udp.c 37.6 KB
Newer Older
L
Linus Torvalds 已提交
1 2
/*
 *	UDP over IPv6
3
 *	Linux INET6 implementation
L
Linus Torvalds 已提交
4 5
 *
 *	Authors:
6
 *	Pedro Roque		<roque@di.fc.ul.pt>
L
Linus Torvalds 已提交
7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34
 *
 *	Based on linux/ipv4/udp.c
 *
 *	Fixes:
 *	Hideaki YOSHIFUJI	:	sin6_scope_id support
 *	YOSHIFUJI Hideaki @USAGI and:	Support IPV6_V6ONLY socket option, which
 *	Alexey Kuznetsov		allow both IPv4 and IPv6 sockets to bind
 *					a single port at the same time.
 *      Kazunori MIYAZAWA @USAGI:       change process style to use ip6_append_data
 *      YOSHIFUJI Hideaki @USAGI:	convert /proc/net/udp6 to seq_file.
 *
 *	This program is free software; you can redistribute it and/or
 *      modify it under the terms of the GNU General Public License
 *      as published by the Free Software Foundation; either version
 *      2 of the License, or (at your option) any later version.
 */

#include <linux/errno.h>
#include <linux/types.h>
#include <linux/socket.h>
#include <linux/sockios.h>
#include <linux/net.h>
#include <linux/in6.h>
#include <linux/netdevice.h>
#include <linux/if_arp.h>
#include <linux/ipv6.h>
#include <linux/icmpv6.h>
#include <linux/init.h>
35
#include <linux/module.h>
36
#include <linux/skbuff.h>
37
#include <linux/slab.h>
L
Linus Torvalds 已提交
38 39
#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
/* 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)
334 335 336
struct sock *udp6_lib_lookup(struct net *net, const struct in6_addr *saddr, __be16 sport,
			     const struct in6_addr *daddr, __be16 dport, int dif)
{
337 338 339 340 341 342 343
	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;
344 345
}
EXPORT_SYMBOL_GPL(udp6_lib_lookup);
346
#endif
347

L
Linus Torvalds 已提交
348
/*
349 350
 *	This should be easy, if there is something there we
 *	return it, otherwise we block.
L
Linus Torvalds 已提交
351 352
 */

353
int udpv6_recvmsg(struct sock *sk, struct msghdr *msg, size_t len,
L
Linus Torvalds 已提交
354 355 356 357
		  int noblock, int flags, int *addr_len)
{
	struct ipv6_pinfo *np = inet6_sk(sk);
	struct inet_sock *inet = inet_sk(sk);
358
	struct sk_buff *skb;
359
	unsigned int ulen, copied;
360
	int peeked, off = 0;
361 362
	int err;
	int is_udplite = IS_UDPLITE(sk);
363
	bool checksum_valid = false;
364
	int is_udp4;
365
	bool slow;
L
Linus Torvalds 已提交
366 367

	if (flags & MSG_ERRQUEUE)
368
		return ipv6_recv_error(sk, msg, len, addr_len);
L
Linus Torvalds 已提交
369

370
	if (np->rxpmtu && np->rxopt.bits.rxpmtu)
371
		return ipv6_recv_rxpmtu(sk, msg, len, addr_len);
372

L
Linus Torvalds 已提交
373
try_again:
374
	skb = __skb_recv_datagram(sk, flags | (noblock ? MSG_DONTWAIT : 0),
375
				  &peeked, &off, &err);
L
Linus Torvalds 已提交
376 377 378
	if (!skb)
		goto out;

379
	ulen = skb->len;
380 381 382 383
	copied = len;
	if (copied > ulen)
		copied = ulen;
	else if (copied < ulen)
384
		msg->msg_flags |= MSG_TRUNC;
L
Linus Torvalds 已提交
385

386 387
	is_udp4 = (skb->protocol == htons(ETH_P_IP));

388
	/*
389 390 391
	 * 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.
392 393
	 */

394
	if (copied < ulen || UDP_SKB_CB(skb)->partial_cov) {
395 396
		checksum_valid = !udp_lib_checksum_complete(skb);
		if (!checksum_valid)
L
Linus Torvalds 已提交
397
			goto csum_copy_err;
398 399
	}

400
	if (checksum_valid || skb_csum_unnecessary(skb))
401
		err = skb_copy_datagram_msg(skb, 0, msg, copied);
402
	else {
403
		err = skb_copy_and_csum_datagram_msg(skb, 0, msg);
L
Linus Torvalds 已提交
404 405 406
		if (err == -EINVAL)
			goto csum_copy_err;
	}
407 408
	if (unlikely(err)) {
		trace_kfree_skb(skb, udpv6_recvmsg);
409 410 411 412 413 414 415 416 417 418 419
		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
420
		goto out_free;
421
	}
422 423 424 425 426 427 428 429
	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);
	}
430

431
	sock_recv_ts_and_drops(msg, sk, skb);
L
Linus Torvalds 已提交
432 433 434

	/* Copy the address. */
	if (msg->msg_name) {
435
		DECLARE_SOCKADDR(struct sockaddr_in6 *, sin6, msg->msg_name);
L
Linus Torvalds 已提交
436
		sin6->sin6_family = AF_INET6;
437
		sin6->sin6_port = udp_hdr(skb)->source;
L
Linus Torvalds 已提交
438 439
		sin6->sin6_flowinfo = 0;

440
		if (is_udp4) {
B
Brian Haley 已提交
441 442
			ipv6_addr_set_v4mapped(ip_hdr(skb)->saddr,
					       &sin6->sin6_addr);
443 444
			sin6->sin6_scope_id = 0;
		} else {
A
Alexey Dobriyan 已提交
445
			sin6->sin6_addr = ipv6_hdr(skb)->saddr;
446 447
			sin6->sin6_scope_id =
				ipv6_iface_scope_id(&sin6->sin6_addr,
448
						    inet6_iif(skb));
L
Linus Torvalds 已提交
449
		}
450
		*addr_len = sizeof(*sin6);
L
Linus Torvalds 已提交
451
	}
452 453 454 455

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

456
	if (is_udp4) {
L
Linus Torvalds 已提交
457 458 459 460
		if (inet->cmsg_flags)
			ip_cmsg_recv(msg, skb);
	} else {
		if (np->rxopt.all)
461
			ip6_datagram_recv_specific_ctl(sk, msg, skb);
462
	}
L
Linus Torvalds 已提交
463

464
	err = copied;
L
Linus Torvalds 已提交
465
	if (flags & MSG_TRUNC)
466
		err = ulen;
L
Linus Torvalds 已提交
467 468

out_free:
469
	skb_free_datagram_locked(sk, skb);
L
Linus Torvalds 已提交
470 471 472 473
out:
	return err;

csum_copy_err:
474
	slow = lock_sock_fast(sk);
475
	if (!skb_kill_datagram(sk, skb, flags)) {
476 477 478
		if (is_udp4) {
			UDP_INC_STATS_USER(sock_net(sk),
					UDP_MIB_CSUMERRORS, is_udplite);
479 480
			UDP_INC_STATS_USER(sock_net(sk),
					UDP_MIB_INERRORS, is_udplite);
481 482 483
		} else {
			UDP6_INC_STATS_USER(sock_net(sk),
					UDP_MIB_CSUMERRORS, is_udplite);
484 485
			UDP6_INC_STATS_USER(sock_net(sk),
					UDP_MIB_INERRORS, is_udplite);
486
		}
487
	}
488
	unlock_sock_fast(sk, slow);
L
Linus Torvalds 已提交
489

490 491
	/* starting over for a new packet, but check if we need to yield */
	cond_resched();
492
	msg->msg_flags &= ~MSG_TRUNC;
L
Linus Torvalds 已提交
493 494 495
	goto try_again;
}

496
void __udp6_lib_err(struct sk_buff *skb, struct inet6_skb_parm *opt,
497
		    u8 type, u8 code, int offset, __be32 info,
498
		    struct udp_table *udptable)
L
Linus Torvalds 已提交
499 500
{
	struct ipv6_pinfo *np;
501 502 503
	const struct ipv6hdr *hdr = (const struct ipv6hdr *)skb->data;
	const struct in6_addr *saddr = &hdr->saddr;
	const struct in6_addr *daddr = &hdr->daddr;
504
	struct udphdr *uh = (struct udphdr *)(skb->data+offset);
L
Linus Torvalds 已提交
505
	struct sock *sk;
506
	int harderr;
L
Linus Torvalds 已提交
507
	int err;
508
	struct net *net = dev_net(skb->dev);
L
Linus Torvalds 已提交
509

510
	sk = __udp6_lib_lookup(net, daddr, uh->dest, saddr, uh->source,
511
			       inet6_iif(skb), udptable, skb);
512
	if (!sk) {
513 514
		ICMP6_INC_STATS_BH(net, __in6_dev_get(skb->dev),
				   ICMP6_MIB_INERRORS);
L
Linus Torvalds 已提交
515
		return;
516
	}
L
Linus Torvalds 已提交
517

518 519 520
	harderr = icmpv6_err_convert(type, code, &err);
	np = inet6_sk(sk);

521 522 523
	if (type == ICMPV6_PKT_TOOBIG) {
		if (!ip6_sk_accept_pmtu(sk))
			goto out;
524
		ip6_sk_update_pmtu(skb, sk, info);
525 526
		if (np->pmtudisc != IPV6_PMTUDISC_DONT)
			harderr = 1;
527
	}
528
	if (type == NDISC_REDIRECT) {
529
		ip6_sk_redirect(skb, sk);
530 531
		goto out;
	}
532

533 534 535 536
	if (!np->recverr) {
		if (!harderr || sk->sk_state != TCP_ESTABLISHED)
			goto out;
	} else {
L
Linus Torvalds 已提交
537
		ipv6_icmp_error(sk, skb, err, uh->dest, ntohl(info), (u8 *)(uh+1));
538
	}
539

L
Linus Torvalds 已提交
540 541 542
	sk->sk_err = err;
	sk->sk_error_report(sk);
out:
543
	return;
L
Linus Torvalds 已提交
544 545
}

546 547 548 549
static int __udpv6_queue_rcv_skb(struct sock *sk, struct sk_buff *skb)
{
	int rc;

550
	if (!ipv6_addr_any(&sk->sk_v6_daddr)) {
551
		sock_rps_save_rxhash(sk, skb);
552
		sk_mark_napi_id(sk, skb);
E
Eric Dumazet 已提交
553
		sk_incoming_cpu_update(sk);
554
	}
555

556
	rc = __sock_queue_rcv_skb(sk, skb);
557 558 559 560 561 562 563 564 565 566 567 568 569 570
	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;
}

571
static __inline__ void udpv6_err(struct sk_buff *skb,
572
				 struct inet6_skb_parm *opt, u8 type,
573
				 u8 code, int offset, __be32 info)
574
{
575
	__udp6_lib_err(skb, opt, type, code, offset, info, &udp_table);
576 577
}

578 579 580 581 582 583 584 585
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);

586
int udpv6_queue_rcv_skb(struct sock *sk, struct sk_buff *skb)
L
Linus Torvalds 已提交
587
{
588
	struct udp_sock *up = udp_sk(sk);
589
	int rc;
W
Wang Chen 已提交
590
	int is_udplite = IS_UDPLITE(sk);
591

592 593
	if (!xfrm6_policy_check(sk, XFRM_POLICY_IN, skb))
		goto drop;
L
Linus Torvalds 已提交
594

595 596 597 598 599 600 601 602 603 604 605 606 607 608 609 610
	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);
611
		if (skb->len > sizeof(struct udphdr) && encap_rcv) {
612 613
			int ret;

614 615 616 617
			/* Verify checksum before giving to encap */
			if (udp_lib_checksum_complete(skb))
				goto csum_error;

618 619 620 621 622 623 624 625 626 627 628 629
			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 */
	}

630 631 632
	/*
	 * UDP-Lite specific tests, ignored on UDP sockets (see net/ipv4/udp.c).
	 */
W
Wang Chen 已提交
633
	if ((is_udplite & UDPLITE_RECV_CC)  &&  UDP_SKB_CB(skb)->partial_cov) {
634 635

		if (up->pcrlen == 0) {          /* full coverage was set  */
636 637
			net_dbg_ratelimited("UDPLITE6: partial coverage %d while full coverage %d requested\n",
					    UDP_SKB_CB(skb)->cscov, skb->len);
638 639 640
			goto drop;
		}
		if (UDP_SKB_CB(skb)->cscov  <  up->pcrlen) {
641 642
			net_dbg_ratelimited("UDPLITE6: coverage %d too small, need min %d\n",
					    UDP_SKB_CB(skb)->cscov, up->pcrlen);
643 644
			goto drop;
		}
L
Linus Torvalds 已提交
645 646
	}

647
	if (rcu_access_pointer(sk->sk_filter)) {
648
		if (udp_lib_checksum_complete(skb))
649
			goto csum_error;
650 651
		if (sk_filter(sk, skb))
			goto drop;
652
	}
653

654
	udp_csum_pull_header(skb);
655
	if (sk_rcvqueues_full(sk, sk->sk_rcvbuf)) {
656 657
		UDP6_INC_STATS_BH(sock_net(sk),
				  UDP_MIB_RCVBUFERRORS, is_udplite);
658
		goto drop;
659
	}
660

661
	skb_dst_drop(skb);
662

663 664 665 666 667 668 669 670 671
	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);
672 673

	return rc;
674

675 676
csum_error:
	UDP6_INC_STATS_BH(sock_net(sk), UDP_MIB_CSUMERRORS, is_udplite);
677
drop:
678
	UDP6_INC_STATS_BH(sock_net(sk), UDP_MIB_INERRORS, is_udplite);
679
	atomic_inc(&sk->sk_drops);
680 681
	kfree_skb(skb);
	return -1;
L
Linus Torvalds 已提交
682 683
}

684 685 686 687
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
688
{
689
	struct inet_sock *inet = inet_sk(sk);
L
Linus Torvalds 已提交
690

691 692 693 694 695 696 697 698
	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)) ||
699 700 701
	    (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)))
702 703 704 705
		return false;
	if (!inet6_mc_check(sk, loc_addr, rmt_addr))
		return false;
	return true;
L
Linus Torvalds 已提交
706 707
}

708 709 710 711 712
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.
	 */
713 714 715
	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));
716 717
}

L
Linus Torvalds 已提交
718 719 720 721
/*
 * Note: called only from the BH handler context,
 * so we don't need to lock the hashes.
 */
722
static int __udp6_lib_mcast_deliver(struct net *net, struct sk_buff *skb,
723
		const struct in6_addr *saddr, const struct in6_addr *daddr,
724
		struct udp_table *udptable, int proto)
L
Linus Torvalds 已提交
725
{
726
	struct sock *sk, *first = NULL;
727
	const struct udphdr *uh = udp_hdr(skb);
728 729
	unsigned short hnum = ntohs(uh->dest);
	struct udp_hslot *hslot = udp_hashslot(udptable, net, hnum);
730
	unsigned int offset = offsetof(typeof(*sk), sk_node);
731
	unsigned int hash2 = 0, hash2_any = 0, use_hash2 = (hslot->count > 10);
732 733 734
	int dif = inet6_iif(skb);
	struct hlist_node *node;
	struct sk_buff *nskb;
735 736 737 738 739 740 741 742 743

	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
744

745 746 747 748 749 750 751 752 753 754 755 756 757 758 759 760 761 762 763 764 765
	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);
			UDP6_INC_STATS_BH(net, UDP_MIB_RCVBUFERRORS,
					  IS_UDPLITE(sk));
			UDP6_INC_STATS_BH(net, UDP_MIB_INERRORS,
					  IS_UDPLITE(sk));
			continue;
H
Hideo Aoki 已提交
766
		}
767

768 769 770
		if (udpv6_queue_rcv_skb(sk, nskb) > 0)
			consume_skb(nskb);
	}
771

772 773 774 775 776 777
	/* 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;
	}

778 779 780
	if (first) {
		if (udpv6_queue_rcv_skb(first, skb) > 0)
			consume_skb(skb);
781
	} else {
782 783 784
		kfree_skb(skb);
		UDP6_INC_STATS_BH(net, UDP_MIB_IGNOREDMULTI,
				  proto == IPPROTO_UDPLITE);
785
	}
786
	return 0;
L
Linus Torvalds 已提交
787 788
}

789
int __udp6_lib_rcv(struct sk_buff *skb, struct udp_table *udptable,
790
		   int proto)
L
Linus Torvalds 已提交
791
{
792
	const struct in6_addr *saddr, *daddr;
793
	struct net *net = dev_net(skb->dev);
794
	struct udphdr *uh;
795
	struct sock *sk;
L
Linus Torvalds 已提交
796 797 798
	u32 ulen = 0;

	if (!pskb_may_pull(skb, sizeof(struct udphdr)))
799
		goto discard;
L
Linus Torvalds 已提交
800

801 802
	saddr = &ipv6_hdr(skb)->saddr;
	daddr = &ipv6_hdr(skb)->daddr;
803
	uh = udp_hdr(skb);
L
Linus Torvalds 已提交
804 805

	ulen = ntohs(uh->len);
806 807
	if (ulen > skb->len)
		goto short_packet;
L
Linus Torvalds 已提交
808

809 810
	if (proto == IPPROTO_UDP) {
		/* UDP validates ulen. */
L
Linus Torvalds 已提交
811

812 813 814
		/* Check for jumbo payload */
		if (ulen == 0)
			ulen = skb->len;
L
Linus Torvalds 已提交
815

816 817
		if (ulen < sizeof(*uh))
			goto short_packet;
L
Linus Torvalds 已提交
818

819 820 821
		if (ulen < skb->len) {
			if (pskb_trim_rcsum(skb, ulen))
				goto short_packet;
822 823
			saddr = &ipv6_hdr(skb)->saddr;
			daddr = &ipv6_hdr(skb)->daddr;
824
			uh = udp_hdr(skb);
825 826
		}
	}
L
Linus Torvalds 已提交
827

828
	if (udp6_csum_init(skb, uh, proto))
829
		goto csum_error;
830

831 832
	/*
	 *	Multicast receive code
L
Linus Torvalds 已提交
833
	 */
834
	if (ipv6_addr_is_multicast(daddr))
835
		return __udp6_lib_mcast_deliver(net, skb,
836
				saddr, daddr, udptable, proto);
L
Linus Torvalds 已提交
837 838 839

	/* Unicast */

840
	/*
L
Linus Torvalds 已提交
841 842 843
	 * check socket cache ... must talk to Alan about his plans
	 * for sock caches... i'll skip this for now.
	 */
844
	sk = __udp6_lib_lookup_skb(skb, uh->source, uh->dest, udptable);
845
	if (sk) {
846 847
		int ret;

848
		if (!uh->check && !udp_sk(sk)->no_check6_rx) {
849 850 851 852
			udp6_csum_zero_error(skb);
			goto csum_error;
		}

853
		if (inet_get_convert_csum(sk) && uh->check && !IS_UDPLITE(sk))
854 855 856
			skb_checksum_try_convert(skb, IPPROTO_UDP, uh->check,
						 ip6_compute_pseudo);

857
		ret = udpv6_queue_rcv_skb(sk, skb);
L
Linus Torvalds 已提交
858

859
		/* a return value > 0 means to resubmit the input */
860
		if (ret > 0)
861
			return ret;
L
Linus Torvalds 已提交
862

S
Stephen Hemminger 已提交
863
		return 0;
L
Linus Torvalds 已提交
864
	}
865

866 867 868 869 870
	if (!uh->check) {
		udp6_csum_zero_error(skb);
		goto csum_error;
	}

871
	if (!xfrm6_policy_check(NULL, XFRM_POLICY_IN, skb))
872
		goto discard;
873 874

	if (udp_lib_checksum_complete(skb))
875
		goto csum_error;
876 877 878 879 880

	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
881
	return 0;
L
Linus Torvalds 已提交
882

883
short_packet:
884 885 886 887 888
	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));
889 890 891
	goto discard;
csum_error:
	UDP6_INC_STATS_BH(net, UDP_MIB_CSUMERRORS, proto == IPPROTO_UDPLITE);
L
Linus Torvalds 已提交
892
discard:
893
	UDP6_INC_STATS_BH(net, UDP_MIB_INERRORS, proto == IPPROTO_UDPLITE);
L
Linus Torvalds 已提交
894
	kfree_skb(skb);
S
Stephen Hemminger 已提交
895
	return 0;
L
Linus Torvalds 已提交
896
}
897

898
static __inline__ int udpv6_rcv(struct sk_buff *skb)
899
{
900
	return __udp6_lib_rcv(skb, &udp_table, IPPROTO_UDP);
901 902
}

L
Linus Torvalds 已提交
903 904 905 906 907 908 909
/*
 * 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);

910 911 912
	if (up->pending == AF_INET)
		udp_flush_pending_frames(sk);
	else if (up->pending) {
L
Linus Torvalds 已提交
913 914 915
		up->len = 0;
		up->pending = 0;
		ip6_flush_pending_frames(sk);
916
	}
L
Linus Torvalds 已提交
917 918
}

919
/**
920 921 922 923
 *	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)
924 925 926 927 928 929 930
 */
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
931
	struct sk_buff *frags = skb_shinfo(skb)->frag_list;
932 933
	__wsum csum = 0;

V
Vlad Yasevich 已提交
934
	if (!frags) {
935 936 937 938 939 940 941 942 943 944 945 946 947 948 949
		/* 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
950 951 952
		do {
			csum = csum_add(csum, frags->csum);
		} while ((frags = frags->next));
953 954 955 956 957 958 959 960

		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
961 962 963 964
/*
 *	Sending
 */

V
Vlad Yasevich 已提交
965
static int udp_v6_send_skb(struct sk_buff *skb, struct flowi6 *fl6)
L
Linus Torvalds 已提交
966
{
V
Vlad Yasevich 已提交
967
	struct sock *sk = skb->sk;
L
Linus Torvalds 已提交
968 969
	struct udphdr *uh;
	int err = 0;
W
Wang Chen 已提交
970
	int is_udplite = IS_UDPLITE(sk);
971
	__wsum csum = 0;
V
Vlad Yasevich 已提交
972 973
	int offset = skb_transport_offset(skb);
	int len = skb->len - offset;
L
Linus Torvalds 已提交
974 975 976 977

	/*
	 * Create a UDP header
	 */
978
	uh = udp_hdr(skb);
979 980
	uh->source = fl6->fl6_sport;
	uh->dest = fl6->fl6_dport;
V
Vlad Yasevich 已提交
981
	uh->len = htons(len);
L
Linus Torvalds 已提交
982 983
	uh->check = 0;

W
Wang Chen 已提交
984
	if (is_udplite)
V
Vlad Yasevich 已提交
985 986
		csum = udplite_csum(skb);
	else if (udp_sk(sk)->no_check6_tx) {   /* UDP csum disabled */
987 988 989
		skb->ip_summed = CHECKSUM_NONE;
		goto send;
	} else if (skb->ip_summed == CHECKSUM_PARTIAL) { /* UDP hardware csum */
V
Vlad Yasevich 已提交
990
		udp6_hwcsum_outgoing(sk, skb, &fl6->saddr, &fl6->daddr, len);
991 992
		goto send;
	} else
V
Vlad Yasevich 已提交
993
		csum = udp_csum(skb);
L
Linus Torvalds 已提交
994

995
	/* add protocol-dependent pseudo-header */
996
	uh->check = csum_ipv6_magic(&fl6->saddr, &fl6->daddr,
V
Vlad Yasevich 已提交
997
				    len, fl6->flowi6_proto, csum);
L
Linus Torvalds 已提交
998
	if (uh->check == 0)
999
		uh->check = CSUM_MANGLED_0;
L
Linus Torvalds 已提交
1000

1001
send:
V
Vlad Yasevich 已提交
1002
	err = ip6_send_skb(skb);
E
Eric Dumazet 已提交
1003 1004 1005 1006 1007 1008 1009 1010 1011
	if (err) {
		if (err == -ENOBUFS && !inet6_sk(sk)->recverr) {
			UDP6_INC_STATS_USER(sock_net(sk),
					    UDP_MIB_SNDBUFERRORS, is_udplite);
			err = 0;
		}
	} else
		UDP6_INC_STATS_USER(sock_net(sk),
				    UDP_MIB_OUTDATAGRAMS, is_udplite);
V
Vlad Yasevich 已提交
1012 1013 1014 1015 1016 1017 1018 1019 1020 1021 1022 1023 1024 1025 1026 1027 1028 1029 1030 1031 1032 1033 1034 1035
	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
1036 1037 1038 1039 1040 1041
out:
	up->len = 0;
	up->pending = 0;
	return err;
}

1042
int udpv6_sendmsg(struct sock *sk, struct msghdr *msg, size_t len)
L
Linus Torvalds 已提交
1043 1044 1045 1046 1047
{
	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);
1048
	DECLARE_SOCKADDR(struct sockaddr_in6 *, sin6, msg->msg_name);
1049
	struct in6_addr *daddr, *final_p, final;
L
Linus Torvalds 已提交
1050
	struct ipv6_txoptions *opt = NULL;
1051
	struct ipv6_txoptions *opt_to_free = NULL;
L
Linus Torvalds 已提交
1052
	struct ip6_flowlabel *flowlabel = NULL;
1053
	struct flowi6 fl6;
L
Linus Torvalds 已提交
1054 1055 1056 1057
	struct dst_entry *dst;
	int addr_len = msg->msg_namelen;
	int ulen = len;
	int hlimit = -1;
1058
	int tclass = -1;
1059
	int dontfrag = -1;
L
Linus Torvalds 已提交
1060 1061
	int corkreq = up->corkflag || msg->msg_flags&MSG_MORE;
	int err;
1062
	int connected = 0;
W
Wang Chen 已提交
1063
	int is_udplite = IS_UDPLITE(sk);
1064
	int (*getfrag)(void *, char *, int, int, int, struct sk_buff *);
1065
	struct sockcm_cookie sockc;
L
Linus Torvalds 已提交
1066 1067 1068 1069 1070 1071 1072 1073 1074 1075 1076 1077 1078 1079 1080 1081 1082 1083 1084 1085 1086 1087 1088 1089 1090

	/* 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;
1091
		daddr = &sk->sk_v6_daddr;
1092
	} else
L
Linus Torvalds 已提交
1093 1094 1095
		daddr = NULL;

	if (daddr) {
B
Brian Haley 已提交
1096
		if (ipv6_addr_v4mapped(daddr)) {
L
Linus Torvalds 已提交
1097 1098
			struct sockaddr_in sin;
			sin.sin_family = AF_INET;
E
Eric Dumazet 已提交
1099
			sin.sin_port = sin6 ? sin6->sin6_port : inet->inet_dport;
L
Linus Torvalds 已提交
1100 1101 1102 1103 1104 1105
			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;
1106
			return udp_sendmsg(sk, msg, len);
L
Linus Torvalds 已提交
1107 1108 1109 1110
		}
	}

	if (up->pending == AF_INET)
1111
		return udp_sendmsg(sk, msg, len);
L
Linus Torvalds 已提交
1112 1113

	/* Rough check on arithmetic overflow,
1114
	   better check is made in ip6_append_data().
L
Linus Torvalds 已提交
1115 1116 1117
	   */
	if (len > INT_MAX - sizeof(struct udphdr))
		return -EMSGSIZE;
1118

1119
	getfrag  =  is_udplite ?  udplite_getfrag : ip_generic_getfrag;
L
Linus Torvalds 已提交
1120 1121 1122 1123 1124 1125 1126 1127 1128 1129 1130 1131 1132 1133 1134 1135 1136 1137
	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);

1138
	memset(&fl6, 0, sizeof(fl6));
L
Linus Torvalds 已提交
1139 1140 1141 1142 1143

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

1144
		fl6.fl6_dport = sin6->sin6_port;
L
Linus Torvalds 已提交
1145 1146 1147
		daddr = &sin6->sin6_addr;

		if (np->sndflow) {
1148 1149 1150
			fl6.flowlabel = sin6->sin6_flowinfo&IPV6_FLOWINFO_MASK;
			if (fl6.flowlabel&IPV6_FLOWLABEL_MASK) {
				flowlabel = fl6_sock_lookup(sk, fl6.flowlabel);
1151
				if (!flowlabel)
L
Linus Torvalds 已提交
1152 1153 1154 1155 1156 1157 1158 1159 1160
					return -EINVAL;
			}
		}

		/*
		 * Otherwise it will be difficult to maintain
		 * sk->sk_dst_cache.
		 */
		if (sk->sk_state == TCP_ESTABLISHED &&
1161 1162
		    ipv6_addr_equal(daddr, &sk->sk_v6_daddr))
			daddr = &sk->sk_v6_daddr;
L
Linus Torvalds 已提交
1163 1164 1165

		if (addr_len >= sizeof(struct sockaddr_in6) &&
		    sin6->sin6_scope_id &&
1166
		    __ipv6_addr_needs_scope_id(__ipv6_addr_type(daddr)))
1167
			fl6.flowi6_oif = sin6->sin6_scope_id;
L
Linus Torvalds 已提交
1168 1169 1170 1171
	} else {
		if (sk->sk_state != TCP_ESTABLISHED)
			return -EDESTADDRREQ;

1172
		fl6.fl6_dport = inet->inet_dport;
1173
		daddr = &sk->sk_v6_daddr;
1174
		fl6.flowlabel = np->flow_label;
1175
		connected = 1;
L
Linus Torvalds 已提交
1176 1177
	}

1178 1179
	if (!fl6.flowi6_oif)
		fl6.flowi6_oif = sk->sk_bound_dev_if;
L
Linus Torvalds 已提交
1180

1181 1182
	if (!fl6.flowi6_oif)
		fl6.flowi6_oif = np->sticky_pktinfo.ipi6_ifindex;
1183

1184
	fl6.flowi6_mark = sk->sk_mark;
1185
	sockc.tsflags = sk->sk_tsflags;
B
Brian Haley 已提交
1186

L
Linus Torvalds 已提交
1187 1188 1189 1190 1191
	if (msg->msg_controllen) {
		opt = &opt_space;
		memset(opt, 0, sizeof(struct ipv6_txoptions));
		opt->tot_len = sizeof(*opt);

1192
		err = ip6_datagram_send_ctl(sock_net(sk), sk, msg, &fl6, opt,
1193 1194
					    &hlimit, &tclass, &dontfrag,
					    &sockc);
L
Linus Torvalds 已提交
1195 1196 1197 1198
		if (err < 0) {
			fl6_sock_release(flowlabel);
			return err;
		}
1199 1200
		if ((fl6.flowlabel&IPV6_FLOWLABEL_MASK) && !flowlabel) {
			flowlabel = fl6_sock_lookup(sk, fl6.flowlabel);
1201
			if (!flowlabel)
L
Linus Torvalds 已提交
1202 1203 1204 1205
				return -EINVAL;
		}
		if (!(opt->opt_nflen|opt->opt_flen))
			opt = NULL;
1206
		connected = 0;
L
Linus Torvalds 已提交
1207
	}
1208 1209 1210 1211
	if (!opt) {
		opt = txopt_get(np);
		opt_to_free = opt;
	}
1212 1213 1214
	if (flowlabel)
		opt = fl6_merge_options(&opt_space, flowlabel, opt);
	opt = ipv6_fixup_options(&opt_space, opt);
L
Linus Torvalds 已提交
1215

1216
	fl6.flowi6_proto = sk->sk_protocol;
1217
	if (!ipv6_addr_any(daddr))
A
Alexey Dobriyan 已提交
1218
		fl6.daddr = *daddr;
1219
	else
1220 1221
		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
1222
		fl6.saddr = np->saddr;
1223
	fl6.fl6_sport = inet->inet_sport;
1224

1225
	final_p = fl6_update_dst(&fl6, opt, &final);
1226
	if (final_p)
1227
		connected = 0;
L
Linus Torvalds 已提交
1228

1229 1230
	if (!fl6.flowi6_oif && ipv6_addr_is_multicast(&fl6.daddr)) {
		fl6.flowi6_oif = np->mcast_oif;
1231
		connected = 0;
1232 1233
	} else if (!fl6.flowi6_oif)
		fl6.flowi6_oif = np->ucast_oif;
L
Linus Torvalds 已提交
1234

1235
	security_sk_classify_flow(sk, flowi6_to_flowi(&fl6));
V
Venkat Yekkirala 已提交
1236

1237
	dst = ip6_sk_dst_lookup_flow(sk, &fl6, final_p);
1238 1239 1240
	if (IS_ERR(dst)) {
		err = PTR_ERR(dst);
		dst = NULL;
L
Linus Torvalds 已提交
1241
		goto out;
1242
	}
L
Linus Torvalds 已提交
1243

1244 1245
	if (hlimit < 0)
		hlimit = ip6_sk_dst_hoplimit(np, &fl6, dst);
L
Linus Torvalds 已提交
1246

G
Gerrit Renker 已提交
1247
	if (tclass < 0)
1248 1249
		tclass = np->tclass;

L
Linus Torvalds 已提交
1250 1251 1252 1253
	if (msg->msg_flags&MSG_CONFIRM)
		goto do_confirm;
back_from_confirm:

1254 1255 1256 1257 1258 1259 1260
	/* 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,
1261
				   msg->msg_flags, dontfrag, &sockc);
1262 1263 1264 1265 1266 1267
		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
1268 1269 1270 1271 1272 1273
	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);

1274
		net_dbg_ratelimited("udp cork app bug 2\n");
L
Linus Torvalds 已提交
1275 1276 1277 1278 1279 1280 1281
		err = -EINVAL;
		goto out;
	}

	up->pending = AF_INET6;

do_append_data:
1282 1283
	if (dontfrag < 0)
		dontfrag = np->dontfrag;
L
Linus Torvalds 已提交
1284
	up->len += ulen;
1285
	err = ip6_append_data(sk, getfrag, msg, ulen,
1286
		sizeof(struct udphdr), hlimit, tclass, opt, &fl6,
1287
		(struct rt6_info *)dst,
1288 1289
		corkreq ? msg->msg_flags|MSG_MORE : msg->msg_flags, dontfrag,
		&sockc);
L
Linus Torvalds 已提交
1290 1291 1292
	if (err)
		udp_v6_flush_pending_frames(sk);
	else if (!corkreq)
1293
		err = udp_v6_push_pending_frames(sk);
H
Herbert Xu 已提交
1294 1295
	else if (unlikely(skb_queue_empty(&sk->sk_write_queue)))
		up->pending = 0;
L
Linus Torvalds 已提交
1296

1297 1298 1299 1300 1301
	if (err > 0)
		err = np->recverr ? net_xmit_errno(err) : 0;
	release_sock(sk);

release_dst:
1302 1303 1304
	if (dst) {
		if (connected) {
			ip6_dst_store(sk, dst,
1305 1306
				      ipv6_addr_equal(&fl6.daddr, &sk->sk_v6_daddr) ?
				      &sk->sk_v6_daddr : NULL,
1307
#ifdef CONFIG_IPV6_SUBTREES
1308
				      ipv6_addr_equal(&fl6.saddr, &np->saddr) ?
1309 1310 1311
				      &np->saddr :
#endif
				      NULL);
1312 1313 1314
		} else {
			dst_release(dst);
		}
1315
		dst = NULL;
1316 1317
	}

L
Linus Torvalds 已提交
1318
out:
1319
	dst_release(dst);
L
Linus Torvalds 已提交
1320
	fl6_sock_release(flowlabel);
1321
	txopt_put(opt_to_free);
1322
	if (!err)
L
Linus Torvalds 已提交
1323
		return len;
1324 1325 1326 1327 1328 1329 1330 1331
	/*
	 * 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)) {
1332 1333
		UDP6_INC_STATS_USER(sock_net(sk),
				UDP_MIB_SNDBUFERRORS, is_udplite);
1334
	}
L
Linus Torvalds 已提交
1335 1336 1337 1338 1339 1340 1341 1342 1343 1344
	return err;

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

1345
void udpv6_destroy_sock(struct sock *sk)
L
Linus Torvalds 已提交
1346
{
T
Tom Parkin 已提交
1347
	struct udp_sock *up = udp_sk(sk);
L
Linus Torvalds 已提交
1348 1349 1350 1351
	lock_sock(sk);
	udp_v6_flush_pending_frames(sk);
	release_sock(sk);

T
Tom Parkin 已提交
1352 1353 1354 1355 1356 1357 1358
	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
1359 1360 1361 1362 1363 1364
	inet6_destroy_sock(sk);
}

/*
 *	Socket option code for UDP
 */
1365
int udpv6_setsockopt(struct sock *sk, int level, int optname,
1366
		     char __user *optval, unsigned int optlen)
1367
{
1368
	if (level == SOL_UDP  ||  level == SOL_UDPLITE)
1369 1370
		return udp_lib_setsockopt(sk, level, optname, optval, optlen,
					  udp_v6_push_pending_frames);
1371
	return ipv6_setsockopt(sk, level, optname, optval, optlen);
1372 1373 1374
}

#ifdef CONFIG_COMPAT
1375
int compat_udpv6_setsockopt(struct sock *sk, int level, int optname,
1376
			    char __user *optval, unsigned int optlen)
1377
{
1378
	if (level == SOL_UDP  ||  level == SOL_UDPLITE)
1379 1380
		return udp_lib_setsockopt(sk, level, optname, optval, optlen,
					  udp_v6_push_pending_frames);
1381
	return compat_ipv6_setsockopt(sk, level, optname, optval, optlen);
1382 1383 1384
}
#endif

1385 1386
int udpv6_getsockopt(struct sock *sk, int level, int optname,
		     char __user *optval, int __user *optlen)
1387
{
1388
	if (level == SOL_UDP  ||  level == SOL_UDPLITE)
1389
		return udp_lib_getsockopt(sk, level, optname, optval, optlen);
1390
	return ipv6_getsockopt(sk, level, optname, optval, optlen);
1391 1392 1393
}

#ifdef CONFIG_COMPAT
1394 1395
int compat_udpv6_getsockopt(struct sock *sk, int level, int optname,
			    char __user *optval, int __user *optlen)
1396
{
1397
	if (level == SOL_UDP  ||  level == SOL_UDPLITE)
1398
		return udp_lib_getsockopt(sk, level, optname, optval, optlen);
1399
	return compat_ipv6_getsockopt(sk, level, optname, optval, optlen);
1400 1401 1402
}
#endif

1403
static const struct inet6_protocol udpv6_protocol = {
L
Linus Torvalds 已提交
1404 1405 1406 1407 1408 1409 1410
	.handler	=	udpv6_rcv,
	.err_handler	=	udpv6_err,
	.flags		=	INET6_PROTO_NOPOLICY|INET6_PROTO_FINAL,
};

/* ------------------------------------------------------------------------ */
#ifdef CONFIG_PROC_FS
1411
int udp6_seq_show(struct seq_file *seq, void *v)
L
Linus Torvalds 已提交
1412
{
1413 1414 1415 1416 1417 1418 1419 1420 1421
	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
1422 1423 1424
	return 0;
}

1425 1426 1427 1428 1429 1430 1431 1432
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
1433 1434 1435
static struct udp_seq_afinfo udp6_seq_afinfo = {
	.name		= "udp6",
	.family		= AF_INET6,
1436
	.udp_table	= &udp_table,
1437
	.seq_fops	= &udp6_afinfo_seq_fops,
1438 1439 1440
	.seq_ops	= {
		.show		= udp6_seq_show,
	},
L
Linus Torvalds 已提交
1441 1442
};

1443
int __net_init udp6_proc_init(struct net *net)
L
Linus Torvalds 已提交
1444
{
1445
	return udp_proc_register(net, &udp6_seq_afinfo);
L
Linus Torvalds 已提交
1446 1447
}

I
Ian Morris 已提交
1448 1449
void udp6_proc_exit(struct net *net)
{
1450
	udp_proc_unregister(net, &udp6_seq_afinfo);
L
Linus Torvalds 已提交
1451 1452 1453
}
#endif /* CONFIG_PROC_FS */

E
Eric Dumazet 已提交
1454 1455 1456 1457 1458 1459 1460 1461 1462 1463 1464
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
1465 1466 1467
/* ------------------------------------------------------------------------ */

struct proto udpv6_prot = {
1468 1469
	.name		   = "UDPv6",
	.owner		   = THIS_MODULE,
1470
	.close		   = udp_lib_close,
1471 1472 1473 1474 1475 1476 1477 1478
	.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,
1479
	.backlog_rcv	   = __udpv6_queue_rcv_skb,
1480 1481
	.hash		   = udp_lib_hash,
	.unhash		   = udp_lib_unhash,
E
Eric Dumazet 已提交
1482
	.rehash		   = udp_v6_rehash,
1483
	.get_port	   = udp_v6_get_port,
H
Hideo Aoki 已提交
1484 1485 1486 1487
	.memory_allocated  = &udp_memory_allocated,
	.sysctl_mem	   = sysctl_udp_mem,
	.sysctl_wmem	   = &sysctl_udp_wmem_min,
	.sysctl_rmem	   = &sysctl_udp_rmem_min,
1488
	.obj_size	   = sizeof(struct udp6_sock),
1489
	.slab_flags	   = SLAB_DESTROY_BY_RCU,
1490
	.h.udp_table	   = &udp_table,
1491
#ifdef CONFIG_COMPAT
1492 1493
	.compat_setsockopt = compat_udpv6_setsockopt,
	.compat_getsockopt = compat_udpv6_getsockopt,
1494
#endif
E
Eric Dumazet 已提交
1495
	.clear_sk	   = udp_v6_clear_sk,
L
Linus Torvalds 已提交
1496 1497 1498 1499 1500 1501 1502 1503 1504 1505
};

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

1506
int __init udpv6_init(void)
L
Linus Torvalds 已提交
1507
{
1508 1509
	int ret;

1510 1511
	ret = inet6_add_protocol(&udpv6_protocol, IPPROTO_UDP);
	if (ret)
1512
		goto out;
1513

1514 1515 1516 1517 1518 1519 1520 1521 1522 1523 1524
	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;
}

1525
void udpv6_exit(void)
1526 1527 1528
{
	inet6_unregister_protosw(&udpv6_protosw);
	inet6_del_protocol(&udpv6_protocol, IPPROTO_UDP);
L
Linus Torvalds 已提交
1529
}