udp.c 63.5 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

226
		rand = prandom_u32();
227
		first = reciprocal_scale(rand, remaining) + low;
228 229 230
		/*
		 * 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
				    !test_bit(snum >> udptable->log, bitmap) &&
249
				    !inet_is_local_reserved_port(net, 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
				matches = 1;
			}
		} else if (score == badness && reuseport) {
			matches++;
451
			if (reciprocal_scale(hash, matches) == 0)
452 453
				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
				matches = 1;
			}
		} else if (score == badness && reuseport) {
			matches++;
532
			if (reciprocal_scale(hash, matches) == 0)
533 534
				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)
562 563 564
{
	const struct iphdr *iph = ip_hdr(skb);

565 566 567
	return __udp4_lib_lookup(dev_net(skb_dst(skb)->dev), iph->saddr, sport,
				 iph->daddr, dport, inet_iif(skb),
				 udptable);
568 569
}

570 571 572
struct sock *udp4_lib_lookup(struct net *net, __be32 saddr, __be16 sport,
			     __be32 daddr, __be16 dport, int dif)
{
573
	return __udp4_lib_lookup(net, saddr, sport, daddr, dport, dif, &udp_table);
574 575 576
}
EXPORT_SYMBOL_GPL(udp4_lib_lookup);

S
Shawn Bohrer 已提交
577 578 579 580 581 582 583 584 585 586 587 588 589 590 591 592 593 594 595 596
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;
}

597 598 599 600 601 602 603 604 605 606 607
/*
 * 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.
 */

608
void __udp4_lib_err(struct sk_buff *skb, u32 info, struct udp_table *udptable)
609 610
{
	struct inet_sock *inet;
611
	const struct iphdr *iph = (const struct iphdr *)skb->data;
E
Eric Dumazet 已提交
612
	struct udphdr *uh = (struct udphdr *)(skb->data+(iph->ihl<<2));
613 614 615 616 617
	const int type = icmp_hdr(skb)->type;
	const int code = icmp_hdr(skb)->code;
	struct sock *sk;
	int harderr;
	int err;
618
	struct net *net = dev_net(skb->dev);
619

620
	sk = __udp4_lib_lookup(net, iph->daddr, uh->dest,
621 622
			iph->saddr, uh->source, skb->dev->ifindex, udptable);
	if (sk == NULL) {
623
		ICMP_INC_STATS_BH(net, ICMP_MIB_INERRORS);
624 625 626 627 628 629 630 631 632 633 634 635 636 637 638 639 640 641 642 643
		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 */
644
			ipv4_sk_update_pmtu(skb, sk, info);
645 646 647 648 649 650 651 652 653 654 655 656 657
			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;
658 659
	case ICMP_REDIRECT:
		ipv4_sk_redirect(skb, sk);
660
		goto out;
661 662 663 664 665 666 667 668 669
	}

	/*
	 *      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;
670
	} else
E
Eric Dumazet 已提交
671
		ip_icmp_error(sk, skb, err, uh->dest, info, (u8 *)(uh+1));
672

673 674 675 676 677 678 679 680
	sk->sk_err = err;
	sk->sk_error_report(sk);
out:
	sock_put(sk);
}

void udp_err(struct sk_buff *skb, u32 info)
{
681
	__udp4_lib_err(skb, info, &udp_table);
682 683 684 685 686
}

/*
 * Throw away all pending data and cancel the corking. Socket is locked.
 */
687
void udp_flush_pending_frames(struct sock *sk)
688 689 690 691 692 693 694 695 696
{
	struct udp_sock *up = udp_sk(sk);

	if (up->pending) {
		up->len = 0;
		up->pending = 0;
		ip_flush_pending_frames(sk);
	}
}
697
EXPORT_SYMBOL(udp_flush_pending_frames);
698 699

/**
H
Herbert Xu 已提交
700
 * 	udp4_hwcsum  -  handle outgoing HW checksumming
701 702
 * 	@skb: 	sk_buff containing the filled-in UDP header
 * 	        (checksum field must be zeroed out)
H
Herbert Xu 已提交
703 704
 *	@src:	source IP address
 *	@dst:	destination IP address
705
 */
706
void udp4_hwcsum(struct sk_buff *skb, __be32 src, __be32 dst)
707 708
{
	struct udphdr *uh = udp_hdr(skb);
H
Herbert Xu 已提交
709 710 711
	int offset = skb_transport_offset(skb);
	int len = skb->len - offset;
	int hlen = len;
712 713
	__wsum csum = 0;

714
	if (!skb_has_frag_list(skb)) {
715 716 717 718 719
		/*
		 * 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
720 721
		uh->check = ~csum_tcpudp_magic(src, dst, len,
					       IPPROTO_UDP, 0);
722
	} else {
723 724
		struct sk_buff *frags;

725 726 727 728 729
		/*
		 * 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
		 */
730
		skb_walk_frags(skb, frags) {
H
Herbert Xu 已提交
731 732
			csum = csum_add(csum, frags->csum);
			hlen -= frags->len;
733
		}
734

H
Herbert Xu 已提交
735
		csum = skb_checksum(skb, offset, hlen, csum);
736 737 738 739 740 741 742
		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;
	}
}
743
EXPORT_SYMBOL_GPL(udp4_hwcsum);
744

745 746 747 748 749 750 751 752 753 754 755 756 757 758 759 760 761 762 763 764 765 766 767 768 769 770 771 772 773 774 775 776 777 778 779 780 781
/* 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);

782
static int udp_send_skb(struct sk_buff *skb, struct flowi4 *fl4)
783
{
H
Herbert Xu 已提交
784
	struct sock *sk = skb->sk;
785 786 787 788
	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
789 790
	int offset = skb_transport_offset(skb);
	int len = skb->len - offset;
791 792 793 794 795 796
	__wsum csum = 0;

	/*
	 * Create a UDP header
	 */
	uh = udp_hdr(skb);
H
Herbert Xu 已提交
797
	uh->source = inet->inet_sport;
798
	uh->dest = fl4->fl4_dport;
H
Herbert Xu 已提交
799
	uh->len = htons(len);
800 801 802
	uh->check = 0;

	if (is_udplite)  				 /*     UDP-Lite      */
H
Herbert Xu 已提交
803
		csum = udplite_csum(skb);
804

805
	else if (sk->sk_no_check_tx) {   /* UDP csum disabled */
806 807 808 809 810 811

		skb->ip_summed = CHECKSUM_NONE;
		goto send;

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

812
		udp4_hwcsum(skb, fl4->saddr, fl4->daddr);
813 814
		goto send;

H
Herbert Xu 已提交
815 816
	} else
		csum = udp_csum(skb);
817 818

	/* add protocol-dependent pseudo-header */
819
	uh->check = csum_tcpudp_magic(fl4->saddr, fl4->daddr, len,
E
Eric Dumazet 已提交
820
				      sk->sk_protocol, csum);
821 822 823 824
	if (uh->check == 0)
		uh->check = CSUM_MANGLED_0;

send:
E
Eric Dumazet 已提交
825
	err = ip_send_skb(sock_net(sk), skb);
E
Eric Dumazet 已提交
826 827 828 829 830 831 832 833 834
	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
835 836 837 838 839 840
	return err;
}

/*
 * Push out all pending data as one UDP datagram. Socket is locked.
 */
841
int udp_push_pending_frames(struct sock *sk)
H
Herbert Xu 已提交
842 843 844
{
	struct udp_sock  *up = udp_sk(sk);
	struct inet_sock *inet = inet_sk(sk);
D
David S. Miller 已提交
845
	struct flowi4 *fl4 = &inet->cork.fl.u.ip4;
H
Herbert Xu 已提交
846 847 848
	struct sk_buff *skb;
	int err = 0;

849
	skb = ip_finish_skb(sk, fl4);
H
Herbert Xu 已提交
850 851 852
	if (!skb)
		goto out;

853
	err = udp_send_skb(skb, fl4);
H
Herbert Xu 已提交
854

855 856 857 858 859
out:
	up->len = 0;
	up->pending = 0;
	return err;
}
860
EXPORT_SYMBOL(udp_push_pending_frames);
861 862 863 864 865 866

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);
867
	struct flowi4 fl4_stack;
D
David S. Miller 已提交
868
	struct flowi4 *fl4;
869 870 871 872 873 874 875 876 877 878 879
	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
880
	struct sk_buff *skb;
881
	struct ip_options_data opt_copy;
882 883 884 885 886 887 888 889

	if (len > 0xFFFF)
		return -EMSGSIZE;

	/*
	 *	Check the flags.
	 */

E
Eric Dumazet 已提交
890
	if (msg->msg_flags & MSG_OOB) /* Mirror BSD error message compatibility */
891 892 893
		return -EOPNOTSUPP;

	ipc.opt = NULL;
894
	ipc.tx_flags = 0;
895 896
	ipc.ttl = 0;
	ipc.tos = -1;
897

H
Herbert Xu 已提交
898 899
	getfrag = is_udplite ? udplite_getfrag : ip_generic_getfrag;

900
	fl4 = &inet->cork.fl.u.ip4;
901 902 903 904 905 906 907 908 909 910 911 912 913 914 915 916 917 918 919 920 921
	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) {
922
		DECLARE_SOCKADDR(struct sockaddr_in *, usin, msg->msg_name);
923 924 925 926 927 928 929 930 931 932 933 934 935 936
		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
937 938
		daddr = inet->inet_daddr;
		dport = inet->inet_dport;
939 940 941 942 943
		/* 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
944
	ipc.addr = inet->inet_saddr;
945 946

	ipc.oif = sk->sk_bound_dev_if;
947 948 949

	sock_tx_timestamp(sk, &ipc.tx_flags);

950
	if (msg->msg_controllen) {
951 952
		err = ip_cmsg_send(sock_net(sk), msg, &ipc,
				   sk->sk_family == AF_INET6);
953 954 955 956 957 958
		if (err)
			return err;
		if (ipc.opt)
			free = 1;
		connected = 0;
	}
959 960 961 962 963 964 965 966 967 968 969 970
	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();
	}
971 972 973 974

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

975
	if (ipc.opt && ipc.opt->opt.srr) {
976 977
		if (!daddr)
			return -EINVAL;
978
		faddr = ipc.opt->opt.faddr;
979 980
		connected = 0;
	}
981
	tos = get_rttos(&ipc, inet);
982 983
	if (sock_flag(sk, SOCK_LOCALROUTE) ||
	    (msg->msg_flags & MSG_DONTROUTE) ||
984
	    (ipc.opt && ipc.opt->opt.is_strictroute)) {
985 986 987 988 989 990 991 992 993 994
		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;
995 996
	} else if (!ipc.oif)
		ipc.oif = inet->uc_index;
997 998

	if (connected)
E
Eric Dumazet 已提交
999
		rt = (struct rtable *)sk_dst_check(sk, 0);
1000 1001

	if (rt == NULL) {
1002 1003
		struct net *net = sock_net(sk);

1004 1005
		fl4 = &fl4_stack;
		flowi4_init_output(fl4, ipc.oif, sk->sk_mark, tos,
1006
				   RT_SCOPE_UNIVERSE, sk->sk_protocol,
1007
				   inet_sk_flowi_flags(sk),
1008 1009
				   faddr, saddr, dport, inet->inet_sport);

1010 1011
		security_sk_classify_flow(sk, flowi4_to_flowi(fl4));
		rt = ip_route_output_flow(net, fl4, sk);
1012 1013
		if (IS_ERR(rt)) {
			err = PTR_ERR(rt);
1014
			rt = NULL;
1015
			if (err == -ENETUNREACH)
1016
				IP_INC_STATS(net, IPSTATS_MIB_OUTNOROUTES);
1017 1018 1019 1020 1021 1022 1023 1024
			goto out;
		}

		err = -EACCES;
		if ((rt->rt_flags & RTCF_BROADCAST) &&
		    !sock_flag(sk, SOCK_BROADCAST))
			goto out;
		if (connected)
1025
			sk_dst_set(sk, dst_clone(&rt->dst));
1026 1027 1028 1029 1030 1031
	}

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

1032
	saddr = fl4->saddr;
1033
	if (!ipc.addr)
1034
		daddr = ipc.addr = fl4->daddr;
1035

H
Herbert Xu 已提交
1036 1037
	/* Lockless fast path for the non-corking case. */
	if (!corkreq) {
1038
		skb = ip_make_skb(sk, fl4, getfrag, msg->msg_iov, ulen,
H
Herbert Xu 已提交
1039 1040 1041
				  sizeof(struct udphdr), &ipc, &rt,
				  msg->msg_flags);
		err = PTR_ERR(skb);
1042
		if (!IS_ERR_OR_NULL(skb))
1043
			err = udp_send_skb(skb, fl4);
H
Herbert Xu 已提交
1044 1045 1046
		goto out;
	}

1047 1048 1049 1050 1051 1052
	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);

1053
		LIMIT_NETDEBUG(KERN_DEBUG pr_fmt("cork app bug 2\n"));
1054 1055 1056 1057 1058 1059
		err = -EINVAL;
		goto out;
	}
	/*
	 *	Now cork the socket to pend data.
	 */
D
David S. Miller 已提交
1060 1061 1062
	fl4 = &inet->cork.fl.u.ip4;
	fl4->daddr = daddr;
	fl4->saddr = saddr;
1063 1064
	fl4->fl4_dport = dport;
	fl4->fl4_sport = inet->inet_sport;
1065 1066 1067 1068
	up->pending = AF_INET;

do_append_data:
	up->len += ulen;
1069 1070 1071
	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);
1072 1073 1074 1075 1076 1077 1078 1079 1080 1081 1082 1083 1084 1085 1086 1087 1088 1089 1090 1091 1092 1093
	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)) {
1094 1095
		UDP_INC_STATS_USER(sock_net(sk),
				UDP_MIB_SNDBUFERRORS, is_udplite);
1096 1097 1098 1099
	}
	return err;

do_confirm:
1100
	dst_confirm(&rt->dst);
1101 1102 1103 1104 1105
	if (!(msg->msg_flags&MSG_PROBE) || len)
		goto back_from_confirm;
	err = 0;
	goto out;
}
E
Eric Dumazet 已提交
1106
EXPORT_SYMBOL(udp_sendmsg);
1107 1108 1109 1110

int udp_sendpage(struct sock *sk, struct page *page, int offset,
		 size_t size, int flags)
{
1111
	struct inet_sock *inet = inet_sk(sk);
1112 1113 1114
	struct udp_sock *up = udp_sk(sk);
	int ret;

1115 1116 1117
	if (flags & MSG_SENDPAGE_NOTLAST)
		flags |= MSG_MORE;

1118 1119 1120 1121 1122 1123 1124 1125 1126 1127 1128 1129 1130 1131 1132 1133 1134
	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);

1135
		LIMIT_NETDEBUG(KERN_DEBUG pr_fmt("udp cork app bug 3\n"));
1136 1137 1138
		return -EINVAL;
	}

1139 1140
	ret = ip_append_page(sk, &inet->cork.fl.u.ip4,
			     page, offset, size, flags);
1141 1142 1143 1144 1145 1146 1147 1148 1149 1150 1151 1152 1153 1154 1155 1156 1157 1158 1159 1160
	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
1161 1162 1163 1164 1165 1166 1167 1168 1169 1170 1171 1172 1173 1174 1175 1176 1177 1178 1179

/**
 *	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)) {
1180 1181
		UDP_INC_STATS_BH(sock_net(sk), UDP_MIB_CSUMERRORS,
				 IS_UDPLITE(sk));
E
Eric Dumazet 已提交
1182 1183
		UDP_INC_STATS_BH(sock_net(sk), UDP_MIB_INERRORS,
				 IS_UDPLITE(sk));
E
Eric Dumazet 已提交
1184
		atomic_inc(&sk->sk_drops);
E
Eric Dumazet 已提交
1185 1186 1187 1188 1189 1190 1191
		__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)) {
1192 1193
		bool slow = lock_sock_fast(sk);

E
Eric Dumazet 已提交
1194 1195
		__skb_queue_purge(&list_kill);
		sk_mem_reclaim_partial(sk);
1196
		unlock_sock_fast(sk, slow);
E
Eric Dumazet 已提交
1197 1198 1199 1200
	}
	return res;
}

L
Linus Torvalds 已提交
1201 1202 1203
/*
 *	IOCTL requests applicable to the UDP protocol
 */
1204

L
Linus Torvalds 已提交
1205 1206
int udp_ioctl(struct sock *sk, int cmd, unsigned long arg)
{
1207 1208
	switch (cmd) {
	case SIOCOUTQ:
L
Linus Torvalds 已提交
1209
	{
1210 1211
		int amount = sk_wmem_alloc_get(sk);

1212 1213
		return put_user(amount, (int __user *)arg);
	}
L
Linus Torvalds 已提交
1214

1215 1216
	case SIOCINQ:
	{
E
Eric Dumazet 已提交
1217
		unsigned int amount = first_packet_length(sk);
1218

E
Eric Dumazet 已提交
1219
		if (amount)
1220 1221 1222 1223 1224
			/*
			 * We will only return the amount
			 * of this packet since that is all
			 * that will be read.
			 */
E
Eric Dumazet 已提交
1225 1226
			amount -= sizeof(struct udphdr);

1227 1228
		return put_user(amount, (int __user *)arg);
	}
L
Linus Torvalds 已提交
1229

1230 1231
	default:
		return -ENOIOCTLCMD;
L
Linus Torvalds 已提交
1232
	}
1233 1234

	return 0;
L
Linus Torvalds 已提交
1235
}
E
Eric Dumazet 已提交
1236
EXPORT_SYMBOL(udp_ioctl);
L
Linus Torvalds 已提交
1237

1238 1239 1240 1241 1242 1243 1244 1245 1246
/*
 * 	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);
1247
	DECLARE_SOCKADDR(struct sockaddr_in *, sin, msg->msg_name);
1248
	struct sk_buff *skb;
1249
	unsigned int ulen, copied;
1250
	int peeked, off = 0;
1251 1252
	int err;
	int is_udplite = IS_UDPLITE(sk);
1253
	bool slow;
1254 1255

	if (flags & MSG_ERRQUEUE)
1256
		return ip_recv_error(sk, msg, len, addr_len);
1257 1258 1259

try_again:
	skb = __skb_recv_datagram(sk, flags | (noblock ? MSG_DONTWAIT : 0),
1260
				  &peeked, &off, &err);
1261 1262 1263 1264
	if (!skb)
		goto out;

	ulen = skb->len - sizeof(struct udphdr);
1265 1266 1267 1268
	copied = len;
	if (copied > ulen)
		copied = ulen;
	else if (copied < ulen)
1269 1270 1271 1272 1273 1274 1275 1276
		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.
	 */

1277
	if (copied < ulen || UDP_SKB_CB(skb)->partial_cov) {
1278 1279 1280 1281 1282 1283
		if (udp_lib_checksum_complete(skb))
			goto csum_copy_err;
	}

	if (skb_csum_unnecessary(skb))
		err = skb_copy_datagram_iovec(skb, sizeof(struct udphdr),
1284
					      msg->msg_iov, copied);
1285
	else {
E
Eric Dumazet 已提交
1286 1287 1288
		err = skb_copy_and_csum_datagram_iovec(skb,
						       sizeof(struct udphdr),
						       msg->msg_iov);
1289 1290 1291 1292 1293

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

1294 1295
	if (unlikely(err)) {
		trace_kfree_skb(skb, udp_recvmsg);
1296 1297 1298 1299 1300
		if (!peeked) {
			atomic_inc(&sk->sk_drops);
			UDP_INC_STATS_USER(sock_net(sk),
					   UDP_MIB_INERRORS, is_udplite);
		}
1301
		goto out_free;
1302
	}
1303 1304

	if (!peeked)
1305 1306
		UDP_INC_STATS_USER(sock_net(sk),
				UDP_MIB_INDATAGRAMS, is_udplite);
1307

1308
	sock_recv_ts_and_drops(msg, sk, skb);
1309 1310

	/* Copy the address. */
E
Eric Dumazet 已提交
1311
	if (sin) {
1312 1313 1314 1315
		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));
1316
		*addr_len = sizeof(*sin);
1317 1318 1319 1320
	}
	if (inet->cmsg_flags)
		ip_cmsg_recv(msg, skb);

1321
	err = copied;
1322 1323 1324 1325
	if (flags & MSG_TRUNC)
		err = ulen;

out_free:
1326
	skb_free_datagram_locked(sk, skb);
1327 1328 1329 1330
out:
	return err;

csum_copy_err:
1331
	slow = lock_sock_fast(sk);
1332 1333
	if (!skb_kill_datagram(sk, skb, flags)) {
		UDP_INC_STATS_USER(sock_net(sk), UDP_MIB_CSUMERRORS, is_udplite);
1334
		UDP_INC_STATS_USER(sock_net(sk), UDP_MIB_INERRORS, is_udplite);
1335
	}
1336
	unlock_sock_fast(sk, slow);
1337 1338 1339

	if (noblock)
		return -EAGAIN;
1340 1341 1342

	/* starting over for a new packet */
	msg->msg_flags &= ~MSG_TRUNC;
1343 1344 1345 1346
	goto try_again;
}


L
Linus Torvalds 已提交
1347 1348 1349 1350 1351 1352
int udp_disconnect(struct sock *sk, int flags)
{
	struct inet_sock *inet = inet_sk(sk);
	/*
	 *	1003.1g - break association.
	 */
1353

L
Linus Torvalds 已提交
1354
	sk->sk_state = TCP_CLOSE;
E
Eric Dumazet 已提交
1355 1356
	inet->inet_daddr = 0;
	inet->inet_dport = 0;
1357
	sock_rps_reset_rxhash(sk);
L
Linus Torvalds 已提交
1358 1359 1360 1361 1362 1363
	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
1364
		inet->inet_sport = 0;
L
Linus Torvalds 已提交
1365 1366 1367 1368
	}
	sk_dst_reset(sk);
	return 0;
}
E
Eric Dumazet 已提交
1369
EXPORT_SYMBOL(udp_disconnect);
L
Linus Torvalds 已提交
1370

1371 1372
void udp_lib_unhash(struct sock *sk)
{
1373 1374
	if (sk_hashed(sk)) {
		struct udp_table *udptable = sk->sk_prot->h.udp_table;
1375 1376 1377 1378 1379
		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);
1380

1381 1382
		spin_lock_bh(&hslot->lock);
		if (sk_nulls_del_node_init_rcu(sk)) {
E
Eric Dumazet 已提交
1383
			hslot->count--;
E
Eric Dumazet 已提交
1384
			inet_sk(sk)->inet_num = 0;
1385
			sock_prot_inuse_add(sock_net(sk), sk->sk_prot, -1);
1386 1387 1388 1389 1390

			spin_lock(&hslot2->lock);
			hlist_nulls_del_init_rcu(&udp_sk(sk)->udp_portaddr_node);
			hslot2->count--;
			spin_unlock(&hslot2->lock);
1391 1392
		}
		spin_unlock_bh(&hslot->lock);
1393 1394 1395 1396
	}
}
EXPORT_SYMBOL(udp_lib_unhash);

E
Eric Dumazet 已提交
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 1423 1424 1425 1426 1427 1428 1429 1430 1431 1432 1433 1434 1435 1436 1437 1438 1439
/*
 * 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
1440 1441
static int __udp_queue_rcv_skb(struct sock *sk, struct sk_buff *skb)
{
T
Tom Herbert 已提交
1442
	int rc;
E
Eric Dumazet 已提交
1443

1444
	if (inet_sk(sk)->inet_daddr) {
1445
		sock_rps_save_rxhash(sk, skb);
1446 1447
		sk_mark_napi_id(sk, skb);
	}
T
Tom Herbert 已提交
1448

1449
	rc = sock_queue_rcv_skb(sk, skb);
E
Eric Dumazet 已提交
1450 1451
	if (rc < 0) {
		int is_udplite = IS_UDPLITE(sk);
H
Herbert Xu 已提交
1452 1453

		/* Note that an ENOMEM error is charged twice */
E
Eric Dumazet 已提交
1454
		if (rc == -ENOMEM)
H
Herbert Xu 已提交
1455 1456
			UDP_INC_STATS_BH(sock_net(sk), UDP_MIB_RCVBUFERRORS,
					 is_udplite);
E
Eric Dumazet 已提交
1457 1458
		UDP_INC_STATS_BH(sock_net(sk), UDP_MIB_INERRORS, is_udplite);
		kfree_skb(skb);
1459
		trace_udp_fail_queue_rcv_skb(rc, sk);
E
Eric Dumazet 已提交
1460
		return -1;
H
Herbert Xu 已提交
1461 1462 1463 1464 1465 1466
	}

	return 0;

}

1467 1468 1469 1470 1471 1472 1473 1474
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);

1475 1476 1477 1478 1479 1480 1481 1482
/* 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
1483
int udp_queue_rcv_skb(struct sock *sk, struct sk_buff *skb)
1484 1485 1486 1487 1488 1489 1490 1491 1492 1493 1494 1495
{
	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);

1496
	if (static_key_false(&udp_encap_needed) && up->encap_type) {
1497 1498
		int (*encap_rcv)(struct sock *sk, struct sk_buff *skb);

1499 1500 1501 1502 1503 1504 1505 1506 1507 1508 1509 1510
		/*
		 * 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 */
1511 1512
		encap_rcv = ACCESS_ONCE(up->encap_rcv);
		if (skb->len > sizeof(struct udphdr) && encap_rcv != NULL) {
1513 1514
			int ret;

1515 1516 1517 1518
			/* Verify checksum before giving to encap */
			if (udp_lib_checksum_complete(skb))
				goto csum_error;

1519
			ret = encap_rcv(sk, skb);
1520
			if (ret <= 0) {
1521 1522
				UDP_INC_STATS_BH(sock_net(sk),
						 UDP_MIB_INDATAGRAMS,
1523 1524 1525 1526 1527 1528 1529 1530 1531 1532 1533 1534 1535 1536 1537 1538 1539 1540 1541 1542 1543 1544 1545 1546 1547
						 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  */
1548 1549
			LIMIT_NETDEBUG(KERN_WARNING "UDPLite: partial coverage %d while full coverage %d requested\n",
				       UDP_SKB_CB(skb)->cscov, skb->len);
1550 1551 1552 1553 1554 1555 1556 1557 1558
			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) {
1559 1560
			LIMIT_NETDEBUG(KERN_WARNING "UDPLite: coverage %d too small, need min %d\n",
				       UDP_SKB_CB(skb)->cscov, up->pcrlen);
1561 1562 1563 1564
			goto drop;
		}
	}

1565 1566
	if (rcu_access_pointer(sk->sk_filter) &&
	    udp_lib_checksum_complete(skb))
1567
		goto csum_error;
1568

1569

1570
	if (sk_rcvqueues_full(sk, sk->sk_rcvbuf)) {
1571 1572
		UDP_INC_STATS_BH(sock_net(sk), UDP_MIB_RCVBUFERRORS,
				 is_udplite);
1573
		goto drop;
1574
	}
1575

H
Herbert Xu 已提交
1576
	rc = 0;
1577

1578
	ipv4_pktinfo_prepare(sk, skb);
H
Herbert Xu 已提交
1579 1580 1581
	bh_lock_sock(sk);
	if (!sock_owned_by_user(sk))
		rc = __udp_queue_rcv_skb(sk, skb);
1582
	else if (sk_add_backlog(sk, skb, sk->sk_rcvbuf)) {
Z
Zhu Yi 已提交
1583 1584 1585
		bh_unlock_sock(sk);
		goto drop;
	}
H
Herbert Xu 已提交
1586 1587 1588
	bh_unlock_sock(sk);

	return rc;
1589

1590 1591
csum_error:
	UDP_INC_STATS_BH(sock_net(sk), UDP_MIB_CSUMERRORS, is_udplite);
1592
drop:
1593
	UDP_INC_STATS_BH(sock_net(sk), UDP_MIB_INERRORS, is_udplite);
E
Eric Dumazet 已提交
1594
	atomic_inc(&sk->sk_drops);
1595 1596 1597 1598
	kfree_skb(skb);
	return -1;
}

1599 1600 1601 1602 1603 1604

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;
1605
	struct sock *sk;
1606 1607

	for (i = 0; i < count; i++) {
1608
		sk = stack[i];
1609 1610 1611
		if (likely(skb1 == NULL))
			skb1 = (i == final) ? skb : skb_clone(skb, GFP_ATOMIC);

1612 1613 1614 1615 1616 1617 1618 1619 1620
		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)
1621
			skb1 = NULL;
1622 1623

		sock_put(sk);
1624 1625 1626 1627 1628
	}
	if (unlikely(skb1))
		kfree_skb(skb1);
}

1629
/* For TCP sockets, sk_rx_dst is protected by socket lock
1630
 * For UDP, we use xchg() to guard against concurrent changes.
1631 1632
 */
static void udp_sk_rx_dst_set(struct sock *sk, struct dst_entry *dst)
S
Shawn Bohrer 已提交
1633
{
1634 1635
	struct dst_entry *old;

1636 1637 1638
	dst_hold(dst);
	old = xchg(&sk->sk_rx_dst, dst);
	dst_release(old);
S
Shawn Bohrer 已提交
1639 1640
}

1641 1642 1643
/*
 *	Multicasts and broadcasts go to each listener.
 *
1644
 *	Note: called only from the BH handler context.
1645
 */
1646
static int __udp4_lib_mcast_deliver(struct net *net, struct sk_buff *skb,
1647 1648
				    struct udphdr  *uh,
				    __be32 saddr, __be32 daddr,
1649
				    struct udp_table *udptable)
1650
{
1651
	struct sock *sk, *stack[256 / sizeof(struct sock *)];
1652 1653 1654 1655
	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;
1656 1657 1658 1659 1660 1661 1662 1663 1664 1665 1666
	unsigned int count = 0, offset = offsetof(typeof(*sk), sk_nulls_node);
	unsigned int hash2 = 0, hash2_any = 0, use_hash2 = (hslot->count > 10);

	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);
	}
1667

1668
	spin_lock(&hslot->lock);
1669
	sk_nulls_for_each_entry_offset(sk, node, &hslot->head, offset) {
1670 1671 1672 1673 1674 1675 1676 1677 1678
		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);
				count = 0;
			}
			stack[count++] = sk;
1679
			sock_hold(sk);
1680 1681 1682
		}
	}

1683
	spin_unlock(&hslot->lock);
1684

1685 1686 1687 1688 1689 1690
	/* 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;
	}

1691 1692 1693 1694 1695 1696 1697 1698
	/*
	 * do the slow work with no lock held
	 */
	if (count) {
		flush_stack(stack, count, skb, count - 1);
	} else {
		kfree_skb(skb);
	}
1699 1700 1701 1702 1703 1704 1705 1706 1707 1708 1709 1710 1711 1712 1713 1714 1715 1716 1717 1718 1719 1720
	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;
	}

1721 1722
	return skb_checksum_init_zero_check(skb, proto, uh->check,
					    inet_compute_pseudo);
1723 1724 1725 1726 1727 1728
}

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

1729
int __udp4_lib_rcv(struct sk_buff *skb, struct udp_table *udptable,
1730 1731 1732
		   int proto)
{
	struct sock *sk;
1733
	struct udphdr *uh;
1734
	unsigned short ulen;
E
Eric Dumazet 已提交
1735
	struct rtable *rt = skb_rtable(skb);
1736
	__be32 saddr, daddr;
1737
	struct net *net = dev_net(skb->dev);
1738 1739 1740 1741 1742 1743 1744

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

1745
	uh   = udp_hdr(skb);
1746
	ulen = ntohs(uh->len);
1747 1748 1749
	saddr = ip_hdr(skb)->saddr;
	daddr = ip_hdr(skb)->daddr;

1750 1751 1752 1753 1754 1755 1756 1757 1758 1759 1760 1761 1762
	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;

1763 1764
	sk = skb_steal_sock(skb);
	if (sk) {
1765
		struct dst_entry *dst = skb_dst(skb);
S
Shawn Bohrer 已提交
1766 1767
		int ret;

1768 1769
		if (unlikely(sk->sk_rx_dst != dst))
			udp_sk_rx_dst_set(sk, dst);
1770

S
Shawn Bohrer 已提交
1771
		ret = udp_queue_rcv_skb(sk, skb);
1772
		sock_put(sk);
S
Shawn Bohrer 已提交
1773 1774 1775 1776 1777 1778 1779 1780 1781 1782 1783 1784 1785
		/* 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);
	}
1786 1787

	if (sk != NULL) {
1788 1789 1790
		int ret;

		ret = udp_queue_rcv_skb(sk, skb);
1791 1792 1793 1794 1795 1796 1797 1798 1799 1800 1801 1802 1803 1804 1805 1806 1807 1808
		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;

1809
	UDP_INC_STATS_BH(net, UDP_MIB_NOPORTS, proto == IPPROTO_UDPLITE);
1810 1811 1812 1813 1814 1815 1816 1817 1818 1819
	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:
1820
	LIMIT_NETDEBUG(KERN_DEBUG "UDP%s: short packet: From %pI4:%u %d/%d to %pI4:%u\n",
1821 1822 1823 1824
		       proto == IPPROTO_UDPLITE ? "Lite" : "",
		       &saddr, ntohs(uh->source),
		       ulen, skb->len,
		       &daddr, ntohs(uh->dest));
1825 1826 1827 1828 1829 1830 1831
	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).
	 */
1832
	LIMIT_NETDEBUG(KERN_DEBUG "UDP%s: bad checksum. From %pI4:%u to %pI4:%u ulen %d\n",
1833 1834
		       proto == IPPROTO_UDPLITE ? "Lite" : "",
		       &saddr, ntohs(uh->source), &daddr, ntohs(uh->dest),
1835
		       ulen);
1836
	UDP_INC_STATS_BH(net, UDP_MIB_CSUMERRORS, proto == IPPROTO_UDPLITE);
1837
drop:
1838
	UDP_INC_STATS_BH(net, UDP_MIB_INERRORS, proto == IPPROTO_UDPLITE);
1839 1840 1841 1842
	kfree_skb(skb);
	return 0;
}

S
Shawn Bohrer 已提交
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];

1857 1858 1859 1860
	/* Do not bother scanning a too big list */
	if (hslot->count > 10)
		return NULL;

S
Shawn Bohrer 已提交
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
	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;
1886
		else if (unlikely(!__udp_is_mcast_sock(net, result,
S
Shawn Bohrer 已提交
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
						       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];
1913
	INET_ADDR_COOKIE(acookie, rmt_addr, loc_addr);
S
Shawn Bohrer 已提交
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
	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)
{
1942 1943 1944
	struct net *net = dev_net(skb->dev);
	const struct iphdr *iph;
	const struct udphdr *uh;
S
Shawn Bohrer 已提交
1945 1946 1947 1948 1949 1950 1951 1952
	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;

1953 1954 1955
	iph = ip_hdr(skb);
	uh = udp_hdr(skb);

S
Shawn Bohrer 已提交
1956 1957 1958 1959 1960 1961 1962 1963 1964 1965 1966 1967 1968 1969 1970 1971 1972 1973 1974 1975 1976 1977 1978
	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);
}

1979 1980
int udp_rcv(struct sk_buff *skb)
{
1981
	return __udp4_lib_rcv(skb, &udp_table, IPPROTO_UDP);
1982 1983
}

1984
void udp_destroy_sock(struct sock *sk)
1985
{
T
Tom Parkin 已提交
1986
	struct udp_sock *up = udp_sk(sk);
1987
	bool slow = lock_sock_fast(sk);
1988
	udp_flush_pending_frames(sk);
1989
	unlock_sock_fast(sk, slow);
T
Tom Parkin 已提交
1990 1991 1992 1993 1994 1995
	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);
	}
1996 1997
}

L
Linus Torvalds 已提交
1998 1999 2000
/*
 *	Socket option code for UDP
 */
2001
int udp_lib_setsockopt(struct sock *sk, int level, int optname,
2002
		       char __user *optval, unsigned int optlen,
2003
		       int (*push_pending_frames)(struct sock *))
L
Linus Torvalds 已提交
2004 2005
{
	struct udp_sock *up = udp_sk(sk);
2006
	int val, valbool;
L
Linus Torvalds 已提交
2007
	int err = 0;
W
Wang Chen 已提交
2008
	int is_udplite = IS_UDPLITE(sk);
L
Linus Torvalds 已提交
2009

E
Eric Dumazet 已提交
2010
	if (optlen < sizeof(int))
L
Linus Torvalds 已提交
2011 2012 2013 2014 2015
		return -EINVAL;

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

2016 2017
	valbool = val ? 1 : 0;

2018
	switch (optname) {
L
Linus Torvalds 已提交
2019 2020 2021 2022 2023 2024
	case UDP_CORK:
		if (val != 0) {
			up->corkflag = 1;
		} else {
			up->corkflag = 0;
			lock_sock(sk);
2025
			(*push_pending_frames)(sk);
L
Linus Torvalds 已提交
2026 2027 2028
			release_sock(sk);
		}
		break;
2029

L
Linus Torvalds 已提交
2030 2031 2032 2033 2034
	case UDP_ENCAP:
		switch (val) {
		case 0:
		case UDP_ENCAP_ESPINUDP:
		case UDP_ENCAP_ESPINUDP_NON_IKE:
2035 2036
			up->encap_rcv = xfrm4_udp_encap_rcv;
			/* FALLTHROUGH */
2037
		case UDP_ENCAP_L2TPINUDP:
L
Linus Torvalds 已提交
2038
			up->encap_type = val;
2039
			udp_encap_enable();
L
Linus Torvalds 已提交
2040 2041 2042 2043 2044 2045 2046
			break;
		default:
			err = -ENOPROTOOPT;
			break;
		}
		break;

2047 2048 2049 2050 2051 2052 2053 2054
	case UDP_NO_CHECK6_TX:
		up->no_check6_tx = valbool;
		break;

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

2055 2056 2057 2058 2059 2060
	/*
	 * 	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
2061
		if (!is_udplite)         /* Disable the option on UDP sockets */
2062 2063 2064
			return -ENOPROTOOPT;
		if (val != 0 && val < 8) /* Illegal coverage: use default (8) */
			val = 8;
2065 2066
		else if (val > USHRT_MAX)
			val = USHRT_MAX;
2067 2068 2069 2070
		up->pcslen = val;
		up->pcflag |= UDPLITE_SEND_CC;
		break;

2071 2072
	/* 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
2073 2074
	 * used, this again means full checksum coverage.                     */
	case UDPLITE_RECV_CSCOV:
W
Wang Chen 已提交
2075
		if (!is_udplite)         /* Disable the option on UDP sockets */
2076 2077 2078
			return -ENOPROTOOPT;
		if (val != 0 && val < 8) /* Avoid silly minimal values.       */
			val = 8;
2079 2080
		else if (val > USHRT_MAX)
			val = USHRT_MAX;
2081 2082 2083 2084
		up->pcrlen = val;
		up->pcflag |= UDPLITE_RECV_CC;
		break;

L
Linus Torvalds 已提交
2085 2086 2087
	default:
		err = -ENOPROTOOPT;
		break;
2088
	}
L
Linus Torvalds 已提交
2089 2090 2091

	return err;
}
E
Eric Dumazet 已提交
2092
EXPORT_SYMBOL(udp_lib_setsockopt);
L
Linus Torvalds 已提交
2093

2094
int udp_setsockopt(struct sock *sk, int level, int optname,
2095
		   char __user *optval, unsigned int optlen)
2096 2097 2098 2099 2100 2101 2102 2103 2104
{
	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,
2105
			  char __user *optval, unsigned int optlen)
2106 2107 2108 2109 2110 2111 2112 2113
{
	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

2114 2115
int udp_lib_getsockopt(struct sock *sk, int level, int optname,
		       char __user *optval, int __user *optlen)
L
Linus Torvalds 已提交
2116 2117 2118 2119
{
	struct udp_sock *up = udp_sk(sk);
	int val, len;

E
Eric Dumazet 已提交
2120
	if (get_user(len, optlen))
L
Linus Torvalds 已提交
2121 2122 2123
		return -EFAULT;

	len = min_t(unsigned int, len, sizeof(int));
2124

2125
	if (len < 0)
L
Linus Torvalds 已提交
2126 2127
		return -EINVAL;

2128
	switch (optname) {
L
Linus Torvalds 已提交
2129 2130 2131 2132 2133 2134 2135 2136
	case UDP_CORK:
		val = up->corkflag;
		break;

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

2137 2138 2139 2140 2141 2142 2143 2144
	case UDP_NO_CHECK6_TX:
		val = up->no_check6_tx;
		break;

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

2145 2146 2147 2148 2149 2150 2151 2152 2153 2154
	/* 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
2155 2156
	default:
		return -ENOPROTOOPT;
2157
	}
L
Linus Torvalds 已提交
2158

2159
	if (put_user(len, optlen))
2160
		return -EFAULT;
E
Eric Dumazet 已提交
2161
	if (copy_to_user(optval, &val, len))
L
Linus Torvalds 已提交
2162
		return -EFAULT;
2163
	return 0;
L
Linus Torvalds 已提交
2164
}
E
Eric Dumazet 已提交
2165
EXPORT_SYMBOL(udp_lib_getsockopt);
L
Linus Torvalds 已提交
2166

2167 2168 2169 2170 2171 2172 2173 2174 2175 2176 2177 2178 2179 2180 2181 2182 2183
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
2184 2185 2186 2187 2188 2189
/**
 * 	udp_poll - wait for a UDP event.
 *	@file - file struct
 *	@sock - socket
 *	@wait - poll table
 *
2190
 *	This is same as datagram poll, except for the special case of
L
Linus Torvalds 已提交
2191 2192 2193 2194 2195 2196 2197 2198 2199 2200
 *	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;
2201

2202 2203
	sock_rps_record_flow(sk);

L
Linus Torvalds 已提交
2204
	/* Check for false positives due to checksum errors */
E
Eric Dumazet 已提交
2205 2206 2207
	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
2208 2209

	return mask;
2210

L
Linus Torvalds 已提交
2211
}
E
Eric Dumazet 已提交
2212
EXPORT_SYMBOL(udp_poll);
L
Linus Torvalds 已提交
2213

2214 2215 2216 2217 2218 2219 2220 2221 2222 2223 2224 2225 2226
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
2227
	.backlog_rcv	   = __udp_queue_rcv_skb,
2228
	.release_cb	   = ip4_datagram_release_cb,
2229 2230
	.hash		   = udp_lib_hash,
	.unhash		   = udp_lib_unhash,
E
Eric Dumazet 已提交
2231
	.rehash		   = udp_v4_rehash,
2232 2233 2234 2235 2236 2237
	.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),
2238
	.slab_flags	   = SLAB_DESTROY_BY_RCU,
2239
	.h.udp_table	   = &udp_table,
2240 2241 2242 2243
#ifdef CONFIG_COMPAT
	.compat_setsockopt = compat_udp_setsockopt,
	.compat_getsockopt = compat_udp_getsockopt,
#endif
2244
	.clear_sk	   = sk_prot_clear_portaddr_nulls,
2245
};
E
Eric Dumazet 已提交
2246
EXPORT_SYMBOL(udp_prot);
L
Linus Torvalds 已提交
2247 2248 2249 2250

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

2251
static struct sock *udp_get_first(struct seq_file *seq, int start)
L
Linus Torvalds 已提交
2252 2253 2254
{
	struct sock *sk;
	struct udp_iter_state *state = seq->private;
2255
	struct net *net = seq_file_net(seq);
L
Linus Torvalds 已提交
2256

2257 2258
	for (state->bucket = start; state->bucket <= state->udp_table->mask;
	     ++state->bucket) {
2259
		struct hlist_nulls_node *node;
2260
		struct udp_hslot *hslot = &state->udp_table->hash[state->bucket];
2261 2262 2263 2264

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

2265
		spin_lock_bh(&hslot->lock);
2266
		sk_nulls_for_each(sk, node, &hslot->head) {
2267
			if (!net_eq(sock_net(sk), net))
2268
				continue;
L
Linus Torvalds 已提交
2269 2270 2271
			if (sk->sk_family == state->family)
				goto found;
		}
2272
		spin_unlock_bh(&hslot->lock);
L
Linus Torvalds 已提交
2273 2274 2275 2276 2277 2278 2279 2280 2281
	}
	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;
2282
	struct net *net = seq_file_net(seq);
L
Linus Torvalds 已提交
2283 2284

	do {
2285
		sk = sk_nulls_next(sk);
2286
	} while (sk && (!net_eq(sock_net(sk), net) || sk->sk_family != state->family));
L
Linus Torvalds 已提交
2287

2288
	if (!sk) {
2289
		if (state->bucket <= state->udp_table->mask)
2290
			spin_unlock_bh(&state->udp_table->hash[state->bucket].lock);
2291
		return udp_get_first(seq, state->bucket + 1);
L
Linus Torvalds 已提交
2292 2293 2294 2295 2296 2297
	}
	return sk;
}

static struct sock *udp_get_idx(struct seq_file *seq, loff_t pos)
{
2298
	struct sock *sk = udp_get_first(seq, 0);
L
Linus Torvalds 已提交
2299 2300

	if (sk)
2301
		while (pos && (sk = udp_get_next(seq, sk)) != NULL)
L
Linus Torvalds 已提交
2302 2303 2304 2305 2306 2307
			--pos;
	return pos ? NULL : sk;
}

static void *udp_seq_start(struct seq_file *seq, loff_t *pos)
{
2308
	struct udp_iter_state *state = seq->private;
2309
	state->bucket = MAX_UDP_PORTS;
2310

2311
	return *pos ? udp_get_idx(seq, *pos-1) : SEQ_START_TOKEN;
L
Linus Torvalds 已提交
2312 2313 2314 2315 2316 2317
}

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

2318
	if (v == SEQ_START_TOKEN)
L
Linus Torvalds 已提交
2319 2320 2321 2322 2323 2324 2325 2326 2327 2328
		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)
{
2329 2330
	struct udp_iter_state *state = seq->private;

2331
	if (state->bucket <= state->udp_table->mask)
2332
		spin_unlock_bh(&state->udp_table->hash[state->bucket].lock);
L
Linus Torvalds 已提交
2333 2334
}

2335
int udp_seq_open(struct inode *inode, struct file *file)
L
Linus Torvalds 已提交
2336
{
A
Al Viro 已提交
2337
	struct udp_seq_afinfo *afinfo = PDE_DATA(inode);
2338 2339
	struct udp_iter_state *s;
	int err;
2340

2341 2342 2343 2344
	err = seq_open_net(inode, file, &afinfo->seq_ops,
			   sizeof(struct udp_iter_state));
	if (err < 0)
		return err;
2345

2346
	s = ((struct seq_file *)file->private_data)->private;
L
Linus Torvalds 已提交
2347
	s->family		= afinfo->family;
2348
	s->udp_table		= afinfo->udp_table;
2349
	return err;
2350
}
2351
EXPORT_SYMBOL(udp_seq_open);
2352

L
Linus Torvalds 已提交
2353
/* ------------------------------------------------------------------------ */
2354
int udp_proc_register(struct net *net, struct udp_seq_afinfo *afinfo)
L
Linus Torvalds 已提交
2355 2356 2357 2358
{
	struct proc_dir_entry *p;
	int rc = 0;

2359 2360 2361 2362
	afinfo->seq_ops.start		= udp_seq_start;
	afinfo->seq_ops.next		= udp_seq_next;
	afinfo->seq_ops.stop		= udp_seq_stop;

2363
	p = proc_create_data(afinfo->name, S_IRUGO, net->proc_net,
2364
			     afinfo->seq_fops, afinfo);
2365
	if (!p)
L
Linus Torvalds 已提交
2366 2367 2368
		rc = -ENOMEM;
	return rc;
}
E
Eric Dumazet 已提交
2369
EXPORT_SYMBOL(udp_proc_register);
L
Linus Torvalds 已提交
2370

2371
void udp_proc_unregister(struct net *net, struct udp_seq_afinfo *afinfo)
L
Linus Torvalds 已提交
2372
{
2373
	remove_proc_entry(afinfo->name, net->proc_net);
L
Linus Torvalds 已提交
2374
}
E
Eric Dumazet 已提交
2375
EXPORT_SYMBOL(udp_proc_unregister);
2376 2377

/* ------------------------------------------------------------------------ */
2378
static void udp4_format_sock(struct sock *sp, struct seq_file *f,
2379
		int bucket)
2380 2381
{
	struct inet_sock *inet = inet_sk(sp);
E
Eric Dumazet 已提交
2382 2383 2384 2385
	__be32 dest = inet->inet_daddr;
	__be32 src  = inet->inet_rcv_saddr;
	__u16 destp	  = ntohs(inet->inet_dport);
	__u16 srcp	  = ntohs(inet->inet_sport);
2386

2387
	seq_printf(f, "%5d: %08X:%04X %08X:%04X"
2388
		" %02X %08X:%08X %02X:%08lX %08X %5u %8d %lu %d %pK %d",
2389
		bucket, src, srcp, dest, destp, sp->sk_state,
2390 2391
		sk_wmem_alloc_get(sp),
		sk_rmem_alloc_get(sp),
2392 2393 2394
		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
2395
		atomic_read(&sp->sk_refcnt), sp,
2396
		atomic_read(&sp->sk_drops));
2397 2398 2399 2400
}

int udp4_seq_show(struct seq_file *seq, void *v)
{
2401
	seq_setwidth(seq, 127);
2402
	if (v == SEQ_START_TOKEN)
2403
		seq_puts(seq, "  sl  local_address rem_address   st tx_queue "
2404
			   "rx_queue tr tm->when retrnsmt   uid  timeout "
E
Eric Dumazet 已提交
2405
			   "inode ref pointer drops");
2406 2407 2408
	else {
		struct udp_iter_state *state = seq->private;

2409
		udp4_format_sock(v, seq, state->bucket);
2410
	}
2411
	seq_pad(seq, '\n');
2412 2413 2414
	return 0;
}

2415 2416 2417 2418 2419 2420 2421 2422
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
};

2423 2424 2425 2426
/* ------------------------------------------------------------------------ */
static struct udp_seq_afinfo udp4_seq_afinfo = {
	.name		= "udp",
	.family		= AF_INET,
2427
	.udp_table	= &udp_table,
2428
	.seq_fops	= &udp_afinfo_seq_fops,
2429 2430 2431
	.seq_ops	= {
		.show		= udp4_seq_show,
	},
2432 2433
};

2434
static int __net_init udp4_proc_init_net(struct net *net)
2435 2436 2437 2438
{
	return udp_proc_register(net, &udp4_seq_afinfo);
}

2439
static void __net_exit udp4_proc_exit_net(struct net *net)
2440 2441 2442 2443 2444 2445 2446 2447 2448
{
	udp_proc_unregister(net, &udp4_seq_afinfo);
}

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

2449 2450
int __init udp4_proc_init(void)
{
2451
	return register_pernet_subsys(&udp4_net_ops);
2452 2453 2454 2455
}

void udp4_proc_exit(void)
{
2456
	unregister_pernet_subsys(&udp4_net_ops);
2457
}
L
Linus Torvalds 已提交
2458 2459
#endif /* CONFIG_PROC_FS */

2460 2461
static __initdata unsigned long uhash_entries;
static int __init set_uhash_entries(char *str)
2462
{
2463 2464
	ssize_t ret;

2465 2466
	if (!str)
		return 0;
2467 2468 2469 2470 2471

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

2472 2473 2474 2475 2476
	if (uhash_entries && uhash_entries < UDP_HTABLE_SIZE_MIN)
		uhash_entries = UDP_HTABLE_SIZE_MIN;
	return 1;
}
__setup("uhash_entries=", set_uhash_entries);
2477

2478 2479 2480 2481
void __init udp_table_init(struct udp_table *table, const char *name)
{
	unsigned int i;

2482 2483 2484 2485 2486 2487 2488 2489 2490 2491
	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);

2492
	table->hash2 = table->hash + (table->mask + 1);
2493
	for (i = 0; i <= table->mask; i++) {
2494
		INIT_HLIST_NULLS_HEAD(&table->hash[i].head, i);
E
Eric Dumazet 已提交
2495
		table->hash[i].count = 0;
2496 2497
		spin_lock_init(&table->hash[i].lock);
	}
2498 2499 2500 2501 2502
	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);
	}
2503 2504
}

H
Hideo Aoki 已提交
2505 2506
void __init udp_init(void)
{
2507
	unsigned long limit;
H
Hideo Aoki 已提交
2508

2509
	udp_table_init(&udp_table, "UDP");
2510
	limit = nr_free_buffer_pages() / 8;
H
Hideo Aoki 已提交
2511 2512 2513 2514 2515 2516 2517 2518
	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;
}