udp.c 38.1 KB
Newer Older
L
Linus Torvalds 已提交
1 2
/*
 *	UDP over IPv6
3
 *	Linux INET6 implementation
L
Linus Torvalds 已提交
4 5
 *
 *	Authors:
6
 *	Pedro Roque		<roque@di.fc.ul.pt>
L
Linus Torvalds 已提交
7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34
 *
 *	Based on linux/ipv4/udp.c
 *
 *	Fixes:
 *	Hideaki YOSHIFUJI	:	sin6_scope_id support
 *	YOSHIFUJI Hideaki @USAGI and:	Support IPV6_V6ONLY socket option, which
 *	Alexey Kuznetsov		allow both IPv4 and IPv6 sockets to bind
 *					a single port at the same time.
 *      Kazunori MIYAZAWA @USAGI:       change process style to use ip6_append_data
 *      YOSHIFUJI Hideaki @USAGI:	convert /proc/net/udp6 to seq_file.
 *
 *	This program is free software; you can redistribute it and/or
 *      modify it under the terms of the GNU General Public License
 *      as published by the Free Software Foundation; either version
 *      2 of the License, or (at your option) any later version.
 */

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

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

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

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

	u32 lhash, fhash;

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

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

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

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

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

	return hash ^ port;
}

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

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

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

	udp_lib_rehash(sk, new_hash);
}

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

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

	score = 0;
	inet = inet_sk(sk);

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

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

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

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

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

162 163 164
	return score;
}

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

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

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

	score = 0;
	inet = inet_sk(sk);

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

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

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

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

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

/* called with read_rcu_lock() */
static struct sock *udp6_lib_lookup2(struct net *net,
		const struct in6_addr *saddr, __be16 sport,
		const struct in6_addr *daddr, unsigned int hnum, int dif,
212 213
		struct udp_hslot *hslot2, unsigned int slot2,
		struct sk_buff *skb)
E
Eric Dumazet 已提交
214 215
{
	struct sock *sk, *result;
216 217
	int score, badness, matches = 0, reuseport = 0;
	u32 hash = 0;
E
Eric Dumazet 已提交
218 219 220

	result = NULL;
	badness = -1;
221
	udp_portaddr_for_each_entry_rcu(sk, &hslot2->head) {
E
Eric Dumazet 已提交
222 223 224
		score = compute_score2(sk, net, saddr, sport,
				      daddr, hnum, dif);
		if (score > badness) {
225 226
			reuseport = sk->sk_reuseport;
			if (reuseport) {
227 228
				hash = udp6_ehashfn(net, daddr, hnum,
						    saddr, sport);
229

230
				result = reuseport_select_sock(sk, hash, skb,
231
							sizeof(struct udphdr));
232 233
				if (result)
					return result;
234
				matches = 1;
235
			}
236 237
			result = sk;
			badness = score;
238 239
		} else if (score == badness && reuseport) {
			matches++;
240
			if (reciprocal_scale(hash, matches) == 0)
241 242
				result = sk;
			hash = next_pseudo_random32(hash);
E
Eric Dumazet 已提交
243 244 245 246 247
		}
	}
	return result;
}

248
/* rcu_read_lock() must be held */
249
struct sock *__udp6_lib_lookup(struct net *net,
250 251
				      const struct in6_addr *saddr, __be16 sport,
				      const struct in6_addr *daddr, __be16 dport,
252 253
				      int dif, struct udp_table *udptable,
				      struct sk_buff *skb)
L
Linus Torvalds 已提交
254
{
255
	struct sock *sk, *result;
L
Linus Torvalds 已提交
256
	unsigned short hnum = ntohs(dport);
E
Eric Dumazet 已提交
257 258
	unsigned int hash2, slot2, slot = udp_hashfn(net, hnum, udptable->mask);
	struct udp_hslot *hslot2, *hslot = &udptable->hash[slot];
259 260
	int score, badness, matches = 0, reuseport = 0;
	u32 hash = 0;
261

E
Eric Dumazet 已提交
262 263 264 265 266 267 268 269 270
	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,
271
					  hslot2, slot2, skb);
E
Eric Dumazet 已提交
272 273 274 275 276 277 278
		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;

279 280
			result = udp6_lib_lookup2(net, saddr, sport,
						  &in6addr_any, hnum, dif,
281
						  hslot2, slot2, skb);
E
Eric Dumazet 已提交
282 283 284
		}
		return result;
	}
285 286 287
begin:
	result = NULL;
	badness = -1;
288
	sk_for_each_rcu(sk, &hslot->head) {
289 290
		score = compute_score(sk, net, hnum, saddr, sport, daddr, dport, dif);
		if (score > badness) {
291 292
			reuseport = sk->sk_reuseport;
			if (reuseport) {
293 294
				hash = udp6_ehashfn(net, daddr, hnum,
						    saddr, sport);
295
				result = reuseport_select_sock(sk, hash, skb,
296
							sizeof(struct udphdr));
297 298
				if (result)
					return result;
299 300
				matches = 1;
			}
301 302
			result = sk;
			badness = score;
303 304
		} else if (score == badness && reuseport) {
			matches++;
305
			if (reciprocal_scale(hash, matches) == 0)
306 307
				result = sk;
			hash = next_pseudo_random32(hash);
L
Linus Torvalds 已提交
308 309 310 311
		}
	}
	return result;
}
312
EXPORT_SYMBOL_GPL(__udp6_lib_lookup);
L
Linus Torvalds 已提交
313

314 315
static struct sock *__udp6_lib_lookup_skb(struct sk_buff *skb,
					  __be16 sport, __be16 dport,
316
					  struct udp_table *udptable)
317
{
318
	struct sock *sk;
319
	const struct ipv6hdr *iph = ipv6_hdr(skb);
320

321 322
	sk = skb_steal_sock(skb);
	if (unlikely(sk))
323
		return sk;
E
Eric Dumazet 已提交
324 325
	return __udp6_lib_lookup(dev_net(skb_dst(skb)->dev), &iph->saddr, sport,
				 &iph->daddr, dport, inet6_iif(skb),
326
				 udptable, skb);
327 328
}

329 330 331 332 333 334 335 336 337 338 339 340 341
struct sock *udp6_lib_lookup_skb(struct sk_buff *skb,
				 __be16 sport, __be16 dport)
{
	const struct ipv6hdr *iph = ipv6_hdr(skb);
	const struct net_device *dev =
	    skb_dst(skb) ? skb_dst(skb)->dev : skb->dev;

	return __udp6_lib_lookup(dev_net(dev), &iph->saddr, sport,
				 &iph->daddr, dport, inet6_iif(skb),
				 &udp_table, skb);
}
EXPORT_SYMBOL_GPL(udp6_lib_lookup_skb);

342 343 344 345 346
/* Must be called under rcu_read_lock().
 * Does increment socket refcount.
 */
#if IS_ENABLED(CONFIG_NETFILTER_XT_MATCH_SOCKET) || \
    IS_ENABLED(CONFIG_NETFILTER_XT_TARGET_TPROXY)
347 348 349
struct sock *udp6_lib_lookup(struct net *net, const struct in6_addr *saddr, __be16 sport,
			     const struct in6_addr *daddr, __be16 dport, int dif)
{
350 351 352 353 354 355 356
	struct sock *sk;

	sk =  __udp6_lib_lookup(net, saddr, sport, daddr, dport,
				dif, &udp_table, NULL);
	if (sk && !atomic_inc_not_zero(&sk->sk_refcnt))
		sk = NULL;
	return sk;
357 358
}
EXPORT_SYMBOL_GPL(udp6_lib_lookup);
359
#endif
360

L
Linus Torvalds 已提交
361
/*
362 363
 *	This should be easy, if there is something there we
 *	return it, otherwise we block.
L
Linus Torvalds 已提交
364 365
 */

366
int udpv6_recvmsg(struct sock *sk, struct msghdr *msg, size_t len,
L
Linus Torvalds 已提交
367 368 369 370
		  int noblock, int flags, int *addr_len)
{
	struct ipv6_pinfo *np = inet6_sk(sk);
	struct inet_sock *inet = inet_sk(sk);
371
	struct sk_buff *skb;
372
	unsigned int ulen, copied;
373
	int peeked, peeking, off;
374 375
	int err;
	int is_udplite = IS_UDPLITE(sk);
376
	bool checksum_valid = false;
377
	int is_udp4;
378
	bool slow;
L
Linus Torvalds 已提交
379 380

	if (flags & MSG_ERRQUEUE)
381
		return ipv6_recv_error(sk, msg, len, addr_len);
L
Linus Torvalds 已提交
382

383
	if (np->rxpmtu && np->rxopt.bits.rxpmtu)
384
		return ipv6_recv_rxpmtu(sk, msg, len, addr_len);
385

L
Linus Torvalds 已提交
386
try_again:
387
	peeking = off = sk_peek_offset(sk, flags);
388
	skb = __skb_recv_datagram(sk, flags | (noblock ? MSG_DONTWAIT : 0),
389
				  &peeked, &off, &err);
L
Linus Torvalds 已提交
390
	if (!skb)
391
		return err;
L
Linus Torvalds 已提交
392

393
	ulen = skb->len;
394
	copied = len;
395 396
	if (copied > ulen - off)
		copied = ulen - off;
397
	else if (copied < ulen)
398
		msg->msg_flags |= MSG_TRUNC;
L
Linus Torvalds 已提交
399

400 401
	is_udp4 = (skb->protocol == htons(ETH_P_IP));

402
	/*
403 404 405
	 * 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.
406 407
	 */

408
	if (copied < ulen || UDP_SKB_CB(skb)->partial_cov || peeking) {
409 410
		checksum_valid = !udp_lib_checksum_complete(skb);
		if (!checksum_valid)
L
Linus Torvalds 已提交
411
			goto csum_copy_err;
412 413
	}

414
	if (checksum_valid || skb_csum_unnecessary(skb))
415
		err = skb_copy_datagram_msg(skb, off, msg, copied);
416
	else {
417
		err = skb_copy_and_csum_datagram_msg(skb, off, msg);
L
Linus Torvalds 已提交
418 419 420
		if (err == -EINVAL)
			goto csum_copy_err;
	}
421 422
	if (unlikely(err)) {
		trace_kfree_skb(skb, udpv6_recvmsg);
423 424 425
		if (!peeked) {
			atomic_inc(&sk->sk_drops);
			if (is_udp4)
426 427
				UDP_INC_STATS(sock_net(sk), UDP_MIB_INERRORS,
					      is_udplite);
428
			else
429 430
				UDP6_INC_STATS(sock_net(sk), UDP_MIB_INERRORS,
					       is_udplite);
431
		}
432 433
		skb_free_datagram_locked(sk, skb);
		return err;
434
	}
435 436
	if (!peeked) {
		if (is_udp4)
437 438
			UDP_INC_STATS(sock_net(sk), UDP_MIB_INDATAGRAMS,
				      is_udplite);
439
		else
440 441
			UDP6_INC_STATS(sock_net(sk), UDP_MIB_INDATAGRAMS,
				       is_udplite);
442
	}
443

444
	sock_recv_ts_and_drops(msg, sk, skb);
L
Linus Torvalds 已提交
445 446 447

	/* Copy the address. */
	if (msg->msg_name) {
448
		DECLARE_SOCKADDR(struct sockaddr_in6 *, sin6, msg->msg_name);
L
Linus Torvalds 已提交
449
		sin6->sin6_family = AF_INET6;
450
		sin6->sin6_port = udp_hdr(skb)->source;
L
Linus Torvalds 已提交
451 452
		sin6->sin6_flowinfo = 0;

453
		if (is_udp4) {
B
Brian Haley 已提交
454 455
			ipv6_addr_set_v4mapped(ip_hdr(skb)->saddr,
					       &sin6->sin6_addr);
456 457
			sin6->sin6_scope_id = 0;
		} else {
A
Alexey Dobriyan 已提交
458
			sin6->sin6_addr = ipv6_hdr(skb)->saddr;
459 460
			sin6->sin6_scope_id =
				ipv6_iface_scope_id(&sin6->sin6_addr,
461
						    inet6_iif(skb));
L
Linus Torvalds 已提交
462
		}
463
		*addr_len = sizeof(*sin6);
L
Linus Torvalds 已提交
464
	}
465 466 467 468

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

469
	if (is_udp4) {
L
Linus Torvalds 已提交
470 471 472 473
		if (inet->cmsg_flags)
			ip_cmsg_recv(msg, skb);
	} else {
		if (np->rxopt.all)
474
			ip6_datagram_recv_specific_ctl(sk, msg, skb);
475
	}
L
Linus Torvalds 已提交
476

477
	err = copied;
L
Linus Torvalds 已提交
478
	if (flags & MSG_TRUNC)
479
		err = ulen;
L
Linus Torvalds 已提交
480

481
	__skb_free_datagram_locked(sk, skb, peeking ? -err : err);
L
Linus Torvalds 已提交
482 483 484
	return err;

csum_copy_err:
485
	slow = lock_sock_fast(sk);
486
	if (!skb_kill_datagram(sk, skb, flags)) {
487
		if (is_udp4) {
488 489 490 491
			UDP_INC_STATS(sock_net(sk),
				      UDP_MIB_CSUMERRORS, is_udplite);
			UDP_INC_STATS(sock_net(sk),
				      UDP_MIB_INERRORS, is_udplite);
492
		} else {
493 494 495 496
			UDP6_INC_STATS(sock_net(sk),
				       UDP_MIB_CSUMERRORS, is_udplite);
			UDP6_INC_STATS(sock_net(sk),
				       UDP_MIB_INERRORS, is_udplite);
497
		}
498
	}
499
	unlock_sock_fast(sk, slow);
L
Linus Torvalds 已提交
500

501 502
	/* starting over for a new packet, but check if we need to yield */
	cond_resched();
503
	msg->msg_flags &= ~MSG_TRUNC;
L
Linus Torvalds 已提交
504 505 506
	goto try_again;
}

507
void __udp6_lib_err(struct sk_buff *skb, struct inet6_skb_parm *opt,
508
		    u8 type, u8 code, int offset, __be32 info,
509
		    struct udp_table *udptable)
L
Linus Torvalds 已提交
510 511
{
	struct ipv6_pinfo *np;
512 513 514
	const struct ipv6hdr *hdr = (const struct ipv6hdr *)skb->data;
	const struct in6_addr *saddr = &hdr->saddr;
	const struct in6_addr *daddr = &hdr->daddr;
515
	struct udphdr *uh = (struct udphdr *)(skb->data+offset);
L
Linus Torvalds 已提交
516
	struct sock *sk;
517
	int harderr;
L
Linus Torvalds 已提交
518
	int err;
519
	struct net *net = dev_net(skb->dev);
L
Linus Torvalds 已提交
520

521
	sk = __udp6_lib_lookup(net, daddr, uh->dest, saddr, uh->source,
522
			       inet6_iif(skb), udptable, skb);
523
	if (!sk) {
E
Eric Dumazet 已提交
524 525
		__ICMP6_INC_STATS(net, __in6_dev_get(skb->dev),
				  ICMP6_MIB_INERRORS);
L
Linus Torvalds 已提交
526
		return;
527
	}
L
Linus Torvalds 已提交
528

529 530 531
	harderr = icmpv6_err_convert(type, code, &err);
	np = inet6_sk(sk);

532 533 534
	if (type == ICMPV6_PKT_TOOBIG) {
		if (!ip6_sk_accept_pmtu(sk))
			goto out;
535
		ip6_sk_update_pmtu(skb, sk, info);
536 537
		if (np->pmtudisc != IPV6_PMTUDISC_DONT)
			harderr = 1;
538
	}
539
	if (type == NDISC_REDIRECT) {
540
		ip6_sk_redirect(skb, sk);
541 542
		goto out;
	}
543

544 545 546 547
	if (!np->recverr) {
		if (!harderr || sk->sk_state != TCP_ESTABLISHED)
			goto out;
	} else {
L
Linus Torvalds 已提交
548
		ipv6_icmp_error(sk, skb, err, uh->dest, ntohl(info), (u8 *)(uh+1));
549
	}
550

L
Linus Torvalds 已提交
551 552 553
	sk->sk_err = err;
	sk->sk_error_report(sk);
out:
554
	return;
L
Linus Torvalds 已提交
555 556
}

557 558 559 560
static int __udpv6_queue_rcv_skb(struct sock *sk, struct sk_buff *skb)
{
	int rc;

561
	if (!ipv6_addr_any(&sk->sk_v6_daddr)) {
562
		sock_rps_save_rxhash(sk, skb);
563
		sk_mark_napi_id(sk, skb);
E
Eric Dumazet 已提交
564
		sk_incoming_cpu_update(sk);
565
	}
566

567
	rc = __sock_queue_rcv_skb(sk, skb);
568 569 570 571 572
	if (rc < 0) {
		int is_udplite = IS_UDPLITE(sk);

		/* Note that an ENOMEM error is charged twice */
		if (rc == -ENOMEM)
573
			UDP6_INC_STATS(sock_net(sk),
574
					 UDP_MIB_RCVBUFERRORS, is_udplite);
575
		UDP6_INC_STATS(sock_net(sk), UDP_MIB_INERRORS, is_udplite);
576 577 578 579 580 581
		kfree_skb(skb);
		return -1;
	}
	return 0;
}

582
static __inline__ void udpv6_err(struct sk_buff *skb,
583
				 struct inet6_skb_parm *opt, u8 type,
584
				 u8 code, int offset, __be32 info)
585
{
586
	__udp6_lib_err(skb, opt, type, code, offset, info, &udp_table);
587 588
}

589 590 591 592 593 594 595 596
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);

597
int udpv6_queue_rcv_skb(struct sock *sk, struct sk_buff *skb)
L
Linus Torvalds 已提交
598
{
599
	struct udp_sock *up = udp_sk(sk);
600
	int rc;
W
Wang Chen 已提交
601
	int is_udplite = IS_UDPLITE(sk);
602

603 604
	if (!xfrm6_policy_check(sk, XFRM_POLICY_IN, skb))
		goto drop;
L
Linus Torvalds 已提交
605

606 607 608 609 610 611 612 613 614 615 616 617 618 619 620 621
	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);
622
		if (skb->len > sizeof(struct udphdr) && encap_rcv) {
623 624
			int ret;

625 626 627 628
			/* Verify checksum before giving to encap */
			if (udp_lib_checksum_complete(skb))
				goto csum_error;

629 630
			ret = encap_rcv(sk, skb);
			if (ret <= 0) {
631 632 633
				__UDP_INC_STATS(sock_net(sk),
						UDP_MIB_INDATAGRAMS,
						is_udplite);
634 635 636 637 638 639 640
				return -ret;
			}
		}

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

641 642 643
	/*
	 * UDP-Lite specific tests, ignored on UDP sockets (see net/ipv4/udp.c).
	 */
W
Wang Chen 已提交
644
	if ((is_udplite & UDPLITE_RECV_CC)  &&  UDP_SKB_CB(skb)->partial_cov) {
645 646

		if (up->pcrlen == 0) {          /* full coverage was set  */
647 648
			net_dbg_ratelimited("UDPLITE6: partial coverage %d while full coverage %d requested\n",
					    UDP_SKB_CB(skb)->cscov, skb->len);
649 650 651
			goto drop;
		}
		if (UDP_SKB_CB(skb)->cscov  <  up->pcrlen) {
652 653
			net_dbg_ratelimited("UDPLITE6: coverage %d too small, need min %d\n",
					    UDP_SKB_CB(skb)->cscov, up->pcrlen);
654 655
			goto drop;
		}
L
Linus Torvalds 已提交
656 657
	}

658
	if (rcu_access_pointer(sk->sk_filter)) {
659
		if (udp_lib_checksum_complete(skb))
660
			goto csum_error;
661 662
		if (sk_filter(sk, skb))
			goto drop;
663
	}
664

665
	udp_csum_pull_header(skb);
666
	if (sk_rcvqueues_full(sk, sk->sk_rcvbuf)) {
667 668
		__UDP6_INC_STATS(sock_net(sk),
				 UDP_MIB_RCVBUFERRORS, is_udplite);
669
		goto drop;
670
	}
671

672
	skb_dst_drop(skb);
673

674 675 676 677 678 679 680 681 682
	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);
683 684

	return rc;
685

686
csum_error:
687
	__UDP6_INC_STATS(sock_net(sk), UDP_MIB_CSUMERRORS, is_udplite);
688
drop:
689
	__UDP6_INC_STATS(sock_net(sk), UDP_MIB_INERRORS, is_udplite);
690
	atomic_inc(&sk->sk_drops);
691 692
	kfree_skb(skb);
	return -1;
L
Linus Torvalds 已提交
693 694
}

695 696 697 698
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
699
{
700
	struct inet_sock *inet = inet_sk(sk);
L
Linus Torvalds 已提交
701

702 703 704 705 706 707 708 709
	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)) ||
710 711 712
	    (sk->sk_bound_dev_if && sk->sk_bound_dev_if != dif) ||
	    (!ipv6_addr_any(&sk->sk_v6_rcv_saddr) &&
		    !ipv6_addr_equal(&sk->sk_v6_rcv_saddr, loc_addr)))
713 714 715 716
		return false;
	if (!inet6_mc_check(sk, loc_addr, rmt_addr))
		return false;
	return true;
L
Linus Torvalds 已提交
717 718
}

719 720 721 722 723
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.
	 */
724 725 726
	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));
727 728
}

L
Linus Torvalds 已提交
729 730 731 732
/*
 * Note: called only from the BH handler context,
 * so we don't need to lock the hashes.
 */
733
static int __udp6_lib_mcast_deliver(struct net *net, struct sk_buff *skb,
734
		const struct in6_addr *saddr, const struct in6_addr *daddr,
735
		struct udp_table *udptable, int proto)
L
Linus Torvalds 已提交
736
{
737
	struct sock *sk, *first = NULL;
738
	const struct udphdr *uh = udp_hdr(skb);
739 740
	unsigned short hnum = ntohs(uh->dest);
	struct udp_hslot *hslot = udp_hashslot(udptable, net, hnum);
741
	unsigned int offset = offsetof(typeof(*sk), sk_node);
742
	unsigned int hash2 = 0, hash2_any = 0, use_hash2 = (hslot->count > 10);
743 744 745
	int dif = inet6_iif(skb);
	struct hlist_node *node;
	struct sk_buff *nskb;
746 747 748 749 750 751 752 753 754

	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
755

756 757 758 759 760 761 762 763 764 765 766 767 768 769 770 771
	sk_for_each_entry_offset_rcu(sk, node, &hslot->head, offset) {
		if (!__udp_v6_is_mcast_sock(net, sk, uh->dest, daddr,
					    uh->source, saddr, dif, hnum))
			continue;
		/* If zero checksum and no_check is not on for
		 * the socket then skip it.
		 */
		if (!uh->check && !udp_sk(sk)->no_check6_rx)
			continue;
		if (!first) {
			first = sk;
			continue;
		}
		nskb = skb_clone(skb, GFP_ATOMIC);
		if (unlikely(!nskb)) {
			atomic_inc(&sk->sk_drops);
772 773 774 775
			__UDP6_INC_STATS(net, UDP_MIB_RCVBUFERRORS,
					 IS_UDPLITE(sk));
			__UDP6_INC_STATS(net, UDP_MIB_INERRORS,
					 IS_UDPLITE(sk));
776
			continue;
H
Hideo Aoki 已提交
777
		}
778

779 780 781
		if (udpv6_queue_rcv_skb(sk, nskb) > 0)
			consume_skb(nskb);
	}
782

783 784 785 786 787 788
	/* 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;
	}

789 790 791
	if (first) {
		if (udpv6_queue_rcv_skb(first, skb) > 0)
			consume_skb(skb);
792
	} else {
793
		kfree_skb(skb);
794 795
		__UDP6_INC_STATS(net, UDP_MIB_IGNOREDMULTI,
				 proto == IPPROTO_UDPLITE);
796
	}
797
	return 0;
L
Linus Torvalds 已提交
798 799
}

800
int __udp6_lib_rcv(struct sk_buff *skb, struct udp_table *udptable,
801
		   int proto)
L
Linus Torvalds 已提交
802
{
803
	const struct in6_addr *saddr, *daddr;
804
	struct net *net = dev_net(skb->dev);
805
	struct udphdr *uh;
806
	struct sock *sk;
L
Linus Torvalds 已提交
807 808 809
	u32 ulen = 0;

	if (!pskb_may_pull(skb, sizeof(struct udphdr)))
810
		goto discard;
L
Linus Torvalds 已提交
811

812 813
	saddr = &ipv6_hdr(skb)->saddr;
	daddr = &ipv6_hdr(skb)->daddr;
814
	uh = udp_hdr(skb);
L
Linus Torvalds 已提交
815 816

	ulen = ntohs(uh->len);
817 818
	if (ulen > skb->len)
		goto short_packet;
L
Linus Torvalds 已提交
819

820 821
	if (proto == IPPROTO_UDP) {
		/* UDP validates ulen. */
L
Linus Torvalds 已提交
822

823 824 825
		/* Check for jumbo payload */
		if (ulen == 0)
			ulen = skb->len;
L
Linus Torvalds 已提交
826

827 828
		if (ulen < sizeof(*uh))
			goto short_packet;
L
Linus Torvalds 已提交
829

830 831 832
		if (ulen < skb->len) {
			if (pskb_trim_rcsum(skb, ulen))
				goto short_packet;
833 834
			saddr = &ipv6_hdr(skb)->saddr;
			daddr = &ipv6_hdr(skb)->daddr;
835
			uh = udp_hdr(skb);
836 837
		}
	}
L
Linus Torvalds 已提交
838

839
	if (udp6_csum_init(skb, uh, proto))
840
		goto csum_error;
841

842 843
	/*
	 *	Multicast receive code
L
Linus Torvalds 已提交
844
	 */
845
	if (ipv6_addr_is_multicast(daddr))
846
		return __udp6_lib_mcast_deliver(net, skb,
847
				saddr, daddr, udptable, proto);
L
Linus Torvalds 已提交
848 849 850

	/* Unicast */

851
	/*
L
Linus Torvalds 已提交
852 853 854
	 * check socket cache ... must talk to Alan about his plans
	 * for sock caches... i'll skip this for now.
	 */
855
	sk = __udp6_lib_lookup_skb(skb, uh->source, uh->dest, udptable);
856
	if (sk) {
857 858
		int ret;

859
		if (!uh->check && !udp_sk(sk)->no_check6_rx) {
860 861 862 863
			udp6_csum_zero_error(skb);
			goto csum_error;
		}

864
		if (inet_get_convert_csum(sk) && uh->check && !IS_UDPLITE(sk))
865 866 867
			skb_checksum_try_convert(skb, IPPROTO_UDP, uh->check,
						 ip6_compute_pseudo);

868
		ret = udpv6_queue_rcv_skb(sk, skb);
L
Linus Torvalds 已提交
869

870
		/* a return value > 0 means to resubmit the input */
871
		if (ret > 0)
872
			return ret;
L
Linus Torvalds 已提交
873

S
Stephen Hemminger 已提交
874
		return 0;
L
Linus Torvalds 已提交
875
	}
876

877 878 879 880 881
	if (!uh->check) {
		udp6_csum_zero_error(skb);
		goto csum_error;
	}

882
	if (!xfrm6_policy_check(NULL, XFRM_POLICY_IN, skb))
883
		goto discard;
884 885

	if (udp_lib_checksum_complete(skb))
886
		goto csum_error;
887

888
	__UDP6_INC_STATS(net, UDP_MIB_NOPORTS, proto == IPPROTO_UDPLITE);
889 890 891
	icmpv6_send(skb, ICMPV6_DEST_UNREACH, ICMPV6_PORT_UNREACH, 0);

	kfree_skb(skb);
S
Stephen Hemminger 已提交
892
	return 0;
L
Linus Torvalds 已提交
893

894
short_packet:
895 896 897 898 899
	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));
900 901
	goto discard;
csum_error:
902
	__UDP6_INC_STATS(net, UDP_MIB_CSUMERRORS, proto == IPPROTO_UDPLITE);
L
Linus Torvalds 已提交
903
discard:
904
	__UDP6_INC_STATS(net, UDP_MIB_INERRORS, proto == IPPROTO_UDPLITE);
L
Linus Torvalds 已提交
905
	kfree_skb(skb);
S
Stephen Hemminger 已提交
906
	return 0;
L
Linus Torvalds 已提交
907
}
908

909
static __inline__ int udpv6_rcv(struct sk_buff *skb)
910
{
911
	return __udp6_lib_rcv(skb, &udp_table, IPPROTO_UDP);
912 913
}

L
Linus Torvalds 已提交
914 915 916 917 918 919 920
/*
 * 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);

921 922 923
	if (up->pending == AF_INET)
		udp_flush_pending_frames(sk);
	else if (up->pending) {
L
Linus Torvalds 已提交
924 925 926
		up->len = 0;
		up->pending = 0;
		ip6_flush_pending_frames(sk);
927
	}
L
Linus Torvalds 已提交
928 929
}

930
/**
931 932 933 934
 *	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)
935 936 937 938 939 940 941
 */
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
942
	struct sk_buff *frags = skb_shinfo(skb)->frag_list;
943 944
	__wsum csum = 0;

V
Vlad Yasevich 已提交
945
	if (!frags) {
946 947 948 949 950 951 952 953 954 955 956 957 958 959 960
		/* 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
961 962 963
		do {
			csum = csum_add(csum, frags->csum);
		} while ((frags = frags->next));
964 965 966 967 968 969 970 971

		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
972 973 974 975
/*
 *	Sending
 */

V
Vlad Yasevich 已提交
976
static int udp_v6_send_skb(struct sk_buff *skb, struct flowi6 *fl6)
L
Linus Torvalds 已提交
977
{
V
Vlad Yasevich 已提交
978
	struct sock *sk = skb->sk;
L
Linus Torvalds 已提交
979 980
	struct udphdr *uh;
	int err = 0;
W
Wang Chen 已提交
981
	int is_udplite = IS_UDPLITE(sk);
982
	__wsum csum = 0;
V
Vlad Yasevich 已提交
983 984
	int offset = skb_transport_offset(skb);
	int len = skb->len - offset;
L
Linus Torvalds 已提交
985 986 987 988

	/*
	 * Create a UDP header
	 */
989
	uh = udp_hdr(skb);
990 991
	uh->source = fl6->fl6_sport;
	uh->dest = fl6->fl6_dport;
V
Vlad Yasevich 已提交
992
	uh->len = htons(len);
L
Linus Torvalds 已提交
993 994
	uh->check = 0;

W
Wang Chen 已提交
995
	if (is_udplite)
V
Vlad Yasevich 已提交
996 997
		csum = udplite_csum(skb);
	else if (udp_sk(sk)->no_check6_tx) {   /* UDP csum disabled */
998 999 1000
		skb->ip_summed = CHECKSUM_NONE;
		goto send;
	} else if (skb->ip_summed == CHECKSUM_PARTIAL) { /* UDP hardware csum */
V
Vlad Yasevich 已提交
1001
		udp6_hwcsum_outgoing(sk, skb, &fl6->saddr, &fl6->daddr, len);
1002 1003
		goto send;
	} else
V
Vlad Yasevich 已提交
1004
		csum = udp_csum(skb);
L
Linus Torvalds 已提交
1005

1006
	/* add protocol-dependent pseudo-header */
1007
	uh->check = csum_ipv6_magic(&fl6->saddr, &fl6->daddr,
V
Vlad Yasevich 已提交
1008
				    len, fl6->flowi6_proto, csum);
L
Linus Torvalds 已提交
1009
	if (uh->check == 0)
1010
		uh->check = CSUM_MANGLED_0;
L
Linus Torvalds 已提交
1011

1012
send:
V
Vlad Yasevich 已提交
1013
	err = ip6_send_skb(skb);
E
Eric Dumazet 已提交
1014 1015
	if (err) {
		if (err == -ENOBUFS && !inet6_sk(sk)->recverr) {
1016 1017
			UDP6_INC_STATS(sock_net(sk),
				       UDP_MIB_SNDBUFERRORS, is_udplite);
E
Eric Dumazet 已提交
1018 1019
			err = 0;
		}
1020 1021 1022 1023
	} else {
		UDP6_INC_STATS(sock_net(sk),
			       UDP_MIB_OUTDATAGRAMS, is_udplite);
	}
V
Vlad Yasevich 已提交
1024 1025 1026 1027 1028 1029 1030 1031 1032 1033 1034 1035 1036 1037 1038 1039 1040 1041 1042 1043 1044 1045 1046 1047
	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
1048 1049 1050 1051 1052 1053
out:
	up->len = 0;
	up->pending = 0;
	return err;
}

1054
int udpv6_sendmsg(struct sock *sk, struct msghdr *msg, size_t len)
L
Linus Torvalds 已提交
1055 1056 1057 1058 1059
{
	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);
1060
	DECLARE_SOCKADDR(struct sockaddr_in6 *, sin6, msg->msg_name);
1061
	struct in6_addr *daddr, *final_p, final;
L
Linus Torvalds 已提交
1062
	struct ipv6_txoptions *opt = NULL;
1063
	struct ipv6_txoptions *opt_to_free = NULL;
L
Linus Torvalds 已提交
1064
	struct ip6_flowlabel *flowlabel = NULL;
1065
	struct flowi6 fl6;
L
Linus Torvalds 已提交
1066 1067 1068 1069
	struct dst_entry *dst;
	int addr_len = msg->msg_namelen;
	int ulen = len;
	int hlimit = -1;
1070
	int tclass = -1;
1071
	int dontfrag = -1;
L
Linus Torvalds 已提交
1072 1073
	int corkreq = up->corkflag || msg->msg_flags&MSG_MORE;
	int err;
1074
	int connected = 0;
W
Wang Chen 已提交
1075
	int is_udplite = IS_UDPLITE(sk);
1076
	int (*getfrag)(void *, char *, int, int, int, struct sk_buff *);
1077
	struct sockcm_cookie sockc;
L
Linus Torvalds 已提交
1078 1079 1080 1081 1082 1083 1084 1085 1086 1087 1088 1089 1090 1091 1092 1093 1094 1095 1096 1097 1098 1099 1100 1101 1102

	/* 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;
1103
		daddr = &sk->sk_v6_daddr;
1104
	} else
L
Linus Torvalds 已提交
1105 1106 1107
		daddr = NULL;

	if (daddr) {
B
Brian Haley 已提交
1108
		if (ipv6_addr_v4mapped(daddr)) {
L
Linus Torvalds 已提交
1109 1110
			struct sockaddr_in sin;
			sin.sin_family = AF_INET;
E
Eric Dumazet 已提交
1111
			sin.sin_port = sin6 ? sin6->sin6_port : inet->inet_dport;
L
Linus Torvalds 已提交
1112 1113 1114 1115 1116 1117
			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;
1118
			return udp_sendmsg(sk, msg, len);
L
Linus Torvalds 已提交
1119 1120 1121 1122
		}
	}

	if (up->pending == AF_INET)
1123
		return udp_sendmsg(sk, msg, len);
L
Linus Torvalds 已提交
1124 1125

	/* Rough check on arithmetic overflow,
1126
	   better check is made in ip6_append_data().
L
Linus Torvalds 已提交
1127 1128 1129
	   */
	if (len > INT_MAX - sizeof(struct udphdr))
		return -EMSGSIZE;
1130

1131
	getfrag  =  is_udplite ?  udplite_getfrag : ip_generic_getfrag;
L
Linus Torvalds 已提交
1132 1133 1134 1135 1136 1137 1138 1139 1140 1141 1142 1143 1144 1145 1146 1147 1148 1149
	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);

1150
	memset(&fl6, 0, sizeof(fl6));
L
Linus Torvalds 已提交
1151 1152 1153 1154 1155

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

1156
		fl6.fl6_dport = sin6->sin6_port;
L
Linus Torvalds 已提交
1157 1158 1159
		daddr = &sin6->sin6_addr;

		if (np->sndflow) {
1160 1161 1162
			fl6.flowlabel = sin6->sin6_flowinfo&IPV6_FLOWINFO_MASK;
			if (fl6.flowlabel&IPV6_FLOWLABEL_MASK) {
				flowlabel = fl6_sock_lookup(sk, fl6.flowlabel);
1163
				if (!flowlabel)
L
Linus Torvalds 已提交
1164 1165 1166 1167 1168 1169 1170 1171 1172
					return -EINVAL;
			}
		}

		/*
		 * Otherwise it will be difficult to maintain
		 * sk->sk_dst_cache.
		 */
		if (sk->sk_state == TCP_ESTABLISHED &&
1173 1174
		    ipv6_addr_equal(daddr, &sk->sk_v6_daddr))
			daddr = &sk->sk_v6_daddr;
L
Linus Torvalds 已提交
1175 1176 1177

		if (addr_len >= sizeof(struct sockaddr_in6) &&
		    sin6->sin6_scope_id &&
1178
		    __ipv6_addr_needs_scope_id(__ipv6_addr_type(daddr)))
1179
			fl6.flowi6_oif = sin6->sin6_scope_id;
L
Linus Torvalds 已提交
1180 1181 1182 1183
	} else {
		if (sk->sk_state != TCP_ESTABLISHED)
			return -EDESTADDRREQ;

1184
		fl6.fl6_dport = inet->inet_dport;
1185
		daddr = &sk->sk_v6_daddr;
1186
		fl6.flowlabel = np->flow_label;
1187
		connected = 1;
L
Linus Torvalds 已提交
1188 1189
	}

1190 1191
	if (!fl6.flowi6_oif)
		fl6.flowi6_oif = sk->sk_bound_dev_if;
L
Linus Torvalds 已提交
1192

1193 1194
	if (!fl6.flowi6_oif)
		fl6.flowi6_oif = np->sticky_pktinfo.ipi6_ifindex;
1195

1196
	fl6.flowi6_mark = sk->sk_mark;
1197
	sockc.tsflags = sk->sk_tsflags;
B
Brian Haley 已提交
1198

L
Linus Torvalds 已提交
1199 1200 1201 1202 1203
	if (msg->msg_controllen) {
		opt = &opt_space;
		memset(opt, 0, sizeof(struct ipv6_txoptions));
		opt->tot_len = sizeof(*opt);

1204
		err = ip6_datagram_send_ctl(sock_net(sk), sk, msg, &fl6, opt,
1205 1206
					    &hlimit, &tclass, &dontfrag,
					    &sockc);
L
Linus Torvalds 已提交
1207 1208 1209 1210
		if (err < 0) {
			fl6_sock_release(flowlabel);
			return err;
		}
1211 1212
		if ((fl6.flowlabel&IPV6_FLOWLABEL_MASK) && !flowlabel) {
			flowlabel = fl6_sock_lookup(sk, fl6.flowlabel);
1213
			if (!flowlabel)
L
Linus Torvalds 已提交
1214 1215 1216 1217
				return -EINVAL;
		}
		if (!(opt->opt_nflen|opt->opt_flen))
			opt = NULL;
1218
		connected = 0;
L
Linus Torvalds 已提交
1219
	}
1220 1221 1222 1223
	if (!opt) {
		opt = txopt_get(np);
		opt_to_free = opt;
	}
1224 1225 1226
	if (flowlabel)
		opt = fl6_merge_options(&opt_space, flowlabel, opt);
	opt = ipv6_fixup_options(&opt_space, opt);
L
Linus Torvalds 已提交
1227

1228
	fl6.flowi6_proto = sk->sk_protocol;
1229
	if (!ipv6_addr_any(daddr))
A
Alexey Dobriyan 已提交
1230
		fl6.daddr = *daddr;
1231
	else
1232 1233
		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
1234
		fl6.saddr = np->saddr;
1235
	fl6.fl6_sport = inet->inet_sport;
1236

1237
	final_p = fl6_update_dst(&fl6, opt, &final);
1238
	if (final_p)
1239
		connected = 0;
L
Linus Torvalds 已提交
1240

1241 1242
	if (!fl6.flowi6_oif && ipv6_addr_is_multicast(&fl6.daddr)) {
		fl6.flowi6_oif = np->mcast_oif;
1243
		connected = 0;
1244 1245
	} else if (!fl6.flowi6_oif)
		fl6.flowi6_oif = np->ucast_oif;
L
Linus Torvalds 已提交
1246

1247
	security_sk_classify_flow(sk, flowi6_to_flowi(&fl6));
V
Venkat Yekkirala 已提交
1248

1249
	dst = ip6_sk_dst_lookup_flow(sk, &fl6, final_p);
1250 1251 1252
	if (IS_ERR(dst)) {
		err = PTR_ERR(dst);
		dst = NULL;
L
Linus Torvalds 已提交
1253
		goto out;
1254
	}
L
Linus Torvalds 已提交
1255

1256 1257
	if (hlimit < 0)
		hlimit = ip6_sk_dst_hoplimit(np, &fl6, dst);
L
Linus Torvalds 已提交
1258

G
Gerrit Renker 已提交
1259
	if (tclass < 0)
1260 1261
		tclass = np->tclass;

L
Linus Torvalds 已提交
1262 1263 1264 1265
	if (msg->msg_flags&MSG_CONFIRM)
		goto do_confirm;
back_from_confirm:

1266 1267 1268 1269 1270 1271 1272
	/* 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,
1273
				   msg->msg_flags, dontfrag, &sockc);
1274 1275 1276 1277 1278 1279
		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
1280 1281 1282 1283 1284 1285
	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);

1286
		net_dbg_ratelimited("udp cork app bug 2\n");
L
Linus Torvalds 已提交
1287 1288 1289 1290 1291 1292 1293
		err = -EINVAL;
		goto out;
	}

	up->pending = AF_INET6;

do_append_data:
1294 1295
	if (dontfrag < 0)
		dontfrag = np->dontfrag;
L
Linus Torvalds 已提交
1296
	up->len += ulen;
1297
	err = ip6_append_data(sk, getfrag, msg, ulen,
1298
		sizeof(struct udphdr), hlimit, tclass, opt, &fl6,
1299
		(struct rt6_info *)dst,
1300 1301
		corkreq ? msg->msg_flags|MSG_MORE : msg->msg_flags, dontfrag,
		&sockc);
L
Linus Torvalds 已提交
1302 1303 1304
	if (err)
		udp_v6_flush_pending_frames(sk);
	else if (!corkreq)
1305
		err = udp_v6_push_pending_frames(sk);
H
Herbert Xu 已提交
1306 1307
	else if (unlikely(skb_queue_empty(&sk->sk_write_queue)))
		up->pending = 0;
L
Linus Torvalds 已提交
1308

1309 1310 1311 1312 1313
	if (err > 0)
		err = np->recverr ? net_xmit_errno(err) : 0;
	release_sock(sk);

release_dst:
1314 1315 1316
	if (dst) {
		if (connected) {
			ip6_dst_store(sk, dst,
1317 1318
				      ipv6_addr_equal(&fl6.daddr, &sk->sk_v6_daddr) ?
				      &sk->sk_v6_daddr : NULL,
1319
#ifdef CONFIG_IPV6_SUBTREES
1320
				      ipv6_addr_equal(&fl6.saddr, &np->saddr) ?
1321 1322 1323
				      &np->saddr :
#endif
				      NULL);
1324 1325 1326
		} else {
			dst_release(dst);
		}
1327
		dst = NULL;
1328 1329
	}

L
Linus Torvalds 已提交
1330
out:
1331
	dst_release(dst);
L
Linus Torvalds 已提交
1332
	fl6_sock_release(flowlabel);
1333
	txopt_put(opt_to_free);
1334
	if (!err)
L
Linus Torvalds 已提交
1335
		return len;
1336 1337 1338 1339 1340 1341 1342 1343
	/*
	 * 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)) {
1344 1345
		UDP6_INC_STATS(sock_net(sk),
			       UDP_MIB_SNDBUFERRORS, is_udplite);
1346
	}
L
Linus Torvalds 已提交
1347 1348 1349 1350 1351 1352 1353 1354 1355 1356
	return err;

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

1357
void udpv6_destroy_sock(struct sock *sk)
L
Linus Torvalds 已提交
1358
{
T
Tom Parkin 已提交
1359
	struct udp_sock *up = udp_sk(sk);
L
Linus Torvalds 已提交
1360 1361 1362 1363
	lock_sock(sk);
	udp_v6_flush_pending_frames(sk);
	release_sock(sk);

T
Tom Parkin 已提交
1364 1365 1366 1367 1368 1369 1370
	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
1371 1372 1373 1374 1375 1376
	inet6_destroy_sock(sk);
}

/*
 *	Socket option code for UDP
 */
1377
int udpv6_setsockopt(struct sock *sk, int level, int optname,
1378
		     char __user *optval, unsigned int optlen)
1379
{
1380
	if (level == SOL_UDP  ||  level == SOL_UDPLITE)
1381 1382
		return udp_lib_setsockopt(sk, level, optname, optval, optlen,
					  udp_v6_push_pending_frames);
1383
	return ipv6_setsockopt(sk, level, optname, optval, optlen);
1384 1385 1386
}

#ifdef CONFIG_COMPAT
1387
int compat_udpv6_setsockopt(struct sock *sk, int level, int optname,
1388
			    char __user *optval, unsigned int optlen)
1389
{
1390
	if (level == SOL_UDP  ||  level == SOL_UDPLITE)
1391 1392
		return udp_lib_setsockopt(sk, level, optname, optval, optlen,
					  udp_v6_push_pending_frames);
1393
	return compat_ipv6_setsockopt(sk, level, optname, optval, optlen);
1394 1395 1396
}
#endif

1397 1398
int udpv6_getsockopt(struct sock *sk, int level, int optname,
		     char __user *optval, int __user *optlen)
1399
{
1400
	if (level == SOL_UDP  ||  level == SOL_UDPLITE)
1401
		return udp_lib_getsockopt(sk, level, optname, optval, optlen);
1402
	return ipv6_getsockopt(sk, level, optname, optval, optlen);
1403 1404 1405
}

#ifdef CONFIG_COMPAT
1406 1407
int compat_udpv6_getsockopt(struct sock *sk, int level, int optname,
			    char __user *optval, int __user *optlen)
1408
{
1409
	if (level == SOL_UDP  ||  level == SOL_UDPLITE)
1410
		return udp_lib_getsockopt(sk, level, optname, optval, optlen);
1411
	return compat_ipv6_getsockopt(sk, level, optname, optval, optlen);
1412 1413 1414
}
#endif

1415
static const struct inet6_protocol udpv6_protocol = {
L
Linus Torvalds 已提交
1416 1417 1418 1419 1420 1421 1422
	.handler	=	udpv6_rcv,
	.err_handler	=	udpv6_err,
	.flags		=	INET6_PROTO_NOPOLICY|INET6_PROTO_FINAL,
};

/* ------------------------------------------------------------------------ */
#ifdef CONFIG_PROC_FS
1423
int udp6_seq_show(struct seq_file *seq, void *v)
L
Linus Torvalds 已提交
1424
{
1425 1426 1427 1428 1429 1430 1431 1432 1433
	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
1434 1435 1436
	return 0;
}

1437 1438 1439 1440 1441 1442 1443 1444
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
1445 1446 1447
static struct udp_seq_afinfo udp6_seq_afinfo = {
	.name		= "udp6",
	.family		= AF_INET6,
1448
	.udp_table	= &udp_table,
1449
	.seq_fops	= &udp6_afinfo_seq_fops,
1450 1451 1452
	.seq_ops	= {
		.show		= udp6_seq_show,
	},
L
Linus Torvalds 已提交
1453 1454
};

1455
int __net_init udp6_proc_init(struct net *net)
L
Linus Torvalds 已提交
1456
{
1457
	return udp_proc_register(net, &udp6_seq_afinfo);
L
Linus Torvalds 已提交
1458 1459
}

I
Ian Morris 已提交
1460 1461
void udp6_proc_exit(struct net *net)
{
1462
	udp_proc_unregister(net, &udp6_seq_afinfo);
L
Linus Torvalds 已提交
1463 1464 1465
}
#endif /* CONFIG_PROC_FS */

E
Eric Dumazet 已提交
1466 1467 1468 1469 1470 1471 1472 1473 1474 1475 1476
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
1477 1478 1479
/* ------------------------------------------------------------------------ */

struct proto udpv6_prot = {
1480 1481
	.name		   = "UDPv6",
	.owner		   = THIS_MODULE,
1482
	.close		   = udp_lib_close,
1483 1484 1485 1486 1487 1488 1489 1490
	.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,
1491
	.backlog_rcv	   = __udpv6_queue_rcv_skb,
1492
	.release_cb	   = ip6_datagram_release_cb,
1493 1494
	.hash		   = udp_lib_hash,
	.unhash		   = udp_lib_unhash,
E
Eric Dumazet 已提交
1495
	.rehash		   = udp_v6_rehash,
1496
	.get_port	   = udp_v6_get_port,
H
Hideo Aoki 已提交
1497 1498 1499 1500
	.memory_allocated  = &udp_memory_allocated,
	.sysctl_mem	   = sysctl_udp_mem,
	.sysctl_wmem	   = &sysctl_udp_wmem_min,
	.sysctl_rmem	   = &sysctl_udp_rmem_min,
1501
	.obj_size	   = sizeof(struct udp6_sock),
1502
	.slab_flags	   = SLAB_DESTROY_BY_RCU,
1503
	.h.udp_table	   = &udp_table,
1504
#ifdef CONFIG_COMPAT
1505 1506
	.compat_setsockopt = compat_udpv6_setsockopt,
	.compat_getsockopt = compat_udpv6_getsockopt,
1507
#endif
E
Eric Dumazet 已提交
1508
	.clear_sk	   = udp_v6_clear_sk,
L
Linus Torvalds 已提交
1509 1510 1511 1512 1513 1514 1515 1516 1517 1518
};

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

1519
int __init udpv6_init(void)
L
Linus Torvalds 已提交
1520
{
1521 1522
	int ret;

1523 1524
	ret = inet6_add_protocol(&udpv6_protocol, IPPROTO_UDP);
	if (ret)
1525
		goto out;
1526

1527 1528 1529 1530 1531 1532 1533 1534 1535 1536 1537
	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;
}

1538
void udpv6_exit(void)
1539 1540 1541
{
	inet6_unregister_protosw(&udpv6_protosw);
	inet6_del_protocol(&udpv6_protocol, IPPROTO_UDP);
L
Linus Torvalds 已提交
1542
}