udp.c 44.5 KB
Newer Older
1
// SPDX-License-Identifier: GPL-2.0-or-later
L
Linus Torvalds 已提交
2 3
/*
 *	UDP over IPv6
4
 *	Linux INET6 implementation
L
Linus Torvalds 已提交
5 6
 *
 *	Authors:
7
 *	Pedro Roque		<roque@di.fc.ul.pt>
L
Linus Torvalds 已提交
8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30
 *
 *	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.
 */

#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>
31
#include <linux/module.h>
32
#include <linux/skbuff.h>
33
#include <linux/slab.h>
34
#include <linux/uaccess.h>
35
#include <linux/indirect_call_wrapper.h>
L
Linus Torvalds 已提交
36

37
#include <net/addrconf.h>
L
Linus Torvalds 已提交
38 39 40 41 42
#include <net/ndisc.h>
#include <net/protocol.h>
#include <net/transp_v6.h>
#include <net/ip6_route.h>
#include <net/raw.h>
43
#include <net/tcp_states.h>
L
Linus Torvalds 已提交
44
#include <net/ip6_checksum.h>
45
#include <net/ip6_tunnel.h>
L
Linus Torvalds 已提交
46
#include <net/xfrm.h>
47
#include <net/inet_hashtables.h>
48
#include <net/inet6_hashtables.h>
49
#include <net/busy_poll.h>
50
#include <net/sock_reuseport.h>
L
Linus Torvalds 已提交
51 52 53

#include <linux/proc_fs.h>
#include <linux/seq_file.h>
54
#include <trace/events/skb.h>
55
#include "udp_impl.h"
L
Linus Torvalds 已提交
56

57 58 59 60 61
static u32 udp6_ehashfn(const struct net *net,
			const struct in6_addr *laddr,
			const u16 lport,
			const struct in6_addr *faddr,
			const __be16 fport)
62
{
63 64 65 66 67 68 69 70 71 72 73 74 75
	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);

76
	return __inet6_ehashfn(lhash, lport, fhash, fport,
77
			       udp_ipv6_hash_secret + net_hash_mix(net));
78 79
}

80
int udp_v6_get_port(struct sock *sk, unsigned short snum)
L
Linus Torvalds 已提交
81
{
E
Eric Dumazet 已提交
82
	unsigned int hash2_nulladdr =
83
		ipv6_portaddr_hash(sock_net(sk), &in6addr_any, snum);
84
	unsigned int hash2_partial =
85
		ipv6_portaddr_hash(sock_net(sk), &sk->sk_v6_rcv_saddr, 0);
E
Eric Dumazet 已提交
86

87
	/* precompute partial secondary hash */
E
Eric Dumazet 已提交
88
	udp_sk(sk)->udp_portaddr_hash = hash2_partial;
89
	return udp_lib_get_port(sk, snum, hash2_nulladdr);
L
Linus Torvalds 已提交
90 91
}

92
void udp_v6_rehash(struct sock *sk)
E
Eric Dumazet 已提交
93
{
94
	u16 new_hash = ipv6_portaddr_hash(sock_net(sk),
95
					  &sk->sk_v6_rcv_saddr,
E
Eric Dumazet 已提交
96 97 98 99 100
					  inet_sk(sk)->inet_num);

	udp_lib_rehash(sk, new_hash);
}

101 102 103
static int compute_score(struct sock *sk, struct net *net,
			 const struct in6_addr *saddr, __be16 sport,
			 const struct in6_addr *daddr, unsigned short hnum,
104
			 int dif, int sdif)
105
{
106 107
	int score;
	struct inet_sock *inet;
108
	bool dev_match;
109 110 111 112 113 114

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

115 116 117
	if (!ipv6_addr_equal(&sk->sk_v6_rcv_saddr, daddr))
		return -1;

118 119 120 121 122 123 124 125 126 127 128 129 130 131 132
	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))
			return -1;
		score++;
	}

133 134 135 136
	dev_match = udp_sk_bound_dev_eq(net, sk->sk_bound_dev_if, dif, sdif);
	if (!dev_match)
		return -1;
	score++;
137

138
	if (READ_ONCE(sk->sk_incoming_cpu) == raw_smp_processor_id())
139 140
		score++;

141 142 143
	return score;
}

144 145 146 147 148 149
static struct sock *lookup_reuseport(struct net *net, struct sock *sk,
				     struct sk_buff *skb,
				     const struct in6_addr *saddr,
				     __be16 sport,
				     const struct in6_addr *daddr,
				     unsigned int hnum)
150 151 152 153 154 155 156 157 158 159 160 161
{
	struct sock *reuse_sk = NULL;
	u32 hash;

	if (sk->sk_reuseport && sk->sk_state != TCP_ESTABLISHED) {
		hash = udp6_ehashfn(net, daddr, hnum, saddr, sport);
		reuse_sk = reuseport_select_sock(sk, hash, skb,
						 sizeof(struct udphdr));
	}
	return reuse_sk;
}

162
/* called with rcu_read_lock() */
E
Eric Dumazet 已提交
163 164
static struct sock *udp6_lib_lookup2(struct net *net,
		const struct in6_addr *saddr, __be16 sport,
165
		const struct in6_addr *daddr, unsigned int hnum,
166 167
		int dif, int sdif, struct udp_hslot *hslot2,
		struct sk_buff *skb)
E
Eric Dumazet 已提交
168 169
{
	struct sock *sk, *result;
P
Paolo Abeni 已提交
170
	int score, badness;
E
Eric Dumazet 已提交
171 172 173

	result = NULL;
	badness = -1;
174
	udp_portaddr_for_each_entry_rcu(sk, &hslot2->head) {
175
		score = compute_score(sk, net, saddr, sport,
176
				      daddr, hnum, dif, sdif);
E
Eric Dumazet 已提交
177
		if (score > badness) {
178 179
			result = lookup_reuseport(net, sk, skb,
						  saddr, sport, daddr, hnum);
180 181
			/* Fall back to scoring if group has connections */
			if (result && !reuseport_has_conns(sk, false))
182 183
				return result;

184
			result = result ? : sk;
185
			badness = score;
E
Eric Dumazet 已提交
186 187 188 189 190
		}
	}
	return result;
}

191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210
static inline struct sock *udp6_lookup_run_bpf(struct net *net,
					       struct udp_table *udptable,
					       struct sk_buff *skb,
					       const struct in6_addr *saddr,
					       __be16 sport,
					       const struct in6_addr *daddr,
					       u16 hnum)
{
	struct sock *sk, *reuse_sk;
	bool no_reuseport;

	if (udptable != &udp_table)
		return NULL; /* only UDP is supported */

	no_reuseport = bpf_sk_lookup_run_v6(net, IPPROTO_UDP,
					    saddr, sport, daddr, hnum, &sk);
	if (no_reuseport || IS_ERR_OR_NULL(sk))
		return sk;

	reuse_sk = lookup_reuseport(net, sk, skb, saddr, sport, daddr, hnum);
211
	if (reuse_sk && !reuseport_has_conns(sk, false))
212 213 214 215
		sk = reuse_sk;
	return sk;
}

216
/* rcu_read_lock() must be held */
217
struct sock *__udp6_lib_lookup(struct net *net,
218 219 220 221
			       const struct in6_addr *saddr, __be16 sport,
			       const struct in6_addr *daddr, __be16 dport,
			       int dif, int sdif, struct udp_table *udptable,
			       struct sk_buff *skb)
L
Linus Torvalds 已提交
222 223
{
	unsigned short hnum = ntohs(dport);
224 225
	unsigned int hash2, slot2;
	struct udp_hslot *hslot2;
226
	struct sock *result, *sk;
227

228 229 230 231
	hash2 = ipv6_portaddr_hash(net, daddr, hnum);
	slot2 = hash2 & udptable->mask;
	hslot2 = &udptable->hash2[slot2];

232
	/* Lookup connected or non-wildcard sockets */
233
	result = udp6_lib_lookup2(net, saddr, sport,
234
				  daddr, hnum, dif, sdif,
235
				  hslot2, skb);
236 237 238 239 240 241 242 243 244 245 246 247
	if (!IS_ERR_OR_NULL(result) && result->sk_state == TCP_ESTABLISHED)
		goto done;

	/* Lookup redirect from BPF */
	if (static_branch_unlikely(&bpf_sk_lookup_enabled)) {
		sk = udp6_lookup_run_bpf(net, udptable, skb,
					 saddr, sport, daddr, hnum);
		if (sk) {
			result = sk;
			goto done;
		}
	}
248

249 250 251
	/* Got non-wildcard socket or error on first lookup */
	if (result)
		goto done;
E
Eric Dumazet 已提交
252

253 254 255 256 257 258 259 260 261
	/* Lookup wildcard sockets */
	hash2 = ipv6_portaddr_hash(net, &in6addr_any, hnum);
	slot2 = hash2 & udptable->mask;
	hslot2 = &udptable->hash2[slot2];

	result = udp6_lib_lookup2(net, saddr, sport,
				  &in6addr_any, hnum, dif, sdif,
				  hslot2, skb);
done:
262
	if (IS_ERR(result))
263
		return NULL;
L
Linus Torvalds 已提交
264 265
	return result;
}
266
EXPORT_SYMBOL_GPL(__udp6_lib_lookup);
L
Linus Torvalds 已提交
267

268 269
static struct sock *__udp6_lib_lookup_skb(struct sk_buff *skb,
					  __be16 sport, __be16 dport,
270
					  struct udp_table *udptable)
271
{
272
	const struct ipv6hdr *iph = ipv6_hdr(skb);
273

274
	return __udp6_lib_lookup(dev_net(skb->dev), &iph->saddr, sport,
E
Eric Dumazet 已提交
275
				 &iph->daddr, dport, inet6_iif(skb),
276
				 inet6_sdif(skb), udptable, skb);
277 278
}

279 280 281 282 283
struct sock *udp6_lib_lookup_skb(struct sk_buff *skb,
				 __be16 sport, __be16 dport)
{
	const struct ipv6hdr *iph = ipv6_hdr(skb);

284
	return __udp6_lib_lookup(dev_net(skb->dev), &iph->saddr, sport,
285
				 &iph->daddr, dport, inet6_iif(skb),
286
				 inet6_sdif(skb), &udp_table, NULL);
287 288 289
}
EXPORT_SYMBOL_GPL(udp6_lib_lookup_skb);

290 291 292
/* Must be called under rcu_read_lock().
 * Does increment socket refcount.
 */
293
#if IS_ENABLED(CONFIG_NF_TPROXY_IPV6) || IS_ENABLED(CONFIG_NF_SOCKET_IPV6)
294 295 296
struct sock *udp6_lib_lookup(struct net *net, const struct in6_addr *saddr, __be16 sport,
			     const struct in6_addr *daddr, __be16 dport, int dif)
{
297 298 299
	struct sock *sk;

	sk =  __udp6_lib_lookup(net, saddr, sport, daddr, dport,
300
				dif, 0, &udp_table, NULL);
301
	if (sk && !refcount_inc_not_zero(&sk->sk_refcnt))
302 303
		sk = NULL;
	return sk;
304 305
}
EXPORT_SYMBOL_GPL(udp6_lib_lookup);
306
#endif
307

P
Paolo Abeni 已提交
308 309 310 311 312 313 314 315 316
/* do not use the scratch area len for jumbogram: their length execeeds the
 * scratch area space; note that the IP6CB flags is still in the first
 * cacheline, so checking for jumbograms is cheap
 */
static int udp6_skb_len(struct sk_buff *skb)
{
	return unlikely(inet6_is_jumbogram(skb)) ? skb->len : udp_skb_len(skb);
}

L
Linus Torvalds 已提交
317
/*
318 319
 *	This should be easy, if there is something there we
 *	return it, otherwise we block.
L
Linus Torvalds 已提交
320 321
 */

322
int udpv6_recvmsg(struct sock *sk, struct msghdr *msg, size_t len,
L
Linus Torvalds 已提交
323 324 325 326
		  int noblock, int flags, int *addr_len)
{
	struct ipv6_pinfo *np = inet6_sk(sk);
	struct inet_sock *inet = inet_sk(sk);
327
	struct sk_buff *skb;
328
	unsigned int ulen, copied;
329
	int off, err, peeking = flags & MSG_PEEK;
330
	int is_udplite = IS_UDPLITE(sk);
331
	struct udp_mib __percpu *mib;
332
	bool checksum_valid = false;
333
	int is_udp4;
L
Linus Torvalds 已提交
334 335

	if (flags & MSG_ERRQUEUE)
336
		return ipv6_recv_error(sk, msg, len, addr_len);
L
Linus Torvalds 已提交
337

338
	if (np->rxpmtu && np->rxopt.bits.rxpmtu)
339
		return ipv6_recv_rxpmtu(sk, msg, len, addr_len);
340

L
Linus Torvalds 已提交
341
try_again:
342
	off = sk_peek_offset(sk, flags);
343
	skb = __skb_recv_udp(sk, flags, noblock, &off, &err);
L
Linus Torvalds 已提交
344
	if (!skb)
345
		return err;
L
Linus Torvalds 已提交
346

P
Paolo Abeni 已提交
347
	ulen = udp6_skb_len(skb);
348
	copied = len;
349 350
	if (copied > ulen - off)
		copied = ulen - off;
351
	else if (copied < ulen)
352
		msg->msg_flags |= MSG_TRUNC;
L
Linus Torvalds 已提交
353

354
	is_udp4 = (skb->protocol == htons(ETH_P_IP));
355
	mib = __UDPX_MIB(sk, is_udp4);
356

357
	/*
358 359 360
	 * 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.
361 362
	 */

363 364
	if (copied < ulen || peeking ||
	    (is_udplite && UDP_SKB_CB(skb)->partial_cov)) {
365 366
		checksum_valid = udp_skb_csum_unnecessary(skb) ||
				!__udp_lib_checksum_complete(skb);
367
		if (!checksum_valid)
L
Linus Torvalds 已提交
368
			goto csum_copy_err;
369 370
	}

371 372 373 374 375 376
	if (checksum_valid || udp_skb_csum_unnecessary(skb)) {
		if (udp_skb_is_linear(skb))
			err = copy_linear_skb(skb, copied, off, &msg->msg_iter);
		else
			err = skb_copy_datagram_msg(skb, off, msg, copied);
	} else {
377
		err = skb_copy_and_csum_datagram_msg(skb, off, msg);
L
Linus Torvalds 已提交
378 379 380
		if (err == -EINVAL)
			goto csum_copy_err;
	}
381
	if (unlikely(err)) {
382
		if (!peeking) {
383
			atomic_inc(&sk->sk_drops);
384
			SNMP_INC_STATS(mib, UDP_MIB_INERRORS);
385
		}
386
		kfree_skb(skb);
387
		return err;
388
	}
389
	if (!peeking)
390
		SNMP_INC_STATS(mib, UDP_MIB_INDATAGRAMS);
391

392
	sock_recv_ts_and_drops(msg, sk, skb);
L
Linus Torvalds 已提交
393 394 395

	/* Copy the address. */
	if (msg->msg_name) {
396
		DECLARE_SOCKADDR(struct sockaddr_in6 *, sin6, msg->msg_name);
L
Linus Torvalds 已提交
397
		sin6->sin6_family = AF_INET6;
398
		sin6->sin6_port = udp_hdr(skb)->source;
L
Linus Torvalds 已提交
399 400
		sin6->sin6_flowinfo = 0;

401
		if (is_udp4) {
B
Brian Haley 已提交
402 403
			ipv6_addr_set_v4mapped(ip_hdr(skb)->saddr,
					       &sin6->sin6_addr);
404 405
			sin6->sin6_scope_id = 0;
		} else {
A
Alexey Dobriyan 已提交
406
			sin6->sin6_addr = ipv6_hdr(skb)->saddr;
407 408
			sin6->sin6_scope_id =
				ipv6_iface_scope_id(&sin6->sin6_addr,
409
						    inet6_iif(skb));
L
Linus Torvalds 已提交
410
		}
411
		*addr_len = sizeof(*sin6);
D
Daniel Borkmann 已提交
412 413 414 415

		if (cgroup_bpf_enabled)
			BPF_CGROUP_RUN_PROG_UDP6_RECVMSG_LOCK(sk,
						(struct sockaddr *)sin6);
L
Linus Torvalds 已提交
416
	}
417

P
Paolo Abeni 已提交
418 419 420
	if (udp_sk(sk)->gro_enabled)
		udp_cmsg_recv(msg, sk, skb);

421 422 423
	if (np->rxopt.all)
		ip6_datagram_recv_common_ctl(sk, msg, skb);

424
	if (is_udp4) {
L
Linus Torvalds 已提交
425
		if (inet->cmsg_flags)
426
			ip_cmsg_recv_offset(msg, sk, skb,
E
Eric Dumazet 已提交
427
					    sizeof(struct udphdr), off);
L
Linus Torvalds 已提交
428 429
	} else {
		if (np->rxopt.all)
430
			ip6_datagram_recv_specific_ctl(sk, msg, skb);
431
	}
L
Linus Torvalds 已提交
432

433
	err = copied;
L
Linus Torvalds 已提交
434
	if (flags & MSG_TRUNC)
435
		err = ulen;
L
Linus Torvalds 已提交
436

437
	skb_consume_udp(sk, skb, peeking ? -err : err);
L
Linus Torvalds 已提交
438 439 440
	return err;

csum_copy_err:
441 442
	if (!__sk_queue_drop_skb(sk, &udp_sk(sk)->reader_queue, skb, flags,
				 udp_skb_destructor)) {
443 444
		SNMP_INC_STATS(mib, UDP_MIB_CSUMERRORS);
		SNMP_INC_STATS(mib, UDP_MIB_INERRORS);
445
	}
446
	kfree_skb(skb);
L
Linus Torvalds 已提交
447

448 449
	/* starting over for a new packet, but check if we need to yield */
	cond_resched();
450
	msg->msg_flags &= ~MSG_TRUNC;
L
Linus Torvalds 已提交
451 452 453
	goto try_again;
}

454 455 456
DEFINE_STATIC_KEY_FALSE(udpv6_encap_needed_key);
void udpv6_encap_enable(void)
{
P
Paolo Abeni 已提交
457
	static_branch_inc(&udpv6_encap_needed_key);
458 459 460
}
EXPORT_SYMBOL(udpv6_encap_enable);

461 462 463 464 465
/* Handler for tunnels with arbitrary destination ports: no socket lookup, go
 * through error handlers in encapsulations looking for a match.
 */
static int __udp6_lib_err_encap_no_sk(struct sk_buff *skb,
				      struct inet6_skb_parm *opt,
466
				      u8 type, u8 code, int offset, __be32 info)
467 468 469 470 471
{
	int i;

	for (i = 0; i < MAX_IPTUN_ENCAP_OPS; i++) {
		int (*handler)(struct sk_buff *skb, struct inet6_skb_parm *opt,
472 473
			       u8 type, u8 code, int offset, __be32 info);
		const struct ip6_tnl_encap_ops *encap;
474

475 476
		encap = rcu_dereference(ip6tun_encaps[i]);
		if (!encap)
477
			continue;
478
		handler = encap->err_handler;
479 480 481 482 483 484 485
		if (handler && !handler(skb, opt, type, code, offset, info))
			return 0;
	}

	return -ENOENT;
}

486 487 488 489 490 491 492
/* Try to match ICMP errors to UDP tunnels by looking up a socket without
 * reversing source and destination port: this will match tunnels that force the
 * same destination port on both endpoints (e.g. VXLAN, GENEVE). Note that
 * lwtunnels might actually break this assumption by being configured with
 * different destination ports on endpoints, in this case we won't be able to
 * trace ICMP messages back to them.
 *
493 494 495 496
 * If this doesn't match any socket, probe tunnels with arbitrary destination
 * ports (e.g. FoU, GUE): there, the receiving socket is useless, as the port
 * we've sent packets to won't necessarily match the local destination port.
 *
497 498 499
 * Then ask the tunnel implementation to match the error against a valid
 * association.
 *
500 501
 * Return an error if we can't find a match, the socket if we need further
 * processing, zero otherwise.
502 503 504 505 506
 */
static struct sock *__udp6_lib_err_encap(struct net *net,
					 const struct ipv6hdr *hdr, int offset,
					 struct udphdr *uh,
					 struct udp_table *udptable,
507 508 509
					 struct sk_buff *skb,
					 struct inet6_skb_parm *opt,
					 u8 type, u8 code, __be32 info)
510 511 512 513 514 515 516 517 518 519 520 521 522
{
	int network_offset, transport_offset;
	struct sock *sk;

	network_offset = skb_network_offset(skb);
	transport_offset = skb_transport_offset(skb);

	/* Network header needs to point to the outer IPv6 header inside ICMP */
	skb_reset_network_header(skb);

	/* Transport header needs to point to the UDP header */
	skb_set_transport_header(skb, offset);

523 524 525 526 527 528 529 530 531 532 533 534 535 536 537 538
	sk = __udp6_lib_lookup(net, &hdr->daddr, uh->source,
			       &hdr->saddr, uh->dest,
			       inet6_iif(skb), 0, udptable, skb);
	if (sk) {
		int (*lookup)(struct sock *sk, struct sk_buff *skb);
		struct udp_sock *up = udp_sk(sk);

		lookup = READ_ONCE(up->encap_err_lookup);
		if (!lookup || lookup(sk, skb))
			sk = NULL;
	}

	if (!sk) {
		sk = ERR_PTR(__udp6_lib_err_encap_no_sk(skb, opt, type, code,
							offset, info));
	}
539 540 541

	skb_set_transport_header(skb, transport_offset);
	skb_set_network_header(skb, network_offset);
542

543 544 545
	return sk;
}

546 547 548
int __udp6_lib_err(struct sk_buff *skb, struct inet6_skb_parm *opt,
		   u8 type, u8 code, int offset, __be32 info,
		   struct udp_table *udptable)
L
Linus Torvalds 已提交
549 550
{
	struct ipv6_pinfo *np;
551 552 553
	const struct ipv6hdr *hdr = (const struct ipv6hdr *)skb->data;
	const struct in6_addr *saddr = &hdr->saddr;
	const struct in6_addr *daddr = &hdr->daddr;
554
	struct udphdr *uh = (struct udphdr *)(skb->data+offset);
555
	bool tunnel = false;
L
Linus Torvalds 已提交
556
	struct sock *sk;
557
	int harderr;
L
Linus Torvalds 已提交
558
	int err;
559
	struct net *net = dev_net(skb->dev);
L
Linus Torvalds 已提交
560

561
	sk = __udp6_lib_lookup(net, daddr, uh->dest, saddr, uh->source,
562
			       inet6_iif(skb), inet6_sdif(skb), udptable, NULL);
563
	if (!sk) {
564
		/* No socket for error: try tunnels before discarding */
565
		sk = ERR_PTR(-ENOENT);
566 567
		if (static_branch_unlikely(&udpv6_encap_needed_key)) {
			sk = __udp6_lib_err_encap(net, hdr, offset, uh,
568 569 570 571
						  udptable, skb,
						  opt, type, code, info);
			if (!sk)
				return 0;
572 573
		}

574
		if (IS_ERR(sk)) {
575 576
			__ICMP6_INC_STATS(net, __in6_dev_get(skb->dev),
					  ICMP6_MIB_INERRORS);
577
			return PTR_ERR(sk);
578
		}
579

580
		tunnel = true;
581
	}
L
Linus Torvalds 已提交
582

583 584 585
	harderr = icmpv6_err_convert(type, code, &err);
	np = inet6_sk(sk);

586 587 588
	if (type == ICMPV6_PKT_TOOBIG) {
		if (!ip6_sk_accept_pmtu(sk))
			goto out;
589
		ip6_sk_update_pmtu(skb, sk, info);
590 591
		if (np->pmtudisc != IPV6_PMTUDISC_DONT)
			harderr = 1;
592
	}
593
	if (type == NDISC_REDIRECT) {
594 595 596 597 598 599
		if (tunnel) {
			ip6_redirect(skb, sock_net(sk), inet6_iif(skb),
				     sk->sk_mark, sk->sk_uid);
		} else {
			ip6_sk_redirect(skb, sk);
		}
600 601
		goto out;
	}
602

603 604 605 606
	/* Tunnels don't have an application socket: don't pass errors back */
	if (tunnel)
		goto out;

607 608 609 610
	if (!np->recverr) {
		if (!harderr || sk->sk_state != TCP_ESTABLISHED)
			goto out;
	} else {
L
Linus Torvalds 已提交
611
		ipv6_icmp_error(sk, skb, err, uh->dest, ntohl(info), (u8 *)(uh+1));
612
	}
613

L
Linus Torvalds 已提交
614 615 616
	sk->sk_err = err;
	sk->sk_error_report(sk);
out:
617
	return 0;
L
Linus Torvalds 已提交
618 619
}

620
static int __udpv6_queue_rcv_skb(struct sock *sk, struct sk_buff *skb)
621 622 623
{
	int rc;

624
	if (!ipv6_addr_any(&sk->sk_v6_daddr)) {
625
		sock_rps_save_rxhash(sk, skb);
626
		sk_mark_napi_id(sk, skb);
E
Eric Dumazet 已提交
627
		sk_incoming_cpu_update(sk);
628 629
	} else {
		sk_mark_napi_id_once(sk, skb);
630
	}
631

632
	rc = __udp_enqueue_schedule_skb(sk, skb);
633 634 635 636 637
	if (rc < 0) {
		int is_udplite = IS_UDPLITE(sk);

		/* Note that an ENOMEM error is charged twice */
		if (rc == -ENOMEM)
638
			UDP6_INC_STATS(sock_net(sk),
639
					 UDP_MIB_RCVBUFERRORS, is_udplite);
640
		UDP6_INC_STATS(sock_net(sk), UDP_MIB_INERRORS, is_udplite);
641 642 643
		kfree_skb(skb);
		return -1;
	}
644

645 646 647
	return 0;
}

648 649 650
static __inline__ int udpv6_err(struct sk_buff *skb,
				struct inet6_skb_parm *opt, u8 type,
				u8 code, int offset, __be32 info)
651
{
652
	return __udp6_lib_err(skb, opt, type, code, offset, info, &udp_table);
653 654
}

655
static int udpv6_queue_rcv_one_skb(struct sock *sk, struct sk_buff *skb)
L
Linus Torvalds 已提交
656
{
657
	struct udp_sock *up = udp_sk(sk);
W
Wang Chen 已提交
658
	int is_udplite = IS_UDPLITE(sk);
659

660 661
	if (!xfrm6_policy_check(sk, XFRM_POLICY_IN, skb))
		goto drop;
L
Linus Torvalds 已提交
662

663
	if (static_branch_unlikely(&udpv6_encap_needed_key) && up->encap_type) {
664 665 666 667 668 669 670 671 672 673 674 675 676 677
		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 */
678
		encap_rcv = READ_ONCE(up->encap_rcv);
679
		if (encap_rcv) {
680 681
			int ret;

682 683 684 685
			/* Verify checksum before giving to encap */
			if (udp_lib_checksum_complete(skb))
				goto csum_error;

686 687
			ret = encap_rcv(sk, skb);
			if (ret <= 0) {
688 689 690
				__UDP_INC_STATS(sock_net(sk),
						UDP_MIB_INDATAGRAMS,
						is_udplite);
691 692 693 694 695 696 697
				return -ret;
			}
		}

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

698 699 700
	/*
	 * UDP-Lite specific tests, ignored on UDP sockets (see net/ipv4/udp.c).
	 */
701
	if ((up->pcflag & UDPLITE_RECV_CC)  &&  UDP_SKB_CB(skb)->partial_cov) {
702 703

		if (up->pcrlen == 0) {          /* full coverage was set  */
704 705
			net_dbg_ratelimited("UDPLITE6: partial coverage %d while full coverage %d requested\n",
					    UDP_SKB_CB(skb)->cscov, skb->len);
706 707 708
			goto drop;
		}
		if (UDP_SKB_CB(skb)->cscov  <  up->pcrlen) {
709 710
			net_dbg_ratelimited("UDPLITE6: coverage %d too small, need min %d\n",
					    UDP_SKB_CB(skb)->cscov, up->pcrlen);
711 712
			goto drop;
		}
L
Linus Torvalds 已提交
713 714
	}

715
	prefetch(&sk->sk_rmem_alloc);
716 717 718 719
	if (rcu_access_pointer(sk->sk_filter) &&
	    udp_lib_checksum_complete(skb))
		goto csum_error;

720
	if (sk_filter_trim_cap(sk, skb, sizeof(struct udphdr)))
721
		goto drop;
722

723
	udp_csum_pull_header(skb);
724

725
	skb_dst_drop(skb);
726

727
	return __udpv6_queue_rcv_skb(sk, skb);
728

729
csum_error:
730
	__UDP6_INC_STATS(sock_net(sk), UDP_MIB_CSUMERRORS, is_udplite);
731
drop:
732
	__UDP6_INC_STATS(sock_net(sk), UDP_MIB_INERRORS, is_udplite);
733
	atomic_inc(&sk->sk_drops);
734 735
	kfree_skb(skb);
	return -1;
L
Linus Torvalds 已提交
736 737
}

738 739 740 741 742 743 744 745 746 747
static int udpv6_queue_rcv_skb(struct sock *sk, struct sk_buff *skb)
{
	struct sk_buff *next, *segs;
	int ret;

	if (likely(!udp_unexpected_gso(sk, skb)))
		return udpv6_queue_rcv_one_skb(sk, skb);

	__skb_push(skb, -skb_mac_offset(skb));
	segs = udp_rcv_segment(sk, skb, false);
748
	skb_list_walk_safe(segs, skb, next) {
749 750 751 752 753 754 755 756 757 758
		__skb_pull(skb, skb_transport_offset(skb));

		ret = udpv6_queue_rcv_one_skb(sk, skb);
		if (ret > 0)
			ip6_protocol_deliver_rcu(dev_net(skb->dev), skb, ret,
						 true);
	}
	return 0;
}

759 760 761
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,
762
				   int dif, int sdif, unsigned short hnum)
L
Linus Torvalds 已提交
763
{
764
	struct inet_sock *inet = inet_sk(sk);
L
Linus Torvalds 已提交
765

766 767 768 769 770 771 772 773
	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)) ||
774
	    !udp_sk_bound_dev_eq(net, sk->sk_bound_dev_if, dif, sdif) ||
775 776
	    (!ipv6_addr_any(&sk->sk_v6_rcv_saddr) &&
		    !ipv6_addr_equal(&sk->sk_v6_rcv_saddr, loc_addr)))
777 778 779 780
		return false;
	if (!inet6_mc_check(sk, loc_addr, rmt_addr))
		return false;
	return true;
L
Linus Torvalds 已提交
781 782
}

783 784 785 786 787
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.
	 */
788 789 790
	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));
791 792
}

L
Linus Torvalds 已提交
793 794 795 796
/*
 * Note: called only from the BH handler context,
 * so we don't need to lock the hashes.
 */
797
static int __udp6_lib_mcast_deliver(struct net *net, struct sk_buff *skb,
798
		const struct in6_addr *saddr, const struct in6_addr *daddr,
799
		struct udp_table *udptable, int proto)
L
Linus Torvalds 已提交
800
{
801
	struct sock *sk, *first = NULL;
802
	const struct udphdr *uh = udp_hdr(skb);
803 804
	unsigned short hnum = ntohs(uh->dest);
	struct udp_hslot *hslot = udp_hashslot(udptable, net, hnum);
805
	unsigned int offset = offsetof(typeof(*sk), sk_node);
806
	unsigned int hash2 = 0, hash2_any = 0, use_hash2 = (hslot->count > 10);
807
	int dif = inet6_iif(skb);
808
	int sdif = inet6_sdif(skb);
809 810
	struct hlist_node *node;
	struct sk_buff *nskb;
811 812

	if (use_hash2) {
813
		hash2_any = ipv6_portaddr_hash(net, &in6addr_any, hnum) &
814
			    udptable->mask;
815
		hash2 = ipv6_portaddr_hash(net, daddr, hnum) & udptable->mask;
816
start_lookup:
817
		hslot = &udptable->hash2[hash2];
818 819
		offset = offsetof(typeof(*sk), __sk_common.skc_portaddr_node);
	}
L
Linus Torvalds 已提交
820

821 822
	sk_for_each_entry_offset_rcu(sk, node, &hslot->head, offset) {
		if (!__udp_v6_is_mcast_sock(net, sk, uh->dest, daddr,
823 824
					    uh->source, saddr, dif, sdif,
					    hnum))
825 826 827 828 829 830 831 832 833 834 835 836 837
			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);
838 839 840 841
			__UDP6_INC_STATS(net, UDP_MIB_RCVBUFERRORS,
					 IS_UDPLITE(sk));
			__UDP6_INC_STATS(net, UDP_MIB_INERRORS,
					 IS_UDPLITE(sk));
842
			continue;
H
Hideo Aoki 已提交
843
		}
844

845 846 847
		if (udpv6_queue_rcv_skb(sk, nskb) > 0)
			consume_skb(nskb);
	}
848

849 850 851 852 853 854
	/* 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;
	}

855 856 857
	if (first) {
		if (udpv6_queue_rcv_skb(first, skb) > 0)
			consume_skb(skb);
858
	} else {
859
		kfree_skb(skb);
860 861
		__UDP6_INC_STATS(net, UDP_MIB_IGNOREDMULTI,
				 proto == IPPROTO_UDPLITE);
862
	}
863
	return 0;
L
Linus Torvalds 已提交
864 865
}

866 867 868 869 870 871 872 873 874
static void udp6_sk_rx_dst_set(struct sock *sk, struct dst_entry *dst)
{
	if (udp_sk_rx_dst_set(sk, dst)) {
		const struct rt6_info *rt = (const struct rt6_info *)dst;

		inet6_sk(sk)->rx_dst_cookie = rt6_get_cookie(rt);
	}
}

875 876 877 878 879 880 881 882 883
/* wrapper for udp_queue_rcv_skb tacking care of csum conversion and
 * return code conversion for ip layer consumption
 */
static int udp6_unicast_rcv_skb(struct sock *sk, struct sk_buff *skb,
				struct udphdr *uh)
{
	int ret;

	if (inet_get_convert_csum(sk) && uh->check && !IS_UDPLITE(sk))
884
		skb_checksum_try_convert(skb, IPPROTO_UDP, ip6_compute_pseudo);
885 886 887

	ret = udpv6_queue_rcv_skb(sk, skb);

888
	/* a return value > 0 means to resubmit the input */
889
	if (ret > 0)
890
		return ret;
891 892 893
	return 0;
}

894
int __udp6_lib_rcv(struct sk_buff *skb, struct udp_table *udptable,
895
		   int proto)
L
Linus Torvalds 已提交
896
{
897
	const struct in6_addr *saddr, *daddr;
898
	struct net *net = dev_net(skb->dev);
899
	struct udphdr *uh;
900
	struct sock *sk;
901
	bool refcounted;
L
Linus Torvalds 已提交
902 903 904
	u32 ulen = 0;

	if (!pskb_may_pull(skb, sizeof(struct udphdr)))
905
		goto discard;
L
Linus Torvalds 已提交
906

907 908
	saddr = &ipv6_hdr(skb)->saddr;
	daddr = &ipv6_hdr(skb)->daddr;
909
	uh = udp_hdr(skb);
L
Linus Torvalds 已提交
910 911

	ulen = ntohs(uh->len);
912 913
	if (ulen > skb->len)
		goto short_packet;
L
Linus Torvalds 已提交
914

915 916
	if (proto == IPPROTO_UDP) {
		/* UDP validates ulen. */
L
Linus Torvalds 已提交
917

918 919 920
		/* Check for jumbo payload */
		if (ulen == 0)
			ulen = skb->len;
L
Linus Torvalds 已提交
921

922 923
		if (ulen < sizeof(*uh))
			goto short_packet;
L
Linus Torvalds 已提交
924

925 926 927
		if (ulen < skb->len) {
			if (pskb_trim_rcsum(skb, ulen))
				goto short_packet;
928 929
			saddr = &ipv6_hdr(skb)->saddr;
			daddr = &ipv6_hdr(skb)->daddr;
930
			uh = udp_hdr(skb);
931 932
		}
	}
L
Linus Torvalds 已提交
933

934
	if (udp6_csum_init(skb, uh, proto))
935
		goto csum_error;
936

937
	/* Check if the socket is already available, e.g. due to early demux */
938
	sk = skb_steal_sock(skb, &refcounted);
939 940 941 942 943
	if (sk) {
		struct dst_entry *dst = skb_dst(skb);
		int ret;

		if (unlikely(sk->sk_rx_dst != dst))
944
			udp6_sk_rx_dst_set(sk, dst);
945

946
		if (!uh->check && !udp_sk(sk)->no_check6_rx) {
947 948
			if (refcounted)
				sock_put(sk);
949 950
			goto report_csum_error;
		}
951

952
		ret = udp6_unicast_rcv_skb(sk, skb, uh);
953 954
		if (refcounted)
			sock_put(sk);
955
		return ret;
956 957
	}

958 959
	/*
	 *	Multicast receive code
L
Linus Torvalds 已提交
960
	 */
961
	if (ipv6_addr_is_multicast(daddr))
962
		return __udp6_lib_mcast_deliver(net, skb,
963
				saddr, daddr, udptable, proto);
L
Linus Torvalds 已提交
964 965

	/* Unicast */
966
	sk = __udp6_lib_lookup_skb(skb, uh->source, uh->dest, udptable);
967
	if (sk) {
968 969 970
		if (!uh->check && !udp_sk(sk)->no_check6_rx)
			goto report_csum_error;
		return udp6_unicast_rcv_skb(sk, skb, uh);
L
Linus Torvalds 已提交
971
	}
972

973 974
	if (!uh->check)
		goto report_csum_error;
975

976
	if (!xfrm6_policy_check(NULL, XFRM_POLICY_IN, skb))
977
		goto discard;
978 979

	if (udp_lib_checksum_complete(skb))
980
		goto csum_error;
981

982
	__UDP6_INC_STATS(net, UDP_MIB_NOPORTS, proto == IPPROTO_UDPLITE);
983 984 985
	icmpv6_send(skb, ICMPV6_DEST_UNREACH, ICMPV6_PORT_UNREACH, 0);

	kfree_skb(skb);
S
Stephen Hemminger 已提交
986
	return 0;
L
Linus Torvalds 已提交
987

988
short_packet:
989 990 991 992 993
	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));
994
	goto discard;
995 996 997

report_csum_error:
	udp6_csum_zero_error(skb);
998
csum_error:
999
	__UDP6_INC_STATS(net, UDP_MIB_CSUMERRORS, proto == IPPROTO_UDPLITE);
L
Linus Torvalds 已提交
1000
discard:
1001
	__UDP6_INC_STATS(net, UDP_MIB_INERRORS, proto == IPPROTO_UDPLITE);
L
Linus Torvalds 已提交
1002
	kfree_skb(skb);
S
Stephen Hemminger 已提交
1003
	return 0;
L
Linus Torvalds 已提交
1004
}
1005

1006

1007 1008 1009
static struct sock *__udp6_lib_demux_lookup(struct net *net,
			__be16 loc_port, const struct in6_addr *loc_addr,
			__be16 rmt_port, const struct in6_addr *rmt_addr,
1010
			int dif, int sdif)
1011
{
1012
	unsigned short hnum = ntohs(loc_port);
1013
	unsigned int hash2 = ipv6_portaddr_hash(net, loc_addr, hnum);
1014 1015 1016
	unsigned int slot2 = hash2 & udp_table.mask;
	struct udp_hslot *hslot2 = &udp_table.hash2[slot2];
	const __portpair ports = INET_COMBINED_PORTS(rmt_port, hnum);
1017 1018
	struct sock *sk;

1019
	udp_portaddr_for_each_entry_rcu(sk, &hslot2->head) {
1020
		if (sk->sk_state == TCP_ESTABLISHED &&
1021
		    INET6_MATCH(sk, net, rmt_addr, loc_addr, ports, dif, sdif))
1022 1023 1024 1025 1026
			return sk;
		/* Only check first socket in chain */
		break;
	}
	return NULL;
1027 1028
}

1029
INDIRECT_CALLABLE_SCOPE void udp_v6_early_demux(struct sk_buff *skb)
1030 1031 1032 1033 1034 1035
{
	struct net *net = dev_net(skb->dev);
	const struct udphdr *uh;
	struct sock *sk;
	struct dst_entry *dst;
	int dif = skb->dev->ifindex;
1036
	int sdif = inet6_sdif(skb);
1037 1038 1039 1040 1041 1042 1043 1044 1045 1046 1047

	if (!pskb_may_pull(skb, skb_transport_offset(skb) +
	    sizeof(struct udphdr)))
		return;

	uh = udp_hdr(skb);

	if (skb->pkt_type == PACKET_HOST)
		sk = __udp6_lib_demux_lookup(net, uh->dest,
					     &ipv6_hdr(skb)->daddr,
					     uh->source, &ipv6_hdr(skb)->saddr,
1048
					     dif, sdif);
1049 1050 1051
	else
		return;

1052
	if (!sk || !refcount_inc_not_zero(&sk->sk_refcnt))
1053 1054 1055 1056 1057 1058 1059 1060 1061
		return;

	skb->sk = sk;
	skb->destructor = sock_efree;
	dst = READ_ONCE(sk->sk_rx_dst);

	if (dst)
		dst = dst_check(dst, inet6_sk(sk)->rx_dst_cookie);
	if (dst) {
1062 1063 1064 1065 1066
		/* set noref for now.
		 * any place which wants to hold dst has to call
		 * dst_hold_safe()
		 */
		skb_dst_set_noref(skb, dst);
1067 1068 1069
	}
}

1070
INDIRECT_CALLABLE_SCOPE int udpv6_rcv(struct sk_buff *skb)
1071
{
1072
	return __udp6_lib_rcv(skb, &udp_table, IPPROTO_UDP);
1073 1074
}

L
Linus Torvalds 已提交
1075 1076 1077 1078 1079 1080 1081
/*
 * 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);

1082 1083 1084
	if (up->pending == AF_INET)
		udp_flush_pending_frames(sk);
	else if (up->pending) {
L
Linus Torvalds 已提交
1085 1086 1087
		up->len = 0;
		up->pending = 0;
		ip6_flush_pending_frames(sk);
1088
	}
L
Linus Torvalds 已提交
1089 1090
}

A
Andrey Ignatov 已提交
1091 1092 1093
static int udpv6_pre_connect(struct sock *sk, struct sockaddr *uaddr,
			     int addr_len)
{
1094 1095
	if (addr_len < offsetofend(struct sockaddr, sa_family))
		return -EINVAL;
A
Andrey Ignatov 已提交
1096 1097 1098 1099 1100 1101 1102 1103 1104 1105 1106 1107 1108 1109 1110 1111
	/* The following checks are replicated from __ip6_datagram_connect()
	 * and intended to prevent BPF program called below from accessing
	 * bytes that are out of the bound specified by user in addr_len.
	 */
	if (uaddr->sa_family == AF_INET) {
		if (__ipv6_only_sock(sk))
			return -EAFNOSUPPORT;
		return udp_pre_connect(sk, uaddr, addr_len);
	}

	if (addr_len < SIN6_LEN_RFC2133)
		return -EINVAL;

	return BPF_CGROUP_RUN_PROG_INET6_CONNECT_LOCK(sk, uaddr);
}

1112
/**
1113 1114 1115 1116
 *	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)
A
Andrew Lunn 已提交
1117 1118 1119
 *	@saddr: source address
 *	@daddr: destination address
 *	@len:	length of packet
1120 1121 1122 1123 1124 1125 1126
 */
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
1127
	struct sk_buff *frags = skb_shinfo(skb)->frag_list;
1128 1129
	__wsum csum = 0;

V
Vlad Yasevich 已提交
1130
	if (!frags) {
1131 1132 1133 1134 1135 1136 1137 1138 1139 1140 1141 1142
		/* 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);
1143
		csum = skb->csum;
1144 1145 1146

		skb->ip_summed = CHECKSUM_NONE;

V
Vlad Yasevich 已提交
1147 1148 1149
		do {
			csum = csum_add(csum, frags->csum);
		} while ((frags = frags->next));
1150 1151 1152 1153 1154 1155 1156 1157

		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
1158 1159 1160 1161
/*
 *	Sending
 */

1162 1163
static int udp_v6_send_skb(struct sk_buff *skb, struct flowi6 *fl6,
			   struct inet_cork *cork)
L
Linus Torvalds 已提交
1164
{
V
Vlad Yasevich 已提交
1165
	struct sock *sk = skb->sk;
L
Linus Torvalds 已提交
1166 1167
	struct udphdr *uh;
	int err = 0;
W
Wang Chen 已提交
1168
	int is_udplite = IS_UDPLITE(sk);
1169
	__wsum csum = 0;
V
Vlad Yasevich 已提交
1170 1171
	int offset = skb_transport_offset(skb);
	int len = skb->len - offset;
J
Josh Hunt 已提交
1172
	int datalen = len - sizeof(*uh);
L
Linus Torvalds 已提交
1173 1174 1175 1176

	/*
	 * Create a UDP header
	 */
1177
	uh = udp_hdr(skb);
1178 1179
	uh->source = fl6->fl6_sport;
	uh->dest = fl6->fl6_dport;
V
Vlad Yasevich 已提交
1180
	uh->len = htons(len);
L
Linus Torvalds 已提交
1181 1182
	uh->check = 0;

1183 1184 1185 1186
	if (cork->gso_size) {
		const int hlen = skb_network_header_len(skb) +
				 sizeof(struct udphdr);

1187 1188
		if (hlen + cork->gso_size > cork->fragsize) {
			kfree_skb(skb);
1189
			return -EINVAL;
1190 1191 1192
		}
		if (skb->len > cork->gso_size * UDP_MAX_SEGMENTS) {
			kfree_skb(skb);
1193
			return -EINVAL;
1194 1195 1196
		}
		if (udp_sk(sk)->no_check6_tx) {
			kfree_skb(skb);
1197
			return -EINVAL;
1198
		}
1199
		if (skb->ip_summed != CHECKSUM_PARTIAL || is_udplite ||
1200 1201
		    dst_xfrm(skb_dst(skb))) {
			kfree_skb(skb);
1202
			return -EIO;
1203
		}
1204

J
Josh Hunt 已提交
1205 1206 1207 1208 1209 1210
		if (datalen > cork->gso_size) {
			skb_shinfo(skb)->gso_size = cork->gso_size;
			skb_shinfo(skb)->gso_type = SKB_GSO_UDP_L4;
			skb_shinfo(skb)->gso_segs = DIV_ROUND_UP(datalen,
								 cork->gso_size);
		}
1211
		goto csum_partial;
1212 1213
	}

W
Wang Chen 已提交
1214
	if (is_udplite)
V
Vlad Yasevich 已提交
1215 1216
		csum = udplite_csum(skb);
	else if (udp_sk(sk)->no_check6_tx) {   /* UDP csum disabled */
1217 1218 1219
		skb->ip_summed = CHECKSUM_NONE;
		goto send;
	} else if (skb->ip_summed == CHECKSUM_PARTIAL) { /* UDP hardware csum */
1220
csum_partial:
V
Vlad Yasevich 已提交
1221
		udp6_hwcsum_outgoing(sk, skb, &fl6->saddr, &fl6->daddr, len);
1222 1223
		goto send;
	} else
V
Vlad Yasevich 已提交
1224
		csum = udp_csum(skb);
L
Linus Torvalds 已提交
1225

1226
	/* add protocol-dependent pseudo-header */
1227
	uh->check = csum_ipv6_magic(&fl6->saddr, &fl6->daddr,
V
Vlad Yasevich 已提交
1228
				    len, fl6->flowi6_proto, csum);
L
Linus Torvalds 已提交
1229
	if (uh->check == 0)
1230
		uh->check = CSUM_MANGLED_0;
L
Linus Torvalds 已提交
1231

1232
send:
V
Vlad Yasevich 已提交
1233
	err = ip6_send_skb(skb);
E
Eric Dumazet 已提交
1234 1235
	if (err) {
		if (err == -ENOBUFS && !inet6_sk(sk)->recverr) {
1236 1237
			UDP6_INC_STATS(sock_net(sk),
				       UDP_MIB_SNDBUFERRORS, is_udplite);
E
Eric Dumazet 已提交
1238 1239
			err = 0;
		}
1240 1241 1242 1243
	} else {
		UDP6_INC_STATS(sock_net(sk),
			       UDP_MIB_OUTDATAGRAMS, is_udplite);
	}
V
Vlad Yasevich 已提交
1244 1245 1246 1247 1248 1249 1250 1251 1252 1253 1254 1255 1256 1257 1258 1259 1260 1261 1262 1263 1264 1265
	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;

1266
	err = udp_v6_send_skb(skb, &fl6, &inet_sk(sk)->cork.base);
V
Vlad Yasevich 已提交
1267

L
Linus Torvalds 已提交
1268 1269 1270 1271 1272 1273
out:
	up->len = 0;
	up->pending = 0;
	return err;
}

1274
int udpv6_sendmsg(struct sock *sk, struct msghdr *msg, size_t len)
L
Linus Torvalds 已提交
1275 1276 1277 1278 1279
{
	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);
1280
	DECLARE_SOCKADDR(struct sockaddr_in6 *, sin6, msg->msg_name);
1281
	struct in6_addr *daddr, *final_p, final;
L
Linus Torvalds 已提交
1282
	struct ipv6_txoptions *opt = NULL;
1283
	struct ipv6_txoptions *opt_to_free = NULL;
L
Linus Torvalds 已提交
1284
	struct ip6_flowlabel *flowlabel = NULL;
1285
	struct flowi6 fl6;
L
Linus Torvalds 已提交
1286
	struct dst_entry *dst;
W
Wei Wang 已提交
1287
	struct ipcm6_cookie ipc6;
L
Linus Torvalds 已提交
1288
	int addr_len = msg->msg_namelen;
1289
	bool connected = false;
L
Linus Torvalds 已提交
1290 1291 1292
	int ulen = len;
	int corkreq = up->corkflag || msg->msg_flags&MSG_MORE;
	int err;
W
Wang Chen 已提交
1293
	int is_udplite = IS_UDPLITE(sk);
1294
	int (*getfrag)(void *, char *, int, int, int, struct sk_buff *);
L
Linus Torvalds 已提交
1295

1296
	ipcm6_init(&ipc6);
1297
	ipc6.gso_size = up->gso_size;
1298
	ipc6.sockc.tsflags = sk->sk_tsflags;
W
Willem de Bruijn 已提交
1299
	ipc6.sockc.mark = sk->sk_mark;
W
Wei Wang 已提交
1300

L
Linus Torvalds 已提交
1301 1302 1303 1304 1305 1306 1307 1308 1309 1310
	/* 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;
1311 1312 1313 1314
			if (ipv6_addr_any(daddr) &&
			    ipv6_addr_v4mapped(&np->saddr))
				ipv6_addr_set_v4mapped(htonl(INADDR_LOOPBACK),
						       daddr);
L
Linus Torvalds 已提交
1315 1316 1317 1318 1319 1320 1321 1322 1323 1324 1325 1326 1327 1328
			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;
1329
		daddr = &sk->sk_v6_daddr;
1330
	} else
L
Linus Torvalds 已提交
1331 1332 1333
		daddr = NULL;

	if (daddr) {
B
Brian Haley 已提交
1334
		if (ipv6_addr_v4mapped(daddr)) {
L
Linus Torvalds 已提交
1335 1336
			struct sockaddr_in sin;
			sin.sin_family = AF_INET;
E
Eric Dumazet 已提交
1337
			sin.sin_port = sin6 ? sin6->sin6_port : inet->inet_dport;
L
Linus Torvalds 已提交
1338 1339 1340 1341 1342 1343
			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;
1344
			return udp_sendmsg(sk, msg, len);
L
Linus Torvalds 已提交
1345 1346 1347 1348
		}
	}

	if (up->pending == AF_INET)
1349
		return udp_sendmsg(sk, msg, len);
L
Linus Torvalds 已提交
1350 1351

	/* Rough check on arithmetic overflow,
1352
	   better check is made in ip6_append_data().
L
Linus Torvalds 已提交
1353 1354 1355
	   */
	if (len > INT_MAX - sizeof(struct udphdr))
		return -EMSGSIZE;
1356

1357
	getfrag  =  is_udplite ?  udplite_getfrag : ip_generic_getfrag;
L
Linus Torvalds 已提交
1358 1359 1360 1361 1362 1363 1364 1365 1366 1367 1368 1369 1370 1371 1372 1373 1374 1375
	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);

1376
	memset(&fl6, 0, sizeof(fl6));
L
Linus Torvalds 已提交
1377 1378 1379 1380 1381

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

1382
		fl6.fl6_dport = sin6->sin6_port;
L
Linus Torvalds 已提交
1383 1384 1385
		daddr = &sin6->sin6_addr;

		if (np->sndflow) {
1386 1387 1388
			fl6.flowlabel = sin6->sin6_flowinfo&IPV6_FLOWINFO_MASK;
			if (fl6.flowlabel&IPV6_FLOWLABEL_MASK) {
				flowlabel = fl6_sock_lookup(sk, fl6.flowlabel);
1389
				if (IS_ERR(flowlabel))
L
Linus Torvalds 已提交
1390 1391 1392 1393 1394 1395 1396 1397 1398
					return -EINVAL;
			}
		}

		/*
		 * Otherwise it will be difficult to maintain
		 * sk->sk_dst_cache.
		 */
		if (sk->sk_state == TCP_ESTABLISHED &&
1399 1400
		    ipv6_addr_equal(daddr, &sk->sk_v6_daddr))
			daddr = &sk->sk_v6_daddr;
L
Linus Torvalds 已提交
1401 1402 1403

		if (addr_len >= sizeof(struct sockaddr_in6) &&
		    sin6->sin6_scope_id &&
1404
		    __ipv6_addr_needs_scope_id(__ipv6_addr_type(daddr)))
1405
			fl6.flowi6_oif = sin6->sin6_scope_id;
L
Linus Torvalds 已提交
1406 1407 1408 1409
	} else {
		if (sk->sk_state != TCP_ESTABLISHED)
			return -EDESTADDRREQ;

1410
		fl6.fl6_dport = inet->inet_dport;
1411
		daddr = &sk->sk_v6_daddr;
1412
		fl6.flowlabel = np->flow_label;
1413
		connected = true;
L
Linus Torvalds 已提交
1414 1415
	}

1416 1417
	if (!fl6.flowi6_oif)
		fl6.flowi6_oif = sk->sk_bound_dev_if;
L
Linus Torvalds 已提交
1418

1419 1420
	if (!fl6.flowi6_oif)
		fl6.flowi6_oif = np->sticky_pktinfo.ipi6_ifindex;
1421

W
Willem de Bruijn 已提交
1422
	fl6.flowi6_mark = ipc6.sockc.mark;
1423
	fl6.flowi6_uid = sk->sk_uid;
B
Brian Haley 已提交
1424

L
Linus Torvalds 已提交
1425 1426 1427 1428
	if (msg->msg_controllen) {
		opt = &opt_space;
		memset(opt, 0, sizeof(struct ipv6_txoptions));
		opt->tot_len = sizeof(*opt);
W
Wei Wang 已提交
1429
		ipc6.opt = opt;
L
Linus Torvalds 已提交
1430

W
Willem de Bruijn 已提交
1431 1432 1433
		err = udp_cmsg_send(sk, msg, &ipc6.gso_size);
		if (err > 0)
			err = ip6_datagram_send_ctl(sock_net(sk), sk, msg, &fl6,
1434
						    &ipc6);
L
Linus Torvalds 已提交
1435 1436 1437 1438
		if (err < 0) {
			fl6_sock_release(flowlabel);
			return err;
		}
1439 1440
		if ((fl6.flowlabel&IPV6_FLOWLABEL_MASK) && !flowlabel) {
			flowlabel = fl6_sock_lookup(sk, fl6.flowlabel);
1441
			if (IS_ERR(flowlabel))
L
Linus Torvalds 已提交
1442 1443 1444 1445
				return -EINVAL;
		}
		if (!(opt->opt_nflen|opt->opt_flen))
			opt = NULL;
1446
		connected = false;
L
Linus Torvalds 已提交
1447
	}
1448 1449 1450 1451
	if (!opt) {
		opt = txopt_get(np);
		opt_to_free = opt;
	}
1452 1453 1454
	if (flowlabel)
		opt = fl6_merge_options(&opt_space, flowlabel, opt);
	opt = ipv6_fixup_options(&opt_space, opt);
W
Wei Wang 已提交
1455
	ipc6.opt = opt;
L
Linus Torvalds 已提交
1456

1457
	fl6.flowi6_proto = sk->sk_protocol;
1458
	fl6.daddr = *daddr;
1459
	if (ipv6_addr_any(&fl6.saddr) && !ipv6_addr_any(&np->saddr))
A
Alexey Dobriyan 已提交
1460
		fl6.saddr = np->saddr;
1461
	fl6.fl6_sport = inet->inet_sport;
1462

A
Andrey Ignatov 已提交
1463 1464 1465 1466 1467 1468 1469 1470 1471 1472 1473 1474 1475 1476 1477 1478 1479 1480 1481 1482 1483 1484 1485
	if (cgroup_bpf_enabled && !connected) {
		err = BPF_CGROUP_RUN_PROG_UDP6_SENDMSG_LOCK(sk,
					   (struct sockaddr *)sin6, &fl6.saddr);
		if (err)
			goto out_no_dst;
		if (sin6) {
			if (ipv6_addr_v4mapped(&sin6->sin6_addr)) {
				/* BPF program rewrote IPv6-only by IPv4-mapped
				 * IPv6. It's currently unsupported.
				 */
				err = -ENOTSUPP;
				goto out_no_dst;
			}
			if (sin6->sin6_port == 0) {
				/* BPF program set invalid port. Reject it. */
				err = -EINVAL;
				goto out_no_dst;
			}
			fl6.fl6_dport = sin6->sin6_port;
			fl6.daddr = sin6->sin6_addr;
		}
	}

1486 1487 1488
	if (ipv6_addr_any(&fl6.daddr))
		fl6.daddr.s6_addr[15] = 0x1; /* :: means loopback (BSD'ism) */

1489
	final_p = fl6_update_dst(&fl6, opt, &final);
1490
	if (final_p)
1491
		connected = false;
L
Linus Torvalds 已提交
1492

1493 1494
	if (!fl6.flowi6_oif && ipv6_addr_is_multicast(&fl6.daddr)) {
		fl6.flowi6_oif = np->mcast_oif;
1495
		connected = false;
1496 1497
	} else if (!fl6.flowi6_oif)
		fl6.flowi6_oif = np->ucast_oif;
L
Linus Torvalds 已提交
1498

1499
	security_sk_classify_flow(sk, flowi6_to_flowi(&fl6));
V
Venkat Yekkirala 已提交
1500

1501 1502 1503 1504 1505
	if (ipc6.tclass < 0)
		ipc6.tclass = np->tclass;

	fl6.flowlabel = ip6_make_flowinfo(ipc6.tclass, fl6.flowlabel);

1506
	dst = ip6_sk_dst_lookup_flow(sk, &fl6, final_p, connected);
1507 1508 1509
	if (IS_ERR(dst)) {
		err = PTR_ERR(dst);
		dst = NULL;
L
Linus Torvalds 已提交
1510
		goto out;
1511
	}
L
Linus Torvalds 已提交
1512

W
Wei Wang 已提交
1513 1514
	if (ipc6.hlimit < 0)
		ipc6.hlimit = ip6_sk_dst_hoplimit(np, &fl6, dst);
L
Linus Torvalds 已提交
1515 1516 1517 1518 1519

	if (msg->msg_flags&MSG_CONFIRM)
		goto do_confirm;
back_from_confirm:

1520 1521
	/* Lockless fast path for the non-corking case */
	if (!corkreq) {
W
Willem de Bruijn 已提交
1522
		struct inet_cork_full cork;
1523 1524 1525
		struct sk_buff *skb;

		skb = ip6_make_skb(sk, getfrag, msg, ulen,
W
Wei Wang 已提交
1526
				   sizeof(struct udphdr), &ipc6,
1527
				   &fl6, (struct rt6_info *)dst,
1528
				   msg->msg_flags, &cork);
1529 1530
		err = PTR_ERR(skb);
		if (!IS_ERR_OR_NULL(skb))
1531
			err = udp_v6_send_skb(skb, &fl6, &cork.base);
1532
		goto out;
1533 1534
	}

L
Linus Torvalds 已提交
1535 1536 1537 1538 1539 1540
	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);

1541
		net_dbg_ratelimited("udp cork app bug 2\n");
L
Linus Torvalds 已提交
1542 1543 1544 1545 1546 1547 1548
		err = -EINVAL;
		goto out;
	}

	up->pending = AF_INET6;

do_append_data:
W
Wei Wang 已提交
1549 1550
	if (ipc6.dontfrag < 0)
		ipc6.dontfrag = np->dontfrag;
L
Linus Torvalds 已提交
1551
	up->len += ulen;
W
Wei Wang 已提交
1552 1553
	err = ip6_append_data(sk, getfrag, msg, ulen, sizeof(struct udphdr),
			      &ipc6, &fl6, (struct rt6_info *)dst,
1554
			      corkreq ? msg->msg_flags|MSG_MORE : msg->msg_flags);
L
Linus Torvalds 已提交
1555 1556 1557
	if (err)
		udp_v6_flush_pending_frames(sk);
	else if (!corkreq)
1558
		err = udp_v6_push_pending_frames(sk);
H
Herbert Xu 已提交
1559 1560
	else if (unlikely(skb_queue_empty(&sk->sk_write_queue)))
		up->pending = 0;
L
Linus Torvalds 已提交
1561

1562 1563 1564 1565
	if (err > 0)
		err = np->recverr ? net_xmit_errno(err) : 0;
	release_sock(sk);

L
Linus Torvalds 已提交
1566
out:
1567
	dst_release(dst);
A
Andrey Ignatov 已提交
1568
out_no_dst:
L
Linus Torvalds 已提交
1569
	fl6_sock_release(flowlabel);
1570
	txopt_put(opt_to_free);
1571
	if (!err)
L
Linus Torvalds 已提交
1572
		return len;
1573 1574 1575 1576 1577 1578 1579 1580
	/*
	 * 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)) {
1581 1582
		UDP6_INC_STATS(sock_net(sk),
			       UDP_MIB_SNDBUFERRORS, is_udplite);
1583
	}
L
Linus Torvalds 已提交
1584 1585 1586
	return err;

do_confirm:
1587 1588
	if (msg->msg_flags & MSG_PROBE)
		dst_confirm_neigh(dst, &fl6.daddr);
L
Linus Torvalds 已提交
1589 1590 1591 1592 1593 1594
	if (!(msg->msg_flags&MSG_PROBE) || len)
		goto back_from_confirm;
	err = 0;
	goto out;
}

1595
void udpv6_destroy_sock(struct sock *sk)
L
Linus Torvalds 已提交
1596
{
T
Tom Parkin 已提交
1597
	struct udp_sock *up = udp_sk(sk);
L
Linus Torvalds 已提交
1598 1599 1600 1601
	lock_sock(sk);
	udp_v6_flush_pending_frames(sk);
	release_sock(sk);

1602 1603 1604 1605 1606 1607 1608 1609
	if (static_branch_unlikely(&udpv6_encap_needed_key)) {
		if (up->encap_type) {
			void (*encap_destroy)(struct sock *sk);
			encap_destroy = READ_ONCE(up->encap_destroy);
			if (encap_destroy)
				encap_destroy(sk);
		}
		if (up->encap_enabled)
P
Paolo Abeni 已提交
1610
			static_branch_dec(&udpv6_encap_needed_key);
T
Tom Parkin 已提交
1611 1612
	}

L
Linus Torvalds 已提交
1613 1614 1615 1616 1617 1618
	inet6_destroy_sock(sk);
}

/*
 *	Socket option code for UDP
 */
1619 1620
int udpv6_setsockopt(struct sock *sk, int level, int optname, sockptr_t optval,
		     unsigned int optlen)
1621
{
1622
	if (level == SOL_UDP  ||  level == SOL_UDPLITE)
1623
		return udp_lib_setsockopt(sk, level, optname,
1624
					  optval, optlen,
1625
					  udp_v6_push_pending_frames);
1626
	return ipv6_setsockopt(sk, level, optname, optval, optlen);
1627 1628
}

1629 1630
int udpv6_getsockopt(struct sock *sk, int level, int optname,
		     char __user *optval, int __user *optlen)
1631
{
1632
	if (level == SOL_UDP  ||  level == SOL_UDPLITE)
1633
		return udp_lib_getsockopt(sk, level, optname, optval, optlen);
1634
	return ipv6_getsockopt(sk, level, optname, optval, optlen);
1635 1636
}

1637 1638 1639
/* thinking of making this const? Don't.
 * early_demux can change based on sysctl.
 */
1640
static struct inet6_protocol udpv6_protocol = {
1641
	.early_demux	=	udp_v6_early_demux,
1642
	.early_demux_handler =  udp_v6_early_demux,
L
Linus Torvalds 已提交
1643 1644 1645 1646 1647 1648 1649
	.handler	=	udpv6_rcv,
	.err_handler	=	udpv6_err,
	.flags		=	INET6_PROTO_NOPOLICY|INET6_PROTO_FINAL,
};

/* ------------------------------------------------------------------------ */
#ifdef CONFIG_PROC_FS
1650
int udp6_seq_show(struct seq_file *seq, void *v)
L
Linus Torvalds 已提交
1651
{
1652 1653 1654 1655 1656 1657 1658
	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);
1659 1660
		__ip6_dgram_sock_seq_show(seq, v, srcp, destp,
					  udp_rqueue_get(v), bucket);
1661
	}
L
Linus Torvalds 已提交
1662 1663 1664
	return 0;
}

1665
const struct seq_operations udp6_seq_ops = {
1666 1667 1668 1669 1670
	.start		= udp_seq_start,
	.next		= udp_seq_next,
	.stop		= udp_seq_stop,
	.show		= udp6_seq_show,
};
1671
EXPORT_SYMBOL(udp6_seq_ops);
1672

L
Linus Torvalds 已提交
1673 1674
static struct udp_seq_afinfo udp6_seq_afinfo = {
	.family		= AF_INET6,
1675
	.udp_table	= &udp_table,
L
Linus Torvalds 已提交
1676 1677
};

1678
int __net_init udp6_proc_init(struct net *net)
L
Linus Torvalds 已提交
1679
{
1680 1681
	if (!proc_create_net_data("udp6", 0444, net->proc_net, &udp6_seq_ops,
			sizeof(struct udp_iter_state), &udp6_seq_afinfo))
1682 1683
		return -ENOMEM;
	return 0;
L
Linus Torvalds 已提交
1684 1685
}

I
Ian Morris 已提交
1686 1687
void udp6_proc_exit(struct net *net)
{
1688
	remove_proc_entry("udp6", net->proc_net);
L
Linus Torvalds 已提交
1689 1690 1691 1692 1693 1694
}
#endif /* CONFIG_PROC_FS */

/* ------------------------------------------------------------------------ */

struct proto udpv6_prot = {
1695 1696 1697
	.name			= "UDPv6",
	.owner			= THIS_MODULE,
	.close			= udp_lib_close,
A
Andrey Ignatov 已提交
1698
	.pre_connect		= udpv6_pre_connect,
1699 1700 1701 1702 1703 1704 1705 1706 1707 1708 1709 1710 1711 1712 1713 1714 1715 1716 1717 1718 1719
	.connect		= ip6_datagram_connect,
	.disconnect		= udp_disconnect,
	.ioctl			= udp_ioctl,
	.init			= udp_init_sock,
	.destroy		= udpv6_destroy_sock,
	.setsockopt		= udpv6_setsockopt,
	.getsockopt		= udpv6_getsockopt,
	.sendmsg		= udpv6_sendmsg,
	.recvmsg		= udpv6_recvmsg,
	.release_cb		= ip6_datagram_release_cb,
	.hash			= udp_lib_hash,
	.unhash			= udp_lib_unhash,
	.rehash			= udp_v6_rehash,
	.get_port		= udp_v6_get_port,
	.memory_allocated	= &udp_memory_allocated,
	.sysctl_mem		= sysctl_udp_mem,
	.sysctl_wmem_offset     = offsetof(struct net, ipv4.sysctl_udp_wmem_min),
	.sysctl_rmem_offset     = offsetof(struct net, ipv4.sysctl_udp_rmem_min),
	.obj_size		= sizeof(struct udp6_sock),
	.h.udp_table		= &udp_table,
	.diag_destroy		= udp_abort,
L
Linus Torvalds 已提交
1720 1721 1722 1723 1724 1725 1726 1727 1728 1729
};

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

1730
int __init udpv6_init(void)
L
Linus Torvalds 已提交
1731
{
1732 1733
	int ret;

1734 1735
	ret = inet6_add_protocol(&udpv6_protocol, IPPROTO_UDP);
	if (ret)
1736
		goto out;
1737

1738 1739 1740 1741 1742 1743 1744 1745 1746 1747 1748
	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;
}

1749
void udpv6_exit(void)
1750 1751 1752
{
	inet6_unregister_protosw(&udpv6_protosw);
	inet6_del_protocol(&udpv6_protocol, IPPROTO_UDP);
L
Linus Torvalds 已提交
1753
}