udp.c 64.3 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

80 81
#define pr_fmt(fmt) "UDP: " fmt

L
Linus Torvalds 已提交
82 83
#include <asm/uaccess.h>
#include <asm/ioctls.h>
H
Hideo Aoki 已提交
84
#include <linux/bootmem.h>
85 86
#include <linux/highmem.h>
#include <linux/swap.h>
L
Linus Torvalds 已提交
87 88 89 90 91
#include <linux/types.h>
#include <linux/fcntl.h>
#include <linux/module.h>
#include <linux/socket.h>
#include <linux/sockios.h>
92
#include <linux/igmp.h>
L
Linus Torvalds 已提交
93 94 95 96 97 98
#include <linux/in.h>
#include <linux/errno.h>
#include <linux/timer.h>
#include <linux/mm.h>
#include <linux/inet.h>
#include <linux/netdevice.h>
99
#include <linux/slab.h>
100
#include <net/tcp_states.h>
L
Linus Torvalds 已提交
101 102 103
#include <linux/skbuff.h>
#include <linux/proc_fs.h>
#include <linux/seq_file.h>
104
#include <net/net_namespace.h>
L
Linus Torvalds 已提交
105
#include <net/icmp.h>
S
Shawn Bohrer 已提交
106
#include <net/inet_hashtables.h>
L
Linus Torvalds 已提交
107 108 109
#include <net/route.h>
#include <net/checksum.h>
#include <net/xfrm.h>
110
#include <trace/events/udp.h>
111
#include <linux/static_key.h>
112
#include <trace/events/skb.h>
113
#include <net/busy_poll.h>
114
#include "udp_impl.h"
L
Linus Torvalds 已提交
115

116
struct udp_table udp_table __read_mostly;
117
EXPORT_SYMBOL(udp_table);
L
Linus Torvalds 已提交
118

E
Eric Dumazet 已提交
119
long sysctl_udp_mem[3] __read_mostly;
H
Hideo Aoki 已提交
120
EXPORT_SYMBOL(sysctl_udp_mem);
E
Eric Dumazet 已提交
121 122

int sysctl_udp_rmem_min __read_mostly;
H
Hideo Aoki 已提交
123
EXPORT_SYMBOL(sysctl_udp_rmem_min);
E
Eric Dumazet 已提交
124 125

int sysctl_udp_wmem_min __read_mostly;
H
Hideo Aoki 已提交
126 127
EXPORT_SYMBOL(sysctl_udp_wmem_min);

E
Eric Dumazet 已提交
128
atomic_long_t udp_memory_allocated;
H
Hideo Aoki 已提交
129 130
EXPORT_SYMBOL(udp_memory_allocated);

131 132
#define MAX_UDP_PORTS 65536
#define PORTS_PER_CHAIN (MAX_UDP_PORTS / UDP_HTABLE_SIZE_MIN)
133

134
static int udp_lib_lport_inuse(struct net *net, __u16 num,
135
			       const struct udp_hslot *hslot,
136
			       unsigned long *bitmap,
137 138
			       struct sock *sk,
			       int (*saddr_comp)(const struct sock *sk1,
139 140
						 const struct sock *sk2),
			       unsigned int log)
L
Linus Torvalds 已提交
141
{
142
	struct sock *sk2;
143
	struct hlist_nulls_node *node;
144
	kuid_t uid = sock_i_uid(sk);
145

146
	sk_nulls_for_each(sk2, node, &hslot->head)
147 148
		if (net_eq(sock_net(sk2), net) &&
		    sk2 != sk &&
149
		    (bitmap || udp_sk(sk2)->udp_port_hash == num) &&
150 151 152
		    (!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) &&
153 154
		    (!sk2->sk_reuseport || !sk->sk_reuseport ||
		      !uid_eq(uid, sock_i_uid(sk2))) &&
155 156
		    (*saddr_comp)(sk, sk2)) {
			if (bitmap)
157 158
				__set_bit(udp_sk(sk2)->udp_port_hash >> log,
					  bitmap);
159 160 161
			else
				return 1;
		}
162 163 164
	return 0;
}

E
Eric Dumazet 已提交
165 166 167 168 169 170 171 172 173 174 175 176
/*
 * 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;
177
	kuid_t uid = sock_i_uid(sk);
E
Eric Dumazet 已提交
178 179 180 181
	int res = 0;

	spin_lock(&hslot2->lock);
	udp_portaddr_for_each_entry(sk2, node, &hslot2->head)
182 183 184 185 186 187
		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) &&
188 189
		    (!sk2->sk_reuseport || !sk->sk_reuseport ||
		      !uid_eq(uid, sock_i_uid(sk2))) &&
E
Eric Dumazet 已提交
190 191 192 193 194 195 196 197
		    (*saddr_comp)(sk, sk2)) {
			res = 1;
			break;
		}
	spin_unlock(&hslot2->lock);
	return res;
}

198
/**
199
 *  udp_lib_get_port  -  UDP/-Lite port lookup for IPv4 and IPv6
200 201 202
 *
 *  @sk:          socket struct in question
 *  @snum:        port number to look up
203
 *  @saddr_comp:  AF-dependent comparison of bound local IP addresses
L
Lucas De Marchi 已提交
204
 *  @hash2_nulladdr: AF-dependent hash value in secondary hash chains,
E
Eric Dumazet 已提交
205
 *                   with NULL address
206
 */
207
int udp_lib_get_port(struct sock *sk, unsigned short snum,
208
		       int (*saddr_comp)(const struct sock *sk1,
E
Eric Dumazet 已提交
209 210
					 const struct sock *sk2),
		     unsigned int hash2_nulladdr)
211
{
212
	struct udp_hslot *hslot, *hslot2;
213
	struct udp_table *udptable = sk->sk_prot->h.udp_table;
214
	int    error = 1;
215
	struct net *net = sock_net(sk);
L
Linus Torvalds 已提交
216

217
	if (!snum) {
E
Eric Dumazet 已提交
218
		int low, high, remaining;
219
		unsigned int rand;
220 221
		unsigned short first, last;
		DECLARE_BITMAP(bitmap, PORTS_PER_CHAIN);
222

223
		inet_get_local_port_range(net, &low, &high);
224
		remaining = (high - low) + 1;
225

E
Eric Dumazet 已提交
226
		rand = net_random();
227 228 229 230
		first = (((u64)rand * remaining) >> 32) + low;
		/*
		 * force rand to be an odd multiple of UDP_HTABLE_SIZE
		 */
231
		rand = (rand | 1) * (udptable->mask + 1);
E
Eric Dumazet 已提交
232 233
		last = first + udptable->mask + 1;
		do {
234
			hslot = udp_hashslot(udptable, net, first);
235
			bitmap_zero(bitmap, PORTS_PER_CHAIN);
236
			spin_lock_bh(&hslot->lock);
237
			udp_lib_lport_inuse(net, snum, hslot, bitmap, sk,
238
					    saddr_comp, udptable->log);
239 240 241 242 243 244 245

			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
246
			do {
247
				if (low <= snum && snum <= high &&
248 249
				    !test_bit(snum >> udptable->log, bitmap) &&
				    !inet_is_reserved_local_port(snum))
250 251 252 253
					goto found;
				snum += rand;
			} while (snum != first);
			spin_unlock_bh(&hslot->lock);
E
Eric Dumazet 已提交
254
		} while (++first != last);
255
		goto fail;
256
	} else {
257
		hslot = udp_hashslot(udptable, net, snum);
258
		spin_lock_bh(&hslot->lock);
E
Eric Dumazet 已提交
259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282
		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:
283 284
		if (udp_lib_lport_inuse(net, snum, hslot, NULL, sk,
					saddr_comp, 0))
285 286
			goto fail_unlock;
	}
287
found:
E
Eric Dumazet 已提交
288
	inet_sk(sk)->inet_num = snum;
289 290
	udp_sk(sk)->udp_port_hash = snum;
	udp_sk(sk)->udp_portaddr_hash ^= snum;
L
Linus Torvalds 已提交
291
	if (sk_unhashed(sk)) {
292
		sk_nulls_add_node_rcu(sk, &hslot->head);
E
Eric Dumazet 已提交
293
		hslot->count++;
294
		sock_prot_inuse_add(sock_net(sk), sk->sk_prot, 1);
295 296 297 298 299 300 301

		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
302
	}
303
	error = 0;
304 305
fail_unlock:
	spin_unlock_bh(&hslot->lock);
L
Linus Torvalds 已提交
306
fail:
307 308
	return error;
}
E
Eric Dumazet 已提交
309
EXPORT_SYMBOL(udp_lib_get_port);
310

311
static int ipv4_rcv_saddr_equal(const struct sock *sk1, const struct sock *sk2)
312 313 314
{
	struct inet_sock *inet1 = inet_sk(sk1), *inet2 = inet_sk(sk2);

E
Eric Dumazet 已提交
315
	return 	(!ipv6_only_sock(sk2)  &&
E
Eric Dumazet 已提交
316 317
		 (!inet1->inet_rcv_saddr || !inet2->inet_rcv_saddr ||
		   inet1->inet_rcv_saddr == inet2->inet_rcv_saddr));
318 319
}

320 321 322
static unsigned int udp4_portaddr_hash(struct net *net, __be32 saddr,
				       unsigned int port)
{
323
	return jhash_1word((__force u32)saddr, net_hash_mix(net)) ^ port;
324 325
}

326
int udp_v4_get_port(struct sock *sk, unsigned short snum)
327
{
E
Eric Dumazet 已提交
328
	unsigned int hash2_nulladdr =
329
		udp4_portaddr_hash(sock_net(sk), htonl(INADDR_ANY), snum);
E
Eric Dumazet 已提交
330 331 332
	unsigned int hash2_partial =
		udp4_portaddr_hash(sock_net(sk), inet_sk(sk)->inet_rcv_saddr, 0);

333
	/* precompute partial secondary hash */
E
Eric Dumazet 已提交
334 335
	udp_sk(sk)->udp_portaddr_hash = hash2_partial;
	return udp_lib_get_port(sk, snum, ipv4_rcv_saddr_equal, hash2_nulladdr);
336 337
}

338 339 340 341 342 343
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;

344
	if (net_eq(sock_net(sk), net) && udp_sk(sk)->udp_port_hash == hnum &&
345 346 347
			!ipv6_only_sock(sk)) {
		struct inet_sock *inet = inet_sk(sk);

348
		score = (sk->sk_family == PF_INET ? 2 : 1);
E
Eric Dumazet 已提交
349 350
		if (inet->inet_rcv_saddr) {
			if (inet->inet_rcv_saddr != daddr)
351
				return -1;
352
			score += 4;
353
		}
E
Eric Dumazet 已提交
354 355
		if (inet->inet_daddr) {
			if (inet->inet_daddr != saddr)
356
				return -1;
357
			score += 4;
358
		}
E
Eric Dumazet 已提交
359 360
		if (inet->inet_dport) {
			if (inet->inet_dport != sport)
361
				return -1;
362
			score += 4;
363 364 365 366
		}
		if (sk->sk_bound_dev_if) {
			if (sk->sk_bound_dev_if != dif)
				return -1;
367
			score += 4;
368 369 370 371 372
		}
	}
	return score;
}

E
Eric Dumazet 已提交
373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389
/*
 * In this second variant, we check (daddr, dport) matches (inet_rcv_sadd, inet_num)
 */
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;

390
		score = (sk->sk_family == PF_INET ? 2 : 1);
E
Eric Dumazet 已提交
391 392 393
		if (inet->inet_daddr) {
			if (inet->inet_daddr != saddr)
				return -1;
394
			score += 4;
E
Eric Dumazet 已提交
395 396 397 398
		}
		if (inet->inet_dport) {
			if (inet->inet_dport != sport)
				return -1;
399
			score += 4;
E
Eric Dumazet 已提交
400 401 402 403
		}
		if (sk->sk_bound_dev_if) {
			if (sk->sk_bound_dev_if != dif)
				return -1;
404
			score += 4;
E
Eric Dumazet 已提交
405 406 407 408 409
		}
	}
	return score;
}

410 411 412 413
static unsigned int udp_ehashfn(struct net *net, const __be32 laddr,
				 const __u16 lport, const __be32 faddr,
				 const __be16 fport)
{
414 415 416 417
	static u32 udp_ehash_secret __read_mostly;

	net_get_random_once(&udp_ehash_secret, sizeof(udp_ehash_secret));

418
	return __inet_ehashfn(laddr, lport, faddr, fport,
419
			      udp_ehash_secret + net_hash_mix(net));
420 421
}

E
Eric Dumazet 已提交
422 423 424 425 426 427 428 429 430

/* 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;
431 432
	int score, badness, matches = 0, reuseport = 0;
	u32 hash = 0;
E
Eric Dumazet 已提交
433 434 435

begin:
	result = NULL;
436
	badness = 0;
E
Eric Dumazet 已提交
437 438 439 440 441 442
	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;
443 444
			reuseport = sk->sk_reuseport;
			if (reuseport) {
445 446
				hash = udp_ehashfn(net, daddr, hnum,
						   saddr, sport);
447 448 449 450 451 452 453
				matches = 1;
			}
		} else if (score == badness && reuseport) {
			matches++;
			if (((u64)hash * matches) >> 32 == 0)
				result = sk;
			hash = next_pseudo_random32(hash);
E
Eric Dumazet 已提交
454 455 456 457 458 459 460 461 462 463
		}
	}
	/*
	 * 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) {
E
Eric Dumazet 已提交
464
		if (unlikely(!atomic_inc_not_zero_hint(&result->sk_refcnt, 2)))
E
Eric Dumazet 已提交
465 466 467 468 469 470 471 472 473 474
			result = NULL;
		else if (unlikely(compute_score2(result, net, saddr, sport,
				  daddr, hnum, dif) < badness)) {
			sock_put(result);
			goto begin;
		}
	}
	return result;
}

475 476 477
/* UDP is nearly always wildcards out the wazoo, it makes no sense to try
 * harder than this. -DaveM
 */
478
struct sock *__udp4_lib_lookup(struct net *net, __be32 saddr,
479
		__be16 sport, __be32 daddr, __be16 dport,
480
		int dif, struct udp_table *udptable)
481
{
482
	struct sock *sk, *result;
483
	struct hlist_nulls_node *node;
484
	unsigned short hnum = ntohs(dport);
E
Eric Dumazet 已提交
485 486
	unsigned int hash2, slot2, slot = udp_hashfn(net, hnum, udptable->mask);
	struct udp_hslot *hslot2, *hslot = &udptable->hash[slot];
487 488
	int score, badness, matches = 0, reuseport = 0;
	u32 hash = 0;
489

490
	rcu_read_lock();
E
Eric Dumazet 已提交
491 492 493 494 495 496 497 498 499 500 501
	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) {
502
			hash2 = udp4_portaddr_hash(net, htonl(INADDR_ANY), hnum);
E
Eric Dumazet 已提交
503 504 505 506 507
			slot2 = hash2 & udptable->mask;
			hslot2 = &udptable->hash2[slot2];
			if (hslot->count < hslot2->count)
				goto begin;

508
			result = udp4_lib_lookup2(net, saddr, sport,
509
						  htonl(INADDR_ANY), hnum, dif,
E
Eric Dumazet 已提交
510 511 512 513 514
						  hslot2, slot2);
		}
		rcu_read_unlock();
		return result;
	}
515 516
begin:
	result = NULL;
517
	badness = 0;
518
	sk_nulls_for_each_rcu(sk, node, &hslot->head) {
519 520 521 522 523
		score = compute_score(sk, net, saddr, hnum, sport,
				      daddr, dport, dif);
		if (score > badness) {
			result = sk;
			badness = score;
524 525
			reuseport = sk->sk_reuseport;
			if (reuseport) {
526 527
				hash = udp_ehashfn(net, daddr, hnum,
						   saddr, sport);
528 529 530 531 532 533 534
				matches = 1;
			}
		} else if (score == badness && reuseport) {
			matches++;
			if (((u64)hash * matches) >> 32 == 0)
				result = sk;
			hash = next_pseudo_random32(hash);
535 536
		}
	}
537 538 539 540 541
	/*
	 * 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
542
	if (get_nulls_value(node) != slot)
543 544
		goto begin;

545
	if (result) {
E
Eric Dumazet 已提交
546
		if (unlikely(!atomic_inc_not_zero_hint(&result->sk_refcnt, 2)))
547 548 549 550 551 552 553 554
			result = NULL;
		else if (unlikely(compute_score(result, net, saddr, hnum, sport,
				  daddr, dport, dif) < badness)) {
			sock_put(result);
			goto begin;
		}
	}
	rcu_read_unlock();
555 556
	return result;
}
557
EXPORT_SYMBOL_GPL(__udp4_lib_lookup);
558

559 560
static inline struct sock *__udp4_lib_lookup_skb(struct sk_buff *skb,
						 __be16 sport, __be16 dport,
561
						 struct udp_table *udptable)
562
{
563
	struct sock *sk;
564 565
	const struct iphdr *iph = ip_hdr(skb);

566 567 568
	if (unlikely(sk = skb_steal_sock(skb)))
		return sk;
	else
E
Eric Dumazet 已提交
569
		return __udp4_lib_lookup(dev_net(skb_dst(skb)->dev), iph->saddr, sport,
570 571
					 iph->daddr, dport, inet_iif(skb),
					 udptable);
572 573
}

574 575 576
struct sock *udp4_lib_lookup(struct net *net, __be32 saddr, __be16 sport,
			     __be32 daddr, __be16 dport, int dif)
{
577
	return __udp4_lib_lookup(net, saddr, sport, daddr, dport, dif, &udp_table);
578 579 580
}
EXPORT_SYMBOL_GPL(udp4_lib_lookup);

S
Shawn Bohrer 已提交
581 582 583 584 585 586 587 588 589 590 591 592 593 594 595 596 597 598 599 600
static inline bool __udp_is_mcast_sock(struct net *net, struct sock *sk,
				       __be16 loc_port, __be32 loc_addr,
				       __be16 rmt_port, __be32 rmt_addr,
				       int dif, unsigned short hnum)
{
	struct inet_sock *inet = inet_sk(sk);

	if (!net_eq(sock_net(sk), net) ||
	    udp_sk(sk)->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(sk) ||
	    (sk->sk_bound_dev_if && sk->sk_bound_dev_if != dif))
		return false;
	if (!ip_mc_sf_allow(sk, loc_addr, rmt_addr, dif))
		return false;
	return true;
}

601
static inline struct sock *udp_v4_mcast_next(struct net *net, struct sock *sk,
602 603 604 605
					     __be16 loc_port, __be32 loc_addr,
					     __be16 rmt_port, __be32 rmt_addr,
					     int dif)
{
606
	struct hlist_nulls_node *node;
607 608 609
	struct sock *s = sk;
	unsigned short hnum = ntohs(loc_port);

610
	sk_nulls_for_each_from(s, node) {
S
Shawn Bohrer 已提交
611 612 613 614 615
		if (__udp_is_mcast_sock(net, s,
					loc_port, loc_addr,
					rmt_port, rmt_addr,
					dif, hnum))
			goto found;
616 617 618 619 620 621 622 623 624 625 626 627 628 629 630 631 632
	}
	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.
 */

633
void __udp4_lib_err(struct sk_buff *skb, u32 info, struct udp_table *udptable)
634 635
{
	struct inet_sock *inet;
636
	const struct iphdr *iph = (const struct iphdr *)skb->data;
E
Eric Dumazet 已提交
637
	struct udphdr *uh = (struct udphdr *)(skb->data+(iph->ihl<<2));
638 639 640 641 642
	const int type = icmp_hdr(skb)->type;
	const int code = icmp_hdr(skb)->code;
	struct sock *sk;
	int harderr;
	int err;
643
	struct net *net = dev_net(skb->dev);
644

645
	sk = __udp4_lib_lookup(net, iph->daddr, uh->dest,
646 647
			iph->saddr, uh->source, skb->dev->ifindex, udptable);
	if (sk == NULL) {
648
		ICMP_INC_STATS_BH(net, ICMP_MIB_INERRORS);
649 650 651 652 653 654 655 656 657 658 659 660 661 662 663 664 665 666 667 668
		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 */
669
			ipv4_sk_update_pmtu(skb, sk, info);
670 671 672 673 674 675 676 677 678 679 680 681 682
			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;
683 684
	case ICMP_REDIRECT:
		ipv4_sk_redirect(skb, sk);
685
		goto out;
686 687 688 689 690 691 692 693 694
	}

	/*
	 *      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;
695
	} else
E
Eric Dumazet 已提交
696
		ip_icmp_error(sk, skb, err, uh->dest, info, (u8 *)(uh+1));
697

698 699 700 701 702 703 704 705
	sk->sk_err = err;
	sk->sk_error_report(sk);
out:
	sock_put(sk);
}

void udp_err(struct sk_buff *skb, u32 info)
{
706
	__udp4_lib_err(skb, info, &udp_table);
707 708 709 710 711
}

/*
 * Throw away all pending data and cancel the corking. Socket is locked.
 */
712
void udp_flush_pending_frames(struct sock *sk)
713 714 715 716 717 718 719 720 721
{
	struct udp_sock *up = udp_sk(sk);

	if (up->pending) {
		up->len = 0;
		up->pending = 0;
		ip_flush_pending_frames(sk);
	}
}
722
EXPORT_SYMBOL(udp_flush_pending_frames);
723 724

/**
H
Herbert Xu 已提交
725
 * 	udp4_hwcsum  -  handle outgoing HW checksumming
726 727
 * 	@skb: 	sk_buff containing the filled-in UDP header
 * 	        (checksum field must be zeroed out)
H
Herbert Xu 已提交
728 729
 *	@src:	source IP address
 *	@dst:	destination IP address
730
 */
731
void udp4_hwcsum(struct sk_buff *skb, __be32 src, __be32 dst)
732 733
{
	struct udphdr *uh = udp_hdr(skb);
H
Herbert Xu 已提交
734 735 736 737
	struct sk_buff *frags = skb_shinfo(skb)->frag_list;
	int offset = skb_transport_offset(skb);
	int len = skb->len - offset;
	int hlen = len;
738 739
	__wsum csum = 0;

H
Herbert Xu 已提交
740
	if (!frags) {
741 742 743 744 745
		/*
		 * 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
746 747
		uh->check = ~csum_tcpudp_magic(src, dst, len,
					       IPPROTO_UDP, 0);
748 749 750 751 752 753
	} 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
754 755 756 757
		do {
			csum = csum_add(csum, frags->csum);
			hlen -= frags->len;
		} while ((frags = frags->next));
758

H
Herbert Xu 已提交
759
		csum = skb_checksum(skb, offset, hlen, csum);
760 761 762 763 764 765 766
		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;
	}
}
767
EXPORT_SYMBOL_GPL(udp4_hwcsum);
768

769
static int udp_send_skb(struct sk_buff *skb, struct flowi4 *fl4)
770
{
H
Herbert Xu 已提交
771
	struct sock *sk = skb->sk;
772 773 774 775
	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
776 777
	int offset = skb_transport_offset(skb);
	int len = skb->len - offset;
778 779 780 781 782 783
	__wsum csum = 0;

	/*
	 * Create a UDP header
	 */
	uh = udp_hdr(skb);
H
Herbert Xu 已提交
784
	uh->source = inet->inet_sport;
785
	uh->dest = fl4->fl4_dport;
H
Herbert Xu 已提交
786
	uh->len = htons(len);
787 788 789
	uh->check = 0;

	if (is_udplite)  				 /*     UDP-Lite      */
H
Herbert Xu 已提交
790
		csum = udplite_csum(skb);
791 792 793 794 795 796 797 798

	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 */

799
		udp4_hwcsum(skb, fl4->saddr, fl4->daddr);
800 801
		goto send;

H
Herbert Xu 已提交
802 803
	} else
		csum = udp_csum(skb);
804 805

	/* add protocol-dependent pseudo-header */
806
	uh->check = csum_tcpudp_magic(fl4->saddr, fl4->daddr, len,
E
Eric Dumazet 已提交
807
				      sk->sk_protocol, csum);
808 809 810 811
	if (uh->check == 0)
		uh->check = CSUM_MANGLED_0;

send:
E
Eric Dumazet 已提交
812
	err = ip_send_skb(sock_net(sk), skb);
E
Eric Dumazet 已提交
813 814 815 816 817 818 819 820 821
	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
822 823 824 825 826 827
	return err;
}

/*
 * Push out all pending data as one UDP datagram. Socket is locked.
 */
828
int udp_push_pending_frames(struct sock *sk)
H
Herbert Xu 已提交
829 830 831
{
	struct udp_sock  *up = udp_sk(sk);
	struct inet_sock *inet = inet_sk(sk);
D
David S. Miller 已提交
832
	struct flowi4 *fl4 = &inet->cork.fl.u.ip4;
H
Herbert Xu 已提交
833 834 835
	struct sk_buff *skb;
	int err = 0;

836
	skb = ip_finish_skb(sk, fl4);
H
Herbert Xu 已提交
837 838 839
	if (!skb)
		goto out;

840
	err = udp_send_skb(skb, fl4);
H
Herbert Xu 已提交
841

842 843 844 845 846
out:
	up->len = 0;
	up->pending = 0;
	return err;
}
847
EXPORT_SYMBOL(udp_push_pending_frames);
848 849 850 851 852 853

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);
854
	struct flowi4 fl4_stack;
D
David S. Miller 已提交
855
	struct flowi4 *fl4;
856 857 858 859 860 861 862 863 864 865 866
	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
867
	struct sk_buff *skb;
868
	struct ip_options_data opt_copy;
869 870 871 872 873 874 875 876

	if (len > 0xFFFF)
		return -EMSGSIZE;

	/*
	 *	Check the flags.
	 */

E
Eric Dumazet 已提交
877
	if (msg->msg_flags & MSG_OOB) /* Mirror BSD error message compatibility */
878 879 880
		return -EOPNOTSUPP;

	ipc.opt = NULL;
881
	ipc.tx_flags = 0;
882 883
	ipc.ttl = 0;
	ipc.tos = -1;
884

H
Herbert Xu 已提交
885 886
	getfrag = is_udplite ? udplite_getfrag : ip_generic_getfrag;

887
	fl4 = &inet->cork.fl.u.ip4;
888 889 890 891 892 893 894 895 896 897 898 899 900 901 902 903 904 905 906 907 908
	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) {
D
Daniel Baluta 已提交
909
		struct sockaddr_in *usin = (struct sockaddr_in *)msg->msg_name;
910 911 912 913 914 915 916 917 918 919 920 921 922 923
		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
924 925
		daddr = inet->inet_daddr;
		dport = inet->inet_dport;
926 927 928 929 930
		/* 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
931
	ipc.addr = inet->inet_saddr;
932 933

	ipc.oif = sk->sk_bound_dev_if;
934 935 936

	sock_tx_timestamp(sk, &ipc.tx_flags);

937
	if (msg->msg_controllen) {
938
		err = ip_cmsg_send(sock_net(sk), msg, &ipc);
939 940 941 942 943 944
		if (err)
			return err;
		if (ipc.opt)
			free = 1;
		connected = 0;
	}
945 946 947 948 949 950 951 952 953 954 955 956
	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();
	}
957 958 959 960

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

961
	if (ipc.opt && ipc.opt->opt.srr) {
962 963
		if (!daddr)
			return -EINVAL;
964
		faddr = ipc.opt->opt.faddr;
965 966
		connected = 0;
	}
967
	tos = get_rttos(&ipc, inet);
968 969
	if (sock_flag(sk, SOCK_LOCALROUTE) ||
	    (msg->msg_flags & MSG_DONTROUTE) ||
970
	    (ipc.opt && ipc.opt->opt.is_strictroute)) {
971 972 973 974 975 976 977 978 979 980
		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;
981 982
	} else if (!ipc.oif)
		ipc.oif = inet->uc_index;
983 984

	if (connected)
E
Eric Dumazet 已提交
985
		rt = (struct rtable *)sk_dst_check(sk, 0);
986 987

	if (rt == NULL) {
988 989
		struct net *net = sock_net(sk);

990 991
		fl4 = &fl4_stack;
		flowi4_init_output(fl4, ipc.oif, sk->sk_mark, tos,
992 993 994 995
				   RT_SCOPE_UNIVERSE, sk->sk_protocol,
				   inet_sk_flowi_flags(sk)|FLOWI_FLAG_CAN_SLEEP,
				   faddr, saddr, dport, inet->inet_sport);

996 997
		security_sk_classify_flow(sk, flowi4_to_flowi(fl4));
		rt = ip_route_output_flow(net, fl4, sk);
998 999
		if (IS_ERR(rt)) {
			err = PTR_ERR(rt);
1000
			rt = NULL;
1001
			if (err == -ENETUNREACH)
1002
				IP_INC_STATS_BH(net, IPSTATS_MIB_OUTNOROUTES);
1003 1004 1005 1006 1007 1008 1009 1010
			goto out;
		}

		err = -EACCES;
		if ((rt->rt_flags & RTCF_BROADCAST) &&
		    !sock_flag(sk, SOCK_BROADCAST))
			goto out;
		if (connected)
1011
			sk_dst_set(sk, dst_clone(&rt->dst));
1012 1013 1014 1015 1016 1017
	}

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

1018
	saddr = fl4->saddr;
1019
	if (!ipc.addr)
1020
		daddr = ipc.addr = fl4->daddr;
1021

H
Herbert Xu 已提交
1022 1023
	/* Lockless fast path for the non-corking case. */
	if (!corkreq) {
1024
		skb = ip_make_skb(sk, fl4, getfrag, msg->msg_iov, ulen,
H
Herbert Xu 已提交
1025 1026 1027
				  sizeof(struct udphdr), &ipc, &rt,
				  msg->msg_flags);
		err = PTR_ERR(skb);
1028
		if (!IS_ERR_OR_NULL(skb))
1029
			err = udp_send_skb(skb, fl4);
H
Herbert Xu 已提交
1030 1031 1032
		goto out;
	}

1033 1034 1035 1036 1037 1038
	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);

1039
		LIMIT_NETDEBUG(KERN_DEBUG pr_fmt("cork app bug 2\n"));
1040 1041 1042 1043 1044 1045
		err = -EINVAL;
		goto out;
	}
	/*
	 *	Now cork the socket to pend data.
	 */
D
David S. Miller 已提交
1046 1047 1048
	fl4 = &inet->cork.fl.u.ip4;
	fl4->daddr = daddr;
	fl4->saddr = saddr;
1049 1050
	fl4->fl4_dport = dport;
	fl4->fl4_sport = inet->inet_sport;
1051 1052 1053 1054
	up->pending = AF_INET;

do_append_data:
	up->len += ulen;
1055 1056 1057
	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);
1058 1059 1060 1061 1062 1063 1064 1065 1066 1067 1068 1069 1070 1071 1072 1073 1074 1075 1076 1077 1078 1079
	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)) {
1080 1081
		UDP_INC_STATS_USER(sock_net(sk),
				UDP_MIB_SNDBUFERRORS, is_udplite);
1082 1083 1084 1085
	}
	return err;

do_confirm:
1086
	dst_confirm(&rt->dst);
1087 1088 1089 1090 1091
	if (!(msg->msg_flags&MSG_PROBE) || len)
		goto back_from_confirm;
	err = 0;
	goto out;
}
E
Eric Dumazet 已提交
1092
EXPORT_SYMBOL(udp_sendmsg);
1093 1094 1095 1096

int udp_sendpage(struct sock *sk, struct page *page, int offset,
		 size_t size, int flags)
{
1097
	struct inet_sock *inet = inet_sk(sk);
1098 1099 1100 1101 1102 1103 1104 1105 1106 1107 1108 1109 1110 1111 1112 1113 1114 1115 1116 1117
	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);

1118
		LIMIT_NETDEBUG(KERN_DEBUG pr_fmt("udp cork app bug 3\n"));
1119 1120 1121
		return -EINVAL;
	}

1122 1123
	ret = ip_append_page(sk, &inet->cork.fl.u.ip4,
			     page, offset, size, flags);
1124 1125 1126 1127 1128 1129 1130 1131 1132 1133 1134 1135 1136 1137 1138 1139 1140 1141 1142 1143
	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
1144 1145 1146 1147 1148 1149 1150 1151 1152 1153 1154 1155 1156 1157 1158 1159 1160 1161 1162

/**
 *	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)) {
1163 1164
		UDP_INC_STATS_BH(sock_net(sk), UDP_MIB_CSUMERRORS,
				 IS_UDPLITE(sk));
E
Eric Dumazet 已提交
1165 1166
		UDP_INC_STATS_BH(sock_net(sk), UDP_MIB_INERRORS,
				 IS_UDPLITE(sk));
E
Eric Dumazet 已提交
1167
		atomic_inc(&sk->sk_drops);
E
Eric Dumazet 已提交
1168 1169 1170 1171 1172 1173 1174
		__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)) {
1175 1176
		bool slow = lock_sock_fast(sk);

E
Eric Dumazet 已提交
1177 1178
		__skb_queue_purge(&list_kill);
		sk_mem_reclaim_partial(sk);
1179
		unlock_sock_fast(sk, slow);
E
Eric Dumazet 已提交
1180 1181 1182 1183
	}
	return res;
}

L
Linus Torvalds 已提交
1184 1185 1186
/*
 *	IOCTL requests applicable to the UDP protocol
 */
1187

L
Linus Torvalds 已提交
1188 1189
int udp_ioctl(struct sock *sk, int cmd, unsigned long arg)
{
1190 1191
	switch (cmd) {
	case SIOCOUTQ:
L
Linus Torvalds 已提交
1192
	{
1193 1194
		int amount = sk_wmem_alloc_get(sk);

1195 1196
		return put_user(amount, (int __user *)arg);
	}
L
Linus Torvalds 已提交
1197

1198 1199
	case SIOCINQ:
	{
E
Eric Dumazet 已提交
1200
		unsigned int amount = first_packet_length(sk);
1201

E
Eric Dumazet 已提交
1202
		if (amount)
1203 1204 1205 1206 1207
			/*
			 * We will only return the amount
			 * of this packet since that is all
			 * that will be read.
			 */
E
Eric Dumazet 已提交
1208 1209
			amount -= sizeof(struct udphdr);

1210 1211
		return put_user(amount, (int __user *)arg);
	}
L
Linus Torvalds 已提交
1212

1213 1214
	default:
		return -ENOIOCTLCMD;
L
Linus Torvalds 已提交
1215
	}
1216 1217

	return 0;
L
Linus Torvalds 已提交
1218
}
E
Eric Dumazet 已提交
1219
EXPORT_SYMBOL(udp_ioctl);
L
Linus Torvalds 已提交
1220

1221 1222 1223 1224 1225 1226 1227 1228 1229 1230 1231
/*
 * 	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;
1232
	unsigned int ulen, copied;
1233
	int peeked, off = 0;
1234 1235
	int err;
	int is_udplite = IS_UDPLITE(sk);
1236
	bool slow;
1237 1238 1239 1240 1241 1242

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

try_again:
	skb = __skb_recv_datagram(sk, flags | (noblock ? MSG_DONTWAIT : 0),
1243
				  &peeked, &off, &err);
1244 1245 1246 1247
	if (!skb)
		goto out;

	ulen = skb->len - sizeof(struct udphdr);
1248 1249 1250 1251
	copied = len;
	if (copied > ulen)
		copied = ulen;
	else if (copied < ulen)
1252 1253 1254 1255 1256 1257 1258 1259
		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.
	 */

1260
	if (copied < ulen || UDP_SKB_CB(skb)->partial_cov) {
1261 1262 1263 1264 1265 1266
		if (udp_lib_checksum_complete(skb))
			goto csum_copy_err;
	}

	if (skb_csum_unnecessary(skb))
		err = skb_copy_datagram_iovec(skb, sizeof(struct udphdr),
1267
					      msg->msg_iov, copied);
1268
	else {
E
Eric Dumazet 已提交
1269 1270 1271
		err = skb_copy_and_csum_datagram_iovec(skb,
						       sizeof(struct udphdr),
						       msg->msg_iov);
1272 1273 1274 1275 1276

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

1277 1278
	if (unlikely(err)) {
		trace_kfree_skb(skb, udp_recvmsg);
1279 1280 1281 1282 1283
		if (!peeked) {
			atomic_inc(&sk->sk_drops);
			UDP_INC_STATS_USER(sock_net(sk),
					   UDP_MIB_INERRORS, is_udplite);
		}
1284
		goto out_free;
1285
	}
1286 1287

	if (!peeked)
1288 1289
		UDP_INC_STATS_USER(sock_net(sk),
				UDP_MIB_INDATAGRAMS, is_udplite);
1290

1291
	sock_recv_ts_and_drops(msg, sk, skb);
1292 1293

	/* Copy the address. */
E
Eric Dumazet 已提交
1294
	if (sin) {
1295 1296 1297 1298
		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));
1299
		*addr_len = sizeof(*sin);
1300 1301 1302 1303
	}
	if (inet->cmsg_flags)
		ip_cmsg_recv(msg, skb);

1304
	err = copied;
1305 1306 1307 1308
	if (flags & MSG_TRUNC)
		err = ulen;

out_free:
1309
	skb_free_datagram_locked(sk, skb);
1310 1311 1312 1313
out:
	return err;

csum_copy_err:
1314
	slow = lock_sock_fast(sk);
1315 1316
	if (!skb_kill_datagram(sk, skb, flags)) {
		UDP_INC_STATS_USER(sock_net(sk), UDP_MIB_CSUMERRORS, is_udplite);
1317
		UDP_INC_STATS_USER(sock_net(sk), UDP_MIB_INERRORS, is_udplite);
1318
	}
1319
	unlock_sock_fast(sk, slow);
1320 1321 1322

	if (noblock)
		return -EAGAIN;
1323 1324 1325

	/* starting over for a new packet */
	msg->msg_flags &= ~MSG_TRUNC;
1326 1327 1328 1329
	goto try_again;
}


L
Linus Torvalds 已提交
1330 1331 1332 1333 1334 1335
int udp_disconnect(struct sock *sk, int flags)
{
	struct inet_sock *inet = inet_sk(sk);
	/*
	 *	1003.1g - break association.
	 */
1336

L
Linus Torvalds 已提交
1337
	sk->sk_state = TCP_CLOSE;
E
Eric Dumazet 已提交
1338 1339
	inet->inet_daddr = 0;
	inet->inet_dport = 0;
1340
	sock_rps_reset_rxhash(sk);
L
Linus Torvalds 已提交
1341 1342 1343 1344 1345 1346
	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
1347
		inet->inet_sport = 0;
L
Linus Torvalds 已提交
1348 1349 1350 1351
	}
	sk_dst_reset(sk);
	return 0;
}
E
Eric Dumazet 已提交
1352
EXPORT_SYMBOL(udp_disconnect);
L
Linus Torvalds 已提交
1353

1354 1355
void udp_lib_unhash(struct sock *sk)
{
1356 1357
	if (sk_hashed(sk)) {
		struct udp_table *udptable = sk->sk_prot->h.udp_table;
1358 1359 1360 1361 1362
		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);
1363

1364 1365
		spin_lock_bh(&hslot->lock);
		if (sk_nulls_del_node_init_rcu(sk)) {
E
Eric Dumazet 已提交
1366
			hslot->count--;
E
Eric Dumazet 已提交
1367
			inet_sk(sk)->inet_num = 0;
1368
			sock_prot_inuse_add(sock_net(sk), sk->sk_prot, -1);
1369 1370 1371 1372 1373

			spin_lock(&hslot2->lock);
			hlist_nulls_del_init_rcu(&udp_sk(sk)->udp_portaddr_node);
			hslot2->count--;
			spin_unlock(&hslot2->lock);
1374 1375
		}
		spin_unlock_bh(&hslot->lock);
1376 1377 1378 1379
	}
}
EXPORT_SYMBOL(udp_lib_unhash);

E
Eric Dumazet 已提交
1380 1381 1382 1383 1384 1385 1386 1387 1388 1389 1390 1391 1392 1393 1394 1395 1396 1397 1398 1399 1400 1401 1402 1403 1404 1405 1406 1407 1408 1409 1410 1411 1412 1413 1414 1415 1416 1417 1418 1419 1420 1421 1422
/*
 * 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
1423 1424
static int __udp_queue_rcv_skb(struct sock *sk, struct sk_buff *skb)
{
T
Tom Herbert 已提交
1425
	int rc;
E
Eric Dumazet 已提交
1426

1427
	if (inet_sk(sk)->inet_daddr) {
1428
		sock_rps_save_rxhash(sk, skb);
1429 1430
		sk_mark_napi_id(sk, skb);
	}
T
Tom Herbert 已提交
1431

1432
	rc = sock_queue_rcv_skb(sk, skb);
E
Eric Dumazet 已提交
1433 1434
	if (rc < 0) {
		int is_udplite = IS_UDPLITE(sk);
H
Herbert Xu 已提交
1435 1436

		/* Note that an ENOMEM error is charged twice */
E
Eric Dumazet 已提交
1437
		if (rc == -ENOMEM)
H
Herbert Xu 已提交
1438 1439
			UDP_INC_STATS_BH(sock_net(sk), UDP_MIB_RCVBUFERRORS,
					 is_udplite);
E
Eric Dumazet 已提交
1440 1441
		UDP_INC_STATS_BH(sock_net(sk), UDP_MIB_INERRORS, is_udplite);
		kfree_skb(skb);
1442
		trace_udp_fail_queue_rcv_skb(rc, sk);
E
Eric Dumazet 已提交
1443
		return -1;
H
Herbert Xu 已提交
1444 1445 1446 1447 1448 1449
	}

	return 0;

}

1450 1451 1452 1453 1454 1455 1456 1457
static struct static_key udp_encap_needed __read_mostly;
void udp_encap_enable(void)
{
	if (!static_key_enabled(&udp_encap_needed))
		static_key_slow_inc(&udp_encap_needed);
}
EXPORT_SYMBOL(udp_encap_enable);

1458 1459 1460 1461 1462 1463 1464 1465
/* 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
1466
int udp_queue_rcv_skb(struct sock *sk, struct sk_buff *skb)
1467 1468 1469 1470 1471 1472 1473 1474 1475 1476 1477 1478
{
	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);

1479
	if (static_key_false(&udp_encap_needed) && up->encap_type) {
1480 1481
		int (*encap_rcv)(struct sock *sk, struct sk_buff *skb);

1482 1483 1484 1485 1486 1487 1488 1489 1490 1491 1492 1493
		/*
		 * 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 */
1494 1495
		encap_rcv = ACCESS_ONCE(up->encap_rcv);
		if (skb->len > sizeof(struct udphdr) && encap_rcv != NULL) {
1496 1497
			int ret;

1498
			ret = encap_rcv(sk, skb);
1499
			if (ret <= 0) {
1500 1501
				UDP_INC_STATS_BH(sock_net(sk),
						 UDP_MIB_INDATAGRAMS,
1502 1503 1504 1505 1506 1507 1508 1509 1510 1511 1512 1513 1514 1515 1516 1517 1518 1519 1520 1521 1522 1523 1524 1525 1526
						 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  */
1527 1528
			LIMIT_NETDEBUG(KERN_WARNING "UDPLite: partial coverage %d while full coverage %d requested\n",
				       UDP_SKB_CB(skb)->cscov, skb->len);
1529 1530 1531 1532 1533 1534 1535 1536 1537
			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) {
1538 1539
			LIMIT_NETDEBUG(KERN_WARNING "UDPLite: coverage %d too small, need min %d\n",
				       UDP_SKB_CB(skb)->cscov, up->pcrlen);
1540 1541 1542 1543
			goto drop;
		}
	}

1544 1545
	if (rcu_access_pointer(sk->sk_filter) &&
	    udp_lib_checksum_complete(skb))
1546
		goto csum_error;
1547

1548

1549
	if (sk_rcvqueues_full(sk, skb, sk->sk_rcvbuf))
1550 1551
		goto drop;

H
Herbert Xu 已提交
1552
	rc = 0;
1553

1554
	ipv4_pktinfo_prepare(sk, skb);
H
Herbert Xu 已提交
1555 1556 1557
	bh_lock_sock(sk);
	if (!sock_owned_by_user(sk))
		rc = __udp_queue_rcv_skb(sk, skb);
1558
	else if (sk_add_backlog(sk, skb, sk->sk_rcvbuf)) {
Z
Zhu Yi 已提交
1559 1560 1561
		bh_unlock_sock(sk);
		goto drop;
	}
H
Herbert Xu 已提交
1562 1563 1564
	bh_unlock_sock(sk);

	return rc;
1565

1566 1567
csum_error:
	UDP_INC_STATS_BH(sock_net(sk), UDP_MIB_CSUMERRORS, is_udplite);
1568
drop:
1569
	UDP_INC_STATS_BH(sock_net(sk), UDP_MIB_INERRORS, is_udplite);
E
Eric Dumazet 已提交
1570
	atomic_inc(&sk->sk_drops);
1571 1572 1573 1574
	kfree_skb(skb);
	return -1;
}

1575 1576 1577 1578 1579 1580

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;
1581
	struct sock *sk;
1582 1583

	for (i = 0; i < count; i++) {
1584
		sk = stack[i];
1585 1586 1587
		if (likely(skb1 == NULL))
			skb1 = (i == final) ? skb : skb_clone(skb, GFP_ATOMIC);

1588 1589 1590 1591 1592 1593 1594 1595 1596
		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)
1597 1598 1599 1600 1601 1602
			skb1 = NULL;
	}
	if (unlikely(skb1))
		kfree_skb(skb1);
}

S
Shawn Bohrer 已提交
1603 1604 1605 1606 1607 1608 1609 1610
static void udp_sk_rx_dst_set(struct sock *sk, const struct sk_buff *skb)
{
	struct dst_entry *dst = skb_dst(skb);

	dst_hold(dst);
	sk->sk_rx_dst = dst;
}

1611 1612 1613
/*
 *	Multicasts and broadcasts go to each listener.
 *
1614
 *	Note: called only from the BH handler context.
1615
 */
1616
static int __udp4_lib_mcast_deliver(struct net *net, struct sk_buff *skb,
1617 1618
				    struct udphdr  *uh,
				    __be32 saddr, __be32 daddr,
1619
				    struct udp_table *udptable)
1620
{
1621
	struct sock *sk, *stack[256 / sizeof(struct sock *)];
1622
	struct udp_hslot *hslot = udp_hashslot(udptable, net, ntohs(uh->dest));
1623
	int dif;
1624
	unsigned int i, count = 0;
1625

1626
	spin_lock(&hslot->lock);
1627
	sk = sk_nulls_head(&hslot->head);
1628
	dif = skb->dev->ifindex;
1629
	sk = udp_v4_mcast_next(net, sk, uh->dest, daddr, uh->source, saddr, dif);
1630 1631 1632 1633 1634 1635 1636 1637 1638 1639 1640 1641 1642 1643 1644 1645 1646
	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]);

1647
	spin_unlock(&hslot->lock);
1648 1649 1650 1651 1652 1653 1654 1655 1656 1657 1658 1659

	/*
	 * 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);
	}
1660 1661 1662 1663 1664 1665 1666 1667 1668 1669 1670 1671 1672 1673 1674 1675 1676 1677 1678 1679 1680 1681 1682 1683 1684 1685 1686
	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
1687
		if (!csum_tcpudp_magic(iph->saddr, iph->daddr, skb->len,
1688 1689 1690 1691 1692 1693 1694 1695 1696 1697 1698 1699 1700 1701 1702 1703 1704
				      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.
 */

1705
int __udp4_lib_rcv(struct sk_buff *skb, struct udp_table *udptable,
1706 1707 1708
		   int proto)
{
	struct sock *sk;
1709
	struct udphdr *uh;
1710
	unsigned short ulen;
E
Eric Dumazet 已提交
1711
	struct rtable *rt = skb_rtable(skb);
1712
	__be32 saddr, daddr;
1713
	struct net *net = dev_net(skb->dev);
1714 1715 1716 1717 1718 1719 1720

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

1721
	uh   = udp_hdr(skb);
1722
	ulen = ntohs(uh->len);
1723 1724 1725
	saddr = ip_hdr(skb)->saddr;
	daddr = ip_hdr(skb)->daddr;

1726 1727 1728 1729 1730 1731 1732 1733 1734 1735 1736 1737 1738
	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;

S
Shawn Bohrer 已提交
1739 1740 1741 1742 1743 1744
	if (skb->sk) {
		int ret;
		sk = skb->sk;

		if (unlikely(sk->sk_rx_dst == NULL))
			udp_sk_rx_dst_set(sk, skb);
1745

S
Shawn Bohrer 已提交
1746 1747 1748 1749 1750 1751 1752 1753 1754 1755 1756 1757 1758 1759 1760
		ret = udp_queue_rcv_skb(sk, skb);

		/* 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;
	} else {
		if (rt->rt_flags & (RTCF_BROADCAST|RTCF_MULTICAST))
			return __udp4_lib_mcast_deliver(net, skb, uh,
					saddr, daddr, udptable);

		sk = __udp4_lib_lookup_skb(skb, uh->source, uh->dest, udptable);
	}
1761 1762

	if (sk != NULL) {
1763 1764 1765
		int ret;

		ret = udp_queue_rcv_skb(sk, skb);
1766 1767 1768 1769 1770 1771 1772 1773 1774 1775 1776 1777 1778 1779 1780 1781 1782 1783
		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;

1784
	UDP_INC_STATS_BH(net, UDP_MIB_NOPORTS, proto == IPPROTO_UDPLITE);
1785 1786 1787 1788 1789 1790 1791 1792 1793 1794
	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:
1795
	LIMIT_NETDEBUG(KERN_DEBUG "UDP%s: short packet: From %pI4:%u %d/%d to %pI4:%u\n",
1796 1797 1798 1799
		       proto == IPPROTO_UDPLITE ? "Lite" : "",
		       &saddr, ntohs(uh->source),
		       ulen, skb->len,
		       &daddr, ntohs(uh->dest));
1800 1801 1802 1803 1804 1805 1806
	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).
	 */
1807
	LIMIT_NETDEBUG(KERN_DEBUG "UDP%s: bad checksum. From %pI4:%u to %pI4:%u ulen %d\n",
1808 1809
		       proto == IPPROTO_UDPLITE ? "Lite" : "",
		       &saddr, ntohs(uh->source), &daddr, ntohs(uh->dest),
1810
		       ulen);
1811
	UDP_INC_STATS_BH(net, UDP_MIB_CSUMERRORS, proto == IPPROTO_UDPLITE);
1812
drop:
1813
	UDP_INC_STATS_BH(net, UDP_MIB_INERRORS, proto == IPPROTO_UDPLITE);
1814 1815 1816 1817
	kfree_skb(skb);
	return 0;
}

S
Shawn Bohrer 已提交
1818 1819 1820 1821 1822 1823 1824 1825 1826 1827 1828 1829 1830 1831 1832 1833 1834 1835 1836 1837 1838 1839 1840 1841 1842 1843 1844 1845 1846 1847 1848 1849 1850 1851 1852 1853 1854 1855 1856
/* We can only early demux multicast if there is a single matching socket.
 * If more than one socket found returns NULL
 */
static struct sock *__udp4_lib_mcast_demux_lookup(struct net *net,
						  __be16 loc_port, __be32 loc_addr,
						  __be16 rmt_port, __be32 rmt_addr,
						  int dif)
{
	struct sock *sk, *result;
	struct hlist_nulls_node *node;
	unsigned short hnum = ntohs(loc_port);
	unsigned int count, slot = udp_hashfn(net, hnum, udp_table.mask);
	struct udp_hslot *hslot = &udp_table.hash[slot];

	rcu_read_lock();
begin:
	count = 0;
	result = NULL;
	sk_nulls_for_each_rcu(sk, node, &hslot->head) {
		if (__udp_is_mcast_sock(net, sk,
					loc_port, loc_addr,
					rmt_port, rmt_addr,
					dif, hnum)) {
			result = sk;
			++count;
		}
	}
	/*
	 * 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) != slot)
		goto begin;

	if (result) {
		if (count != 1 ||
		    unlikely(!atomic_inc_not_zero_hint(&result->sk_refcnt, 2)))
			result = NULL;
1857
		else if (unlikely(!__udp_is_mcast_sock(net, result,
S
Shawn Bohrer 已提交
1858 1859 1860 1861 1862 1863 1864 1865 1866 1867 1868 1869 1870 1871 1872 1873 1874 1875 1876 1877 1878 1879 1880 1881 1882 1883 1884 1885 1886 1887 1888 1889 1890 1891 1892 1893 1894 1895 1896 1897 1898 1899 1900 1901 1902 1903 1904 1905 1906 1907 1908 1909 1910 1911 1912 1913 1914 1915 1916 1917 1918 1919 1920 1921 1922 1923 1924 1925 1926 1927 1928 1929 1930 1931 1932 1933 1934 1935 1936 1937 1938 1939 1940 1941 1942 1943 1944 1945 1946
						       loc_port, loc_addr,
						       rmt_port, rmt_addr,
						       dif, hnum))) {
			sock_put(result);
			result = NULL;
		}
	}
	rcu_read_unlock();
	return result;
}

/* For unicast we should only early demux connected sockets or we can
 * break forwarding setups.  The chains here can be long so only check
 * if the first socket is an exact match and if not move on.
 */
static struct sock *__udp4_lib_demux_lookup(struct net *net,
					    __be16 loc_port, __be32 loc_addr,
					    __be16 rmt_port, __be32 rmt_addr,
					    int dif)
{
	struct sock *sk, *result;
	struct hlist_nulls_node *node;
	unsigned short hnum = ntohs(loc_port);
	unsigned int hash2 = udp4_portaddr_hash(net, loc_addr, hnum);
	unsigned int slot2 = hash2 & udp_table.mask;
	struct udp_hslot *hslot2 = &udp_table.hash2[slot2];
	INET_ADDR_COOKIE(acookie, rmt_addr, loc_addr)
	const __portpair ports = INET_COMBINED_PORTS(rmt_port, hnum);

	rcu_read_lock();
	result = NULL;
	udp_portaddr_for_each_entry_rcu(sk, node, &hslot2->head) {
		if (INET_MATCH(sk, net, acookie,
			       rmt_addr, loc_addr, ports, dif))
			result = sk;
		/* Only check first socket in chain */
		break;
	}

	if (result) {
		if (unlikely(!atomic_inc_not_zero_hint(&result->sk_refcnt, 2)))
			result = NULL;
		else if (unlikely(!INET_MATCH(sk, net, acookie,
					      rmt_addr, loc_addr,
					      ports, dif))) {
			sock_put(result);
			result = NULL;
		}
	}
	rcu_read_unlock();
	return result;
}

void udp_v4_early_demux(struct sk_buff *skb)
{
	const struct iphdr *iph = ip_hdr(skb);
	const struct udphdr *uh = udp_hdr(skb);
	struct sock *sk;
	struct dst_entry *dst;
	struct net *net = dev_net(skb->dev);
	int dif = skb->dev->ifindex;

	/* validate the packet */
	if (!pskb_may_pull(skb, skb_transport_offset(skb) + sizeof(struct udphdr)))
		return;

	if (skb->pkt_type == PACKET_BROADCAST ||
	    skb->pkt_type == PACKET_MULTICAST)
		sk = __udp4_lib_mcast_demux_lookup(net, uh->dest, iph->daddr,
						   uh->source, iph->saddr, dif);
	else if (skb->pkt_type == PACKET_HOST)
		sk = __udp4_lib_demux_lookup(net, uh->dest, iph->daddr,
					     uh->source, iph->saddr, dif);
	else
		return;

	if (!sk)
		return;

	skb->sk = sk;
	skb->destructor = sock_edemux;
	dst = sk->sk_rx_dst;

	if (dst)
		dst = dst_check(dst, 0);
	if (dst)
		skb_dst_set_noref(skb, dst);
}

1947 1948
int udp_rcv(struct sk_buff *skb)
{
1949
	return __udp4_lib_rcv(skb, &udp_table, IPPROTO_UDP);
1950 1951
}

1952
void udp_destroy_sock(struct sock *sk)
1953
{
T
Tom Parkin 已提交
1954
	struct udp_sock *up = udp_sk(sk);
1955
	bool slow = lock_sock_fast(sk);
1956
	udp_flush_pending_frames(sk);
1957
	unlock_sock_fast(sk, slow);
T
Tom Parkin 已提交
1958 1959 1960 1961 1962 1963
	if (static_key_false(&udp_encap_needed) && up->encap_type) {
		void (*encap_destroy)(struct sock *sk);
		encap_destroy = ACCESS_ONCE(up->encap_destroy);
		if (encap_destroy)
			encap_destroy(sk);
	}
1964 1965
}

L
Linus Torvalds 已提交
1966 1967 1968
/*
 *	Socket option code for UDP
 */
1969
int udp_lib_setsockopt(struct sock *sk, int level, int optname,
1970
		       char __user *optval, unsigned int optlen,
1971
		       int (*push_pending_frames)(struct sock *))
L
Linus Torvalds 已提交
1972 1973 1974 1975
{
	struct udp_sock *up = udp_sk(sk);
	int val;
	int err = 0;
W
Wang Chen 已提交
1976
	int is_udplite = IS_UDPLITE(sk);
L
Linus Torvalds 已提交
1977

E
Eric Dumazet 已提交
1978
	if (optlen < sizeof(int))
L
Linus Torvalds 已提交
1979 1980 1981 1982 1983
		return -EINVAL;

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

1984
	switch (optname) {
L
Linus Torvalds 已提交
1985 1986 1987 1988 1989 1990
	case UDP_CORK:
		if (val != 0) {
			up->corkflag = 1;
		} else {
			up->corkflag = 0;
			lock_sock(sk);
1991
			(*push_pending_frames)(sk);
L
Linus Torvalds 已提交
1992 1993 1994
			release_sock(sk);
		}
		break;
1995

L
Linus Torvalds 已提交
1996 1997 1998 1999 2000
	case UDP_ENCAP:
		switch (val) {
		case 0:
		case UDP_ENCAP_ESPINUDP:
		case UDP_ENCAP_ESPINUDP_NON_IKE:
2001 2002
			up->encap_rcv = xfrm4_udp_encap_rcv;
			/* FALLTHROUGH */
2003
		case UDP_ENCAP_L2TPINUDP:
L
Linus Torvalds 已提交
2004
			up->encap_type = val;
2005
			udp_encap_enable();
L
Linus Torvalds 已提交
2006 2007 2008 2009 2010 2011 2012
			break;
		default:
			err = -ENOPROTOOPT;
			break;
		}
		break;

2013 2014 2015 2016 2017 2018
	/*
	 * 	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
2019
		if (!is_udplite)         /* Disable the option on UDP sockets */
2020 2021 2022
			return -ENOPROTOOPT;
		if (val != 0 && val < 8) /* Illegal coverage: use default (8) */
			val = 8;
2023 2024
		else if (val > USHRT_MAX)
			val = USHRT_MAX;
2025 2026 2027 2028
		up->pcslen = val;
		up->pcflag |= UDPLITE_SEND_CC;
		break;

2029 2030
	/* 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
2031 2032
	 * used, this again means full checksum coverage.                     */
	case UDPLITE_RECV_CSCOV:
W
Wang Chen 已提交
2033
		if (!is_udplite)         /* Disable the option on UDP sockets */
2034 2035 2036
			return -ENOPROTOOPT;
		if (val != 0 && val < 8) /* Avoid silly minimal values.       */
			val = 8;
2037 2038
		else if (val > USHRT_MAX)
			val = USHRT_MAX;
2039 2040 2041 2042
		up->pcrlen = val;
		up->pcflag |= UDPLITE_RECV_CC;
		break;

L
Linus Torvalds 已提交
2043 2044 2045
	default:
		err = -ENOPROTOOPT;
		break;
2046
	}
L
Linus Torvalds 已提交
2047 2048 2049

	return err;
}
E
Eric Dumazet 已提交
2050
EXPORT_SYMBOL(udp_lib_setsockopt);
L
Linus Torvalds 已提交
2051

2052
int udp_setsockopt(struct sock *sk, int level, int optname,
2053
		   char __user *optval, unsigned int optlen)
2054 2055 2056 2057 2058 2059 2060 2061 2062
{
	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,
2063
			  char __user *optval, unsigned int optlen)
2064 2065 2066 2067 2068 2069 2070 2071
{
	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

2072 2073
int udp_lib_getsockopt(struct sock *sk, int level, int optname,
		       char __user *optval, int __user *optlen)
L
Linus Torvalds 已提交
2074 2075 2076 2077
{
	struct udp_sock *up = udp_sk(sk);
	int val, len;

E
Eric Dumazet 已提交
2078
	if (get_user(len, optlen))
L
Linus Torvalds 已提交
2079 2080 2081
		return -EFAULT;

	len = min_t(unsigned int, len, sizeof(int));
2082

2083
	if (len < 0)
L
Linus Torvalds 已提交
2084 2085
		return -EINVAL;

2086
	switch (optname) {
L
Linus Torvalds 已提交
2087 2088 2089 2090 2091 2092 2093 2094
	case UDP_CORK:
		val = up->corkflag;
		break;

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

2095 2096 2097 2098 2099 2100 2101 2102 2103 2104
	/* 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
2105 2106
	default:
		return -ENOPROTOOPT;
2107
	}
L
Linus Torvalds 已提交
2108

2109
	if (put_user(len, optlen))
2110
		return -EFAULT;
E
Eric Dumazet 已提交
2111
	if (copy_to_user(optval, &val, len))
L
Linus Torvalds 已提交
2112
		return -EFAULT;
2113
	return 0;
L
Linus Torvalds 已提交
2114
}
E
Eric Dumazet 已提交
2115
EXPORT_SYMBOL(udp_lib_getsockopt);
L
Linus Torvalds 已提交
2116

2117 2118 2119 2120 2121 2122 2123 2124 2125 2126 2127 2128 2129 2130 2131 2132 2133
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
2134 2135 2136 2137 2138 2139
/**
 * 	udp_poll - wait for a UDP event.
 *	@file - file struct
 *	@sock - socket
 *	@wait - poll table
 *
2140
 *	This is same as datagram poll, except for the special case of
L
Linus Torvalds 已提交
2141 2142 2143 2144 2145 2146 2147 2148 2149 2150
 *	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;
2151

2152 2153
	sock_rps_record_flow(sk);

L
Linus Torvalds 已提交
2154
	/* Check for false positives due to checksum errors */
E
Eric Dumazet 已提交
2155 2156 2157
	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
2158 2159

	return mask;
2160

L
Linus Torvalds 已提交
2161
}
E
Eric Dumazet 已提交
2162
EXPORT_SYMBOL(udp_poll);
L
Linus Torvalds 已提交
2163

2164 2165 2166 2167 2168 2169 2170 2171 2172 2173 2174 2175 2176
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
2177
	.backlog_rcv	   = __udp_queue_rcv_skb,
2178
	.release_cb	   = ip4_datagram_release_cb,
2179 2180
	.hash		   = udp_lib_hash,
	.unhash		   = udp_lib_unhash,
E
Eric Dumazet 已提交
2181
	.rehash		   = udp_v4_rehash,
2182 2183 2184 2185 2186 2187
	.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),
2188
	.slab_flags	   = SLAB_DESTROY_BY_RCU,
2189
	.h.udp_table	   = &udp_table,
2190 2191 2192 2193
#ifdef CONFIG_COMPAT
	.compat_setsockopt = compat_udp_setsockopt,
	.compat_getsockopt = compat_udp_getsockopt,
#endif
2194
	.clear_sk	   = sk_prot_clear_portaddr_nulls,
2195
};
E
Eric Dumazet 已提交
2196
EXPORT_SYMBOL(udp_prot);
L
Linus Torvalds 已提交
2197 2198 2199 2200

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

2201
static struct sock *udp_get_first(struct seq_file *seq, int start)
L
Linus Torvalds 已提交
2202 2203 2204
{
	struct sock *sk;
	struct udp_iter_state *state = seq->private;
2205
	struct net *net = seq_file_net(seq);
L
Linus Torvalds 已提交
2206

2207 2208
	for (state->bucket = start; state->bucket <= state->udp_table->mask;
	     ++state->bucket) {
2209
		struct hlist_nulls_node *node;
2210
		struct udp_hslot *hslot = &state->udp_table->hash[state->bucket];
2211 2212 2213 2214

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

2215
		spin_lock_bh(&hslot->lock);
2216
		sk_nulls_for_each(sk, node, &hslot->head) {
2217
			if (!net_eq(sock_net(sk), net))
2218
				continue;
L
Linus Torvalds 已提交
2219 2220 2221
			if (sk->sk_family == state->family)
				goto found;
		}
2222
		spin_unlock_bh(&hslot->lock);
L
Linus Torvalds 已提交
2223 2224 2225 2226 2227 2228 2229 2230 2231
	}
	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;
2232
	struct net *net = seq_file_net(seq);
L
Linus Torvalds 已提交
2233 2234

	do {
2235
		sk = sk_nulls_next(sk);
2236
	} while (sk && (!net_eq(sock_net(sk), net) || sk->sk_family != state->family));
L
Linus Torvalds 已提交
2237

2238
	if (!sk) {
2239
		if (state->bucket <= state->udp_table->mask)
2240
			spin_unlock_bh(&state->udp_table->hash[state->bucket].lock);
2241
		return udp_get_first(seq, state->bucket + 1);
L
Linus Torvalds 已提交
2242 2243 2244 2245 2246 2247
	}
	return sk;
}

static struct sock *udp_get_idx(struct seq_file *seq, loff_t pos)
{
2248
	struct sock *sk = udp_get_first(seq, 0);
L
Linus Torvalds 已提交
2249 2250

	if (sk)
2251
		while (pos && (sk = udp_get_next(seq, sk)) != NULL)
L
Linus Torvalds 已提交
2252 2253 2254 2255 2256 2257
			--pos;
	return pos ? NULL : sk;
}

static void *udp_seq_start(struct seq_file *seq, loff_t *pos)
{
2258
	struct udp_iter_state *state = seq->private;
2259
	state->bucket = MAX_UDP_PORTS;
2260

2261
	return *pos ? udp_get_idx(seq, *pos-1) : SEQ_START_TOKEN;
L
Linus Torvalds 已提交
2262 2263 2264 2265 2266 2267
}

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

2268
	if (v == SEQ_START_TOKEN)
L
Linus Torvalds 已提交
2269 2270 2271 2272 2273 2274 2275 2276 2277 2278
		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)
{
2279 2280
	struct udp_iter_state *state = seq->private;

2281
	if (state->bucket <= state->udp_table->mask)
2282
		spin_unlock_bh(&state->udp_table->hash[state->bucket].lock);
L
Linus Torvalds 已提交
2283 2284
}

2285
int udp_seq_open(struct inode *inode, struct file *file)
L
Linus Torvalds 已提交
2286
{
A
Al Viro 已提交
2287
	struct udp_seq_afinfo *afinfo = PDE_DATA(inode);
2288 2289
	struct udp_iter_state *s;
	int err;
2290

2291 2292 2293 2294
	err = seq_open_net(inode, file, &afinfo->seq_ops,
			   sizeof(struct udp_iter_state));
	if (err < 0)
		return err;
2295

2296
	s = ((struct seq_file *)file->private_data)->private;
L
Linus Torvalds 已提交
2297
	s->family		= afinfo->family;
2298
	s->udp_table		= afinfo->udp_table;
2299
	return err;
2300
}
2301
EXPORT_SYMBOL(udp_seq_open);
2302

L
Linus Torvalds 已提交
2303
/* ------------------------------------------------------------------------ */
2304
int udp_proc_register(struct net *net, struct udp_seq_afinfo *afinfo)
L
Linus Torvalds 已提交
2305 2306 2307 2308
{
	struct proc_dir_entry *p;
	int rc = 0;

2309 2310 2311 2312
	afinfo->seq_ops.start		= udp_seq_start;
	afinfo->seq_ops.next		= udp_seq_next;
	afinfo->seq_ops.stop		= udp_seq_stop;

2313
	p = proc_create_data(afinfo->name, S_IRUGO, net->proc_net,
2314
			     afinfo->seq_fops, afinfo);
2315
	if (!p)
L
Linus Torvalds 已提交
2316 2317 2318
		rc = -ENOMEM;
	return rc;
}
E
Eric Dumazet 已提交
2319
EXPORT_SYMBOL(udp_proc_register);
L
Linus Torvalds 已提交
2320

2321
void udp_proc_unregister(struct net *net, struct udp_seq_afinfo *afinfo)
L
Linus Torvalds 已提交
2322
{
2323
	remove_proc_entry(afinfo->name, net->proc_net);
L
Linus Torvalds 已提交
2324
}
E
Eric Dumazet 已提交
2325
EXPORT_SYMBOL(udp_proc_unregister);
2326 2327

/* ------------------------------------------------------------------------ */
2328
static void udp4_format_sock(struct sock *sp, struct seq_file *f,
2329
		int bucket)
2330 2331
{
	struct inet_sock *inet = inet_sk(sp);
E
Eric Dumazet 已提交
2332 2333 2334 2335
	__be32 dest = inet->inet_daddr;
	__be32 src  = inet->inet_rcv_saddr;
	__u16 destp	  = ntohs(inet->inet_dport);
	__u16 srcp	  = ntohs(inet->inet_sport);
2336

2337
	seq_printf(f, "%5d: %08X:%04X %08X:%04X"
2338
		" %02X %08X:%08X %02X:%08lX %08X %5u %8d %lu %d %pK %d",
2339
		bucket, src, srcp, dest, destp, sp->sk_state,
2340 2341
		sk_wmem_alloc_get(sp),
		sk_rmem_alloc_get(sp),
2342 2343 2344
		0, 0L, 0,
		from_kuid_munged(seq_user_ns(f), sock_i_uid(sp)),
		0, sock_i_ino(sp),
E
Eric Dumazet 已提交
2345
		atomic_read(&sp->sk_refcnt), sp,
2346
		atomic_read(&sp->sk_drops));
2347 2348 2349 2350
}

int udp4_seq_show(struct seq_file *seq, void *v)
{
2351
	seq_setwidth(seq, 127);
2352
	if (v == SEQ_START_TOKEN)
2353
		seq_puts(seq, "  sl  local_address rem_address   st tx_queue "
2354
			   "rx_queue tr tm->when retrnsmt   uid  timeout "
E
Eric Dumazet 已提交
2355
			   "inode ref pointer drops");
2356 2357 2358
	else {
		struct udp_iter_state *state = seq->private;

2359
		udp4_format_sock(v, seq, state->bucket);
2360
	}
2361
	seq_pad(seq, '\n');
2362 2363 2364
	return 0;
}

2365 2366 2367 2368 2369 2370 2371 2372
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
};

2373 2374 2375 2376
/* ------------------------------------------------------------------------ */
static struct udp_seq_afinfo udp4_seq_afinfo = {
	.name		= "udp",
	.family		= AF_INET,
2377
	.udp_table	= &udp_table,
2378
	.seq_fops	= &udp_afinfo_seq_fops,
2379 2380 2381
	.seq_ops	= {
		.show		= udp4_seq_show,
	},
2382 2383
};

2384
static int __net_init udp4_proc_init_net(struct net *net)
2385 2386 2387 2388
{
	return udp_proc_register(net, &udp4_seq_afinfo);
}

2389
static void __net_exit udp4_proc_exit_net(struct net *net)
2390 2391 2392 2393 2394 2395 2396 2397 2398
{
	udp_proc_unregister(net, &udp4_seq_afinfo);
}

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

2399 2400
int __init udp4_proc_init(void)
{
2401
	return register_pernet_subsys(&udp4_net_ops);
2402 2403 2404 2405
}

void udp4_proc_exit(void)
{
2406
	unregister_pernet_subsys(&udp4_net_ops);
2407
}
L
Linus Torvalds 已提交
2408 2409
#endif /* CONFIG_PROC_FS */

2410 2411
static __initdata unsigned long uhash_entries;
static int __init set_uhash_entries(char *str)
2412
{
2413 2414
	ssize_t ret;

2415 2416
	if (!str)
		return 0;
2417 2418 2419 2420 2421

	ret = kstrtoul(str, 0, &uhash_entries);
	if (ret)
		return 0;

2422 2423 2424 2425 2426
	if (uhash_entries && uhash_entries < UDP_HTABLE_SIZE_MIN)
		uhash_entries = UDP_HTABLE_SIZE_MIN;
	return 1;
}
__setup("uhash_entries=", set_uhash_entries);
2427

2428 2429 2430 2431
void __init udp_table_init(struct udp_table *table, const char *name)
{
	unsigned int i;

2432 2433 2434 2435 2436 2437 2438 2439 2440 2441
	table->hash = alloc_large_system_hash(name,
					      2 * sizeof(struct udp_hslot),
					      uhash_entries,
					      21, /* one slot per 2 MB */
					      0,
					      &table->log,
					      &table->mask,
					      UDP_HTABLE_SIZE_MIN,
					      64 * 1024);

2442
	table->hash2 = table->hash + (table->mask + 1);
2443
	for (i = 0; i <= table->mask; i++) {
2444
		INIT_HLIST_NULLS_HEAD(&table->hash[i].head, i);
E
Eric Dumazet 已提交
2445
		table->hash[i].count = 0;
2446 2447
		spin_lock_init(&table->hash[i].lock);
	}
2448 2449 2450 2451 2452
	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);
	}
2453 2454
}

H
Hideo Aoki 已提交
2455 2456
void __init udp_init(void)
{
2457
	unsigned long limit;
H
Hideo Aoki 已提交
2458

2459
	udp_table_init(&udp_table, "UDP");
2460
	limit = nr_free_buffer_pages() / 8;
H
Hideo Aoki 已提交
2461 2462 2463 2464 2465 2466 2467 2468 2469
	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;
}

2470 2471
struct sk_buff *skb_udp_tunnel_segment(struct sk_buff *skb,
				       netdev_features_t features)
2472 2473 2474 2475
{
	struct sk_buff *segs = ERR_PTR(-EINVAL);
	int mac_len = skb->mac_len;
	int tnl_hlen = skb_inner_mac_header(skb) - skb_transport_header(skb);
P
Pravin B Shelar 已提交
2476
	__be16 protocol = skb->protocol;
2477
	netdev_features_t enc_features;
P
Pravin B Shelar 已提交
2478
	int outer_hlen;
2479 2480 2481 2482 2483 2484 2485 2486 2487

	if (unlikely(!pskb_may_pull(skb, tnl_hlen)))
		goto out;

	skb->encapsulation = 0;
	__skb_pull(skb, tnl_hlen);
	skb_reset_mac_header(skb);
	skb_set_network_header(skb, skb_inner_network_offset(skb));
	skb->mac_len = skb_inner_network_offset(skb);
2488
	skb->protocol = htons(ETH_P_TEB);
2489 2490 2491 2492 2493 2494 2495 2496 2497 2498 2499 2500 2501

	/* segment inner packet. */
	enc_features = skb->dev->hw_enc_features & netif_skb_features(skb);
	segs = skb_mac_gso_segment(skb, enc_features);
	if (!segs || IS_ERR(segs))
		goto out;

	outer_hlen = skb_tnl_header_len(skb);
	skb = segs;
	do {
		struct udphdr *uh;
		int udp_offset = outer_hlen - tnl_hlen;

2502 2503 2504
		skb_reset_inner_headers(skb);
		skb->encapsulation = 1;

2505 2506 2507 2508 2509 2510 2511 2512 2513 2514
		skb->mac_len = mac_len;

		skb_push(skb, outer_hlen);
		skb_reset_mac_header(skb);
		skb_set_network_header(skb, mac_len);
		skb_set_transport_header(skb, udp_offset);
		uh = udp_hdr(skb);
		uh->len = htons(skb->len - udp_offset);

		/* csum segment if tunnel sets skb with csum. */
2515
		if (protocol == htons(ETH_P_IP) && unlikely(uh->check)) {
2516 2517 2518 2519 2520 2521 2522 2523 2524 2525
			struct iphdr *iph = ip_hdr(skb);

			uh->check = ~csum_tcpudp_magic(iph->saddr, iph->daddr,
						       skb->len - udp_offset,
						       IPPROTO_UDP, 0);
			uh->check = csum_fold(skb_checksum(skb, udp_offset,
							   skb->len - udp_offset, 0));
			if (uh->check == 0)
				uh->check = CSUM_MANGLED_0;

2526 2527 2528 2529 2530 2531 2532 2533 2534 2535
		} else if (protocol == htons(ETH_P_IPV6)) {
			struct ipv6hdr *ipv6h = ipv6_hdr(skb);
			u32 len = skb->len - udp_offset;

			uh->check = ~csum_ipv6_magic(&ipv6h->saddr, &ipv6h->daddr,
						     len, IPPROTO_UDP, 0);
			uh->check = csum_fold(skb_checksum(skb, udp_offset, len, 0));
			if (uh->check == 0)
				uh->check = CSUM_MANGLED_0;
			skb->ip_summed = CHECKSUM_NONE;
2536
		}
2537

P
Pravin B Shelar 已提交
2538
		skb->protocol = protocol;
2539 2540 2541 2542
	} while ((skb = skb->next));
out:
	return segs;
}