udp.c 45.0 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
 *
 *	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.
 */

20
#include <linux/bpf-cgroup.h>
L
Linus Torvalds 已提交
21 22 23 24 25 26 27 28 29 30 31
#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>
32
#include <linux/module.h>
33
#include <linux/skbuff.h>
34
#include <linux/slab.h>
35
#include <linux/uaccess.h>
36
#include <linux/indirect_call_wrapper.h>
L
Linus Torvalds 已提交
37

38
#include <net/addrconf.h>
L
Linus Torvalds 已提交
39 40 41 42 43
#include <net/ndisc.h>
#include <net/protocol.h>
#include <net/transp_v6.h>
#include <net/ip6_route.h>
#include <net/raw.h>
44
#include <net/tcp_states.h>
L
Linus Torvalds 已提交
45
#include <net/ip6_checksum.h>
46
#include <net/ip6_tunnel.h>
L
Linus Torvalds 已提交
47
#include <net/xfrm.h>
48
#include <net/inet_hashtables.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
int udp_v6_get_port(struct sock *sk, unsigned short snum)
L
Linus Torvalds 已提交
82
{
E
Eric Dumazet 已提交
83
	unsigned int hash2_nulladdr =
84
		ipv6_portaddr_hash(sock_net(sk), &in6addr_any, snum);
85
	unsigned int hash2_partial =
86
		ipv6_portaddr_hash(sock_net(sk), &sk->sk_v6_rcv_saddr, 0);
E
Eric Dumazet 已提交
87

88
	/* precompute partial secondary hash */
E
Eric Dumazet 已提交
89
	udp_sk(sk)->udp_portaddr_hash = hash2_partial;
90
	return udp_lib_get_port(sk, snum, hash2_nulladdr);
L
Linus Torvalds 已提交
91 92
}

93
void udp_v6_rehash(struct sock *sk)
94
{
95
	u16 new_hash = ipv6_portaddr_hash(sock_net(sk),
96
					  &sk->sk_v6_rcv_saddr,
97 98 99 100 101
					  inet_sk(sk)->inet_num);

	udp_lib_rehash(sk, new_hash);
}

102 103 104
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,
105
			 int dif, int sdif)
106
{
107 108
	int score;
	struct inet_sock *inet;
109
	bool dev_match;
110 111 112 113 114 115

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

116 117 118
	if (!ipv6_addr_equal(&sk->sk_v6_rcv_saddr, daddr))
		return -1;

119 120 121 122 123 124 125 126 127 128 129 130 131 132 133
	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++;
	}

134 135 136
	dev_match = udp_sk_bound_dev_eq(net, sk->sk_bound_dev_if, dif, sdif);
	if (!dev_match)
		return -1;
137 138
	if (sk->sk_bound_dev_if)
		score++;
139

140
	if (READ_ONCE(sk->sk_incoming_cpu) == raw_smp_processor_id())
141 142
		score++;

143 144 145
	return score;
}

146 147 148 149 150 151
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)
152 153 154 155 156 157 158 159 160 161 162 163
{
	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;
}

164
/* called with rcu_read_lock() */
165 166
static struct sock *udp6_lib_lookup2(struct net *net,
		const struct in6_addr *saddr, __be16 sport,
167
		const struct in6_addr *daddr, unsigned int hnum,
168 169
		int dif, int sdif, struct udp_hslot *hslot2,
		struct sk_buff *skb)
170 171
{
	struct sock *sk, *result;
172
	int score, badness;
173 174 175

	result = NULL;
	badness = -1;
176
	udp_portaddr_for_each_entry_rcu(sk, &hslot2->head) {
177
		score = compute_score(sk, net, saddr, sport,
178
				      daddr, hnum, dif, sdif);
179
		if (score > badness) {
180 181
			result = lookup_reuseport(net, sk, skb,
						  saddr, sport, daddr, hnum);
182 183
			/* Fall back to scoring if group has connections */
			if (result && !reuseport_has_conns(sk, false))
184 185
				return result;

186
			result = result ? : sk;
187
			badness = score;
188 189 190 191 192
		}
	}
	return result;
}

193 194 195 196 197 198
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,
199
					       u16 hnum, const int dif)
200 201 202 203 204 205 206
{
	struct sock *sk, *reuse_sk;
	bool no_reuseport;

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

207 208
	no_reuseport = bpf_sk_lookup_run_v6(net, IPPROTO_UDP, saddr, sport,
					    daddr, hnum, dif, &sk);
209 210 211 212
	if (no_reuseport || IS_ERR_OR_NULL(sk))
		return sk;

	reuse_sk = lookup_reuseport(net, sk, skb, saddr, sport, daddr, hnum);
213
	if (reuse_sk)
214 215 216 217
		sk = reuse_sk;
	return sk;
}

218
/* rcu_read_lock() must be held */
219
struct sock *__udp6_lib_lookup(struct net *net,
220 221 222 223
			       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
224 225
{
	unsigned short hnum = ntohs(dport);
226 227
	unsigned int hash2, slot2;
	struct udp_hslot *hslot2;
228
	struct sock *result, *sk;
229

230 231 232 233
	hash2 = ipv6_portaddr_hash(net, daddr, hnum);
	slot2 = hash2 & udptable->mask;
	hslot2 = &udptable->hash2[slot2];

234
	/* Lookup connected or non-wildcard sockets */
235
	result = udp6_lib_lookup2(net, saddr, sport,
236
				  daddr, hnum, dif, sdif,
237
				  hslot2, skb);
238 239 240 241 242 243
	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,
244
					 saddr, sport, daddr, hnum, dif);
245 246 247 248 249
		if (sk) {
			result = sk;
			goto done;
		}
	}
250

251 252 253
	/* Got non-wildcard socket or error on first lookup */
	if (result)
		goto done;
254

255 256 257 258 259 260 261 262 263
	/* 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:
264
	if (IS_ERR(result))
265
		return NULL;
L
Linus Torvalds 已提交
266 267
	return result;
}
268
EXPORT_SYMBOL_GPL(__udp6_lib_lookup);
L
Linus Torvalds 已提交
269

270 271
static struct sock *__udp6_lib_lookup_skb(struct sk_buff *skb,
					  __be16 sport, __be16 dport,
272
					  struct udp_table *udptable)
273
{
274
	const struct ipv6hdr *iph = ipv6_hdr(skb);
275

276
	return __udp6_lib_lookup(dev_net(skb->dev), &iph->saddr, sport,
E
Eric Dumazet 已提交
277
				 &iph->daddr, dport, inet6_iif(skb),
278
				 inet6_sdif(skb), udptable, skb);
279 280
}

281
struct sock *udp6_lib_lookup_skb(const struct sk_buff *skb,
282 283 284 285
				 __be16 sport, __be16 dport)
{
	const struct ipv6hdr *iph = ipv6_hdr(skb);

286
	return __udp6_lib_lookup(dev_net(skb->dev), &iph->saddr, sport,
287
				 &iph->daddr, dport, inet6_iif(skb),
288
				 inet6_sdif(skb), &udp_table, NULL);
289 290
}

291 292 293
/* Must be called under rcu_read_lock().
 * Does increment socket refcount.
 */
294
#if IS_ENABLED(CONFIG_NF_TPROXY_IPV6) || IS_ENABLED(CONFIG_NF_SOCKET_IPV6)
295 296 297
struct sock *udp6_lib_lookup(struct net *net, const struct in6_addr *saddr, __be16 sport,
			     const struct in6_addr *daddr, __be16 dport, int dif)
{
298 299 300
	struct sock *sk;

	sk =  __udp6_lib_lookup(net, saddr, sport, daddr, dport,
301
				dif, 0, &udp_table, NULL);
302
	if (sk && !refcount_inc_not_zero(&sk->sk_refcnt))
303 304
		sk = NULL;
	return sk;
305 306
}
EXPORT_SYMBOL_GPL(udp6_lib_lookup);
307
#endif
308

309 310 311 312 313 314 315 316 317
/* 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
318
/*
319 320
 *	This should be easy, if there is something there we
 *	return it, otherwise we block.
L
Linus Torvalds 已提交
321 322
 */

323
int udpv6_recvmsg(struct sock *sk, struct msghdr *msg, size_t len,
L
Linus Torvalds 已提交
324 325 326 327
		  int noblock, int flags, int *addr_len)
{
	struct ipv6_pinfo *np = inet6_sk(sk);
	struct inet_sock *inet = inet_sk(sk);
328
	struct sk_buff *skb;
329
	unsigned int ulen, copied;
330
	int off, err, peeking = flags & MSG_PEEK;
331
	int is_udplite = IS_UDPLITE(sk);
332
	struct udp_mib __percpu *mib;
333
	bool checksum_valid = false;
334
	int is_udp4;
L
Linus Torvalds 已提交
335 336

	if (flags & MSG_ERRQUEUE)
337
		return ipv6_recv_error(sk, msg, len, addr_len);
L
Linus Torvalds 已提交
338

339
	if (np->rxpmtu && np->rxopt.bits.rxpmtu)
340
		return ipv6_recv_rxpmtu(sk, msg, len, addr_len);
341

L
Linus Torvalds 已提交
342
try_again:
343
	off = sk_peek_offset(sk, flags);
344
	skb = __skb_recv_udp(sk, flags, noblock, &off, &err);
L
Linus Torvalds 已提交
345
	if (!skb)
346
		return err;
L
Linus Torvalds 已提交
347

348
	ulen = udp6_skb_len(skb);
349
	copied = len;
350 351
	if (copied > ulen - off)
		copied = ulen - off;
352
	else if (copied < ulen)
353
		msg->msg_flags |= MSG_TRUNC;
L
Linus Torvalds 已提交
354

355
	is_udp4 = (skb->protocol == htons(ETH_P_IP));
356
	mib = __UDPX_MIB(sk, is_udp4);
357

358
	/*
359 360 361
	 * 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.
362 363
	 */

364 365
	if (copied < ulen || peeking ||
	    (is_udplite && UDP_SKB_CB(skb)->partial_cov)) {
366 367
		checksum_valid = udp_skb_csum_unnecessary(skb) ||
				!__udp_lib_checksum_complete(skb);
368
		if (!checksum_valid)
L
Linus Torvalds 已提交
369
			goto csum_copy_err;
370 371
	}

372 373 374 375 376 377
	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 {
378
		err = skb_copy_and_csum_datagram_msg(skb, off, msg);
L
Linus Torvalds 已提交
379 380 381
		if (err == -EINVAL)
			goto csum_copy_err;
	}
382
	if (unlikely(err)) {
383
		if (!peeking) {
384
			atomic_inc(&sk->sk_drops);
385
			SNMP_INC_STATS(mib, UDP_MIB_INERRORS);
386
		}
387
		kfree_skb(skb);
388
		return err;
389
	}
390
	if (!peeking)
391
		SNMP_INC_STATS(mib, UDP_MIB_INDATAGRAMS);
392

393
	sock_recv_ts_and_drops(msg, sk, skb);
L
Linus Torvalds 已提交
394 395 396

	/* Copy the address. */
	if (msg->msg_name) {
397
		DECLARE_SOCKADDR(struct sockaddr_in6 *, sin6, msg->msg_name);
L
Linus Torvalds 已提交
398
		sin6->sin6_family = AF_INET6;
399
		sin6->sin6_port = udp_hdr(skb)->source;
L
Linus Torvalds 已提交
400 401
		sin6->sin6_flowinfo = 0;

402
		if (is_udp4) {
403 404
			ipv6_addr_set_v4mapped(ip_hdr(skb)->saddr,
					       &sin6->sin6_addr);
405 406
			sin6->sin6_scope_id = 0;
		} else {
407
			sin6->sin6_addr = ipv6_hdr(skb)->saddr;
408 409
			sin6->sin6_scope_id =
				ipv6_iface_scope_id(&sin6->sin6_addr,
410
						    inet6_iif(skb));
L
Linus Torvalds 已提交
411
		}
412
		*addr_len = sizeof(*sin6);
413

414 415
		BPF_CGROUP_RUN_PROG_UDP6_RECVMSG_LOCK(sk,
						      (struct sockaddr *)sin6);
L
Linus Torvalds 已提交
416
	}
417

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,
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)
{
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
					 struct sock *sk,
508 509 510
					 struct sk_buff *skb,
					 struct inet6_skb_parm *opt,
					 u8 type, u8 code, __be32 info)
511
{
512
	int (*lookup)(struct sock *sk, struct sk_buff *skb);
513
	int network_offset, transport_offset;
514
	struct udp_sock *up;
515 516 517 518 519 520 521 522 523 524

	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);

525 526 527 528 529 530 531 532 533 534
	if (sk) {
		up = udp_sk(sk);

		lookup = READ_ONCE(up->encap_err_lookup);
		if (lookup && lookup(sk, skb))
			sk = NULL;

		goto out;
	}

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) {
539
		up = udp_sk(sk);
540 541 542 543 544 545

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

546
out:
547 548 549 550
	if (!sk) {
		sk = ERR_PTR(__udp6_lib_err_encap_no_sk(skb, opt, type, code,
							offset, info));
	}
551 552 553

	skb_set_transport_header(skb, transport_offset);
	skb_set_network_header(skb, network_offset);
554

555 556 557
	return sk;
}

558 559 560
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
561 562
{
	struct ipv6_pinfo *np;
563 564 565
	const struct ipv6hdr *hdr = (const struct ipv6hdr *)skb->data;
	const struct in6_addr *saddr = &hdr->saddr;
	const struct in6_addr *daddr = &hdr->daddr;
566
	struct udphdr *uh = (struct udphdr *)(skb->data+offset);
567
	bool tunnel = false;
L
Linus Torvalds 已提交
568
	struct sock *sk;
569
	int harderr;
L
Linus Torvalds 已提交
570
	int err;
571
	struct net *net = dev_net(skb->dev);
L
Linus Torvalds 已提交
572

573
	sk = __udp6_lib_lookup(net, daddr, uh->dest, saddr, uh->source,
574
			       inet6_iif(skb), inet6_sdif(skb), udptable, NULL);
575

576
	if (!sk || udp_sk(sk)->encap_type) {
577 578 579
		/* No socket for error: try tunnels before discarding */
		if (static_branch_unlikely(&udpv6_encap_needed_key)) {
			sk = __udp6_lib_err_encap(net, hdr, offset, uh,
580
						  udptable, sk, skb,
581 582 583
						  opt, type, code, info);
			if (!sk)
				return 0;
584 585
		} else
			sk = ERR_PTR(-ENOENT);
586

587
		if (IS_ERR(sk)) {
588 589
			__ICMP6_INC_STATS(net, __in6_dev_get(skb->dev),
					  ICMP6_MIB_INERRORS);
590
			return PTR_ERR(sk);
591
		}
592

593
		tunnel = true;
594
	}
L
Linus Torvalds 已提交
595

596 597 598
	harderr = icmpv6_err_convert(type, code, &err);
	np = inet6_sk(sk);

599 600 601
	if (type == ICMPV6_PKT_TOOBIG) {
		if (!ip6_sk_accept_pmtu(sk))
			goto out;
602
		ip6_sk_update_pmtu(skb, sk, info);
603 604
		if (np->pmtudisc != IPV6_PMTUDISC_DONT)
			harderr = 1;
605
	}
606
	if (type == NDISC_REDIRECT) {
607 608 609 610 611 612
		if (tunnel) {
			ip6_redirect(skb, sock_net(sk), inet6_iif(skb),
				     sk->sk_mark, sk->sk_uid);
		} else {
			ip6_sk_redirect(skb, sk);
		}
613 614
		goto out;
	}
615

616 617 618 619
	/* Tunnels don't have an application socket: don't pass errors back */
	if (tunnel)
		goto out;

620 621 622 623
	if (!np->recverr) {
		if (!harderr || sk->sk_state != TCP_ESTABLISHED)
			goto out;
	} else {
L
Linus Torvalds 已提交
624
		ipv6_icmp_error(sk, skb, err, uh->dest, ntohl(info), (u8 *)(uh+1));
625
	}
626

L
Linus Torvalds 已提交
627
	sk->sk_err = err;
628
	sk_error_report(sk);
L
Linus Torvalds 已提交
629
out:
630
	return 0;
L
Linus Torvalds 已提交
631 632
}

633
static int __udpv6_queue_rcv_skb(struct sock *sk, struct sk_buff *skb)
634 635 636
{
	int rc;

637
	if (!ipv6_addr_any(&sk->sk_v6_daddr)) {
638
		sock_rps_save_rxhash(sk, skb);
639
		sk_mark_napi_id(sk, skb);
640
		sk_incoming_cpu_update(sk);
641 642
	} else {
		sk_mark_napi_id_once(sk, skb);
643
	}
644

645
	rc = __udp_enqueue_schedule_skb(sk, skb);
646 647 648 649 650
	if (rc < 0) {
		int is_udplite = IS_UDPLITE(sk);

		/* Note that an ENOMEM error is charged twice */
		if (rc == -ENOMEM)
651
			UDP6_INC_STATS(sock_net(sk),
652
					 UDP_MIB_RCVBUFERRORS, is_udplite);
653 654 655
		else
			UDP6_INC_STATS(sock_net(sk),
				       UDP_MIB_MEMERRORS, is_udplite);
656
		UDP6_INC_STATS(sock_net(sk), UDP_MIB_INERRORS, is_udplite);
657 658 659
		kfree_skb(skb);
		return -1;
	}
660

661 662 663
	return 0;
}

664 665 666
static __inline__ int udpv6_err(struct sk_buff *skb,
				struct inet6_skb_parm *opt, u8 type,
				u8 code, int offset, __be32 info)
667
{
668
	return __udp6_lib_err(skb, opt, type, code, offset, info, &udp_table);
669 670
}

671
static int udpv6_queue_rcv_one_skb(struct sock *sk, struct sk_buff *skb)
L
Linus Torvalds 已提交
672
{
673
	struct udp_sock *up = udp_sk(sk);
674
	int is_udplite = IS_UDPLITE(sk);
675

676 677
	if (!xfrm6_policy_check(sk, XFRM_POLICY_IN, skb))
		goto drop;
L
Linus Torvalds 已提交
678

679
	if (static_branch_unlikely(&udpv6_encap_needed_key) && up->encap_type) {
680 681 682 683 684 685 686 687 688 689 690 691 692 693
		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 */
694
		encap_rcv = READ_ONCE(up->encap_rcv);
695
		if (encap_rcv) {
696 697
			int ret;

698 699 700 701
			/* Verify checksum before giving to encap */
			if (udp_lib_checksum_complete(skb))
				goto csum_error;

702 703
			ret = encap_rcv(sk, skb);
			if (ret <= 0) {
704 705 706
				__UDP6_INC_STATS(sock_net(sk),
						 UDP_MIB_INDATAGRAMS,
						 is_udplite);
707 708 709 710 711 712 713
				return -ret;
			}
		}

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

714 715 716
	/*
	 * UDP-Lite specific tests, ignored on UDP sockets (see net/ipv4/udp.c).
	 */
717
	if ((up->pcflag & UDPLITE_RECV_CC)  &&  UDP_SKB_CB(skb)->partial_cov) {
718 719

		if (up->pcrlen == 0) {          /* full coverage was set  */
720 721
			net_dbg_ratelimited("UDPLITE6: partial coverage %d while full coverage %d requested\n",
					    UDP_SKB_CB(skb)->cscov, skb->len);
722 723 724
			goto drop;
		}
		if (UDP_SKB_CB(skb)->cscov  <  up->pcrlen) {
725 726
			net_dbg_ratelimited("UDPLITE6: coverage %d too small, need min %d\n",
					    UDP_SKB_CB(skb)->cscov, up->pcrlen);
727 728
			goto drop;
		}
L
Linus Torvalds 已提交
729 730
	}

731
	prefetch(&sk->sk_rmem_alloc);
732 733 734 735
	if (rcu_access_pointer(sk->sk_filter) &&
	    udp_lib_checksum_complete(skb))
		goto csum_error;

736
	if (sk_filter_trim_cap(sk, skb, sizeof(struct udphdr)))
737
		goto drop;
738

739
	udp_csum_pull_header(skb);
740

741
	skb_dst_drop(skb);
742

743
	return __udpv6_queue_rcv_skb(sk, skb);
744

745
csum_error:
746
	__UDP6_INC_STATS(sock_net(sk), UDP_MIB_CSUMERRORS, is_udplite);
747
drop:
748
	__UDP6_INC_STATS(sock_net(sk), UDP_MIB_INERRORS, is_udplite);
749
	atomic_inc(&sk->sk_drops);
750 751
	kfree_skb(skb);
	return -1;
L
Linus Torvalds 已提交
752 753
}

754 755 756 757 758 759 760 761 762 763
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);
764
	skb_list_walk_safe(segs, skb, next) {
765 766
		__skb_pull(skb, skb_transport_offset(skb));

767
		udp_post_segment_fix_csum(skb);
768 769 770 771 772 773 774 775
		ret = udpv6_queue_rcv_one_skb(sk, skb);
		if (ret > 0)
			ip6_protocol_deliver_rcu(dev_net(skb->dev), skb, ret,
						 true);
	}
	return 0;
}

776 777 778
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,
779
				   int dif, int sdif, unsigned short hnum)
L
Linus Torvalds 已提交
780
{
781
	struct inet_sock *inet = inet_sk(sk);
L
Linus Torvalds 已提交
782

783 784 785 786 787 788 789 790
	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)) ||
791
	    !udp_sk_bound_dev_eq(net, sk->sk_bound_dev_if, dif, sdif) ||
792 793
	    (!ipv6_addr_any(&sk->sk_v6_rcv_saddr) &&
		    !ipv6_addr_equal(&sk->sk_v6_rcv_saddr, loc_addr)))
794 795 796 797
		return false;
	if (!inet6_mc_check(sk, loc_addr, rmt_addr))
		return false;
	return true;
L
Linus Torvalds 已提交
798 799
}

800 801 802 803 804
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.
	 */
805 806 807
	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));
808 809
}

L
Linus Torvalds 已提交
810 811 812 813
/*
 * Note: called only from the BH handler context,
 * so we don't need to lock the hashes.
 */
814
static int __udp6_lib_mcast_deliver(struct net *net, struct sk_buff *skb,
815
		const struct in6_addr *saddr, const struct in6_addr *daddr,
816
		struct udp_table *udptable, int proto)
L
Linus Torvalds 已提交
817
{
818
	struct sock *sk, *first = NULL;
819
	const struct udphdr *uh = udp_hdr(skb);
820 821
	unsigned short hnum = ntohs(uh->dest);
	struct udp_hslot *hslot = udp_hashslot(udptable, net, hnum);
822
	unsigned int offset = offsetof(typeof(*sk), sk_node);
823
	unsigned int hash2 = 0, hash2_any = 0, use_hash2 = (hslot->count > 10);
824
	int dif = inet6_iif(skb);
825
	int sdif = inet6_sdif(skb);
826 827
	struct hlist_node *node;
	struct sk_buff *nskb;
828 829

	if (use_hash2) {
830
		hash2_any = ipv6_portaddr_hash(net, &in6addr_any, hnum) &
831
			    udptable->mask;
832
		hash2 = ipv6_portaddr_hash(net, daddr, hnum) & udptable->mask;
833
start_lookup:
834
		hslot = &udptable->hash2[hash2];
835 836
		offset = offsetof(typeof(*sk), __sk_common.skc_portaddr_node);
	}
L
Linus Torvalds 已提交
837

838 839
	sk_for_each_entry_offset_rcu(sk, node, &hslot->head, offset) {
		if (!__udp_v6_is_mcast_sock(net, sk, uh->dest, daddr,
840 841
					    uh->source, saddr, dif, sdif,
					    hnum))
842 843 844 845 846 847 848 849 850 851 852 853 854
			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);
855 856 857 858
			__UDP6_INC_STATS(net, UDP_MIB_RCVBUFERRORS,
					 IS_UDPLITE(sk));
			__UDP6_INC_STATS(net, UDP_MIB_INERRORS,
					 IS_UDPLITE(sk));
859
			continue;
860
		}
861

862 863 864
		if (udpv6_queue_rcv_skb(sk, nskb) > 0)
			consume_skb(nskb);
	}
865

866 867 868 869 870 871
	/* 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;
	}

872 873 874
	if (first) {
		if (udpv6_queue_rcv_skb(first, skb) > 0)
			consume_skb(skb);
875
	} else {
876
		kfree_skb(skb);
877 878
		__UDP6_INC_STATS(net, UDP_MIB_IGNOREDMULTI,
				 proto == IPPROTO_UDPLITE);
879
	}
880
	return 0;
L
Linus Torvalds 已提交
881 882
}

883 884 885 886 887
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;

888
		sk->sk_rx_dst_cookie = rt6_get_cookie(rt);
889 890 891
	}
}

892 893 894 895 896 897 898 899 900
/* 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))
901
		skb_checksum_try_convert(skb, IPPROTO_UDP, ip6_compute_pseudo);
902 903 904

	ret = udpv6_queue_rcv_skb(sk, skb);

905
	/* a return value > 0 means to resubmit the input */
906
	if (ret > 0)
907
		return ret;
908 909 910
	return 0;
}

911
int __udp6_lib_rcv(struct sk_buff *skb, struct udp_table *udptable,
912
		   int proto)
L
Linus Torvalds 已提交
913
{
914
	const struct in6_addr *saddr, *daddr;
915
	struct net *net = dev_net(skb->dev);
916
	struct udphdr *uh;
917
	struct sock *sk;
918
	bool refcounted;
L
Linus Torvalds 已提交
919 920 921
	u32 ulen = 0;

	if (!pskb_may_pull(skb, sizeof(struct udphdr)))
922
		goto discard;
L
Linus Torvalds 已提交
923

924 925
	saddr = &ipv6_hdr(skb)->saddr;
	daddr = &ipv6_hdr(skb)->daddr;
926
	uh = udp_hdr(skb);
L
Linus Torvalds 已提交
927 928

	ulen = ntohs(uh->len);
929 930
	if (ulen > skb->len)
		goto short_packet;
L
Linus Torvalds 已提交
931

932 933
	if (proto == IPPROTO_UDP) {
		/* UDP validates ulen. */
L
Linus Torvalds 已提交
934

935 936 937
		/* Check for jumbo payload */
		if (ulen == 0)
			ulen = skb->len;
L
Linus Torvalds 已提交
938

939 940
		if (ulen < sizeof(*uh))
			goto short_packet;
L
Linus Torvalds 已提交
941

942 943 944
		if (ulen < skb->len) {
			if (pskb_trim_rcsum(skb, ulen))
				goto short_packet;
945 946
			saddr = &ipv6_hdr(skb)->saddr;
			daddr = &ipv6_hdr(skb)->daddr;
947
			uh = udp_hdr(skb);
948 949
		}
	}
L
Linus Torvalds 已提交
950

951
	if (udp6_csum_init(skb, uh, proto))
952
		goto csum_error;
953

954
	/* Check if the socket is already available, e.g. due to early demux */
955
	sk = skb_steal_sock(skb, &refcounted);
956 957 958 959 960
	if (sk) {
		struct dst_entry *dst = skb_dst(skb);
		int ret;

		if (unlikely(sk->sk_rx_dst != dst))
961
			udp6_sk_rx_dst_set(sk, dst);
962

963
		if (!uh->check && !udp_sk(sk)->no_check6_rx) {
964 965
			if (refcounted)
				sock_put(sk);
966 967
			goto report_csum_error;
		}
968

969
		ret = udp6_unicast_rcv_skb(sk, skb, uh);
970 971
		if (refcounted)
			sock_put(sk);
972
		return ret;
973 974
	}

975 976
	/*
	 *	Multicast receive code
L
Linus Torvalds 已提交
977
	 */
978
	if (ipv6_addr_is_multicast(daddr))
979
		return __udp6_lib_mcast_deliver(net, skb,
980
				saddr, daddr, udptable, proto);
L
Linus Torvalds 已提交
981 982

	/* Unicast */
983
	sk = __udp6_lib_lookup_skb(skb, uh->source, uh->dest, udptable);
984
	if (sk) {
985 986 987
		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
988
	}
989

990 991
	if (!uh->check)
		goto report_csum_error;
992

993
	if (!xfrm6_policy_check(NULL, XFRM_POLICY_IN, skb))
994
		goto discard;
995 996

	if (udp_lib_checksum_complete(skb))
997
		goto csum_error;
998

999
	__UDP6_INC_STATS(net, UDP_MIB_NOPORTS, proto == IPPROTO_UDPLITE);
1000 1001 1002
	icmpv6_send(skb, ICMPV6_DEST_UNREACH, ICMPV6_PORT_UNREACH, 0);

	kfree_skb(skb);
1003
	return 0;
L
Linus Torvalds 已提交
1004

1005
short_packet:
1006 1007 1008 1009 1010
	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));
1011
	goto discard;
1012 1013 1014

report_csum_error:
	udp6_csum_zero_error(skb);
1015
csum_error:
1016
	__UDP6_INC_STATS(net, UDP_MIB_CSUMERRORS, proto == IPPROTO_UDPLITE);
L
Linus Torvalds 已提交
1017
discard:
1018
	__UDP6_INC_STATS(net, UDP_MIB_INERRORS, proto == IPPROTO_UDPLITE);
L
Linus Torvalds 已提交
1019
	kfree_skb(skb);
1020
	return 0;
L
Linus Torvalds 已提交
1021
}
1022

1023

1024 1025 1026
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,
1027
			int dif, int sdif)
1028
{
1029
	unsigned short hnum = ntohs(loc_port);
1030
	unsigned int hash2 = ipv6_portaddr_hash(net, loc_addr, hnum);
1031 1032 1033
	unsigned int slot2 = hash2 & udp_table.mask;
	struct udp_hslot *hslot2 = &udp_table.hash2[slot2];
	const __portpair ports = INET_COMBINED_PORTS(rmt_port, hnum);
1034 1035
	struct sock *sk;

1036
	udp_portaddr_for_each_entry_rcu(sk, &hslot2->head) {
1037
		if (sk->sk_state == TCP_ESTABLISHED &&
1038
		    INET6_MATCH(sk, net, rmt_addr, loc_addr, ports, dif, sdif))
1039 1040 1041 1042 1043
			return sk;
		/* Only check first socket in chain */
		break;
	}
	return NULL;
1044 1045
}

1046
INDIRECT_CALLABLE_SCOPE void udp_v6_early_demux(struct sk_buff *skb)
1047 1048 1049 1050 1051 1052
{
	struct net *net = dev_net(skb->dev);
	const struct udphdr *uh;
	struct sock *sk;
	struct dst_entry *dst;
	int dif = skb->dev->ifindex;
1053
	int sdif = inet6_sdif(skb);
1054 1055 1056 1057 1058 1059 1060 1061 1062 1063 1064

	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,
1065
					     dif, sdif);
1066 1067 1068
	else
		return;

1069
	if (!sk || !refcount_inc_not_zero(&sk->sk_refcnt))
1070 1071 1072 1073 1074 1075 1076
		return;

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

	if (dst)
1077
		dst = dst_check(dst, sk->sk_rx_dst_cookie);
1078
	if (dst) {
1079 1080 1081 1082 1083
		/* set noref for now.
		 * any place which wants to hold dst has to call
		 * dst_hold_safe()
		 */
		skb_dst_set_noref(skb, dst);
1084 1085 1086
	}
}

1087
INDIRECT_CALLABLE_SCOPE int udpv6_rcv(struct sk_buff *skb)
1088
{
1089
	return __udp6_lib_rcv(skb, &udp_table, IPPROTO_UDP);
1090 1091
}

L
Linus Torvalds 已提交
1092 1093 1094 1095 1096 1097 1098
/*
 * 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);

1099 1100 1101
	if (up->pending == AF_INET)
		udp_flush_pending_frames(sk);
	else if (up->pending) {
L
Linus Torvalds 已提交
1102 1103 1104
		up->len = 0;
		up->pending = 0;
		ip6_flush_pending_frames(sk);
1105
	}
L
Linus Torvalds 已提交
1106 1107
}

1108 1109 1110
static int udpv6_pre_connect(struct sock *sk, struct sockaddr *uaddr,
			     int addr_len)
{
1111 1112
	if (addr_len < offsetofend(struct sockaddr, sa_family))
		return -EINVAL;
1113 1114 1115 1116 1117 1118 1119 1120 1121 1122 1123 1124 1125 1126 1127 1128
	/* 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);
}

1129
/**
1130 1131 1132 1133
 *	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)
1134 1135 1136
 *	@saddr: source address
 *	@daddr: destination address
 *	@len:	length of packet
1137 1138 1139 1140 1141 1142 1143
 */
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);
1144
	struct sk_buff *frags = skb_shinfo(skb)->frag_list;
1145 1146
	__wsum csum = 0;

1147
	if (!frags) {
1148 1149 1150 1151 1152 1153 1154 1155 1156 1157 1158 1159
		/* 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);
1160
		csum = skb->csum;
1161 1162 1163

		skb->ip_summed = CHECKSUM_NONE;

1164 1165 1166
		do {
			csum = csum_add(csum, frags->csum);
		} while ((frags = frags->next));
1167 1168 1169 1170 1171 1172 1173 1174

		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
1175 1176 1177 1178
/*
 *	Sending
 */

1179 1180
static int udp_v6_send_skb(struct sk_buff *skb, struct flowi6 *fl6,
			   struct inet_cork *cork)
L
Linus Torvalds 已提交
1181
{
1182
	struct sock *sk = skb->sk;
L
Linus Torvalds 已提交
1183 1184
	struct udphdr *uh;
	int err = 0;
1185
	int is_udplite = IS_UDPLITE(sk);
1186
	__wsum csum = 0;
1187 1188
	int offset = skb_transport_offset(skb);
	int len = skb->len - offset;
1189
	int datalen = len - sizeof(*uh);
L
Linus Torvalds 已提交
1190 1191 1192 1193

	/*
	 * Create a UDP header
	 */
1194
	uh = udp_hdr(skb);
1195 1196
	uh->source = fl6->fl6_sport;
	uh->dest = fl6->fl6_dport;
1197
	uh->len = htons(len);
L
Linus Torvalds 已提交
1198 1199
	uh->check = 0;

1200 1201 1202 1203
	if (cork->gso_size) {
		const int hlen = skb_network_header_len(skb) +
				 sizeof(struct udphdr);

1204 1205
		if (hlen + cork->gso_size > cork->fragsize) {
			kfree_skb(skb);
1206
			return -EINVAL;
1207 1208 1209
		}
		if (skb->len > cork->gso_size * UDP_MAX_SEGMENTS) {
			kfree_skb(skb);
1210
			return -EINVAL;
1211 1212 1213
		}
		if (udp_sk(sk)->no_check6_tx) {
			kfree_skb(skb);
1214
			return -EINVAL;
1215
		}
1216
		if (skb->ip_summed != CHECKSUM_PARTIAL || is_udplite ||
1217 1218
		    dst_xfrm(skb_dst(skb))) {
			kfree_skb(skb);
1219
			return -EIO;
1220
		}
1221

1222 1223 1224 1225 1226 1227
		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);
		}
1228
		goto csum_partial;
1229 1230
	}

1231
	if (is_udplite)
1232 1233
		csum = udplite_csum(skb);
	else if (udp_sk(sk)->no_check6_tx) {   /* UDP csum disabled */
1234 1235 1236
		skb->ip_summed = CHECKSUM_NONE;
		goto send;
	} else if (skb->ip_summed == CHECKSUM_PARTIAL) { /* UDP hardware csum */
1237
csum_partial:
1238
		udp6_hwcsum_outgoing(sk, skb, &fl6->saddr, &fl6->daddr, len);
1239 1240
		goto send;
	} else
1241
		csum = udp_csum(skb);
L
Linus Torvalds 已提交
1242

1243
	/* add protocol-dependent pseudo-header */
1244
	uh->check = csum_ipv6_magic(&fl6->saddr, &fl6->daddr,
1245
				    len, fl6->flowi6_proto, csum);
L
Linus Torvalds 已提交
1246
	if (uh->check == 0)
1247
		uh->check = CSUM_MANGLED_0;
L
Linus Torvalds 已提交
1248

1249
send:
1250
	err = ip6_send_skb(skb);
1251 1252
	if (err) {
		if (err == -ENOBUFS && !inet6_sk(sk)->recverr) {
1253 1254
			UDP6_INC_STATS(sock_net(sk),
				       UDP_MIB_SNDBUFERRORS, is_udplite);
1255 1256
			err = 0;
		}
1257 1258 1259 1260
	} else {
		UDP6_INC_STATS(sock_net(sk),
			       UDP_MIB_OUTDATAGRAMS, is_udplite);
	}
1261 1262 1263 1264 1265 1266 1267 1268 1269 1270 1271 1272 1273 1274 1275 1276 1277 1278 1279 1280 1281 1282
	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;

1283
	err = udp_v6_send_skb(skb, &fl6, &inet_sk(sk)->cork.base);
1284

L
Linus Torvalds 已提交
1285 1286 1287 1288 1289 1290
out:
	up->len = 0;
	up->pending = 0;
	return err;
}

1291
int udpv6_sendmsg(struct sock *sk, struct msghdr *msg, size_t len)
L
Linus Torvalds 已提交
1292 1293 1294 1295 1296
{
	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);
1297
	DECLARE_SOCKADDR(struct sockaddr_in6 *, sin6, msg->msg_name);
1298
	struct in6_addr *daddr, *final_p, final;
L
Linus Torvalds 已提交
1299
	struct ipv6_txoptions *opt = NULL;
1300
	struct ipv6_txoptions *opt_to_free = NULL;
L
Linus Torvalds 已提交
1301
	struct ip6_flowlabel *flowlabel = NULL;
1302
	struct flowi6 fl6;
L
Linus Torvalds 已提交
1303
	struct dst_entry *dst;
1304
	struct ipcm6_cookie ipc6;
L
Linus Torvalds 已提交
1305
	int addr_len = msg->msg_namelen;
1306
	bool connected = false;
L
Linus Torvalds 已提交
1307
	int ulen = len;
1308
	int corkreq = READ_ONCE(up->corkflag) || msg->msg_flags&MSG_MORE;
L
Linus Torvalds 已提交
1309
	int err;
1310
	int is_udplite = IS_UDPLITE(sk);
1311
	int (*getfrag)(void *, char *, int, int, int, struct sk_buff *);
L
Linus Torvalds 已提交
1312

1313
	ipcm6_init(&ipc6);
1314
	ipc6.gso_size = READ_ONCE(up->gso_size);
1315
	ipc6.sockc.tsflags = sk->sk_tsflags;
1316
	ipc6.sockc.mark = sk->sk_mark;
1317

L
Linus Torvalds 已提交
1318 1319 1320 1321 1322 1323 1324 1325 1326 1327
	/* 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;
1328 1329 1330 1331
			if (ipv6_addr_any(daddr) &&
			    ipv6_addr_v4mapped(&np->saddr))
				ipv6_addr_set_v4mapped(htonl(INADDR_LOOPBACK),
						       daddr);
L
Linus Torvalds 已提交
1332 1333 1334 1335 1336 1337 1338 1339 1340 1341 1342 1343 1344 1345
			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;
1346
		daddr = &sk->sk_v6_daddr;
1347
	} else
L
Linus Torvalds 已提交
1348 1349 1350
		daddr = NULL;

	if (daddr) {
1351
		if (ipv6_addr_v4mapped(daddr)) {
L
Linus Torvalds 已提交
1352 1353
			struct sockaddr_in sin;
			sin.sin_family = AF_INET;
1354
			sin.sin_port = sin6 ? sin6->sin6_port : inet->inet_dport;
L
Linus Torvalds 已提交
1355 1356 1357 1358 1359 1360
			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;
1361
			return udp_sendmsg(sk, msg, len);
L
Linus Torvalds 已提交
1362 1363 1364 1365
		}
	}

	if (up->pending == AF_INET)
1366
		return udp_sendmsg(sk, msg, len);
L
Linus Torvalds 已提交
1367 1368

	/* Rough check on arithmetic overflow,
1369
	   better check is made in ip6_append_data().
L
Linus Torvalds 已提交
1370 1371 1372
	   */
	if (len > INT_MAX - sizeof(struct udphdr))
		return -EMSGSIZE;
1373

1374
	getfrag  =  is_udplite ?  udplite_getfrag : ip_generic_getfrag;
L
Linus Torvalds 已提交
1375 1376 1377 1378 1379 1380 1381 1382 1383 1384 1385 1386 1387 1388 1389 1390 1391 1392
	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);

1393
	memset(&fl6, 0, sizeof(fl6));
L
Linus Torvalds 已提交
1394 1395 1396 1397 1398

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

1399
		fl6.fl6_dport = sin6->sin6_port;
L
Linus Torvalds 已提交
1400 1401 1402
		daddr = &sin6->sin6_addr;

		if (np->sndflow) {
1403 1404 1405
			fl6.flowlabel = sin6->sin6_flowinfo&IPV6_FLOWINFO_MASK;
			if (fl6.flowlabel&IPV6_FLOWLABEL_MASK) {
				flowlabel = fl6_sock_lookup(sk, fl6.flowlabel);
1406
				if (IS_ERR(flowlabel))
L
Linus Torvalds 已提交
1407 1408 1409 1410 1411 1412 1413 1414 1415
					return -EINVAL;
			}
		}

		/*
		 * Otherwise it will be difficult to maintain
		 * sk->sk_dst_cache.
		 */
		if (sk->sk_state == TCP_ESTABLISHED &&
1416 1417
		    ipv6_addr_equal(daddr, &sk->sk_v6_daddr))
			daddr = &sk->sk_v6_daddr;
L
Linus Torvalds 已提交
1418 1419 1420

		if (addr_len >= sizeof(struct sockaddr_in6) &&
		    sin6->sin6_scope_id &&
1421
		    __ipv6_addr_needs_scope_id(__ipv6_addr_type(daddr)))
1422
			fl6.flowi6_oif = sin6->sin6_scope_id;
L
Linus Torvalds 已提交
1423 1424 1425 1426
	} else {
		if (sk->sk_state != TCP_ESTABLISHED)
			return -EDESTADDRREQ;

1427
		fl6.fl6_dport = inet->inet_dport;
1428
		daddr = &sk->sk_v6_daddr;
1429
		fl6.flowlabel = np->flow_label;
1430
		connected = true;
L
Linus Torvalds 已提交
1431 1432
	}

1433 1434
	if (!fl6.flowi6_oif)
		fl6.flowi6_oif = sk->sk_bound_dev_if;
L
Linus Torvalds 已提交
1435

1436 1437
	if (!fl6.flowi6_oif)
		fl6.flowi6_oif = np->sticky_pktinfo.ipi6_ifindex;
1438

1439
	fl6.flowi6_uid = sk->sk_uid;
1440

L
Linus Torvalds 已提交
1441 1442 1443 1444
	if (msg->msg_controllen) {
		opt = &opt_space;
		memset(opt, 0, sizeof(struct ipv6_txoptions));
		opt->tot_len = sizeof(*opt);
1445
		ipc6.opt = opt;
L
Linus Torvalds 已提交
1446

1447 1448 1449
		err = udp_cmsg_send(sk, msg, &ipc6.gso_size);
		if (err > 0)
			err = ip6_datagram_send_ctl(sock_net(sk), sk, msg, &fl6,
1450
						    &ipc6);
L
Linus Torvalds 已提交
1451 1452 1453 1454
		if (err < 0) {
			fl6_sock_release(flowlabel);
			return err;
		}
1455 1456
		if ((fl6.flowlabel&IPV6_FLOWLABEL_MASK) && !flowlabel) {
			flowlabel = fl6_sock_lookup(sk, fl6.flowlabel);
1457
			if (IS_ERR(flowlabel))
L
Linus Torvalds 已提交
1458 1459 1460 1461
				return -EINVAL;
		}
		if (!(opt->opt_nflen|opt->opt_flen))
			opt = NULL;
1462
		connected = false;
L
Linus Torvalds 已提交
1463
	}
1464 1465 1466 1467
	if (!opt) {
		opt = txopt_get(np);
		opt_to_free = opt;
	}
1468 1469 1470
	if (flowlabel)
		opt = fl6_merge_options(&opt_space, flowlabel, opt);
	opt = ipv6_fixup_options(&opt_space, opt);
1471
	ipc6.opt = opt;
L
Linus Torvalds 已提交
1472

1473
	fl6.flowi6_proto = sk->sk_protocol;
1474
	fl6.flowi6_mark = ipc6.sockc.mark;
1475
	fl6.daddr = *daddr;
1476
	if (ipv6_addr_any(&fl6.saddr) && !ipv6_addr_any(&np->saddr))
1477
		fl6.saddr = np->saddr;
1478
	fl6.fl6_sport = inet->inet_sport;
1479

1480
	if (cgroup_bpf_enabled(CGROUP_UDP6_SENDMSG) && !connected) {
1481 1482 1483 1484 1485 1486 1487 1488 1489 1490 1491 1492 1493 1494 1495 1496 1497 1498 1499 1500 1501 1502
		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;
		}
	}

1503 1504 1505
	if (ipv6_addr_any(&fl6.daddr))
		fl6.daddr.s6_addr[15] = 0x1; /* :: means loopback (BSD'ism) */

1506
	final_p = fl6_update_dst(&fl6, opt, &final);
1507
	if (final_p)
1508
		connected = false;
L
Linus Torvalds 已提交
1509

1510 1511
	if (!fl6.flowi6_oif && ipv6_addr_is_multicast(&fl6.daddr)) {
		fl6.flowi6_oif = np->mcast_oif;
1512
		connected = false;
1513 1514
	} else if (!fl6.flowi6_oif)
		fl6.flowi6_oif = np->ucast_oif;
L
Linus Torvalds 已提交
1515

1516
	security_sk_classify_flow(sk, flowi6_to_flowi_common(&fl6));
1517

1518 1519 1520 1521 1522
	if (ipc6.tclass < 0)
		ipc6.tclass = np->tclass;

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

1523
	dst = ip6_sk_dst_lookup_flow(sk, &fl6, final_p, connected);
1524 1525 1526
	if (IS_ERR(dst)) {
		err = PTR_ERR(dst);
		dst = NULL;
L
Linus Torvalds 已提交
1527
		goto out;
1528
	}
L
Linus Torvalds 已提交
1529

1530 1531
	if (ipc6.hlimit < 0)
		ipc6.hlimit = ip6_sk_dst_hoplimit(np, &fl6, dst);
L
Linus Torvalds 已提交
1532 1533 1534 1535 1536

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

1537 1538
	/* Lockless fast path for the non-corking case */
	if (!corkreq) {
1539
		struct inet_cork_full cork;
1540 1541 1542
		struct sk_buff *skb;

		skb = ip6_make_skb(sk, getfrag, msg, ulen,
1543
				   sizeof(struct udphdr), &ipc6,
1544
				   &fl6, (struct rt6_info *)dst,
1545
				   msg->msg_flags, &cork);
1546 1547
		err = PTR_ERR(skb);
		if (!IS_ERR_OR_NULL(skb))
1548
			err = udp_v6_send_skb(skb, &fl6, &cork.base);
1549
		goto out;
1550 1551
	}

L
Linus Torvalds 已提交
1552 1553 1554 1555 1556 1557
	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);

1558
		net_dbg_ratelimited("udp cork app bug 2\n");
L
Linus Torvalds 已提交
1559 1560 1561 1562 1563 1564 1565
		err = -EINVAL;
		goto out;
	}

	up->pending = AF_INET6;

do_append_data:
1566 1567
	if (ipc6.dontfrag < 0)
		ipc6.dontfrag = np->dontfrag;
L
Linus Torvalds 已提交
1568
	up->len += ulen;
1569 1570
	err = ip6_append_data(sk, getfrag, msg, ulen, sizeof(struct udphdr),
			      &ipc6, &fl6, (struct rt6_info *)dst,
1571
			      corkreq ? msg->msg_flags|MSG_MORE : msg->msg_flags);
L
Linus Torvalds 已提交
1572 1573 1574
	if (err)
		udp_v6_flush_pending_frames(sk);
	else if (!corkreq)
1575
		err = udp_v6_push_pending_frames(sk);
H
Herbert Xu 已提交
1576 1577
	else if (unlikely(skb_queue_empty(&sk->sk_write_queue)))
		up->pending = 0;
L
Linus Torvalds 已提交
1578

1579 1580 1581 1582
	if (err > 0)
		err = np->recverr ? net_xmit_errno(err) : 0;
	release_sock(sk);

L
Linus Torvalds 已提交
1583
out:
1584
	dst_release(dst);
1585
out_no_dst:
L
Linus Torvalds 已提交
1586
	fl6_sock_release(flowlabel);
1587
	txopt_put(opt_to_free);
1588
	if (!err)
L
Linus Torvalds 已提交
1589
		return len;
1590 1591 1592 1593 1594 1595 1596 1597
	/*
	 * 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)) {
1598 1599
		UDP6_INC_STATS(sock_net(sk),
			       UDP_MIB_SNDBUFERRORS, is_udplite);
1600
	}
L
Linus Torvalds 已提交
1601 1602 1603
	return err;

do_confirm:
1604 1605
	if (msg->msg_flags & MSG_PROBE)
		dst_confirm_neigh(dst, &fl6.daddr);
L
Linus Torvalds 已提交
1606 1607 1608 1609 1610 1611
	if (!(msg->msg_flags&MSG_PROBE) || len)
		goto back_from_confirm;
	err = 0;
	goto out;
}

1612
void udpv6_destroy_sock(struct sock *sk)
L
Linus Torvalds 已提交
1613
{
1614
	struct udp_sock *up = udp_sk(sk);
L
Linus Torvalds 已提交
1615
	lock_sock(sk);
1616 1617 1618

	/* protects from races with udp_abort() */
	sock_set_flag(sk, SOCK_DEAD);
L
Linus Torvalds 已提交
1619 1620 1621
	udp_v6_flush_pending_frames(sk);
	release_sock(sk);

1622 1623 1624 1625 1626 1627 1628
	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);
		}
1629
		if (up->encap_enabled) {
1630
			static_branch_dec(&udpv6_encap_needed_key);
1631 1632
			udp_encap_disable();
		}
1633 1634
	}

L
Linus Torvalds 已提交
1635 1636 1637 1638 1639 1640
	inet6_destroy_sock(sk);
}

/*
 *	Socket option code for UDP
 */
1641 1642
int udpv6_setsockopt(struct sock *sk, int level, int optname, sockptr_t optval,
		     unsigned int optlen)
1643
{
1644
	if (level == SOL_UDP  ||  level == SOL_UDPLITE)
1645
		return udp_lib_setsockopt(sk, level, optname,
1646
					  optval, optlen,
1647
					  udp_v6_push_pending_frames);
1648
	return ipv6_setsockopt(sk, level, optname, optval, optlen);
1649 1650
}

1651 1652
int udpv6_getsockopt(struct sock *sk, int level, int optname,
		     char __user *optval, int __user *optlen)
1653
{
1654
	if (level == SOL_UDP  ||  level == SOL_UDPLITE)
1655
		return udp_lib_getsockopt(sk, level, optname, optval, optlen);
1656
	return ipv6_getsockopt(sk, level, optname, optval, optlen);
1657 1658
}

1659 1660 1661
/* thinking of making this const? Don't.
 * early_demux can change based on sysctl.
 */
1662
static struct inet6_protocol udpv6_protocol = {
1663
	.early_demux	=	udp_v6_early_demux,
1664
	.early_demux_handler =  udp_v6_early_demux,
L
Linus Torvalds 已提交
1665 1666 1667 1668 1669 1670 1671
	.handler	=	udpv6_rcv,
	.err_handler	=	udpv6_err,
	.flags		=	INET6_PROTO_NOPOLICY|INET6_PROTO_FINAL,
};

/* ------------------------------------------------------------------------ */
#ifdef CONFIG_PROC_FS
1672
int udp6_seq_show(struct seq_file *seq, void *v)
L
Linus Torvalds 已提交
1673
{
1674 1675 1676 1677 1678 1679 1680
	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);
1681 1682
		__ip6_dgram_sock_seq_show(seq, v, srcp, destp,
					  udp_rqueue_get(v), bucket);
1683
	}
L
Linus Torvalds 已提交
1684 1685 1686
	return 0;
}

1687
const struct seq_operations udp6_seq_ops = {
1688 1689 1690 1691 1692
	.start		= udp_seq_start,
	.next		= udp_seq_next,
	.stop		= udp_seq_stop,
	.show		= udp6_seq_show,
};
1693
EXPORT_SYMBOL(udp6_seq_ops);
1694

L
Linus Torvalds 已提交
1695 1696
static struct udp_seq_afinfo udp6_seq_afinfo = {
	.family		= AF_INET6,
1697
	.udp_table	= &udp_table,
L
Linus Torvalds 已提交
1698 1699
};

1700
int __net_init udp6_proc_init(struct net *net)
L
Linus Torvalds 已提交
1701
{
1702 1703
	if (!proc_create_net_data("udp6", 0444, net->proc_net, &udp6_seq_ops,
			sizeof(struct udp_iter_state), &udp6_seq_afinfo))
1704 1705
		return -ENOMEM;
	return 0;
L
Linus Torvalds 已提交
1706 1707
}

1708 1709
void udp6_proc_exit(struct net *net)
{
1710
	remove_proc_entry("udp6", net->proc_net);
L
Linus Torvalds 已提交
1711 1712 1713 1714 1715 1716
}
#endif /* CONFIG_PROC_FS */

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

struct proto udpv6_prot = {
1717 1718 1719
	.name			= "UDPv6",
	.owner			= THIS_MODULE,
	.close			= udp_lib_close,
1720
	.pre_connect		= udpv6_pre_connect,
1721 1722 1723 1724 1725 1726 1727 1728 1729 1730 1731 1732 1733 1734
	.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,
1735 1736 1737
#ifdef CONFIG_BPF_SYSCALL
	.psock_update_sk_prot	= udp_bpf_update_proto,
#endif
1738 1739 1740 1741 1742 1743 1744
	.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
1745 1746 1747 1748 1749 1750 1751 1752 1753 1754
};

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

1755
int __init udpv6_init(void)
L
Linus Torvalds 已提交
1756
{
1757 1758
	int ret;

1759 1760
	ret = inet6_add_protocol(&udpv6_protocol, IPPROTO_UDP);
	if (ret)
1761
		goto out;
1762

1763 1764 1765 1766 1767 1768 1769 1770 1771 1772 1773
	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;
}

1774
void udpv6_exit(void)
1775 1776 1777
{
	inet6_unregister_protosw(&udpv6_protosw);
	inet6_del_protocol(&udpv6_protocol, IPPROTO_UDP);
L
Linus Torvalds 已提交
1778
}
新手
引导
客服 返回
顶部