af_netlink.c 43.4 KB
Newer Older
L
Linus Torvalds 已提交
1 2 3 4 5 6 7 8 9 10
/*
 * NETLINK      Kernel-user communication protocol.
 *
 * 		Authors:	Alan Cox <alan@redhat.com>
 * 				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/selinux.h>
58
#include <linux/mutex.h>
A
Andrew Morton 已提交
59

60
#include <net/net_namespace.h>
L
Linus Torvalds 已提交
61 62
#include <net/sock.h>
#include <net/scm.h>
63
#include <net/netlink.h>
L
Linus Torvalds 已提交
64

65
#define NLGRPSZ(x)	(ALIGN(x, sizeof(unsigned long) * 8) / 8)
66
#define NLGRPLONGS(x)	(NLGRPSZ(x)/sizeof(unsigned long))
L
Linus Torvalds 已提交
67 68 69 70 71 72

struct netlink_sock {
	/* struct sock has to be the first member of netlink_sock */
	struct sock		sk;
	u32			pid;
	u32			dst_pid;
73
	u32			dst_group;
74 75 76 77
	u32			flags;
	u32			subscriptions;
	u32			ngroups;
	unsigned long		*groups;
L
Linus Torvalds 已提交
78 79 80
	unsigned long		state;
	wait_queue_head_t	wait;
	struct netlink_callback	*cb;
81 82
	struct mutex		*cb_mutex;
	struct mutex		cb_def_mutex;
83
	void			(*netlink_rcv)(struct sk_buff *skb);
84
	struct module		*module;
L
Linus Torvalds 已提交
85 86
};

87
#define NETLINK_KERNEL_SOCKET	0x1
88
#define NETLINK_RECV_PKTINFO	0x2
89

L
Linus Torvalds 已提交
90 91
static inline struct netlink_sock *nlk_sk(struct sock *sk)
{
D
Denis Cheng 已提交
92
	return container_of(sk, struct netlink_sock, sk);
L
Linus Torvalds 已提交
93 94
}

95 96 97 98 99
static inline int netlink_is_kernel(struct sock *sk)
{
	return nlk_sk(sk)->flags & NETLINK_KERNEL_SOCKET;
}

L
Linus Torvalds 已提交
100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115
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;
116
	unsigned long *listeners;
L
Linus Torvalds 已提交
117
	unsigned int nl_nonroot;
118
	unsigned int groups;
119
	struct mutex *cb_mutex;
120
	struct module *module;
121
	int registered;
L
Linus Torvalds 已提交
122 123 124 125 126 127 128 129 130 131 132 133
};

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);

134
static ATOMIC_NOTIFIER_HEAD(netlink_chain);
L
Linus Torvalds 已提交
135

136 137 138 139 140
static u32 netlink_group_mask(u32 group)
{
	return group ? 1 << (group - 1) : 0;
}

L
Linus Torvalds 已提交
141 142 143 144 145 146 147
static struct hlist_head *nl_pid_hashfn(struct nl_pid_hash *hash, u32 pid)
{
	return &hash->table[jhash_1word(pid, hash->rnd) & hash->mask];
}

static void netlink_sock_destruct(struct sock *sk)
{
148 149 150 151 152 153 154 155
	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
156 157 158
	skb_queue_purge(&sk->sk_receive_queue);

	if (!sock_flag(sk, SOCK_DEAD)) {
159
		printk(KERN_ERR "Freeing alive netlink socket %p\n", sk);
L
Linus Torvalds 已提交
160 161 162 163
		return;
	}
	BUG_TRAP(!atomic_read(&sk->sk_rmem_alloc));
	BUG_TRAP(!atomic_read(&sk->sk_wmem_alloc));
164
	BUG_TRAP(!nlk_sk(sk)->groups);
L
Linus Torvalds 已提交
165 166
}

167 168
/* 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
169 170 171 172 173
 * immediately hit write lock and grab all the cpus. Exclusive sleep solves
 * this, _but_ remember, it adds useless work on UP machines.
 */

static void netlink_table_grab(void)
174
	__acquires(nl_table_lock)
L
Linus Torvalds 已提交
175
{
176
	write_lock_irq(&nl_table_lock);
L
Linus Torvalds 已提交
177 178 179 180 181

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

		add_wait_queue_exclusive(&nl_table_wait, &wait);
182
		for (;;) {
L
Linus Torvalds 已提交
183 184 185
			set_current_state(TASK_UNINTERRUPTIBLE);
			if (atomic_read(&nl_table_users) == 0)
				break;
186
			write_unlock_irq(&nl_table_lock);
L
Linus Torvalds 已提交
187
			schedule();
188
			write_lock_irq(&nl_table_lock);
L
Linus Torvalds 已提交
189 190 191 192 193 194 195
		}

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

196
static void netlink_table_ungrab(void)
197
	__releases(nl_table_lock)
L
Linus Torvalds 已提交
198
{
199
	write_unlock_irq(&nl_table_lock);
L
Linus Torvalds 已提交
200 201 202
	wake_up(&nl_table_wait);
}

203
static inline void
L
Linus Torvalds 已提交
204 205 206 207 208 209 210 211 212
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);
}

213
static inline void
L
Linus Torvalds 已提交
214 215 216 217 218 219
netlink_unlock_table(void)
{
	if (atomic_dec_and_test(&nl_table_users))
		wake_up(&nl_table_wait);
}

220 221
static inline struct sock *netlink_lookup(struct net *net, int protocol,
					  u32 pid)
L
Linus Torvalds 已提交
222 223 224 225 226 227 228 229 230
{
	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) {
231
		if ((sk->sk_net == net) && (nlk_sk(sk)->pid == pid)) {
L
Linus Torvalds 已提交
232 233 234 235 236 237 238 239 240 241
			sock_hold(sk);
			goto found;
		}
	}
	sk = NULL;
found:
	read_unlock(&nl_table_lock);
	return sk;
}

E
Eric Dumazet 已提交
242
static inline struct hlist_head *nl_pid_hash_zalloc(size_t size)
L
Linus Torvalds 已提交
243 244
{
	if (size <= PAGE_SIZE)
E
Eric Dumazet 已提交
245
		return kzalloc(size, GFP_ATOMIC);
L
Linus Torvalds 已提交
246 247
	else
		return (struct hlist_head *)
E
Eric Dumazet 已提交
248 249
			__get_free_pages(GFP_ATOMIC | __GFP_ZERO,
					 get_order(size));
L
Linus Torvalds 已提交
250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277
}

static inline void nl_pid_hash_free(struct hlist_head *table, size_t size)
{
	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
278
	table = nl_pid_hash_zalloc(size);
L
Linus Torvalds 已提交
279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315
	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;
}

316
static const struct proto_ops netlink_ops;
L
Linus Torvalds 已提交
317

318 319 320 321 322 323 324 325
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;

326
	for (i = 0; i < NLGRPLONGS(tbl->groups); i++) {
327
		mask = 0;
328 329 330 331
		sk_for_each_bound(sk, node, &tbl->mc_list) {
			if (i < NLGRPLONGS(nlk_sk(sk)->ngroups))
				mask |= nlk_sk(sk)->groups[i];
		}
332 333 334 335 336 337
		tbl->listeners[i] = mask;
	}
	/* this function is only called with the netlink table "grabbed", which
	 * makes sure updates are visible before bind or setsockopt return. */
}

338
static int netlink_insert(struct sock *sk, struct net *net, u32 pid)
L
Linus Torvalds 已提交
339 340 341 342 343 344 345 346 347 348 349 350
{
	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) {
351
		if ((osk->sk_net == net) && (nlk_sk(osk)->pid == pid))
L
Linus Torvalds 已提交
352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380
			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();
381 382
	if (sk_del_node_init(sk))
		nl_table[sk->sk_protocol].hash.entries--;
383
	if (nlk_sk(sk)->subscriptions)
L
Linus Torvalds 已提交
384 385 386 387 388 389 390 391 392 393
		__sk_del_bind_node(sk);
	netlink_table_ungrab();
}

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

394 395
static int __netlink_create(struct net *net, struct socket *sock,
			    struct mutex *cb_mutex, int protocol)
L
Linus Torvalds 已提交
396 397 398
{
	struct sock *sk;
	struct netlink_sock *nlk;
399 400 401

	sock->ops = &netlink_ops;

402
	sk = sk_alloc(net, PF_NETLINK, GFP_KERNEL, &netlink_proto);
403 404 405 406 407 408
	if (!sk)
		return -ENOMEM;

	sock_init_data(sock, sk);

	nlk = nlk_sk(sk);
409 410 411 412 413 414
	if (cb_mutex)
		nlk->cb_mutex = cb_mutex;
	else {
		nlk->cb_mutex = &nlk->cb_def_mutex;
		mutex_init(nlk->cb_mutex);
	}
415 416 417 418 419 420 421
	init_waitqueue_head(&nlk->wait);

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

422
static int netlink_create(struct net *net, struct socket *sock, int protocol)
423 424
{
	struct module *module = NULL;
425
	struct mutex *cb_mutex;
426
	struct netlink_sock *nlk;
427
	int err = 0;
L
Linus Torvalds 已提交
428 429 430 431 432 433

	sock->state = SS_UNCONNECTED;

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

434
	if (protocol < 0 || protocol >= MAX_LINKS)
L
Linus Torvalds 已提交
435 436
		return -EPROTONOSUPPORT;

437
	netlink_lock_table();
438
#ifdef CONFIG_KMOD
439
	if (!nl_table[protocol].registered) {
440
		netlink_unlock_table();
441
		request_module("net-pf-%d-proto-%d", PF_NETLINK, protocol);
442
		netlink_lock_table();
443
	}
444 445 446 447
#endif
	if (nl_table[protocol].registered &&
	    try_module_get(nl_table[protocol].module))
		module = nl_table[protocol].module;
448
	cb_mutex = nl_table[protocol].cb_mutex;
449
	netlink_unlock_table();
450

451 452
	err = __netlink_create(net, sock, cb_mutex, protocol);
	if (err < 0)
453 454 455 456
		goto out_module;

	nlk = nlk_sk(sock->sk);
	nlk->module = module;
457 458
out:
	return err;
L
Linus Torvalds 已提交
459

460 461 462
out_module:
	module_put(module);
	goto out;
L
Linus Torvalds 已提交
463 464 465 466 467 468 469 470 471 472 473
}

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

	if (!sk)
		return 0;

	netlink_remove(sk);
474
	sock_orphan(sk);
L
Linus Torvalds 已提交
475 476
	nlk = nlk_sk(sk);

477 478 479 480
	/*
	 * OK. Socket is unlinked, any packets that arrive now
	 * will be purged.
	 */
L
Linus Torvalds 已提交
481 482 483 484 485 486

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

	skb_queue_purge(&sk->sk_write_queue);

487
	if (nlk->pid && !nlk->subscriptions) {
L
Linus Torvalds 已提交
488
		struct netlink_notify n = {
489
						.net = sk->sk_net,
L
Linus Torvalds 已提交
490 491 492
						.protocol = sk->sk_protocol,
						.pid = nlk->pid,
					  };
493 494
		atomic_notifier_call_chain(&netlink_chain,
				NETLINK_URELEASE, &n);
495
	}
496

497
	module_put(nlk->module);
498

499
	netlink_table_grab();
500
	if (netlink_is_kernel(sk)) {
501 502 503 504 505 506
		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;
		}
507 508 509
	} else if (nlk->subscriptions)
		netlink_update_listeners(sk);
	netlink_table_ungrab();
510

511 512 513
	kfree(nlk->groups);
	nlk->groups = NULL;

L
Linus Torvalds 已提交
514 515 516 517 518 519 520
	sock_put(sk);
	return 0;
}

static int netlink_autobind(struct socket *sock)
{
	struct sock *sk = sock->sk;
521
	struct net *net = sk->sk_net;
L
Linus Torvalds 已提交
522 523 524 525
	struct nl_pid_hash *hash = &nl_table[sk->sk_protocol].hash;
	struct hlist_head *head;
	struct sock *osk;
	struct hlist_node *node;
526
	s32 pid = current->tgid;
L
Linus Torvalds 已提交
527 528 529 530 531 532 533 534
	int err;
	static s32 rover = -4097;

retry:
	cond_resched();
	netlink_table_grab();
	head = nl_pid_hashfn(hash, pid);
	sk_for_each(osk, node, head) {
535 536
		if ((osk->sk_net != net))
			continue;
L
Linus Torvalds 已提交
537 538 539 540 541 542 543 544 545 546 547
		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();

548
	err = netlink_insert(sk, net, pid);
L
Linus Torvalds 已提交
549 550
	if (err == -EADDRINUSE)
		goto retry;
551 552 553 554 555 556

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

	return err;
L
Linus Torvalds 已提交
557 558
}

559 560
static inline int netlink_capable(struct socket *sock, unsigned int flag)
{
L
Linus Torvalds 已提交
561 562
	return (nl_table[sock->sk->sk_protocol].nl_nonroot & flag) ||
	       capable(CAP_NET_ADMIN);
563
}
L
Linus Torvalds 已提交
564

565 566 567 568 569 570 571 572 573 574 575 576
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;
}

577
static int netlink_realloc_groups(struct sock *sk)
578 579 580
{
	struct netlink_sock *nlk = nlk_sk(sk);
	unsigned int groups;
581
	unsigned long *new_groups;
582 583
	int err = 0;

584 585
	netlink_table_grab();

586
	groups = nl_table[sk->sk_protocol].groups;
587
	if (!nl_table[sk->sk_protocol].registered) {
588
		err = -ENOENT;
589 590
		goto out_unlock;
	}
591

592 593
	if (nlk->ngroups >= groups)
		goto out_unlock;
594

595 596 597 598 599
	new_groups = krealloc(nlk->groups, NLGRPSZ(groups), GFP_ATOMIC);
	if (new_groups == NULL) {
		err = -ENOMEM;
		goto out_unlock;
	}
600
	memset((char *)new_groups + NLGRPSZ(nlk->ngroups), 0,
601 602 603
	       NLGRPSZ(groups) - NLGRPSZ(nlk->ngroups));

	nlk->groups = new_groups;
604
	nlk->ngroups = groups;
605 606 607
 out_unlock:
	netlink_table_ungrab();
	return err;
608 609
}

610 611
static int netlink_bind(struct socket *sock, struct sockaddr *addr,
			int addr_len)
L
Linus Torvalds 已提交
612 613
{
	struct sock *sk = sock->sk;
614
	struct net *net = sk->sk_net;
L
Linus Torvalds 已提交
615 616 617
	struct netlink_sock *nlk = nlk_sk(sk);
	struct sockaddr_nl *nladdr = (struct sockaddr_nl *)addr;
	int err;
618

L
Linus Torvalds 已提交
619 620 621 622
	if (nladdr->nl_family != AF_NETLINK)
		return -EINVAL;

	/* Only superuser is allowed to listen multicasts */
623 624 625
	if (nladdr->nl_groups) {
		if (!netlink_capable(sock, NL_NONROOT_RECV))
			return -EPERM;
626 627 628
		err = netlink_realloc_groups(sk);
		if (err)
			return err;
629
	}
L
Linus Torvalds 已提交
630 631 632 633 634 635

	if (nlk->pid) {
		if (nladdr->nl_pid != nlk->pid)
			return -EINVAL;
	} else {
		err = nladdr->nl_pid ?
636
			netlink_insert(sk, net, nladdr->nl_pid) :
L
Linus Torvalds 已提交
637 638 639 640 641
			netlink_autobind(sock);
		if (err)
			return err;
	}

642
	if (!nladdr->nl_groups && (nlk->groups == NULL || !(u32)nlk->groups[0]))
L
Linus Torvalds 已提交
643 644 645
		return 0;

	netlink_table_grab();
646
	netlink_update_subscriptions(sk, nlk->subscriptions +
647 648 649
					 hweight32(nladdr->nl_groups) -
					 hweight32(nlk->groups[0]));
	nlk->groups[0] = (nlk->groups[0] & ~0xffffffffUL) | nladdr->nl_groups;
650
	netlink_update_listeners(sk);
L
Linus Torvalds 已提交
651 652 653 654 655 656 657 658 659 660 661
	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);
662
	struct sockaddr_nl *nladdr = (struct sockaddr_nl *)addr;
L
Linus Torvalds 已提交
663 664 665 666

	if (addr->sa_family == AF_UNSPEC) {
		sk->sk_state	= NETLINK_UNCONNECTED;
		nlk->dst_pid	= 0;
667
		nlk->dst_group  = 0;
L
Linus Torvalds 已提交
668 669 670 671 672 673 674 675 676 677 678 679 680 681 682
		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;
683
		nlk->dst_group  = ffs(nladdr->nl_groups);
L
Linus Torvalds 已提交
684 685 686 687 688
	}

	return err;
}

689 690
static int netlink_getname(struct socket *sock, struct sockaddr *addr,
			   int *addr_len, int peer)
L
Linus Torvalds 已提交
691 692 693
{
	struct sock *sk = sock->sk;
	struct netlink_sock *nlk = nlk_sk(sk);
694
	struct sockaddr_nl *nladdr = (struct sockaddr_nl *)addr;
695

L
Linus Torvalds 已提交
696 697 698 699 700 701
	nladdr->nl_family = AF_NETLINK;
	nladdr->nl_pad = 0;
	*addr_len = sizeof(*nladdr);

	if (peer) {
		nladdr->nl_pid = nlk->dst_pid;
702
		nladdr->nl_groups = netlink_group_mask(nlk->dst_group);
L
Linus Torvalds 已提交
703 704
	} else {
		nladdr->nl_pid = nlk->pid;
705
		nladdr->nl_groups = nlk->groups ? nlk->groups[0] : 0;
L
Linus Torvalds 已提交
706 707 708 709 710 711 712 713 714 715 716 717 718 719 720 721 722
	}
	return 0;
}

static void netlink_overrun(struct sock *sk)
{
	if (!test_and_set_bit(0, &nlk_sk(sk)->state)) {
		sk->sk_err = ENOBUFS;
		sk->sk_error_report(sk);
	}
}

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

723
	sock = netlink_lookup(ssk->sk_net, ssk->sk_protocol, pid);
L
Linus Torvalds 已提交
724 725 726 727 728
	if (!sock)
		return ERR_PTR(-ECONNREFUSED);

	/* Don't bother queuing skb if kernel socket has no input function */
	nlk = nlk_sk(sock);
729 730
	if (sock->sk_state == NETLINK_CONNECTED &&
	    nlk->dst_pid != nlk_sk(ssk)->pid) {
L
Linus Torvalds 已提交
731 732 733 734 735 736 737 738
		sock_put(sock);
		return ERR_PTR(-ECONNREFUSED);
	}
	return sock;
}

struct sock *netlink_getsockbyfilp(struct file *filp)
{
J
Josef Sipek 已提交
739
	struct inode *inode = filp->f_path.dentry->d_inode;
L
Linus Torvalds 已提交
740 741 742 743 744 745 746 747 748 749 750 751 752 753 754 755 756 757 758 759 760 761 762
	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.
 */
A
Alexey Kuznetsov 已提交
763
int netlink_attachskb(struct sock *sk, struct sk_buff *skb, int nonblock,
P
Patrick McHardy 已提交
764
		      long *timeo, struct sock *ssk)
L
Linus Torvalds 已提交
765 766 767 768 769 770 771 772
{
	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
773
		if (!*timeo) {
774
			if (!ssk || netlink_is_kernel(ssk))
L
Linus Torvalds 已提交
775 776 777 778 779 780 781 782 783 784 785 786
				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
787
			*timeo = schedule_timeout(*timeo);
L
Linus Torvalds 已提交
788 789 790 791 792 793 794

		__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
795
			return sock_intr_errno(*timeo);
L
Linus Torvalds 已提交
796 797 798 799 800 801 802
		}
		return 1;
	}
	skb_set_owner_r(skb, sk);
	return 0;
}

803
int netlink_sendskb(struct sock *sk, struct sk_buff *skb)
L
Linus Torvalds 已提交
804 805 806 807 808 809 810 811 812 813 814 815 816 817 818
{
	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);
}

819
static inline struct sk_buff *netlink_trim(struct sk_buff *skb,
A
Al Viro 已提交
820
					   gfp_t allocation)
L
Linus Torvalds 已提交
821 822 823 824 825
{
	int delta;

	skb_orphan(skb);

826
	delta = skb->end - skb->tail;
L
Linus Torvalds 已提交
827 828 829 830 831 832 833 834 835 836 837 838 839 840 841 842 843
	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;
}

844 845 846 847 848 849 850 851 852 853 854 855 856 857 858 859 860 861 862 863 864 865 866 867 868 869 870 871
static inline void netlink_rcv_wake(struct sock *sk)
{
	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);
}

static inline int netlink_unicast_kernel(struct sock *sk, struct sk_buff *skb)
{
	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
872 873 874 875 876 877 878 879 880 881 882 883 884 885
{
	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);
	}
886 887 888
	if (netlink_is_kernel(sk))
		return netlink_unicast_kernel(sk, skb);

P
Patrick McHardy 已提交
889
	err = netlink_attachskb(sk, skb, nonblock, &timeo, ssk);
L
Linus Torvalds 已提交
890 891 892 893 894
	if (err == 1)
		goto retry;
	if (err)
		return err;

895
	return netlink_sendskb(sk, skb);
L
Linus Torvalds 已提交
896
}
897
EXPORT_SYMBOL(netlink_unicast);
L
Linus Torvalds 已提交
898

899 900 901
int netlink_has_listeners(struct sock *sk, unsigned int group)
{
	int res = 0;
902
	unsigned long *listeners;
903

904
	BUG_ON(!netlink_is_kernel(sk));
905 906 907 908

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

909
	if (group - 1 < nl_table[sk->sk_protocol].groups)
910 911 912 913
		res = test_bit(group - 1, listeners);

	rcu_read_unlock();

914 915 916 917
	return res;
}
EXPORT_SYMBOL_GPL(netlink_has_listeners);

918 919
static inline int netlink_broadcast_deliver(struct sock *sk,
					    struct sk_buff *skb)
L
Linus Torvalds 已提交
920 921 922 923 924 925 926 927 928 929 930 931 932 933 934
{
	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);
		return atomic_read(&sk->sk_rmem_alloc) > sk->sk_rcvbuf;
	}
	return -1;
}

struct netlink_broadcast_data {
	struct sock *exclude_sk;
935
	struct net *net;
L
Linus Torvalds 已提交
936 937 938 939 940
	u32 pid;
	u32 group;
	int failure;
	int congested;
	int delivered;
A
Al Viro 已提交
941
	gfp_t allocation;
L
Linus Torvalds 已提交
942 943 944 945 946 947 948 949 950 951 952 953
	struct sk_buff *skb, *skb2;
};

static inline int do_one_broadcast(struct sock *sk,
				   struct netlink_broadcast_data *p)
{
	struct netlink_sock *nlk = nlk_sk(sk);
	int val;

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

954 955
	if (nlk->pid == p->pid || p->group - 1 >= nlk->ngroups ||
	    !test_bit(p->group - 1, nlk->groups))
L
Linus Torvalds 已提交
956 957
		goto out;

958 959 960
	if ((sk->sk_net != p->net))
		goto out;

L
Linus Torvalds 已提交
961 962 963 964 965 966 967
	if (p->failure) {
		netlink_overrun(sk);
		goto out;
	}

	sock_hold(sk);
	if (p->skb2 == NULL) {
968
		if (skb_shared(p->skb)) {
L
Linus Torvalds 已提交
969 970
			p->skb2 = skb_clone(p->skb, p->allocation);
		} else {
971 972 973 974 975 976
			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
977 978 979 980 981 982 983 984 985 986 987 988 989 990 991 992 993 994 995 996
		}
	}
	if (p->skb2 == NULL) {
		netlink_overrun(sk);
		/* Clone failed. Notify ALL listeners. */
		p->failure = 1;
	} else if ((val = netlink_broadcast_deliver(sk, p->skb2)) < 0) {
		netlink_overrun(sk);
	} else {
		p->congested |= val;
		p->delivered = 1;
		p->skb2 = NULL;
	}
	sock_put(sk);

out:
	return 0;
}

int netlink_broadcast(struct sock *ssk, struct sk_buff *skb, u32 pid,
A
Al Viro 已提交
997
		      u32 group, gfp_t allocation)
L
Linus Torvalds 已提交
998
{
999
	struct net *net = ssk->sk_net;
L
Linus Torvalds 已提交
1000 1001 1002 1003 1004 1005 1006
	struct netlink_broadcast_data info;
	struct hlist_node *node;
	struct sock *sk;

	skb = netlink_trim(skb, allocation);

	info.exclude_sk = ssk;
1007
	info.net = net;
L
Linus Torvalds 已提交
1008 1009 1010 1011 1012 1013 1014 1015 1016 1017 1018 1019 1020 1021 1022 1023
	info.pid = pid;
	info.group = group;
	info.failure = 0;
	info.congested = 0;
	info.delivered = 0;
	info.allocation = allocation;
	info.skb = skb;
	info.skb2 = NULL;

	/* 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);

1024 1025
	kfree_skb(skb);

L
Linus Torvalds 已提交
1026 1027 1028 1029 1030 1031 1032 1033 1034 1035 1036 1037 1038 1039
	netlink_unlock_table();

	if (info.skb2)
		kfree_skb(info.skb2);

	if (info.delivered) {
		if (info.congested && (allocation & __GFP_WAIT))
			yield();
		return 0;
	}
	if (info.failure)
		return -ENOBUFS;
	return -ESRCH;
}
1040
EXPORT_SYMBOL(netlink_broadcast);
L
Linus Torvalds 已提交
1041 1042 1043 1044 1045 1046 1047 1048 1049 1050 1051 1052 1053 1054 1055 1056

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

static inline int do_one_set_err(struct sock *sk,
				 struct netlink_set_err_data *p)
{
	struct netlink_sock *nlk = nlk_sk(sk);

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

1057 1058 1059
	if (sk->sk_net != p->exclude_sk->sk_net)
		goto out;

1060 1061
	if (nlk->pid == p->pid || p->group - 1 >= nlk->ngroups ||
	    !test_bit(p->group - 1, nlk->groups))
L
Linus Torvalds 已提交
1062 1063 1064 1065 1066 1067 1068 1069 1070 1071 1072 1073 1074 1075 1076 1077 1078 1079 1080 1081 1082 1083 1084 1085 1086 1087 1088
		goto out;

	sk->sk_err = p->code;
	sk->sk_error_report(sk);
out:
	return 0;
}

void netlink_set_err(struct sock *ssk, u32 pid, u32 group, int code)
{
	struct netlink_set_err_data info;
	struct hlist_node *node;
	struct sock *sk;

	info.exclude_sk = ssk;
	info.pid = pid;
	info.group = group;
	info.code = code;

	read_lock(&nl_table_lock);

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

	read_unlock(&nl_table_lock);
}

1089 1090 1091 1092 1093 1094 1095 1096 1097 1098 1099 1100 1101 1102 1103 1104 1105
/* 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);
}

1106
static int netlink_setsockopt(struct socket *sock, int level, int optname,
1107
			      char __user *optval, int optlen)
1108 1109 1110
{
	struct sock *sk = sock->sk;
	struct netlink_sock *nlk = nlk_sk(sk);
1111 1112
	unsigned int val = 0;
	int err;
1113 1114 1115 1116 1117

	if (level != SOL_NETLINK)
		return -ENOPROTOOPT;

	if (optlen >= sizeof(int) &&
1118
	    get_user(val, (unsigned int __user *)optval))
1119 1120 1121 1122 1123 1124 1125 1126 1127 1128 1129 1130 1131 1132
		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;
1133 1134 1135
		err = netlink_realloc_groups(sk);
		if (err)
			return err;
1136 1137 1138
		if (!val || val - 1 >= nlk->ngroups)
			return -EINVAL;
		netlink_table_grab();
1139 1140
		netlink_update_socket_mc(nlk, val,
					 optname == NETLINK_ADD_MEMBERSHIP);
1141 1142 1143 1144 1145 1146 1147 1148 1149 1150 1151
		netlink_table_ungrab();
		err = 0;
		break;
	}
	default:
		err = -ENOPROTOOPT;
	}
	return err;
}

static int netlink_getsockopt(struct socket *sock, int level, int optname,
1152
			      char __user *optval, int __user *optlen)
1153 1154 1155 1156 1157 1158 1159 1160 1161 1162 1163 1164 1165 1166 1167 1168 1169 1170 1171
{
	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
1172 1173 1174
		if (put_user(len, optlen) ||
		    put_user(val, optval))
			return -EFAULT;
1175 1176 1177 1178 1179 1180 1181 1182 1183 1184 1185 1186 1187 1188 1189 1190
		err = 0;
		break;
	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
1191 1192 1193 1194 1195 1196
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);
1197
	struct sockaddr_nl *addr = msg->msg_name;
L
Linus Torvalds 已提交
1198
	u32 dst_pid;
1199
	u32 dst_group;
L
Linus Torvalds 已提交
1200 1201 1202 1203 1204 1205 1206 1207 1208 1209 1210 1211 1212 1213 1214 1215 1216
	struct sk_buff *skb;
	int err;
	struct scm_cookie scm;

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

	if (NULL == siocb->scm)
		siocb->scm = &scm;
	err = scm_send(sock, msg, siocb->scm);
	if (err < 0)
		return err;

	if (msg->msg_namelen) {
		if (addr->nl_family != AF_NETLINK)
			return -EINVAL;
		dst_pid = addr->nl_pid;
1217 1218
		dst_group = ffs(addr->nl_groups);
		if (dst_group && !netlink_capable(sock, NL_NONROOT_SEND))
L
Linus Torvalds 已提交
1219 1220 1221
			return -EPERM;
	} else {
		dst_pid = nlk->dst_pid;
1222
		dst_group = nlk->dst_group;
L
Linus Torvalds 已提交
1223 1224 1225 1226 1227 1228 1229 1230 1231 1232 1233 1234
	}

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

	err = -EMSGSIZE;
	if (len > sk->sk_sndbuf - 32)
		goto out;
	err = -ENOBUFS;
1235
	skb = alloc_skb(len, GFP_KERNEL);
1236
	if (skb == NULL)
L
Linus Torvalds 已提交
1237 1238 1239
		goto out;

	NETLINK_CB(skb).pid	= nlk->pid;
1240
	NETLINK_CB(skb).dst_group = dst_group;
1241
	NETLINK_CB(skb).loginuid = audit_get_loginuid(current->audit_context);
1242
	selinux_get_task_sid(current, &(NETLINK_CB(skb).sid));
L
Linus Torvalds 已提交
1243 1244 1245 1246 1247 1248 1249 1250 1251
	memcpy(NETLINK_CREDS(skb), &siocb->scm->creds, sizeof(struct ucred));

	/* What can I do? Netlink is asynchronous, so that
	   we will have to save current capabilities to
	   check them, when this message will be delivered
	   to corresponding kernel module.   --ANK (980802)
	 */

	err = -EFAULT;
1252
	if (memcpy_fromiovec(skb_put(skb, len), msg->msg_iov, len)) {
L
Linus Torvalds 已提交
1253 1254 1255 1256 1257 1258 1259 1260 1261 1262
		kfree_skb(skb);
		goto out;
	}

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

1263
	if (dst_group) {
L
Linus Torvalds 已提交
1264
		atomic_inc(&skb->users);
1265
		netlink_broadcast(sk, skb, dst_pid, dst_group, GFP_KERNEL);
L
Linus Torvalds 已提交
1266 1267 1268 1269 1270 1271 1272 1273 1274 1275 1276 1277 1278 1279 1280 1281 1282 1283 1284 1285 1286 1287 1288 1289 1290
	}
	err = netlink_unicast(sk, skb, dst_pid, msg->msg_flags&MSG_DONTWAIT);

out:
	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;
	struct sk_buff *skb;
	int err;

	if (flags&MSG_OOB)
		return -EOPNOTSUPP;

	copied = 0;

1291 1292
	skb = skb_recv_datagram(sk, flags, noblock, &err);
	if (skb == NULL)
L
Linus Torvalds 已提交
1293 1294 1295 1296 1297 1298 1299 1300 1301 1302
		goto out;

	msg->msg_namelen = 0;

	copied = skb->len;
	if (len < copied) {
		msg->msg_flags |= MSG_TRUNC;
		copied = len;
	}

1303
	skb_reset_transport_header(skb);
L
Linus Torvalds 已提交
1304 1305 1306
	err = skb_copy_datagram_iovec(skb, 0, msg->msg_iov, copied);

	if (msg->msg_name) {
1307
		struct sockaddr_nl *addr = (struct sockaddr_nl *)msg->msg_name;
L
Linus Torvalds 已提交
1308 1309 1310
		addr->nl_family = AF_NETLINK;
		addr->nl_pad    = 0;
		addr->nl_pid	= NETLINK_CB(skb).pid;
1311
		addr->nl_groups	= netlink_group_mask(NETLINK_CB(skb).dst_group);
L
Linus Torvalds 已提交
1312 1313 1314
		msg->msg_namelen = sizeof(*addr);
	}

1315 1316 1317
	if (nlk->flags & NETLINK_RECV_PKTINFO)
		netlink_cmsg_recv_pktinfo(msg, skb);

L
Linus Torvalds 已提交
1318 1319 1320 1321 1322
	if (NULL == siocb->scm) {
		memset(&scm, 0, sizeof(scm));
		siocb->scm = &scm;
	}
	siocb->scm->creds = *NETLINK_CREDS(skb);
1323 1324
	if (flags & MSG_TRUNC)
		copied = skb->len;
L
Linus Torvalds 已提交
1325 1326 1327 1328 1329 1330 1331 1332 1333 1334 1335 1336 1337
	skb_free_datagram(sk, skb);

	if (nlk->cb && atomic_read(&sk->sk_rmem_alloc) <= sk->sk_rcvbuf / 2)
		netlink_dump(sk);

	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)
{
1338
	BUG();
L
Linus Torvalds 已提交
1339 1340 1341
}

/*
1342
 *	We export these functions to other modules. They provide a
L
Linus Torvalds 已提交
1343 1344 1345 1346 1347
 *	complete set of kernel non-blocking support for message
 *	queueing.
 */

struct sock *
1348
netlink_kernel_create(struct net *net, int unit, unsigned int groups,
1349
		      void (*input)(struct sk_buff *skb),
1350
		      struct mutex *cb_mutex, struct module *module)
L
Linus Torvalds 已提交
1351 1352 1353
{
	struct socket *sock;
	struct sock *sk;
1354
	struct netlink_sock *nlk;
1355
	unsigned long *listeners = NULL;
L
Linus Torvalds 已提交
1356

1357
	BUG_ON(!nl_table);
L
Linus Torvalds 已提交
1358

1359
	if (unit < 0 || unit >= MAX_LINKS)
L
Linus Torvalds 已提交
1360 1361 1362 1363 1364
		return NULL;

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

1365
	if (__netlink_create(net, sock, cb_mutex, unit) < 0)
1366
		goto out_sock_release;
1367

1368 1369 1370 1371 1372 1373 1374
	if (groups < 32)
		groups = 32;

	listeners = kzalloc(NLGRPSZ(groups), GFP_KERNEL);
	if (!listeners)
		goto out_sock_release;

L
Linus Torvalds 已提交
1375 1376 1377
	sk = sock->sk;
	sk->sk_data_ready = netlink_data_ready;
	if (input)
1378
		nlk_sk(sk)->netlink_rcv = input;
L
Linus Torvalds 已提交
1379

1380
	if (netlink_insert(sk, net, 0))
1381
		goto out_sock_release;
1382

1383 1384
	nlk = nlk_sk(sk);
	nlk->flags |= NETLINK_KERNEL_SOCKET;
1385 1386

	netlink_table_grab();
1387 1388 1389 1390 1391 1392
	if (!nl_table[unit].registered) {
		nl_table[unit].groups = groups;
		nl_table[unit].listeners = listeners;
		nl_table[unit].cb_mutex = cb_mutex;
		nl_table[unit].module = module;
		nl_table[unit].registered = 1;
1393 1394
	} else {
		kfree(listeners);
1395
		nl_table[unit].registered++;
1396
	}
1397
	netlink_table_ungrab();
1398

1399 1400 1401
	/* Do not hold an extra referrence to a namespace as this socket is
	 * internal to a namespace and does not prevent it to stop. */
	put_net(net);
1402 1403
	return sk;

1404
out_sock_release:
1405
	kfree(listeners);
1406
	sock_release(sock);
1407
	return NULL;
L
Linus Torvalds 已提交
1408
}
1409
EXPORT_SYMBOL(netlink_kernel_create);
L
Linus Torvalds 已提交
1410

1411 1412 1413 1414 1415 1416

void
netlink_kernel_release(struct sock *sk)
{
	if (sk == NULL || sk->sk_socket == NULL)
		return;
1417 1418 1419 1420 1421 1422 1423 1424 1425

	/*
	 * Last sock_put should drop referrence to sk->sk_net. It has already
	 * been dropped in netlink_kernel_create. Taking referrence to stopping
	 * namespace is not an option.
	 * Take referrence to a socket to remove it from netlink lookup table
	 * _alive_ and after that destroy it in the context of init_net.
	 */
	sock_hold(sk);
1426
	sock_release(sk->sk_socket);
1427 1428 1429

	sk->sk_net = get_net(&init_net);
	sock_put(sk);
1430 1431 1432 1433
}
EXPORT_SYMBOL(netlink_kernel_release);


1434 1435 1436 1437 1438
/**
 * 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
1439 1440 1441
 * 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.
1442 1443 1444 1445 1446 1447 1448 1449 1450 1451 1452 1453 1454 1455 1456 1457 1458 1459 1460 1461 1462 1463 1464 1465 1466 1467 1468 1469 1470 1471 1472 1473 1474 1475
 *
 * @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)
{
	unsigned long *listeners, *old = NULL;
	struct netlink_table *tbl = &nl_table[sk->sk_protocol];
	int err = 0;

	if (groups < 32)
		groups = 32;

	netlink_table_grab();
	if (NLGRPSZ(tbl->groups) < NLGRPSZ(groups)) {
		listeners = kzalloc(NLGRPSZ(groups), GFP_ATOMIC);
		if (!listeners) {
			err = -ENOMEM;
			goto out_ungrab;
		}
		old = tbl->listeners;
		memcpy(listeners, old, NLGRPSZ(tbl->groups));
		rcu_assign_pointer(tbl->listeners, listeners);
	}
	tbl->groups = groups;

 out_ungrab:
	netlink_table_ungrab();
	synchronize_rcu();
	kfree(old);
	return err;
}
EXPORT_SYMBOL(netlink_change_ngroups);

1476 1477 1478 1479 1480 1481 1482 1483 1484 1485 1486 1487 1488 1489 1490 1491 1492 1493 1494 1495 1496 1497 1498
/**
 * 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)
{
	struct sock *sk;
	struct hlist_node *node;
	struct netlink_table *tbl = &nl_table[ksk->sk_protocol];

	netlink_table_grab();

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

	netlink_table_ungrab();
}
EXPORT_SYMBOL(netlink_clear_multicast_users);

L
Linus Torvalds 已提交
1499
void netlink_set_nonroot(int protocol, unsigned int flags)
1500 1501
{
	if ((unsigned int)protocol < MAX_LINKS)
L
Linus Torvalds 已提交
1502
		nl_table[protocol].nl_nonroot = flags;
1503
}
1504
EXPORT_SYMBOL(netlink_set_nonroot);
L
Linus Torvalds 已提交
1505 1506 1507 1508 1509 1510 1511 1512 1513 1514 1515 1516 1517 1518 1519 1520 1521 1522 1523

static void netlink_destroy_callback(struct netlink_callback *cb)
{
	if (cb->skb)
		kfree_skb(cb->skb);
	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;
	struct sk_buff *skb;
	struct nlmsghdr *nlh;
1524
	int len, err = -ENOBUFS;
1525

L
Linus Torvalds 已提交
1526 1527
	skb = sock_rmalloc(sk, NLMSG_GOODSIZE, 0, GFP_KERNEL);
	if (!skb)
1528
		goto errout;
L
Linus Torvalds 已提交
1529

1530
	mutex_lock(nlk->cb_mutex);
L
Linus Torvalds 已提交
1531 1532 1533

	cb = nlk->cb;
	if (cb == NULL) {
1534 1535
		err = -EINVAL;
		goto errout_skb;
L
Linus Torvalds 已提交
1536 1537 1538 1539 1540
	}

	len = cb->dump(skb, cb);

	if (len > 0) {
1541
		mutex_unlock(nlk->cb_mutex);
L
Linus Torvalds 已提交
1542 1543 1544 1545 1546
		skb_queue_tail(&sk->sk_receive_queue, skb);
		sk->sk_data_ready(sk, len);
		return 0;
	}

1547 1548 1549 1550 1551 1552
	nlh = nlmsg_put_answer(skb, cb, NLMSG_DONE, sizeof(len), NLM_F_MULTI);
	if (!nlh)
		goto errout_skb;

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

L
Linus Torvalds 已提交
1553 1554 1555
	skb_queue_tail(&sk->sk_receive_queue, skb);
	sk->sk_data_ready(sk, skb->len);

1556 1557
	if (cb->done)
		cb->done(cb);
L
Linus Torvalds 已提交
1558
	nlk->cb = NULL;
1559
	mutex_unlock(nlk->cb_mutex);
L
Linus Torvalds 已提交
1560 1561 1562

	netlink_destroy_callback(cb);
	return 0;
1563

1564
errout_skb:
1565
	mutex_unlock(nlk->cb_mutex);
1566 1567 1568
	kfree_skb(skb);
errout:
	return err;
L
Linus Torvalds 已提交
1569 1570 1571 1572
}

int netlink_dump_start(struct sock *ssk, struct sk_buff *skb,
		       struct nlmsghdr *nlh,
1573 1574 1575
		       int (*dump)(struct sk_buff *skb,
				   struct netlink_callback *),
		       int (*done)(struct netlink_callback *))
L
Linus Torvalds 已提交
1576 1577 1578 1579 1580
{
	struct netlink_callback *cb;
	struct sock *sk;
	struct netlink_sock *nlk;

1581
	cb = kzalloc(sizeof(*cb), GFP_KERNEL);
L
Linus Torvalds 已提交
1582 1583 1584 1585 1586 1587 1588 1589 1590
	if (cb == NULL)
		return -ENOBUFS;

	cb->dump = dump;
	cb->done = done;
	cb->nlh = nlh;
	atomic_inc(&skb->users);
	cb->skb = skb;

1591
	sk = netlink_lookup(ssk->sk_net, ssk->sk_protocol, NETLINK_CB(skb).pid);
L
Linus Torvalds 已提交
1592 1593 1594 1595 1596
	if (sk == NULL) {
		netlink_destroy_callback(cb);
		return -ECONNREFUSED;
	}
	nlk = nlk_sk(sk);
1597
	/* A dump is in progress... */
1598
	mutex_lock(nlk->cb_mutex);
1599
	if (nlk->cb) {
1600
		mutex_unlock(nlk->cb_mutex);
L
Linus Torvalds 已提交
1601 1602 1603 1604 1605
		netlink_destroy_callback(cb);
		sock_put(sk);
		return -EBUSY;
	}
	nlk->cb = cb;
1606
	mutex_unlock(nlk->cb_mutex);
L
Linus Torvalds 已提交
1607 1608 1609

	netlink_dump(sk);
	sock_put(sk);
1610 1611 1612 1613 1614

	/* 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
1615
}
1616
EXPORT_SYMBOL(netlink_dump_start);
L
Linus Torvalds 已提交
1617 1618 1619 1620 1621 1622

void netlink_ack(struct sk_buff *in_skb, struct nlmsghdr *nlh, int err)
{
	struct sk_buff *skb;
	struct nlmsghdr *rep;
	struct nlmsgerr *errmsg;
1623
	size_t payload = sizeof(*errmsg);
L
Linus Torvalds 已提交
1624

1625 1626 1627
	/* error messages get the original request appened */
	if (err)
		payload += nlmsg_len(nlh);
L
Linus Torvalds 已提交
1628

1629
	skb = nlmsg_new(payload, GFP_KERNEL);
L
Linus Torvalds 已提交
1630 1631 1632
	if (!skb) {
		struct sock *sk;

1633 1634
		sk = netlink_lookup(in_skb->sk->sk_net,
				    in_skb->sk->sk_protocol,
L
Linus Torvalds 已提交
1635 1636 1637 1638 1639 1640 1641 1642 1643 1644
				    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,
1645
			  NLMSG_ERROR, sizeof(struct nlmsgerr), 0);
1646
	errmsg = nlmsg_data(rep);
L
Linus Torvalds 已提交
1647
	errmsg->error = err;
1648
	memcpy(&errmsg->msg, nlh, err ? nlh->nlmsg_len : sizeof(*nlh));
L
Linus Torvalds 已提交
1649 1650
	netlink_unicast(in_skb->sk, skb, NETLINK_CB(in_skb).pid, MSG_DONTWAIT);
}
1651
EXPORT_SYMBOL(netlink_ack);
L
Linus Torvalds 已提交
1652

1653
int netlink_rcv_skb(struct sk_buff *skb, int (*cb)(struct sk_buff *,
1654
						     struct nlmsghdr *))
1655 1656 1657 1658 1659
{
	struct nlmsghdr *nlh;
	int err;

	while (skb->len >= nlmsg_total_size(0)) {
1660 1661
		int msglen;

1662
		nlh = nlmsg_hdr(skb);
1663
		err = 0;
1664

1665
		if (nlh->nlmsg_len < NLMSG_HDRLEN || skb->len < nlh->nlmsg_len)
1666 1667
			return 0;

1668 1669
		/* Only requests are handled by the kernel */
		if (!(nlh->nlmsg_flags & NLM_F_REQUEST))
1670
			goto ack;
1671 1672 1673

		/* Skip control messages */
		if (nlh->nlmsg_type < NLMSG_MIN_TYPE)
1674
			goto ack;
1675

1676
		err = cb(skb, nlh);
1677 1678 1679 1680
		if (err == -EINTR)
			goto skip;

ack:
1681
		if (nlh->nlmsg_flags & NLM_F_ACK || err)
1682 1683
			netlink_ack(skb, nlh, err);

1684
skip:
1685
		msglen = NLMSG_ALIGN(nlh->nlmsg_len);
1686 1687 1688
		if (msglen > skb->len)
			msglen = skb->len;
		skb_pull(skb, msglen);
1689 1690 1691 1692
	}

	return 0;
}
1693
EXPORT_SYMBOL(netlink_rcv_skb);
1694

1695 1696 1697 1698 1699 1700 1701 1702 1703 1704 1705 1706 1707 1708 1709 1710 1711 1712 1713 1714 1715 1716 1717 1718 1719 1720 1721 1722 1723 1724 1725
/**
 * 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;
		}

		/* errors reported via destination sk->sk_err */
		nlmsg_multicast(sk, skb, exclude_pid, group, flags);
	}

	if (report)
		err = nlmsg_unicast(sk, skb, pid);

	return err;
}
1726
EXPORT_SYMBOL(nlmsg_notify);
1727

L
Linus Torvalds 已提交
1728 1729
#ifdef CONFIG_PROC_FS
struct nl_seq_iter {
1730
	struct seq_net_private p;
L
Linus Torvalds 已提交
1731 1732 1733 1734 1735 1736 1737 1738 1739 1740 1741 1742
	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;

1743
	for (i = 0; i < MAX_LINKS; i++) {
L
Linus Torvalds 已提交
1744 1745 1746 1747
		struct nl_pid_hash *hash = &nl_table[i].hash;

		for (j = 0; j <= hash->mask; j++) {
			sk_for_each(s, node, &hash->table[j]) {
1748
				if (iter->p.net != s->sk_net)
1749
					continue;
L
Linus Torvalds 已提交
1750 1751 1752 1753 1754 1755 1756 1757 1758 1759 1760 1761 1762
				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)
1763
	__acquires(nl_table_lock)
L
Linus Torvalds 已提交
1764 1765 1766 1767 1768 1769 1770 1771 1772 1773 1774 1775 1776 1777 1778
{
	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);
1779

1780 1781 1782 1783
	iter = seq->private;
	s = v;
	do {
		s = sk_next(s);
1784
	} while (s && (iter->p.net != s->sk_net));
L
Linus Torvalds 已提交
1785 1786 1787 1788 1789 1790 1791 1792 1793 1794 1795
	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]);
1796
			while (s && (iter->p.net != s->sk_net))
1797
				s = sk_next(s);
L
Linus Torvalds 已提交
1798 1799 1800 1801 1802 1803 1804 1805 1806 1807 1808 1809 1810 1811
			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)
1812
	__releases(nl_table_lock)
L
Linus Torvalds 已提交
1813 1814 1815 1816 1817 1818 1819 1820 1821 1822 1823 1824 1825 1826 1827 1828 1829 1830 1831
{
	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   "
			 "Rmem     Wmem     Dump     Locks\n");
	else {
		struct sock *s = v;
		struct netlink_sock *nlk = nlk_sk(s);

		seq_printf(seq, "%p %-3d %-6d %08x %-8d %-8d %p %d\n",
			   s,
			   s->sk_protocol,
			   nlk->pid,
1832
			   nlk->groups ? (u32)nlk->groups[0] : 0,
L
Linus Torvalds 已提交
1833 1834 1835 1836 1837 1838 1839 1840 1841 1842
			   atomic_read(&s->sk_rmem_alloc),
			   atomic_read(&s->sk_wmem_alloc),
			   nlk->cb,
			   atomic_read(&s->sk_refcnt)
			);

	}
	return 0;
}

1843
static const struct seq_operations netlink_seq_ops = {
L
Linus Torvalds 已提交
1844 1845 1846 1847 1848 1849 1850 1851 1852
	.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)
{
1853 1854
	return seq_open_net(inode, file, &netlink_seq_ops,
				sizeof(struct nl_seq_iter));
1855 1856
}

1857
static const struct file_operations netlink_seq_fops = {
L
Linus Torvalds 已提交
1858 1859 1860 1861
	.owner		= THIS_MODULE,
	.open		= netlink_seq_open,
	.read		= seq_read,
	.llseek		= seq_lseek,
1862
	.release	= seq_release_net,
L
Linus Torvalds 已提交
1863 1864 1865 1866 1867 1868
};

#endif

int netlink_register_notifier(struct notifier_block *nb)
{
1869
	return atomic_notifier_chain_register(&netlink_chain, nb);
L
Linus Torvalds 已提交
1870
}
1871
EXPORT_SYMBOL(netlink_register_notifier);
L
Linus Torvalds 已提交
1872 1873 1874

int netlink_unregister_notifier(struct notifier_block *nb)
{
1875
	return atomic_notifier_chain_unregister(&netlink_chain, nb);
L
Linus Torvalds 已提交
1876
}
1877
EXPORT_SYMBOL(netlink_unregister_notifier);
1878

1879
static const struct proto_ops netlink_ops = {
L
Linus Torvalds 已提交
1880 1881 1882 1883 1884 1885 1886 1887 1888 1889 1890 1891
	.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,
1892 1893
	.setsockopt =	netlink_setsockopt,
	.getsockopt =	netlink_getsockopt,
L
Linus Torvalds 已提交
1894 1895 1896 1897 1898 1899 1900 1901 1902 1903 1904 1905
	.sendmsg =	netlink_sendmsg,
	.recvmsg =	netlink_recvmsg,
	.mmap =		sock_no_mmap,
	.sendpage =	sock_no_sendpage,
};

static struct net_proto_family netlink_family_ops = {
	.family = PF_NETLINK,
	.create = netlink_create,
	.owner	= THIS_MODULE,	/* for consistency 8) */
};

1906
static int __net_init netlink_net_init(struct net *net)
1907 1908 1909 1910 1911 1912 1913 1914
{
#ifdef CONFIG_PROC_FS
	if (!proc_net_fops_create(net, "netlink", 0, &netlink_seq_fops))
		return -ENOMEM;
#endif
	return 0;
}

1915
static void __net_exit netlink_net_exit(struct net *net)
1916 1917 1918 1919 1920 1921
{
#ifdef CONFIG_PROC_FS
	proc_net_remove(net, "netlink");
#endif
}

1922
static struct pernet_operations __net_initdata netlink_net_ops = {
1923 1924 1925 1926
	.init = netlink_net_init,
	.exit = netlink_net_exit,
};

L
Linus Torvalds 已提交
1927 1928 1929 1930
static int __init netlink_proto_init(void)
{
	struct sk_buff *dummy_skb;
	int i;
1931
	unsigned long limit;
L
Linus Torvalds 已提交
1932 1933 1934 1935 1936 1937
	unsigned int order;
	int err = proto_register(&netlink_proto, 0);

	if (err != 0)
		goto out;

1938
	BUILD_BUG_ON(sizeof(struct netlink_skb_parms) > sizeof(dummy_skb->cb));
L
Linus Torvalds 已提交
1939

1940
	nl_table = kcalloc(MAX_LINKS, sizeof(*nl_table), GFP_KERNEL);
1941 1942
	if (!nl_table)
		goto panic;
L
Linus Torvalds 已提交
1943 1944

	if (num_physpages >= (128 * 1024))
1945
		limit = num_physpages >> (21 - PAGE_SHIFT);
L
Linus Torvalds 已提交
1946
	else
1947
		limit = num_physpages >> (23 - PAGE_SHIFT);
L
Linus Torvalds 已提交
1948

1949 1950 1951
	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
1952 1953 1954 1955

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

E
Eric Dumazet 已提交
1956
		hash->table = nl_pid_hash_zalloc(1 * sizeof(*hash->table));
L
Linus Torvalds 已提交
1957 1958 1959 1960 1961
		if (!hash->table) {
			while (i-- > 0)
				nl_pid_hash_free(nl_table[i].hash.table,
						 1 * sizeof(*hash->table));
			kfree(nl_table);
1962
			goto panic;
L
Linus Torvalds 已提交
1963 1964 1965 1966 1967 1968 1969 1970
		}
		hash->max_shift = order;
		hash->shift = 0;
		hash->mask = 0;
		hash->rehash_time = jiffies;
	}

	sock_register(&netlink_family_ops);
1971
	register_pernet_subsys(&netlink_net_ops);
1972
	/* The netlink device handler may be needed early. */
L
Linus Torvalds 已提交
1973 1974 1975
	rtnetlink_init();
out:
	return err;
1976 1977
panic:
	panic("netlink_init: Cannot allocate nl_table\n");
L
Linus Torvalds 已提交
1978 1979 1980
}

core_initcall(netlink_proto_init);