udp.c 63.7 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
#include <linux/skbuff.h>
T
Tom Herbert 已提交
102
#include <linux/netdevice.h>
L
Linus Torvalds 已提交
103 104
#include <linux/proc_fs.h>
#include <linux/seq_file.h>
105
#include <net/net_namespace.h>
L
Linus Torvalds 已提交
106
#include <net/icmp.h>
S
Shawn Bohrer 已提交
107
#include <net/inet_hashtables.h>
L
Linus Torvalds 已提交
108 109 110
#include <net/route.h>
#include <net/checksum.h>
#include <net/xfrm.h>
111
#include <trace/events/udp.h>
112
#include <linux/static_key.h>
113
#include <trace/events/skb.h>
114
#include <net/busy_poll.h>
115
#include "udp_impl.h"
L
Linus Torvalds 已提交
116

117
struct udp_table udp_table __read_mostly;
118
EXPORT_SYMBOL(udp_table);
L
Linus Torvalds 已提交
119

E
Eric Dumazet 已提交
120
long sysctl_udp_mem[3] __read_mostly;
H
Hideo Aoki 已提交
121
EXPORT_SYMBOL(sysctl_udp_mem);
E
Eric Dumazet 已提交
122 123

int sysctl_udp_rmem_min __read_mostly;
H
Hideo Aoki 已提交
124
EXPORT_SYMBOL(sysctl_udp_rmem_min);
E
Eric Dumazet 已提交
125 126

int sysctl_udp_wmem_min __read_mostly;
H
Hideo Aoki 已提交
127 128
EXPORT_SYMBOL(sysctl_udp_wmem_min);

E
Eric Dumazet 已提交
129
atomic_long_t udp_memory_allocated;
H
Hideo Aoki 已提交
130 131
EXPORT_SYMBOL(udp_memory_allocated);

132 133
#define MAX_UDP_PORTS 65536
#define PORTS_PER_CHAIN (MAX_UDP_PORTS / UDP_HTABLE_SIZE_MIN)
134

135
static int udp_lib_lport_inuse(struct net *net, __u16 num,
136
			       const struct udp_hslot *hslot,
137
			       unsigned long *bitmap,
138 139
			       struct sock *sk,
			       int (*saddr_comp)(const struct sock *sk1,
140 141
						 const struct sock *sk2),
			       unsigned int log)
L
Linus Torvalds 已提交
142
{
143
	struct sock *sk2;
144
	struct hlist_nulls_node *node;
145
	kuid_t uid = sock_i_uid(sk);
146

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

E
Eric Dumazet 已提交
165 166 167 168 169
/*
 * 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,
170 171 172 173
				struct udp_hslot *hslot2,
				struct sock *sk,
				int (*saddr_comp)(const struct sock *sk1,
						  const struct sock *sk2))
E
Eric Dumazet 已提交
174 175 176
{
	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
	int res = 0;

	spin_lock(&hslot2->lock);
181
	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
		    (!sk2->sk_reuseport || !sk->sk_reuseport ||
189 190
		     !uid_eq(uid, sock_i_uid(sk2))) &&
		    saddr_comp(sk, sk2)) {
E
Eric Dumazet 已提交
191 192 193
			res = 1;
			break;
		}
194
	}
E
Eric Dumazet 已提交
195 196 197 198
	spin_unlock(&hslot2->lock);
	return res;
}

199
/**
200
 *  udp_lib_get_port  -  UDP/-Lite port lookup for IPv4 and IPv6
201 202 203
 *
 *  @sk:          socket struct in question
 *  @snum:        port number to look up
204
 *  @saddr_comp:  AF-dependent comparison of bound local IP addresses
L
Lucas De Marchi 已提交
205
 *  @hash2_nulladdr: AF-dependent hash value in secondary hash chains,
E
Eric Dumazet 已提交
206
 *                   with NULL address
207
 */
208
int udp_lib_get_port(struct sock *sk, unsigned short snum,
209 210
		     int (*saddr_comp)(const struct sock *sk1,
				       const struct sock *sk2),
E
Eric Dumazet 已提交
211
		     unsigned int hash2_nulladdr)
212
{
213
	struct udp_hslot *hslot, *hslot2;
214
	struct udp_table *udptable = sk->sk_prot->h.udp_table;
215
	int    error = 1;
216
	struct net *net = sock_net(sk);
L
Linus Torvalds 已提交
217

218
	if (!snum) {
E
Eric Dumazet 已提交
219
		int low, high, remaining;
220
		unsigned int rand;
221 222
		unsigned short first, last;
		DECLARE_BITMAP(bitmap, PORTS_PER_CHAIN);
223

224
		inet_get_local_port_range(net, &low, &high);
225
		remaining = (high - low) + 1;
226

227
		rand = prandom_u32();
228
		first = reciprocal_scale(rand, remaining) + low;
229 230 231
		/*
		 * force rand to be an odd multiple of UDP_HTABLE_SIZE
		 */
232
		rand = (rand | 1) * (udptable->mask + 1);
E
Eric Dumazet 已提交
233 234
		last = first + udptable->mask + 1;
		do {
235
			hslot = udp_hashslot(udptable, net, first);
236
			bitmap_zero(bitmap, PORTS_PER_CHAIN);
237
			spin_lock_bh(&hslot->lock);
238
			udp_lib_lport_inuse(net, snum, hslot, bitmap, sk,
239
					    saddr_comp, udptable->log);
240 241 242 243 244 245 246

			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
247
			do {
248
				if (low <= snum && snum <= high &&
249
				    !test_bit(snum >> udptable->log, bitmap) &&
250
				    !inet_is_local_reserved_port(net, snum))
251 252 253 254
					goto found;
				snum += rand;
			} while (snum != first);
			spin_unlock_bh(&hslot->lock);
E
Eric Dumazet 已提交
255
		} while (++first != last);
256
		goto fail;
257
	} else {
258
		hslot = udp_hashslot(udptable, net, snum);
259
		spin_lock_bh(&hslot->lock);
E
Eric Dumazet 已提交
260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283
		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:
284 285
		if (udp_lib_lport_inuse(net, snum, hslot, NULL, sk,
					saddr_comp, 0))
286 287
			goto fail_unlock;
	}
288
found:
E
Eric Dumazet 已提交
289
	inet_sk(sk)->inet_num = snum;
290 291
	udp_sk(sk)->udp_port_hash = snum;
	udp_sk(sk)->udp_portaddr_hash ^= snum;
L
Linus Torvalds 已提交
292
	if (sk_unhashed(sk)) {
293
		sk_nulls_add_node_rcu(sk, &hslot->head);
E
Eric Dumazet 已提交
294
		hslot->count++;
295
		sock_prot_inuse_add(sock_net(sk), sk->sk_prot, 1);
296 297 298 299 300 301 302

		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
303
	}
304
	error = 0;
305 306
fail_unlock:
	spin_unlock_bh(&hslot->lock);
L
Linus Torvalds 已提交
307
fail:
308 309
	return error;
}
E
Eric Dumazet 已提交
310
EXPORT_SYMBOL(udp_lib_get_port);
311

312
static int ipv4_rcv_saddr_equal(const struct sock *sk1, const struct sock *sk2)
313 314 315
{
	struct inet_sock *inet1 = inet_sk(sk1), *inet2 = inet_sk(sk2);

E
Eric Dumazet 已提交
316
	return 	(!ipv6_only_sock(sk2)  &&
E
Eric Dumazet 已提交
317 318
		 (!inet1->inet_rcv_saddr || !inet2->inet_rcv_saddr ||
		   inet1->inet_rcv_saddr == inet2->inet_rcv_saddr));
319 320
}

321 322 323
static unsigned int udp4_portaddr_hash(struct net *net, __be32 saddr,
				       unsigned int port)
{
324
	return jhash_1word((__force u32)saddr, net_hash_mix(net)) ^ port;
325 326
}

327
int udp_v4_get_port(struct sock *sk, unsigned short snum)
328
{
E
Eric Dumazet 已提交
329
	unsigned int hash2_nulladdr =
330
		udp4_portaddr_hash(sock_net(sk), htonl(INADDR_ANY), snum);
E
Eric Dumazet 已提交
331 332 333
	unsigned int hash2_partial =
		udp4_portaddr_hash(sock_net(sk), inet_sk(sk)->inet_rcv_saddr, 0);

334
	/* precompute partial secondary hash */
E
Eric Dumazet 已提交
335 336
	udp_sk(sk)->udp_portaddr_hash = hash2_partial;
	return udp_lib_get_port(sk, snum, ipv4_rcv_saddr_equal, hash2_nulladdr);
337 338
}

339 340 341
static inline int compute_score(struct sock *sk, struct net *net,
				__be32 saddr, unsigned short hnum, __be16 sport,
				__be32 daddr, __be16 dport, int dif)
342
{
343 344
	int score;
	struct inet_sock *inet;
345

346 347 348 349
	if (!net_eq(sock_net(sk), net) ||
	    udp_sk(sk)->udp_port_hash != hnum ||
	    ipv6_only_sock(sk))
		return -1;
350

351 352 353 354 355 356 357
	score = (sk->sk_family == PF_INET) ? 2 : 1;
	inet = inet_sk(sk);

	if (inet->inet_rcv_saddr) {
		if (inet->inet_rcv_saddr != daddr)
			return -1;
		score += 4;
358
	}
359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377

	if (inet->inet_daddr) {
		if (inet->inet_daddr != saddr)
			return -1;
		score += 4;
	}

	if (inet->inet_dport) {
		if (inet->inet_dport != sport)
			return -1;
		score += 4;
	}

	if (sk->sk_bound_dev_if) {
		if (sk->sk_bound_dev_if != dif)
			return -1;
		score += 4;
	}

378 379 380
	return score;
}

E
Eric Dumazet 已提交
381 382 383 384 385 386 387
/*
 * 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)
{
388 389
	int score;
	struct inet_sock *inet;
E
Eric Dumazet 已提交
390

391 392 393
	if (!net_eq(sock_net(sk), net) ||
	    ipv6_only_sock(sk))
		return -1;
E
Eric Dumazet 已提交
394

395 396 397 398 399 400 401 402 403 404
	inet = inet_sk(sk);

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

	score = (sk->sk_family == PF_INET) ? 2 : 1;

	if (inet->inet_daddr) {
		if (inet->inet_daddr != saddr)
E
Eric Dumazet 已提交
405
			return -1;
406 407 408 409 410
		score += 4;
	}

	if (inet->inet_dport) {
		if (inet->inet_dport != sport)
E
Eric Dumazet 已提交
411
			return -1;
412 413
		score += 4;
	}
E
Eric Dumazet 已提交
414

415 416 417 418
	if (sk->sk_bound_dev_if) {
		if (sk->sk_bound_dev_if != dif)
			return -1;
		score += 4;
E
Eric Dumazet 已提交
419
	}
420

E
Eric Dumazet 已提交
421 422 423
	return score;
}

424 425 426 427
static unsigned int udp_ehashfn(struct net *net, const __be32 laddr,
				 const __u16 lport, const __be32 faddr,
				 const __be16 fport)
{
428 429 430 431
	static u32 udp_ehash_secret __read_mostly;

	net_get_random_once(&udp_ehash_secret, sizeof(udp_ehash_secret));

432
	return __inet_ehashfn(laddr, lport, faddr, fport,
433
			      udp_ehash_secret + net_hash_mix(net));
434 435
}

E
Eric Dumazet 已提交
436 437 438 439 440 441 442 443 444

/* 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;
445 446
	int score, badness, matches = 0, reuseport = 0;
	u32 hash = 0;
E
Eric Dumazet 已提交
447 448 449

begin:
	result = NULL;
450
	badness = 0;
E
Eric Dumazet 已提交
451 452 453 454 455 456
	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;
457 458
			reuseport = sk->sk_reuseport;
			if (reuseport) {
459 460
				hash = udp_ehashfn(net, daddr, hnum,
						   saddr, sport);
461 462 463 464
				matches = 1;
			}
		} else if (score == badness && reuseport) {
			matches++;
465
			if (reciprocal_scale(hash, matches) == 0)
466 467
				result = sk;
			hash = next_pseudo_random32(hash);
E
Eric Dumazet 已提交
468 469 470 471 472 473 474 475 476 477
		}
	}
	/*
	 * 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
478
		if (unlikely(!atomic_inc_not_zero_hint(&result->sk_refcnt, 2)))
E
Eric Dumazet 已提交
479 480 481 482 483 484 485 486 487 488
			result = NULL;
		else if (unlikely(compute_score2(result, net, saddr, sport,
				  daddr, hnum, dif) < badness)) {
			sock_put(result);
			goto begin;
		}
	}
	return result;
}

489 490 491
/* UDP is nearly always wildcards out the wazoo, it makes no sense to try
 * harder than this. -DaveM
 */
492
struct sock *__udp4_lib_lookup(struct net *net, __be32 saddr,
493
		__be16 sport, __be32 daddr, __be16 dport,
494
		int dif, struct udp_table *udptable)
495
{
496
	struct sock *sk, *result;
497
	struct hlist_nulls_node *node;
498
	unsigned short hnum = ntohs(dport);
E
Eric Dumazet 已提交
499 500
	unsigned int hash2, slot2, slot = udp_hashfn(net, hnum, udptable->mask);
	struct udp_hslot *hslot2, *hslot = &udptable->hash[slot];
501 502
	int score, badness, matches = 0, reuseport = 0;
	u32 hash = 0;
503

504
	rcu_read_lock();
E
Eric Dumazet 已提交
505 506 507 508 509 510 511 512 513 514 515
	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) {
516
			hash2 = udp4_portaddr_hash(net, htonl(INADDR_ANY), hnum);
E
Eric Dumazet 已提交
517 518 519 520 521
			slot2 = hash2 & udptable->mask;
			hslot2 = &udptable->hash2[slot2];
			if (hslot->count < hslot2->count)
				goto begin;

522
			result = udp4_lib_lookup2(net, saddr, sport,
523
						  htonl(INADDR_ANY), hnum, dif,
E
Eric Dumazet 已提交
524 525 526 527 528
						  hslot2, slot2);
		}
		rcu_read_unlock();
		return result;
	}
529 530
begin:
	result = NULL;
531
	badness = 0;
532
	sk_nulls_for_each_rcu(sk, node, &hslot->head) {
533 534 535 536 537
		score = compute_score(sk, net, saddr, hnum, sport,
				      daddr, dport, dif);
		if (score > badness) {
			result = sk;
			badness = score;
538 539
			reuseport = sk->sk_reuseport;
			if (reuseport) {
540 541
				hash = udp_ehashfn(net, daddr, hnum,
						   saddr, sport);
542 543 544 545
				matches = 1;
			}
		} else if (score == badness && reuseport) {
			matches++;
546
			if (reciprocal_scale(hash, matches) == 0)
547 548
				result = sk;
			hash = next_pseudo_random32(hash);
549 550
		}
	}
551 552 553 554 555
	/*
	 * 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
556
	if (get_nulls_value(node) != slot)
557 558
		goto begin;

559
	if (result) {
E
Eric Dumazet 已提交
560
		if (unlikely(!atomic_inc_not_zero_hint(&result->sk_refcnt, 2)))
561 562 563 564 565 566 567 568
			result = NULL;
		else if (unlikely(compute_score(result, net, saddr, hnum, sport,
				  daddr, dport, dif) < badness)) {
			sock_put(result);
			goto begin;
		}
	}
	rcu_read_unlock();
569 570
	return result;
}
571
EXPORT_SYMBOL_GPL(__udp4_lib_lookup);
572

573 574
static inline struct sock *__udp4_lib_lookup_skb(struct sk_buff *skb,
						 __be16 sport, __be16 dport,
575
						 struct udp_table *udptable)
576 577 578
{
	const struct iphdr *iph = ip_hdr(skb);

579 580 581
	return __udp4_lib_lookup(dev_net(skb_dst(skb)->dev), iph->saddr, sport,
				 iph->daddr, dport, inet_iif(skb),
				 udptable);
582 583
}

584 585 586
struct sock *udp4_lib_lookup(struct net *net, __be32 saddr, __be16 sport,
			     __be32 daddr, __be16 dport, int dif)
{
587
	return __udp4_lib_lookup(net, saddr, sport, daddr, dport, dif, &udp_table);
588 589 590
}
EXPORT_SYMBOL_GPL(udp4_lib_lookup);

S
Shawn Bohrer 已提交
591 592 593 594 595 596 597 598 599 600 601 602 603 604 605 606 607 608 609 610
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;
}

611 612 613 614 615 616 617 618 619 620 621
/*
 * 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.
 */

622
void __udp4_lib_err(struct sk_buff *skb, u32 info, struct udp_table *udptable)
623 624
{
	struct inet_sock *inet;
625
	const struct iphdr *iph = (const struct iphdr *)skb->data;
E
Eric Dumazet 已提交
626
	struct udphdr *uh = (struct udphdr *)(skb->data+(iph->ihl<<2));
627 628 629 630 631
	const int type = icmp_hdr(skb)->type;
	const int code = icmp_hdr(skb)->code;
	struct sock *sk;
	int harderr;
	int err;
632
	struct net *net = dev_net(skb->dev);
633

634
	sk = __udp4_lib_lookup(net, iph->daddr, uh->dest,
635 636
			iph->saddr, uh->source, skb->dev->ifindex, udptable);
	if (sk == NULL) {
637
		ICMP_INC_STATS_BH(net, ICMP_MIB_INERRORS);
638 639 640 641 642 643 644 645 646 647 648 649 650 651 652 653 654 655 656 657
		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 */
658
			ipv4_sk_update_pmtu(skb, sk, info);
659 660 661 662 663 664 665 666 667 668 669 670 671
			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;
672 673
	case ICMP_REDIRECT:
		ipv4_sk_redirect(skb, sk);
674
		goto out;
675 676 677 678 679 680 681 682 683
	}

	/*
	 *      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;
684
	} else
E
Eric Dumazet 已提交
685
		ip_icmp_error(sk, skb, err, uh->dest, info, (u8 *)(uh+1));
686

687 688 689 690 691 692 693 694
	sk->sk_err = err;
	sk->sk_error_report(sk);
out:
	sock_put(sk);
}

void udp_err(struct sk_buff *skb, u32 info)
{
695
	__udp4_lib_err(skb, info, &udp_table);
696 697 698 699 700
}

/*
 * Throw away all pending data and cancel the corking. Socket is locked.
 */
701
void udp_flush_pending_frames(struct sock *sk)
702 703 704 705 706 707 708 709 710
{
	struct udp_sock *up = udp_sk(sk);

	if (up->pending) {
		up->len = 0;
		up->pending = 0;
		ip_flush_pending_frames(sk);
	}
}
711
EXPORT_SYMBOL(udp_flush_pending_frames);
712 713

/**
H
Herbert Xu 已提交
714
 * 	udp4_hwcsum  -  handle outgoing HW checksumming
715 716
 * 	@skb: 	sk_buff containing the filled-in UDP header
 * 	        (checksum field must be zeroed out)
H
Herbert Xu 已提交
717 718
 *	@src:	source IP address
 *	@dst:	destination IP address
719
 */
720
void udp4_hwcsum(struct sk_buff *skb, __be32 src, __be32 dst)
721 722
{
	struct udphdr *uh = udp_hdr(skb);
H
Herbert Xu 已提交
723 724 725
	int offset = skb_transport_offset(skb);
	int len = skb->len - offset;
	int hlen = len;
726 727
	__wsum csum = 0;

728
	if (!skb_has_frag_list(skb)) {
729 730 731 732 733
		/*
		 * 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
734 735
		uh->check = ~csum_tcpudp_magic(src, dst, len,
					       IPPROTO_UDP, 0);
736
	} else {
737 738
		struct sk_buff *frags;

739 740 741 742 743
		/*
		 * 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
		 */
744
		skb_walk_frags(skb, frags) {
H
Herbert Xu 已提交
745 746
			csum = csum_add(csum, frags->csum);
			hlen -= frags->len;
747
		}
748

H
Herbert Xu 已提交
749
		csum = skb_checksum(skb, offset, hlen, csum);
750 751 752 753 754 755 756
		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;
	}
}
757
EXPORT_SYMBOL_GPL(udp4_hwcsum);
758

759 760 761 762 763 764 765 766 767 768 769 770 771 772 773 774 775 776 777 778 779 780 781 782 783 784 785 786 787 788 789 790 791 792 793 794 795
/* Function to set UDP checksum for an IPv4 UDP packet. This is intended
 * for the simple case like when setting the checksum for a UDP tunnel.
 */
void udp_set_csum(bool nocheck, struct sk_buff *skb,
		  __be32 saddr, __be32 daddr, int len)
{
	struct udphdr *uh = udp_hdr(skb);

	if (nocheck)
		uh->check = 0;
	else if (skb_is_gso(skb))
		uh->check = ~udp_v4_check(len, saddr, daddr, 0);
	else if (skb_dst(skb) && skb_dst(skb)->dev &&
		 (skb_dst(skb)->dev->features & NETIF_F_V4_CSUM)) {

		BUG_ON(skb->ip_summed == CHECKSUM_PARTIAL);

		skb->ip_summed = CHECKSUM_PARTIAL;
		skb->csum_start = skb_transport_header(skb) - skb->head;
		skb->csum_offset = offsetof(struct udphdr, check);
		uh->check = ~udp_v4_check(len, saddr, daddr, 0);
	} else {
		__wsum csum;

		BUG_ON(skb->ip_summed == CHECKSUM_PARTIAL);

		uh->check = 0;
		csum = skb_checksum(skb, 0, len, 0);
		uh->check = udp_v4_check(len, saddr, daddr, csum);
		if (uh->check == 0)
			uh->check = CSUM_MANGLED_0;

		skb->ip_summed = CHECKSUM_UNNECESSARY;
	}
}
EXPORT_SYMBOL(udp_set_csum);

796
static int udp_send_skb(struct sk_buff *skb, struct flowi4 *fl4)
797
{
H
Herbert Xu 已提交
798
	struct sock *sk = skb->sk;
799 800 801 802
	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
803 804
	int offset = skb_transport_offset(skb);
	int len = skb->len - offset;
805 806 807 808 809 810
	__wsum csum = 0;

	/*
	 * Create a UDP header
	 */
	uh = udp_hdr(skb);
H
Herbert Xu 已提交
811
	uh->source = inet->inet_sport;
812
	uh->dest = fl4->fl4_dport;
H
Herbert Xu 已提交
813
	uh->len = htons(len);
814 815 816
	uh->check = 0;

	if (is_udplite)  				 /*     UDP-Lite      */
H
Herbert Xu 已提交
817
		csum = udplite_csum(skb);
818

819
	else if (sk->sk_no_check_tx) {   /* UDP csum disabled */
820 821 822 823 824 825

		skb->ip_summed = CHECKSUM_NONE;
		goto send;

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

826
		udp4_hwcsum(skb, fl4->saddr, fl4->daddr);
827 828
		goto send;

H
Herbert Xu 已提交
829 830
	} else
		csum = udp_csum(skb);
831 832

	/* add protocol-dependent pseudo-header */
833
	uh->check = csum_tcpudp_magic(fl4->saddr, fl4->daddr, len,
E
Eric Dumazet 已提交
834
				      sk->sk_protocol, csum);
835 836 837 838
	if (uh->check == 0)
		uh->check = CSUM_MANGLED_0;

send:
E
Eric Dumazet 已提交
839
	err = ip_send_skb(sock_net(sk), skb);
E
Eric Dumazet 已提交
840 841 842 843 844 845 846 847 848
	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
849 850 851 852 853 854
	return err;
}

/*
 * Push out all pending data as one UDP datagram. Socket is locked.
 */
855
int udp_push_pending_frames(struct sock *sk)
H
Herbert Xu 已提交
856 857 858
{
	struct udp_sock  *up = udp_sk(sk);
	struct inet_sock *inet = inet_sk(sk);
D
David S. Miller 已提交
859
	struct flowi4 *fl4 = &inet->cork.fl.u.ip4;
H
Herbert Xu 已提交
860 861 862
	struct sk_buff *skb;
	int err = 0;

863
	skb = ip_finish_skb(sk, fl4);
H
Herbert Xu 已提交
864 865 866
	if (!skb)
		goto out;

867
	err = udp_send_skb(skb, fl4);
H
Herbert Xu 已提交
868

869 870 871 872 873
out:
	up->len = 0;
	up->pending = 0;
	return err;
}
874
EXPORT_SYMBOL(udp_push_pending_frames);
875 876 877 878 879 880

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);
881
	struct flowi4 fl4_stack;
D
David S. Miller 已提交
882
	struct flowi4 *fl4;
883 884 885 886 887 888 889 890 891 892 893
	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
894
	struct sk_buff *skb;
895
	struct ip_options_data opt_copy;
896 897 898 899 900 901 902 903

	if (len > 0xFFFF)
		return -EMSGSIZE;

	/*
	 *	Check the flags.
	 */

E
Eric Dumazet 已提交
904
	if (msg->msg_flags & MSG_OOB) /* Mirror BSD error message compatibility */
905 906 907
		return -EOPNOTSUPP;

	ipc.opt = NULL;
908
	ipc.tx_flags = 0;
909 910
	ipc.ttl = 0;
	ipc.tos = -1;
911

H
Herbert Xu 已提交
912 913
	getfrag = is_udplite ? udplite_getfrag : ip_generic_getfrag;

914
	fl4 = &inet->cork.fl.u.ip4;
915 916 917 918 919 920 921 922 923 924 925 926 927 928 929 930 931 932 933 934 935
	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) {
936
		DECLARE_SOCKADDR(struct sockaddr_in *, usin, msg->msg_name);
937 938 939 940 941 942 943 944 945 946 947 948 949 950
		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
951 952
		daddr = inet->inet_daddr;
		dport = inet->inet_dport;
953 954 955 956 957
		/* 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
958
	ipc.addr = inet->inet_saddr;
959 960

	ipc.oif = sk->sk_bound_dev_if;
961 962 963

	sock_tx_timestamp(sk, &ipc.tx_flags);

964
	if (msg->msg_controllen) {
965 966
		err = ip_cmsg_send(sock_net(sk), msg, &ipc,
				   sk->sk_family == AF_INET6);
967 968 969 970 971 972
		if (err)
			return err;
		if (ipc.opt)
			free = 1;
		connected = 0;
	}
973 974 975 976 977 978 979 980 981 982 983 984
	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();
	}
985 986 987 988

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

989
	if (ipc.opt && ipc.opt->opt.srr) {
990 991
		if (!daddr)
			return -EINVAL;
992
		faddr = ipc.opt->opt.faddr;
993 994
		connected = 0;
	}
995
	tos = get_rttos(&ipc, inet);
996 997
	if (sock_flag(sk, SOCK_LOCALROUTE) ||
	    (msg->msg_flags & MSG_DONTROUTE) ||
998
	    (ipc.opt && ipc.opt->opt.is_strictroute)) {
999 1000 1001 1002 1003 1004 1005 1006 1007 1008
		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;
1009 1010
	} else if (!ipc.oif)
		ipc.oif = inet->uc_index;
1011 1012

	if (connected)
E
Eric Dumazet 已提交
1013
		rt = (struct rtable *)sk_dst_check(sk, 0);
1014 1015

	if (rt == NULL) {
1016 1017
		struct net *net = sock_net(sk);

1018 1019
		fl4 = &fl4_stack;
		flowi4_init_output(fl4, ipc.oif, sk->sk_mark, tos,
1020
				   RT_SCOPE_UNIVERSE, sk->sk_protocol,
1021
				   inet_sk_flowi_flags(sk),
1022 1023
				   faddr, saddr, dport, inet->inet_sport);

1024 1025
		security_sk_classify_flow(sk, flowi4_to_flowi(fl4));
		rt = ip_route_output_flow(net, fl4, sk);
1026 1027
		if (IS_ERR(rt)) {
			err = PTR_ERR(rt);
1028
			rt = NULL;
1029
			if (err == -ENETUNREACH)
1030
				IP_INC_STATS(net, IPSTATS_MIB_OUTNOROUTES);
1031 1032 1033 1034 1035 1036 1037 1038
			goto out;
		}

		err = -EACCES;
		if ((rt->rt_flags & RTCF_BROADCAST) &&
		    !sock_flag(sk, SOCK_BROADCAST))
			goto out;
		if (connected)
1039
			sk_dst_set(sk, dst_clone(&rt->dst));
1040 1041 1042 1043 1044 1045
	}

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

1046
	saddr = fl4->saddr;
1047
	if (!ipc.addr)
1048
		daddr = ipc.addr = fl4->daddr;
1049

H
Herbert Xu 已提交
1050 1051
	/* Lockless fast path for the non-corking case. */
	if (!corkreq) {
1052
		skb = ip_make_skb(sk, fl4, getfrag, msg, ulen,
H
Herbert Xu 已提交
1053 1054 1055
				  sizeof(struct udphdr), &ipc, &rt,
				  msg->msg_flags);
		err = PTR_ERR(skb);
1056
		if (!IS_ERR_OR_NULL(skb))
1057
			err = udp_send_skb(skb, fl4);
H
Herbert Xu 已提交
1058 1059 1060
		goto out;
	}

1061 1062 1063 1064 1065 1066
	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);

1067
		net_dbg_ratelimited("cork app bug 2\n");
1068 1069 1070 1071 1072 1073
		err = -EINVAL;
		goto out;
	}
	/*
	 *	Now cork the socket to pend data.
	 */
D
David S. Miller 已提交
1074 1075 1076
	fl4 = &inet->cork.fl.u.ip4;
	fl4->daddr = daddr;
	fl4->saddr = saddr;
1077 1078
	fl4->fl4_dport = dport;
	fl4->fl4_sport = inet->inet_sport;
1079 1080 1081 1082
	up->pending = AF_INET;

do_append_data:
	up->len += ulen;
1083
	err = ip_append_data(sk, fl4, getfrag, msg, ulen,
1084 1085
			     sizeof(struct udphdr), &ipc, &rt,
			     corkreq ? msg->msg_flags|MSG_MORE : msg->msg_flags);
1086 1087 1088 1089 1090 1091 1092 1093 1094 1095 1096 1097 1098 1099 1100 1101 1102 1103 1104 1105 1106 1107
	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)) {
1108 1109
		UDP_INC_STATS_USER(sock_net(sk),
				UDP_MIB_SNDBUFERRORS, is_udplite);
1110 1111 1112 1113
	}
	return err;

do_confirm:
1114
	dst_confirm(&rt->dst);
1115 1116 1117 1118 1119
	if (!(msg->msg_flags&MSG_PROBE) || len)
		goto back_from_confirm;
	err = 0;
	goto out;
}
E
Eric Dumazet 已提交
1120
EXPORT_SYMBOL(udp_sendmsg);
1121 1122 1123 1124

int udp_sendpage(struct sock *sk, struct page *page, int offset,
		 size_t size, int flags)
{
1125
	struct inet_sock *inet = inet_sk(sk);
1126 1127 1128
	struct udp_sock *up = udp_sk(sk);
	int ret;

1129 1130 1131
	if (flags & MSG_SENDPAGE_NOTLAST)
		flags |= MSG_MORE;

1132 1133 1134 1135 1136 1137 1138 1139 1140 1141 1142 1143 1144 1145 1146 1147 1148
	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);

1149
		net_dbg_ratelimited("udp cork app bug 3\n");
1150 1151 1152
		return -EINVAL;
	}

1153 1154
	ret = ip_append_page(sk, &inet->cork.fl.u.ip4,
			     page, offset, size, flags);
1155 1156 1157 1158 1159 1160 1161 1162 1163 1164 1165 1166 1167 1168 1169 1170 1171 1172 1173 1174
	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
1175 1176 1177 1178 1179 1180 1181 1182 1183 1184 1185 1186 1187 1188 1189 1190 1191 1192 1193

/**
 *	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)) {
1194 1195
		UDP_INC_STATS_BH(sock_net(sk), UDP_MIB_CSUMERRORS,
				 IS_UDPLITE(sk));
E
Eric Dumazet 已提交
1196 1197
		UDP_INC_STATS_BH(sock_net(sk), UDP_MIB_INERRORS,
				 IS_UDPLITE(sk));
E
Eric Dumazet 已提交
1198
		atomic_inc(&sk->sk_drops);
E
Eric Dumazet 已提交
1199 1200 1201 1202 1203 1204 1205
		__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)) {
1206 1207
		bool slow = lock_sock_fast(sk);

E
Eric Dumazet 已提交
1208 1209
		__skb_queue_purge(&list_kill);
		sk_mem_reclaim_partial(sk);
1210
		unlock_sock_fast(sk, slow);
E
Eric Dumazet 已提交
1211 1212 1213 1214
	}
	return res;
}

L
Linus Torvalds 已提交
1215 1216 1217
/*
 *	IOCTL requests applicable to the UDP protocol
 */
1218

L
Linus Torvalds 已提交
1219 1220
int udp_ioctl(struct sock *sk, int cmd, unsigned long arg)
{
1221 1222
	switch (cmd) {
	case SIOCOUTQ:
L
Linus Torvalds 已提交
1223
	{
1224 1225
		int amount = sk_wmem_alloc_get(sk);

1226 1227
		return put_user(amount, (int __user *)arg);
	}
L
Linus Torvalds 已提交
1228

1229 1230
	case SIOCINQ:
	{
E
Eric Dumazet 已提交
1231
		unsigned int amount = first_packet_length(sk);
1232

E
Eric Dumazet 已提交
1233
		if (amount)
1234 1235 1236 1237 1238
			/*
			 * We will only return the amount
			 * of this packet since that is all
			 * that will be read.
			 */
E
Eric Dumazet 已提交
1239 1240
			amount -= sizeof(struct udphdr);

1241 1242
		return put_user(amount, (int __user *)arg);
	}
L
Linus Torvalds 已提交
1243

1244 1245
	default:
		return -ENOIOCTLCMD;
L
Linus Torvalds 已提交
1246
	}
1247 1248

	return 0;
L
Linus Torvalds 已提交
1249
}
E
Eric Dumazet 已提交
1250
EXPORT_SYMBOL(udp_ioctl);
L
Linus Torvalds 已提交
1251

1252 1253 1254 1255 1256 1257 1258 1259 1260
/*
 * 	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);
1261
	DECLARE_SOCKADDR(struct sockaddr_in *, sin, msg->msg_name);
1262
	struct sk_buff *skb;
1263
	unsigned int ulen, copied;
1264
	int peeked, off = 0;
1265 1266
	int err;
	int is_udplite = IS_UDPLITE(sk);
1267
	bool slow;
1268 1269

	if (flags & MSG_ERRQUEUE)
1270
		return ip_recv_error(sk, msg, len, addr_len);
1271 1272 1273

try_again:
	skb = __skb_recv_datagram(sk, flags | (noblock ? MSG_DONTWAIT : 0),
1274
				  &peeked, &off, &err);
1275 1276 1277 1278
	if (!skb)
		goto out;

	ulen = skb->len - sizeof(struct udphdr);
1279 1280 1281 1282
	copied = len;
	if (copied > ulen)
		copied = ulen;
	else if (copied < ulen)
1283 1284 1285 1286 1287 1288 1289 1290
		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.
	 */

1291
	if (copied < ulen || UDP_SKB_CB(skb)->partial_cov) {
1292 1293 1294 1295 1296
		if (udp_lib_checksum_complete(skb))
			goto csum_copy_err;
	}

	if (skb_csum_unnecessary(skb))
1297 1298
		err = skb_copy_datagram_msg(skb, sizeof(struct udphdr),
					    msg, copied);
1299
	else {
1300 1301
		err = skb_copy_and_csum_datagram_msg(skb, sizeof(struct udphdr),
						     msg);
1302 1303 1304 1305 1306

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

1307 1308
	if (unlikely(err)) {
		trace_kfree_skb(skb, udp_recvmsg);
1309 1310 1311 1312 1313
		if (!peeked) {
			atomic_inc(&sk->sk_drops);
			UDP_INC_STATS_USER(sock_net(sk),
					   UDP_MIB_INERRORS, is_udplite);
		}
1314
		goto out_free;
1315
	}
1316 1317

	if (!peeked)
1318 1319
		UDP_INC_STATS_USER(sock_net(sk),
				UDP_MIB_INDATAGRAMS, is_udplite);
1320

1321
	sock_recv_ts_and_drops(msg, sk, skb);
1322 1323

	/* Copy the address. */
E
Eric Dumazet 已提交
1324
	if (sin) {
1325 1326 1327 1328
		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));
1329
		*addr_len = sizeof(*sin);
1330 1331 1332 1333
	}
	if (inet->cmsg_flags)
		ip_cmsg_recv(msg, skb);

1334
	err = copied;
1335 1336 1337 1338
	if (flags & MSG_TRUNC)
		err = ulen;

out_free:
1339
	skb_free_datagram_locked(sk, skb);
1340 1341 1342 1343
out:
	return err;

csum_copy_err:
1344
	slow = lock_sock_fast(sk);
1345 1346
	if (!skb_kill_datagram(sk, skb, flags)) {
		UDP_INC_STATS_USER(sock_net(sk), UDP_MIB_CSUMERRORS, is_udplite);
1347
		UDP_INC_STATS_USER(sock_net(sk), UDP_MIB_INERRORS, is_udplite);
1348
	}
1349
	unlock_sock_fast(sk, slow);
1350 1351 1352

	if (noblock)
		return -EAGAIN;
1353 1354 1355

	/* starting over for a new packet */
	msg->msg_flags &= ~MSG_TRUNC;
1356 1357 1358 1359
	goto try_again;
}


L
Linus Torvalds 已提交
1360 1361 1362 1363 1364 1365
int udp_disconnect(struct sock *sk, int flags)
{
	struct inet_sock *inet = inet_sk(sk);
	/*
	 *	1003.1g - break association.
	 */
1366

L
Linus Torvalds 已提交
1367
	sk->sk_state = TCP_CLOSE;
E
Eric Dumazet 已提交
1368 1369
	inet->inet_daddr = 0;
	inet->inet_dport = 0;
1370
	sock_rps_reset_rxhash(sk);
L
Linus Torvalds 已提交
1371 1372 1373 1374 1375 1376
	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
1377
		inet->inet_sport = 0;
L
Linus Torvalds 已提交
1378 1379 1380 1381
	}
	sk_dst_reset(sk);
	return 0;
}
E
Eric Dumazet 已提交
1382
EXPORT_SYMBOL(udp_disconnect);
L
Linus Torvalds 已提交
1383

1384 1385
void udp_lib_unhash(struct sock *sk)
{
1386 1387
	if (sk_hashed(sk)) {
		struct udp_table *udptable = sk->sk_prot->h.udp_table;
1388 1389 1390 1391 1392
		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);
1393

1394 1395
		spin_lock_bh(&hslot->lock);
		if (sk_nulls_del_node_init_rcu(sk)) {
E
Eric Dumazet 已提交
1396
			hslot->count--;
E
Eric Dumazet 已提交
1397
			inet_sk(sk)->inet_num = 0;
1398
			sock_prot_inuse_add(sock_net(sk), sk->sk_prot, -1);
1399 1400 1401 1402 1403

			spin_lock(&hslot2->lock);
			hlist_nulls_del_init_rcu(&udp_sk(sk)->udp_portaddr_node);
			hslot2->count--;
			spin_unlock(&hslot2->lock);
1404 1405
		}
		spin_unlock_bh(&hslot->lock);
1406 1407 1408 1409
	}
}
EXPORT_SYMBOL(udp_lib_unhash);

E
Eric Dumazet 已提交
1410 1411 1412 1413 1414 1415 1416 1417 1418 1419 1420 1421 1422 1423 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
/*
 * 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
1453 1454
static int __udp_queue_rcv_skb(struct sock *sk, struct sk_buff *skb)
{
T
Tom Herbert 已提交
1455
	int rc;
E
Eric Dumazet 已提交
1456

1457
	if (inet_sk(sk)->inet_daddr) {
1458
		sock_rps_save_rxhash(sk, skb);
1459
		sk_mark_napi_id(sk, skb);
E
Eric Dumazet 已提交
1460
		sk_incoming_cpu_update(sk);
1461
	}
T
Tom Herbert 已提交
1462

1463
	rc = sock_queue_rcv_skb(sk, skb);
E
Eric Dumazet 已提交
1464 1465
	if (rc < 0) {
		int is_udplite = IS_UDPLITE(sk);
H
Herbert Xu 已提交
1466 1467

		/* Note that an ENOMEM error is charged twice */
E
Eric Dumazet 已提交
1468
		if (rc == -ENOMEM)
H
Herbert Xu 已提交
1469 1470
			UDP_INC_STATS_BH(sock_net(sk), UDP_MIB_RCVBUFERRORS,
					 is_udplite);
E
Eric Dumazet 已提交
1471 1472
		UDP_INC_STATS_BH(sock_net(sk), UDP_MIB_INERRORS, is_udplite);
		kfree_skb(skb);
1473
		trace_udp_fail_queue_rcv_skb(rc, sk);
E
Eric Dumazet 已提交
1474
		return -1;
H
Herbert Xu 已提交
1475 1476 1477 1478 1479 1480
	}

	return 0;

}

1481 1482 1483 1484 1485 1486 1487 1488
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);

1489 1490 1491 1492 1493 1494 1495 1496
/* 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
1497
int udp_queue_rcv_skb(struct sock *sk, struct sk_buff *skb)
1498 1499 1500 1501 1502 1503 1504 1505 1506 1507 1508 1509
{
	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);

1510
	if (static_key_false(&udp_encap_needed) && up->encap_type) {
1511 1512
		int (*encap_rcv)(struct sock *sk, struct sk_buff *skb);

1513 1514 1515 1516 1517 1518 1519 1520 1521 1522 1523 1524
		/*
		 * 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 */
1525 1526
		encap_rcv = ACCESS_ONCE(up->encap_rcv);
		if (skb->len > sizeof(struct udphdr) && encap_rcv != NULL) {
1527 1528
			int ret;

1529 1530 1531 1532
			/* Verify checksum before giving to encap */
			if (udp_lib_checksum_complete(skb))
				goto csum_error;

1533
			ret = encap_rcv(sk, skb);
1534
			if (ret <= 0) {
1535 1536
				UDP_INC_STATS_BH(sock_net(sk),
						 UDP_MIB_INDATAGRAMS,
1537 1538 1539 1540 1541 1542 1543 1544 1545 1546 1547 1548 1549 1550 1551 1552 1553 1554 1555 1556 1557 1558 1559 1560 1561
						 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  */
1562 1563
			net_dbg_ratelimited("UDPLite: partial coverage %d while full coverage %d requested\n",
					    UDP_SKB_CB(skb)->cscov, skb->len);
1564 1565 1566 1567 1568 1569 1570 1571 1572
			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) {
1573 1574
			net_dbg_ratelimited("UDPLite: coverage %d too small, need min %d\n",
					    UDP_SKB_CB(skb)->cscov, up->pcrlen);
1575 1576 1577 1578
			goto drop;
		}
	}

1579 1580
	if (rcu_access_pointer(sk->sk_filter) &&
	    udp_lib_checksum_complete(skb))
1581
		goto csum_error;
1582

1583

1584
	if (sk_rcvqueues_full(sk, sk->sk_rcvbuf)) {
1585 1586
		UDP_INC_STATS_BH(sock_net(sk), UDP_MIB_RCVBUFERRORS,
				 is_udplite);
1587
		goto drop;
1588
	}
1589

H
Herbert Xu 已提交
1590
	rc = 0;
1591

1592
	ipv4_pktinfo_prepare(sk, skb);
H
Herbert Xu 已提交
1593 1594 1595
	bh_lock_sock(sk);
	if (!sock_owned_by_user(sk))
		rc = __udp_queue_rcv_skb(sk, skb);
1596
	else if (sk_add_backlog(sk, skb, sk->sk_rcvbuf)) {
Z
Zhu Yi 已提交
1597 1598 1599
		bh_unlock_sock(sk);
		goto drop;
	}
H
Herbert Xu 已提交
1600 1601 1602
	bh_unlock_sock(sk);

	return rc;
1603

1604 1605
csum_error:
	UDP_INC_STATS_BH(sock_net(sk), UDP_MIB_CSUMERRORS, is_udplite);
1606
drop:
1607
	UDP_INC_STATS_BH(sock_net(sk), UDP_MIB_INERRORS, is_udplite);
E
Eric Dumazet 已提交
1608
	atomic_inc(&sk->sk_drops);
1609 1610 1611 1612
	kfree_skb(skb);
	return -1;
}

1613 1614 1615 1616 1617 1618

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;
1619
	struct sock *sk;
1620 1621

	for (i = 0; i < count; i++) {
1622
		sk = stack[i];
1623 1624 1625
		if (likely(skb1 == NULL))
			skb1 = (i == final) ? skb : skb_clone(skb, GFP_ATOMIC);

1626 1627 1628 1629 1630 1631 1632 1633 1634
		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)
1635
			skb1 = NULL;
1636 1637

		sock_put(sk);
1638 1639 1640 1641 1642
	}
	if (unlikely(skb1))
		kfree_skb(skb1);
}

1643
/* For TCP sockets, sk_rx_dst is protected by socket lock
1644
 * For UDP, we use xchg() to guard against concurrent changes.
1645 1646
 */
static void udp_sk_rx_dst_set(struct sock *sk, struct dst_entry *dst)
S
Shawn Bohrer 已提交
1647
{
1648 1649
	struct dst_entry *old;

1650 1651 1652
	dst_hold(dst);
	old = xchg(&sk->sk_rx_dst, dst);
	dst_release(old);
S
Shawn Bohrer 已提交
1653 1654
}

1655 1656 1657
/*
 *	Multicasts and broadcasts go to each listener.
 *
1658
 *	Note: called only from the BH handler context.
1659
 */
1660
static int __udp4_lib_mcast_deliver(struct net *net, struct sk_buff *skb,
1661 1662
				    struct udphdr  *uh,
				    __be32 saddr, __be32 daddr,
1663 1664
				    struct udp_table *udptable,
				    int proto)
1665
{
1666
	struct sock *sk, *stack[256 / sizeof(struct sock *)];
1667 1668 1669 1670
	struct hlist_nulls_node *node;
	unsigned short hnum = ntohs(uh->dest);
	struct udp_hslot *hslot = udp_hashslot(udptable, net, hnum);
	int dif = skb->dev->ifindex;
1671 1672
	unsigned int count = 0, offset = offsetof(typeof(*sk), sk_nulls_node);
	unsigned int hash2 = 0, hash2_any = 0, use_hash2 = (hslot->count > 10);
1673
	bool inner_flushed = false;
1674 1675 1676 1677 1678 1679 1680 1681 1682

	if (use_hash2) {
		hash2_any = udp4_portaddr_hash(net, htonl(INADDR_ANY), hnum) &
			    udp_table.mask;
		hash2 = udp4_portaddr_hash(net, daddr, hnum) & udp_table.mask;
start_lookup:
		hslot = &udp_table.hash2[hash2];
		offset = offsetof(typeof(*sk), __sk_common.skc_portaddr_node);
	}
1683

1684
	spin_lock(&hslot->lock);
1685
	sk_nulls_for_each_entry_offset(sk, node, &hslot->head, offset) {
1686 1687 1688 1689 1690 1691
		if (__udp_is_mcast_sock(net, sk,
					uh->dest, daddr,
					uh->source, saddr,
					dif, hnum)) {
			if (unlikely(count == ARRAY_SIZE(stack))) {
				flush_stack(stack, count, skb, ~0);
1692
				inner_flushed = true;
1693 1694 1695
				count = 0;
			}
			stack[count++] = sk;
1696
			sock_hold(sk);
1697 1698 1699
		}
	}

1700
	spin_unlock(&hslot->lock);
1701

1702 1703 1704 1705 1706 1707
	/* Also lookup *:port if we are using hash2 and haven't done so yet. */
	if (use_hash2 && hash2 != hash2_any) {
		hash2 = hash2_any;
		goto start_lookup;
	}

1708 1709 1710 1711 1712 1713
	/*
	 * do the slow work with no lock held
	 */
	if (count) {
		flush_stack(stack, count, skb, count - 1);
	} else {
1714 1715 1716 1717
		if (!inner_flushed)
			UDP_INC_STATS_BH(net, UDP_MIB_IGNOREDMULTI,
					 proto == IPPROTO_UDPLITE);
		consume_skb(skb);
1718
	}
1719 1720 1721 1722 1723 1724 1725 1726 1727 1728 1729 1730 1731 1732 1733 1734 1735 1736 1737 1738 1739 1740
	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)
{
	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;
	}

1741 1742
	return skb_checksum_init_zero_check(skb, proto, uh->check,
					    inet_compute_pseudo);
1743 1744 1745 1746 1747 1748
}

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

1749
int __udp4_lib_rcv(struct sk_buff *skb, struct udp_table *udptable,
1750 1751 1752
		   int proto)
{
	struct sock *sk;
1753
	struct udphdr *uh;
1754
	unsigned short ulen;
E
Eric Dumazet 已提交
1755
	struct rtable *rt = skb_rtable(skb);
1756
	__be32 saddr, daddr;
1757
	struct net *net = dev_net(skb->dev);
1758 1759 1760 1761 1762 1763 1764

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

1765
	uh   = udp_hdr(skb);
1766
	ulen = ntohs(uh->len);
1767 1768 1769
	saddr = ip_hdr(skb)->saddr;
	daddr = ip_hdr(skb)->daddr;

1770 1771 1772 1773 1774 1775 1776 1777 1778 1779 1780 1781 1782
	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;

1783 1784
	sk = skb_steal_sock(skb);
	if (sk) {
1785
		struct dst_entry *dst = skb_dst(skb);
S
Shawn Bohrer 已提交
1786 1787
		int ret;

1788 1789
		if (unlikely(sk->sk_rx_dst != dst))
			udp_sk_rx_dst_set(sk, dst);
1790

S
Shawn Bohrer 已提交
1791
		ret = udp_queue_rcv_skb(sk, skb);
1792
		sock_put(sk);
S
Shawn Bohrer 已提交
1793 1794 1795 1796 1797 1798 1799
		/* 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;
	}
1800

F
Fabian Frederick 已提交
1801 1802
	if (rt->rt_flags & (RTCF_BROADCAST|RTCF_MULTICAST))
		return __udp4_lib_mcast_deliver(net, skb, uh,
1803
						saddr, daddr, udptable, proto);
F
Fabian Frederick 已提交
1804 1805

	sk = __udp4_lib_lookup_skb(skb, uh->source, uh->dest, udptable);
1806
	if (sk != NULL) {
1807 1808
		int ret;

1809 1810 1811 1812
		if (udp_sk(sk)->convert_csum && uh->check && !IS_UDPLITE(sk))
			skb_checksum_try_convert(skb, IPPROTO_UDP, uh->check,
						 inet_compute_pseudo);

1813
		ret = udp_queue_rcv_skb(sk, skb);
1814 1815 1816 1817 1818 1819 1820 1821 1822 1823 1824 1825 1826 1827 1828 1829 1830 1831
		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;

1832
	UDP_INC_STATS_BH(net, UDP_MIB_NOPORTS, proto == IPPROTO_UDPLITE);
1833 1834 1835 1836 1837 1838 1839 1840 1841 1842
	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:
1843 1844 1845 1846 1847
	net_dbg_ratelimited("UDP%s: short packet: From %pI4:%u %d/%d to %pI4:%u\n",
			    proto == IPPROTO_UDPLITE ? "Lite" : "",
			    &saddr, ntohs(uh->source),
			    ulen, skb->len,
			    &daddr, ntohs(uh->dest));
1848 1849 1850 1851 1852 1853 1854
	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).
	 */
1855 1856 1857 1858
	net_dbg_ratelimited("UDP%s: bad checksum. From %pI4:%u to %pI4:%u ulen %d\n",
			    proto == IPPROTO_UDPLITE ? "Lite" : "",
			    &saddr, ntohs(uh->source), &daddr, ntohs(uh->dest),
			    ulen);
1859
	UDP_INC_STATS_BH(net, UDP_MIB_CSUMERRORS, proto == IPPROTO_UDPLITE);
1860
drop:
1861
	UDP_INC_STATS_BH(net, UDP_MIB_INERRORS, proto == IPPROTO_UDPLITE);
1862 1863 1864 1865
	kfree_skb(skb);
	return 0;
}

S
Shawn Bohrer 已提交
1866 1867 1868 1869 1870 1871 1872 1873 1874 1875 1876 1877 1878 1879
/* 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];

1880 1881 1882 1883
	/* Do not bother scanning a too big list */
	if (hslot->count > 10)
		return NULL;

S
Shawn Bohrer 已提交
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
	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;
1909
		else if (unlikely(!__udp_is_mcast_sock(net, result,
S
Shawn Bohrer 已提交
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
						       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];
1936
	INET_ADDR_COOKIE(acookie, rmt_addr, loc_addr);
S
Shawn Bohrer 已提交
1937 1938 1939 1940 1941 1942 1943 1944 1945 1946 1947 1948 1949 1950 1951 1952 1953 1954 1955 1956 1957 1958 1959 1960 1961 1962 1963 1964
	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)
{
1965 1966 1967
	struct net *net = dev_net(skb->dev);
	const struct iphdr *iph;
	const struct udphdr *uh;
S
Shawn Bohrer 已提交
1968 1969 1970 1971 1972 1973 1974 1975
	struct sock *sk;
	struct dst_entry *dst;
	int dif = skb->dev->ifindex;

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

1976 1977 1978
	iph = ip_hdr(skb);
	uh = udp_hdr(skb);

S
Shawn Bohrer 已提交
1979 1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992
	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;
1993
	skb->destructor = sock_efree;
S
Shawn Bohrer 已提交
1994 1995 1996 1997 1998 1999 2000 2001
	dst = sk->sk_rx_dst;

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

2002 2003
int udp_rcv(struct sk_buff *skb)
{
2004
	return __udp4_lib_rcv(skb, &udp_table, IPPROTO_UDP);
2005 2006
}

2007
void udp_destroy_sock(struct sock *sk)
2008
{
T
Tom Parkin 已提交
2009
	struct udp_sock *up = udp_sk(sk);
2010
	bool slow = lock_sock_fast(sk);
2011
	udp_flush_pending_frames(sk);
2012
	unlock_sock_fast(sk, slow);
T
Tom Parkin 已提交
2013 2014 2015 2016 2017 2018
	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);
	}
2019 2020
}

L
Linus Torvalds 已提交
2021 2022 2023
/*
 *	Socket option code for UDP
 */
2024
int udp_lib_setsockopt(struct sock *sk, int level, int optname,
2025
		       char __user *optval, unsigned int optlen,
2026
		       int (*push_pending_frames)(struct sock *))
L
Linus Torvalds 已提交
2027 2028
{
	struct udp_sock *up = udp_sk(sk);
2029
	int val, valbool;
L
Linus Torvalds 已提交
2030
	int err = 0;
W
Wang Chen 已提交
2031
	int is_udplite = IS_UDPLITE(sk);
L
Linus Torvalds 已提交
2032

E
Eric Dumazet 已提交
2033
	if (optlen < sizeof(int))
L
Linus Torvalds 已提交
2034 2035 2036 2037 2038
		return -EINVAL;

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

2039 2040
	valbool = val ? 1 : 0;

2041
	switch (optname) {
L
Linus Torvalds 已提交
2042 2043 2044 2045 2046 2047
	case UDP_CORK:
		if (val != 0) {
			up->corkflag = 1;
		} else {
			up->corkflag = 0;
			lock_sock(sk);
2048
			push_pending_frames(sk);
L
Linus Torvalds 已提交
2049 2050 2051
			release_sock(sk);
		}
		break;
2052

L
Linus Torvalds 已提交
2053 2054 2055 2056 2057
	case UDP_ENCAP:
		switch (val) {
		case 0:
		case UDP_ENCAP_ESPINUDP:
		case UDP_ENCAP_ESPINUDP_NON_IKE:
2058 2059
			up->encap_rcv = xfrm4_udp_encap_rcv;
			/* FALLTHROUGH */
2060
		case UDP_ENCAP_L2TPINUDP:
L
Linus Torvalds 已提交
2061
			up->encap_type = val;
2062
			udp_encap_enable();
L
Linus Torvalds 已提交
2063 2064 2065 2066 2067 2068 2069
			break;
		default:
			err = -ENOPROTOOPT;
			break;
		}
		break;

2070 2071 2072 2073 2074 2075 2076 2077
	case UDP_NO_CHECK6_TX:
		up->no_check6_tx = valbool;
		break;

	case UDP_NO_CHECK6_RX:
		up->no_check6_rx = valbool;
		break;

2078 2079 2080 2081 2082 2083
	/*
	 * 	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
2084
		if (!is_udplite)         /* Disable the option on UDP sockets */
2085 2086 2087
			return -ENOPROTOOPT;
		if (val != 0 && val < 8) /* Illegal coverage: use default (8) */
			val = 8;
2088 2089
		else if (val > USHRT_MAX)
			val = USHRT_MAX;
2090 2091 2092 2093
		up->pcslen = val;
		up->pcflag |= UDPLITE_SEND_CC;
		break;

2094 2095
	/* 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
2096 2097
	 * used, this again means full checksum coverage.                     */
	case UDPLITE_RECV_CSCOV:
W
Wang Chen 已提交
2098
		if (!is_udplite)         /* Disable the option on UDP sockets */
2099 2100 2101
			return -ENOPROTOOPT;
		if (val != 0 && val < 8) /* Avoid silly minimal values.       */
			val = 8;
2102 2103
		else if (val > USHRT_MAX)
			val = USHRT_MAX;
2104 2105 2106 2107
		up->pcrlen = val;
		up->pcflag |= UDPLITE_RECV_CC;
		break;

L
Linus Torvalds 已提交
2108 2109 2110
	default:
		err = -ENOPROTOOPT;
		break;
2111
	}
L
Linus Torvalds 已提交
2112 2113 2114

	return err;
}
E
Eric Dumazet 已提交
2115
EXPORT_SYMBOL(udp_lib_setsockopt);
L
Linus Torvalds 已提交
2116

2117
int udp_setsockopt(struct sock *sk, int level, int optname,
2118
		   char __user *optval, unsigned int optlen)
2119 2120 2121 2122 2123 2124 2125 2126 2127
{
	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,
2128
			  char __user *optval, unsigned int optlen)
2129 2130 2131 2132 2133 2134 2135 2136
{
	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

2137 2138
int udp_lib_getsockopt(struct sock *sk, int level, int optname,
		       char __user *optval, int __user *optlen)
L
Linus Torvalds 已提交
2139 2140 2141 2142
{
	struct udp_sock *up = udp_sk(sk);
	int val, len;

E
Eric Dumazet 已提交
2143
	if (get_user(len, optlen))
L
Linus Torvalds 已提交
2144 2145 2146
		return -EFAULT;

	len = min_t(unsigned int, len, sizeof(int));
2147

2148
	if (len < 0)
L
Linus Torvalds 已提交
2149 2150
		return -EINVAL;

2151
	switch (optname) {
L
Linus Torvalds 已提交
2152 2153 2154 2155 2156 2157 2158 2159
	case UDP_CORK:
		val = up->corkflag;
		break;

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

2160 2161 2162 2163 2164 2165 2166 2167
	case UDP_NO_CHECK6_TX:
		val = up->no_check6_tx;
		break;

	case UDP_NO_CHECK6_RX:
		val = up->no_check6_rx;
		break;

2168 2169 2170 2171 2172 2173 2174 2175 2176 2177
	/* 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
2178 2179
	default:
		return -ENOPROTOOPT;
2180
	}
L
Linus Torvalds 已提交
2181

2182
	if (put_user(len, optlen))
2183
		return -EFAULT;
E
Eric Dumazet 已提交
2184
	if (copy_to_user(optval, &val, len))
L
Linus Torvalds 已提交
2185
		return -EFAULT;
2186
	return 0;
L
Linus Torvalds 已提交
2187
}
E
Eric Dumazet 已提交
2188
EXPORT_SYMBOL(udp_lib_getsockopt);
L
Linus Torvalds 已提交
2189

2190 2191 2192 2193 2194 2195 2196 2197 2198 2199 2200 2201 2202 2203 2204 2205 2206
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
2207 2208 2209 2210 2211 2212
/**
 * 	udp_poll - wait for a UDP event.
 *	@file - file struct
 *	@sock - socket
 *	@wait - poll table
 *
2213
 *	This is same as datagram poll, except for the special case of
L
Linus Torvalds 已提交
2214 2215 2216 2217 2218 2219 2220 2221 2222 2223
 *	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;
2224

2225 2226
	sock_rps_record_flow(sk);

L
Linus Torvalds 已提交
2227
	/* Check for false positives due to checksum errors */
E
Eric Dumazet 已提交
2228 2229 2230
	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
2231 2232

	return mask;
2233

L
Linus Torvalds 已提交
2234
}
E
Eric Dumazet 已提交
2235
EXPORT_SYMBOL(udp_poll);
L
Linus Torvalds 已提交
2236

2237 2238 2239 2240 2241 2242 2243 2244 2245 2246 2247 2248 2249
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
2250
	.backlog_rcv	   = __udp_queue_rcv_skb,
2251
	.release_cb	   = ip4_datagram_release_cb,
2252 2253
	.hash		   = udp_lib_hash,
	.unhash		   = udp_lib_unhash,
E
Eric Dumazet 已提交
2254
	.rehash		   = udp_v4_rehash,
2255 2256 2257 2258 2259 2260
	.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),
2261
	.slab_flags	   = SLAB_DESTROY_BY_RCU,
2262
	.h.udp_table	   = &udp_table,
2263 2264 2265 2266
#ifdef CONFIG_COMPAT
	.compat_setsockopt = compat_udp_setsockopt,
	.compat_getsockopt = compat_udp_getsockopt,
#endif
2267
	.clear_sk	   = sk_prot_clear_portaddr_nulls,
2268
};
E
Eric Dumazet 已提交
2269
EXPORT_SYMBOL(udp_prot);
L
Linus Torvalds 已提交
2270 2271 2272 2273

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

2274
static struct sock *udp_get_first(struct seq_file *seq, int start)
L
Linus Torvalds 已提交
2275 2276 2277
{
	struct sock *sk;
	struct udp_iter_state *state = seq->private;
2278
	struct net *net = seq_file_net(seq);
L
Linus Torvalds 已提交
2279

2280 2281
	for (state->bucket = start; state->bucket <= state->udp_table->mask;
	     ++state->bucket) {
2282
		struct hlist_nulls_node *node;
2283
		struct udp_hslot *hslot = &state->udp_table->hash[state->bucket];
2284 2285 2286 2287

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

2288
		spin_lock_bh(&hslot->lock);
2289
		sk_nulls_for_each(sk, node, &hslot->head) {
2290
			if (!net_eq(sock_net(sk), net))
2291
				continue;
L
Linus Torvalds 已提交
2292 2293 2294
			if (sk->sk_family == state->family)
				goto found;
		}
2295
		spin_unlock_bh(&hslot->lock);
L
Linus Torvalds 已提交
2296 2297 2298 2299 2300 2301 2302 2303 2304
	}
	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;
2305
	struct net *net = seq_file_net(seq);
L
Linus Torvalds 已提交
2306 2307

	do {
2308
		sk = sk_nulls_next(sk);
2309
	} while (sk && (!net_eq(sock_net(sk), net) || sk->sk_family != state->family));
L
Linus Torvalds 已提交
2310

2311
	if (!sk) {
2312
		if (state->bucket <= state->udp_table->mask)
2313
			spin_unlock_bh(&state->udp_table->hash[state->bucket].lock);
2314
		return udp_get_first(seq, state->bucket + 1);
L
Linus Torvalds 已提交
2315 2316 2317 2318 2319 2320
	}
	return sk;
}

static struct sock *udp_get_idx(struct seq_file *seq, loff_t pos)
{
2321
	struct sock *sk = udp_get_first(seq, 0);
L
Linus Torvalds 已提交
2322 2323

	if (sk)
2324
		while (pos && (sk = udp_get_next(seq, sk)) != NULL)
L
Linus Torvalds 已提交
2325 2326 2327 2328 2329 2330
			--pos;
	return pos ? NULL : sk;
}

static void *udp_seq_start(struct seq_file *seq, loff_t *pos)
{
2331
	struct udp_iter_state *state = seq->private;
2332
	state->bucket = MAX_UDP_PORTS;
2333

2334
	return *pos ? udp_get_idx(seq, *pos-1) : SEQ_START_TOKEN;
L
Linus Torvalds 已提交
2335 2336 2337 2338 2339 2340
}

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

2341
	if (v == SEQ_START_TOKEN)
L
Linus Torvalds 已提交
2342 2343 2344 2345 2346 2347 2348 2349 2350 2351
		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)
{
2352 2353
	struct udp_iter_state *state = seq->private;

2354
	if (state->bucket <= state->udp_table->mask)
2355
		spin_unlock_bh(&state->udp_table->hash[state->bucket].lock);
L
Linus Torvalds 已提交
2356 2357
}

2358
int udp_seq_open(struct inode *inode, struct file *file)
L
Linus Torvalds 已提交
2359
{
A
Al Viro 已提交
2360
	struct udp_seq_afinfo *afinfo = PDE_DATA(inode);
2361 2362
	struct udp_iter_state *s;
	int err;
2363

2364 2365 2366 2367
	err = seq_open_net(inode, file, &afinfo->seq_ops,
			   sizeof(struct udp_iter_state));
	if (err < 0)
		return err;
2368

2369
	s = ((struct seq_file *)file->private_data)->private;
L
Linus Torvalds 已提交
2370
	s->family		= afinfo->family;
2371
	s->udp_table		= afinfo->udp_table;
2372
	return err;
2373
}
2374
EXPORT_SYMBOL(udp_seq_open);
2375

L
Linus Torvalds 已提交
2376
/* ------------------------------------------------------------------------ */
2377
int udp_proc_register(struct net *net, struct udp_seq_afinfo *afinfo)
L
Linus Torvalds 已提交
2378 2379 2380 2381
{
	struct proc_dir_entry *p;
	int rc = 0;

2382 2383 2384 2385
	afinfo->seq_ops.start		= udp_seq_start;
	afinfo->seq_ops.next		= udp_seq_next;
	afinfo->seq_ops.stop		= udp_seq_stop;

2386
	p = proc_create_data(afinfo->name, S_IRUGO, net->proc_net,
2387
			     afinfo->seq_fops, afinfo);
2388
	if (!p)
L
Linus Torvalds 已提交
2389 2390 2391
		rc = -ENOMEM;
	return rc;
}
E
Eric Dumazet 已提交
2392
EXPORT_SYMBOL(udp_proc_register);
L
Linus Torvalds 已提交
2393

2394
void udp_proc_unregister(struct net *net, struct udp_seq_afinfo *afinfo)
L
Linus Torvalds 已提交
2395
{
2396
	remove_proc_entry(afinfo->name, net->proc_net);
L
Linus Torvalds 已提交
2397
}
E
Eric Dumazet 已提交
2398
EXPORT_SYMBOL(udp_proc_unregister);
2399 2400

/* ------------------------------------------------------------------------ */
2401
static void udp4_format_sock(struct sock *sp, struct seq_file *f,
2402
		int bucket)
2403 2404
{
	struct inet_sock *inet = inet_sk(sp);
E
Eric Dumazet 已提交
2405 2406 2407 2408
	__be32 dest = inet->inet_daddr;
	__be32 src  = inet->inet_rcv_saddr;
	__u16 destp	  = ntohs(inet->inet_dport);
	__u16 srcp	  = ntohs(inet->inet_sport);
2409

2410
	seq_printf(f, "%5d: %08X:%04X %08X:%04X"
2411
		" %02X %08X:%08X %02X:%08lX %08X %5u %8d %lu %d %pK %d",
2412
		bucket, src, srcp, dest, destp, sp->sk_state,
2413 2414
		sk_wmem_alloc_get(sp),
		sk_rmem_alloc_get(sp),
2415 2416 2417
		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
2418
		atomic_read(&sp->sk_refcnt), sp,
2419
		atomic_read(&sp->sk_drops));
2420 2421 2422 2423
}

int udp4_seq_show(struct seq_file *seq, void *v)
{
2424
	seq_setwidth(seq, 127);
2425
	if (v == SEQ_START_TOKEN)
2426
		seq_puts(seq, "  sl  local_address rem_address   st tx_queue "
2427
			   "rx_queue tr tm->when retrnsmt   uid  timeout "
E
Eric Dumazet 已提交
2428
			   "inode ref pointer drops");
2429 2430 2431
	else {
		struct udp_iter_state *state = seq->private;

2432
		udp4_format_sock(v, seq, state->bucket);
2433
	}
2434
	seq_pad(seq, '\n');
2435 2436 2437
	return 0;
}

2438 2439 2440 2441 2442 2443 2444 2445
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
};

2446 2447 2448 2449
/* ------------------------------------------------------------------------ */
static struct udp_seq_afinfo udp4_seq_afinfo = {
	.name		= "udp",
	.family		= AF_INET,
2450
	.udp_table	= &udp_table,
2451
	.seq_fops	= &udp_afinfo_seq_fops,
2452 2453 2454
	.seq_ops	= {
		.show		= udp4_seq_show,
	},
2455 2456
};

2457
static int __net_init udp4_proc_init_net(struct net *net)
2458 2459 2460 2461
{
	return udp_proc_register(net, &udp4_seq_afinfo);
}

2462
static void __net_exit udp4_proc_exit_net(struct net *net)
2463 2464 2465 2466 2467 2468 2469 2470 2471
{
	udp_proc_unregister(net, &udp4_seq_afinfo);
}

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

2472 2473
int __init udp4_proc_init(void)
{
2474
	return register_pernet_subsys(&udp4_net_ops);
2475 2476 2477 2478
}

void udp4_proc_exit(void)
{
2479
	unregister_pernet_subsys(&udp4_net_ops);
2480
}
L
Linus Torvalds 已提交
2481 2482
#endif /* CONFIG_PROC_FS */

2483 2484
static __initdata unsigned long uhash_entries;
static int __init set_uhash_entries(char *str)
2485
{
2486 2487
	ssize_t ret;

2488 2489
	if (!str)
		return 0;
2490 2491 2492 2493 2494

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

2495 2496 2497 2498 2499
	if (uhash_entries && uhash_entries < UDP_HTABLE_SIZE_MIN)
		uhash_entries = UDP_HTABLE_SIZE_MIN;
	return 1;
}
__setup("uhash_entries=", set_uhash_entries);
2500

2501 2502 2503 2504
void __init udp_table_init(struct udp_table *table, const char *name)
{
	unsigned int i;

2505 2506 2507 2508 2509 2510 2511 2512 2513 2514
	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);

2515
	table->hash2 = table->hash + (table->mask + 1);
2516
	for (i = 0; i <= table->mask; i++) {
2517
		INIT_HLIST_NULLS_HEAD(&table->hash[i].head, i);
E
Eric Dumazet 已提交
2518
		table->hash[i].count = 0;
2519 2520
		spin_lock_init(&table->hash[i].lock);
	}
2521 2522 2523 2524 2525
	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);
	}
2526 2527
}

H
Hideo Aoki 已提交
2528 2529
void __init udp_init(void)
{
2530
	unsigned long limit;
H
Hideo Aoki 已提交
2531

2532
	udp_table_init(&udp_table, "UDP");
2533
	limit = nr_free_buffer_pages() / 8;
H
Hideo Aoki 已提交
2534 2535 2536 2537 2538 2539 2540 2541
	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;
}