udp.c 57.3 KB
Newer Older
L
Linus Torvalds 已提交
1 2 3 4 5 6 7
/*
 * INET		An implementation of the TCP/IP protocol suite for the LINUX
 *		operating system.  INET is implemented using the  BSD Socket
 *		interface as the means of communication with the user level.
 *
 *		The User Datagram Protocol (UDP).
 *
8
 * Authors:	Ross Biro
L
Linus Torvalds 已提交
9 10
 *		Fred N. van Kempen, <waltje@uWalt.NL.Mugnet.ORG>
 *		Arnt Gulbrandsen, <agulbra@nvg.unit.no>
11
 *		Alan Cox, <alan@lxorguk.ukuu.org.uk>
L
Linus Torvalds 已提交
12 13 14 15 16 17 18 19 20
 *		Hirokazu Takahashi, <taka@valinux.co.jp>
 *
 * Fixes:
 *		Alan Cox	:	verify_area() calls
 *		Alan Cox	: 	stopped close while in use off icmp
 *					messages. Not a fix but a botch that
 *					for udp at least is 'valid'.
 *		Alan Cox	:	Fixed icmp handling properly
 *		Alan Cox	: 	Correct error for oversized datagrams
21 22
 *		Alan Cox	:	Tidied select() semantics.
 *		Alan Cox	:	udp_err() fixed properly, also now
L
Linus Torvalds 已提交
23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56
 *					select and read wake correctly on errors
 *		Alan Cox	:	udp_send verify_area moved to avoid mem leak
 *		Alan Cox	:	UDP can count its memory
 *		Alan Cox	:	send to an unknown connection causes
 *					an ECONNREFUSED off the icmp, but
 *					does NOT close.
 *		Alan Cox	:	Switched to new sk_buff handlers. No more backlog!
 *		Alan Cox	:	Using generic datagram code. Even smaller and the PEEK
 *					bug no longer crashes it.
 *		Fred Van Kempen	: 	Net2e support for sk->broadcast.
 *		Alan Cox	:	Uses skb_free_datagram
 *		Alan Cox	:	Added get/set sockopt support.
 *		Alan Cox	:	Broadcasting without option set returns EACCES.
 *		Alan Cox	:	No wakeup calls. Instead we now use the callbacks.
 *		Alan Cox	:	Use ip_tos and ip_ttl
 *		Alan Cox	:	SNMP Mibs
 *		Alan Cox	:	MSG_DONTROUTE, and 0.0.0.0 support.
 *		Matt Dillon	:	UDP length checks.
 *		Alan Cox	:	Smarter af_inet used properly.
 *		Alan Cox	:	Use new kernel side addressing.
 *		Alan Cox	:	Incorrect return on truncated datagram receive.
 *	Arnt Gulbrandsen 	:	New udp_send and stuff
 *		Alan Cox	:	Cache last socket
 *		Alan Cox	:	Route cache
 *		Jon Peatfield	:	Minor efficiency fix to sendto().
 *		Mike Shaver	:	RFC1122 checks.
 *		Alan Cox	:	Nonblocking error fix.
 *	Willy Konynenberg	:	Transparent proxying support.
 *		Mike McLagan	:	Routing by source
 *		David S. Miller	:	New socket lookup architecture.
 *					Last socket cache retained as it
 *					does have a high hit rate.
 *		Olaf Kirch	:	Don't linearise iovec on sendmsg.
 *		Andi Kleen	:	Some cleanups, cache destination entry
57
 *					for connect.
L
Linus Torvalds 已提交
58 59 60 61 62 63 64 65 66 67 68 69 70
 *	Vitaly E. Lavrov	:	Transparent proxy revived after year coma.
 *		Melvin Smith	:	Check msg_name not msg_namelen in sendto(),
 *					return ENOTCONN for unconnected sockets (POSIX)
 *		Janos Farkas	:	don't deliver multi/broadcasts to a different
 *					bound-to-device socket
 *	Hirokazu Takahashi	:	HW checksumming for outgoing UDP
 *					datagrams.
 *	Hirokazu Takahashi	:	sendfile() on UDP works now.
 *		Arnaldo C. Melo :	convert /proc/net/udp to seq_file
 *	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.
 *	Derek Atkins <derek@ihtfp.com>: Add Encapulation Support
71
 *	James Chapman		:	Add L2TP encapsulation type.
L
Linus Torvalds 已提交
72 73 74 75 76 77 78
 *
 *
 *		This program is free software; you can redistribute it and/or
 *		modify it under the terms of the GNU General Public License
 *		as published by the Free Software Foundation; either version
 *		2 of the License, or (at your option) any later version.
 */
79

L
Linus Torvalds 已提交
80 81 82
#include <asm/system.h>
#include <asm/uaccess.h>
#include <asm/ioctls.h>
H
Hideo Aoki 已提交
83
#include <linux/bootmem.h>
84 85
#include <linux/highmem.h>
#include <linux/swap.h>
L
Linus Torvalds 已提交
86 87 88 89 90
#include <linux/types.h>
#include <linux/fcntl.h>
#include <linux/module.h>
#include <linux/socket.h>
#include <linux/sockios.h>
91
#include <linux/igmp.h>
L
Linus Torvalds 已提交
92 93 94 95 96 97
#include <linux/in.h>
#include <linux/errno.h>
#include <linux/timer.h>
#include <linux/mm.h>
#include <linux/inet.h>
#include <linux/netdevice.h>
98
#include <linux/slab.h>
99
#include <net/tcp_states.h>
L
Linus Torvalds 已提交
100 101 102
#include <linux/skbuff.h>
#include <linux/proc_fs.h>
#include <linux/seq_file.h>
103
#include <net/net_namespace.h>
L
Linus Torvalds 已提交
104 105 106 107
#include <net/icmp.h>
#include <net/route.h>
#include <net/checksum.h>
#include <net/xfrm.h>
108
#include <trace/events/udp.h>
109
#include "udp_impl.h"
L
Linus Torvalds 已提交
110

111
struct udp_table udp_table __read_mostly;
112
EXPORT_SYMBOL(udp_table);
L
Linus Torvalds 已提交
113

E
Eric Dumazet 已提交
114
long sysctl_udp_mem[3] __read_mostly;
H
Hideo Aoki 已提交
115
EXPORT_SYMBOL(sysctl_udp_mem);
E
Eric Dumazet 已提交
116 117

int sysctl_udp_rmem_min __read_mostly;
H
Hideo Aoki 已提交
118
EXPORT_SYMBOL(sysctl_udp_rmem_min);
E
Eric Dumazet 已提交
119 120

int sysctl_udp_wmem_min __read_mostly;
H
Hideo Aoki 已提交
121 122
EXPORT_SYMBOL(sysctl_udp_wmem_min);

E
Eric Dumazet 已提交
123
atomic_long_t udp_memory_allocated;
H
Hideo Aoki 已提交
124 125
EXPORT_SYMBOL(udp_memory_allocated);

126 127
#define MAX_UDP_PORTS 65536
#define PORTS_PER_CHAIN (MAX_UDP_PORTS / UDP_HTABLE_SIZE_MIN)
128

129
static int udp_lib_lport_inuse(struct net *net, __u16 num,
130
			       const struct udp_hslot *hslot,
131
			       unsigned long *bitmap,
132 133
			       struct sock *sk,
			       int (*saddr_comp)(const struct sock *sk1,
134 135
						 const struct sock *sk2),
			       unsigned int log)
L
Linus Torvalds 已提交
136
{
137
	struct sock *sk2;
138
	struct hlist_nulls_node *node;
139

140
	sk_nulls_for_each(sk2, node, &hslot->head)
141 142
		if (net_eq(sock_net(sk2), net) &&
		    sk2 != sk &&
143
		    (bitmap || udp_sk(sk2)->udp_port_hash == num) &&
144 145 146
		    (!sk2->sk_reuse || !sk->sk_reuse) &&
		    (!sk2->sk_bound_dev_if || !sk->sk_bound_dev_if ||
		     sk2->sk_bound_dev_if == sk->sk_bound_dev_if) &&
147 148
		    (*saddr_comp)(sk, sk2)) {
			if (bitmap)
149 150
				__set_bit(udp_sk(sk2)->udp_port_hash >> log,
					  bitmap);
151 152 153
			else
				return 1;
		}
154 155 156
	return 0;
}

E
Eric Dumazet 已提交
157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172
/*
 * Note: we still hold spinlock of primary hash chain, so no other writer
 * can insert/delete a socket with local_port == num
 */
static int udp_lib_lport_inuse2(struct net *net, __u16 num,
			       struct udp_hslot *hslot2,
			       struct sock *sk,
			       int (*saddr_comp)(const struct sock *sk1,
						 const struct sock *sk2))
{
	struct sock *sk2;
	struct hlist_nulls_node *node;
	int res = 0;

	spin_lock(&hslot2->lock);
	udp_portaddr_for_each_entry(sk2, node, &hslot2->head)
173 174 175 176 177 178
		if (net_eq(sock_net(sk2), net) &&
		    sk2 != sk &&
		    (udp_sk(sk2)->udp_port_hash == num) &&
		    (!sk2->sk_reuse || !sk->sk_reuse) &&
		    (!sk2->sk_bound_dev_if || !sk->sk_bound_dev_if ||
		     sk2->sk_bound_dev_if == sk->sk_bound_dev_if) &&
E
Eric Dumazet 已提交
179 180 181 182 183 184 185 186
		    (*saddr_comp)(sk, sk2)) {
			res = 1;
			break;
		}
	spin_unlock(&hslot2->lock);
	return res;
}

187
/**
188
 *  udp_lib_get_port  -  UDP/-Lite port lookup for IPv4 and IPv6
189 190 191
 *
 *  @sk:          socket struct in question
 *  @snum:        port number to look up
192
 *  @saddr_comp:  AF-dependent comparison of bound local IP addresses
L
Lucas De Marchi 已提交
193
 *  @hash2_nulladdr: AF-dependent hash value in secondary hash chains,
E
Eric Dumazet 已提交
194
 *                   with NULL address
195
 */
196
int udp_lib_get_port(struct sock *sk, unsigned short snum,
197
		       int (*saddr_comp)(const struct sock *sk1,
E
Eric Dumazet 已提交
198 199
					 const struct sock *sk2),
		     unsigned int hash2_nulladdr)
200
{
201
	struct udp_hslot *hslot, *hslot2;
202
	struct udp_table *udptable = sk->sk_prot->h.udp_table;
203
	int    error = 1;
204
	struct net *net = sock_net(sk);
L
Linus Torvalds 已提交
205

206
	if (!snum) {
E
Eric Dumazet 已提交
207 208
		int low, high, remaining;
		unsigned rand;
209 210
		unsigned short first, last;
		DECLARE_BITMAP(bitmap, PORTS_PER_CHAIN);
211

212
		inet_get_local_port_range(&low, &high);
213
		remaining = (high - low) + 1;
214

E
Eric Dumazet 已提交
215
		rand = net_random();
216 217 218 219
		first = (((u64)rand * remaining) >> 32) + low;
		/*
		 * force rand to be an odd multiple of UDP_HTABLE_SIZE
		 */
220
		rand = (rand | 1) * (udptable->mask + 1);
E
Eric Dumazet 已提交
221 222
		last = first + udptable->mask + 1;
		do {
223
			hslot = udp_hashslot(udptable, net, first);
224
			bitmap_zero(bitmap, PORTS_PER_CHAIN);
225
			spin_lock_bh(&hslot->lock);
226
			udp_lib_lport_inuse(net, snum, hslot, bitmap, sk,
227
					    saddr_comp, udptable->log);
228 229 230 231 232 233 234

			snum = first;
			/*
			 * Iterate on all possible values of snum for this hash.
			 * Using steps of an odd multiple of UDP_HTABLE_SIZE
			 * give us randomization and full range coverage.
			 */
E
Eric Dumazet 已提交
235
			do {
236
				if (low <= snum && snum <= high &&
237 238
				    !test_bit(snum >> udptable->log, bitmap) &&
				    !inet_is_reserved_local_port(snum))
239 240 241 242
					goto found;
				snum += rand;
			} while (snum != first);
			spin_unlock_bh(&hslot->lock);
E
Eric Dumazet 已提交
243
		} while (++first != last);
244
		goto fail;
245
	} else {
246
		hslot = udp_hashslot(udptable, net, snum);
247
		spin_lock_bh(&hslot->lock);
E
Eric Dumazet 已提交
248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271
		if (hslot->count > 10) {
			int exist;
			unsigned int slot2 = udp_sk(sk)->udp_portaddr_hash ^ snum;

			slot2          &= udptable->mask;
			hash2_nulladdr &= udptable->mask;

			hslot2 = udp_hashslot2(udptable, slot2);
			if (hslot->count < hslot2->count)
				goto scan_primary_hash;

			exist = udp_lib_lport_inuse2(net, snum, hslot2,
						     sk, saddr_comp);
			if (!exist && (hash2_nulladdr != slot2)) {
				hslot2 = udp_hashslot2(udptable, hash2_nulladdr);
				exist = udp_lib_lport_inuse2(net, snum, hslot2,
							     sk, saddr_comp);
			}
			if (exist)
				goto fail_unlock;
			else
				goto found;
		}
scan_primary_hash:
272 273
		if (udp_lib_lport_inuse(net, snum, hslot, NULL, sk,
					saddr_comp, 0))
274 275
			goto fail_unlock;
	}
276
found:
E
Eric Dumazet 已提交
277
	inet_sk(sk)->inet_num = snum;
278 279
	udp_sk(sk)->udp_port_hash = snum;
	udp_sk(sk)->udp_portaddr_hash ^= snum;
L
Linus Torvalds 已提交
280
	if (sk_unhashed(sk)) {
281
		sk_nulls_add_node_rcu(sk, &hslot->head);
E
Eric Dumazet 已提交
282
		hslot->count++;
283
		sock_prot_inuse_add(sock_net(sk), sk->sk_prot, 1);
284 285 286 287 288 289 290

		hslot2 = udp_hashslot2(udptable, udp_sk(sk)->udp_portaddr_hash);
		spin_lock(&hslot2->lock);
		hlist_nulls_add_head_rcu(&udp_sk(sk)->udp_portaddr_node,
					 &hslot2->head);
		hslot2->count++;
		spin_unlock(&hslot2->lock);
L
Linus Torvalds 已提交
291
	}
292
	error = 0;
293 294
fail_unlock:
	spin_unlock_bh(&hslot->lock);
L
Linus Torvalds 已提交
295
fail:
296 297
	return error;
}
E
Eric Dumazet 已提交
298
EXPORT_SYMBOL(udp_lib_get_port);
299

300
static int ipv4_rcv_saddr_equal(const struct sock *sk1, const struct sock *sk2)
301 302 303
{
	struct inet_sock *inet1 = inet_sk(sk1), *inet2 = inet_sk(sk2);

E
Eric Dumazet 已提交
304
	return 	(!ipv6_only_sock(sk2)  &&
E
Eric Dumazet 已提交
305 306
		 (!inet1->inet_rcv_saddr || !inet2->inet_rcv_saddr ||
		   inet1->inet_rcv_saddr == inet2->inet_rcv_saddr));
307 308
}

309 310 311
static unsigned int udp4_portaddr_hash(struct net *net, __be32 saddr,
				       unsigned int port)
{
312
	return jhash_1word((__force u32)saddr, net_hash_mix(net)) ^ port;
313 314
}

315
int udp_v4_get_port(struct sock *sk, unsigned short snum)
316
{
E
Eric Dumazet 已提交
317
	unsigned int hash2_nulladdr =
318
		udp4_portaddr_hash(sock_net(sk), htonl(INADDR_ANY), snum);
E
Eric Dumazet 已提交
319 320 321
	unsigned int hash2_partial =
		udp4_portaddr_hash(sock_net(sk), inet_sk(sk)->inet_rcv_saddr, 0);

322
	/* precompute partial secondary hash */
E
Eric Dumazet 已提交
323 324
	udp_sk(sk)->udp_portaddr_hash = hash2_partial;
	return udp_lib_get_port(sk, snum, ipv4_rcv_saddr_equal, hash2_nulladdr);
325 326
}

327 328 329 330 331 332
static inline int compute_score(struct sock *sk, struct net *net, __be32 saddr,
			 unsigned short hnum,
			 __be16 sport, __be32 daddr, __be16 dport, int dif)
{
	int score = -1;

333
	if (net_eq(sock_net(sk), net) && udp_sk(sk)->udp_port_hash == hnum &&
334 335 336 337
			!ipv6_only_sock(sk)) {
		struct inet_sock *inet = inet_sk(sk);

		score = (sk->sk_family == PF_INET ? 1 : 0);
E
Eric Dumazet 已提交
338 339
		if (inet->inet_rcv_saddr) {
			if (inet->inet_rcv_saddr != daddr)
340 341 342
				return -1;
			score += 2;
		}
E
Eric Dumazet 已提交
343 344
		if (inet->inet_daddr) {
			if (inet->inet_daddr != saddr)
345 346 347
				return -1;
			score += 2;
		}
E
Eric Dumazet 已提交
348 349
		if (inet->inet_dport) {
			if (inet->inet_dport != sport)
350 351 352 353 354 355 356 357 358 359 360 361
				return -1;
			score += 2;
		}
		if (sk->sk_bound_dev_if) {
			if (sk->sk_bound_dev_if != dif)
				return -1;
			score += 2;
		}
	}
	return score;
}

E
Eric Dumazet 已提交
362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433
/*
 * In this second variant, we check (daddr, dport) matches (inet_rcv_sadd, inet_num)
 */
#define SCORE2_MAX (1 + 2 + 2 + 2)
static inline int compute_score2(struct sock *sk, struct net *net,
				 __be32 saddr, __be16 sport,
				 __be32 daddr, unsigned int hnum, int dif)
{
	int score = -1;

	if (net_eq(sock_net(sk), net) && !ipv6_only_sock(sk)) {
		struct inet_sock *inet = inet_sk(sk);

		if (inet->inet_rcv_saddr != daddr)
			return -1;
		if (inet->inet_num != hnum)
			return -1;

		score = (sk->sk_family == PF_INET ? 1 : 0);
		if (inet->inet_daddr) {
			if (inet->inet_daddr != saddr)
				return -1;
			score += 2;
		}
		if (inet->inet_dport) {
			if (inet->inet_dport != sport)
				return -1;
			score += 2;
		}
		if (sk->sk_bound_dev_if) {
			if (sk->sk_bound_dev_if != dif)
				return -1;
			score += 2;
		}
	}
	return score;
}


/* called with read_rcu_lock() */
static struct sock *udp4_lib_lookup2(struct net *net,
		__be32 saddr, __be16 sport,
		__be32 daddr, unsigned int hnum, int dif,
		struct udp_hslot *hslot2, unsigned int slot2)
{
	struct sock *sk, *result;
	struct hlist_nulls_node *node;
	int score, badness;

begin:
	result = NULL;
	badness = -1;
	udp_portaddr_for_each_entry_rcu(sk, node, &hslot2->head) {
		score = compute_score2(sk, net, saddr, sport,
				      daddr, hnum, dif);
		if (score > badness) {
			result = sk;
			badness = score;
			if (score == SCORE2_MAX)
				goto exact_match;
		}
	}
	/*
	 * if the nulls value we got at the end of this lookup is
	 * not the expected one, we must restart lookup.
	 * We probably met an item that was moved to another chain.
	 */
	if (get_nulls_value(node) != slot2)
		goto begin;

	if (result) {
exact_match:
E
Eric Dumazet 已提交
434
		if (unlikely(!atomic_inc_not_zero_hint(&result->sk_refcnt, 2)))
E
Eric Dumazet 已提交
435 436 437 438 439 440 441 442 443 444
			result = NULL;
		else if (unlikely(compute_score2(result, net, saddr, sport,
				  daddr, hnum, dif) < badness)) {
			sock_put(result);
			goto begin;
		}
	}
	return result;
}

445 446 447
/* UDP is nearly always wildcards out the wazoo, it makes no sense to try
 * harder than this. -DaveM
 */
448
struct sock *__udp4_lib_lookup(struct net *net, __be32 saddr,
449
		__be16 sport, __be32 daddr, __be16 dport,
450
		int dif, struct udp_table *udptable)
451
{
452
	struct sock *sk, *result;
453
	struct hlist_nulls_node *node;
454
	unsigned short hnum = ntohs(dport);
E
Eric Dumazet 已提交
455 456
	unsigned int hash2, slot2, slot = udp_hashfn(net, hnum, udptable->mask);
	struct udp_hslot *hslot2, *hslot = &udptable->hash[slot];
457
	int score, badness;
458

459
	rcu_read_lock();
E
Eric Dumazet 已提交
460 461 462 463 464 465 466 467 468 469 470
	if (hslot->count > 10) {
		hash2 = udp4_portaddr_hash(net, daddr, hnum);
		slot2 = hash2 & udptable->mask;
		hslot2 = &udptable->hash2[slot2];
		if (hslot->count < hslot2->count)
			goto begin;

		result = udp4_lib_lookup2(net, saddr, sport,
					  daddr, hnum, dif,
					  hslot2, slot2);
		if (!result) {
471
			hash2 = udp4_portaddr_hash(net, htonl(INADDR_ANY), hnum);
E
Eric Dumazet 已提交
472 473 474 475 476
			slot2 = hash2 & udptable->mask;
			hslot2 = &udptable->hash2[slot2];
			if (hslot->count < hslot2->count)
				goto begin;

477
			result = udp4_lib_lookup2(net, saddr, sport,
478
						  htonl(INADDR_ANY), hnum, dif,
E
Eric Dumazet 已提交
479 480 481 482 483
						  hslot2, slot2);
		}
		rcu_read_unlock();
		return result;
	}
484 485 486
begin:
	result = NULL;
	badness = -1;
487
	sk_nulls_for_each_rcu(sk, node, &hslot->head) {
488 489 490 491 492
		score = compute_score(sk, net, saddr, hnum, sport,
				      daddr, dport, dif);
		if (score > badness) {
			result = sk;
			badness = score;
493 494
		}
	}
495 496 497 498 499
	/*
	 * if the nulls value we got at the end of this lookup is
	 * not the expected one, we must restart lookup.
	 * We probably met an item that was moved to another chain.
	 */
E
Eric Dumazet 已提交
500
	if (get_nulls_value(node) != slot)
501 502
		goto begin;

503
	if (result) {
E
Eric Dumazet 已提交
504
		if (unlikely(!atomic_inc_not_zero_hint(&result->sk_refcnt, 2)))
505 506 507 508 509 510 511 512
			result = NULL;
		else if (unlikely(compute_score(result, net, saddr, hnum, sport,
				  daddr, dport, dif) < badness)) {
			sock_put(result);
			goto begin;
		}
	}
	rcu_read_unlock();
513 514
	return result;
}
515
EXPORT_SYMBOL_GPL(__udp4_lib_lookup);
516

517 518
static inline struct sock *__udp4_lib_lookup_skb(struct sk_buff *skb,
						 __be16 sport, __be16 dport,
519
						 struct udp_table *udptable)
520
{
521
	struct sock *sk;
522 523
	const struct iphdr *iph = ip_hdr(skb);

524 525 526
	if (unlikely(sk = skb_steal_sock(skb)))
		return sk;
	else
E
Eric Dumazet 已提交
527
		return __udp4_lib_lookup(dev_net(skb_dst(skb)->dev), iph->saddr, sport,
528 529
					 iph->daddr, dport, inet_iif(skb),
					 udptable);
530 531
}

532 533 534
struct sock *udp4_lib_lookup(struct net *net, __be32 saddr, __be16 sport,
			     __be32 daddr, __be16 dport, int dif)
{
535
	return __udp4_lib_lookup(net, saddr, sport, daddr, dport, dif, &udp_table);
536 537 538
}
EXPORT_SYMBOL_GPL(udp4_lib_lookup);

539
static inline struct sock *udp_v4_mcast_next(struct net *net, struct sock *sk,
540 541 542 543
					     __be16 loc_port, __be32 loc_addr,
					     __be16 rmt_port, __be32 rmt_addr,
					     int dif)
{
544
	struct hlist_nulls_node *node;
545 546 547
	struct sock *s = sk;
	unsigned short hnum = ntohs(loc_port);

548
	sk_nulls_for_each_from(s, node) {
549 550
		struct inet_sock *inet = inet_sk(s);

551 552 553 554 555 556 557
		if (!net_eq(sock_net(s), net) ||
		    udp_sk(s)->udp_port_hash != hnum ||
		    (inet->inet_daddr && inet->inet_daddr != rmt_addr) ||
		    (inet->inet_dport != rmt_port && inet->inet_dport) ||
		    (inet->inet_rcv_saddr &&
		     inet->inet_rcv_saddr != loc_addr) ||
		    ipv6_only_sock(s) ||
558 559 560 561 562 563 564 565 566 567 568 569 570 571 572 573 574 575 576 577 578 579
		    (s->sk_bound_dev_if && s->sk_bound_dev_if != dif))
			continue;
		if (!ip_mc_sf_allow(s, loc_addr, rmt_addr, dif))
			continue;
		goto found;
	}
	s = NULL;
found:
	return s;
}

/*
 * This routine is called by the ICMP module when it gets some
 * sort of error condition.  If err < 0 then the socket should
 * be closed and the error returned to the user.  If err > 0
 * it's just the icmp type << 8 | icmp code.
 * Header points to the ip header of the error packet. We move
 * on past this. Then (as it used to claim before adjustment)
 * header points to the first 8 bytes of the udp header.  We need
 * to find the appropriate port.
 */

580
void __udp4_lib_err(struct sk_buff *skb, u32 info, struct udp_table *udptable)
581 582
{
	struct inet_sock *inet;
583
	const struct iphdr *iph = (const struct iphdr *)skb->data;
E
Eric Dumazet 已提交
584
	struct udphdr *uh = (struct udphdr *)(skb->data+(iph->ihl<<2));
585 586 587 588 589
	const int type = icmp_hdr(skb)->type;
	const int code = icmp_hdr(skb)->code;
	struct sock *sk;
	int harderr;
	int err;
590
	struct net *net = dev_net(skb->dev);
591

592
	sk = __udp4_lib_lookup(net, iph->daddr, uh->dest,
593 594
			iph->saddr, uh->source, skb->dev->ifindex, udptable);
	if (sk == NULL) {
595
		ICMP_INC_STATS_BH(net, ICMP_MIB_INERRORS);
596 597 598 599 600 601 602 603 604 605 606 607 608 609 610 611 612 613 614 615 616 617 618 619 620 621 622 623 624 625 626 627 628 629 630 631 632 633 634 635 636 637
		return;	/* No socket for error */
	}

	err = 0;
	harderr = 0;
	inet = inet_sk(sk);

	switch (type) {
	default:
	case ICMP_TIME_EXCEEDED:
		err = EHOSTUNREACH;
		break;
	case ICMP_SOURCE_QUENCH:
		goto out;
	case ICMP_PARAMETERPROB:
		err = EPROTO;
		harderr = 1;
		break;
	case ICMP_DEST_UNREACH:
		if (code == ICMP_FRAG_NEEDED) { /* Path MTU discovery */
			if (inet->pmtudisc != IP_PMTUDISC_DONT) {
				err = EMSGSIZE;
				harderr = 1;
				break;
			}
			goto out;
		}
		err = EHOSTUNREACH;
		if (code <= NR_ICMP_UNREACH) {
			harderr = icmp_err_convert[code].fatal;
			err = icmp_err_convert[code].errno;
		}
		break;
	}

	/*
	 *      RFC1122: OK.  Passes ICMP errors back to application, as per
	 *	4.1.3.3.
	 */
	if (!inet->recverr) {
		if (!harderr || sk->sk_state != TCP_ESTABLISHED)
			goto out;
638
	} else
E
Eric Dumazet 已提交
639
		ip_icmp_error(sk, skb, err, uh->dest, info, (u8 *)(uh+1));
640

641 642 643 644 645 646 647 648
	sk->sk_err = err;
	sk->sk_error_report(sk);
out:
	sock_put(sk);
}

void udp_err(struct sk_buff *skb, u32 info)
{
649
	__udp4_lib_err(skb, info, &udp_table);
650 651 652 653 654
}

/*
 * Throw away all pending data and cancel the corking. Socket is locked.
 */
655
void udp_flush_pending_frames(struct sock *sk)
656 657 658 659 660 661 662 663 664
{
	struct udp_sock *up = udp_sk(sk);

	if (up->pending) {
		up->len = 0;
		up->pending = 0;
		ip_flush_pending_frames(sk);
	}
}
665
EXPORT_SYMBOL(udp_flush_pending_frames);
666 667

/**
H
Herbert Xu 已提交
668
 * 	udp4_hwcsum  -  handle outgoing HW checksumming
669 670
 * 	@skb: 	sk_buff containing the filled-in UDP header
 * 	        (checksum field must be zeroed out)
H
Herbert Xu 已提交
671 672
 *	@src:	source IP address
 *	@dst:	destination IP address
673
 */
H
Herbert Xu 已提交
674
static void udp4_hwcsum(struct sk_buff *skb, __be32 src, __be32 dst)
675 676
{
	struct udphdr *uh = udp_hdr(skb);
H
Herbert Xu 已提交
677 678 679 680
	struct sk_buff *frags = skb_shinfo(skb)->frag_list;
	int offset = skb_transport_offset(skb);
	int len = skb->len - offset;
	int hlen = len;
681 682
	__wsum csum = 0;

H
Herbert Xu 已提交
683
	if (!frags) {
684 685 686 687 688
		/*
		 * Only one fragment on the socket.
		 */
		skb->csum_start = skb_transport_header(skb) - skb->head;
		skb->csum_offset = offsetof(struct udphdr, check);
H
Herbert Xu 已提交
689 690
		uh->check = ~csum_tcpudp_magic(src, dst, len,
					       IPPROTO_UDP, 0);
691 692 693 694 695 696
	} 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
		 */
H
Herbert Xu 已提交
697 698 699 700
		do {
			csum = csum_add(csum, frags->csum);
			hlen -= frags->len;
		} while ((frags = frags->next));
701

H
Herbert Xu 已提交
702
		csum = skb_checksum(skb, offset, hlen, csum);
703 704 705 706 707 708 709 710
		skb->ip_summed = CHECKSUM_NONE;

		uh->check = csum_tcpudp_magic(src, dst, len, IPPROTO_UDP, csum);
		if (uh->check == 0)
			uh->check = CSUM_MANGLED_0;
	}
}

711
static int udp_send_skb(struct sk_buff *skb, struct flowi4 *fl4)
712
{
H
Herbert Xu 已提交
713
	struct sock *sk = skb->sk;
714 715 716 717
	struct inet_sock *inet = inet_sk(sk);
	struct udphdr *uh;
	int err = 0;
	int is_udplite = IS_UDPLITE(sk);
H
Herbert Xu 已提交
718 719
	int offset = skb_transport_offset(skb);
	int len = skb->len - offset;
720 721 722 723 724 725
	__wsum csum = 0;

	/*
	 * Create a UDP header
	 */
	uh = udp_hdr(skb);
H
Herbert Xu 已提交
726
	uh->source = inet->inet_sport;
727
	uh->dest = fl4->fl4_dport;
H
Herbert Xu 已提交
728
	uh->len = htons(len);
729 730 731
	uh->check = 0;

	if (is_udplite)  				 /*     UDP-Lite      */
H
Herbert Xu 已提交
732
		csum = udplite_csum(skb);
733 734 735 736 737 738 739 740

	else if (sk->sk_no_check == UDP_CSUM_NOXMIT) {   /* UDP csum disabled */

		skb->ip_summed = CHECKSUM_NONE;
		goto send;

	} else if (skb->ip_summed == CHECKSUM_PARTIAL) { /* UDP hardware csum */

741
		udp4_hwcsum(skb, fl4->saddr, fl4->daddr);
742 743
		goto send;

H
Herbert Xu 已提交
744 745
	} else
		csum = udp_csum(skb);
746 747

	/* add protocol-dependent pseudo-header */
748
	uh->check = csum_tcpudp_magic(fl4->saddr, fl4->daddr, len,
E
Eric Dumazet 已提交
749
				      sk->sk_protocol, csum);
750 751 752 753
	if (uh->check == 0)
		uh->check = CSUM_MANGLED_0;

send:
H
Herbert Xu 已提交
754
	err = ip_send_skb(skb);
E
Eric Dumazet 已提交
755 756 757 758 759 760 761 762 763
	if (err) {
		if (err == -ENOBUFS && !inet->recverr) {
			UDP_INC_STATS_USER(sock_net(sk),
					   UDP_MIB_SNDBUFERRORS, is_udplite);
			err = 0;
		}
	} else
		UDP_INC_STATS_USER(sock_net(sk),
				   UDP_MIB_OUTDATAGRAMS, is_udplite);
H
Herbert Xu 已提交
764 765 766 767 768 769 770 771 772 773
	return err;
}

/*
 * Push out all pending data as one UDP datagram. Socket is locked.
 */
static int udp_push_pending_frames(struct sock *sk)
{
	struct udp_sock  *up = udp_sk(sk);
	struct inet_sock *inet = inet_sk(sk);
D
David S. Miller 已提交
774
	struct flowi4 *fl4 = &inet->cork.fl.u.ip4;
H
Herbert Xu 已提交
775 776 777
	struct sk_buff *skb;
	int err = 0;

778
	skb = ip_finish_skb(sk, fl4);
H
Herbert Xu 已提交
779 780 781
	if (!skb)
		goto out;

782
	err = udp_send_skb(skb, fl4);
H
Herbert Xu 已提交
783

784 785 786 787 788 789 790 791 792 793 794
out:
	up->len = 0;
	up->pending = 0;
	return err;
}

int udp_sendmsg(struct kiocb *iocb, struct sock *sk, struct msghdr *msg,
		size_t len)
{
	struct inet_sock *inet = inet_sk(sk);
	struct udp_sock *up = udp_sk(sk);
795
	struct flowi4 fl4_stack;
D
David S. Miller 已提交
796
	struct flowi4 *fl4;
797 798 799 800 801 802 803 804 805 806 807
	int ulen = len;
	struct ipcm_cookie ipc;
	struct rtable *rt = NULL;
	int free = 0;
	int connected = 0;
	__be32 daddr, faddr, saddr;
	__be16 dport;
	u8  tos;
	int err, is_udplite = IS_UDPLITE(sk);
	int corkreq = up->corkflag || msg->msg_flags&MSG_MORE;
	int (*getfrag)(void *, char *, int, int, int, struct sk_buff *);
H
Herbert Xu 已提交
808
	struct sk_buff *skb;
809
	struct ip_options_data opt_copy;
810 811 812 813 814 815 816 817

	if (len > 0xFFFF)
		return -EMSGSIZE;

	/*
	 *	Check the flags.
	 */

E
Eric Dumazet 已提交
818
	if (msg->msg_flags & MSG_OOB) /* Mirror BSD error message compatibility */
819 820 821
		return -EOPNOTSUPP;

	ipc.opt = NULL;
822
	ipc.tx_flags = 0;
823

H
Herbert Xu 已提交
824 825
	getfrag = is_udplite ? udplite_getfrag : ip_generic_getfrag;

826
	fl4 = &inet->cork.fl.u.ip4;
827 828 829 830 831 832 833 834 835 836 837 838 839 840 841 842 843 844 845 846 847
	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_INET)) {
				release_sock(sk);
				return -EINVAL;
			}
			goto do_append_data;
		}
		release_sock(sk);
	}
	ulen += sizeof(struct udphdr);

	/*
	 *	Get and verify the address.
	 */
	if (msg->msg_name) {
E
Eric Dumazet 已提交
848
		struct sockaddr_in * usin = (struct sockaddr_in *)msg->msg_name;
849 850 851 852 853 854 855 856 857 858 859 860 861 862
		if (msg->msg_namelen < sizeof(*usin))
			return -EINVAL;
		if (usin->sin_family != AF_INET) {
			if (usin->sin_family != AF_UNSPEC)
				return -EAFNOSUPPORT;
		}

		daddr = usin->sin_addr.s_addr;
		dport = usin->sin_port;
		if (dport == 0)
			return -EINVAL;
	} else {
		if (sk->sk_state != TCP_ESTABLISHED)
			return -EDESTADDRREQ;
E
Eric Dumazet 已提交
863 864
		daddr = inet->inet_daddr;
		dport = inet->inet_dport;
865 866 867 868 869
		/* Open fast path for connected socket.
		   Route will not be used, if at least one option is set.
		 */
		connected = 1;
	}
E
Eric Dumazet 已提交
870
	ipc.addr = inet->inet_saddr;
871 872

	ipc.oif = sk->sk_bound_dev_if;
873
	err = sock_tx_timestamp(sk, &ipc.tx_flags);
874 875
	if (err)
		return err;
876
	if (msg->msg_controllen) {
877
		err = ip_cmsg_send(sock_net(sk), msg, &ipc);
878 879 880 881 882 883
		if (err)
			return err;
		if (ipc.opt)
			free = 1;
		connected = 0;
	}
884 885 886 887 888 889 890 891 892 893 894 895
	if (!ipc.opt) {
		struct ip_options_rcu *inet_opt;

		rcu_read_lock();
		inet_opt = rcu_dereference(inet->inet_opt);
		if (inet_opt) {
			memcpy(&opt_copy, inet_opt,
			       sizeof(*inet_opt) + inet_opt->opt.optlen);
			ipc.opt = &opt_copy.opt;
		}
		rcu_read_unlock();
	}
896 897 898 899

	saddr = ipc.addr;
	ipc.addr = faddr = daddr;

900
	if (ipc.opt && ipc.opt->opt.srr) {
901 902
		if (!daddr)
			return -EINVAL;
903
		faddr = ipc.opt->opt.faddr;
904 905 906 907 908
		connected = 0;
	}
	tos = RT_TOS(inet->tos);
	if (sock_flag(sk, SOCK_LOCALROUTE) ||
	    (msg->msg_flags & MSG_DONTROUTE) ||
909
	    (ipc.opt && ipc.opt->opt.is_strictroute)) {
910 911 912 913 914 915 916 917 918 919 920 921 922
		tos |= RTO_ONLINK;
		connected = 0;
	}

	if (ipv4_is_multicast(daddr)) {
		if (!ipc.oif)
			ipc.oif = inet->mc_index;
		if (!saddr)
			saddr = inet->mc_addr;
		connected = 0;
	}

	if (connected)
E
Eric Dumazet 已提交
923
		rt = (struct rtable *)sk_dst_check(sk, 0);
924 925

	if (rt == NULL) {
926 927
		struct net *net = sock_net(sk);

928 929
		fl4 = &fl4_stack;
		flowi4_init_output(fl4, ipc.oif, sk->sk_mark, tos,
930 931 932 933
				   RT_SCOPE_UNIVERSE, sk->sk_protocol,
				   inet_sk_flowi_flags(sk)|FLOWI_FLAG_CAN_SLEEP,
				   faddr, saddr, dport, inet->inet_sport);

934 935
		security_sk_classify_flow(sk, flowi4_to_flowi(fl4));
		rt = ip_route_output_flow(net, fl4, sk);
936 937
		if (IS_ERR(rt)) {
			err = PTR_ERR(rt);
938
			rt = NULL;
939
			if (err == -ENETUNREACH)
940
				IP_INC_STATS_BH(net, IPSTATS_MIB_OUTNOROUTES);
941 942 943 944 945 946 947 948
			goto out;
		}

		err = -EACCES;
		if ((rt->rt_flags & RTCF_BROADCAST) &&
		    !sock_flag(sk, SOCK_BROADCAST))
			goto out;
		if (connected)
949
			sk_dst_set(sk, dst_clone(&rt->dst));
950 951 952 953 954 955
	}

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

956
	saddr = fl4->saddr;
957
	if (!ipc.addr)
958
		daddr = ipc.addr = fl4->daddr;
959

H
Herbert Xu 已提交
960 961
	/* Lockless fast path for the non-corking case. */
	if (!corkreq) {
962
		skb = ip_make_skb(sk, fl4, getfrag, msg->msg_iov, ulen,
H
Herbert Xu 已提交
963 964 965 966
				  sizeof(struct udphdr), &ipc, &rt,
				  msg->msg_flags);
		err = PTR_ERR(skb);
		if (skb && !IS_ERR(skb))
967
			err = udp_send_skb(skb, fl4);
H
Herbert Xu 已提交
968 969 970
		goto out;
	}

971 972 973 974 975 976 977 978 979 980 981 982 983
	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);

		LIMIT_NETDEBUG(KERN_DEBUG "udp cork app bug 2\n");
		err = -EINVAL;
		goto out;
	}
	/*
	 *	Now cork the socket to pend data.
	 */
D
David S. Miller 已提交
984 985 986
	fl4 = &inet->cork.fl.u.ip4;
	fl4->daddr = daddr;
	fl4->saddr = saddr;
987 988
	fl4->fl4_dport = dport;
	fl4->fl4_sport = inet->inet_sport;
989 990 991 992
	up->pending = AF_INET;

do_append_data:
	up->len += ulen;
993 994 995
	err = ip_append_data(sk, fl4, getfrag, msg->msg_iov, ulen,
			     sizeof(struct udphdr), &ipc, &rt,
			     corkreq ? msg->msg_flags|MSG_MORE : msg->msg_flags);
996 997 998 999 1000 1001 1002 1003 1004 1005 1006 1007 1008 1009 1010 1011 1012 1013 1014 1015 1016 1017
	if (err)
		udp_flush_pending_frames(sk);
	else if (!corkreq)
		err = udp_push_pending_frames(sk);
	else if (unlikely(skb_queue_empty(&sk->sk_write_queue)))
		up->pending = 0;
	release_sock(sk);

out:
	ip_rt_put(rt);
	if (free)
		kfree(ipc.opt);
	if (!err)
		return len;
	/*
	 * 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)) {
1018 1019
		UDP_INC_STATS_USER(sock_net(sk),
				UDP_MIB_SNDBUFERRORS, is_udplite);
1020 1021 1022 1023
	}
	return err;

do_confirm:
1024
	dst_confirm(&rt->dst);
1025 1026 1027 1028 1029
	if (!(msg->msg_flags&MSG_PROBE) || len)
		goto back_from_confirm;
	err = 0;
	goto out;
}
E
Eric Dumazet 已提交
1030
EXPORT_SYMBOL(udp_sendmsg);
1031 1032 1033 1034

int udp_sendpage(struct sock *sk, struct page *page, int offset,
		 size_t size, int flags)
{
1035
	struct inet_sock *inet = inet_sk(sk);
1036 1037 1038 1039 1040 1041 1042 1043 1044 1045 1046 1047 1048 1049 1050 1051 1052 1053 1054 1055 1056 1057 1058 1059
	struct udp_sock *up = udp_sk(sk);
	int ret;

	if (!up->pending) {
		struct msghdr msg = {	.msg_flags = flags|MSG_MORE };

		/* Call udp_sendmsg to specify destination address which
		 * sendpage interface can't pass.
		 * This will succeed only when the socket is connected.
		 */
		ret = udp_sendmsg(NULL, sk, &msg, 0);
		if (ret < 0)
			return ret;
	}

	lock_sock(sk);

	if (unlikely(!up->pending)) {
		release_sock(sk);

		LIMIT_NETDEBUG(KERN_DEBUG "udp cork app bug 3\n");
		return -EINVAL;
	}

1060 1061
	ret = ip_append_page(sk, &inet->cork.fl.u.ip4,
			     page, offset, size, flags);
1062 1063 1064 1065 1066 1067 1068 1069 1070 1071 1072 1073 1074 1075 1076 1077 1078 1079 1080 1081
	if (ret == -EOPNOTSUPP) {
		release_sock(sk);
		return sock_no_sendpage(sk->sk_socket, page, offset,
					size, flags);
	}
	if (ret < 0) {
		udp_flush_pending_frames(sk);
		goto out;
	}

	up->len += size;
	if (!(up->corkflag || (flags&MSG_MORE)))
		ret = udp_push_pending_frames(sk);
	if (!ret)
		ret = size;
out:
	release_sock(sk);
	return ret;
}

E
Eric Dumazet 已提交
1082 1083 1084 1085 1086 1087 1088 1089 1090 1091 1092 1093 1094 1095 1096 1097 1098 1099 1100 1101 1102

/**
 *	first_packet_length	- return length of first packet in receive queue
 *	@sk: socket
 *
 *	Drops all bad checksum frames, until a valid one is found.
 *	Returns the length of found skb, or 0 if none is found.
 */
static unsigned int first_packet_length(struct sock *sk)
{
	struct sk_buff_head list_kill, *rcvq = &sk->sk_receive_queue;
	struct sk_buff *skb;
	unsigned int res;

	__skb_queue_head_init(&list_kill);

	spin_lock_bh(&rcvq->lock);
	while ((skb = skb_peek(rcvq)) != NULL &&
		udp_lib_checksum_complete(skb)) {
		UDP_INC_STATS_BH(sock_net(sk), UDP_MIB_INERRORS,
				 IS_UDPLITE(sk));
E
Eric Dumazet 已提交
1103
		atomic_inc(&sk->sk_drops);
E
Eric Dumazet 已提交
1104 1105 1106 1107 1108 1109 1110
		__skb_unlink(skb, rcvq);
		__skb_queue_tail(&list_kill, skb);
	}
	res = skb ? skb->len : 0;
	spin_unlock_bh(&rcvq->lock);

	if (!skb_queue_empty(&list_kill)) {
1111 1112
		bool slow = lock_sock_fast(sk);

E
Eric Dumazet 已提交
1113 1114
		__skb_queue_purge(&list_kill);
		sk_mem_reclaim_partial(sk);
1115
		unlock_sock_fast(sk, slow);
E
Eric Dumazet 已提交
1116 1117 1118 1119
	}
	return res;
}

L
Linus Torvalds 已提交
1120 1121 1122
/*
 *	IOCTL requests applicable to the UDP protocol
 */
1123

L
Linus Torvalds 已提交
1124 1125
int udp_ioctl(struct sock *sk, int cmd, unsigned long arg)
{
1126 1127
	switch (cmd) {
	case SIOCOUTQ:
L
Linus Torvalds 已提交
1128
	{
1129 1130
		int amount = sk_wmem_alloc_get(sk);

1131 1132
		return put_user(amount, (int __user *)arg);
	}
L
Linus Torvalds 已提交
1133

1134 1135
	case SIOCINQ:
	{
E
Eric Dumazet 已提交
1136
		unsigned int amount = first_packet_length(sk);
1137

E
Eric Dumazet 已提交
1138
		if (amount)
1139 1140 1141 1142 1143
			/*
			 * We will only return the amount
			 * of this packet since that is all
			 * that will be read.
			 */
E
Eric Dumazet 已提交
1144 1145
			amount -= sizeof(struct udphdr);

1146 1147
		return put_user(amount, (int __user *)arg);
	}
L
Linus Torvalds 已提交
1148

1149 1150
	default:
		return -ENOIOCTLCMD;
L
Linus Torvalds 已提交
1151
	}
1152 1153

	return 0;
L
Linus Torvalds 已提交
1154
}
E
Eric Dumazet 已提交
1155
EXPORT_SYMBOL(udp_ioctl);
L
Linus Torvalds 已提交
1156

1157 1158 1159 1160 1161 1162 1163 1164 1165 1166 1167
/*
 * 	This should be easy, if there is something there we
 * 	return it, otherwise we block.
 */

int udp_recvmsg(struct kiocb *iocb, struct sock *sk, struct msghdr *msg,
		size_t len, int noblock, int flags, int *addr_len)
{
	struct inet_sock *inet = inet_sk(sk);
	struct sockaddr_in *sin = (struct sockaddr_in *)msg->msg_name;
	struct sk_buff *skb;
1168
	unsigned int ulen, copied;
1169 1170 1171
	int peeked;
	int err;
	int is_udplite = IS_UDPLITE(sk);
1172
	bool slow;
1173 1174 1175 1176 1177

	/*
	 *	Check any passed addresses
	 */
	if (addr_len)
E
Eric Dumazet 已提交
1178
		*addr_len = sizeof(*sin);
1179 1180 1181 1182 1183 1184 1185 1186 1187 1188 1189

	if (flags & MSG_ERRQUEUE)
		return ip_recv_error(sk, msg, len);

try_again:
	skb = __skb_recv_datagram(sk, flags | (noblock ? MSG_DONTWAIT : 0),
				  &peeked, &err);
	if (!skb)
		goto out;

	ulen = skb->len - sizeof(struct udphdr);
1190 1191 1192 1193
	copied = len;
	if (copied > ulen)
		copied = ulen;
	else if (copied < ulen)
1194 1195 1196 1197 1198 1199 1200 1201
		msg->msg_flags |= MSG_TRUNC;

	/*
	 * 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.
	 */

1202
	if (copied < ulen || UDP_SKB_CB(skb)->partial_cov) {
1203 1204 1205 1206 1207 1208
		if (udp_lib_checksum_complete(skb))
			goto csum_copy_err;
	}

	if (skb_csum_unnecessary(skb))
		err = skb_copy_datagram_iovec(skb, sizeof(struct udphdr),
1209
					      msg->msg_iov, copied);
1210
	else {
E
Eric Dumazet 已提交
1211 1212 1213
		err = skb_copy_and_csum_datagram_iovec(skb,
						       sizeof(struct udphdr),
						       msg->msg_iov);
1214 1215 1216 1217 1218 1219 1220 1221 1222

		if (err == -EINVAL)
			goto csum_copy_err;
	}

	if (err)
		goto out_free;

	if (!peeked)
1223 1224
		UDP_INC_STATS_USER(sock_net(sk),
				UDP_MIB_INDATAGRAMS, is_udplite);
1225

1226
	sock_recv_ts_and_drops(msg, sk, skb);
1227 1228

	/* Copy the address. */
E
Eric Dumazet 已提交
1229
	if (sin) {
1230 1231 1232 1233 1234 1235 1236 1237
		sin->sin_family = AF_INET;
		sin->sin_port = udp_hdr(skb)->source;
		sin->sin_addr.s_addr = ip_hdr(skb)->saddr;
		memset(sin->sin_zero, 0, sizeof(sin->sin_zero));
	}
	if (inet->cmsg_flags)
		ip_cmsg_recv(msg, skb);

1238
	err = copied;
1239 1240 1241 1242
	if (flags & MSG_TRUNC)
		err = ulen;

out_free:
1243
	skb_free_datagram_locked(sk, skb);
1244 1245 1246 1247
out:
	return err;

csum_copy_err:
1248
	slow = lock_sock_fast(sk);
1249
	if (!skb_kill_datagram(sk, skb, flags))
1250
		UDP_INC_STATS_USER(sock_net(sk), UDP_MIB_INERRORS, is_udplite);
1251
	unlock_sock_fast(sk, slow);
1252 1253 1254

	if (noblock)
		return -EAGAIN;
1255 1256 1257

	/* starting over for a new packet */
	msg->msg_flags &= ~MSG_TRUNC;
1258 1259 1260 1261
	goto try_again;
}


L
Linus Torvalds 已提交
1262 1263 1264 1265 1266 1267
int udp_disconnect(struct sock *sk, int flags)
{
	struct inet_sock *inet = inet_sk(sk);
	/*
	 *	1003.1g - break association.
	 */
1268

L
Linus Torvalds 已提交
1269
	sk->sk_state = TCP_CLOSE;
E
Eric Dumazet 已提交
1270 1271
	inet->inet_daddr = 0;
	inet->inet_dport = 0;
1272
	sock_rps_reset_rxhash(sk);
L
Linus Torvalds 已提交
1273 1274 1275 1276 1277 1278
	sk->sk_bound_dev_if = 0;
	if (!(sk->sk_userlocks & SOCK_BINDADDR_LOCK))
		inet_reset_saddr(sk);

	if (!(sk->sk_userlocks & SOCK_BINDPORT_LOCK)) {
		sk->sk_prot->unhash(sk);
E
Eric Dumazet 已提交
1279
		inet->inet_sport = 0;
L
Linus Torvalds 已提交
1280 1281 1282 1283
	}
	sk_dst_reset(sk);
	return 0;
}
E
Eric Dumazet 已提交
1284
EXPORT_SYMBOL(udp_disconnect);
L
Linus Torvalds 已提交
1285

1286 1287
void udp_lib_unhash(struct sock *sk)
{
1288 1289
	if (sk_hashed(sk)) {
		struct udp_table *udptable = sk->sk_prot->h.udp_table;
1290 1291 1292 1293 1294
		struct udp_hslot *hslot, *hslot2;

		hslot  = udp_hashslot(udptable, sock_net(sk),
				      udp_sk(sk)->udp_port_hash);
		hslot2 = udp_hashslot2(udptable, udp_sk(sk)->udp_portaddr_hash);
1295

1296 1297
		spin_lock_bh(&hslot->lock);
		if (sk_nulls_del_node_init_rcu(sk)) {
E
Eric Dumazet 已提交
1298
			hslot->count--;
E
Eric Dumazet 已提交
1299
			inet_sk(sk)->inet_num = 0;
1300
			sock_prot_inuse_add(sock_net(sk), sk->sk_prot, -1);
1301 1302 1303 1304 1305

			spin_lock(&hslot2->lock);
			hlist_nulls_del_init_rcu(&udp_sk(sk)->udp_portaddr_node);
			hslot2->count--;
			spin_unlock(&hslot2->lock);
1306 1307
		}
		spin_unlock_bh(&hslot->lock);
1308 1309 1310 1311
	}
}
EXPORT_SYMBOL(udp_lib_unhash);

E
Eric Dumazet 已提交
1312 1313 1314 1315 1316 1317 1318 1319 1320 1321 1322 1323 1324 1325 1326 1327 1328 1329 1330 1331 1332 1333 1334 1335 1336 1337 1338 1339 1340 1341 1342 1343 1344 1345 1346 1347 1348 1349 1350 1351 1352 1353 1354
/*
 * inet_rcv_saddr was changed, we must rehash secondary hash
 */
void udp_lib_rehash(struct sock *sk, u16 newhash)
{
	if (sk_hashed(sk)) {
		struct udp_table *udptable = sk->sk_prot->h.udp_table;
		struct udp_hslot *hslot, *hslot2, *nhslot2;

		hslot2 = udp_hashslot2(udptable, udp_sk(sk)->udp_portaddr_hash);
		nhslot2 = udp_hashslot2(udptable, newhash);
		udp_sk(sk)->udp_portaddr_hash = newhash;
		if (hslot2 != nhslot2) {
			hslot = udp_hashslot(udptable, sock_net(sk),
					     udp_sk(sk)->udp_port_hash);
			/* we must lock primary chain too */
			spin_lock_bh(&hslot->lock);

			spin_lock(&hslot2->lock);
			hlist_nulls_del_init_rcu(&udp_sk(sk)->udp_portaddr_node);
			hslot2->count--;
			spin_unlock(&hslot2->lock);

			spin_lock(&nhslot2->lock);
			hlist_nulls_add_head_rcu(&udp_sk(sk)->udp_portaddr_node,
						 &nhslot2->head);
			nhslot2->count++;
			spin_unlock(&nhslot2->lock);

			spin_unlock_bh(&hslot->lock);
		}
	}
}
EXPORT_SYMBOL(udp_lib_rehash);

static void udp_v4_rehash(struct sock *sk)
{
	u16 new_hash = udp4_portaddr_hash(sock_net(sk),
					  inet_sk(sk)->inet_rcv_saddr,
					  inet_sk(sk)->inet_num);
	udp_lib_rehash(sk, new_hash);
}

H
Herbert Xu 已提交
1355 1356
static int __udp_queue_rcv_skb(struct sock *sk, struct sk_buff *skb)
{
T
Tom Herbert 已提交
1357
	int rc;
E
Eric Dumazet 已提交
1358

T
Tom Herbert 已提交
1359
	if (inet_sk(sk)->inet_daddr)
1360
		sock_rps_save_rxhash(sk, skb);
T
Tom Herbert 已提交
1361

1362
	rc = sock_queue_rcv_skb(sk, skb);
E
Eric Dumazet 已提交
1363 1364
	if (rc < 0) {
		int is_udplite = IS_UDPLITE(sk);
H
Herbert Xu 已提交
1365 1366

		/* Note that an ENOMEM error is charged twice */
E
Eric Dumazet 已提交
1367
		if (rc == -ENOMEM)
H
Herbert Xu 已提交
1368 1369
			UDP_INC_STATS_BH(sock_net(sk), UDP_MIB_RCVBUFERRORS,
					 is_udplite);
E
Eric Dumazet 已提交
1370 1371
		UDP_INC_STATS_BH(sock_net(sk), UDP_MIB_INERRORS, is_udplite);
		kfree_skb(skb);
1372
		trace_udp_fail_queue_rcv_skb(rc, sk);
E
Eric Dumazet 已提交
1373
		return -1;
H
Herbert Xu 已提交
1374 1375 1376 1377 1378 1379
	}

	return 0;

}

1380 1381 1382 1383 1384 1385 1386 1387
/* returns:
 *  -1: error
 *   0: success
 *  >0: "udp encap" protocol resubmission
 *
 * Note that in the success and error cases, the skb is assumed to
 * have either been requeued or freed.
 */
E
Eric Dumazet 已提交
1388
int udp_queue_rcv_skb(struct sock *sk, struct sk_buff *skb)
1389 1390 1391 1392 1393 1394 1395 1396 1397 1398 1399 1400 1401
{
	struct udp_sock *up = udp_sk(sk);
	int rc;
	int is_udplite = IS_UDPLITE(sk);

	/*
	 *	Charge it to the socket, dropping if the queue is full.
	 */
	if (!xfrm4_policy_check(sk, XFRM_POLICY_IN, skb))
		goto drop;
	nf_reset(skb);

	if (up->encap_type) {
1402 1403
		int (*encap_rcv)(struct sock *sk, struct sk_buff *skb);

1404 1405 1406 1407 1408 1409 1410 1411 1412 1413 1414 1415
		/*
		 * 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 */
1416 1417
		encap_rcv = ACCESS_ONCE(up->encap_rcv);
		if (skb->len > sizeof(struct udphdr) && encap_rcv != NULL) {
1418 1419
			int ret;

1420
			ret = encap_rcv(sk, skb);
1421
			if (ret <= 0) {
1422 1423
				UDP_INC_STATS_BH(sock_net(sk),
						 UDP_MIB_INDATAGRAMS,
1424 1425 1426 1427 1428 1429 1430 1431 1432 1433 1434 1435 1436 1437 1438 1439 1440 1441 1442 1443 1444 1445 1446 1447 1448 1449 1450 1451 1452 1453 1454 1455 1456 1457 1458 1459 1460 1461 1462 1463 1464 1465 1466 1467
						 is_udplite);
				return -ret;
			}
		}

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

	/*
	 * 	UDP-Lite specific tests, ignored on UDP sockets
	 */
	if ((is_udplite & UDPLITE_RECV_CC)  &&  UDP_SKB_CB(skb)->partial_cov) {

		/*
		 * MIB statistics other than incrementing the error count are
		 * disabled for the following two types of errors: these depend
		 * on the application settings, not on the functioning of the
		 * protocol stack as such.
		 *
		 * RFC 3828 here recommends (sec 3.3): "There should also be a
		 * way ... to ... at least let the receiving application block
		 * delivery of packets with coverage values less than a value
		 * provided by the application."
		 */
		if (up->pcrlen == 0) {          /* full coverage was set  */
			LIMIT_NETDEBUG(KERN_WARNING "UDPLITE: partial coverage "
				"%d while full coverage %d requested\n",
				UDP_SKB_CB(skb)->cscov, skb->len);
			goto drop;
		}
		/* The next case involves violating the min. coverage requested
		 * by the receiver. This is subtle: if receiver wants x and x is
		 * greater than the buffersize/MTU then receiver will complain
		 * that it wants x while sender emits packets of smaller size y.
		 * Therefore the above ...()->partial_cov statement is essential.
		 */
		if (UDP_SKB_CB(skb)->cscov  <  up->pcrlen) {
			LIMIT_NETDEBUG(KERN_WARNING
				"UDPLITE: coverage %d too small, need min %d\n",
				UDP_SKB_CB(skb)->cscov, up->pcrlen);
			goto drop;
		}
	}

1468 1469 1470
	if (rcu_access_pointer(sk->sk_filter) &&
	    udp_lib_checksum_complete(skb))
		goto drop;
1471

1472 1473 1474 1475

	if (sk_rcvqueues_full(sk, skb))
		goto drop;

H
Herbert Xu 已提交
1476
	rc = 0;
1477

1478
	ipv4_pktinfo_prepare(skb);
H
Herbert Xu 已提交
1479 1480 1481
	bh_lock_sock(sk);
	if (!sock_owned_by_user(sk))
		rc = __udp_queue_rcv_skb(sk, skb);
Z
Zhu Yi 已提交
1482
	else if (sk_add_backlog(sk, skb)) {
Z
Zhu Yi 已提交
1483 1484 1485
		bh_unlock_sock(sk);
		goto drop;
	}
H
Herbert Xu 已提交
1486 1487 1488
	bh_unlock_sock(sk);

	return rc;
1489 1490

drop:
1491
	UDP_INC_STATS_BH(sock_net(sk), UDP_MIB_INERRORS, is_udplite);
E
Eric Dumazet 已提交
1492
	atomic_inc(&sk->sk_drops);
1493 1494 1495 1496
	kfree_skb(skb);
	return -1;
}

1497 1498 1499 1500 1501 1502

static void flush_stack(struct sock **stack, unsigned int count,
			struct sk_buff *skb, unsigned int final)
{
	unsigned int i;
	struct sk_buff *skb1 = NULL;
1503
	struct sock *sk;
1504 1505

	for (i = 0; i < count; i++) {
1506
		sk = stack[i];
1507 1508 1509
		if (likely(skb1 == NULL))
			skb1 = (i == final) ? skb : skb_clone(skb, GFP_ATOMIC);

1510 1511 1512 1513 1514 1515 1516 1517 1518
		if (!skb1) {
			atomic_inc(&sk->sk_drops);
			UDP_INC_STATS_BH(sock_net(sk), UDP_MIB_RCVBUFERRORS,
					 IS_UDPLITE(sk));
			UDP_INC_STATS_BH(sock_net(sk), UDP_MIB_INERRORS,
					 IS_UDPLITE(sk));
		}

		if (skb1 && udp_queue_rcv_skb(sk, skb1) <= 0)
1519 1520 1521 1522 1523 1524
			skb1 = NULL;
	}
	if (unlikely(skb1))
		kfree_skb(skb1);
}

1525 1526 1527
/*
 *	Multicasts and broadcasts go to each listener.
 *
1528
 *	Note: called only from the BH handler context.
1529
 */
1530
static int __udp4_lib_mcast_deliver(struct net *net, struct sk_buff *skb,
1531 1532
				    struct udphdr  *uh,
				    __be32 saddr, __be32 daddr,
1533
				    struct udp_table *udptable)
1534
{
1535
	struct sock *sk, *stack[256 / sizeof(struct sock *)];
1536
	struct udp_hslot *hslot = udp_hashslot(udptable, net, ntohs(uh->dest));
1537
	int dif;
1538
	unsigned int i, count = 0;
1539

1540
	spin_lock(&hslot->lock);
1541
	sk = sk_nulls_head(&hslot->head);
1542
	dif = skb->dev->ifindex;
1543
	sk = udp_v4_mcast_next(net, sk, uh->dest, daddr, uh->source, saddr, dif);
1544 1545 1546 1547 1548 1549 1550 1551 1552 1553 1554 1555 1556 1557 1558 1559 1560
	while (sk) {
		stack[count++] = sk;
		sk = udp_v4_mcast_next(net, sk_nulls_next(sk), uh->dest,
				       daddr, uh->source, saddr, dif);
		if (unlikely(count == ARRAY_SIZE(stack))) {
			if (!sk)
				break;
			flush_stack(stack, count, skb, ~0);
			count = 0;
		}
	}
	/*
	 * before releasing chain lock, we must take a reference on sockets
	 */
	for (i = 0; i < count; i++)
		sock_hold(stack[i]);

1561
	spin_unlock(&hslot->lock);
1562 1563 1564 1565 1566 1567 1568 1569 1570 1571 1572 1573

	/*
	 * do the slow work with no lock held
	 */
	if (count) {
		flush_stack(stack, count, skb, count - 1);

		for (i = 0; i < count; i++)
			sock_put(stack[i]);
	} else {
		kfree_skb(skb);
	}
1574 1575 1576 1577 1578 1579 1580 1581 1582 1583 1584 1585 1586 1587 1588 1589 1590 1591 1592 1593 1594 1595 1596 1597 1598 1599 1600
	return 0;
}

/* Initialize UDP checksum. If exited with zero value (success),
 * CHECKSUM_UNNECESSARY means, that no more checks are required.
 * Otherwise, csum completion requires chacksumming packet body,
 * including udp header and folding it to skb->csum.
 */
static inline int udp4_csum_init(struct sk_buff *skb, struct udphdr *uh,
				 int proto)
{
	const struct iphdr *iph;
	int err;

	UDP_SKB_CB(skb)->partial_cov = 0;
	UDP_SKB_CB(skb)->cscov = skb->len;

	if (proto == IPPROTO_UDPLITE) {
		err = udplite_checksum_init(skb, uh);
		if (err)
			return err;
	}

	iph = ip_hdr(skb);
	if (uh->check == 0) {
		skb->ip_summed = CHECKSUM_UNNECESSARY;
	} else if (skb->ip_summed == CHECKSUM_COMPLETE) {
E
Eric Dumazet 已提交
1601
		if (!csum_tcpudp_magic(iph->saddr, iph->daddr, skb->len,
1602 1603 1604 1605 1606 1607 1608 1609 1610 1611 1612 1613 1614 1615 1616 1617 1618
				      proto, skb->csum))
			skb->ip_summed = CHECKSUM_UNNECESSARY;
	}
	if (!skb_csum_unnecessary(skb))
		skb->csum = csum_tcpudp_nofold(iph->saddr, iph->daddr,
					       skb->len, proto, 0);
	/* Probably, we should checksum udp header (it should be in cache
	 * in any case) and data in tiny packets (< rx copybreak).
	 */

	return 0;
}

/*
 *	All we need to do is get the socket, and then do a checksum.
 */

1619
int __udp4_lib_rcv(struct sk_buff *skb, struct udp_table *udptable,
1620 1621 1622
		   int proto)
{
	struct sock *sk;
1623
	struct udphdr *uh;
1624
	unsigned short ulen;
E
Eric Dumazet 已提交
1625
	struct rtable *rt = skb_rtable(skb);
1626
	__be32 saddr, daddr;
1627
	struct net *net = dev_net(skb->dev);
1628 1629 1630 1631 1632 1633 1634

	/*
	 *  Validate the packet.
	 */
	if (!pskb_may_pull(skb, sizeof(struct udphdr)))
		goto drop;		/* No space for header. */

1635
	uh   = udp_hdr(skb);
1636
	ulen = ntohs(uh->len);
1637 1638 1639
	saddr = ip_hdr(skb)->saddr;
	daddr = ip_hdr(skb)->daddr;

1640 1641 1642 1643 1644 1645 1646 1647 1648 1649 1650 1651 1652 1653
	if (ulen > skb->len)
		goto short_packet;

	if (proto == IPPROTO_UDP) {
		/* UDP validates ulen. */
		if (ulen < sizeof(*uh) || pskb_trim_rcsum(skb, ulen))
			goto short_packet;
		uh = udp_hdr(skb);
	}

	if (udp4_csum_init(skb, uh, proto))
		goto csum_error;

	if (rt->rt_flags & (RTCF_BROADCAST|RTCF_MULTICAST))
1654 1655
		return __udp4_lib_mcast_deliver(net, skb, uh,
				saddr, daddr, udptable);
1656

1657
	sk = __udp4_lib_lookup_skb(skb, uh->source, uh->dest, udptable);
1658 1659

	if (sk != NULL) {
H
Herbert Xu 已提交
1660
		int ret = udp_queue_rcv_skb(sk, skb);
1661 1662 1663 1664 1665 1666 1667 1668 1669 1670 1671 1672 1673 1674 1675 1676 1677 1678
		sock_put(sk);

		/* a return value > 0 means to resubmit the input, but
		 * it wants the return to be -protocol, or 0
		 */
		if (ret > 0)
			return -ret;
		return 0;
	}

	if (!xfrm4_policy_check(NULL, XFRM_POLICY_IN, skb))
		goto drop;
	nf_reset(skb);

	/* No socket. Drop packet silently, if checksum is wrong */
	if (udp_lib_checksum_complete(skb))
		goto csum_error;

1679
	UDP_INC_STATS_BH(net, UDP_MIB_NOPORTS, proto == IPPROTO_UDPLITE);
1680 1681 1682 1683 1684 1685 1686 1687 1688 1689
	icmp_send(skb, ICMP_DEST_UNREACH, ICMP_PORT_UNREACH, 0);

	/*
	 * Hmm.  We got an UDP packet to a port to which we
	 * don't wanna listen.  Ignore it.
	 */
	kfree_skb(skb);
	return 0;

short_packet:
1690
	LIMIT_NETDEBUG(KERN_DEBUG "UDP%s: short packet: From %pI4:%u %d/%d to %pI4:%u\n",
1691
		       proto == IPPROTO_UDPLITE ? "-Lite" : "",
1692
		       &saddr,
1693 1694 1695
		       ntohs(uh->source),
		       ulen,
		       skb->len,
1696
		       &daddr,
1697 1698 1699 1700 1701 1702 1703 1704
		       ntohs(uh->dest));
	goto drop;

csum_error:
	/*
	 * RFC1122: OK.  Discards the bad packet silently (as far as
	 * the network is concerned, anyway) as per 4.1.3.4 (MUST).
	 */
1705
	LIMIT_NETDEBUG(KERN_DEBUG "UDP%s: bad checksum. From %pI4:%u to %pI4:%u ulen %d\n",
1706
		       proto == IPPROTO_UDPLITE ? "-Lite" : "",
1707
		       &saddr,
1708
		       ntohs(uh->source),
1709
		       &daddr,
1710 1711 1712
		       ntohs(uh->dest),
		       ulen);
drop:
1713
	UDP_INC_STATS_BH(net, UDP_MIB_INERRORS, proto == IPPROTO_UDPLITE);
1714 1715 1716 1717 1718 1719
	kfree_skb(skb);
	return 0;
}

int udp_rcv(struct sk_buff *skb)
{
1720
	return __udp4_lib_rcv(skb, &udp_table, IPPROTO_UDP);
1721 1722
}

1723
void udp_destroy_sock(struct sock *sk)
1724
{
1725
	bool slow = lock_sock_fast(sk);
1726
	udp_flush_pending_frames(sk);
1727
	unlock_sock_fast(sk, slow);
1728 1729
}

L
Linus Torvalds 已提交
1730 1731 1732
/*
 *	Socket option code for UDP
 */
1733
int udp_lib_setsockopt(struct sock *sk, int level, int optname,
1734
		       char __user *optval, unsigned int optlen,
1735
		       int (*push_pending_frames)(struct sock *))
L
Linus Torvalds 已提交
1736 1737 1738 1739
{
	struct udp_sock *up = udp_sk(sk);
	int val;
	int err = 0;
W
Wang Chen 已提交
1740
	int is_udplite = IS_UDPLITE(sk);
L
Linus Torvalds 已提交
1741

E
Eric Dumazet 已提交
1742
	if (optlen < sizeof(int))
L
Linus Torvalds 已提交
1743 1744 1745 1746 1747
		return -EINVAL;

	if (get_user(val, (int __user *)optval))
		return -EFAULT;

1748
	switch (optname) {
L
Linus Torvalds 已提交
1749 1750 1751 1752 1753 1754
	case UDP_CORK:
		if (val != 0) {
			up->corkflag = 1;
		} else {
			up->corkflag = 0;
			lock_sock(sk);
1755
			(*push_pending_frames)(sk);
L
Linus Torvalds 已提交
1756 1757 1758
			release_sock(sk);
		}
		break;
1759

L
Linus Torvalds 已提交
1760 1761 1762 1763 1764
	case UDP_ENCAP:
		switch (val) {
		case 0:
		case UDP_ENCAP_ESPINUDP:
		case UDP_ENCAP_ESPINUDP_NON_IKE:
1765 1766
			up->encap_rcv = xfrm4_udp_encap_rcv;
			/* FALLTHROUGH */
1767
		case UDP_ENCAP_L2TPINUDP:
L
Linus Torvalds 已提交
1768 1769 1770 1771 1772 1773 1774 1775
			up->encap_type = val;
			break;
		default:
			err = -ENOPROTOOPT;
			break;
		}
		break;

1776 1777 1778 1779 1780 1781
	/*
	 * 	UDP-Lite's partial checksum coverage (RFC 3828).
	 */
	/* The sender sets actual checksum coverage length via this option.
	 * The case coverage > packet length is handled by send module. */
	case UDPLITE_SEND_CSCOV:
W
Wang Chen 已提交
1782
		if (!is_udplite)         /* Disable the option on UDP sockets */
1783 1784 1785
			return -ENOPROTOOPT;
		if (val != 0 && val < 8) /* Illegal coverage: use default (8) */
			val = 8;
1786 1787
		else if (val > USHRT_MAX)
			val = USHRT_MAX;
1788 1789 1790 1791
		up->pcslen = val;
		up->pcflag |= UDPLITE_SEND_CC;
		break;

1792 1793
	/* The receiver specifies a minimum checksum coverage value. To make
	 * sense, this should be set to at least 8 (as done below). If zero is
1794 1795
	 * used, this again means full checksum coverage.                     */
	case UDPLITE_RECV_CSCOV:
W
Wang Chen 已提交
1796
		if (!is_udplite)         /* Disable the option on UDP sockets */
1797 1798 1799
			return -ENOPROTOOPT;
		if (val != 0 && val < 8) /* Avoid silly minimal values.       */
			val = 8;
1800 1801
		else if (val > USHRT_MAX)
			val = USHRT_MAX;
1802 1803 1804 1805
		up->pcrlen = val;
		up->pcflag |= UDPLITE_RECV_CC;
		break;

L
Linus Torvalds 已提交
1806 1807 1808
	default:
		err = -ENOPROTOOPT;
		break;
1809
	}
L
Linus Torvalds 已提交
1810 1811 1812

	return err;
}
E
Eric Dumazet 已提交
1813
EXPORT_SYMBOL(udp_lib_setsockopt);
L
Linus Torvalds 已提交
1814

1815
int udp_setsockopt(struct sock *sk, int level, int optname,
1816
		   char __user *optval, unsigned int optlen)
1817 1818 1819 1820 1821 1822 1823 1824 1825
{
	if (level == SOL_UDP  ||  level == SOL_UDPLITE)
		return udp_lib_setsockopt(sk, level, optname, optval, optlen,
					  udp_push_pending_frames);
	return ip_setsockopt(sk, level, optname, optval, optlen);
}

#ifdef CONFIG_COMPAT
int compat_udp_setsockopt(struct sock *sk, int level, int optname,
1826
			  char __user *optval, unsigned int optlen)
1827 1828 1829 1830 1831 1832 1833 1834
{
	if (level == SOL_UDP  ||  level == SOL_UDPLITE)
		return udp_lib_setsockopt(sk, level, optname, optval, optlen,
					  udp_push_pending_frames);
	return compat_ip_setsockopt(sk, level, optname, optval, optlen);
}
#endif

1835 1836
int udp_lib_getsockopt(struct sock *sk, int level, int optname,
		       char __user *optval, int __user *optlen)
L
Linus Torvalds 已提交
1837 1838 1839 1840
{
	struct udp_sock *up = udp_sk(sk);
	int val, len;

E
Eric Dumazet 已提交
1841
	if (get_user(len, optlen))
L
Linus Torvalds 已提交
1842 1843 1844
		return -EFAULT;

	len = min_t(unsigned int, len, sizeof(int));
1845

1846
	if (len < 0)
L
Linus Torvalds 已提交
1847 1848
		return -EINVAL;

1849
	switch (optname) {
L
Linus Torvalds 已提交
1850 1851 1852 1853 1854 1855 1856 1857
	case UDP_CORK:
		val = up->corkflag;
		break;

	case UDP_ENCAP:
		val = up->encap_type;
		break;

1858 1859 1860 1861 1862 1863 1864 1865 1866 1867
	/* The following two cannot be changed on UDP sockets, the return is
	 * always 0 (which corresponds to the full checksum coverage of UDP). */
	case UDPLITE_SEND_CSCOV:
		val = up->pcslen;
		break;

	case UDPLITE_RECV_CSCOV:
		val = up->pcrlen;
		break;

L
Linus Torvalds 已提交
1868 1869
	default:
		return -ENOPROTOOPT;
1870
	}
L
Linus Torvalds 已提交
1871

1872
	if (put_user(len, optlen))
1873
		return -EFAULT;
E
Eric Dumazet 已提交
1874
	if (copy_to_user(optval, &val, len))
L
Linus Torvalds 已提交
1875
		return -EFAULT;
1876
	return 0;
L
Linus Torvalds 已提交
1877
}
E
Eric Dumazet 已提交
1878
EXPORT_SYMBOL(udp_lib_getsockopt);
L
Linus Torvalds 已提交
1879

1880 1881 1882 1883 1884 1885 1886 1887 1888 1889 1890 1891 1892 1893 1894 1895 1896
int udp_getsockopt(struct sock *sk, int level, int optname,
		   char __user *optval, int __user *optlen)
{
	if (level == SOL_UDP  ||  level == SOL_UDPLITE)
		return udp_lib_getsockopt(sk, level, optname, optval, optlen);
	return ip_getsockopt(sk, level, optname, optval, optlen);
}

#ifdef CONFIG_COMPAT
int compat_udp_getsockopt(struct sock *sk, int level, int optname,
				 char __user *optval, int __user *optlen)
{
	if (level == SOL_UDP  ||  level == SOL_UDPLITE)
		return udp_lib_getsockopt(sk, level, optname, optval, optlen);
	return compat_ip_getsockopt(sk, level, optname, optval, optlen);
}
#endif
L
Linus Torvalds 已提交
1897 1898 1899 1900 1901 1902
/**
 * 	udp_poll - wait for a UDP event.
 *	@file - file struct
 *	@sock - socket
 *	@wait - poll table
 *
1903
 *	This is same as datagram poll, except for the special case of
L
Linus Torvalds 已提交
1904 1905 1906 1907 1908 1909 1910 1911 1912 1913
 *	blocking sockets. If application is using a blocking fd
 *	and a packet with checksum error is in the queue;
 *	then it could get return from select indicating data available
 *	but then block when reading it. Add special case code
 *	to work around these arguably broken applications.
 */
unsigned int udp_poll(struct file *file, struct socket *sock, poll_table *wait)
{
	unsigned int mask = datagram_poll(file, sock, wait);
	struct sock *sk = sock->sk;
1914

L
Linus Torvalds 已提交
1915
	/* Check for false positives due to checksum errors */
E
Eric Dumazet 已提交
1916 1917 1918
	if ((mask & POLLRDNORM) && !(file->f_flags & O_NONBLOCK) &&
	    !(sk->sk_shutdown & RCV_SHUTDOWN) && !first_packet_length(sk))
		mask &= ~(POLLIN | POLLRDNORM);
L
Linus Torvalds 已提交
1919 1920

	return mask;
1921

L
Linus Torvalds 已提交
1922
}
E
Eric Dumazet 已提交
1923
EXPORT_SYMBOL(udp_poll);
L
Linus Torvalds 已提交
1924

1925 1926 1927 1928 1929 1930 1931 1932 1933 1934 1935 1936 1937
struct proto udp_prot = {
	.name		   = "UDP",
	.owner		   = THIS_MODULE,
	.close		   = udp_lib_close,
	.connect	   = ip4_datagram_connect,
	.disconnect	   = udp_disconnect,
	.ioctl		   = udp_ioctl,
	.destroy	   = udp_destroy_sock,
	.setsockopt	   = udp_setsockopt,
	.getsockopt	   = udp_getsockopt,
	.sendmsg	   = udp_sendmsg,
	.recvmsg	   = udp_recvmsg,
	.sendpage	   = udp_sendpage,
H
Herbert Xu 已提交
1938
	.backlog_rcv	   = __udp_queue_rcv_skb,
1939 1940
	.hash		   = udp_lib_hash,
	.unhash		   = udp_lib_unhash,
E
Eric Dumazet 已提交
1941
	.rehash		   = udp_v4_rehash,
1942 1943 1944 1945 1946 1947
	.get_port	   = udp_v4_get_port,
	.memory_allocated  = &udp_memory_allocated,
	.sysctl_mem	   = sysctl_udp_mem,
	.sysctl_wmem	   = &sysctl_udp_wmem_min,
	.sysctl_rmem	   = &sysctl_udp_rmem_min,
	.obj_size	   = sizeof(struct udp_sock),
1948
	.slab_flags	   = SLAB_DESTROY_BY_RCU,
1949
	.h.udp_table	   = &udp_table,
1950 1951 1952 1953
#ifdef CONFIG_COMPAT
	.compat_setsockopt = compat_udp_setsockopt,
	.compat_getsockopt = compat_udp_getsockopt,
#endif
1954
	.clear_sk	   = sk_prot_clear_portaddr_nulls,
1955
};
E
Eric Dumazet 已提交
1956
EXPORT_SYMBOL(udp_prot);
L
Linus Torvalds 已提交
1957 1958 1959 1960

/* ------------------------------------------------------------------------ */
#ifdef CONFIG_PROC_FS

1961
static struct sock *udp_get_first(struct seq_file *seq, int start)
L
Linus Torvalds 已提交
1962 1963 1964
{
	struct sock *sk;
	struct udp_iter_state *state = seq->private;
1965
	struct net *net = seq_file_net(seq);
L
Linus Torvalds 已提交
1966

1967 1968
	for (state->bucket = start; state->bucket <= state->udp_table->mask;
	     ++state->bucket) {
1969
		struct hlist_nulls_node *node;
1970
		struct udp_hslot *hslot = &state->udp_table->hash[state->bucket];
1971 1972 1973 1974

		if (hlist_nulls_empty(&hslot->head))
			continue;

1975
		spin_lock_bh(&hslot->lock);
1976
		sk_nulls_for_each(sk, node, &hslot->head) {
1977
			if (!net_eq(sock_net(sk), net))
1978
				continue;
L
Linus Torvalds 已提交
1979 1980 1981
			if (sk->sk_family == state->family)
				goto found;
		}
1982
		spin_unlock_bh(&hslot->lock);
L
Linus Torvalds 已提交
1983 1984 1985 1986 1987 1988 1989 1990 1991
	}
	sk = NULL;
found:
	return sk;
}

static struct sock *udp_get_next(struct seq_file *seq, struct sock *sk)
{
	struct udp_iter_state *state = seq->private;
1992
	struct net *net = seq_file_net(seq);
L
Linus Torvalds 已提交
1993 1994

	do {
1995
		sk = sk_nulls_next(sk);
1996
	} while (sk && (!net_eq(sock_net(sk), net) || sk->sk_family != state->family));
L
Linus Torvalds 已提交
1997

1998
	if (!sk) {
1999
		if (state->bucket <= state->udp_table->mask)
2000
			spin_unlock_bh(&state->udp_table->hash[state->bucket].lock);
2001
		return udp_get_first(seq, state->bucket + 1);
L
Linus Torvalds 已提交
2002 2003 2004 2005 2006 2007
	}
	return sk;
}

static struct sock *udp_get_idx(struct seq_file *seq, loff_t pos)
{
2008
	struct sock *sk = udp_get_first(seq, 0);
L
Linus Torvalds 已提交
2009 2010

	if (sk)
2011
		while (pos && (sk = udp_get_next(seq, sk)) != NULL)
L
Linus Torvalds 已提交
2012 2013 2014 2015 2016 2017
			--pos;
	return pos ? NULL : sk;
}

static void *udp_seq_start(struct seq_file *seq, loff_t *pos)
{
2018
	struct udp_iter_state *state = seq->private;
2019
	state->bucket = MAX_UDP_PORTS;
2020

2021
	return *pos ? udp_get_idx(seq, *pos-1) : SEQ_START_TOKEN;
L
Linus Torvalds 已提交
2022 2023 2024 2025 2026 2027
}

static void *udp_seq_next(struct seq_file *seq, void *v, loff_t *pos)
{
	struct sock *sk;

2028
	if (v == SEQ_START_TOKEN)
L
Linus Torvalds 已提交
2029 2030 2031 2032 2033 2034 2035 2036 2037 2038
		sk = udp_get_idx(seq, 0);
	else
		sk = udp_get_next(seq, v);

	++*pos;
	return sk;
}

static void udp_seq_stop(struct seq_file *seq, void *v)
{
2039 2040
	struct udp_iter_state *state = seq->private;

2041
	if (state->bucket <= state->udp_table->mask)
2042
		spin_unlock_bh(&state->udp_table->hash[state->bucket].lock);
L
Linus Torvalds 已提交
2043 2044
}

2045
int udp_seq_open(struct inode *inode, struct file *file)
L
Linus Torvalds 已提交
2046 2047
{
	struct udp_seq_afinfo *afinfo = PDE(inode)->data;
2048 2049
	struct udp_iter_state *s;
	int err;
2050

2051 2052 2053 2054
	err = seq_open_net(inode, file, &afinfo->seq_ops,
			   sizeof(struct udp_iter_state));
	if (err < 0)
		return err;
2055

2056
	s = ((struct seq_file *)file->private_data)->private;
L
Linus Torvalds 已提交
2057
	s->family		= afinfo->family;
2058
	s->udp_table		= afinfo->udp_table;
2059
	return err;
2060
}
2061
EXPORT_SYMBOL(udp_seq_open);
2062

L
Linus Torvalds 已提交
2063
/* ------------------------------------------------------------------------ */
2064
int udp_proc_register(struct net *net, struct udp_seq_afinfo *afinfo)
L
Linus Torvalds 已提交
2065 2066 2067 2068
{
	struct proc_dir_entry *p;
	int rc = 0;

2069 2070 2071 2072
	afinfo->seq_ops.start		= udp_seq_start;
	afinfo->seq_ops.next		= udp_seq_next;
	afinfo->seq_ops.stop		= udp_seq_stop;

2073
	p = proc_create_data(afinfo->name, S_IRUGO, net->proc_net,
2074
			     afinfo->seq_fops, afinfo);
2075
	if (!p)
L
Linus Torvalds 已提交
2076 2077 2078
		rc = -ENOMEM;
	return rc;
}
E
Eric Dumazet 已提交
2079
EXPORT_SYMBOL(udp_proc_register);
L
Linus Torvalds 已提交
2080

2081
void udp_proc_unregister(struct net *net, struct udp_seq_afinfo *afinfo)
L
Linus Torvalds 已提交
2082
{
2083
	proc_net_remove(net, afinfo->name);
L
Linus Torvalds 已提交
2084
}
E
Eric Dumazet 已提交
2085
EXPORT_SYMBOL(udp_proc_unregister);
2086 2087

/* ------------------------------------------------------------------------ */
2088 2089
static void udp4_format_sock(struct sock *sp, struct seq_file *f,
		int bucket, int *len)
2090 2091
{
	struct inet_sock *inet = inet_sk(sp);
E
Eric Dumazet 已提交
2092 2093 2094 2095
	__be32 dest = inet->inet_daddr;
	__be32 src  = inet->inet_rcv_saddr;
	__u16 destp	  = ntohs(inet->inet_dport);
	__u16 srcp	  = ntohs(inet->inet_sport);
2096

2097
	seq_printf(f, "%5d: %08X:%04X %08X:%04X"
D
Dan Rosenberg 已提交
2098
		" %02X %08X:%08X %02X:%08lX %08X %5d %8d %lu %d %pK %d%n",
2099
		bucket, src, srcp, dest, destp, sp->sk_state,
2100 2101
		sk_wmem_alloc_get(sp),
		sk_rmem_alloc_get(sp),
2102
		0, 0L, 0, sock_i_uid(sp), 0, sock_i_ino(sp),
E
Eric Dumazet 已提交
2103 2104
		atomic_read(&sp->sk_refcnt), sp,
		atomic_read(&sp->sk_drops), len);
2105 2106 2107 2108 2109 2110 2111 2112
}

int udp4_seq_show(struct seq_file *seq, void *v)
{
	if (v == SEQ_START_TOKEN)
		seq_printf(seq, "%-127s\n",
			   "  sl  local_address rem_address   st tx_queue "
			   "rx_queue tr tm->when retrnsmt   uid  timeout "
E
Eric Dumazet 已提交
2113
			   "inode ref pointer drops");
2114 2115
	else {
		struct udp_iter_state *state = seq->private;
2116
		int len;
2117

2118
		udp4_format_sock(v, seq, state->bucket, &len);
E
Eric Dumazet 已提交
2119
		seq_printf(seq, "%*s\n", 127 - len, "");
2120 2121 2122 2123
	}
	return 0;
}

2124 2125 2126 2127 2128 2129 2130 2131
static const struct file_operations udp_afinfo_seq_fops = {
	.owner    = THIS_MODULE,
	.open     = udp_seq_open,
	.read     = seq_read,
	.llseek   = seq_lseek,
	.release  = seq_release_net
};

2132 2133 2134 2135
/* ------------------------------------------------------------------------ */
static struct udp_seq_afinfo udp4_seq_afinfo = {
	.name		= "udp",
	.family		= AF_INET,
2136
	.udp_table	= &udp_table,
2137
	.seq_fops	= &udp_afinfo_seq_fops,
2138 2139 2140
	.seq_ops	= {
		.show		= udp4_seq_show,
	},
2141 2142
};

2143
static int __net_init udp4_proc_init_net(struct net *net)
2144 2145 2146 2147
{
	return udp_proc_register(net, &udp4_seq_afinfo);
}

2148
static void __net_exit udp4_proc_exit_net(struct net *net)
2149 2150 2151 2152 2153 2154 2155 2156 2157
{
	udp_proc_unregister(net, &udp4_seq_afinfo);
}

static struct pernet_operations udp4_net_ops = {
	.init = udp4_proc_init_net,
	.exit = udp4_proc_exit_net,
};

2158 2159
int __init udp4_proc_init(void)
{
2160
	return register_pernet_subsys(&udp4_net_ops);
2161 2162 2163 2164
}

void udp4_proc_exit(void)
{
2165
	unregister_pernet_subsys(&udp4_net_ops);
2166
}
L
Linus Torvalds 已提交
2167 2168
#endif /* CONFIG_PROC_FS */

2169 2170
static __initdata unsigned long uhash_entries;
static int __init set_uhash_entries(char *str)
2171
{
2172 2173 2174 2175 2176 2177 2178 2179
	if (!str)
		return 0;
	uhash_entries = simple_strtoul(str, &str, 0);
	if (uhash_entries && uhash_entries < UDP_HTABLE_SIZE_MIN)
		uhash_entries = UDP_HTABLE_SIZE_MIN;
	return 1;
}
__setup("uhash_entries=", set_uhash_entries);
2180

2181 2182 2183 2184 2185 2186
void __init udp_table_init(struct udp_table *table, const char *name)
{
	unsigned int i;

	if (!CONFIG_BASE_SMALL)
		table->hash = alloc_large_system_hash(name,
2187
			2 * sizeof(struct udp_hslot),
2188 2189 2190 2191 2192 2193 2194 2195 2196 2197 2198
			uhash_entries,
			21, /* one slot per 2 MB */
			0,
			&table->log,
			&table->mask,
			64 * 1024);
	/*
	 * Make sure hash table has the minimum size
	 */
	if (CONFIG_BASE_SMALL || table->mask < UDP_HTABLE_SIZE_MIN - 1) {
		table->hash = kmalloc(UDP_HTABLE_SIZE_MIN *
2199
				      2 * sizeof(struct udp_hslot), GFP_KERNEL);
2200 2201 2202 2203 2204
		if (!table->hash)
			panic(name);
		table->log = ilog2(UDP_HTABLE_SIZE_MIN);
		table->mask = UDP_HTABLE_SIZE_MIN - 1;
	}
2205
	table->hash2 = table->hash + (table->mask + 1);
2206
	for (i = 0; i <= table->mask; i++) {
2207
		INIT_HLIST_NULLS_HEAD(&table->hash[i].head, i);
E
Eric Dumazet 已提交
2208
		table->hash[i].count = 0;
2209 2210
		spin_lock_init(&table->hash[i].lock);
	}
2211 2212 2213 2214 2215
	for (i = 0; i <= table->mask; i++) {
		INIT_HLIST_NULLS_HEAD(&table->hash2[i].head, i);
		table->hash2[i].count = 0;
		spin_lock_init(&table->hash2[i].lock);
	}
2216 2217
}

H
Hideo Aoki 已提交
2218 2219
void __init udp_init(void)
{
2220
	unsigned long limit;
H
Hideo Aoki 已提交
2221

2222
	udp_table_init(&udp_table, "UDP");
2223
	limit = nr_free_buffer_pages() / 8;
H
Hideo Aoki 已提交
2224 2225 2226 2227 2228 2229 2230 2231 2232
	limit = max(limit, 128UL);
	sysctl_udp_mem[0] = limit / 4 * 3;
	sysctl_udp_mem[1] = limit;
	sysctl_udp_mem[2] = sysctl_udp_mem[0] * 2;

	sysctl_udp_rmem_min = SK_MEM_QUANTUM;
	sysctl_udp_wmem_min = SK_MEM_QUANTUM;
}

2233 2234 2235 2236 2237 2238 2239 2240 2241 2242 2243 2244 2245 2246 2247 2248 2249 2250 2251
int udp4_ufo_send_check(struct sk_buff *skb)
{
	const struct iphdr *iph;
	struct udphdr *uh;

	if (!pskb_may_pull(skb, sizeof(*uh)))
		return -EINVAL;

	iph = ip_hdr(skb);
	uh = udp_hdr(skb);

	uh->check = ~csum_tcpudp_magic(iph->saddr, iph->daddr, skb->len,
				       IPPROTO_UDP, 0);
	skb->csum_start = skb_transport_header(skb) - skb->head;
	skb->csum_offset = offsetof(struct udphdr, check);
	skb->ip_summed = CHECKSUM_PARTIAL;
	return 0;
}

2252 2253
struct sk_buff *udp4_ufo_fragment(struct sk_buff *skb,
	netdev_features_t features)
2254 2255 2256 2257 2258 2259 2260 2261 2262 2263 2264 2265 2266 2267 2268 2269 2270 2271 2272 2273 2274 2275 2276 2277 2278 2279 2280
{
	struct sk_buff *segs = ERR_PTR(-EINVAL);
	unsigned int mss;
	int offset;
	__wsum csum;

	mss = skb_shinfo(skb)->gso_size;
	if (unlikely(skb->len <= mss))
		goto out;

	if (skb_gso_ok(skb, features | NETIF_F_GSO_ROBUST)) {
		/* Packet is from an untrusted source, reset gso_segs. */
		int type = skb_shinfo(skb)->gso_type;

		if (unlikely(type & ~(SKB_GSO_UDP | SKB_GSO_DODGY) ||
			     !(type & (SKB_GSO_UDP))))
			goto out;

		skb_shinfo(skb)->gso_segs = DIV_ROUND_UP(skb->len, mss);

		segs = NULL;
		goto out;
	}

	/* Do software UFO. Complete and fill in the UDP checksum as HW cannot
	 * do checksum of UDP packets sent as multiple IP fragments.
	 */
2281
	offset = skb_checksum_start_offset(skb);
E
Eric Dumazet 已提交
2282
	csum = skb_checksum(skb, offset, skb->len - offset, 0);
2283 2284 2285 2286 2287 2288 2289 2290 2291 2292 2293 2294
	offset += skb->csum_offset;
	*(__sum16 *)(skb->data + offset) = csum_fold(csum);
	skb->ip_summed = CHECKSUM_NONE;

	/* Fragment the skb. IP headers of the fragments are updated in
	 * inet_gso_segment()
	 */
	segs = skb_segment(skb, features);
out:
	return segs;
}