L
Linus Torvalds 已提交
831
{
832
	struct net *net = dev_net(skb->dev);
L
Linus Torvalds 已提交
833
	struct sock *sk;
834
	struct udphdr *uh;
835
	const struct in6_addr *saddr, *daddr;
L
Linus Torvalds 已提交
836 837 838
	u32 ulen = 0;

	if (!pskb_may_pull(skb, sizeof(struct udphdr)))
839
		goto discard;
L
Linus Torvalds 已提交
840

841 842
	saddr = &ipv6_hdr(skb)->saddr;
	daddr = &ipv6_hdr(skb)->daddr;
843
	uh = udp_hdr(skb);
L
Linus Torvalds 已提交
844 845

	ulen = ntohs(uh->len);
846 847
	if (ulen > skb->len)
		goto short_packet;
L
Linus Torvalds 已提交
848

849 850
	if (proto == IPPROTO_UDP) {
		/* UDP validates ulen. */
L
Linus Torvalds 已提交
851

852 853 854
		/* Check for jumbo payload */
		if (ulen == 0)
			ulen = skb->len;
L
Linus Torvalds 已提交
855

856 857
		if (ulen < sizeof(*uh))
			goto short_packet;
L
Linus Torvalds 已提交
858

859 860 861
		if (ulen < skb->len) {
			if (pskb_trim_rcsum(skb, ulen))
				goto short_packet;
862 863
			saddr = &ipv6_hdr(skb)->saddr;
			daddr = &ipv6_hdr(skb)->daddr;
864
			uh = udp_hdr(skb);
865 866
		}
	}
L
Linus Torvalds 已提交
867

868
	if (udp6_csum_init(skb, uh, proto))
869
		goto csum_error;
870

871 872
	/*
	 *	Multicast receive code
L
Linus Torvalds 已提交
873
	 */
874
	if (ipv6_addr_is_multicast(daddr))
875 876
		return __udp6_lib_mcast_deliver(net, skb,
				saddr, daddr, udptable);
L
Linus Torvalds 已提交
877 878 879

	/* Unicast */

880
	/*
L
Linus Torvalds 已提交
881 882 883
	 * check socket cache ... must talk to Alan about his plans
	 * for sock caches... i'll skip this for now.
	 */
884
	sk = __udp6_lib_lookup_skb(skb, uh->source, uh->dest, udptable);
885
	if (sk != NULL) {
886 887
		int ret;

888
		if (!uh->check && !udp_sk(sk)->no_check6_rx) {
889
			sock_put(sk);
890 891 892 893
			udp6_csum_zero_error(skb);
			goto csum_error;
		}

894
		ret = udpv6_queue_rcv_skb(sk, skb);
895
		sock_put(sk);
L
Linus Torvalds 已提交
896

897 898 899 900 901
		/* 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
902

S
Stephen Hemminger 已提交
903
		return 0;
L
Linus Torvalds 已提交
904
	}
905

906 907 908 909 910
	if (!uh->check) {
		udp6_csum_zero_error(skb);
		goto csum_error;
	}

911
	if (!xfrm6_policy_check(NULL, XFRM_POLICY_IN, skb))
912
		goto discard;
913 914

	if (udp_lib_checksum_complete(skb))
915
		goto csum_error;
916 917 918 919 920

	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
921
	return 0;
L
Linus Torvalds 已提交
922

923
short_packet:
924
	LIMIT_NETDEBUG(KERN_DEBUG "UDP%sv6: short packet: From [%pI6c]:%u %d/%d to [%pI6c]:%u\n",
925
		       proto == IPPROTO_UDPLITE ? "-Lite" : "",
926 927 928 929 930 931
		       saddr,
		       ntohs(uh->source),
		       ulen,
		       skb->len,
		       daddr,
		       ntohs(uh->dest));
932 933 934
	goto discard;
csum_error:
	UDP6_INC_STATS_BH(net, UDP_MIB_CSUMERRORS, proto == IPPROTO_UDPLITE);
L
Linus Torvalds 已提交
935
discard:
936
	UDP6_INC_STATS_BH(net, UDP_MIB_INERRORS, proto == IPPROTO_UDPLITE);
L
Linus Torvalds 已提交
937
	kfree_skb(skb);
S
Stephen Hemminger 已提交
938
	return 0;
L
Linus Torvalds 已提交
939
}
940

941
static __inline__ int udpv6_rcv(struct sk_buff *skb)
942
{
943
	return __udp6_lib_rcv(skb, &udp_table, IPPROTO_UDP);
944 945
}

L
Linus Torvalds 已提交
946 947 948 949 950 951 952
/*
 * 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);

953 954 955
	if (up->pending == AF_INET)
		udp_flush_pending_frames(sk);
	else if (up->pending) {
L
Linus Torvalds 已提交
956 957 958
		up->len = 0;
		up->pending = 0;
		ip6_flush_pending_frames(sk);
959
	}
L
Linus Torvalds 已提交
960 961
}

962
/**
963 964 965 966
 *	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)
967 968 969 970 971 972 973 974 975 976 977 978 979 980 981 982 983 984 985 986 987 988 989 990 991 992 993 994 995 996 997 998 999 1000 1001 1002
 */
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
1003 1004 1005 1006
/*
 *	Sending
 */

1007
static int udp_v6_push_pending_frames(struct sock *sk)
L
Linus Torvalds 已提交
1008 1009 1010
{
	struct sk_buff *skb;
	struct udphdr *uh;
1011
	struct udp_sock  *up = udp_sk(sk);
L
Linus Torvalds 已提交
1012
	struct inet_sock *inet = inet_sk(sk);
1013
	struct flowi6 *fl6;
L
Linus Torvalds 已提交
1014
	int err = 0;
W
Wang Chen 已提交
1015
	int is_udplite = IS_UDPLITE(sk);
1016
	__wsum csum = 0;
L
Linus Torvalds 已提交
1017

1018 1019 1020 1021 1022
	if (up->pending == AF_INET)
		return udp_push_pending_frames(sk);

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

L
Linus Torvalds 已提交
1023 1024 1025 1026 1027 1028 1029
	/* Grab the skbuff where UDP header space exists. */
	if ((skb = skb_peek(&sk->sk_write_queue)) == NULL)
		goto out;

	/*
	 * Create a UDP header
	 */
1030
	uh = udp_hdr(skb);
1031 1032
	uh->source = fl6->fl6_sport;
	uh->dest = fl6->fl6_dport;
L
Linus Torvalds 已提交
1033 1034 1035
	uh->len = htons(up->len);
	uh->check = 0;

W
Wang Chen 已提交
1036
	if (is_udplite)
1037
		csum = udplite_csum_outgoing(sk, skb);
1038
	else if (up->no_check6_tx) {   /* UDP csum disabled */
1039 1040 1041
		skb->ip_summed = CHECKSUM_NONE;
		goto send;
	} else if (skb->ip_summed == CHECKSUM_PARTIAL) { /* UDP hardware csum */
1042
		udp6_hwcsum_outgoing(sk, skb, &fl6->saddr, &fl6->daddr,
1043 1044 1045
				     up->len);
		goto send;
	} else
1046
		csum = udp_csum_outgoing(sk, skb);
L
Linus Torvalds 已提交
1047

1048
	/* add protocol-dependent pseudo-header */
1049 1050
	uh->check = csum_ipv6_magic(&fl6->saddr, &fl6->daddr,
				    up->len, fl6->flowi6_proto, csum);
L
Linus Torvalds 已提交
1051
	if (uh->check == 0)
1052
		uh->check = CSUM_MANGLED_0;
L
Linus Torvalds 已提交
1053

1054
send:
L
Linus Torvalds 已提交
1055
	err = ip6_push_pending_frames(sk);
E
Eric Dumazet 已提交
1056 1057 1058 1059 1060 1061 1062 1063 1064
	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
1065 1066 1067 1068 1069 1070
out:
	up->len = 0;
	up->pending = 0;
	return err;
}

1071
int udpv6_sendmsg(struct kiocb *iocb, struct sock *sk,
L
Linus Torvalds 已提交
1072 1073 1074 1075 1076 1077
		  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);
1078
	DECLARE_SOCKADDR(struct sockaddr_in6 *, sin6, msg->msg_name);
1079
	struct in6_addr *daddr, *final_p, final;
L
Linus Torvalds 已提交
1080 1081
	struct ipv6_txoptions *opt = NULL;
	struct ip6_flowlabel *flowlabel = NULL;
1082
	struct flowi6 fl6;
L
Linus Torvalds 已提交
1083 1084 1085 1086
	struct dst_entry *dst;
	int addr_len = msg->msg_namelen;
	int ulen = len;
	int hlimit = -1;
1087
	int tclass = -1;
1088
	int dontfrag = -1;
L
Linus Torvalds 已提交
1089 1090
	int corkreq = up->corkflag || msg->msg_flags&MSG_MORE;
	int err;
1091
	int connected = 0;
W
Wang Chen 已提交
1092
	int is_udplite = IS_UDPLITE(sk);
1093
	int (*getfrag)(void *, char *, int, int, int, struct sk_buff *);
L
Linus Torvalds 已提交
1094 1095 1096 1097 1098 1099 1100 1101 1102 1103 1104 1105 1106 1107 1108 1109 1110 1111 1112 1113 1114 1115 1116 1117 1118

	/* 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;
1119
		daddr = &sk->sk_v6_daddr;
1120
	} else
L
Linus Torvalds 已提交
1121 1122 1123
		daddr = NULL;

	if (daddr) {
B
Brian Haley 已提交
1124
		if (ipv6_addr_v4mapped(daddr)) {
L
Linus Torvalds 已提交
1125 1126
			struct sockaddr_in sin;
			sin.sin_family = AF_INET;
E
Eric Dumazet 已提交
1127
			sin.sin_port = sin6 ? sin6->sin6_port : inet->inet_dport;
L
Linus Torvalds 已提交
1128 1129 1130 1131 1132 1133 1134 1135 1136 1137 1138 1139 1140 1141
			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,
1142
	   better check is made in ip6_append_data().
L
Linus Torvalds 已提交
1143 1144 1145
	   */
	if (len > INT_MAX - sizeof(struct udphdr))
		return -EMSGSIZE;
1146

L
Linus Torvalds 已提交
1147 1148 1149 1150 1151 1152 1153 1154 1155 1156 1157 1158 1159 1160 1161 1162 1163 1164
	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);

1165
	memset(&fl6, 0, sizeof(fl6));
L
Linus Torvalds 已提交
1166 1167 1168 1169 1170

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

1171
		fl6.fl6_dport = sin6->sin6_port;
L
Linus Torvalds 已提交
1172 1173 1174
		daddr = &sin6->sin6_addr;

		if (np->sndflow) {
1175 1176 1177
			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
1178 1179 1180 1181 1182 1183 1184 1185 1186 1187
				if (flowlabel == NULL)
					return -EINVAL;
			}
		}

		/*
		 * Otherwise it will be difficult to maintain
		 * sk->sk_dst_cache.
		 */
		if (sk->sk_state == TCP_ESTABLISHED &&
1188 1189
		    ipv6_addr_equal(daddr, &sk->sk_v6_daddr))
			daddr = &sk->sk_v6_daddr;
L
Linus Torvalds 已提交
1190 1191 1192

		if (addr_len >= sizeof(struct sockaddr_in6) &&
		    sin6->sin6_scope_id &&
1193
		    __ipv6_addr_needs_scope_id(__ipv6_addr_type(daddr)))
1194
			fl6.flowi6_oif = sin6->sin6_scope_id;
L
Linus Torvalds 已提交
1195 1196 1197 1198
	} else {
		if (sk->sk_state != TCP_ESTABLISHED)
			return -EDESTADDRREQ;

1199
		fl6.fl6_dport = inet->inet_dport;
1200
		daddr = &sk->sk_v6_daddr;
1201
		fl6.flowlabel = np->flow_label;
1202
		connected = 1;
L
Linus Torvalds 已提交
1203 1204
	}

1205 1206
	if (!fl6.flowi6_oif)
		fl6.flowi6_oif = sk->sk_bound_dev_if;
L
Linus Torvalds 已提交
1207

1208 1209
	if (!fl6.flowi6_oif)
		fl6.flowi6_oif = np->sticky_pktinfo.ipi6_ifindex;
1210

1211
	fl6.flowi6_mark = sk->sk_mark;
B
Brian Haley 已提交
1212

L
Linus Torvalds 已提交
1213 1214 1215 1216 1217
	if (msg->msg_controllen) {
		opt = &opt_space;
		memset(opt, 0, sizeof(struct ipv6_txoptions));
		opt->tot_len = sizeof(*opt);

1218 1219
		err = ip6_datagram_send_ctl(sock_net(sk), sk, msg, &fl6, opt,
					    &hlimit, &tclass, &dontfrag);
L
Linus Torvalds 已提交
1220 1221 1222 1223
		if (err < 0) {
			fl6_sock_release(flowlabel);
			return err;
		}
1224 1225
		if ((fl6.flowlabel&IPV6_FLOWLABEL_MASK) && !flowlabel) {
			flowlabel = fl6_sock_lookup(sk, fl6.flowlabel);
L
Linus Torvalds 已提交
1226 1227 1228 1229 1230
			if (flowlabel == NULL)
				return -EINVAL;
		}
		if (!(opt->opt_nflen|opt->opt_flen))
			opt = NULL;
1231
		connected = 0;
L
Linus Torvalds 已提交
1232 1233 1234
	}
	if (opt == NULL)
		opt = np->opt;
1235 1236 1237
	if (flowlabel)
		opt = fl6_merge_options(&opt_space, flowlabel, opt);
	opt = ipv6_fixup_options(&opt_space, opt);
L
Linus Torvalds 已提交
1238

1239
	fl6.flowi6_proto = sk->sk_protocol;
1240
	if (!ipv6_addr_any(daddr))
A
Alexey Dobriyan 已提交
1241
		fl6.daddr = *daddr;
1242
	else
1243 1244
		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
1245
		fl6.saddr = np->saddr;
1246
	fl6.fl6_sport = inet->inet_sport;
1247

1248
	final_p = fl6_update_dst(&fl6, opt, &final);
1249
	if (final_p)
1250
		connected = 0;
L
Linus Torvalds 已提交
1251

1252 1253
	if (!fl6.flowi6_oif && ipv6_addr_is_multicast(&fl6.daddr)) {
		fl6.flowi6_oif = np->mcast_oif;
1254
		connected = 0;
1255 1256
	} else if (!fl6.flowi6_oif)
		fl6.flowi6_oif = np->ucast_oif;
L
Linus Torvalds 已提交
1257

1258
	security_sk_classify_flow(sk, flowi6_to_flowi(&fl6));
V
Venkat Yekkirala 已提交
1259

1260
	dst = ip6_sk_dst_lookup_flow(sk, &fl6, final_p);
1261 1262 1263
	if (IS_ERR(dst)) {
		err = PTR_ERR(dst);
		dst = NULL;
L
Linus Torvalds 已提交
1264
		goto out;
1265
	}
L
Linus Torvalds 已提交
1266

1267 1268
	if (hlimit < 0)
		hlimit = ip6_sk_dst_hoplimit(np, &fl6, dst);
L
Linus Torvalds 已提交
1269

G
Gerrit Renker 已提交
1270
	if (tclass < 0)
1271 1272
		tclass = np->tclass;

L
Linus Torvalds 已提交
1273 1274 1275 1276 1277 1278 1279 1280 1281 1282
	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);

1283
		LIMIT_NETDEBUG(KERN_DEBUG "udp cork app bug 2\n");
L
Linus Torvalds 已提交
1284 1285 1286 1287 1288 1289 1290
		err = -EINVAL;
		goto out;
	}

	up->pending = AF_INET6;

do_append_data:
1291 1292
	if (dontfrag < 0)
		dontfrag = np->dontfrag;
L
Linus Torvalds 已提交
1293
	up->len += ulen;
1294 1295
	getfrag  =  is_udplite ?  udplite_getfrag : ip_generic_getfrag;
	err = ip6_append_data(sk, getfrag, msg->msg_iov, ulen,
1296
		sizeof(struct udphdr), hlimit, tclass, opt, &fl6,
1297
		(struct rt6_info *)dst,
1298
		corkreq ? msg->msg_flags|MSG_MORE : msg->msg_flags, dontfrag);
L
Linus Torvalds 已提交
1299 1300 1301
	if (err)
		udp_v6_flush_pending_frames(sk);
	else if (!corkreq)
1302
		err = udp_v6_push_pending_frames(sk);
H
Herbert Xu 已提交
1303 1304
	else if (unlikely(skb_queue_empty(&sk->sk_write_queue)))
		up->pending = 0;
L
Linus Torvalds 已提交
1305

1306 1307 1308
	if (dst) {
		if (connected) {
			ip6_dst_store(sk, dst,
1309 1310
				      ipv6_addr_equal(&fl6.daddr, &sk->sk_v6_daddr) ?
				      &sk->sk_v6_daddr : NULL,
1311
#ifdef CONFIG_IPV6_SUBTREES
1312
				      ipv6_addr_equal(&fl6.saddr, &np->saddr) ?
1313 1314 1315
				      &np->saddr :
#endif
				      NULL);
1316 1317 1318
		} else {
			dst_release(dst);
		}
1319
		dst = NULL;
1320 1321
	}

L
Linus Torvalds 已提交
1322 1323 1324 1325
	if (err > 0)
		err = np->recverr ? net_xmit_errno(err) : 0;
	release_sock(sk);
out:
1326
	dst_release(dst);
L
Linus Torvalds 已提交
1327
	fl6_sock_release(flowlabel);
1328
	if (!err)
L
Linus Torvalds 已提交
1329
		return len;
1330 1331 1332 1333 1334 1335 1336 1337
	/*
	 * 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)) {
1338 1339
		UDP6_INC_STATS_USER(sock_net(sk),
				UDP_MIB_SNDBUFERRORS, is_udplite);
1340
	}
L
Linus Torvalds 已提交
1341 1342 1343 1344 1345 1346 1347 1348 1349 1350
	return err;

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

1351
void udpv6_destroy_sock(struct sock *sk)
L
Linus Torvalds 已提交
1352
{
T
Tom Parkin 已提交
1353
	struct udp_sock *up = udp_sk(sk);
L
Linus Torvalds 已提交
1354 1355 1356 1357
	lock_sock(sk);
	udp_v6_flush_pending_frames(sk);
	release_sock(sk);

T
Tom Parkin 已提交
1358 1359 1360 1361 1362 1363 1364
	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
1365 1366 1367 1368 1369 1370
	inet6_destroy_sock(sk);
}

/*
 *	Socket option code for UDP
 */
1371
int udpv6_setsockopt(struct sock *sk, int level, int optname,
1372
		     char __user *optval, unsigned int optlen)
1373
{
1374
	if (level == SOL_UDP  ||  level == SOL_UDPLITE)
1375 1376
		return udp_lib_setsockopt(sk, level, optname, optval, optlen,
					  udp_v6_push_pending_frames);
1377
	return ipv6_setsockopt(sk, level, optname, optval, optlen);
1378 1379 1380
}

#ifdef CONFIG_COMPAT
1381
int compat_udpv6_setsockopt(struct sock *sk, int level, int optname,
1382
			    char __user *optval, unsigned int optlen)
1383
{
1384
	if (level == SOL_UDP  ||  level == SOL_UDPLITE)
1385 1386
		return udp_lib_setsockopt(sk, level, optname, optval, optlen,
					  udp_v6_push_pending_frames);
1387
	return compat_ipv6_setsockopt(sk, level, optname, optval, optlen);
1388 1389 1390
}
#endif

1391 1392
int udpv6_getsockopt(struct sock *sk, int level, int optname,
		     char __user *optval, int __user *optlen)
1393
{
1394
	if (level == SOL_UDP  ||  level == SOL_UDPLITE)
1395
		return udp_lib_getsockopt(sk, level, optname, optval, optlen);
1396
	return ipv6_getsockopt(sk, level, optname, optval, optlen);
1397 1398 1399
}

#ifdef CONFIG_COMPAT
1400 1401
int compat_udpv6_getsockopt(struct sock *sk, int level, int optname,
			    char __user *optval, int __user *optlen)
1402
{
1403
	if (level == SOL_UDP  ||  level == SOL_UDPLITE)
1404
		return udp_lib_getsockopt(sk, level, optname, optval, optlen);
1405
	return compat_ipv6_getsockopt(sk, level, optname, optval, optlen);
1406 1407 1408
}
#endif

1409
static const struct inet6_protocol udpv6_protocol = {
L
Linus Torvalds 已提交
1410 1411 1412 1413 1414 1415 1416
	.handler	=	udpv6_rcv,
	.err_handler	=	udpv6_err,
	.flags		=	INET6_PROTO_NOPOLICY|INET6_PROTO_FINAL,
};

/* ------------------------------------------------------------------------ */
#ifdef CONFIG_PROC_FS
1417
int udp6_seq_show(struct seq_file *seq, void *v)
L
Linus Torvalds 已提交
1418
{
1419 1420 1421 1422 1423 1424 1425 1426 1427
	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
1428 1429 1430
	return 0;
}

1431 1432 1433 1434 1435 1436 1437 1438
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
1439 1440 1441
static struct udp_seq_afinfo udp6_seq_afinfo = {
	.name		= "udp6",
	.family		= AF_INET6,
1442
	.udp_table	= &udp_table,
1443
	.seq_fops	= &udp6_afinfo_seq_fops,
1444 1445 1446
	.seq_ops	= {
		.show		= udp6_seq_show,
	},
L
Linus Torvalds 已提交
1447 1448
};

1449
int __net_init udp6_proc_init(struct net *net)
L
Linus Torvalds 已提交
1450
{
1451
	return udp_proc_register(net, &udp6_seq_afinfo);
L
Linus Torvalds 已提交
1452 1453
}

1454 1455
void udp6_proc_exit(struct net *net) {
	udp_proc_unregister(net, &udp6_seq_afinfo);
L
Linus Torvalds 已提交
1456 1457 1458
}
#endif /* CONFIG_PROC_FS */

E
Eric Dumazet 已提交
1459 1460 1461 1462 1463 1464 1465 1466 1467 1468 1469
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
1470 1471 1472
/* ------------------------------------------------------------------------ */

struct proto udpv6_prot = {
1473 1474
	.name		   = "UDPv6",
	.owner		   = THIS_MODULE,
1475
	.close		   = udp_lib_close,
1476 1477 1478 1479 1480 1481 1482 1483
	.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,
1484
	.backlog_rcv	   = __udpv6_queue_rcv_skb,
1485 1486
	.hash		   = udp_lib_hash,
	.unhash		   = udp_lib_unhash,
E
Eric Dumazet 已提交
1487
	.rehash		   = udp_v6_rehash,
1488
	.get_port	   = udp_v6_get_port,
H
Hideo Aoki 已提交
1489 1490 1491 1492
	.memory_allocated  = &udp_memory_allocated,
	.sysctl_mem	   = sysctl_udp_mem,
	.sysctl_wmem	   = &sysctl_udp_wmem_min,
	.sysctl_rmem	   = &sysctl_udp_rmem_min,
1493
	.obj_size	   = sizeof(struct udp6_sock),
1494
	.slab_flags	   = SLAB_DESTROY_BY_RCU,
1495
	.h.udp_table	   = &udp_table,
1496
#ifdef CONFIG_COMPAT
1497 1498
	.compat_setsockopt = compat_udpv6_setsockopt,
	.compat_getsockopt = compat_udpv6_getsockopt,
1499
#endif
E
Eric Dumazet 已提交
1500
	.clear_sk	   = udp_v6_clear_sk,
L
Linus Torvalds 已提交
1501 1502 1503 1504 1505 1506 1507 1508 1509 1510 1511
};

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


1512
int __init udpv6_init(void)
L
Linus Torvalds 已提交
1513
{
1514 1515
	int ret;

1516 1517
	ret = inet6_add_protocol(&udpv6_protocol, IPPROTO_UDP);
	if (ret)
1518
		goto out;
1519

1520 1521 1522 1523 1524 1525 1526 1527 1528 1529 1530
	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;
}

1531
void udpv6_exit(void)
1532 1533 1534
{
	inet6_unregister_protosw(&udpv6_protosw);
	inet6_del_protocol(&udpv6_protocol, IPPROTO_UDP);
L
Linus Torvalds 已提交
1535
}