af_netlink.c 74.4 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
 * 				Alexey Kuznetsov <kuznet@ms2.inr.ac.ru>
6
 * 				Patrick McHardy <kaber@trash.net>
L
Linus Torvalds 已提交
7 8 9 10 11
 *
 *		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.
12
 *
L
Linus Torvalds 已提交
13 14 15 16
 * 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)
17 18 19 20 21 22
 * 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
23 24 25 26
 */

#include <linux/module.h>

27
#include <linux/capability.h>
L
Linus Torvalds 已提交
28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56
#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
57
#include <linux/audit.h>
58
#include <linux/mutex.h>
59
#include <linux/vmalloc.h>
60
#include <linux/if_arp.h>
61
#include <linux/rhashtable.h>
62
#include <asm/cacheflush.h>
63
#include <linux/hash.h>
64
#include <linux/genetlink.h>
A
Andrew Morton 已提交
65

66
#include <net/net_namespace.h>
L
Linus Torvalds 已提交
67 68
#include <net/sock.h>
#include <net/scm.h>
69
#include <net/netlink.h>
L
Linus Torvalds 已提交
70

71
#include "af_netlink.h"
L
Linus Torvalds 已提交
72

73 74 75
struct listeners {
	struct rcu_head		rcu;
	unsigned long		masks[0];
76 77
};

78
/* state bits */
79
#define NETLINK_S_CONGESTED		0x0
80 81

/* flags */
82 83 84 85
#define NETLINK_F_KERNEL_SOCKET		0x1
#define NETLINK_F_RECV_PKTINFO		0x2
#define NETLINK_F_BROADCAST_SEND_ERROR	0x4
#define NETLINK_F_RECV_NO_ENOBUFS	0x8
86
#define NETLINK_F_LISTEN_ALL_NSID	0x10
87

88
static inline int netlink_is_kernel(struct sock *sk)
89
{
90
	return nlk_sk(sk)->flags & NETLINK_F_KERNEL_SOCKET;
91 92
}

93 94
struct netlink_table *nl_table;
EXPORT_SYMBOL_GPL(nl_table);
L
Linus Torvalds 已提交
95 96 97 98

static DECLARE_WAIT_QUEUE_HEAD(nl_table_wait);

static int netlink_dump(struct sock *sk);
99
static void netlink_skb_destructor(struct sk_buff *skb);
L
Linus Torvalds 已提交
100

101
/* nl_table locking explained:
102
 * Lookup and traversal are protected with an RCU read-side lock. Insertion
Y
Ying Xue 已提交
103
 * and removal are protected with per bucket lock while using RCU list
104 105 106 107
 * modification primitives and may run in parallel to RCU protected lookups.
 * Destruction of the Netlink socket may only occur *after* nl_table_lock has
 * been acquired * either during or after the socket has been removed from
 * the list and after an RCU grace period.
108
 */
109 110
DEFINE_RWLOCK(nl_table_lock);
EXPORT_SYMBOL_GPL(nl_table_lock);
L
Linus Torvalds 已提交
111 112
static atomic_t nl_table_users = ATOMIC_INIT(0);

113 114
#define nl_deref_protected(X) rcu_dereference_protected(X, lockdep_is_held(&nl_table_lock));

115
static ATOMIC_NOTIFIER_HEAD(netlink_chain);
L
Linus Torvalds 已提交
116

117 118 119
static DEFINE_SPINLOCK(netlink_tap_lock);
static struct list_head netlink_tap_all __read_mostly;

120 121
static const struct rhashtable_params netlink_rhashtable_params;

122
static inline u32 netlink_group_mask(u32 group)
123 124 125 126
{
	return group ? 1 << (group - 1) : 0;
}

127 128 129 130 131 132 133 134 135
int netlink_add_tap(struct netlink_tap *nt)
{
	if (unlikely(nt->dev->type != ARPHRD_NETLINK))
		return -EINVAL;

	spin_lock(&netlink_tap_lock);
	list_add_rcu(&nt->list, &netlink_tap_all);
	spin_unlock(&netlink_tap_lock);

136
	__module_get(nt->module);
137 138 139 140 141

	return 0;
}
EXPORT_SYMBOL_GPL(netlink_add_tap);

142
static int __netlink_remove_tap(struct netlink_tap *nt)
143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177
{
	bool found = false;
	struct netlink_tap *tmp;

	spin_lock(&netlink_tap_lock);

	list_for_each_entry(tmp, &netlink_tap_all, list) {
		if (nt == tmp) {
			list_del_rcu(&nt->list);
			found = true;
			goto out;
		}
	}

	pr_warn("__netlink_remove_tap: %p not found\n", nt);
out:
	spin_unlock(&netlink_tap_lock);

	if (found && nt->module)
		module_put(nt->module);

	return found ? 0 : -ENODEV;
}

int netlink_remove_tap(struct netlink_tap *nt)
{
	int ret;

	ret = __netlink_remove_tap(nt);
	synchronize_net();

	return ret;
}
EXPORT_SYMBOL_GPL(netlink_remove_tap);

178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193
static bool netlink_filter_tap(const struct sk_buff *skb)
{
	struct sock *sk = skb->sk;

	/* We take the more conservative approach and
	 * whitelist socket protocols that may pass.
	 */
	switch (sk->sk_protocol) {
	case NETLINK_ROUTE:
	case NETLINK_USERSOCK:
	case NETLINK_SOCK_DIAG:
	case NETLINK_NFLOG:
	case NETLINK_XFRM:
	case NETLINK_FIB_LOOKUP:
	case NETLINK_NETFILTER:
	case NETLINK_GENERIC:
V
Varka Bhadram 已提交
194
		return true;
195 196
	}

V
Varka Bhadram 已提交
197
	return false;
198 199
}

200 201 202 203
static int __netlink_deliver_tap_skb(struct sk_buff *skb,
				     struct net_device *dev)
{
	struct sk_buff *nskb;
204
	struct sock *sk = skb->sk;
205 206 207 208 209 210
	int ret = -ENOMEM;

	dev_hold(dev);
	nskb = skb_clone(skb, GFP_ATOMIC);
	if (nskb) {
		nskb->dev = dev;
211
		nskb->protocol = htons((u16) sk->sk_protocol);
212 213
		nskb->pkt_type = netlink_is_kernel(sk) ?
				 PACKET_KERNEL : PACKET_USER;
214
		skb_reset_network_header(nskb);
215 216 217 218 219 220 221 222 223 224 225 226 227 228
		ret = dev_queue_xmit(nskb);
		if (unlikely(ret > 0))
			ret = net_xmit_errno(ret);
	}

	dev_put(dev);
	return ret;
}

static void __netlink_deliver_tap(struct sk_buff *skb)
{
	int ret;
	struct netlink_tap *tmp;

229 230 231
	if (!netlink_filter_tap(skb))
		return;

232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248
	list_for_each_entry_rcu(tmp, &netlink_tap_all, list) {
		ret = __netlink_deliver_tap_skb(skb, tmp->dev);
		if (unlikely(ret))
			break;
	}
}

static void netlink_deliver_tap(struct sk_buff *skb)
{
	rcu_read_lock();

	if (unlikely(!list_empty(&netlink_tap_all)))
		__netlink_deliver_tap(skb);

	rcu_read_unlock();
}

249 250 251 252 253 254 255
static void netlink_deliver_tap_kernel(struct sock *dst, struct sock *src,
				       struct sk_buff *skb)
{
	if (!(netlink_is_kernel(dst) && netlink_is_kernel(src)))
		netlink_deliver_tap(skb);
}

256 257 258 259
static void netlink_overrun(struct sock *sk)
{
	struct netlink_sock *nlk = nlk_sk(sk);

260 261 262
	if (!(nlk->flags & NETLINK_F_RECV_NO_ENOBUFS)) {
		if (!test_and_set_bit(NETLINK_S_CONGESTED,
				      &nlk_sk(sk)->state)) {
263 264 265 266 267 268 269 270 271 272 273 274
			sk->sk_err = ENOBUFS;
			sk->sk_error_report(sk);
		}
	}
	atomic_inc(&sk->sk_drops);
}

static void netlink_rcv_wake(struct sock *sk)
{
	struct netlink_sock *nlk = nlk_sk(sk);

	if (skb_queue_empty(&sk->sk_receive_queue))
275 276
		clear_bit(NETLINK_S_CONGESTED, &nlk->state);
	if (!test_bit(NETLINK_S_CONGESTED, &nlk->state))
277 278 279
		wake_up_interruptible(&nlk->wait);
}

280
#ifdef CONFIG_NETLINK_MMAP
281 282 283 284 285
static bool netlink_skb_is_mmaped(const struct sk_buff *skb)
{
	return NETLINK_CB(skb).flags & NETLINK_SKB_MMAPED;
}

286 287 288 289 290
static bool netlink_rx_is_mmaped(struct sock *sk)
{
	return nlk_sk(sk)->rx_ring.pg_vec != NULL;
}

291 292 293 294 295
static bool netlink_tx_is_mmaped(struct sock *sk)
{
	return nlk_sk(sk)->tx_ring.pg_vec != NULL;
}

296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341
static __pure struct page *pgvec_to_page(const void *addr)
{
	if (is_vmalloc_addr(addr))
		return vmalloc_to_page(addr);
	else
		return virt_to_page(addr);
}

static void free_pg_vec(void **pg_vec, unsigned int order, unsigned int len)
{
	unsigned int i;

	for (i = 0; i < len; i++) {
		if (pg_vec[i] != NULL) {
			if (is_vmalloc_addr(pg_vec[i]))
				vfree(pg_vec[i]);
			else
				free_pages((unsigned long)pg_vec[i], order);
		}
	}
	kfree(pg_vec);
}

static void *alloc_one_pg_vec_page(unsigned long order)
{
	void *buffer;
	gfp_t gfp_flags = GFP_KERNEL | __GFP_COMP | __GFP_ZERO |
			  __GFP_NOWARN | __GFP_NORETRY;

	buffer = (void *)__get_free_pages(gfp_flags, order);
	if (buffer != NULL)
		return buffer;

	buffer = vzalloc((1 << order) * PAGE_SIZE);
	if (buffer != NULL)
		return buffer;

	gfp_flags &= ~__GFP_NORETRY;
	return (void *)__get_free_pages(gfp_flags, order);
}

static void **alloc_pg_vec(struct netlink_sock *nlk,
			   struct nl_mmap_req *req, unsigned int order)
{
	unsigned int block_nr = req->nm_block_nr;
	unsigned int i;
342
	void **pg_vec;
343 344 345 346 347 348

	pg_vec = kcalloc(block_nr, sizeof(void *), GFP_KERNEL);
	if (pg_vec == NULL)
		return NULL;

	for (i = 0; i < block_nr; i++) {
349
		pg_vec[i] = alloc_one_pg_vec_page(order);
350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385
		if (pg_vec[i] == NULL)
			goto err1;
	}

	return pg_vec;
err1:
	free_pg_vec(pg_vec, order, block_nr);
	return NULL;
}

static int netlink_set_ring(struct sock *sk, struct nl_mmap_req *req,
			    bool closing, bool tx_ring)
{
	struct netlink_sock *nlk = nlk_sk(sk);
	struct netlink_ring *ring;
	struct sk_buff_head *queue;
	void **pg_vec = NULL;
	unsigned int order = 0;
	int err;

	ring  = tx_ring ? &nlk->tx_ring : &nlk->rx_ring;
	queue = tx_ring ? &sk->sk_write_queue : &sk->sk_receive_queue;

	if (!closing) {
		if (atomic_read(&nlk->mapped))
			return -EBUSY;
		if (atomic_read(&ring->pending))
			return -EBUSY;
	}

	if (req->nm_block_nr) {
		if (ring->pg_vec != NULL)
			return -EBUSY;

		if ((int)req->nm_block_size <= 0)
			return -EINVAL;
T
Tobias Klauser 已提交
386
		if (!PAGE_ALIGNED(req->nm_block_size))
387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 439 440 441 442 443 444 445 446 447 448 449 450 451 452 453 454 455 456 457 458 459 460 461 462 463 464 465 466 467 468 469 470 471 472 473 474 475 476 477 478 479 480 481 482 483 484 485 486 487 488 489 490 491 492 493 494 495 496 497 498 499 500 501 502 503 504 505 506 507 508 509 510 511 512 513 514 515 516
			return -EINVAL;
		if (req->nm_frame_size < NL_MMAP_HDRLEN)
			return -EINVAL;
		if (!IS_ALIGNED(req->nm_frame_size, NL_MMAP_MSG_ALIGNMENT))
			return -EINVAL;

		ring->frames_per_block = req->nm_block_size /
					 req->nm_frame_size;
		if (ring->frames_per_block == 0)
			return -EINVAL;
		if (ring->frames_per_block * req->nm_block_nr !=
		    req->nm_frame_nr)
			return -EINVAL;

		order = get_order(req->nm_block_size);
		pg_vec = alloc_pg_vec(nlk, req, order);
		if (pg_vec == NULL)
			return -ENOMEM;
	} else {
		if (req->nm_frame_nr)
			return -EINVAL;
	}

	err = -EBUSY;
	mutex_lock(&nlk->pg_vec_lock);
	if (closing || atomic_read(&nlk->mapped) == 0) {
		err = 0;
		spin_lock_bh(&queue->lock);

		ring->frame_max		= req->nm_frame_nr - 1;
		ring->head		= 0;
		ring->frame_size	= req->nm_frame_size;
		ring->pg_vec_pages	= req->nm_block_size / PAGE_SIZE;

		swap(ring->pg_vec_len, req->nm_block_nr);
		swap(ring->pg_vec_order, order);
		swap(ring->pg_vec, pg_vec);

		__skb_queue_purge(queue);
		spin_unlock_bh(&queue->lock);

		WARN_ON(atomic_read(&nlk->mapped));
	}
	mutex_unlock(&nlk->pg_vec_lock);

	if (pg_vec)
		free_pg_vec(pg_vec, order, req->nm_block_nr);
	return err;
}

static void netlink_mm_open(struct vm_area_struct *vma)
{
	struct file *file = vma->vm_file;
	struct socket *sock = file->private_data;
	struct sock *sk = sock->sk;

	if (sk)
		atomic_inc(&nlk_sk(sk)->mapped);
}

static void netlink_mm_close(struct vm_area_struct *vma)
{
	struct file *file = vma->vm_file;
	struct socket *sock = file->private_data;
	struct sock *sk = sock->sk;

	if (sk)
		atomic_dec(&nlk_sk(sk)->mapped);
}

static const struct vm_operations_struct netlink_mmap_ops = {
	.open	= netlink_mm_open,
	.close	= netlink_mm_close,
};

static int netlink_mmap(struct file *file, struct socket *sock,
			struct vm_area_struct *vma)
{
	struct sock *sk = sock->sk;
	struct netlink_sock *nlk = nlk_sk(sk);
	struct netlink_ring *ring;
	unsigned long start, size, expected;
	unsigned int i;
	int err = -EINVAL;

	if (vma->vm_pgoff)
		return -EINVAL;

	mutex_lock(&nlk->pg_vec_lock);

	expected = 0;
	for (ring = &nlk->rx_ring; ring <= &nlk->tx_ring; ring++) {
		if (ring->pg_vec == NULL)
			continue;
		expected += ring->pg_vec_len * ring->pg_vec_pages * PAGE_SIZE;
	}

	if (expected == 0)
		goto out;

	size = vma->vm_end - vma->vm_start;
	if (size != expected)
		goto out;

	start = vma->vm_start;
	for (ring = &nlk->rx_ring; ring <= &nlk->tx_ring; ring++) {
		if (ring->pg_vec == NULL)
			continue;

		for (i = 0; i < ring->pg_vec_len; i++) {
			struct page *page;
			void *kaddr = ring->pg_vec[i];
			unsigned int pg_num;

			for (pg_num = 0; pg_num < ring->pg_vec_pages; pg_num++) {
				page = pgvec_to_page(kaddr);
				err = vm_insert_page(vma, start, page);
				if (err < 0)
					goto out;
				start += PAGE_SIZE;
				kaddr += PAGE_SIZE;
			}
		}
	}

	atomic_inc(&nlk->mapped);
	vma->vm_ops = &netlink_mmap_ops;
	err = 0;
out:
	mutex_unlock(&nlk->pg_vec_lock);
517
	return err;
518
}
519

D
David Miller 已提交
520
static void netlink_frame_flush_dcache(const struct nl_mmap_hdr *hdr, unsigned int nm_len)
521 522 523 524 525 526
{
#if ARCH_IMPLEMENTS_FLUSH_DCACHE_PAGE == 1
	struct page *p_start, *p_end;

	/* First page is flushed through netlink_{get,set}_status */
	p_start = pgvec_to_page(hdr + PAGE_SIZE);
D
David Miller 已提交
527
	p_end   = pgvec_to_page((void *)hdr + NL_MMAP_HDRLEN + nm_len - 1);
528 529 530 531 532 533 534 535 536 537 538 539 540 541 542 543 544
	while (p_start <= p_end) {
		flush_dcache_page(p_start);
		p_start++;
	}
#endif
}

static enum nl_mmap_status netlink_get_status(const struct nl_mmap_hdr *hdr)
{
	smp_rmb();
	flush_dcache_page(pgvec_to_page(hdr));
	return hdr->nm_status;
}

static void netlink_set_status(struct nl_mmap_hdr *hdr,
			       enum nl_mmap_status status)
{
545
	smp_mb();
546 547 548 549 550 551 552 553 554 555 556 557 558 559 560 561 562 563 564 565 566 567 568 569 570 571 572 573 574 575 576 577 578 579 580 581 582 583 584 585 586 587 588 589 590 591 592 593 594 595 596 597 598 599 600 601 602 603 604 605 606 607 608 609 610
	hdr->nm_status = status;
	flush_dcache_page(pgvec_to_page(hdr));
}

static struct nl_mmap_hdr *
__netlink_lookup_frame(const struct netlink_ring *ring, unsigned int pos)
{
	unsigned int pg_vec_pos, frame_off;

	pg_vec_pos = pos / ring->frames_per_block;
	frame_off  = pos % ring->frames_per_block;

	return ring->pg_vec[pg_vec_pos] + (frame_off * ring->frame_size);
}

static struct nl_mmap_hdr *
netlink_lookup_frame(const struct netlink_ring *ring, unsigned int pos,
		     enum nl_mmap_status status)
{
	struct nl_mmap_hdr *hdr;

	hdr = __netlink_lookup_frame(ring, pos);
	if (netlink_get_status(hdr) != status)
		return NULL;

	return hdr;
}

static struct nl_mmap_hdr *
netlink_current_frame(const struct netlink_ring *ring,
		      enum nl_mmap_status status)
{
	return netlink_lookup_frame(ring, ring->head, status);
}

static struct nl_mmap_hdr *
netlink_previous_frame(const struct netlink_ring *ring,
		       enum nl_mmap_status status)
{
	unsigned int prev;

	prev = ring->head ? ring->head - 1 : ring->frame_max;
	return netlink_lookup_frame(ring, prev, status);
}

static void netlink_increment_head(struct netlink_ring *ring)
{
	ring->head = ring->head != ring->frame_max ? ring->head + 1 : 0;
}

static void netlink_forward_ring(struct netlink_ring *ring)
{
	unsigned int head = ring->head, pos = head;
	const struct nl_mmap_hdr *hdr;

	do {
		hdr = __netlink_lookup_frame(ring, pos);
		if (hdr->nm_status == NL_MMAP_STATUS_UNUSED)
			break;
		if (hdr->nm_status != NL_MMAP_STATUS_SKIP)
			break;
		netlink_increment_head(ring);
	} while (ring->head != head);
}

611 612 613 614 615 616 617 618 619 620 621 622 623 624 625 626 627 628 629
static bool netlink_dump_space(struct netlink_sock *nlk)
{
	struct netlink_ring *ring = &nlk->rx_ring;
	struct nl_mmap_hdr *hdr;
	unsigned int n;

	hdr = netlink_current_frame(ring, NL_MMAP_STATUS_UNUSED);
	if (hdr == NULL)
		return false;

	n = ring->head + ring->frame_max / 2;
	if (n > ring->frame_max)
		n -= ring->frame_max;

	hdr = __netlink_lookup_frame(ring, n);

	return hdr->nm_status == NL_MMAP_STATUS_UNUSED;
}

630 631 632 633 634 635
static unsigned int netlink_poll(struct file *file, struct socket *sock,
				 poll_table *wait)
{
	struct sock *sk = sock->sk;
	struct netlink_sock *nlk = nlk_sk(sk);
	unsigned int mask;
636
	int err;
637

638 639 640 641 642
	if (nlk->rx_ring.pg_vec != NULL) {
		/* Memory mapped sockets don't call recvmsg(), so flow control
		 * for dumps is performed here. A dump is allowed to continue
		 * if at least half the ring is unused.
		 */
643
		while (nlk->cb_running && netlink_dump_space(nlk)) {
644 645
			err = netlink_dump(sk);
			if (err < 0) {
646
				sk->sk_err = -err;
647 648 649 650 651 652
				sk->sk_error_report(sk);
				break;
			}
		}
		netlink_rcv_wake(sk);
	}
653

654 655 656 657 658 659 660 661 662 663 664 665 666 667 668 669 670 671 672 673 674 675 676 677 678 679 680 681 682 683 684 685 686 687 688 689 690 691 692 693 694 695 696 697 698
	mask = datagram_poll(file, sock, wait);

	spin_lock_bh(&sk->sk_receive_queue.lock);
	if (nlk->rx_ring.pg_vec) {
		netlink_forward_ring(&nlk->rx_ring);
		if (!netlink_previous_frame(&nlk->rx_ring, NL_MMAP_STATUS_UNUSED))
			mask |= POLLIN | POLLRDNORM;
	}
	spin_unlock_bh(&sk->sk_receive_queue.lock);

	spin_lock_bh(&sk->sk_write_queue.lock);
	if (nlk->tx_ring.pg_vec) {
		if (netlink_current_frame(&nlk->tx_ring, NL_MMAP_STATUS_UNUSED))
			mask |= POLLOUT | POLLWRNORM;
	}
	spin_unlock_bh(&sk->sk_write_queue.lock);

	return mask;
}

static struct nl_mmap_hdr *netlink_mmap_hdr(struct sk_buff *skb)
{
	return (struct nl_mmap_hdr *)(skb->head - NL_MMAP_HDRLEN);
}

static void netlink_ring_setup_skb(struct sk_buff *skb, struct sock *sk,
				   struct netlink_ring *ring,
				   struct nl_mmap_hdr *hdr)
{
	unsigned int size;
	void *data;

	size = ring->frame_size - NL_MMAP_HDRLEN;
	data = (void *)hdr + NL_MMAP_HDRLEN;

	skb->head	= data;
	skb->data	= data;
	skb_reset_tail_pointer(skb);
	skb->end	= skb->tail + size;
	skb->len	= 0;

	skb->destructor	= netlink_skb_destructor;
	NETLINK_CB(skb).flags |= NETLINK_SKB_MMAPED;
	NETLINK_CB(skb).sk = sk;
}
699 700 701

static int netlink_mmap_sendmsg(struct sock *sk, struct msghdr *msg,
				u32 dst_portid, u32 dst_group,
C
Christoph Hellwig 已提交
702
				struct scm_cookie *scm)
703 704 705 706 707 708 709 710 711 712 713 714 715 716
{
	struct netlink_sock *nlk = nlk_sk(sk);
	struct netlink_ring *ring;
	struct nl_mmap_hdr *hdr;
	struct sk_buff *skb;
	unsigned int maxlen;
	int err = 0, len = 0;

	mutex_lock(&nlk->pg_vec_lock);

	ring   = &nlk->tx_ring;
	maxlen = ring->frame_size - NL_MMAP_HDRLEN;

	do {
D
David Miller 已提交
717 718
		unsigned int nm_len;

719 720 721 722 723 724 725
		hdr = netlink_current_frame(ring, NL_MMAP_STATUS_VALID);
		if (hdr == NULL) {
			if (!(msg->msg_flags & MSG_DONTWAIT) &&
			    atomic_read(&nlk->tx_ring.pending))
				schedule();
			continue;
		}
D
David Miller 已提交
726 727 728

		nm_len = ACCESS_ONCE(hdr->nm_len);
		if (nm_len > maxlen) {
729 730 731 732
			err = -EINVAL;
			goto out;
		}

D
David Miller 已提交
733
		netlink_frame_flush_dcache(hdr, nm_len);
734

D
David Miller 已提交
735 736 737 738
		skb = alloc_skb(nm_len, GFP_KERNEL);
		if (skb == NULL) {
			err = -ENOBUFS;
			goto out;
739
		}
D
David Miller 已提交
740 741 742
		__skb_put(skb, nm_len);
		memcpy(skb->data, (void *)hdr + NL_MMAP_HDRLEN, nm_len);
		netlink_set_status(hdr, NL_MMAP_STATUS_UNUSED);
743 744 745 746 747

		netlink_increment_head(ring);

		NETLINK_CB(skb).portid	  = nlk->portid;
		NETLINK_CB(skb).dst_group = dst_group;
C
Christoph Hellwig 已提交
748
		NETLINK_CB(skb).creds	  = scm->creds;
749 750 751 752 753 754 755 756 757 758 759 760 761 762 763 764 765 766 767 768 769 770 771 772 773 774 775 776

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

		if (unlikely(dst_group)) {
			atomic_inc(&skb->users);
			netlink_broadcast(sk, skb, dst_portid, dst_group,
					  GFP_KERNEL);
		}
		err = netlink_unicast(sk, skb, dst_portid,
				      msg->msg_flags & MSG_DONTWAIT);
		if (err < 0)
			goto out;
		len += err;

	} while (hdr != NULL ||
		 (!(msg->msg_flags & MSG_DONTWAIT) &&
		  atomic_read(&nlk->tx_ring.pending)));

	if (len > 0)
		err = len;
out:
	mutex_unlock(&nlk->pg_vec_lock);
	return err;
}
777 778 779 780 781 782 783 784 785

static void netlink_queue_mmaped_skb(struct sock *sk, struct sk_buff *skb)
{
	struct nl_mmap_hdr *hdr;

	hdr = netlink_mmap_hdr(skb);
	hdr->nm_len	= skb->len;
	hdr->nm_group	= NETLINK_CB(skb).dst_group;
	hdr->nm_pid	= NETLINK_CB(skb).creds.pid;
786 787
	hdr->nm_uid	= from_kuid(sk_user_ns(sk), NETLINK_CB(skb).creds.uid);
	hdr->nm_gid	= from_kgid(sk_user_ns(sk), NETLINK_CB(skb).creds.gid);
D
David Miller 已提交
788
	netlink_frame_flush_dcache(hdr, hdr->nm_len);
789 790 791 792 793 794 795 796 797 798 799 800 801 802 803 804 805
	netlink_set_status(hdr, NL_MMAP_STATUS_VALID);

	NETLINK_CB(skb).flags |= NETLINK_SKB_DELIVERED;
	kfree_skb(skb);
}

static void netlink_ring_set_copied(struct sock *sk, struct sk_buff *skb)
{
	struct netlink_sock *nlk = nlk_sk(sk);
	struct netlink_ring *ring = &nlk->rx_ring;
	struct nl_mmap_hdr *hdr;

	spin_lock_bh(&sk->sk_receive_queue.lock);
	hdr = netlink_current_frame(ring, NL_MMAP_STATUS_UNUSED);
	if (hdr == NULL) {
		spin_unlock_bh(&sk->sk_receive_queue.lock);
		kfree_skb(skb);
806
		netlink_overrun(sk);
807 808 809 810 811 812 813 814 815
		return;
	}
	netlink_increment_head(ring);
	__skb_queue_tail(&sk->sk_receive_queue, skb);
	spin_unlock_bh(&sk->sk_receive_queue.lock);

	hdr->nm_len	= skb->len;
	hdr->nm_group	= NETLINK_CB(skb).dst_group;
	hdr->nm_pid	= NETLINK_CB(skb).creds.pid;
816 817
	hdr->nm_uid	= from_kuid(sk_user_ns(sk), NETLINK_CB(skb).creds.uid);
	hdr->nm_gid	= from_kgid(sk_user_ns(sk), NETLINK_CB(skb).creds.gid);
818 819 820
	netlink_set_status(hdr, NL_MMAP_STATUS_COPY);
}

821
#else /* CONFIG_NETLINK_MMAP */
822
#define netlink_skb_is_mmaped(skb)	false
823
#define netlink_rx_is_mmaped(sk)	false
824
#define netlink_tx_is_mmaped(sk)	false
825
#define netlink_mmap			sock_no_mmap
826
#define netlink_poll			datagram_poll
C
Christoph Hellwig 已提交
827
#define netlink_mmap_sendmsg(sk, msg, dst_portid, dst_group, scm)	0
828 829
#endif /* CONFIG_NETLINK_MMAP */

830 831
static void netlink_skb_destructor(struct sk_buff *skb)
{
832 833 834 835 836 837 838 839 840 841 842 843 844 845
#ifdef CONFIG_NETLINK_MMAP
	struct nl_mmap_hdr *hdr;
	struct netlink_ring *ring;
	struct sock *sk;

	/* If a packet from the kernel to userspace was freed because of an
	 * error without being delivered to userspace, the kernel must reset
	 * the status. In the direction userspace to kernel, the status is
	 * always reset here after the packet was processed and freed.
	 */
	if (netlink_skb_is_mmaped(skb)) {
		hdr = netlink_mmap_hdr(skb);
		sk = NETLINK_CB(skb).sk;

846 847 848 849 850 851 852 853 854
		if (NETLINK_CB(skb).flags & NETLINK_SKB_TX) {
			netlink_set_status(hdr, NL_MMAP_STATUS_UNUSED);
			ring = &nlk_sk(sk)->tx_ring;
		} else {
			if (!(NETLINK_CB(skb).flags & NETLINK_SKB_DELIVERED)) {
				hdr->nm_len = 0;
				netlink_set_status(hdr, NL_MMAP_STATUS_VALID);
			}
			ring = &nlk_sk(sk)->rx_ring;
855 856 857 858 859 860
		}

		WARN_ON(atomic_read(&ring->pending) == 0);
		atomic_dec(&ring->pending);
		sock_put(sk);

861
		skb->head = NULL;
862 863
	}
#endif
864
	if (is_vmalloc_addr(skb->head)) {
865 866 867 868
		if (!skb->cloned ||
		    !atomic_dec_return(&(skb_shinfo(skb)->dataref)))
			vfree(skb->head);

869 870
		skb->head = NULL;
	}
871 872
	if (skb->sk != NULL)
		sock_rfree(skb);
873 874 875 876 877 878 879 880 881 882 883
}

static void netlink_skb_set_owner_r(struct sk_buff *skb, struct sock *sk)
{
	WARN_ON(skb->sk != NULL);
	skb->sk = sk;
	skb->destructor = netlink_skb_destructor;
	atomic_add(skb->truesize, &sk->sk_rmem_alloc);
	sk_mem_charge(sk, skb->truesize);
}

L
Linus Torvalds 已提交
884 885
static void netlink_sock_destruct(struct sock *sk)
{
886 887
	struct netlink_sock *nlk = nlk_sk(sk);

888 889 890
	if (nlk->cb_running) {
		if (nlk->cb.done)
			nlk->cb.done(&nlk->cb);
891

892 893
		module_put(nlk->cb.module);
		kfree_skb(nlk->cb.skb);
894 895
	}

L
Linus Torvalds 已提交
896
	skb_queue_purge(&sk->sk_receive_queue);
897 898 899 900 901 902 903 904 905 906 907 908
#ifdef CONFIG_NETLINK_MMAP
	if (1) {
		struct nl_mmap_req req;

		memset(&req, 0, sizeof(req));
		if (nlk->rx_ring.pg_vec)
			netlink_set_ring(sk, &req, true, false);
		memset(&req, 0, sizeof(req));
		if (nlk->tx_ring.pg_vec)
			netlink_set_ring(sk, &req, true, true);
	}
#endif /* CONFIG_NETLINK_MMAP */
L
Linus Torvalds 已提交
909 910

	if (!sock_flag(sk, SOCK_DEAD)) {
911
		printk(KERN_ERR "Freeing alive netlink socket %p\n", sk);
L
Linus Torvalds 已提交
912 913
		return;
	}
914 915 916 917

	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
918 919
}

920 921
/* 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
922 923 924 925
 * immediately hit write lock and grab all the cpus. Exclusive sleep solves
 * this, _but_ remember, it adds useless work on UP machines.
 */

926
void netlink_table_grab(void)
927
	__acquires(nl_table_lock)
L
Linus Torvalds 已提交
928
{
929 930
	might_sleep();

931
	write_lock_irq(&nl_table_lock);
L
Linus Torvalds 已提交
932 933 934 935 936

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

		add_wait_queue_exclusive(&nl_table_wait, &wait);
937
		for (;;) {
L
Linus Torvalds 已提交
938 939 940
			set_current_state(TASK_UNINTERRUPTIBLE);
			if (atomic_read(&nl_table_users) == 0)
				break;
941
			write_unlock_irq(&nl_table_lock);
L
Linus Torvalds 已提交
942
			schedule();
943
			write_lock_irq(&nl_table_lock);
L
Linus Torvalds 已提交
944 945 946 947 948 949 950
		}

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

951
void netlink_table_ungrab(void)
952
	__releases(nl_table_lock)
L
Linus Torvalds 已提交
953
{
954
	write_unlock_irq(&nl_table_lock);
L
Linus Torvalds 已提交
955 956 957
	wake_up(&nl_table_wait);
}

958
static inline void
L
Linus Torvalds 已提交
959 960 961 962 963 964 965 966 967
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);
}

968
static inline void
L
Linus Torvalds 已提交
969 970 971 972 973 974
netlink_unlock_table(void)
{
	if (atomic_dec_and_test(&nl_table_users))
		wake_up(&nl_table_wait);
}

975
struct netlink_compare_arg
L
Linus Torvalds 已提交
976
{
977
	possible_net_t pnet;
978 979
	u32 portid;
};
L
Linus Torvalds 已提交
980

981 982 983
/* Doing sizeof directly may yield 4 extra bytes on 64-bit. */
#define netlink_compare_arg_len \
	(offsetof(struct netlink_compare_arg, portid) + sizeof(u32))
984 985 986

static inline int netlink_compare(struct rhashtable_compare_arg *arg,
				  const void *ptr)
L
Linus Torvalds 已提交
987
{
988 989
	const struct netlink_compare_arg *x = arg->key;
	const struct netlink_sock *nlk = ptr;
L
Linus Torvalds 已提交
990

991 992 993 994 995 996 997 998 999 1000
	return nlk->portid != x->portid ||
	       !net_eq(sock_net(&nlk->sk), read_pnet(&x->pnet));
}

static void netlink_compare_arg_init(struct netlink_compare_arg *arg,
				     struct net *net, u32 portid)
{
	memset(arg, 0, sizeof(*arg));
	write_pnet(&arg->pnet, net);
	arg->portid = portid;
L
Linus Torvalds 已提交
1001 1002
}

1003 1004
static struct sock *__netlink_lookup(struct netlink_table *table, u32 portid,
				     struct net *net)
L
Linus Torvalds 已提交
1005
{
1006
	struct netlink_compare_arg arg;
L
Linus Torvalds 已提交
1007

1008 1009 1010
	netlink_compare_arg_init(&arg, net, portid);
	return rhashtable_lookup_fast(&table->hash, &arg,
				      netlink_rhashtable_params);
L
Linus Torvalds 已提交
1011 1012
}

1013
static int __netlink_insert(struct netlink_table *table, struct sock *sk)
Y
Ying Xue 已提交
1014
{
1015
	struct netlink_compare_arg arg;
Y
Ying Xue 已提交
1016

1017 1018 1019 1020
	netlink_compare_arg_init(&arg, sock_net(sk), nlk_sk(sk)->portid);
	return rhashtable_lookup_insert_key(&table->hash, &arg,
					    &nlk_sk(sk)->node,
					    netlink_rhashtable_params);
Y
Ying Xue 已提交
1021 1022
}

1023
static struct sock *netlink_lookup(struct net *net, int protocol, u32 portid)
L
Linus Torvalds 已提交
1024
{
1025 1026
	struct netlink_table *table = &nl_table[protocol];
	struct sock *sk;
L
Linus Torvalds 已提交
1027

1028 1029 1030 1031 1032
	rcu_read_lock();
	sk = __netlink_lookup(table, portid, net);
	if (sk)
		sock_hold(sk);
	rcu_read_unlock();
L
Linus Torvalds 已提交
1033

1034
	return sk;
L
Linus Torvalds 已提交
1035 1036
}

1037
static const struct proto_ops netlink_ops;
L
Linus Torvalds 已提交
1038

1039 1040 1041 1042 1043 1044
static void
netlink_update_listeners(struct sock *sk)
{
	struct netlink_table *tbl = &nl_table[sk->sk_protocol];
	unsigned long mask;
	unsigned int i;
1045 1046 1047 1048 1049
	struct listeners *listeners;

	listeners = nl_deref_protected(tbl->listeners);
	if (!listeners)
		return;
1050

1051
	for (i = 0; i < NLGRPLONGS(tbl->groups); i++) {
1052
		mask = 0;
1053
		sk_for_each_bound(sk, &tbl->mc_list) {
1054 1055 1056
			if (i < NLGRPLONGS(nlk_sk(sk)->ngroups))
				mask |= nlk_sk(sk)->groups[i];
		}
1057
		listeners->masks[i] = mask;
1058 1059 1060 1061 1062
	}
	/* this function is only called with the netlink table "grabbed", which
	 * makes sure updates are visible before bind or setsockopt return. */
}

1063
static int netlink_insert(struct sock *sk, u32 portid)
L
Linus Torvalds 已提交
1064
{
1065
	struct netlink_table *table = &nl_table[sk->sk_protocol];
1066
	int err;
L
Linus Torvalds 已提交
1067

Y
Ying Xue 已提交
1068
	lock_sock(sk);
L
Linus Torvalds 已提交
1069 1070

	err = -EBUSY;
1071
	if (nlk_sk(sk)->portid)
L
Linus Torvalds 已提交
1072 1073 1074
		goto err;

	err = -ENOMEM;
1075 1076
	if (BITS_PER_LONG > 32 &&
	    unlikely(atomic_read(&table->hash.nelems) >= UINT_MAX))
L
Linus Torvalds 已提交
1077 1078
		goto err;

1079
	nlk_sk(sk)->portid = portid;
1080
	sock_hold(sk);
1081

1082 1083 1084 1085
	err = __netlink_insert(table, sk);
	if (err) {
		if (err == -EEXIST)
			err = -EADDRINUSE;
Y
Ying Xue 已提交
1086
		sock_put(sk);
1087 1088
	}

L
Linus Torvalds 已提交
1089
err:
Y
Ying Xue 已提交
1090
	release_sock(sk);
L
Linus Torvalds 已提交
1091 1092 1093 1094 1095
	return err;
}

static void netlink_remove(struct sock *sk)
{
1096 1097 1098
	struct netlink_table *table;

	table = &nl_table[sk->sk_protocol];
1099 1100
	if (!rhashtable_remove_fast(&table->hash, &nlk_sk(sk)->node,
				    netlink_rhashtable_params)) {
1101 1102 1103 1104
		WARN_ON(atomic_read(&sk->sk_refcnt) == 1);
		__sock_put(sk);
	}

L
Linus Torvalds 已提交
1105
	netlink_table_grab();
1106
	if (nlk_sk(sk)->subscriptions) {
L
Linus Torvalds 已提交
1107
		__sk_del_bind_node(sk);
1108 1109
		netlink_update_listeners(sk);
	}
1110 1111
	if (sk->sk_protocol == NETLINK_GENERIC)
		atomic_inc(&genl_sk_destructing_cnt);
L
Linus Torvalds 已提交
1112 1113 1114 1115 1116 1117 1118 1119 1120
	netlink_table_ungrab();
}

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

1121
static int __netlink_create(struct net *net, struct socket *sock,
1122 1123
			    struct mutex *cb_mutex, int protocol,
			    int kern)
L
Linus Torvalds 已提交
1124 1125 1126
{
	struct sock *sk;
	struct netlink_sock *nlk;
1127 1128 1129

	sock->ops = &netlink_ops;

1130
	sk = sk_alloc(net, PF_NETLINK, GFP_KERNEL, &netlink_proto, kern);
1131 1132 1133 1134 1135 1136
	if (!sk)
		return -ENOMEM;

	sock_init_data(sock, sk);

	nlk = nlk_sk(sk);
E
Eric Dumazet 已提交
1137
	if (cb_mutex) {
1138
		nlk->cb_mutex = cb_mutex;
E
Eric Dumazet 已提交
1139
	} else {
1140 1141 1142
		nlk->cb_mutex = &nlk->cb_def_mutex;
		mutex_init(nlk->cb_mutex);
	}
1143
	init_waitqueue_head(&nlk->wait);
1144 1145 1146
#ifdef CONFIG_NETLINK_MMAP
	mutex_init(&nlk->pg_vec_lock);
#endif
1147 1148 1149 1150 1151 1152

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

1153 1154
static int netlink_create(struct net *net, struct socket *sock, int protocol,
			  int kern)
1155 1156
{
	struct module *module = NULL;
1157
	struct mutex *cb_mutex;
1158
	struct netlink_sock *nlk;
1159 1160
	int (*bind)(struct net *net, int group);
	void (*unbind)(struct net *net, int group);
1161
	int err = 0;
L
Linus Torvalds 已提交
1162 1163 1164 1165 1166 1167

	sock->state = SS_UNCONNECTED;

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

1168
	if (protocol < 0 || protocol >= MAX_LINKS)
L
Linus Torvalds 已提交
1169 1170
		return -EPROTONOSUPPORT;

1171
	netlink_lock_table();
1172
#ifdef CONFIG_MODULES
1173
	if (!nl_table[protocol].registered) {
1174
		netlink_unlock_table();
1175
		request_module("net-pf-%d-proto-%d", PF_NETLINK, protocol);
1176
		netlink_lock_table();
1177
	}
1178 1179 1180 1181
#endif
	if (nl_table[protocol].registered &&
	    try_module_get(nl_table[protocol].module))
		module = nl_table[protocol].module;
1182 1183
	else
		err = -EPROTONOSUPPORT;
1184
	cb_mutex = nl_table[protocol].cb_mutex;
1185
	bind = nl_table[protocol].bind;
1186
	unbind = nl_table[protocol].unbind;
1187
	netlink_unlock_table();
1188

1189 1190 1191
	if (err < 0)
		goto out;

1192
	err = __netlink_create(net, sock, cb_mutex, protocol, kern);
1193
	if (err < 0)
1194 1195
		goto out_module;

1196
	local_bh_disable();
1197
	sock_prot_inuse_add(net, &netlink_proto, 1);
1198 1199
	local_bh_enable();

1200 1201
	nlk = nlk_sk(sock->sk);
	nlk->module = module;
1202
	nlk->netlink_bind = bind;
1203
	nlk->netlink_unbind = unbind;
1204 1205
out:
	return err;
L
Linus Torvalds 已提交
1206

1207 1208 1209
out_module:
	module_put(module);
	goto out;
L
Linus Torvalds 已提交
1210 1211
}

1212 1213 1214 1215 1216 1217 1218
static void deferred_put_nlk_sk(struct rcu_head *head)
{
	struct netlink_sock *nlk = container_of(head, struct netlink_sock, rcu);

	sock_put(&nlk->sk);
}

L
Linus Torvalds 已提交
1219 1220 1221 1222 1223 1224 1225 1226 1227
static int netlink_release(struct socket *sock)
{
	struct sock *sk = sock->sk;
	struct netlink_sock *nlk;

	if (!sk)
		return 0;

	netlink_remove(sk);
1228
	sock_orphan(sk);
L
Linus Torvalds 已提交
1229 1230
	nlk = nlk_sk(sk);

1231 1232 1233 1234
	/*
	 * OK. Socket is unlinked, any packets that arrive now
	 * will be purged.
	 */
L
Linus Torvalds 已提交
1235

1236 1237 1238 1239 1240 1241 1242 1243 1244 1245 1246 1247 1248 1249
	/* must not acquire netlink_table_lock in any way again before unbind
	 * and notifying genetlink is done as otherwise it might deadlock
	 */
	if (nlk->netlink_unbind) {
		int i;

		for (i = 0; i < nlk->ngroups; i++)
			if (test_bit(i, nlk->groups))
				nlk->netlink_unbind(sock_net(sk), i + 1);
	}
	if (sk->sk_protocol == NETLINK_GENERIC &&
	    atomic_dec_return(&genl_sk_destructing_cnt) == 0)
		wake_up(&genl_sk_destructing_waitq);

L
Linus Torvalds 已提交
1250 1251 1252 1253 1254
	sock->sk = NULL;
	wake_up_interruptible_all(&nlk->wait);

	skb_queue_purge(&sk->sk_write_queue);

1255
	if (nlk->portid) {
L
Linus Torvalds 已提交
1256
		struct netlink_notify n = {
1257
						.net = sock_net(sk),
L
Linus Torvalds 已提交
1258
						.protocol = sk->sk_protocol,
1259
						.portid = nlk->portid,
L
Linus Torvalds 已提交
1260
					  };
1261 1262
		atomic_notifier_call_chain(&netlink_chain,
				NETLINK_URELEASE, &n);
1263
	}
1264

1265
	module_put(nlk->module);
1266

1267
	if (netlink_is_kernel(sk)) {
1268
		netlink_table_grab();
1269 1270
		BUG_ON(nl_table[sk->sk_protocol].registered == 0);
		if (--nl_table[sk->sk_protocol].registered == 0) {
1271 1272 1273 1274 1275
			struct listeners *old;

			old = nl_deref_protected(nl_table[sk->sk_protocol].listeners);
			RCU_INIT_POINTER(nl_table[sk->sk_protocol].listeners, NULL);
			kfree_rcu(old, rcu);
1276
			nl_table[sk->sk_protocol].module = NULL;
1277
			nl_table[sk->sk_protocol].bind = NULL;
1278
			nl_table[sk->sk_protocol].unbind = NULL;
1279
			nl_table[sk->sk_protocol].flags = 0;
1280 1281
			nl_table[sk->sk_protocol].registered = 0;
		}
1282
		netlink_table_ungrab();
E
Eric Dumazet 已提交
1283
	}
1284

1285 1286 1287
	kfree(nlk->groups);
	nlk->groups = NULL;

1288
	local_bh_disable();
1289
	sock_prot_inuse_add(sock_net(sk), &netlink_proto, -1);
1290
	local_bh_enable();
1291
	call_rcu(&nlk->rcu, deferred_put_nlk_sk);
L
Linus Torvalds 已提交
1292 1293 1294 1295 1296 1297
	return 0;
}

static int netlink_autobind(struct socket *sock)
{
	struct sock *sk = sock->sk;
1298
	struct net *net = sock_net(sk);
1299
	struct netlink_table *table = &nl_table[sk->sk_protocol];
1300
	s32 portid = task_tgid_vnr(current);
L
Linus Torvalds 已提交
1301 1302 1303 1304 1305
	int err;
	static s32 rover = -4097;

retry:
	cond_resched();
1306 1307 1308 1309 1310 1311 1312 1313
	rcu_read_lock();
	if (__netlink_lookup(table, portid, net)) {
		/* Bind collision, search negative portid values. */
		portid = rover--;
		if (rover > -4097)
			rover = -4097;
		rcu_read_unlock();
		goto retry;
L
Linus Torvalds 已提交
1314
	}
1315
	rcu_read_unlock();
L
Linus Torvalds 已提交
1316

1317
	err = netlink_insert(sk, portid);
L
Linus Torvalds 已提交
1318 1319
	if (err == -EADDRINUSE)
		goto retry;
1320 1321 1322 1323 1324 1325

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

	return err;
L
Linus Torvalds 已提交
1326 1327
}

1328 1329 1330 1331 1332 1333 1334 1335 1336 1337 1338 1339 1340
/**
 * __netlink_ns_capable - General netlink message capability test
 * @nsp: NETLINK_CB of the socket buffer holding a netlink command from userspace.
 * @user_ns: The user namespace of the capability to use
 * @cap: The capability to use
 *
 * Test to see if the opener of the socket we received the message
 * from had when the netlink socket was created and the sender of the
 * message has has the capability @cap in the user namespace @user_ns.
 */
bool __netlink_ns_capable(const struct netlink_skb_parms *nsp,
			struct user_namespace *user_ns, int cap)
{
1341 1342 1343
	return ((nsp->flags & NETLINK_SKB_DST) ||
		file_ns_capable(nsp->sk->sk_socket->file, user_ns, cap)) &&
		ns_capable(user_ns, cap);
1344 1345 1346 1347 1348 1349 1350 1351 1352 1353 1354 1355 1356 1357 1358 1359 1360 1361 1362 1363 1364 1365 1366 1367 1368 1369 1370 1371 1372 1373 1374 1375 1376 1377 1378 1379 1380 1381 1382 1383 1384 1385 1386 1387 1388 1389 1390 1391 1392 1393 1394
}
EXPORT_SYMBOL(__netlink_ns_capable);

/**
 * netlink_ns_capable - General netlink message capability test
 * @skb: socket buffer holding a netlink command from userspace
 * @user_ns: The user namespace of the capability to use
 * @cap: The capability to use
 *
 * Test to see if the opener of the socket we received the message
 * from had when the netlink socket was created and the sender of the
 * message has has the capability @cap in the user namespace @user_ns.
 */
bool netlink_ns_capable(const struct sk_buff *skb,
			struct user_namespace *user_ns, int cap)
{
	return __netlink_ns_capable(&NETLINK_CB(skb), user_ns, cap);
}
EXPORT_SYMBOL(netlink_ns_capable);

/**
 * netlink_capable - Netlink global message capability test
 * @skb: socket buffer holding a netlink command from userspace
 * @cap: The capability to use
 *
 * Test to see if the opener of the socket we received the message
 * from had when the netlink socket was created and the sender of the
 * message has has the capability @cap in all user namespaces.
 */
bool netlink_capable(const struct sk_buff *skb, int cap)
{
	return netlink_ns_capable(skb, &init_user_ns, cap);
}
EXPORT_SYMBOL(netlink_capable);

/**
 * netlink_net_capable - Netlink network namespace message capability test
 * @skb: socket buffer holding a netlink command from userspace
 * @cap: The capability to use
 *
 * Test to see if the opener of the socket we received the message
 * from had when the netlink socket was created and the sender of the
 * message has has the capability @cap over the network namespace of
 * the socket we received the message from.
 */
bool netlink_net_capable(const struct sk_buff *skb, int cap)
{
	return netlink_ns_capable(skb, sock_net(skb->sk)->user_ns, cap);
}
EXPORT_SYMBOL(netlink_net_capable);

1395
static inline int netlink_allowed(const struct socket *sock, unsigned int flag)
1396
{
1397
	return (nl_table[sock->sk->sk_protocol].flags & flag) ||
1398
		ns_capable(sock_net(sock->sk)->user_ns, CAP_NET_ADMIN);
1399
}
L
Linus Torvalds 已提交
1400

1401 1402 1403 1404 1405 1406 1407 1408 1409 1410 1411 1412
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;
}

1413
static int netlink_realloc_groups(struct sock *sk)
1414 1415 1416
{
	struct netlink_sock *nlk = nlk_sk(sk);
	unsigned int groups;
1417
	unsigned long *new_groups;
1418 1419
	int err = 0;

1420 1421
	netlink_table_grab();

1422
	groups = nl_table[sk->sk_protocol].groups;
1423
	if (!nl_table[sk->sk_protocol].registered) {
1424
		err = -ENOENT;
1425 1426
		goto out_unlock;
	}
1427

1428 1429
	if (nlk->ngroups >= groups)
		goto out_unlock;
1430

1431 1432 1433 1434 1435
	new_groups = krealloc(nlk->groups, NLGRPSZ(groups), GFP_ATOMIC);
	if (new_groups == NULL) {
		err = -ENOMEM;
		goto out_unlock;
	}
1436
	memset((char *)new_groups + NLGRPSZ(nlk->ngroups), 0,
1437 1438 1439
	       NLGRPSZ(groups) - NLGRPSZ(nlk->ngroups));

	nlk->groups = new_groups;
1440
	nlk->ngroups = groups;
1441 1442 1443
 out_unlock:
	netlink_table_ungrab();
	return err;
1444 1445
}

1446
static void netlink_undo_bind(int group, long unsigned int groups,
1447
			      struct sock *sk)
1448
{
1449
	struct netlink_sock *nlk = nlk_sk(sk);
1450 1451 1452 1453 1454 1455
	int undo;

	if (!nlk->netlink_unbind)
		return;

	for (undo = 0; undo < group; undo++)
1456
		if (test_bit(undo, &groups))
1457
			nlk->netlink_unbind(sock_net(sk), undo + 1);
1458 1459
}

1460 1461
static int netlink_bind(struct socket *sock, struct sockaddr *addr,
			int addr_len)
L
Linus Torvalds 已提交
1462 1463
{
	struct sock *sk = sock->sk;
1464
	struct net *net = sock_net(sk);
L
Linus Torvalds 已提交
1465 1466 1467
	struct netlink_sock *nlk = nlk_sk(sk);
	struct sockaddr_nl *nladdr = (struct sockaddr_nl *)addr;
	int err;
1468
	long unsigned int groups = nladdr->nl_groups;
1469

1470 1471 1472
	if (addr_len < sizeof(struct sockaddr_nl))
		return -EINVAL;

L
Linus Torvalds 已提交
1473 1474 1475 1476
	if (nladdr->nl_family != AF_NETLINK)
		return -EINVAL;

	/* Only superuser is allowed to listen multicasts */
1477
	if (groups) {
1478
		if (!netlink_allowed(sock, NL_CFG_F_NONROOT_RECV))
1479
			return -EPERM;
1480 1481 1482
		err = netlink_realloc_groups(sk);
		if (err)
			return err;
1483
	}
L
Linus Torvalds 已提交
1484

1485
	if (nlk->portid)
1486
		if (nladdr->nl_pid != nlk->portid)
L
Linus Torvalds 已提交
1487
			return -EINVAL;
1488 1489 1490 1491 1492 1493 1494

	if (nlk->netlink_bind && groups) {
		int group;

		for (group = 0; group < nlk->ngroups; group++) {
			if (!test_bit(group, &groups))
				continue;
1495
			err = nlk->netlink_bind(net, group + 1);
1496 1497
			if (!err)
				continue;
1498
			netlink_undo_bind(group, groups, sk);
1499 1500 1501 1502 1503
			return err;
		}
	}

	if (!nlk->portid) {
L
Linus Torvalds 已提交
1504
		err = nladdr->nl_pid ?
1505
			netlink_insert(sk, nladdr->nl_pid) :
L
Linus Torvalds 已提交
1506
			netlink_autobind(sock);
1507
		if (err) {
1508
			netlink_undo_bind(nlk->ngroups, groups, sk);
L
Linus Torvalds 已提交
1509
			return err;
1510
		}
L
Linus Torvalds 已提交
1511 1512
	}

1513
	if (!groups && (nlk->groups == NULL || !(u32)nlk->groups[0]))
L
Linus Torvalds 已提交
1514 1515 1516
		return 0;

	netlink_table_grab();
1517
	netlink_update_subscriptions(sk, nlk->subscriptions +
1518
					 hweight32(groups) -
1519
					 hweight32(nlk->groups[0]));
1520
	nlk->groups[0] = (nlk->groups[0] & ~0xffffffffUL) | groups;
1521
	netlink_update_listeners(sk);
L
Linus Torvalds 已提交
1522 1523 1524 1525 1526 1527 1528 1529 1530 1531 1532
	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);
1533
	struct sockaddr_nl *nladdr = (struct sockaddr_nl *)addr;
L
Linus Torvalds 已提交
1534

1535 1536 1537
	if (alen < sizeof(addr->sa_family))
		return -EINVAL;

L
Linus Torvalds 已提交
1538 1539
	if (addr->sa_family == AF_UNSPEC) {
		sk->sk_state	= NETLINK_UNCONNECTED;
1540
		nlk->dst_portid	= 0;
1541
		nlk->dst_group  = 0;
L
Linus Torvalds 已提交
1542 1543 1544 1545 1546
		return 0;
	}
	if (addr->sa_family != AF_NETLINK)
		return -EINVAL;

1547
	if ((nladdr->nl_groups || nladdr->nl_pid) &&
1548
	    !netlink_allowed(sock, NL_CFG_F_NONROOT_SEND))
L
Linus Torvalds 已提交
1549 1550
		return -EPERM;

1551
	if (!nlk->portid)
L
Linus Torvalds 已提交
1552 1553 1554 1555
		err = netlink_autobind(sock);

	if (err == 0) {
		sk->sk_state	= NETLINK_CONNECTED;
1556
		nlk->dst_portid = nladdr->nl_pid;
1557
		nlk->dst_group  = ffs(nladdr->nl_groups);
L
Linus Torvalds 已提交
1558 1559 1560 1561 1562
	}

	return err;
}

1563 1564
static int netlink_getname(struct socket *sock, struct sockaddr *addr,
			   int *addr_len, int peer)
L
Linus Torvalds 已提交
1565 1566 1567
{
	struct sock *sk = sock->sk;
	struct netlink_sock *nlk = nlk_sk(sk);
1568
	DECLARE_SOCKADDR(struct sockaddr_nl *, nladdr, addr);
1569

L
Linus Torvalds 已提交
1570 1571 1572 1573 1574
	nladdr->nl_family = AF_NETLINK;
	nladdr->nl_pad = 0;
	*addr_len = sizeof(*nladdr);

	if (peer) {
1575
		nladdr->nl_pid = nlk->dst_portid;
1576
		nladdr->nl_groups = netlink_group_mask(nlk->dst_group);
L
Linus Torvalds 已提交
1577
	} else {
1578
		nladdr->nl_pid = nlk->portid;
1579
		nladdr->nl_groups = nlk->groups ? nlk->groups[0] : 0;
L
Linus Torvalds 已提交
1580 1581 1582 1583
	}
	return 0;
}

1584
static struct sock *netlink_getsockbyportid(struct sock *ssk, u32 portid)
L
Linus Torvalds 已提交
1585 1586 1587 1588
{
	struct sock *sock;
	struct netlink_sock *nlk;

1589
	sock = netlink_lookup(sock_net(ssk), ssk->sk_protocol, portid);
L
Linus Torvalds 已提交
1590 1591 1592 1593 1594
	if (!sock)
		return ERR_PTR(-ECONNREFUSED);

	/* Don't bother queuing skb if kernel socket has no input function */
	nlk = nlk_sk(sock);
1595
	if (sock->sk_state == NETLINK_CONNECTED &&
1596
	    nlk->dst_portid != nlk_sk(ssk)->portid) {
L
Linus Torvalds 已提交
1597 1598 1599 1600 1601 1602 1603 1604
		sock_put(sock);
		return ERR_PTR(-ECONNREFUSED);
	}
	return sock;
}

struct sock *netlink_getsockbyfilp(struct file *filp)
{
A
Al Viro 已提交
1605
	struct inode *inode = file_inode(filp);
L
Linus Torvalds 已提交
1606 1607 1608 1609 1610 1611 1612 1613 1614 1615 1616 1617 1618
	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;
}

1619 1620
static struct sk_buff *netlink_alloc_large_skb(unsigned int size,
					       int broadcast)
1621 1622 1623 1624
{
	struct sk_buff *skb;
	void *data;

1625
	if (size <= NLMSG_GOODSIZE || broadcast)
1626 1627
		return alloc_skb(size, GFP_KERNEL);

1628 1629
	size = SKB_DATA_ALIGN(size) +
	       SKB_DATA_ALIGN(sizeof(struct skb_shared_info));
1630 1631 1632

	data = vmalloc(size);
	if (data == NULL)
1633
		return NULL;
1634

E
Eric Dumazet 已提交
1635
	skb = __build_skb(data, size);
1636 1637
	if (skb == NULL)
		vfree(data);
E
Eric Dumazet 已提交
1638
	else
1639
		skb->destructor = netlink_skb_destructor;
1640 1641 1642 1643

	return skb;
}

L
Linus Torvalds 已提交
1644 1645 1646 1647 1648 1649 1650 1651 1652 1653
/*
 * 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.
 */
1654
int netlink_attachskb(struct sock *sk, struct sk_buff *skb,
P
Patrick McHardy 已提交
1655
		      long *timeo, struct sock *ssk)
L
Linus Torvalds 已提交
1656 1657 1658 1659 1660
{
	struct netlink_sock *nlk;

	nlk = nlk_sk(sk);

1661
	if ((atomic_read(&sk->sk_rmem_alloc) > sk->sk_rcvbuf ||
1662
	     test_bit(NETLINK_S_CONGESTED, &nlk->state)) &&
1663
	    !netlink_skb_is_mmaped(skb)) {
L
Linus Torvalds 已提交
1664
		DECLARE_WAITQUEUE(wait, current);
P
Patrick McHardy 已提交
1665
		if (!*timeo) {
1666
			if (!ssk || netlink_is_kernel(ssk))
L
Linus Torvalds 已提交
1667 1668 1669 1670 1671 1672 1673 1674 1675 1676
				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 ||
1677
		     test_bit(NETLINK_S_CONGESTED, &nlk->state)) &&
L
Linus Torvalds 已提交
1678
		    !sock_flag(sk, SOCK_DEAD))
P
Patrick McHardy 已提交
1679
			*timeo = schedule_timeout(*timeo);
L
Linus Torvalds 已提交
1680 1681 1682 1683 1684 1685 1686

		__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
1687
			return sock_intr_errno(*timeo);
L
Linus Torvalds 已提交
1688 1689 1690
		}
		return 1;
	}
1691
	netlink_skb_set_owner_r(skb, sk);
L
Linus Torvalds 已提交
1692 1693 1694
	return 0;
}

1695
static int __netlink_sendskb(struct sock *sk, struct sk_buff *skb)
L
Linus Torvalds 已提交
1696 1697 1698
{
	int len = skb->len;

1699 1700
	netlink_deliver_tap(skb);

1701 1702 1703 1704 1705 1706 1707 1708
#ifdef CONFIG_NETLINK_MMAP
	if (netlink_skb_is_mmaped(skb))
		netlink_queue_mmaped_skb(sk, skb);
	else if (netlink_rx_is_mmaped(sk))
		netlink_ring_set_copied(sk, skb);
	else
#endif /* CONFIG_NETLINK_MMAP */
		skb_queue_tail(&sk->sk_receive_queue, skb);
1709
	sk->sk_data_ready(sk);
1710 1711 1712 1713 1714 1715 1716
	return len;
}

int netlink_sendskb(struct sock *sk, struct sk_buff *skb)
{
	int len = __netlink_sendskb(sk, skb);

L
Linus Torvalds 已提交
1717 1718 1719 1720 1721 1722 1723 1724 1725 1726
	sock_put(sk);
	return len;
}

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

1727
static struct sk_buff *netlink_trim(struct sk_buff *skb, gfp_t allocation)
L
Linus Torvalds 已提交
1728 1729 1730
{
	int delta;

1731
	WARN_ON(skb->sk != NULL);
1732 1733
	if (netlink_skb_is_mmaped(skb))
		return skb;
L
Linus Torvalds 已提交
1734

1735
	delta = skb->end - skb->tail;
1736
	if (is_vmalloc_addr(skb->head) || delta * 2 < skb->truesize)
L
Linus Torvalds 已提交
1737 1738 1739 1740 1741 1742
		return skb;

	if (skb_shared(skb)) {
		struct sk_buff *nskb = skb_clone(skb, allocation);
		if (!nskb)
			return skb;
1743
		consume_skb(skb);
L
Linus Torvalds 已提交
1744 1745 1746 1747 1748 1749 1750 1751 1752
		skb = nskb;
	}

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

	return skb;
}

1753 1754
static int netlink_unicast_kernel(struct sock *sk, struct sk_buff *skb,
				  struct sock *ssk)
1755 1756 1757 1758 1759 1760 1761
{
	int ret;
	struct netlink_sock *nlk = nlk_sk(sk);

	ret = -ECONNREFUSED;
	if (nlk->netlink_rcv != NULL) {
		ret = skb->len;
1762
		netlink_skb_set_owner_r(skb, sk);
1763
		NETLINK_CB(skb).sk = ssk;
1764
		netlink_deliver_tap_kernel(sk, ssk, skb);
1765
		nlk->netlink_rcv(skb);
1766 1767 1768
		consume_skb(skb);
	} else {
		kfree_skb(skb);
1769 1770 1771 1772 1773 1774
	}
	sock_put(sk);
	return ret;
}

int netlink_unicast(struct sock *ssk, struct sk_buff *skb,
1775
		    u32 portid, int nonblock)
L
Linus Torvalds 已提交
1776 1777 1778 1779 1780 1781 1782 1783 1784
{
	struct sock *sk;
	int err;
	long timeo;

	skb = netlink_trim(skb, gfp_any());

	timeo = sock_sndtimeo(ssk, nonblock);
retry:
1785
	sk = netlink_getsockbyportid(ssk, portid);
L
Linus Torvalds 已提交
1786 1787 1788 1789
	if (IS_ERR(sk)) {
		kfree_skb(skb);
		return PTR_ERR(sk);
	}
1790
	if (netlink_is_kernel(sk))
1791
		return netlink_unicast_kernel(sk, skb, ssk);
1792

1793
	if (sk_filter(sk, skb)) {
W
Wang Chen 已提交
1794
		err = skb->len;
1795 1796 1797 1798 1799
		kfree_skb(skb);
		sock_put(sk);
		return err;
	}

1800
	err = netlink_attachskb(sk, skb, &timeo, ssk);
L
Linus Torvalds 已提交
1801 1802 1803 1804 1805
	if (err == 1)
		goto retry;
	if (err)
		return err;

1806
	return netlink_sendskb(sk, skb);
L
Linus Torvalds 已提交
1807
}
1808
EXPORT_SYMBOL(netlink_unicast);
L
Linus Torvalds 已提交
1809

1810 1811 1812 1813 1814 1815 1816 1817 1818 1819 1820 1821 1822 1823 1824 1825 1826 1827 1828
struct sk_buff *netlink_alloc_skb(struct sock *ssk, unsigned int size,
				  u32 dst_portid, gfp_t gfp_mask)
{
#ifdef CONFIG_NETLINK_MMAP
	struct sock *sk = NULL;
	struct sk_buff *skb;
	struct netlink_ring *ring;
	struct nl_mmap_hdr *hdr;
	unsigned int maxlen;

	sk = netlink_getsockbyportid(ssk, dst_portid);
	if (IS_ERR(sk))
		goto out;

	ring = &nlk_sk(sk)->rx_ring;
	/* fast-path without atomic ops for common case: non-mmaped receiver */
	if (ring->pg_vec == NULL)
		goto out_put;

1829 1830 1831
	if (ring->frame_size - NL_MMAP_HDRLEN < size)
		goto out_put;

1832 1833 1834 1835 1836 1837 1838 1839 1840
	skb = alloc_skb_head(gfp_mask);
	if (skb == NULL)
		goto err1;

	spin_lock_bh(&sk->sk_receive_queue.lock);
	/* check again under lock */
	if (ring->pg_vec == NULL)
		goto out_free;

1841
	/* check again under lock */
1842 1843 1844 1845 1846 1847 1848 1849 1850 1851 1852 1853 1854 1855 1856 1857 1858 1859 1860
	maxlen = ring->frame_size - NL_MMAP_HDRLEN;
	if (maxlen < size)
		goto out_free;

	netlink_forward_ring(ring);
	hdr = netlink_current_frame(ring, NL_MMAP_STATUS_UNUSED);
	if (hdr == NULL)
		goto err2;
	netlink_ring_setup_skb(skb, sk, ring, hdr);
	netlink_set_status(hdr, NL_MMAP_STATUS_RESERVED);
	atomic_inc(&ring->pending);
	netlink_increment_head(ring);

	spin_unlock_bh(&sk->sk_receive_queue.lock);
	return skb;

err2:
	kfree_skb(skb);
	spin_unlock_bh(&sk->sk_receive_queue.lock);
1861
	netlink_overrun(sk);
1862 1863 1864 1865 1866 1867 1868 1869 1870 1871 1872 1873 1874 1875 1876
err1:
	sock_put(sk);
	return NULL;

out_free:
	kfree_skb(skb);
	spin_unlock_bh(&sk->sk_receive_queue.lock);
out_put:
	sock_put(sk);
out:
#endif
	return alloc_skb(size, gfp_mask);
}
EXPORT_SYMBOL_GPL(netlink_alloc_skb);

1877 1878 1879
int netlink_has_listeners(struct sock *sk, unsigned int group)
{
	int res = 0;
1880
	struct listeners *listeners;
1881

1882
	BUG_ON(!netlink_is_kernel(sk));
1883 1884 1885 1886

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

1887
	if (listeners && group - 1 < nl_table[sk->sk_protocol].groups)
1888
		res = test_bit(group - 1, listeners->masks);
1889 1890 1891

	rcu_read_unlock();

1892 1893 1894 1895
	return res;
}
EXPORT_SYMBOL_GPL(netlink_has_listeners);

1896
static int netlink_broadcast_deliver(struct sock *sk, struct sk_buff *skb)
L
Linus Torvalds 已提交
1897 1898 1899 1900
{
	struct netlink_sock *nlk = nlk_sk(sk);

	if (atomic_read(&sk->sk_rmem_alloc) <= sk->sk_rcvbuf &&
1901
	    !test_bit(NETLINK_S_CONGESTED, &nlk->state)) {
1902
		netlink_skb_set_owner_r(skb, sk);
1903
		__netlink_sendskb(sk, skb);
1904
		return atomic_read(&sk->sk_rmem_alloc) > (sk->sk_rcvbuf >> 1);
L
Linus Torvalds 已提交
1905 1906 1907 1908 1909 1910
	}
	return -1;
}

struct netlink_broadcast_data {
	struct sock *exclude_sk;
1911
	struct net *net;
1912
	u32 portid;
L
Linus Torvalds 已提交
1913 1914
	u32 group;
	int failure;
1915
	int delivery_failure;
L
Linus Torvalds 已提交
1916 1917
	int congested;
	int delivered;
A
Al Viro 已提交
1918
	gfp_t allocation;
L
Linus Torvalds 已提交
1919
	struct sk_buff *skb, *skb2;
1920 1921
	int (*tx_filter)(struct sock *dsk, struct sk_buff *skb, void *data);
	void *tx_data;
L
Linus Torvalds 已提交
1922 1923
};

1924 1925
static void do_one_broadcast(struct sock *sk,
				    struct netlink_broadcast_data *p)
L
Linus Torvalds 已提交
1926 1927 1928 1929 1930
{
	struct netlink_sock *nlk = nlk_sk(sk);
	int val;

	if (p->exclude_sk == sk)
1931
		return;
L
Linus Torvalds 已提交
1932

1933
	if (nlk->portid == p->portid || p->group - 1 >= nlk->ngroups ||
1934
	    !test_bit(p->group - 1, nlk->groups))
1935
		return;
L
Linus Torvalds 已提交
1936

1937 1938 1939 1940 1941 1942 1943 1944 1945 1946 1947
	if (!net_eq(sock_net(sk), p->net)) {
		if (!(nlk->flags & NETLINK_F_LISTEN_ALL_NSID))
			return;

		if (!peernet_has_id(sock_net(sk), p->net))
			return;

		if (!file_ns_capable(sk->sk_socket->file, p->net->user_ns,
				     CAP_NET_BROADCAST))
			return;
	}
1948

L
Linus Torvalds 已提交
1949 1950
	if (p->failure) {
		netlink_overrun(sk);
1951
		return;
L
Linus Torvalds 已提交
1952 1953 1954 1955
	}

	sock_hold(sk);
	if (p->skb2 == NULL) {
1956
		if (skb_shared(p->skb)) {
L
Linus Torvalds 已提交
1957 1958
			p->skb2 = skb_clone(p->skb, p->allocation);
		} else {
1959 1960 1961 1962 1963 1964
			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
1965 1966 1967 1968 1969 1970
		}
	}
	if (p->skb2 == NULL) {
		netlink_overrun(sk);
		/* Clone failed. Notify ALL listeners. */
		p->failure = 1;
1971
		if (nlk->flags & NETLINK_F_BROADCAST_SEND_ERROR)
1972
			p->delivery_failure = 1;
1973 1974 1975
		goto out;
	}
	if (p->tx_filter && p->tx_filter(sk, p->skb2, p->tx_data)) {
1976 1977
		kfree_skb(p->skb2);
		p->skb2 = NULL;
1978 1979 1980
		goto out;
	}
	if (sk_filter(sk, p->skb2)) {
1981 1982
		kfree_skb(p->skb2);
		p->skb2 = NULL;
1983 1984 1985 1986 1987 1988
		goto out;
	}
	NETLINK_CB(p->skb2).nsid = peernet2id(sock_net(sk), p->net);
	NETLINK_CB(p->skb2).nsid_is_set = true;
	val = netlink_broadcast_deliver(sk, p->skb2);
	if (val < 0) {
L
Linus Torvalds 已提交
1989
		netlink_overrun(sk);
1990
		if (nlk->flags & NETLINK_F_BROADCAST_SEND_ERROR)
1991
			p->delivery_failure = 1;
L
Linus Torvalds 已提交
1992 1993 1994 1995 1996
	} else {
		p->congested |= val;
		p->delivered = 1;
		p->skb2 = NULL;
	}
1997
out:
L
Linus Torvalds 已提交
1998 1999 2000
	sock_put(sk);
}

2001
int netlink_broadcast_filtered(struct sock *ssk, struct sk_buff *skb, u32 portid,
2002 2003 2004
	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
2005
{
2006
	struct net *net = sock_net(ssk);
L
Linus Torvalds 已提交
2007 2008 2009 2010 2011 2012
	struct netlink_broadcast_data info;
	struct sock *sk;

	skb = netlink_trim(skb, allocation);

	info.exclude_sk = ssk;
2013
	info.net = net;
2014
	info.portid = portid;
L
Linus Torvalds 已提交
2015 2016
	info.group = group;
	info.failure = 0;
2017
	info.delivery_failure = 0;
L
Linus Torvalds 已提交
2018 2019 2020 2021 2022
	info.congested = 0;
	info.delivered = 0;
	info.allocation = allocation;
	info.skb = skb;
	info.skb2 = NULL;
2023 2024
	info.tx_filter = filter;
	info.tx_data = filter_data;
L
Linus Torvalds 已提交
2025 2026 2027 2028 2029

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

	netlink_lock_table();

2030
	sk_for_each_bound(sk, &nl_table[ssk->sk_protocol].mc_list)
L
Linus Torvalds 已提交
2031 2032
		do_one_broadcast(sk, &info);

2033
	consume_skb(skb);
2034

L
Linus Torvalds 已提交
2035 2036
	netlink_unlock_table();

2037 2038
	if (info.delivery_failure) {
		kfree_skb(info.skb2);
2039
		return -ENOBUFS;
E
Eric Dumazet 已提交
2040 2041
	}
	consume_skb(info.skb2);
2042

L
Linus Torvalds 已提交
2043 2044 2045 2046 2047 2048 2049
	if (info.delivered) {
		if (info.congested && (allocation & __GFP_WAIT))
			yield();
		return 0;
	}
	return -ESRCH;
}
2050 2051
EXPORT_SYMBOL(netlink_broadcast_filtered);

2052
int netlink_broadcast(struct sock *ssk, struct sk_buff *skb, u32 portid,
2053 2054
		      u32 group, gfp_t allocation)
{
2055
	return netlink_broadcast_filtered(ssk, skb, portid, group, allocation,
2056 2057
		NULL, NULL);
}
2058
EXPORT_SYMBOL(netlink_broadcast);
L
Linus Torvalds 已提交
2059 2060 2061

struct netlink_set_err_data {
	struct sock *exclude_sk;
2062
	u32 portid;
L
Linus Torvalds 已提交
2063 2064 2065 2066
	u32 group;
	int code;
};

2067
static int do_one_set_err(struct sock *sk, struct netlink_set_err_data *p)
L
Linus Torvalds 已提交
2068 2069
{
	struct netlink_sock *nlk = nlk_sk(sk);
2070
	int ret = 0;
L
Linus Torvalds 已提交
2071 2072 2073 2074

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

O
Octavian Purdila 已提交
2075
	if (!net_eq(sock_net(sk), sock_net(p->exclude_sk)))
2076 2077
		goto out;

2078
	if (nlk->portid == p->portid || p->group - 1 >= nlk->ngroups ||
2079
	    !test_bit(p->group - 1, nlk->groups))
L
Linus Torvalds 已提交
2080 2081
		goto out;

2082
	if (p->code == ENOBUFS && nlk->flags & NETLINK_F_RECV_NO_ENOBUFS) {
2083 2084 2085 2086
		ret = 1;
		goto out;
	}

L
Linus Torvalds 已提交
2087 2088 2089
	sk->sk_err = p->code;
	sk->sk_error_report(sk);
out:
2090
	return ret;
L
Linus Torvalds 已提交
2091 2092
}

2093 2094 2095
/**
 * netlink_set_err - report error to broadcast listeners
 * @ssk: the kernel netlink socket, as returned by netlink_kernel_create()
2096
 * @portid: the PORTID of a process that we want to skip (if any)
2097
 * @group: the broadcast group that will notice the error
2098
 * @code: error code, must be negative (as usual in kernelspace)
2099 2100
 *
 * This function returns the number of broadcast listeners that have set the
2101
 * NETLINK_NO_ENOBUFS socket option.
2102
 */
2103
int netlink_set_err(struct sock *ssk, u32 portid, u32 group, int code)
L
Linus Torvalds 已提交
2104 2105 2106
{
	struct netlink_set_err_data info;
	struct sock *sk;
2107
	int ret = 0;
L
Linus Torvalds 已提交
2108 2109

	info.exclude_sk = ssk;
2110
	info.portid = portid;
L
Linus Torvalds 已提交
2111
	info.group = group;
2112 2113
	/* sk->sk_err wants a positive error value */
	info.code = -code;
L
Linus Torvalds 已提交
2114 2115 2116

	read_lock(&nl_table_lock);

2117
	sk_for_each_bound(sk, &nl_table[ssk->sk_protocol].mc_list)
2118
		ret += do_one_set_err(sk, &info);
L
Linus Torvalds 已提交
2119 2120

	read_unlock(&nl_table_lock);
2121
	return ret;
L
Linus Torvalds 已提交
2122
}
2123
EXPORT_SYMBOL(netlink_set_err);
L
Linus Torvalds 已提交
2124

2125 2126 2127 2128 2129 2130 2131 2132 2133 2134 2135 2136 2137 2138 2139 2140 2141
/* 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);
}

2142
static int netlink_setsockopt(struct socket *sock, int level, int optname,
2143
			      char __user *optval, unsigned int optlen)
2144 2145 2146
{
	struct sock *sk = sock->sk;
	struct netlink_sock *nlk = nlk_sk(sk);
2147 2148
	unsigned int val = 0;
	int err;
2149 2150 2151 2152

	if (level != SOL_NETLINK)
		return -ENOPROTOOPT;

2153 2154
	if (optname != NETLINK_RX_RING && optname != NETLINK_TX_RING &&
	    optlen >= sizeof(int) &&
2155
	    get_user(val, (unsigned int __user *)optval))
2156 2157 2158 2159 2160
		return -EFAULT;

	switch (optname) {
	case NETLINK_PKTINFO:
		if (val)
2161
			nlk->flags |= NETLINK_F_RECV_PKTINFO;
2162
		else
2163
			nlk->flags &= ~NETLINK_F_RECV_PKTINFO;
2164 2165 2166 2167
		err = 0;
		break;
	case NETLINK_ADD_MEMBERSHIP:
	case NETLINK_DROP_MEMBERSHIP: {
2168
		if (!netlink_allowed(sock, NL_CFG_F_NONROOT_RECV))
2169
			return -EPERM;
2170 2171 2172
		err = netlink_realloc_groups(sk);
		if (err)
			return err;
2173 2174
		if (!val || val - 1 >= nlk->ngroups)
			return -EINVAL;
2175
		if (optname == NETLINK_ADD_MEMBERSHIP && nlk->netlink_bind) {
2176
			err = nlk->netlink_bind(sock_net(sk), val);
2177 2178 2179
			if (err)
				return err;
		}
2180
		netlink_table_grab();
2181 2182
		netlink_update_socket_mc(nlk, val,
					 optname == NETLINK_ADD_MEMBERSHIP);
2183
		netlink_table_ungrab();
2184
		if (optname == NETLINK_DROP_MEMBERSHIP && nlk->netlink_unbind)
2185
			nlk->netlink_unbind(sock_net(sk), val);
2186

2187 2188 2189
		err = 0;
		break;
	}
2190 2191
	case NETLINK_BROADCAST_ERROR:
		if (val)
2192
			nlk->flags |= NETLINK_F_BROADCAST_SEND_ERROR;
2193
		else
2194
			nlk->flags &= ~NETLINK_F_BROADCAST_SEND_ERROR;
2195 2196
		err = 0;
		break;
2197 2198
	case NETLINK_NO_ENOBUFS:
		if (val) {
2199 2200
			nlk->flags |= NETLINK_F_RECV_NO_ENOBUFS;
			clear_bit(NETLINK_S_CONGESTED, &nlk->state);
2201
			wake_up_interruptible(&nlk->wait);
E
Eric Dumazet 已提交
2202
		} else {
2203
			nlk->flags &= ~NETLINK_F_RECV_NO_ENOBUFS;
E
Eric Dumazet 已提交
2204
		}
2205 2206
		err = 0;
		break;
2207 2208 2209 2210 2211 2212 2213 2214 2215 2216 2217 2218 2219 2220 2221 2222 2223 2224 2225
#ifdef CONFIG_NETLINK_MMAP
	case NETLINK_RX_RING:
	case NETLINK_TX_RING: {
		struct nl_mmap_req req;

		/* Rings might consume more memory than queue limits, require
		 * CAP_NET_ADMIN.
		 */
		if (!capable(CAP_NET_ADMIN))
			return -EPERM;
		if (optlen < sizeof(req))
			return -EINVAL;
		if (copy_from_user(&req, optval, sizeof(req)))
			return -EFAULT;
		err = netlink_set_ring(sk, &req, false,
				       optname == NETLINK_TX_RING);
		break;
	}
#endif /* CONFIG_NETLINK_MMAP */
2226 2227 2228 2229 2230 2231 2232 2233 2234 2235
	case NETLINK_LISTEN_ALL_NSID:
		if (!ns_capable(sock_net(sk)->user_ns, CAP_NET_BROADCAST))
			return -EPERM;

		if (val)
			nlk->flags |= NETLINK_F_LISTEN_ALL_NSID;
		else
			nlk->flags &= ~NETLINK_F_LISTEN_ALL_NSID;
		err = 0;
		break;
2236 2237 2238 2239 2240 2241 2242
	default:
		err = -ENOPROTOOPT;
	}
	return err;
}

static int netlink_getsockopt(struct socket *sock, int level, int optname,
2243
			      char __user *optval, int __user *optlen)
2244 2245 2246 2247 2248 2249 2250 2251 2252 2253 2254 2255 2256 2257 2258 2259 2260 2261
{
	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);
2262
		val = nlk->flags & NETLINK_F_RECV_PKTINFO ? 1 : 0;
H
Heiko Carstens 已提交
2263 2264 2265
		if (put_user(len, optlen) ||
		    put_user(val, optval))
			return -EFAULT;
2266 2267
		err = 0;
		break;
2268 2269 2270 2271
	case NETLINK_BROADCAST_ERROR:
		if (len < sizeof(int))
			return -EINVAL;
		len = sizeof(int);
2272
		val = nlk->flags & NETLINK_F_BROADCAST_SEND_ERROR ? 1 : 0;
2273 2274 2275 2276 2277
		if (put_user(len, optlen) ||
		    put_user(val, optval))
			return -EFAULT;
		err = 0;
		break;
2278 2279 2280 2281
	case NETLINK_NO_ENOBUFS:
		if (len < sizeof(int))
			return -EINVAL;
		len = sizeof(int);
2282
		val = nlk->flags & NETLINK_F_RECV_NO_ENOBUFS ? 1 : 0;
2283 2284 2285 2286 2287
		if (put_user(len, optlen) ||
		    put_user(val, optval))
			return -EFAULT;
		err = 0;
		break;
2288 2289 2290 2291 2292 2293 2294 2295 2296 2297 2298 2299 2300 2301
	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);
}

2302 2303 2304 2305 2306 2307 2308 2309 2310 2311
static void netlink_cmsg_listen_all_nsid(struct sock *sk, struct msghdr *msg,
					 struct sk_buff *skb)
{
	if (!NETLINK_CB(skb).nsid_is_set)
		return;

	put_cmsg(msg, SOL_NETLINK, NETLINK_LISTEN_ALL_NSID, sizeof(int),
		 &NETLINK_CB(skb).nsid);
}

2312
static int netlink_sendmsg(struct socket *sock, struct msghdr *msg, size_t len)
L
Linus Torvalds 已提交
2313 2314 2315
{
	struct sock *sk = sock->sk;
	struct netlink_sock *nlk = nlk_sk(sk);
2316
	DECLARE_SOCKADDR(struct sockaddr_nl *, addr, msg->msg_name);
2317
	u32 dst_portid;
2318
	u32 dst_group;
L
Linus Torvalds 已提交
2319 2320 2321
	struct sk_buff *skb;
	int err;
	struct scm_cookie scm;
2322
	u32 netlink_skb_flags = 0;
L
Linus Torvalds 已提交
2323 2324 2325 2326

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

C
Christoph Hellwig 已提交
2327
	err = scm_send(sock, msg, &scm, true);
L
Linus Torvalds 已提交
2328 2329 2330 2331
	if (err < 0)
		return err;

	if (msg->msg_namelen) {
2332
		err = -EINVAL;
L
Linus Torvalds 已提交
2333
		if (addr->nl_family != AF_NETLINK)
2334
			goto out;
2335
		dst_portid = addr->nl_pid;
2336
		dst_group = ffs(addr->nl_groups);
2337
		err =  -EPERM;
2338
		if ((dst_group || dst_portid) &&
2339
		    !netlink_allowed(sock, NL_CFG_F_NONROOT_SEND))
2340
			goto out;
2341
		netlink_skb_flags |= NETLINK_SKB_DST;
L
Linus Torvalds 已提交
2342
	} else {
2343
		dst_portid = nlk->dst_portid;
2344
		dst_group = nlk->dst_group;
L
Linus Torvalds 已提交
2345 2346
	}

2347
	if (!nlk->portid) {
L
Linus Torvalds 已提交
2348 2349 2350 2351 2352
		err = netlink_autobind(sock);
		if (err)
			goto out;
	}

2353 2354 2355
	/* It's a really convoluted way for userland to ask for mmaped
	 * sendmsg(), but that's what we've got...
	 */
2356
	if (netlink_tx_is_mmaped(sk) &&
2357 2358
	    msg->msg_iter.type == ITER_IOVEC &&
	    msg->msg_iter.nr_segs == 1 &&
A
Al Viro 已提交
2359
	    msg->msg_iter.iov->iov_base == NULL) {
2360
		err = netlink_mmap_sendmsg(sk, msg, dst_portid, dst_group,
C
Christoph Hellwig 已提交
2361
					   &scm);
2362 2363 2364
		goto out;
	}

L
Linus Torvalds 已提交
2365 2366 2367 2368
	err = -EMSGSIZE;
	if (len > sk->sk_sndbuf - 32)
		goto out;
	err = -ENOBUFS;
2369
	skb = netlink_alloc_large_skb(len, dst_group);
2370
	if (skb == NULL)
L
Linus Torvalds 已提交
2371 2372
		goto out;

2373
	NETLINK_CB(skb).portid	= nlk->portid;
2374
	NETLINK_CB(skb).dst_group = dst_group;
C
Christoph Hellwig 已提交
2375
	NETLINK_CB(skb).creds	= scm.creds;
2376
	NETLINK_CB(skb).flags	= netlink_skb_flags;
L
Linus Torvalds 已提交
2377 2378

	err = -EFAULT;
A
Al Viro 已提交
2379
	if (memcpy_from_msg(skb_put(skb, len), msg, len)) {
L
Linus Torvalds 已提交
2380 2381 2382 2383 2384 2385 2386 2387 2388 2389
		kfree_skb(skb);
		goto out;
	}

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

2390
	if (dst_group) {
L
Linus Torvalds 已提交
2391
		atomic_inc(&skb->users);
2392
		netlink_broadcast(sk, skb, dst_portid, dst_group, GFP_KERNEL);
L
Linus Torvalds 已提交
2393
	}
2394
	err = netlink_unicast(sk, skb, dst_portid, msg->msg_flags&MSG_DONTWAIT);
L
Linus Torvalds 已提交
2395 2396

out:
C
Christoph Hellwig 已提交
2397
	scm_destroy(&scm);
L
Linus Torvalds 已提交
2398 2399 2400
	return err;
}

2401
static int netlink_recvmsg(struct socket *sock, struct msghdr *msg, size_t len,
L
Linus Torvalds 已提交
2402 2403 2404 2405 2406 2407 2408
			   int flags)
{
	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
2409
	struct sk_buff *skb, *data_skb;
2410
	int err, ret;
L
Linus Torvalds 已提交
2411 2412 2413 2414 2415 2416

	if (flags&MSG_OOB)
		return -EOPNOTSUPP;

	copied = 0;

2417 2418
	skb = skb_recv_datagram(sk, flags, noblock, &err);
	if (skb == NULL)
L
Linus Torvalds 已提交
2419 2420
		goto out;

J
Johannes Berg 已提交
2421 2422
	data_skb = skb;

2423 2424 2425
#ifdef CONFIG_COMPAT_NETLINK_MESSAGES
	if (unlikely(skb_shinfo(skb)->frag_list)) {
		/*
J
Johannes Berg 已提交
2426 2427 2428
		 * 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.
2429
		 *
J
Johannes Berg 已提交
2430 2431 2432 2433
		 * 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.
2434
		 */
J
Johannes Berg 已提交
2435 2436
		if (flags & MSG_CMSG_COMPAT)
			data_skb = skb_shinfo(skb)->frag_list;
2437 2438 2439
	}
#endif

E
Eric Dumazet 已提交
2440 2441 2442 2443 2444
	/* Record the max length of recvmsg() calls for future allocations */
	nlk->max_recvmsg_len = max(nlk->max_recvmsg_len, len);
	nlk->max_recvmsg_len = min_t(size_t, nlk->max_recvmsg_len,
				     16384);

J
Johannes Berg 已提交
2445
	copied = data_skb->len;
L
Linus Torvalds 已提交
2446 2447 2448 2449 2450
	if (len < copied) {
		msg->msg_flags |= MSG_TRUNC;
		copied = len;
	}

J
Johannes Berg 已提交
2451
	skb_reset_transport_header(data_skb);
2452
	err = skb_copy_datagram_msg(data_skb, 0, msg, copied);
L
Linus Torvalds 已提交
2453 2454

	if (msg->msg_name) {
2455
		DECLARE_SOCKADDR(struct sockaddr_nl *, addr, msg->msg_name);
L
Linus Torvalds 已提交
2456 2457
		addr->nl_family = AF_NETLINK;
		addr->nl_pad    = 0;
2458
		addr->nl_pid	= NETLINK_CB(skb).portid;
2459
		addr->nl_groups	= netlink_group_mask(NETLINK_CB(skb).dst_group);
L
Linus Torvalds 已提交
2460 2461 2462
		msg->msg_namelen = sizeof(*addr);
	}

2463
	if (nlk->flags & NETLINK_F_RECV_PKTINFO)
2464
		netlink_cmsg_recv_pktinfo(msg, skb);
2465 2466
	if (nlk->flags & NETLINK_F_LISTEN_ALL_NSID)
		netlink_cmsg_listen_all_nsid(sk, msg, skb);
2467

C
Christoph Hellwig 已提交
2468 2469
	memset(&scm, 0, sizeof(scm));
	scm.creds = *NETLINK_CREDS(skb);
2470
	if (flags & MSG_TRUNC)
J
Johannes Berg 已提交
2471
		copied = data_skb->len;
2472

L
Linus Torvalds 已提交
2473 2474
	skb_free_datagram(sk, skb);

2475 2476
	if (nlk->cb_running &&
	    atomic_read(&sk->sk_rmem_alloc) <= sk->sk_rcvbuf / 2) {
2477 2478
		ret = netlink_dump(sk);
		if (ret) {
2479
			sk->sk_err = -ret;
2480 2481 2482
			sk->sk_error_report(sk);
		}
	}
L
Linus Torvalds 已提交
2483

C
Christoph Hellwig 已提交
2484
	scm_recv(sock, msg, &scm, flags);
L
Linus Torvalds 已提交
2485 2486 2487 2488 2489
out:
	netlink_rcv_wake(sk);
	return err ? : copied;
}

2490
static void netlink_data_ready(struct sock *sk)
L
Linus Torvalds 已提交
2491
{
2492
	BUG();
L
Linus Torvalds 已提交
2493 2494 2495
}

/*
2496
 *	We export these functions to other modules. They provide a
L
Linus Torvalds 已提交
2497 2498 2499 2500 2501
 *	complete set of kernel non-blocking support for message
 *	queueing.
 */

struct sock *
2502 2503
__netlink_kernel_create(struct net *net, int unit, struct module *module,
			struct netlink_kernel_cfg *cfg)
L
Linus Torvalds 已提交
2504 2505 2506
{
	struct socket *sock;
	struct sock *sk;
2507
	struct netlink_sock *nlk;
2508
	struct listeners *listeners = NULL;
2509 2510
	struct mutex *cb_mutex = cfg ? cfg->cb_mutex : NULL;
	unsigned int groups;
L
Linus Torvalds 已提交
2511

2512
	BUG_ON(!nl_table);
L
Linus Torvalds 已提交
2513

2514
	if (unit < 0 || unit >= MAX_LINKS)
L
Linus Torvalds 已提交
2515 2516 2517 2518
		return NULL;

	if (sock_create_lite(PF_NETLINK, SOCK_DGRAM, unit, &sock))
		return NULL;
2519 2520

	if (__netlink_create(net, sock, cb_mutex, unit, 1) < 0)
2521 2522 2523
		goto out_sock_release_nosk;

	sk = sock->sk;
2524

2525
	if (!cfg || cfg->groups < 32)
2526
		groups = 32;
2527 2528
	else
		groups = cfg->groups;
2529

2530
	listeners = kzalloc(sizeof(*listeners) + NLGRPSZ(groups), GFP_KERNEL);
2531 2532 2533
	if (!listeners)
		goto out_sock_release;

L
Linus Torvalds 已提交
2534
	sk->sk_data_ready = netlink_data_ready;
2535 2536
	if (cfg && cfg->input)
		nlk_sk(sk)->netlink_rcv = cfg->input;
L
Linus Torvalds 已提交
2537

2538
	if (netlink_insert(sk, 0))
2539
		goto out_sock_release;
2540

2541
	nlk = nlk_sk(sk);
2542
	nlk->flags |= NETLINK_F_KERNEL_SOCKET;
2543 2544

	netlink_table_grab();
2545 2546
	if (!nl_table[unit].registered) {
		nl_table[unit].groups = groups;
2547
		rcu_assign_pointer(nl_table[unit].listeners, listeners);
2548 2549
		nl_table[unit].cb_mutex = cb_mutex;
		nl_table[unit].module = module;
2550 2551
		if (cfg) {
			nl_table[unit].bind = cfg->bind;
2552
			nl_table[unit].unbind = cfg->unbind;
2553
			nl_table[unit].flags = cfg->flags;
2554 2555
			if (cfg->compare)
				nl_table[unit].compare = cfg->compare;
2556
		}
2557
		nl_table[unit].registered = 1;
2558 2559
	} else {
		kfree(listeners);
2560
		nl_table[unit].registered++;
2561
	}
2562
	netlink_table_ungrab();
2563 2564
	return sk;

2565
out_sock_release:
2566
	kfree(listeners);
2567
	netlink_kernel_release(sk);
2568 2569 2570
	return NULL;

out_sock_release_nosk:
2571
	sock_release(sock);
2572
	return NULL;
L
Linus Torvalds 已提交
2573
}
2574
EXPORT_SYMBOL(__netlink_kernel_create);
2575 2576 2577 2578

void
netlink_kernel_release(struct sock *sk)
{
2579 2580 2581 2582
	if (sk == NULL || sk->sk_socket == NULL)
		return;

	sock_release(sk->sk_socket);
2583 2584 2585
}
EXPORT_SYMBOL(netlink_kernel_release);

2586
int __netlink_change_ngroups(struct sock *sk, unsigned int groups)
2587
{
2588
	struct listeners *new, *old;
2589 2590 2591 2592 2593 2594
	struct netlink_table *tbl = &nl_table[sk->sk_protocol];

	if (groups < 32)
		groups = 32;

	if (NLGRPSZ(tbl->groups) < NLGRPSZ(groups)) {
2595 2596
		new = kzalloc(sizeof(*new) + NLGRPSZ(groups), GFP_ATOMIC);
		if (!new)
2597
			return -ENOMEM;
2598
		old = nl_deref_protected(tbl->listeners);
2599 2600 2601
		memcpy(new->masks, old->masks, NLGRPSZ(tbl->groups));
		rcu_assign_pointer(tbl->listeners, new);

2602
		kfree_rcu(old, rcu);
2603 2604 2605
	}
	tbl->groups = groups;

2606 2607 2608 2609 2610 2611 2612 2613 2614 2615 2616 2617 2618 2619 2620 2621 2622 2623 2624 2625 2626
	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);
2627
	netlink_table_ungrab();
2628

2629 2630 2631
	return err;
}

2632 2633 2634 2635 2636
void __netlink_clear_multicast_users(struct sock *ksk, unsigned int group)
{
	struct sock *sk;
	struct netlink_table *tbl = &nl_table[ksk->sk_protocol];

2637
	sk_for_each_bound(sk, &tbl->mc_list)
2638 2639 2640
		netlink_update_socket_mc(nlk_sk(sk), group, 0);
}

2641
struct nlmsghdr *
2642
__nlmsg_put(struct sk_buff *skb, u32 portid, u32 seq, int type, int len, int flags)
2643 2644
{
	struct nlmsghdr *nlh;
2645
	int size = nlmsg_msg_size(len);
2646

2647
	nlh = (struct nlmsghdr *)skb_put(skb, NLMSG_ALIGN(size));
2648 2649 2650
	nlh->nlmsg_type = type;
	nlh->nlmsg_len = size;
	nlh->nlmsg_flags = flags;
2651
	nlh->nlmsg_pid = portid;
2652 2653
	nlh->nlmsg_seq = seq;
	if (!__builtin_constant_p(size) || NLMSG_ALIGN(size) - size != 0)
2654
		memset(nlmsg_data(nlh) + len, 0, NLMSG_ALIGN(size) - size);
2655 2656 2657 2658
	return nlh;
}
EXPORT_SYMBOL(__nlmsg_put);

L
Linus Torvalds 已提交
2659 2660 2661 2662 2663 2664 2665 2666 2667
/*
 * 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;
2668
	struct sk_buff *skb = NULL;
L
Linus Torvalds 已提交
2669
	struct nlmsghdr *nlh;
2670
	int len, err = -ENOBUFS;
2671
	int alloc_size;
L
Linus Torvalds 已提交
2672

2673
	mutex_lock(nlk->cb_mutex);
2674
	if (!nlk->cb_running) {
2675 2676
		err = -EINVAL;
		goto errout_skb;
L
Linus Torvalds 已提交
2677 2678
	}

2679
	cb = &nlk->cb;
2680 2681
	alloc_size = max_t(int, cb->min_dump_alloc, NLMSG_GOODSIZE);

2682 2683 2684
	if (!netlink_rx_is_mmaped(sk) &&
	    atomic_read(&sk->sk_rmem_alloc) >= sk->sk_rcvbuf)
		goto errout_skb;
E
Eric Dumazet 已提交
2685 2686 2687 2688 2689 2690 2691 2692 2693 2694 2695 2696 2697 2698 2699 2700 2701 2702 2703 2704 2705

	/* NLMSG_GOODSIZE is small to avoid high order allocations being
	 * required, but it makes sense to _attempt_ a 16K bytes allocation
	 * to reduce number of system calls on dump operations, if user
	 * ever provided a big enough buffer.
	 */
	if (alloc_size < nlk->max_recvmsg_len) {
		skb = netlink_alloc_skb(sk,
					nlk->max_recvmsg_len,
					nlk->portid,
					GFP_KERNEL |
					__GFP_NOWARN |
					__GFP_NORETRY);
		/* available room should be exact amount to avoid MSG_TRUNC */
		if (skb)
			skb_reserve(skb, skb_tailroom(skb) -
					 nlk->max_recvmsg_len);
	}
	if (!skb)
		skb = netlink_alloc_skb(sk, alloc_size, nlk->portid,
					GFP_KERNEL);
2706
	if (!skb)
2707
		goto errout_skb;
2708
	netlink_skb_set_owner_r(skb, sk);
2709

L
Linus Torvalds 已提交
2710 2711 2712
	len = cb->dump(skb, cb);

	if (len > 0) {
2713
		mutex_unlock(nlk->cb_mutex);
2714 2715 2716

		if (sk_filter(sk, skb))
			kfree_skb(skb);
2717 2718
		else
			__netlink_sendskb(sk, skb);
L
Linus Torvalds 已提交
2719 2720 2721
		return 0;
	}

2722 2723 2724 2725
	nlh = nlmsg_put_answer(skb, cb, NLMSG_DONE, sizeof(len), NLM_F_MULTI);
	if (!nlh)
		goto errout_skb;

2726 2727
	nl_dump_check_consistent(cb, nlh);

2728 2729
	memcpy(nlmsg_data(nlh), &len, sizeof(len));

2730 2731
	if (sk_filter(sk, skb))
		kfree_skb(skb);
2732 2733
	else
		__netlink_sendskb(sk, skb);
L
Linus Torvalds 已提交
2734

2735 2736
	if (cb->done)
		cb->done(cb);
L
Linus Torvalds 已提交
2737

2738 2739
	nlk->cb_running = false;
	mutex_unlock(nlk->cb_mutex);
2740
	module_put(cb->module);
2741
	consume_skb(cb->skb);
L
Linus Torvalds 已提交
2742
	return 0;
2743

2744
errout_skb:
2745
	mutex_unlock(nlk->cb_mutex);
2746 2747
	kfree_skb(skb);
	return err;
L
Linus Torvalds 已提交
2748 2749
}

2750 2751 2752
int __netlink_dump_start(struct sock *ssk, struct sk_buff *skb,
			 const struct nlmsghdr *nlh,
			 struct netlink_dump_control *control)
L
Linus Torvalds 已提交
2753 2754 2755 2756
{
	struct netlink_callback *cb;
	struct sock *sk;
	struct netlink_sock *nlk;
2757
	int ret;
L
Linus Torvalds 已提交
2758

2759 2760 2761 2762 2763 2764
	/* Memory mapped dump requests need to be copied to avoid looping
	 * on the pending state in netlink_mmap_sendmsg() while the CB hold
	 * a reference to the skb.
	 */
	if (netlink_skb_is_mmaped(skb)) {
		skb = skb_copy(skb, GFP_KERNEL);
2765
		if (skb == NULL)
2766 2767 2768 2769
			return -ENOBUFS;
	} else
		atomic_inc(&skb->users);

2770
	sk = netlink_lookup(sock_net(ssk), ssk->sk_protocol, NETLINK_CB(skb).portid);
L
Linus Torvalds 已提交
2771
	if (sk == NULL) {
2772 2773
		ret = -ECONNREFUSED;
		goto error_free;
L
Linus Torvalds 已提交
2774
	}
2775

2776
	nlk = nlk_sk(sk);
2777
	mutex_lock(nlk->cb_mutex);
2778
	/* A dump is in progress... */
2779
	if (nlk->cb_running) {
2780
		ret = -EBUSY;
2781
		goto error_unlock;
L
Linus Torvalds 已提交
2782
	}
2783
	/* add reference of module which cb->dump belongs to */
2784
	if (!try_module_get(control->module)) {
2785
		ret = -EPROTONOSUPPORT;
2786
		goto error_unlock;
2787 2788
	}

2789 2790 2791 2792 2793 2794 2795 2796 2797 2798 2799 2800
	cb = &nlk->cb;
	memset(cb, 0, sizeof(*cb));
	cb->dump = control->dump;
	cb->done = control->done;
	cb->nlh = nlh;
	cb->data = control->data;
	cb->module = control->module;
	cb->min_dump_alloc = control->min_dump_alloc;
	cb->skb = skb;

	nlk->cb_running = true;

2801
	mutex_unlock(nlk->cb_mutex);
L
Linus Torvalds 已提交
2802

2803
	ret = netlink_dump(sk);
L
Linus Torvalds 已提交
2804
	sock_put(sk);
2805

2806 2807 2808
	if (ret)
		return ret;

2809 2810 2811 2812
	/* We successfully started a dump, by returning -EINTR we
	 * signal not to send ACK even if it was requested.
	 */
	return -EINTR;
2813 2814 2815 2816 2817 2818 2819

error_unlock:
	sock_put(sk);
	mutex_unlock(nlk->cb_mutex);
error_free:
	kfree_skb(skb);
	return ret;
L
Linus Torvalds 已提交
2820
}
2821
EXPORT_SYMBOL(__netlink_dump_start);
L
Linus Torvalds 已提交
2822 2823 2824 2825 2826 2827

void netlink_ack(struct sk_buff *in_skb, struct nlmsghdr *nlh, int err)
{
	struct sk_buff *skb;
	struct nlmsghdr *rep;
	struct nlmsgerr *errmsg;
2828
	size_t payload = sizeof(*errmsg);
L
Linus Torvalds 已提交
2829

2830 2831 2832
	/* error messages get the original request appened */
	if (err)
		payload += nlmsg_len(nlh);
L
Linus Torvalds 已提交
2833

2834 2835
	skb = netlink_alloc_skb(in_skb->sk, nlmsg_total_size(payload),
				NETLINK_CB(in_skb).portid, GFP_KERNEL);
L
Linus Torvalds 已提交
2836 2837 2838
	if (!skb) {
		struct sock *sk;

2839
		sk = netlink_lookup(sock_net(in_skb->sk),
2840
				    in_skb->sk->sk_protocol,
2841
				    NETLINK_CB(in_skb).portid);
L
Linus Torvalds 已提交
2842 2843 2844 2845 2846 2847 2848 2849
		if (sk) {
			sk->sk_err = ENOBUFS;
			sk->sk_error_report(sk);
			sock_put(sk);
		}
		return;
	}

2850
	rep = __nlmsg_put(skb, NETLINK_CB(in_skb).portid, nlh->nlmsg_seq,
2851
			  NLMSG_ERROR, payload, 0);
2852
	errmsg = nlmsg_data(rep);
L
Linus Torvalds 已提交
2853
	errmsg->error = err;
2854
	memcpy(&errmsg->msg, nlh, err ? nlh->nlmsg_len : sizeof(*nlh));
2855
	netlink_unicast(in_skb->sk, skb, NETLINK_CB(in_skb).portid, MSG_DONTWAIT);
L
Linus Torvalds 已提交
2856
}
2857
EXPORT_SYMBOL(netlink_ack);
L
Linus Torvalds 已提交
2858

2859
int netlink_rcv_skb(struct sk_buff *skb, int (*cb)(struct sk_buff *,
2860
						     struct nlmsghdr *))
2861 2862 2863 2864 2865
{
	struct nlmsghdr *nlh;
	int err;

	while (skb->len >= nlmsg_total_size(0)) {
2866 2867
		int msglen;

2868
		nlh = nlmsg_hdr(skb);
2869
		err = 0;
2870

2871
		if (nlh->nlmsg_len < NLMSG_HDRLEN || skb->len < nlh->nlmsg_len)
2872 2873
			return 0;

2874 2875
		/* Only requests are handled by the kernel */
		if (!(nlh->nlmsg_flags & NLM_F_REQUEST))
2876
			goto ack;
2877 2878 2879

		/* Skip control messages */
		if (nlh->nlmsg_type < NLMSG_MIN_TYPE)
2880
			goto ack;
2881

2882
		err = cb(skb, nlh);
2883 2884 2885 2886
		if (err == -EINTR)
			goto skip;

ack:
2887
		if (nlh->nlmsg_flags & NLM_F_ACK || err)
2888 2889
			netlink_ack(skb, nlh, err);

2890
skip:
2891
		msglen = NLMSG_ALIGN(nlh->nlmsg_len);
2892 2893 2894
		if (msglen > skb->len)
			msglen = skb->len;
		skb_pull(skb, msglen);
2895 2896 2897 2898
	}

	return 0;
}
2899
EXPORT_SYMBOL(netlink_rcv_skb);
2900

2901 2902 2903 2904
/**
 * nlmsg_notify - send a notification netlink message
 * @sk: netlink socket to use
 * @skb: notification message
2905
 * @portid: destination netlink portid for reports or 0
2906 2907 2908 2909
 * @group: destination multicast group or 0
 * @report: 1 to report back, 0 to disable
 * @flags: allocation flags
 */
2910
int nlmsg_notify(struct sock *sk, struct sk_buff *skb, u32 portid,
2911 2912 2913 2914 2915
		 unsigned int group, int report, gfp_t flags)
{
	int err = 0;

	if (group) {
2916
		int exclude_portid = 0;
2917 2918 2919

		if (report) {
			atomic_inc(&skb->users);
2920
			exclude_portid = portid;
2921 2922
		}

2923 2924
		/* errors reported via destination sk->sk_err, but propagate
		 * delivery errors if NETLINK_BROADCAST_ERROR flag is set */
2925
		err = nlmsg_multicast(sk, skb, exclude_portid, group, flags);
2926 2927
	}

2928 2929 2930
	if (report) {
		int err2;

2931
		err2 = nlmsg_unicast(sk, skb, portid);
2932 2933 2934
		if (!err || err == -ESRCH)
			err = err2;
	}
2935 2936 2937

	return err;
}
2938
EXPORT_SYMBOL(nlmsg_notify);
2939

L
Linus Torvalds 已提交
2940 2941
#ifdef CONFIG_PROC_FS
struct nl_seq_iter {
2942
	struct seq_net_private p;
2943
	struct rhashtable_iter hti;
L
Linus Torvalds 已提交
2944 2945 2946
	int link;
};

2947
static int netlink_walk_start(struct nl_seq_iter *iter)
L
Linus Torvalds 已提交
2948
{
2949
	int err;
L
Linus Torvalds 已提交
2950

2951 2952 2953 2954
	err = rhashtable_walk_init(&nl_table[iter->link].hash, &iter->hti);
	if (err) {
		iter->link = MAX_LINKS;
		return err;
L
Linus Torvalds 已提交
2955
	}
2956 2957 2958

	err = rhashtable_walk_start(&iter->hti);
	return err == -EAGAIN ? 0 : err;
L
Linus Torvalds 已提交
2959 2960
}

2961
static void netlink_walk_stop(struct nl_seq_iter *iter)
L
Linus Torvalds 已提交
2962
{
2963 2964
	rhashtable_walk_stop(&iter->hti);
	rhashtable_walk_exit(&iter->hti);
L
Linus Torvalds 已提交
2965 2966
}

2967
static void *__netlink_seq_next(struct seq_file *seq)
L
Linus Torvalds 已提交
2968
{
2969
	struct nl_seq_iter *iter = seq->private;
2970
	struct netlink_sock *nlk;
L
Linus Torvalds 已提交
2971

2972 2973 2974
	do {
		for (;;) {
			int err;
L
Linus Torvalds 已提交
2975

2976
			nlk = rhashtable_walk_next(&iter->hti);
2977

2978 2979 2980
			if (IS_ERR(nlk)) {
				if (PTR_ERR(nlk) == -EAGAIN)
					continue;
2981

2982 2983
				return nlk;
			}
L
Linus Torvalds 已提交
2984

2985 2986
			if (nlk)
				break;
L
Linus Torvalds 已提交
2987

2988 2989 2990
			netlink_walk_stop(iter);
			if (++iter->link >= MAX_LINKS)
				return NULL;
2991

2992 2993 2994
			err = netlink_walk_start(iter);
			if (err)
				return ERR_PTR(err);
L
Linus Torvalds 已提交
2995
		}
2996
	} while (sock_net(&nlk->sk) != seq_file_net(seq));
L
Linus Torvalds 已提交
2997

2998 2999
	return nlk;
}
L
Linus Torvalds 已提交
3000

3001 3002 3003 3004 3005 3006 3007 3008 3009 3010 3011 3012 3013 3014 3015 3016 3017 3018 3019 3020 3021 3022 3023
static void *netlink_seq_start(struct seq_file *seq, loff_t *posp)
{
	struct nl_seq_iter *iter = seq->private;
	void *obj = SEQ_START_TOKEN;
	loff_t pos;
	int err;

	iter->link = 0;

	err = netlink_walk_start(iter);
	if (err)
		return ERR_PTR(err);

	for (pos = *posp; pos && obj && !IS_ERR(obj); pos--)
		obj = __netlink_seq_next(seq);

	return obj;
}

static void *netlink_seq_next(struct seq_file *seq, void *v, loff_t *pos)
{
	++*pos;
	return __netlink_seq_next(seq);
L
Linus Torvalds 已提交
3024 3025 3026 3027
}

static void netlink_seq_stop(struct seq_file *seq, void *v)
{
3028 3029 3030 3031 3032 3033
	struct nl_seq_iter *iter = seq->private;

	if (iter->link >= MAX_LINKS)
		return;

	netlink_walk_stop(iter);
L
Linus Torvalds 已提交
3034 3035 3036 3037 3038
}


static int netlink_seq_show(struct seq_file *seq, void *v)
{
E
Eric Dumazet 已提交
3039
	if (v == SEQ_START_TOKEN) {
L
Linus Torvalds 已提交
3040 3041
		seq_puts(seq,
			 "sk       Eth Pid    Groups   "
3042
			 "Rmem     Wmem     Dump     Locks     Drops     Inode\n");
E
Eric Dumazet 已提交
3043
	} else {
L
Linus Torvalds 已提交
3044 3045 3046
		struct sock *s = v;
		struct netlink_sock *nlk = nlk_sk(s);

3047
		seq_printf(seq, "%pK %-3d %-6u %08x %-8d %-8d %d %-8d %-8d %-8lu\n",
L
Linus Torvalds 已提交
3048 3049
			   s,
			   s->sk_protocol,
3050
			   nlk->portid,
3051
			   nlk->groups ? (u32)nlk->groups[0] : 0,
3052 3053
			   sk_rmem_alloc_get(s),
			   sk_wmem_alloc_get(s),
3054
			   nlk->cb_running,
3055
			   atomic_read(&s->sk_refcnt),
3056 3057
			   atomic_read(&s->sk_drops),
			   sock_i_ino(s)
L
Linus Torvalds 已提交
3058 3059 3060 3061 3062 3063
			);

	}
	return 0;
}

3064
static const struct seq_operations netlink_seq_ops = {
L
Linus Torvalds 已提交
3065 3066 3067 3068 3069 3070 3071 3072 3073
	.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)
{
3074 3075
	return seq_open_net(inode, file, &netlink_seq_ops,
				sizeof(struct nl_seq_iter));
3076 3077
}

3078
static const struct file_operations netlink_seq_fops = {
L
Linus Torvalds 已提交
3079 3080 3081 3082
	.owner		= THIS_MODULE,
	.open		= netlink_seq_open,
	.read		= seq_read,
	.llseek		= seq_lseek,
3083
	.release	= seq_release_net,
L
Linus Torvalds 已提交
3084 3085 3086 3087 3088 3089
};

#endif

int netlink_register_notifier(struct notifier_block *nb)
{
3090
	return atomic_notifier_chain_register(&netlink_chain, nb);
L
Linus Torvalds 已提交
3091
}
3092
EXPORT_SYMBOL(netlink_register_notifier);
L
Linus Torvalds 已提交
3093 3094 3095

int netlink_unregister_notifier(struct notifier_block *nb)
{
3096
	return atomic_notifier_chain_unregister(&netlink_chain, nb);
L
Linus Torvalds 已提交
3097
}
3098
EXPORT_SYMBOL(netlink_unregister_notifier);
3099

3100
static const struct proto_ops netlink_ops = {
L
Linus Torvalds 已提交
3101 3102 3103 3104 3105 3106 3107 3108
	.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,
3109
	.poll =		netlink_poll,
L
Linus Torvalds 已提交
3110 3111 3112
	.ioctl =	sock_no_ioctl,
	.listen =	sock_no_listen,
	.shutdown =	sock_no_shutdown,
3113 3114
	.setsockopt =	netlink_setsockopt,
	.getsockopt =	netlink_getsockopt,
L
Linus Torvalds 已提交
3115 3116
	.sendmsg =	netlink_sendmsg,
	.recvmsg =	netlink_recvmsg,
3117
	.mmap =		netlink_mmap,
L
Linus Torvalds 已提交
3118 3119 3120
	.sendpage =	sock_no_sendpage,
};

3121
static const struct net_proto_family netlink_family_ops = {
L
Linus Torvalds 已提交
3122 3123 3124 3125 3126
	.family = PF_NETLINK,
	.create = netlink_create,
	.owner	= THIS_MODULE,	/* for consistency 8) */
};

3127
static int __net_init netlink_net_init(struct net *net)
3128 3129
{
#ifdef CONFIG_PROC_FS
3130
	if (!proc_create("netlink", 0, net->proc_net, &netlink_seq_fops))
3131 3132 3133 3134 3135
		return -ENOMEM;
#endif
	return 0;
}

3136
static void __net_exit netlink_net_exit(struct net *net)
3137 3138
{
#ifdef CONFIG_PROC_FS
3139
	remove_proc_entry("netlink", net->proc_net);
3140 3141 3142
#endif
}

3143 3144
static void __init netlink_add_usersock_entry(void)
{
3145
	struct listeners *listeners;
3146 3147
	int groups = 32;

3148
	listeners = kzalloc(sizeof(*listeners) + NLGRPSZ(groups), GFP_KERNEL);
3149
	if (!listeners)
3150
		panic("netlink_add_usersock_entry: Cannot allocate listeners\n");
3151 3152 3153 3154

	netlink_table_grab();

	nl_table[NETLINK_USERSOCK].groups = groups;
3155
	rcu_assign_pointer(nl_table[NETLINK_USERSOCK].listeners, listeners);
3156 3157
	nl_table[NETLINK_USERSOCK].module = THIS_MODULE;
	nl_table[NETLINK_USERSOCK].registered = 1;
3158
	nl_table[NETLINK_USERSOCK].flags = NL_CFG_F_NONROOT_SEND;
3159 3160 3161 3162

	netlink_table_ungrab();
}

3163
static struct pernet_operations __net_initdata netlink_net_ops = {
3164 3165 3166 3167
	.init = netlink_net_init,
	.exit = netlink_net_exit,
};

3168
static inline u32 netlink_hash(const void *data, u32 len, u32 seed)
3169 3170 3171 3172 3173
{
	const struct netlink_sock *nlk = data;
	struct netlink_compare_arg arg;

	netlink_compare_arg_init(&arg, sock_net(&nlk->sk), nlk->portid);
3174
	return jhash2((u32 *)&arg, netlink_compare_arg_len / sizeof(u32), seed);
3175 3176 3177 3178 3179 3180 3181
}

static const struct rhashtable_params netlink_rhashtable_params = {
	.head_offset = offsetof(struct netlink_sock, node),
	.key_len = netlink_compare_arg_len,
	.obj_hashfn = netlink_hash,
	.obj_cmpfn = netlink_compare,
3182
	.automatic_shrinking = true,
3183 3184
};

L
Linus Torvalds 已提交
3185 3186 3187 3188 3189 3190 3191 3192
static int __init netlink_proto_init(void)
{
	int i;
	int err = proto_register(&netlink_proto, 0);

	if (err != 0)
		goto out;

3193
	BUILD_BUG_ON(sizeof(struct netlink_skb_parms) > FIELD_SIZEOF(struct sk_buff, cb));
L
Linus Torvalds 已提交
3194

3195
	nl_table = kcalloc(MAX_LINKS, sizeof(*nl_table), GFP_KERNEL);
3196 3197
	if (!nl_table)
		goto panic;
L
Linus Torvalds 已提交
3198 3199

	for (i = 0; i < MAX_LINKS; i++) {
3200 3201
		if (rhashtable_init(&nl_table[i].hash,
				    &netlink_rhashtable_params) < 0) {
3202 3203
			while (--i > 0)
				rhashtable_destroy(&nl_table[i].hash);
L
Linus Torvalds 已提交
3204
			kfree(nl_table);
3205
			goto panic;
L
Linus Torvalds 已提交
3206 3207 3208
		}
	}

3209 3210
	INIT_LIST_HEAD(&netlink_tap_all);

3211 3212
	netlink_add_usersock_entry();

L
Linus Torvalds 已提交
3213
	sock_register(&netlink_family_ops);
3214
	register_pernet_subsys(&netlink_net_ops);
3215
	/* The netlink device handler may be needed early. */
L
Linus Torvalds 已提交
3216 3217 3218
	rtnetlink_init();
out:
	return err;
3219 3220
panic:
	panic("netlink_init: Cannot allocate nl_table\n");
L
Linus Torvalds 已提交
3221 3222 3223
}

core_initcall(netlink_proto_init);