af_netlink.c 47.9 KB
Newer Older
L
Linus Torvalds 已提交
1 2 3
/*
 * NETLINK      Kernel-user communication protocol.
 *
4
 * 		Authors:	Alan Cox <alan@lxorguk.ukuu.org.uk>
L
Linus Torvalds 已提交
5 6 7 8 9 10
 * 				Alexey Kuznetsov <kuznet@ms2.inr.ac.ru>
 *
 *		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.
11
 *
L
Linus Torvalds 已提交
12 13 14 15
 * Tue Jun 26 14:36:48 MEST 2001 Herbert "herp" Rosmanith
 *                               added netlink_proto_exit
 * Tue Jan 22 18:32:44 BRST 2002 Arnaldo C. de Melo <acme@conectiva.com.br>
 * 				 use nlk_sk, as sk->protinfo is on a diet 8)
16 17 18 19 20 21
 * Fri Jul 22 19:51:12 MEST 2005 Harald Welte <laforge@gnumonks.org>
 * 				 - inc module use count of module that owns
 * 				   the kernel socket in case userspace opens
 * 				   socket of same protocol
 * 				 - remove all module support, since netlink is
 * 				   mandatory if CONFIG_NET=y these days
L
Linus Torvalds 已提交
22 23 24 25
 */

#include <linux/module.h>

26
#include <linux/capability.h>
L
Linus Torvalds 已提交
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
#include <linux/kernel.h>
#include <linux/init.h>
#include <linux/signal.h>
#include <linux/sched.h>
#include <linux/errno.h>
#include <linux/string.h>
#include <linux/stat.h>
#include <linux/socket.h>
#include <linux/un.h>
#include <linux/fcntl.h>
#include <linux/termios.h>
#include <linux/sockios.h>
#include <linux/net.h>
#include <linux/fs.h>
#include <linux/slab.h>
#include <asm/uaccess.h>
#include <linux/skbuff.h>
#include <linux/netdevice.h>
#include <linux/rtnetlink.h>
#include <linux/proc_fs.h>
#include <linux/seq_file.h>
#include <linux/notifier.h>
#include <linux/security.h>
#include <linux/jhash.h>
#include <linux/jiffies.h>
#include <linux/random.h>
#include <linux/bitops.h>
#include <linux/mm.h>
#include <linux/types.h>
A
Andrew Morton 已提交
56
#include <linux/audit.h>
57
#include <linux/mutex.h>
A
Andrew Morton 已提交
58

59
#include <net/net_namespace.h>
L
Linus Torvalds 已提交
60 61
#include <net/sock.h>
#include <net/scm.h>
62
#include <net/netlink.h>
L
Linus Torvalds 已提交
63

64
#define NLGRPSZ(x)	(ALIGN(x, sizeof(unsigned long) * 8) / 8)
65
#define NLGRPLONGS(x)	(NLGRPSZ(x)/sizeof(unsigned long))
L
Linus Torvalds 已提交
66 67 68 69 70 71

struct netlink_sock {
	/* struct sock has to be the first member of netlink_sock */
	struct sock		sk;
	u32			pid;
	u32			dst_pid;
72
	u32			dst_group;
73 74 75 76
	u32			flags;
	u32			subscriptions;
	u32			ngroups;
	unsigned long		*groups;
L
Linus Torvalds 已提交
77 78 79
	unsigned long		state;
	wait_queue_head_t	wait;
	struct netlink_callback	*cb;
80 81
	struct mutex		*cb_mutex;
	struct mutex		cb_def_mutex;
82
	void			(*netlink_rcv)(struct sk_buff *skb);
83
	struct module		*module;
L
Linus Torvalds 已提交
84 85
};

86 87 88
struct listeners {
	struct rcu_head		rcu;
	unsigned long		masks[0];
89 90
};

91
#define NETLINK_KERNEL_SOCKET	0x1
92
#define NETLINK_RECV_PKTINFO	0x2
93
#define NETLINK_BROADCAST_SEND_ERROR	0x4
94
#define NETLINK_RECV_NO_ENOBUFS	0x8
95

L
Linus Torvalds 已提交
96 97
static inline struct netlink_sock *nlk_sk(struct sock *sk)
{
D
Denis Cheng 已提交
98
	return container_of(sk, struct netlink_sock, sk);
L
Linus Torvalds 已提交
99 100
}

101
static inline int netlink_is_kernel(struct sock *sk)
102 103 104 105
{
	return nlk_sk(sk)->flags & NETLINK_KERNEL_SOCKET;
}

L
Linus Torvalds 已提交
106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121
struct nl_pid_hash {
	struct hlist_head *table;
	unsigned long rehash_time;

	unsigned int mask;
	unsigned int shift;

	unsigned int entries;
	unsigned int max_shift;

	u32 rnd;
};

struct netlink_table {
	struct nl_pid_hash hash;
	struct hlist_head mc_list;
122
	struct listeners __rcu *listeners;
L
Linus Torvalds 已提交
123
	unsigned int nl_nonroot;
124
	unsigned int groups;
125
	struct mutex *cb_mutex;
126
	struct module *module;
127
	int registered;
L
Linus Torvalds 已提交
128 129 130 131 132 133 134 135 136 137 138 139
};

static struct netlink_table *nl_table;

static DECLARE_WAIT_QUEUE_HEAD(nl_table_wait);

static int netlink_dump(struct sock *sk);
static void netlink_destroy_callback(struct netlink_callback *cb);

static DEFINE_RWLOCK(nl_table_lock);
static atomic_t nl_table_users = ATOMIC_INIT(0);

140
static ATOMIC_NOTIFIER_HEAD(netlink_chain);
L
Linus Torvalds 已提交
141

142
static inline u32 netlink_group_mask(u32 group)
143 144 145 146
{
	return group ? 1 << (group - 1) : 0;
}

147
static inline struct hlist_head *nl_pid_hashfn(struct nl_pid_hash *hash, u32 pid)
L
Linus Torvalds 已提交
148 149 150 151 152 153
{
	return &hash->table[jhash_1word(pid, hash->rnd) & hash->mask];
}

static void netlink_sock_destruct(struct sock *sk)
{
154 155 156 157 158 159 160 161
	struct netlink_sock *nlk = nlk_sk(sk);

	if (nlk->cb) {
		if (nlk->cb->done)
			nlk->cb->done(nlk->cb);
		netlink_destroy_callback(nlk->cb);
	}

L
Linus Torvalds 已提交
162 163 164
	skb_queue_purge(&sk->sk_receive_queue);

	if (!sock_flag(sk, SOCK_DEAD)) {
165
		printk(KERN_ERR "Freeing alive netlink socket %p\n", sk);
L
Linus Torvalds 已提交
166 167
		return;
	}
168 169 170 171

	WARN_ON(atomic_read(&sk->sk_rmem_alloc));
	WARN_ON(atomic_read(&sk->sk_wmem_alloc));
	WARN_ON(nlk_sk(sk)->groups);
L
Linus Torvalds 已提交
172 173
}

174 175
/* This lock without WQ_FLAG_EXCLUSIVE is good on UP and it is _very_ bad on
 * SMP. Look, when several writers sleep and reader wakes them up, all but one
L
Linus Torvalds 已提交
176 177 178 179
 * immediately hit write lock and grab all the cpus. Exclusive sleep solves
 * this, _but_ remember, it adds useless work on UP machines.
 */

180
void netlink_table_grab(void)
181
	__acquires(nl_table_lock)
L
Linus Torvalds 已提交
182
{
183 184
	might_sleep();

185
	write_lock_irq(&nl_table_lock);
L
Linus Torvalds 已提交
186 187 188 189 190

	if (atomic_read(&nl_table_users)) {
		DECLARE_WAITQUEUE(wait, current);

		add_wait_queue_exclusive(&nl_table_wait, &wait);
191
		for (;;) {
L
Linus Torvalds 已提交
192 193 194
			set_current_state(TASK_UNINTERRUPTIBLE);
			if (atomic_read(&nl_table_users) == 0)
				break;
195
			write_unlock_irq(&nl_table_lock);
L
Linus Torvalds 已提交
196
			schedule();
197
			write_lock_irq(&nl_table_lock);
L
Linus Torvalds 已提交
198 199 200 201 202 203 204
		}

		__set_current_state(TASK_RUNNING);
		remove_wait_queue(&nl_table_wait, &wait);
	}
}

205
void netlink_table_ungrab(void)
206
	__releases(nl_table_lock)
L
Linus Torvalds 已提交
207
{
208
	write_unlock_irq(&nl_table_lock);
L
Linus Torvalds 已提交
209 210 211
	wake_up(&nl_table_wait);
}

212
static inline void
L
Linus Torvalds 已提交
213 214 215 216 217 218 219 220 221
netlink_lock_table(void)
{
	/* read_lock() synchronizes us to netlink_table_grab */

	read_lock(&nl_table_lock);
	atomic_inc(&nl_table_users);
	read_unlock(&nl_table_lock);
}

222
static inline void
L
Linus Torvalds 已提交
223 224 225 226 227 228
netlink_unlock_table(void)
{
	if (atomic_dec_and_test(&nl_table_users))
		wake_up(&nl_table_wait);
}

229
static struct sock *netlink_lookup(struct net *net, int protocol, u32 pid)
L
Linus Torvalds 已提交
230 231 232 233 234 235 236 237 238
{
	struct nl_pid_hash *hash = &nl_table[protocol].hash;
	struct hlist_head *head;
	struct sock *sk;
	struct hlist_node *node;

	read_lock(&nl_table_lock);
	head = nl_pid_hashfn(hash, pid);
	sk_for_each(sk, node, head) {
239
		if (net_eq(sock_net(sk), net) && (nlk_sk(sk)->pid == pid)) {
L
Linus Torvalds 已提交
240 241 242 243 244 245 246 247 248 249
			sock_hold(sk);
			goto found;
		}
	}
	sk = NULL;
found:
	read_unlock(&nl_table_lock);
	return sk;
}

250
static struct hlist_head *nl_pid_hash_zalloc(size_t size)
L
Linus Torvalds 已提交
251 252
{
	if (size <= PAGE_SIZE)
E
Eric Dumazet 已提交
253
		return kzalloc(size, GFP_ATOMIC);
L
Linus Torvalds 已提交
254 255
	else
		return (struct hlist_head *)
E
Eric Dumazet 已提交
256 257
			__get_free_pages(GFP_ATOMIC | __GFP_ZERO,
					 get_order(size));
L
Linus Torvalds 已提交
258 259
}

260
static void nl_pid_hash_free(struct hlist_head *table, size_t size)
L
Linus Torvalds 已提交
261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285
{
	if (size <= PAGE_SIZE)
		kfree(table);
	else
		free_pages((unsigned long)table, get_order(size));
}

static int nl_pid_hash_rehash(struct nl_pid_hash *hash, int grow)
{
	unsigned int omask, mask, shift;
	size_t osize, size;
	struct hlist_head *otable, *table;
	int i;

	omask = mask = hash->mask;
	osize = size = (mask + 1) * sizeof(*table);
	shift = hash->shift;

	if (grow) {
		if (++shift > hash->max_shift)
			return 0;
		mask = mask * 2 + 1;
		size *= 2;
	}

E
Eric Dumazet 已提交
286
	table = nl_pid_hash_zalloc(size);
L
Linus Torvalds 已提交
287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323
	if (!table)
		return 0;

	otable = hash->table;
	hash->table = table;
	hash->mask = mask;
	hash->shift = shift;
	get_random_bytes(&hash->rnd, sizeof(hash->rnd));

	for (i = 0; i <= omask; i++) {
		struct sock *sk;
		struct hlist_node *node, *tmp;

		sk_for_each_safe(sk, node, tmp, &otable[i])
			__sk_add_node(sk, nl_pid_hashfn(hash, nlk_sk(sk)->pid));
	}

	nl_pid_hash_free(otable, osize);
	hash->rehash_time = jiffies + 10 * 60 * HZ;
	return 1;
}

static inline int nl_pid_hash_dilute(struct nl_pid_hash *hash, int len)
{
	int avg = hash->entries >> hash->shift;

	if (unlikely(avg > 1) && nl_pid_hash_rehash(hash, 1))
		return 1;

	if (unlikely(len > avg) && time_after(jiffies, hash->rehash_time)) {
		nl_pid_hash_rehash(hash, 0);
		return 1;
	}

	return 0;
}

324
static const struct proto_ops netlink_ops;
L
Linus Torvalds 已提交
325

326 327 328 329 330 331 332 333
static void
netlink_update_listeners(struct sock *sk)
{
	struct netlink_table *tbl = &nl_table[sk->sk_protocol];
	struct hlist_node *node;
	unsigned long mask;
	unsigned int i;

334
	for (i = 0; i < NLGRPLONGS(tbl->groups); i++) {
335
		mask = 0;
336 337 338 339
		sk_for_each_bound(sk, node, &tbl->mc_list) {
			if (i < NLGRPLONGS(nlk_sk(sk)->ngroups))
				mask |= nlk_sk(sk)->groups[i];
		}
340
		tbl->listeners->masks[i] = mask;
341 342 343 344 345
	}
	/* this function is only called with the netlink table "grabbed", which
	 * makes sure updates are visible before bind or setsockopt return. */
}

346
static int netlink_insert(struct sock *sk, struct net *net, u32 pid)
L
Linus Torvalds 已提交
347 348 349 350 351 352 353 354 355 356 357 358
{
	struct nl_pid_hash *hash = &nl_table[sk->sk_protocol].hash;
	struct hlist_head *head;
	int err = -EADDRINUSE;
	struct sock *osk;
	struct hlist_node *node;
	int len;

	netlink_table_grab();
	head = nl_pid_hashfn(hash, pid);
	len = 0;
	sk_for_each(osk, node, head) {
359
		if (net_eq(sock_net(osk), net) && (nlk_sk(osk)->pid == pid))
L
Linus Torvalds 已提交
360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388
			break;
		len++;
	}
	if (node)
		goto err;

	err = -EBUSY;
	if (nlk_sk(sk)->pid)
		goto err;

	err = -ENOMEM;
	if (BITS_PER_LONG > 32 && unlikely(hash->entries >= UINT_MAX))
		goto err;

	if (len && nl_pid_hash_dilute(hash, len))
		head = nl_pid_hashfn(hash, pid);
	hash->entries++;
	nlk_sk(sk)->pid = pid;
	sk_add_node(sk, head);
	err = 0;

err:
	netlink_table_ungrab();
	return err;
}

static void netlink_remove(struct sock *sk)
{
	netlink_table_grab();
389 390
	if (sk_del_node_init(sk))
		nl_table[sk->sk_protocol].hash.entries--;
391
	if (nlk_sk(sk)->subscriptions)
L
Linus Torvalds 已提交
392 393 394 395 396 397 398 399 400 401
		__sk_del_bind_node(sk);
	netlink_table_ungrab();
}

static struct proto netlink_proto = {
	.name	  = "NETLINK",
	.owner	  = THIS_MODULE,
	.obj_size = sizeof(struct netlink_sock),
};

402 403
static int __netlink_create(struct net *net, struct socket *sock,
			    struct mutex *cb_mutex, int protocol)
L
Linus Torvalds 已提交
404 405 406
{
	struct sock *sk;
	struct netlink_sock *nlk;
407 408 409

	sock->ops = &netlink_ops;

410
	sk = sk_alloc(net, PF_NETLINK, GFP_KERNEL, &netlink_proto);
411 412 413 414 415 416
	if (!sk)
		return -ENOMEM;

	sock_init_data(sock, sk);

	nlk = nlk_sk(sk);
417 418 419 420 421 422
	if (cb_mutex)
		nlk->cb_mutex = cb_mutex;
	else {
		nlk->cb_mutex = &nlk->cb_def_mutex;
		mutex_init(nlk->cb_mutex);
	}
423 424 425 426 427 428 429
	init_waitqueue_head(&nlk->wait);

	sk->sk_destruct = netlink_sock_destruct;
	sk->sk_protocol = protocol;
	return 0;
}

430 431
static int netlink_create(struct net *net, struct socket *sock, int protocol,
			  int kern)
432 433
{
	struct module *module = NULL;
434
	struct mutex *cb_mutex;
435
	struct netlink_sock *nlk;
436
	int err = 0;
L
Linus Torvalds 已提交
437 438 439 440 441 442

	sock->state = SS_UNCONNECTED;

	if (sock->type != SOCK_RAW && sock->type != SOCK_DGRAM)
		return -ESOCKTNOSUPPORT;

443
	if (protocol < 0 || protocol >= MAX_LINKS)
L
Linus Torvalds 已提交
444 445
		return -EPROTONOSUPPORT;

446
	netlink_lock_table();
447
#ifdef CONFIG_MODULES
448
	if (!nl_table[protocol].registered) {
449
		netlink_unlock_table();
450
		request_module("net-pf-%d-proto-%d", PF_NETLINK, protocol);
451
		netlink_lock_table();
452
	}
453 454 455 456
#endif
	if (nl_table[protocol].registered &&
	    try_module_get(nl_table[protocol].module))
		module = nl_table[protocol].module;
457 458
	else
		err = -EPROTONOSUPPORT;
459
	cb_mutex = nl_table[protocol].cb_mutex;
460
	netlink_unlock_table();
461

462 463 464
	if (err < 0)
		goto out;

465 466
	err = __netlink_create(net, sock, cb_mutex, protocol);
	if (err < 0)
467 468
		goto out_module;

469
	local_bh_disable();
470
	sock_prot_inuse_add(net, &netlink_proto, 1);
471 472
	local_bh_enable();

473 474
	nlk = nlk_sk(sock->sk);
	nlk->module = module;
475 476
out:
	return err;
L
Linus Torvalds 已提交
477

478 479 480
out_module:
	module_put(module);
	goto out;
L
Linus Torvalds 已提交
481 482 483 484 485 486 487 488 489 490 491
}

static int netlink_release(struct socket *sock)
{
	struct sock *sk = sock->sk;
	struct netlink_sock *nlk;

	if (!sk)
		return 0;

	netlink_remove(sk);
492
	sock_orphan(sk);
L
Linus Torvalds 已提交
493 494
	nlk = nlk_sk(sk);

495 496 497 498
	/*
	 * OK. Socket is unlinked, any packets that arrive now
	 * will be purged.
	 */
L
Linus Torvalds 已提交
499 500 501 502 503 504

	sock->sk = NULL;
	wake_up_interruptible_all(&nlk->wait);

	skb_queue_purge(&sk->sk_write_queue);

505
	if (nlk->pid) {
L
Linus Torvalds 已提交
506
		struct netlink_notify n = {
507
						.net = sock_net(sk),
L
Linus Torvalds 已提交
508 509 510
						.protocol = sk->sk_protocol,
						.pid = nlk->pid,
					  };
511 512
		atomic_notifier_call_chain(&netlink_chain,
				NETLINK_URELEASE, &n);
513
	}
514

515
	module_put(nlk->module);
516

517
	netlink_table_grab();
518
	if (netlink_is_kernel(sk)) {
519 520 521 522 523 524
		BUG_ON(nl_table[sk->sk_protocol].registered == 0);
		if (--nl_table[sk->sk_protocol].registered == 0) {
			kfree(nl_table[sk->sk_protocol].listeners);
			nl_table[sk->sk_protocol].module = NULL;
			nl_table[sk->sk_protocol].registered = 0;
		}
525 526 527
	} else if (nlk->subscriptions)
		netlink_update_listeners(sk);
	netlink_table_ungrab();
528

529 530 531
	kfree(nlk->groups);
	nlk->groups = NULL;

532
	local_bh_disable();
533
	sock_prot_inuse_add(sock_net(sk), &netlink_proto, -1);
534
	local_bh_enable();
L
Linus Torvalds 已提交
535 536 537 538 539 540 541
	sock_put(sk);
	return 0;
}

static int netlink_autobind(struct socket *sock)
{
	struct sock *sk = sock->sk;
542
	struct net *net = sock_net(sk);
L
Linus Torvalds 已提交
543 544 545 546
	struct nl_pid_hash *hash = &nl_table[sk->sk_protocol].hash;
	struct hlist_head *head;
	struct sock *osk;
	struct hlist_node *node;
547
	s32 pid = task_tgid_vnr(current);
L
Linus Torvalds 已提交
548 549 550 551 552 553 554 555
	int err;
	static s32 rover = -4097;

retry:
	cond_resched();
	netlink_table_grab();
	head = nl_pid_hashfn(hash, pid);
	sk_for_each(osk, node, head) {
556
		if (!net_eq(sock_net(osk), net))
557
			continue;
L
Linus Torvalds 已提交
558 559 560 561 562 563 564 565 566 567 568
		if (nlk_sk(osk)->pid == pid) {
			/* Bind collision, search negative pid values. */
			pid = rover--;
			if (rover > -4097)
				rover = -4097;
			netlink_table_ungrab();
			goto retry;
		}
	}
	netlink_table_ungrab();

569
	err = netlink_insert(sk, net, pid);
L
Linus Torvalds 已提交
570 571
	if (err == -EADDRINUSE)
		goto retry;
572 573 574 575 576 577

	/* If 2 threads race to autobind, that is fine.  */
	if (err == -EBUSY)
		err = 0;

	return err;
L
Linus Torvalds 已提交
578 579
}

580
static inline int netlink_capable(const struct socket *sock, unsigned int flag)
581
{
L
Linus Torvalds 已提交
582 583
	return (nl_table[sock->sk->sk_protocol].nl_nonroot & flag) ||
	       capable(CAP_NET_ADMIN);
584
}
L
Linus Torvalds 已提交
585

586 587 588 589 590 591 592 593 594 595 596 597
static void
netlink_update_subscriptions(struct sock *sk, unsigned int subscriptions)
{
	struct netlink_sock *nlk = nlk_sk(sk);

	if (nlk->subscriptions && !subscriptions)
		__sk_del_bind_node(sk);
	else if (!nlk->subscriptions && subscriptions)
		sk_add_bind_node(sk, &nl_table[sk->sk_protocol].mc_list);
	nlk->subscriptions = subscriptions;
}

598
static int netlink_realloc_groups(struct sock *sk)
599 600 601
{
	struct netlink_sock *nlk = nlk_sk(sk);
	unsigned int groups;
602
	unsigned long *new_groups;
603 604
	int err = 0;

605 606
	netlink_table_grab();

607
	groups = nl_table[sk->sk_protocol].groups;
608
	if (!nl_table[sk->sk_protocol].registered) {
609
		err = -ENOENT;
610 611
		goto out_unlock;
	}
612

613 614
	if (nlk->ngroups >= groups)
		goto out_unlock;
615

616 617 618 619 620
	new_groups = krealloc(nlk->groups, NLGRPSZ(groups), GFP_ATOMIC);
	if (new_groups == NULL) {
		err = -ENOMEM;
		goto out_unlock;
	}
621
	memset((char *)new_groups + NLGRPSZ(nlk->ngroups), 0,
622 623 624
	       NLGRPSZ(groups) - NLGRPSZ(nlk->ngroups));

	nlk->groups = new_groups;
625
	nlk->ngroups = groups;
626 627 628
 out_unlock:
	netlink_table_ungrab();
	return err;
629 630
}

631 632
static int netlink_bind(struct socket *sock, struct sockaddr *addr,
			int addr_len)
L
Linus Torvalds 已提交
633 634
{
	struct sock *sk = sock->sk;
635
	struct net *net = sock_net(sk);
L
Linus Torvalds 已提交
636 637 638
	struct netlink_sock *nlk = nlk_sk(sk);
	struct sockaddr_nl *nladdr = (struct sockaddr_nl *)addr;
	int err;
639

L
Linus Torvalds 已提交
640 641 642 643
	if (nladdr->nl_family != AF_NETLINK)
		return -EINVAL;

	/* Only superuser is allowed to listen multicasts */
644 645 646
	if (nladdr->nl_groups) {
		if (!netlink_capable(sock, NL_NONROOT_RECV))
			return -EPERM;
647 648 649
		err = netlink_realloc_groups(sk);
		if (err)
			return err;
650
	}
L
Linus Torvalds 已提交
651 652 653 654 655 656

	if (nlk->pid) {
		if (nladdr->nl_pid != nlk->pid)
			return -EINVAL;
	} else {
		err = nladdr->nl_pid ?
657
			netlink_insert(sk, net, nladdr->nl_pid) :
L
Linus Torvalds 已提交
658 659 660 661 662
			netlink_autobind(sock);
		if (err)
			return err;
	}

663
	if (!nladdr->nl_groups && (nlk->groups == NULL || !(u32)nlk->groups[0]))
L
Linus Torvalds 已提交
664 665 666
		return 0;

	netlink_table_grab();
667
	netlink_update_subscriptions(sk, nlk->subscriptions +
668 669 670
					 hweight32(nladdr->nl_groups) -
					 hweight32(nlk->groups[0]));
	nlk->groups[0] = (nlk->groups[0] & ~0xffffffffUL) | nladdr->nl_groups;
671
	netlink_update_listeners(sk);
L
Linus Torvalds 已提交
672 673 674 675 676 677 678 679 680 681 682
	netlink_table_ungrab();

	return 0;
}

static int netlink_connect(struct socket *sock, struct sockaddr *addr,
			   int alen, int flags)
{
	int err = 0;
	struct sock *sk = sock->sk;
	struct netlink_sock *nlk = nlk_sk(sk);
683
	struct sockaddr_nl *nladdr = (struct sockaddr_nl *)addr;
L
Linus Torvalds 已提交
684

685 686 687
	if (alen < sizeof(addr->sa_family))
		return -EINVAL;

L
Linus Torvalds 已提交
688 689 690
	if (addr->sa_family == AF_UNSPEC) {
		sk->sk_state	= NETLINK_UNCONNECTED;
		nlk->dst_pid	= 0;
691
		nlk->dst_group  = 0;
L
Linus Torvalds 已提交
692 693 694 695 696 697 698 699 700 701 702 703 704 705 706
		return 0;
	}
	if (addr->sa_family != AF_NETLINK)
		return -EINVAL;

	/* Only superuser is allowed to send multicasts */
	if (nladdr->nl_groups && !netlink_capable(sock, NL_NONROOT_SEND))
		return -EPERM;

	if (!nlk->pid)
		err = netlink_autobind(sock);

	if (err == 0) {
		sk->sk_state	= NETLINK_CONNECTED;
		nlk->dst_pid 	= nladdr->nl_pid;
707
		nlk->dst_group  = ffs(nladdr->nl_groups);
L
Linus Torvalds 已提交
708 709 710 711 712
	}

	return err;
}

713 714
static int netlink_getname(struct socket *sock, struct sockaddr *addr,
			   int *addr_len, int peer)
L
Linus Torvalds 已提交
715 716 717
{
	struct sock *sk = sock->sk;
	struct netlink_sock *nlk = nlk_sk(sk);
718
	DECLARE_SOCKADDR(struct sockaddr_nl *, nladdr, addr);
719

L
Linus Torvalds 已提交
720 721 722 723 724 725
	nladdr->nl_family = AF_NETLINK;
	nladdr->nl_pad = 0;
	*addr_len = sizeof(*nladdr);

	if (peer) {
		nladdr->nl_pid = nlk->dst_pid;
726
		nladdr->nl_groups = netlink_group_mask(nlk->dst_group);
L
Linus Torvalds 已提交
727 728
	} else {
		nladdr->nl_pid = nlk->pid;
729
		nladdr->nl_groups = nlk->groups ? nlk->groups[0] : 0;
L
Linus Torvalds 已提交
730 731 732 733 734 735
	}
	return 0;
}

static void netlink_overrun(struct sock *sk)
{
736 737 738 739 740 741 742
	struct netlink_sock *nlk = nlk_sk(sk);

	if (!(nlk->flags & NETLINK_RECV_NO_ENOBUFS)) {
		if (!test_and_set_bit(0, &nlk_sk(sk)->state)) {
			sk->sk_err = ENOBUFS;
			sk->sk_error_report(sk);
		}
L
Linus Torvalds 已提交
743
	}
744
	atomic_inc(&sk->sk_drops);
L
Linus Torvalds 已提交
745 746 747 748 749 750 751
}

static struct sock *netlink_getsockbypid(struct sock *ssk, u32 pid)
{
	struct sock *sock;
	struct netlink_sock *nlk;

752
	sock = netlink_lookup(sock_net(ssk), ssk->sk_protocol, pid);
L
Linus Torvalds 已提交
753 754 755 756 757
	if (!sock)
		return ERR_PTR(-ECONNREFUSED);

	/* Don't bother queuing skb if kernel socket has no input function */
	nlk = nlk_sk(sock);
758 759
	if (sock->sk_state == NETLINK_CONNECTED &&
	    nlk->dst_pid != nlk_sk(ssk)->pid) {
L
Linus Torvalds 已提交
760 761 762 763 764 765 766 767
		sock_put(sock);
		return ERR_PTR(-ECONNREFUSED);
	}
	return sock;
}

struct sock *netlink_getsockbyfilp(struct file *filp)
{
J
Josef Sipek 已提交
768
	struct inode *inode = filp->f_path.dentry->d_inode;
L
Linus Torvalds 已提交
769 770 771 772 773 774 775 776 777 778 779 780 781 782 783 784 785 786 787 788 789 790 791
	struct sock *sock;

	if (!S_ISSOCK(inode->i_mode))
		return ERR_PTR(-ENOTSOCK);

	sock = SOCKET_I(inode)->sk;
	if (sock->sk_family != AF_NETLINK)
		return ERR_PTR(-EINVAL);

	sock_hold(sock);
	return sock;
}

/*
 * Attach a skb to a netlink socket.
 * The caller must hold a reference to the destination socket. On error, the
 * reference is dropped. The skb is not send to the destination, just all
 * all error checks are performed and memory in the queue is reserved.
 * Return values:
 * < 0: error. skb freed, reference to sock dropped.
 * 0: continue
 * 1: repeat lookup - reference dropped while waiting for socket memory.
 */
792
int netlink_attachskb(struct sock *sk, struct sk_buff *skb,
P
Patrick McHardy 已提交
793
		      long *timeo, struct sock *ssk)
L
Linus Torvalds 已提交
794 795 796 797 798 799 800 801
{
	struct netlink_sock *nlk;

	nlk = nlk_sk(sk);

	if (atomic_read(&sk->sk_rmem_alloc) > sk->sk_rcvbuf ||
	    test_bit(0, &nlk->state)) {
		DECLARE_WAITQUEUE(wait, current);
P
Patrick McHardy 已提交
802
		if (!*timeo) {
803
			if (!ssk || netlink_is_kernel(ssk))
L
Linus Torvalds 已提交
804 805 806 807 808 809 810 811 812 813 814 815
				netlink_overrun(sk);
			sock_put(sk);
			kfree_skb(skb);
			return -EAGAIN;
		}

		__set_current_state(TASK_INTERRUPTIBLE);
		add_wait_queue(&nlk->wait, &wait);

		if ((atomic_read(&sk->sk_rmem_alloc) > sk->sk_rcvbuf ||
		     test_bit(0, &nlk->state)) &&
		    !sock_flag(sk, SOCK_DEAD))
P
Patrick McHardy 已提交
816
			*timeo = schedule_timeout(*timeo);
L
Linus Torvalds 已提交
817 818 819 820 821 822 823

		__set_current_state(TASK_RUNNING);
		remove_wait_queue(&nlk->wait, &wait);
		sock_put(sk);

		if (signal_pending(current)) {
			kfree_skb(skb);
P
Patrick McHardy 已提交
824
			return sock_intr_errno(*timeo);
L
Linus Torvalds 已提交
825 826 827 828 829 830 831
		}
		return 1;
	}
	skb_set_owner_r(skb, sk);
	return 0;
}

832
int netlink_sendskb(struct sock *sk, struct sk_buff *skb)
L
Linus Torvalds 已提交
833 834 835 836 837 838 839 840 841 842 843 844 845 846 847
{
	int len = skb->len;

	skb_queue_tail(&sk->sk_receive_queue, skb);
	sk->sk_data_ready(sk, len);
	sock_put(sk);
	return len;
}

void netlink_detachskb(struct sock *sk, struct sk_buff *skb)
{
	kfree_skb(skb);
	sock_put(sk);
}

848
static struct sk_buff *netlink_trim(struct sk_buff *skb, gfp_t allocation)
L
Linus Torvalds 已提交
849 850 851 852 853
{
	int delta;

	skb_orphan(skb);

854
	delta = skb->end - skb->tail;
L
Linus Torvalds 已提交
855 856 857 858 859 860 861 862 863 864 865 866 867 868 869 870 871
	if (delta * 2 < skb->truesize)
		return skb;

	if (skb_shared(skb)) {
		struct sk_buff *nskb = skb_clone(skb, allocation);
		if (!nskb)
			return skb;
		kfree_skb(skb);
		skb = nskb;
	}

	if (!pskb_expand_head(skb, 0, -delta, allocation))
		skb->truesize -= delta;

	return skb;
}

872
static void netlink_rcv_wake(struct sock *sk)
873 874 875 876 877 878 879 880 881
{
	struct netlink_sock *nlk = nlk_sk(sk);

	if (skb_queue_empty(&sk->sk_receive_queue))
		clear_bit(0, &nlk->state);
	if (!test_bit(0, &nlk->state))
		wake_up_interruptible(&nlk->wait);
}

882
static int netlink_unicast_kernel(struct sock *sk, struct sk_buff *skb)
883 884 885 886 887 888 889 890 891 892 893 894 895 896 897 898 899
{
	int ret;
	struct netlink_sock *nlk = nlk_sk(sk);

	ret = -ECONNREFUSED;
	if (nlk->netlink_rcv != NULL) {
		ret = skb->len;
		skb_set_owner_r(skb, sk);
		nlk->netlink_rcv(skb);
	}
	kfree_skb(skb);
	sock_put(sk);
	return ret;
}

int netlink_unicast(struct sock *ssk, struct sk_buff *skb,
		    u32 pid, int nonblock)
L
Linus Torvalds 已提交
900 901 902 903 904 905 906 907 908 909 910 911 912 913
{
	struct sock *sk;
	int err;
	long timeo;

	skb = netlink_trim(skb, gfp_any());

	timeo = sock_sndtimeo(ssk, nonblock);
retry:
	sk = netlink_getsockbypid(ssk, pid);
	if (IS_ERR(sk)) {
		kfree_skb(skb);
		return PTR_ERR(sk);
	}
914 915 916
	if (netlink_is_kernel(sk))
		return netlink_unicast_kernel(sk, skb);

917
	if (sk_filter(sk, skb)) {
W
Wang Chen 已提交
918
		err = skb->len;
919 920 921 922 923
		kfree_skb(skb);
		sock_put(sk);
		return err;
	}

924
	err = netlink_attachskb(sk, skb, &timeo, ssk);
L
Linus Torvalds 已提交
925 926 927 928 929
	if (err == 1)
		goto retry;
	if (err)
		return err;

930
	return netlink_sendskb(sk, skb);
L
Linus Torvalds 已提交
931
}
932
EXPORT_SYMBOL(netlink_unicast);
L
Linus Torvalds 已提交
933

934 935 936
int netlink_has_listeners(struct sock *sk, unsigned int group)
{
	int res = 0;
937
	struct listeners *listeners;
938

939
	BUG_ON(!netlink_is_kernel(sk));
940 941 942 943

	rcu_read_lock();
	listeners = rcu_dereference(nl_table[sk->sk_protocol].listeners);

944
	if (group - 1 < nl_table[sk->sk_protocol].groups)
945
		res = test_bit(group - 1, listeners->masks);
946 947 948

	rcu_read_unlock();

949 950 951 952
	return res;
}
EXPORT_SYMBOL_GPL(netlink_has_listeners);

953
static int netlink_broadcast_deliver(struct sock *sk, struct sk_buff *skb)
L
Linus Torvalds 已提交
954 955 956 957 958 959 960 961
{
	struct netlink_sock *nlk = nlk_sk(sk);

	if (atomic_read(&sk->sk_rmem_alloc) <= sk->sk_rcvbuf &&
	    !test_bit(0, &nlk->state)) {
		skb_set_owner_r(skb, sk);
		skb_queue_tail(&sk->sk_receive_queue, skb);
		sk->sk_data_ready(sk, skb->len);
962
		return atomic_read(&sk->sk_rmem_alloc) > (sk->sk_rcvbuf >> 1);
L
Linus Torvalds 已提交
963 964 965 966 967 968
	}
	return -1;
}

struct netlink_broadcast_data {
	struct sock *exclude_sk;
969
	struct net *net;
L
Linus Torvalds 已提交
970 971 972
	u32 pid;
	u32 group;
	int failure;
973
	int delivery_failure;
L
Linus Torvalds 已提交
974 975
	int congested;
	int delivered;
A
Al Viro 已提交
976
	gfp_t allocation;
L
Linus Torvalds 已提交
977
	struct sk_buff *skb, *skb2;
978 979
	int (*tx_filter)(struct sock *dsk, struct sk_buff *skb, void *data);
	void *tx_data;
L
Linus Torvalds 已提交
980 981
};

982
static int do_one_broadcast(struct sock *sk,
L
Linus Torvalds 已提交
983 984 985 986 987 988 989 990
				   struct netlink_broadcast_data *p)
{
	struct netlink_sock *nlk = nlk_sk(sk);
	int val;

	if (p->exclude_sk == sk)
		goto out;

991 992
	if (nlk->pid == p->pid || p->group - 1 >= nlk->ngroups ||
	    !test_bit(p->group - 1, nlk->groups))
L
Linus Torvalds 已提交
993 994
		goto out;

995
	if (!net_eq(sock_net(sk), p->net))
996 997
		goto out;

L
Linus Torvalds 已提交
998 999 1000 1001 1002 1003 1004
	if (p->failure) {
		netlink_overrun(sk);
		goto out;
	}

	sock_hold(sk);
	if (p->skb2 == NULL) {
1005
		if (skb_shared(p->skb)) {
L
Linus Torvalds 已提交
1006 1007
			p->skb2 = skb_clone(p->skb, p->allocation);
		} else {
1008 1009 1010 1011 1012 1013
			p->skb2 = skb_get(p->skb);
			/*
			 * skb ownership may have been set when
			 * delivered to a previous socket.
			 */
			skb_orphan(p->skb2);
L
Linus Torvalds 已提交
1014 1015 1016 1017 1018 1019
		}
	}
	if (p->skb2 == NULL) {
		netlink_overrun(sk);
		/* Clone failed. Notify ALL listeners. */
		p->failure = 1;
1020 1021
		if (nlk->flags & NETLINK_BROADCAST_SEND_ERROR)
			p->delivery_failure = 1;
1022 1023 1024
	} else if (p->tx_filter && p->tx_filter(sk, p->skb2, p->tx_data)) {
		kfree_skb(p->skb2);
		p->skb2 = NULL;
1025 1026 1027
	} else if (sk_filter(sk, p->skb2)) {
		kfree_skb(p->skb2);
		p->skb2 = NULL;
L
Linus Torvalds 已提交
1028 1029
	} else if ((val = netlink_broadcast_deliver(sk, p->skb2)) < 0) {
		netlink_overrun(sk);
1030 1031
		if (nlk->flags & NETLINK_BROADCAST_SEND_ERROR)
			p->delivery_failure = 1;
L
Linus Torvalds 已提交
1032 1033 1034 1035 1036 1037 1038 1039 1040 1041 1042
	} else {
		p->congested |= val;
		p->delivered = 1;
		p->skb2 = NULL;
	}
	sock_put(sk);

out:
	return 0;
}

1043 1044 1045 1046
int netlink_broadcast_filtered(struct sock *ssk, struct sk_buff *skb, u32 pid,
	u32 group, gfp_t allocation,
	int (*filter)(struct sock *dsk, struct sk_buff *skb, void *data),
	void *filter_data)
L
Linus Torvalds 已提交
1047
{
1048
	struct net *net = sock_net(ssk);
L
Linus Torvalds 已提交
1049 1050 1051 1052 1053 1054 1055
	struct netlink_broadcast_data info;
	struct hlist_node *node;
	struct sock *sk;

	skb = netlink_trim(skb, allocation);

	info.exclude_sk = ssk;
1056
	info.net = net;
L
Linus Torvalds 已提交
1057 1058 1059
	info.pid = pid;
	info.group = group;
	info.failure = 0;
1060
	info.delivery_failure = 0;
L
Linus Torvalds 已提交
1061 1062 1063 1064 1065
	info.congested = 0;
	info.delivered = 0;
	info.allocation = allocation;
	info.skb = skb;
	info.skb2 = NULL;
1066 1067
	info.tx_filter = filter;
	info.tx_data = filter_data;
L
Linus Torvalds 已提交
1068 1069 1070 1071 1072 1073 1074 1075

	/* While we sleep in clone, do not allow to change socket list */

	netlink_lock_table();

	sk_for_each_bound(sk, node, &nl_table[ssk->sk_protocol].mc_list)
		do_one_broadcast(sk, &info);

1076
	consume_skb(skb);
1077

L
Linus Torvalds 已提交
1078 1079
	netlink_unlock_table();

1080 1081
	if (info.delivery_failure) {
		kfree_skb(info.skb2);
1082
		return -ENOBUFS;
1083 1084
	} else
		consume_skb(info.skb2);
1085

L
Linus Torvalds 已提交
1086 1087 1088 1089 1090 1091 1092
	if (info.delivered) {
		if (info.congested && (allocation & __GFP_WAIT))
			yield();
		return 0;
	}
	return -ESRCH;
}
1093 1094 1095 1096 1097 1098 1099 1100
EXPORT_SYMBOL(netlink_broadcast_filtered);

int netlink_broadcast(struct sock *ssk, struct sk_buff *skb, u32 pid,
		      u32 group, gfp_t allocation)
{
	return netlink_broadcast_filtered(ssk, skb, pid, group, allocation,
		NULL, NULL);
}
1101
EXPORT_SYMBOL(netlink_broadcast);
L
Linus Torvalds 已提交
1102 1103 1104 1105 1106 1107 1108 1109

struct netlink_set_err_data {
	struct sock *exclude_sk;
	u32 pid;
	u32 group;
	int code;
};

1110
static int do_one_set_err(struct sock *sk, struct netlink_set_err_data *p)
L
Linus Torvalds 已提交
1111 1112
{
	struct netlink_sock *nlk = nlk_sk(sk);
1113
	int ret = 0;
L
Linus Torvalds 已提交
1114 1115 1116 1117

	if (sk == p->exclude_sk)
		goto out;

O
Octavian Purdila 已提交
1118
	if (!net_eq(sock_net(sk), sock_net(p->exclude_sk)))
1119 1120
		goto out;

1121 1122
	if (nlk->pid == p->pid || p->group - 1 >= nlk->ngroups ||
	    !test_bit(p->group - 1, nlk->groups))
L
Linus Torvalds 已提交
1123 1124
		goto out;

1125 1126 1127 1128 1129
	if (p->code == ENOBUFS && nlk->flags & NETLINK_RECV_NO_ENOBUFS) {
		ret = 1;
		goto out;
	}

L
Linus Torvalds 已提交
1130 1131 1132
	sk->sk_err = p->code;
	sk->sk_error_report(sk);
out:
1133
	return ret;
L
Linus Torvalds 已提交
1134 1135
}

1136 1137 1138 1139 1140 1141
/**
 * netlink_set_err - report error to broadcast listeners
 * @ssk: the kernel netlink socket, as returned by netlink_kernel_create()
 * @pid: the PID of a process that we want to skip (if any)
 * @groups: the broadcast group that will notice the error
 * @code: error code, must be negative (as usual in kernelspace)
1142 1143 1144
 *
 * This function returns the number of broadcast listeners that have set the
 * NETLINK_RECV_NO_ENOBUFS socket option.
1145
 */
1146
int netlink_set_err(struct sock *ssk, u32 pid, u32 group, int code)
L
Linus Torvalds 已提交
1147 1148 1149 1150
{
	struct netlink_set_err_data info;
	struct hlist_node *node;
	struct sock *sk;
1151
	int ret = 0;
L
Linus Torvalds 已提交
1152 1153 1154 1155

	info.exclude_sk = ssk;
	info.pid = pid;
	info.group = group;
1156 1157
	/* sk->sk_err wants a positive error value */
	info.code = -code;
L
Linus Torvalds 已提交
1158 1159 1160 1161

	read_lock(&nl_table_lock);

	sk_for_each_bound(sk, node, &nl_table[ssk->sk_protocol].mc_list)
1162
		ret += do_one_set_err(sk, &info);
L
Linus Torvalds 已提交
1163 1164

	read_unlock(&nl_table_lock);
1165
	return ret;
L
Linus Torvalds 已提交
1166
}
1167
EXPORT_SYMBOL(netlink_set_err);
L
Linus Torvalds 已提交
1168

1169 1170 1171 1172 1173 1174 1175 1176 1177 1178 1179 1180 1181 1182 1183 1184 1185
/* must be called with netlink table grabbed */
static void netlink_update_socket_mc(struct netlink_sock *nlk,
				     unsigned int group,
				     int is_new)
{
	int old, new = !!is_new, subscriptions;

	old = test_bit(group - 1, nlk->groups);
	subscriptions = nlk->subscriptions - old + new;
	if (new)
		__set_bit(group - 1, nlk->groups);
	else
		__clear_bit(group - 1, nlk->groups);
	netlink_update_subscriptions(&nlk->sk, subscriptions);
	netlink_update_listeners(&nlk->sk);
}

1186
static int netlink_setsockopt(struct socket *sock, int level, int optname,
1187
			      char __user *optval, unsigned int optlen)
1188 1189 1190
{
	struct sock *sk = sock->sk;
	struct netlink_sock *nlk = nlk_sk(sk);
1191 1192
	unsigned int val = 0;
	int err;
1193 1194 1195 1196 1197

	if (level != SOL_NETLINK)
		return -ENOPROTOOPT;

	if (optlen >= sizeof(int) &&
1198
	    get_user(val, (unsigned int __user *)optval))
1199 1200 1201 1202 1203 1204 1205 1206 1207 1208 1209 1210 1211 1212
		return -EFAULT;

	switch (optname) {
	case NETLINK_PKTINFO:
		if (val)
			nlk->flags |= NETLINK_RECV_PKTINFO;
		else
			nlk->flags &= ~NETLINK_RECV_PKTINFO;
		err = 0;
		break;
	case NETLINK_ADD_MEMBERSHIP:
	case NETLINK_DROP_MEMBERSHIP: {
		if (!netlink_capable(sock, NL_NONROOT_RECV))
			return -EPERM;
1213 1214 1215
		err = netlink_realloc_groups(sk);
		if (err)
			return err;
1216 1217 1218
		if (!val || val - 1 >= nlk->ngroups)
			return -EINVAL;
		netlink_table_grab();
1219 1220
		netlink_update_socket_mc(nlk, val,
					 optname == NETLINK_ADD_MEMBERSHIP);
1221 1222 1223 1224
		netlink_table_ungrab();
		err = 0;
		break;
	}
1225 1226 1227 1228 1229 1230 1231
	case NETLINK_BROADCAST_ERROR:
		if (val)
			nlk->flags |= NETLINK_BROADCAST_SEND_ERROR;
		else
			nlk->flags &= ~NETLINK_BROADCAST_SEND_ERROR;
		err = 0;
		break;
1232 1233 1234 1235 1236 1237 1238 1239 1240
	case NETLINK_NO_ENOBUFS:
		if (val) {
			nlk->flags |= NETLINK_RECV_NO_ENOBUFS;
			clear_bit(0, &nlk->state);
			wake_up_interruptible(&nlk->wait);
		} else
			nlk->flags &= ~NETLINK_RECV_NO_ENOBUFS;
		err = 0;
		break;
1241 1242 1243 1244 1245 1246 1247
	default:
		err = -ENOPROTOOPT;
	}
	return err;
}

static int netlink_getsockopt(struct socket *sock, int level, int optname,
1248
			      char __user *optval, int __user *optlen)
1249 1250 1251 1252 1253 1254 1255 1256 1257 1258 1259 1260 1261 1262 1263 1264 1265 1266 1267
{
	struct sock *sk = sock->sk;
	struct netlink_sock *nlk = nlk_sk(sk);
	int len, val, err;

	if (level != SOL_NETLINK)
		return -ENOPROTOOPT;

	if (get_user(len, optlen))
		return -EFAULT;
	if (len < 0)
		return -EINVAL;

	switch (optname) {
	case NETLINK_PKTINFO:
		if (len < sizeof(int))
			return -EINVAL;
		len = sizeof(int);
		val = nlk->flags & NETLINK_RECV_PKTINFO ? 1 : 0;
H
Heiko Carstens 已提交
1268 1269 1270
		if (put_user(len, optlen) ||
		    put_user(val, optval))
			return -EFAULT;
1271 1272
		err = 0;
		break;
1273 1274 1275 1276 1277 1278 1279 1280 1281 1282
	case NETLINK_BROADCAST_ERROR:
		if (len < sizeof(int))
			return -EINVAL;
		len = sizeof(int);
		val = nlk->flags & NETLINK_BROADCAST_SEND_ERROR ? 1 : 0;
		if (put_user(len, optlen) ||
		    put_user(val, optval))
			return -EFAULT;
		err = 0;
		break;
1283 1284 1285 1286 1287 1288 1289 1290 1291 1292
	case NETLINK_NO_ENOBUFS:
		if (len < sizeof(int))
			return -EINVAL;
		len = sizeof(int);
		val = nlk->flags & NETLINK_RECV_NO_ENOBUFS ? 1 : 0;
		if (put_user(len, optlen) ||
		    put_user(val, optval))
			return -EFAULT;
		err = 0;
		break;
1293 1294 1295 1296 1297 1298 1299 1300 1301 1302 1303 1304 1305 1306
	default:
		err = -ENOPROTOOPT;
	}
	return err;
}

static void netlink_cmsg_recv_pktinfo(struct msghdr *msg, struct sk_buff *skb)
{
	struct nl_pktinfo info;

	info.group = NETLINK_CB(skb).dst_group;
	put_cmsg(msg, SOL_NETLINK, NETLINK_PKTINFO, sizeof(info), &info);
}

L
Linus Torvalds 已提交
1307 1308 1309 1310 1311 1312
static int netlink_sendmsg(struct kiocb *kiocb, struct socket *sock,
			   struct msghdr *msg, size_t len)
{
	struct sock_iocb *siocb = kiocb_to_siocb(kiocb);
	struct sock *sk = sock->sk;
	struct netlink_sock *nlk = nlk_sk(sk);
1313
	struct sockaddr_nl *addr = msg->msg_name;
L
Linus Torvalds 已提交
1314
	u32 dst_pid;
1315
	u32 dst_group;
L
Linus Torvalds 已提交
1316 1317 1318 1319 1320 1321 1322
	struct sk_buff *skb;
	int err;
	struct scm_cookie scm;

	if (msg->msg_flags&MSG_OOB)
		return -EOPNOTSUPP;

1323
	if (NULL == siocb->scm)
L
Linus Torvalds 已提交
1324
		siocb->scm = &scm;
1325

L
Linus Torvalds 已提交
1326 1327 1328 1329 1330
	err = scm_send(sock, msg, siocb->scm);
	if (err < 0)
		return err;

	if (msg->msg_namelen) {
1331
		err = -EINVAL;
L
Linus Torvalds 已提交
1332
		if (addr->nl_family != AF_NETLINK)
1333
			goto out;
L
Linus Torvalds 已提交
1334
		dst_pid = addr->nl_pid;
1335
		dst_group = ffs(addr->nl_groups);
1336
		err =  -EPERM;
1337
		if (dst_group && !netlink_capable(sock, NL_NONROOT_SEND))
1338
			goto out;
L
Linus Torvalds 已提交
1339 1340
	} else {
		dst_pid = nlk->dst_pid;
1341
		dst_group = nlk->dst_group;
L
Linus Torvalds 已提交
1342 1343 1344 1345 1346 1347 1348 1349 1350 1351 1352 1353
	}

	if (!nlk->pid) {
		err = netlink_autobind(sock);
		if (err)
			goto out;
	}

	err = -EMSGSIZE;
	if (len > sk->sk_sndbuf - 32)
		goto out;
	err = -ENOBUFS;
1354
	skb = alloc_skb(len, GFP_KERNEL);
1355
	if (skb == NULL)
L
Linus Torvalds 已提交
1356 1357 1358
		goto out;

	NETLINK_CB(skb).pid	= nlk->pid;
1359
	NETLINK_CB(skb).dst_group = dst_group;
L
Linus Torvalds 已提交
1360 1361 1362
	memcpy(NETLINK_CREDS(skb), &siocb->scm->creds, sizeof(struct ucred));

	err = -EFAULT;
1363
	if (memcpy_fromiovec(skb_put(skb, len), msg->msg_iov, len)) {
L
Linus Torvalds 已提交
1364 1365 1366 1367 1368 1369 1370 1371 1372 1373
		kfree_skb(skb);
		goto out;
	}

	err = security_netlink_send(sk, skb);
	if (err) {
		kfree_skb(skb);
		goto out;
	}

1374
	if (dst_group) {
L
Linus Torvalds 已提交
1375
		atomic_inc(&skb->users);
1376
		netlink_broadcast(sk, skb, dst_pid, dst_group, GFP_KERNEL);
L
Linus Torvalds 已提交
1377 1378 1379 1380
	}
	err = netlink_unicast(sk, skb, dst_pid, msg->msg_flags&MSG_DONTWAIT);

out:
1381
	scm_destroy(siocb->scm);
L
Linus Torvalds 已提交
1382 1383 1384 1385 1386 1387 1388 1389 1390 1391 1392 1393 1394
	return err;
}

static int netlink_recvmsg(struct kiocb *kiocb, struct socket *sock,
			   struct msghdr *msg, size_t len,
			   int flags)
{
	struct sock_iocb *siocb = kiocb_to_siocb(kiocb);
	struct scm_cookie scm;
	struct sock *sk = sock->sk;
	struct netlink_sock *nlk = nlk_sk(sk);
	int noblock = flags&MSG_DONTWAIT;
	size_t copied;
J
Johannes Berg 已提交
1395
	struct sk_buff *skb, *data_skb;
1396
	int err, ret;
L
Linus Torvalds 已提交
1397 1398 1399 1400 1401 1402

	if (flags&MSG_OOB)
		return -EOPNOTSUPP;

	copied = 0;

1403 1404
	skb = skb_recv_datagram(sk, flags, noblock, &err);
	if (skb == NULL)
L
Linus Torvalds 已提交
1405 1406
		goto out;

J
Johannes Berg 已提交
1407 1408
	data_skb = skb;

1409 1410 1411
#ifdef CONFIG_COMPAT_NETLINK_MESSAGES
	if (unlikely(skb_shinfo(skb)->frag_list)) {
		/*
J
Johannes Berg 已提交
1412 1413 1414
		 * If this skb has a frag_list, then here that means that we
		 * will have to use the frag_list skb's data for compat tasks
		 * and the regular skb's data for normal (non-compat) tasks.
1415
		 *
J
Johannes Berg 已提交
1416 1417 1418 1419
		 * If we need to send the compat skb, assign it to the
		 * 'data_skb' variable so that it will be used below for data
		 * copying. We keep 'skb' for everything else, including
		 * freeing both later.
1420
		 */
J
Johannes Berg 已提交
1421 1422
		if (flags & MSG_CMSG_COMPAT)
			data_skb = skb_shinfo(skb)->frag_list;
1423 1424 1425
	}
#endif

L
Linus Torvalds 已提交
1426 1427
	msg->msg_namelen = 0;

J
Johannes Berg 已提交
1428
	copied = data_skb->len;
L
Linus Torvalds 已提交
1429 1430 1431 1432 1433
	if (len < copied) {
		msg->msg_flags |= MSG_TRUNC;
		copied = len;
	}

J
Johannes Berg 已提交
1434 1435
	skb_reset_transport_header(data_skb);
	err = skb_copy_datagram_iovec(data_skb, 0, msg->msg_iov, copied);
L
Linus Torvalds 已提交
1436 1437

	if (msg->msg_name) {
1438
		struct sockaddr_nl *addr = (struct sockaddr_nl *)msg->msg_name;
L
Linus Torvalds 已提交
1439 1440 1441
		addr->nl_family = AF_NETLINK;
		addr->nl_pad    = 0;
		addr->nl_pid	= NETLINK_CB(skb).pid;
1442
		addr->nl_groups	= netlink_group_mask(NETLINK_CB(skb).dst_group);
L
Linus Torvalds 已提交
1443 1444 1445
		msg->msg_namelen = sizeof(*addr);
	}

1446 1447 1448
	if (nlk->flags & NETLINK_RECV_PKTINFO)
		netlink_cmsg_recv_pktinfo(msg, skb);

L
Linus Torvalds 已提交
1449 1450 1451 1452 1453
	if (NULL == siocb->scm) {
		memset(&scm, 0, sizeof(scm));
		siocb->scm = &scm;
	}
	siocb->scm->creds = *NETLINK_CREDS(skb);
1454
	if (flags & MSG_TRUNC)
J
Johannes Berg 已提交
1455
		copied = data_skb->len;
1456

L
Linus Torvalds 已提交
1457 1458
	skb_free_datagram(sk, skb);

1459 1460 1461 1462 1463 1464 1465
	if (nlk->cb && atomic_read(&sk->sk_rmem_alloc) <= sk->sk_rcvbuf / 2) {
		ret = netlink_dump(sk);
		if (ret) {
			sk->sk_err = ret;
			sk->sk_error_report(sk);
		}
	}
L
Linus Torvalds 已提交
1466 1467 1468 1469 1470 1471 1472 1473 1474

	scm_recv(sock, msg, siocb->scm, flags);
out:
	netlink_rcv_wake(sk);
	return err ? : copied;
}

static void netlink_data_ready(struct sock *sk, int len)
{
1475
	BUG();
L
Linus Torvalds 已提交
1476 1477 1478
}

/*
1479
 *	We export these functions to other modules. They provide a
L
Linus Torvalds 已提交
1480 1481 1482 1483 1484
 *	complete set of kernel non-blocking support for message
 *	queueing.
 */

struct sock *
1485
netlink_kernel_create(struct net *net, int unit, unsigned int groups,
1486
		      void (*input)(struct sk_buff *skb),
1487
		      struct mutex *cb_mutex, struct module *module)
L
Linus Torvalds 已提交
1488 1489 1490
{
	struct socket *sock;
	struct sock *sk;
1491
	struct netlink_sock *nlk;
1492
	struct listeners *listeners = NULL;
L
Linus Torvalds 已提交
1493

1494
	BUG_ON(!nl_table);
L
Linus Torvalds 已提交
1495

1496
	if (unit < 0 || unit >= MAX_LINKS)
L
Linus Torvalds 已提交
1497 1498 1499 1500 1501
		return NULL;

	if (sock_create_lite(PF_NETLINK, SOCK_DGRAM, unit, &sock))
		return NULL;

1502 1503 1504 1505 1506 1507 1508 1509 1510 1511
	/*
	 * We have to just have a reference on the net from sk, but don't
	 * get_net it. Besides, we cannot get and then put the net here.
	 * So we create one inside init_net and the move it to net.
	 */

	if (__netlink_create(&init_net, sock, cb_mutex, unit) < 0)
		goto out_sock_release_nosk;

	sk = sock->sk;
1512
	sk_change_net(sk, net);
1513

1514 1515 1516
	if (groups < 32)
		groups = 32;

1517
	listeners = kzalloc(sizeof(*listeners) + NLGRPSZ(groups), GFP_KERNEL);
1518 1519 1520
	if (!listeners)
		goto out_sock_release;

L
Linus Torvalds 已提交
1521 1522
	sk->sk_data_ready = netlink_data_ready;
	if (input)
1523
		nlk_sk(sk)->netlink_rcv = input;
L
Linus Torvalds 已提交
1524

1525
	if (netlink_insert(sk, net, 0))
1526
		goto out_sock_release;
1527

1528 1529
	nlk = nlk_sk(sk);
	nlk->flags |= NETLINK_KERNEL_SOCKET;
1530 1531

	netlink_table_grab();
1532 1533
	if (!nl_table[unit].registered) {
		nl_table[unit].groups = groups;
1534
		rcu_assign_pointer(nl_table[unit].listeners, listeners);
1535 1536 1537
		nl_table[unit].cb_mutex = cb_mutex;
		nl_table[unit].module = module;
		nl_table[unit].registered = 1;
1538 1539
	} else {
		kfree(listeners);
1540
		nl_table[unit].registered++;
1541
	}
1542
	netlink_table_ungrab();
1543 1544
	return sk;

1545
out_sock_release:
1546
	kfree(listeners);
1547
	netlink_kernel_release(sk);
1548 1549 1550
	return NULL;

out_sock_release_nosk:
1551
	sock_release(sock);
1552
	return NULL;
L
Linus Torvalds 已提交
1553
}
1554
EXPORT_SYMBOL(netlink_kernel_create);
L
Linus Torvalds 已提交
1555

1556 1557 1558 1559

void
netlink_kernel_release(struct sock *sk)
{
1560
	sk_release_kernel(sk);
1561 1562 1563
}
EXPORT_SYMBOL(netlink_kernel_release);

1564
int __netlink_change_ngroups(struct sock *sk, unsigned int groups)
1565
{
1566
	struct listeners *new, *old;
1567 1568 1569 1570 1571 1572
	struct netlink_table *tbl = &nl_table[sk->sk_protocol];

	if (groups < 32)
		groups = 32;

	if (NLGRPSZ(tbl->groups) < NLGRPSZ(groups)) {
1573 1574
		new = kzalloc(sizeof(*new) + NLGRPSZ(groups), GFP_ATOMIC);
		if (!new)
1575
			return -ENOMEM;
1576
		old = rcu_dereference_protected(tbl->listeners, 1);
1577 1578 1579
		memcpy(new->masks, old->masks, NLGRPSZ(tbl->groups));
		rcu_assign_pointer(tbl->listeners, new);

1580
		kfree_rcu(old, rcu);
1581 1582 1583
	}
	tbl->groups = groups;

1584 1585 1586 1587 1588 1589 1590 1591 1592 1593 1594 1595 1596 1597 1598 1599 1600 1601 1602 1603 1604
	return 0;
}

/**
 * netlink_change_ngroups - change number of multicast groups
 *
 * This changes the number of multicast groups that are available
 * on a certain netlink family. Note that it is not possible to
 * change the number of groups to below 32. Also note that it does
 * not implicitly call netlink_clear_multicast_users() when the
 * number of groups is reduced.
 *
 * @sk: The kernel netlink socket, as returned by netlink_kernel_create().
 * @groups: The new number of groups.
 */
int netlink_change_ngroups(struct sock *sk, unsigned int groups)
{
	int err;

	netlink_table_grab();
	err = __netlink_change_ngroups(sk, groups);
1605
	netlink_table_ungrab();
1606

1607 1608 1609
	return err;
}

1610 1611 1612 1613 1614 1615 1616 1617 1618 1619
void __netlink_clear_multicast_users(struct sock *ksk, unsigned int group)
{
	struct sock *sk;
	struct hlist_node *node;
	struct netlink_table *tbl = &nl_table[ksk->sk_protocol];

	sk_for_each_bound(sk, node, &tbl->mc_list)
		netlink_update_socket_mc(nlk_sk(sk), group, 0);
}

1620 1621 1622 1623 1624 1625 1626 1627 1628 1629 1630
/**
 * netlink_clear_multicast_users - kick off multicast listeners
 *
 * This function removes all listeners from the given group.
 * @ksk: The kernel netlink socket, as returned by
 *	netlink_kernel_create().
 * @group: The multicast group to clear.
 */
void netlink_clear_multicast_users(struct sock *ksk, unsigned int group)
{
	netlink_table_grab();
1631
	__netlink_clear_multicast_users(ksk, group);
1632 1633 1634
	netlink_table_ungrab();
}

L
Linus Torvalds 已提交
1635
void netlink_set_nonroot(int protocol, unsigned int flags)
1636 1637
{
	if ((unsigned int)protocol < MAX_LINKS)
L
Linus Torvalds 已提交
1638
		nl_table[protocol].nl_nonroot = flags;
1639
}
1640
EXPORT_SYMBOL(netlink_set_nonroot);
L
Linus Torvalds 已提交
1641 1642 1643

static void netlink_destroy_callback(struct netlink_callback *cb)
{
1644
	kfree_skb(cb->skb);
L
Linus Torvalds 已提交
1645 1646 1647 1648 1649 1650 1651 1652 1653 1654 1655 1656
	kfree(cb);
}

/*
 * It looks a bit ugly.
 * It would be better to create kernel thread.
 */

static int netlink_dump(struct sock *sk)
{
	struct netlink_sock *nlk = nlk_sk(sk);
	struct netlink_callback *cb;
1657
	struct sk_buff *skb = NULL;
L
Linus Torvalds 已提交
1658
	struct nlmsghdr *nlh;
1659
	int len, err = -ENOBUFS;
1660
	int alloc_size;
L
Linus Torvalds 已提交
1661

1662
	mutex_lock(nlk->cb_mutex);
L
Linus Torvalds 已提交
1663 1664 1665

	cb = nlk->cb;
	if (cb == NULL) {
1666 1667
		err = -EINVAL;
		goto errout_skb;
L
Linus Torvalds 已提交
1668 1669
	}

1670 1671 1672 1673
	alloc_size = max_t(int, cb->min_dump_alloc, NLMSG_GOODSIZE);

	skb = sock_rmalloc(sk, alloc_size, 0, GFP_KERNEL);
	if (!skb)
1674
		goto errout_skb;
1675

L
Linus Torvalds 已提交
1676 1677 1678
	len = cb->dump(skb, cb);

	if (len > 0) {
1679
		mutex_unlock(nlk->cb_mutex);
1680 1681 1682 1683 1684 1685 1686

		if (sk_filter(sk, skb))
			kfree_skb(skb);
		else {
			skb_queue_tail(&sk->sk_receive_queue, skb);
			sk->sk_data_ready(sk, skb->len);
		}
L
Linus Torvalds 已提交
1687 1688 1689
		return 0;
	}

1690 1691 1692 1693
	nlh = nlmsg_put_answer(skb, cb, NLMSG_DONE, sizeof(len), NLM_F_MULTI);
	if (!nlh)
		goto errout_skb;

1694 1695
	nl_dump_check_consistent(cb, nlh);

1696 1697
	memcpy(nlmsg_data(nlh), &len, sizeof(len));

1698 1699 1700 1701 1702 1703
	if (sk_filter(sk, skb))
		kfree_skb(skb);
	else {
		skb_queue_tail(&sk->sk_receive_queue, skb);
		sk->sk_data_ready(sk, skb->len);
	}
L
Linus Torvalds 已提交
1704

1705 1706
	if (cb->done)
		cb->done(cb);
L
Linus Torvalds 已提交
1707
	nlk->cb = NULL;
1708
	mutex_unlock(nlk->cb_mutex);
L
Linus Torvalds 已提交
1709 1710 1711

	netlink_destroy_callback(cb);
	return 0;
1712

1713
errout_skb:
1714
	mutex_unlock(nlk->cb_mutex);
1715 1716
	kfree_skb(skb);
	return err;
L
Linus Torvalds 已提交
1717 1718 1719
}

int netlink_dump_start(struct sock *ssk, struct sk_buff *skb,
1720
		       const struct nlmsghdr *nlh,
1721 1722
		       int (*dump)(struct sk_buff *skb,
				   struct netlink_callback *),
1723 1724
		       int (*done)(struct netlink_callback *),
		       u16 min_dump_alloc)
L
Linus Torvalds 已提交
1725 1726 1727 1728
{
	struct netlink_callback *cb;
	struct sock *sk;
	struct netlink_sock *nlk;
1729
	int ret;
L
Linus Torvalds 已提交
1730

1731
	cb = kzalloc(sizeof(*cb), GFP_KERNEL);
L
Linus Torvalds 已提交
1732 1733 1734 1735 1736 1737
	if (cb == NULL)
		return -ENOBUFS;

	cb->dump = dump;
	cb->done = done;
	cb->nlh = nlh;
1738
	cb->min_dump_alloc = min_dump_alloc;
L
Linus Torvalds 已提交
1739 1740 1741
	atomic_inc(&skb->users);
	cb->skb = skb;

1742
	sk = netlink_lookup(sock_net(ssk), ssk->sk_protocol, NETLINK_CB(skb).pid);
L
Linus Torvalds 已提交
1743 1744 1745 1746 1747
	if (sk == NULL) {
		netlink_destroy_callback(cb);
		return -ECONNREFUSED;
	}
	nlk = nlk_sk(sk);
1748
	/* A dump is in progress... */
1749
	mutex_lock(nlk->cb_mutex);
1750
	if (nlk->cb) {
1751
		mutex_unlock(nlk->cb_mutex);
L
Linus Torvalds 已提交
1752 1753 1754 1755 1756
		netlink_destroy_callback(cb);
		sock_put(sk);
		return -EBUSY;
	}
	nlk->cb = cb;
1757
	mutex_unlock(nlk->cb_mutex);
L
Linus Torvalds 已提交
1758

1759 1760
	ret = netlink_dump(sk);

L
Linus Torvalds 已提交
1761
	sock_put(sk);
1762

1763 1764 1765
	if (ret)
		return ret;

1766 1767 1768 1769
	/* We successfully started a dump, by returning -EINTR we
	 * signal not to send ACK even if it was requested.
	 */
	return -EINTR;
L
Linus Torvalds 已提交
1770
}
1771
EXPORT_SYMBOL(netlink_dump_start);
L
Linus Torvalds 已提交
1772 1773 1774 1775 1776 1777

void netlink_ack(struct sk_buff *in_skb, struct nlmsghdr *nlh, int err)
{
	struct sk_buff *skb;
	struct nlmsghdr *rep;
	struct nlmsgerr *errmsg;
1778
	size_t payload = sizeof(*errmsg);
L
Linus Torvalds 已提交
1779

1780 1781 1782
	/* error messages get the original request appened */
	if (err)
		payload += nlmsg_len(nlh);
L
Linus Torvalds 已提交
1783

1784
	skb = nlmsg_new(payload, GFP_KERNEL);
L
Linus Torvalds 已提交
1785 1786 1787
	if (!skb) {
		struct sock *sk;

1788
		sk = netlink_lookup(sock_net(in_skb->sk),
1789
				    in_skb->sk->sk_protocol,
L
Linus Torvalds 已提交
1790 1791 1792 1793 1794 1795 1796 1797 1798 1799
				    NETLINK_CB(in_skb).pid);
		if (sk) {
			sk->sk_err = ENOBUFS;
			sk->sk_error_report(sk);
			sock_put(sk);
		}
		return;
	}

	rep = __nlmsg_put(skb, NETLINK_CB(in_skb).pid, nlh->nlmsg_seq,
1800
			  NLMSG_ERROR, payload, 0);
1801
	errmsg = nlmsg_data(rep);
L
Linus Torvalds 已提交
1802
	errmsg->error = err;
1803
	memcpy(&errmsg->msg, nlh, err ? nlh->nlmsg_len : sizeof(*nlh));
L
Linus Torvalds 已提交
1804 1805
	netlink_unicast(in_skb->sk, skb, NETLINK_CB(in_skb).pid, MSG_DONTWAIT);
}
1806
EXPORT_SYMBOL(netlink_ack);
L
Linus Torvalds 已提交
1807

1808
int netlink_rcv_skb(struct sk_buff *skb, int (*cb)(struct sk_buff *,
1809
						     struct nlmsghdr *))
1810 1811 1812 1813 1814
{
	struct nlmsghdr *nlh;
	int err;

	while (skb->len >= nlmsg_total_size(0)) {
1815 1816
		int msglen;

1817
		nlh = nlmsg_hdr(skb);
1818
		err = 0;
1819

1820
		if (nlh->nlmsg_len < NLMSG_HDRLEN || skb->len < nlh->nlmsg_len)
1821 1822
			return 0;

1823 1824
		/* Only requests are handled by the kernel */
		if (!(nlh->nlmsg_flags & NLM_F_REQUEST))
1825
			goto ack;
1826 1827 1828

		/* Skip control messages */
		if (nlh->nlmsg_type < NLMSG_MIN_TYPE)
1829
			goto ack;
1830

1831
		err = cb(skb, nlh);
1832 1833 1834 1835
		if (err == -EINTR)
			goto skip;

ack:
1836
		if (nlh->nlmsg_flags & NLM_F_ACK || err)
1837 1838
			netlink_ack(skb, nlh, err);

1839
skip:
1840
		msglen = NLMSG_ALIGN(nlh->nlmsg_len);
1841 1842 1843
		if (msglen > skb->len)
			msglen = skb->len;
		skb_pull(skb, msglen);
1844 1845 1846 1847
	}

	return 0;
}
1848
EXPORT_SYMBOL(netlink_rcv_skb);
1849

1850 1851 1852 1853 1854 1855 1856 1857 1858 1859 1860 1861 1862 1863 1864 1865 1866 1867 1868 1869 1870 1871
/**
 * nlmsg_notify - send a notification netlink message
 * @sk: netlink socket to use
 * @skb: notification message
 * @pid: destination netlink pid for reports or 0
 * @group: destination multicast group or 0
 * @report: 1 to report back, 0 to disable
 * @flags: allocation flags
 */
int nlmsg_notify(struct sock *sk, struct sk_buff *skb, u32 pid,
		 unsigned int group, int report, gfp_t flags)
{
	int err = 0;

	if (group) {
		int exclude_pid = 0;

		if (report) {
			atomic_inc(&skb->users);
			exclude_pid = pid;
		}

1872 1873 1874
		/* errors reported via destination sk->sk_err, but propagate
		 * delivery errors if NETLINK_BROADCAST_ERROR flag is set */
		err = nlmsg_multicast(sk, skb, exclude_pid, group, flags);
1875 1876
	}

1877 1878 1879 1880 1881 1882 1883
	if (report) {
		int err2;

		err2 = nlmsg_unicast(sk, skb, pid);
		if (!err || err == -ESRCH)
			err = err2;
	}
1884 1885 1886

	return err;
}
1887
EXPORT_SYMBOL(nlmsg_notify);
1888

L
Linus Torvalds 已提交
1889 1890
#ifdef CONFIG_PROC_FS
struct nl_seq_iter {
1891
	struct seq_net_private p;
L
Linus Torvalds 已提交
1892 1893 1894 1895 1896 1897 1898 1899 1900 1901 1902 1903
	int link;
	int hash_idx;
};

static struct sock *netlink_seq_socket_idx(struct seq_file *seq, loff_t pos)
{
	struct nl_seq_iter *iter = seq->private;
	int i, j;
	struct sock *s;
	struct hlist_node *node;
	loff_t off = 0;

1904
	for (i = 0; i < MAX_LINKS; i++) {
L
Linus Torvalds 已提交
1905 1906 1907 1908
		struct nl_pid_hash *hash = &nl_table[i].hash;

		for (j = 0; j <= hash->mask; j++) {
			sk_for_each(s, node, &hash->table[j]) {
1909
				if (sock_net(s) != seq_file_net(seq))
1910
					continue;
L
Linus Torvalds 已提交
1911 1912 1913 1914 1915 1916 1917 1918 1919 1920 1921 1922 1923
				if (off == pos) {
					iter->link = i;
					iter->hash_idx = j;
					return s;
				}
				++off;
			}
		}
	}
	return NULL;
}

static void *netlink_seq_start(struct seq_file *seq, loff_t *pos)
1924
	__acquires(nl_table_lock)
L
Linus Torvalds 已提交
1925 1926 1927 1928 1929 1930 1931 1932 1933 1934 1935 1936 1937 1938 1939
{
	read_lock(&nl_table_lock);
	return *pos ? netlink_seq_socket_idx(seq, *pos - 1) : SEQ_START_TOKEN;
}

static void *netlink_seq_next(struct seq_file *seq, void *v, loff_t *pos)
{
	struct sock *s;
	struct nl_seq_iter *iter;
	int i, j;

	++*pos;

	if (v == SEQ_START_TOKEN)
		return netlink_seq_socket_idx(seq, 0);
1940

1941 1942 1943 1944
	iter = seq->private;
	s = v;
	do {
		s = sk_next(s);
1945
	} while (s && sock_net(s) != seq_file_net(seq));
L
Linus Torvalds 已提交
1946 1947 1948 1949 1950 1951 1952 1953 1954 1955 1956
	if (s)
		return s;

	i = iter->link;
	j = iter->hash_idx + 1;

	do {
		struct nl_pid_hash *hash = &nl_table[i].hash;

		for (; j <= hash->mask; j++) {
			s = sk_head(&hash->table[j]);
1957
			while (s && sock_net(s) != seq_file_net(seq))
1958
				s = sk_next(s);
L
Linus Torvalds 已提交
1959 1960 1961 1962 1963 1964 1965 1966 1967 1968 1969 1970 1971 1972
			if (s) {
				iter->link = i;
				iter->hash_idx = j;
				return s;
			}
		}

		j = 0;
	} while (++i < MAX_LINKS);

	return NULL;
}

static void netlink_seq_stop(struct seq_file *seq, void *v)
1973
	__releases(nl_table_lock)
L
Linus Torvalds 已提交
1974 1975 1976 1977 1978 1979 1980 1981 1982 1983
{
	read_unlock(&nl_table_lock);
}


static int netlink_seq_show(struct seq_file *seq, void *v)
{
	if (v == SEQ_START_TOKEN)
		seq_puts(seq,
			 "sk       Eth Pid    Groups   "
1984
			 "Rmem     Wmem     Dump     Locks     Drops     Inode\n");
L
Linus Torvalds 已提交
1985 1986 1987 1988
	else {
		struct sock *s = v;
		struct netlink_sock *nlk = nlk_sk(s);

D
Dan Rosenberg 已提交
1989
		seq_printf(seq, "%pK %-3d %-6d %08x %-8d %-8d %pK %-8d %-8d %-8lu\n",
L
Linus Torvalds 已提交
1990 1991 1992
			   s,
			   s->sk_protocol,
			   nlk->pid,
1993
			   nlk->groups ? (u32)nlk->groups[0] : 0,
1994 1995
			   sk_rmem_alloc_get(s),
			   sk_wmem_alloc_get(s),
L
Linus Torvalds 已提交
1996
			   nlk->cb,
1997
			   atomic_read(&s->sk_refcnt),
1998 1999
			   atomic_read(&s->sk_drops),
			   sock_i_ino(s)
L
Linus Torvalds 已提交
2000 2001 2002 2003 2004 2005
			);

	}
	return 0;
}

2006
static const struct seq_operations netlink_seq_ops = {
L
Linus Torvalds 已提交
2007 2008 2009 2010 2011 2012 2013 2014 2015
	.start  = netlink_seq_start,
	.next   = netlink_seq_next,
	.stop   = netlink_seq_stop,
	.show   = netlink_seq_show,
};


static int netlink_seq_open(struct inode *inode, struct file *file)
{
2016 2017
	return seq_open_net(inode, file, &netlink_seq_ops,
				sizeof(struct nl_seq_iter));
2018 2019
}

2020
static const struct file_operations netlink_seq_fops = {
L
Linus Torvalds 已提交
2021 2022 2023 2024
	.owner		= THIS_MODULE,
	.open		= netlink_seq_open,
	.read		= seq_read,
	.llseek		= seq_lseek,
2025
	.release	= seq_release_net,
L
Linus Torvalds 已提交
2026 2027 2028 2029 2030 2031
};

#endif

int netlink_register_notifier(struct notifier_block *nb)
{
2032
	return atomic_notifier_chain_register(&netlink_chain, nb);
L
Linus Torvalds 已提交
2033
}
2034
EXPORT_SYMBOL(netlink_register_notifier);
L
Linus Torvalds 已提交
2035 2036 2037

int netlink_unregister_notifier(struct notifier_block *nb)
{
2038
	return atomic_notifier_chain_unregister(&netlink_chain, nb);
L
Linus Torvalds 已提交
2039
}
2040
EXPORT_SYMBOL(netlink_unregister_notifier);
2041

2042
static const struct proto_ops netlink_ops = {
L
Linus Torvalds 已提交
2043 2044 2045 2046 2047 2048 2049 2050 2051 2052 2053 2054
	.family =	PF_NETLINK,
	.owner =	THIS_MODULE,
	.release =	netlink_release,
	.bind =		netlink_bind,
	.connect =	netlink_connect,
	.socketpair =	sock_no_socketpair,
	.accept =	sock_no_accept,
	.getname =	netlink_getname,
	.poll =		datagram_poll,
	.ioctl =	sock_no_ioctl,
	.listen =	sock_no_listen,
	.shutdown =	sock_no_shutdown,
2055 2056
	.setsockopt =	netlink_setsockopt,
	.getsockopt =	netlink_getsockopt,
L
Linus Torvalds 已提交
2057 2058 2059 2060 2061 2062
	.sendmsg =	netlink_sendmsg,
	.recvmsg =	netlink_recvmsg,
	.mmap =		sock_no_mmap,
	.sendpage =	sock_no_sendpage,
};

2063
static const struct net_proto_family netlink_family_ops = {
L
Linus Torvalds 已提交
2064 2065 2066 2067 2068
	.family = PF_NETLINK,
	.create = netlink_create,
	.owner	= THIS_MODULE,	/* for consistency 8) */
};

2069
static int __net_init netlink_net_init(struct net *net)
2070 2071 2072 2073 2074 2075 2076 2077
{
#ifdef CONFIG_PROC_FS
	if (!proc_net_fops_create(net, "netlink", 0, &netlink_seq_fops))
		return -ENOMEM;
#endif
	return 0;
}

2078
static void __net_exit netlink_net_exit(struct net *net)
2079 2080 2081 2082 2083 2084
{
#ifdef CONFIG_PROC_FS
	proc_net_remove(net, "netlink");
#endif
}

2085 2086
static void __init netlink_add_usersock_entry(void)
{
2087
	struct listeners *listeners;
2088 2089
	int groups = 32;

2090
	listeners = kzalloc(sizeof(*listeners) + NLGRPSZ(groups), GFP_KERNEL);
2091
	if (!listeners)
2092
		panic("netlink_add_usersock_entry: Cannot allocate listeners\n");
2093 2094 2095 2096

	netlink_table_grab();

	nl_table[NETLINK_USERSOCK].groups = groups;
2097
	rcu_assign_pointer(nl_table[NETLINK_USERSOCK].listeners, listeners);
2098 2099 2100 2101 2102 2103
	nl_table[NETLINK_USERSOCK].module = THIS_MODULE;
	nl_table[NETLINK_USERSOCK].registered = 1;

	netlink_table_ungrab();
}

2104
static struct pernet_operations __net_initdata netlink_net_ops = {
2105 2106 2107 2108
	.init = netlink_net_init,
	.exit = netlink_net_exit,
};

L
Linus Torvalds 已提交
2109 2110 2111 2112
static int __init netlink_proto_init(void)
{
	struct sk_buff *dummy_skb;
	int i;
2113
	unsigned long limit;
L
Linus Torvalds 已提交
2114 2115 2116 2117 2118 2119
	unsigned int order;
	int err = proto_register(&netlink_proto, 0);

	if (err != 0)
		goto out;

2120
	BUILD_BUG_ON(sizeof(struct netlink_skb_parms) > sizeof(dummy_skb->cb));
L
Linus Torvalds 已提交
2121

2122
	nl_table = kcalloc(MAX_LINKS, sizeof(*nl_table), GFP_KERNEL);
2123 2124
	if (!nl_table)
		goto panic;
L
Linus Torvalds 已提交
2125

2126 2127
	if (totalram_pages >= (128 * 1024))
		limit = totalram_pages >> (21 - PAGE_SHIFT);
L
Linus Torvalds 已提交
2128
	else
2129
		limit = totalram_pages >> (23 - PAGE_SHIFT);
L
Linus Torvalds 已提交
2130

2131 2132 2133
	order = get_bitmask_order(limit) - 1 + PAGE_SHIFT;
	limit = (1UL << order) / sizeof(struct hlist_head);
	order = get_bitmask_order(min(limit, (unsigned long)UINT_MAX)) - 1;
L
Linus Torvalds 已提交
2134 2135 2136 2137

	for (i = 0; i < MAX_LINKS; i++) {
		struct nl_pid_hash *hash = &nl_table[i].hash;

E
Eric Dumazet 已提交
2138
		hash->table = nl_pid_hash_zalloc(1 * sizeof(*hash->table));
L
Linus Torvalds 已提交
2139 2140 2141 2142 2143
		if (!hash->table) {
			while (i-- > 0)
				nl_pid_hash_free(nl_table[i].hash.table,
						 1 * sizeof(*hash->table));
			kfree(nl_table);
2144
			goto panic;
L
Linus Torvalds 已提交
2145 2146 2147 2148 2149 2150 2151
		}
		hash->max_shift = order;
		hash->shift = 0;
		hash->mask = 0;
		hash->rehash_time = jiffies;
	}

2152 2153
	netlink_add_usersock_entry();

L
Linus Torvalds 已提交
2154
	sock_register(&netlink_family_ops);
2155
	register_pernet_subsys(&netlink_net_ops);
2156
	/* The netlink device handler may be needed early. */
L
Linus Torvalds 已提交
2157 2158 2159
	rtnetlink_init();
out:
	return err;
2160 2161
panic:
	panic("netlink_init: Cannot allocate nl_table\n");
L
Linus Torvalds 已提交
2162 2163 2164
}

core_initcall(netlink_proto_init);