af_unix.c 68.4 KB
Newer Older
L
Linus Torvalds 已提交
1 2 3
/*
 * NET4:	Implementation of BSD Unix domain sockets.
 *
4
 * Authors:	Alan Cox, <alan@lxorguk.ukuu.org.uk>
L
Linus Torvalds 已提交
5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82
 *
 *		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.
 *
 * Fixes:
 *		Linus Torvalds	:	Assorted bug cures.
 *		Niibe Yutaka	:	async I/O support.
 *		Carsten Paeth	:	PF_UNIX check, address fixes.
 *		Alan Cox	:	Limit size of allocated blocks.
 *		Alan Cox	:	Fixed the stupid socketpair bug.
 *		Alan Cox	:	BSD compatibility fine tuning.
 *		Alan Cox	:	Fixed a bug in connect when interrupted.
 *		Alan Cox	:	Sorted out a proper draft version of
 *					file descriptor passing hacked up from
 *					Mike Shaver's work.
 *		Marty Leisner	:	Fixes to fd passing
 *		Nick Nevin	:	recvmsg bugfix.
 *		Alan Cox	:	Started proper garbage collector
 *		Heiko EiBfeldt	:	Missing verify_area check
 *		Alan Cox	:	Started POSIXisms
 *		Andreas Schwab	:	Replace inode by dentry for proper
 *					reference counting
 *		Kirk Petersen	:	Made this a module
 *	    Christoph Rohland	:	Elegant non-blocking accept/connect algorithm.
 *					Lots of bug fixes.
 *	     Alexey Kuznetosv	:	Repaired (I hope) bugs introduces
 *					by above two patches.
 *	     Andrea Arcangeli	:	If possible we block in connect(2)
 *					if the max backlog of the listen socket
 *					is been reached. This won't break
 *					old apps and it will avoid huge amount
 *					of socks hashed (this for unix_gc()
 *					performances reasons).
 *					Security fix that limits the max
 *					number of socks to 2*max_files and
 *					the number of skb queueable in the
 *					dgram receiver.
 *		Artur Skawina   :	Hash function optimizations
 *	     Alexey Kuznetsov   :	Full scale SMP. Lot of bugs are introduced 8)
 *	      Malcolm Beattie   :	Set peercred for socketpair
 *	     Michal Ostrowski   :       Module initialization cleanup.
 *	     Arnaldo C. Melo	:	Remove MOD_{INC,DEC}_USE_COUNT,
 *	     				the core infrastructure is doing that
 *	     				for all net proto families now (2.5.69+)
 *
 *
 * Known differences from reference BSD that was tested:
 *
 *	[TO FIX]
 *	ECONNREFUSED is not returned from one end of a connected() socket to the
 *		other the moment one end closes.
 *	fstat() doesn't return st_dev=0, and give the blksize as high water mark
 *		and a fake inode identifier (nor the BSD first socket fstat twice bug).
 *	[NOT TO FIX]
 *	accept() returns a path name even if the connecting socket has closed
 *		in the meantime (BSD loses the path and gives up).
 *	accept() returns 0 length path for an unbound connector. BSD returns 16
 *		and a null first byte in the path (but not for gethost/peername - BSD bug ??)
 *	socketpair(...SOCK_RAW..) doesn't panic the kernel.
 *	BSD af_unix apparently has connect forgetting to block properly.
 *		(need to check this with the POSIX spec in detail)
 *
 * Differences from 2.0.0-11-... (ANK)
 *	Bug fixes and improvements.
 *		- client shutdown killed server socket.
 *		- removed all useless cli/sti pairs.
 *
 *	Semantic changes/extensions.
 *		- generic control message passing.
 *		- SCM_CREDENTIALS control message.
 *		- "Abstract" (not FS based) socket bindings.
 *		  Abstract names are sequences of bytes (not zero terminated)
 *		  started by 0, so that this name space does not intersect
 *		  with BSD names.
 */

W
wangweidong 已提交
83 84
#define pr_fmt(fmt) KBUILD_MODNAME ": " fmt

L
Linus Torvalds 已提交
85 86 87
#include <linux/module.h>
#include <linux/kernel.h>
#include <linux/signal.h>
88
#include <linux/sched/signal.h>
L
Linus Torvalds 已提交
89 90 91 92 93 94 95 96 97 98 99 100 101 102
#include <linux/errno.h>
#include <linux/string.h>
#include <linux/stat.h>
#include <linux/dcache.h>
#include <linux/namei.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/in.h>
#include <linux/fs.h>
#include <linux/slab.h>
103
#include <linux/uaccess.h>
L
Linus Torvalds 已提交
104 105
#include <linux/skbuff.h>
#include <linux/netdevice.h>
106
#include <net/net_namespace.h>
L
Linus Torvalds 已提交
107
#include <net/sock.h>
108
#include <net/tcp_states.h>
L
Linus Torvalds 已提交
109 110 111 112 113 114 115 116 117 118
#include <net/af_unix.h>
#include <linux/proc_fs.h>
#include <linux/seq_file.h>
#include <net/scm.h>
#include <linux/init.h>
#include <linux/poll.h>
#include <linux/rtnetlink.h>
#include <linux/mount.h>
#include <net/checksum.h>
#include <linux/security.h>
119
#include <linux/freezer.h>
120
#include <linux/file.h>
L
Linus Torvalds 已提交
121

E
Eric Dumazet 已提交
122
struct hlist_head unix_socket_table[2 * UNIX_HASH_SIZE];
123 124 125
EXPORT_SYMBOL_GPL(unix_socket_table);
DEFINE_SPINLOCK(unix_table_lock);
EXPORT_SYMBOL_GPL(unix_table_lock);
E
Eric Dumazet 已提交
126
static atomic_long_t unix_nr_socks;
L
Linus Torvalds 已提交
127 128


E
Eric Dumazet 已提交
129 130 131 132 133 134 135 136 137 138 139
static struct hlist_head *unix_sockets_unbound(void *addr)
{
	unsigned long hash = (unsigned long)addr;

	hash ^= hash >> 16;
	hash ^= hash >> 8;
	hash %= UNIX_HASH_SIZE;
	return &unix_socket_table[UNIX_HASH_SIZE + hash];
}

#define UNIX_ABSTRACT(sk)	(unix_sk(sk)->addr->hash < UNIX_HASH_SIZE)
L
Linus Torvalds 已提交
140

C
Catherine Zhang 已提交
141
#ifdef CONFIG_SECURITY_NETWORK
142
static void unix_get_secdata(struct scm_cookie *scm, struct sk_buff *skb)
C
Catherine Zhang 已提交
143
{
144
	UNIXCB(skb).secid = scm->secid;
C
Catherine Zhang 已提交
145 146 147 148
}

static inline void unix_set_secdata(struct scm_cookie *scm, struct sk_buff *skb)
{
149 150 151 152 153 154
	scm->secid = UNIXCB(skb).secid;
}

static inline bool unix_secdata_eq(struct scm_cookie *scm, struct sk_buff *skb)
{
	return (scm->secid == UNIXCB(skb).secid);
C
Catherine Zhang 已提交
155 156
}
#else
157
static inline void unix_get_secdata(struct scm_cookie *scm, struct sk_buff *skb)
C
Catherine Zhang 已提交
158 159 160 161
{ }

static inline void unix_set_secdata(struct scm_cookie *scm, struct sk_buff *skb)
{ }
162 163 164 165 166

static inline bool unix_secdata_eq(struct scm_cookie *scm, struct sk_buff *skb)
{
	return true;
}
C
Catherine Zhang 已提交
167 168
#endif /* CONFIG_SECURITY_NETWORK */

L
Linus Torvalds 已提交
169 170
/*
 *  SMP locking strategy:
171
 *    hash table is protected with spinlock unix_table_lock
172
 *    each socket state is protected by separate spin lock.
L
Linus Torvalds 已提交
173 174
 */

175
static inline unsigned int unix_hash_fold(__wsum n)
L
Linus Torvalds 已提交
176
{
177
	unsigned int hash = (__force unsigned int)csum_fold(n);
178

L
Linus Torvalds 已提交
179 180 181 182 183 184 185 186 187 188 189 190 191
	hash ^= hash>>8;
	return hash&(UNIX_HASH_SIZE-1);
}

#define unix_peer(sk) (unix_sk(sk)->peer)

static inline int unix_our_peer(struct sock *sk, struct sock *osk)
{
	return unix_peer(osk) == sk;
}

static inline int unix_may_send(struct sock *sk, struct sock *osk)
{
E
Eric Dumazet 已提交
192
	return unix_peer(osk) == NULL || unix_our_peer(sk, osk);
L
Linus Torvalds 已提交
193 194
}

195 196 197 198 199
static inline int unix_recvq_full(struct sock const *sk)
{
	return skb_queue_len(&sk->sk_receive_queue) > sk->sk_max_ack_backlog;
}

200
struct sock *unix_peer_get(struct sock *s)
L
Linus Torvalds 已提交
201 202 203
{
	struct sock *peer;

204
	unix_state_lock(s);
L
Linus Torvalds 已提交
205 206 207
	peer = unix_peer(s);
	if (peer)
		sock_hold(peer);
208
	unix_state_unlock(s);
L
Linus Torvalds 已提交
209 210
	return peer;
}
211
EXPORT_SYMBOL_GPL(unix_peer_get);
L
Linus Torvalds 已提交
212 213 214 215 216 217 218 219 220 221 222 223 224

static inline void unix_release_addr(struct unix_address *addr)
{
	if (atomic_dec_and_test(&addr->refcnt))
		kfree(addr);
}

/*
 *	Check unix socket name:
 *		- should be not zero length.
 *	        - if started by not zero, should be NULL terminated (FS object)
 *		- if started by zero, it is abstract name.
 */
225

226
static int unix_mkname(struct sockaddr_un *sunaddr, int len, unsigned int *hashp)
L
Linus Torvalds 已提交
227 228 229 230 231 232 233 234 235
{
	if (len <= sizeof(short) || len > sizeof(*sunaddr))
		return -EINVAL;
	if (!sunaddr || sunaddr->sun_family != AF_UNIX)
		return -EINVAL;
	if (sunaddr->sun_path[0]) {
		/*
		 * This may look like an off by one error but it is a bit more
		 * subtle. 108 is the longest valid AF_UNIX path for a binding.
L
Lucas De Marchi 已提交
236
		 * sun_path[108] doesn't as such exist.  However in kernel space
L
Linus Torvalds 已提交
237 238 239
		 * we are guaranteed that it is a valid memory location in our
		 * kernel address buffer.
		 */
240
		((char *)sunaddr)[len] = 0;
L
Linus Torvalds 已提交
241 242 243 244
		len = strlen(sunaddr->sun_path)+1+sizeof(short);
		return len;
	}

245
	*hashp = unix_hash_fold(csum_partial(sunaddr, len, 0));
L
Linus Torvalds 已提交
246 247 248 249 250 251 252 253 254 255
	return len;
}

static void __unix_remove_socket(struct sock *sk)
{
	sk_del_node_init(sk);
}

static void __unix_insert_socket(struct hlist_head *list, struct sock *sk)
{
256
	WARN_ON(!sk_unhashed(sk));
L
Linus Torvalds 已提交
257 258 259 260 261
	sk_add_node(sk, list);
}

static inline void unix_remove_socket(struct sock *sk)
{
262
	spin_lock(&unix_table_lock);
L
Linus Torvalds 已提交
263
	__unix_remove_socket(sk);
264
	spin_unlock(&unix_table_lock);
L
Linus Torvalds 已提交
265 266 267 268
}

static inline void unix_insert_socket(struct hlist_head *list, struct sock *sk)
{
269
	spin_lock(&unix_table_lock);
L
Linus Torvalds 已提交
270
	__unix_insert_socket(list, sk);
271
	spin_unlock(&unix_table_lock);
L
Linus Torvalds 已提交
272 273
}

274 275
static struct sock *__unix_find_socket_byname(struct net *net,
					      struct sockaddr_un *sunname,
276
					      int len, int type, unsigned int hash)
L
Linus Torvalds 已提交
277 278 279
{
	struct sock *s;

280
	sk_for_each(s, &unix_socket_table[hash ^ type]) {
L
Linus Torvalds 已提交
281 282
		struct unix_sock *u = unix_sk(s);

283
		if (!net_eq(sock_net(s), net))
284 285
			continue;

L
Linus Torvalds 已提交
286 287 288 289 290 291 292 293 294
		if (u->addr->len == len &&
		    !memcmp(u->addr->name, sunname, len))
			goto found;
	}
	s = NULL;
found:
	return s;
}

295 296
static inline struct sock *unix_find_socket_byname(struct net *net,
						   struct sockaddr_un *sunname,
L
Linus Torvalds 已提交
297
						   int len, int type,
298
						   unsigned int hash)
L
Linus Torvalds 已提交
299 300 301
{
	struct sock *s;

302
	spin_lock(&unix_table_lock);
303
	s = __unix_find_socket_byname(net, sunname, len, type, hash);
L
Linus Torvalds 已提交
304 305
	if (s)
		sock_hold(s);
306
	spin_unlock(&unix_table_lock);
L
Linus Torvalds 已提交
307 308 309
	return s;
}

310
static struct sock *unix_find_socket_byinode(struct inode *i)
L
Linus Torvalds 已提交
311 312 313
{
	struct sock *s;

314
	spin_lock(&unix_table_lock);
315
	sk_for_each(s,
L
Linus Torvalds 已提交
316
		    &unix_socket_table[i->i_ino & (UNIX_HASH_SIZE - 1)]) {
A
Al Viro 已提交
317
		struct dentry *dentry = unix_sk(s)->path.dentry;
L
Linus Torvalds 已提交
318

319
		if (dentry && d_backing_inode(dentry) == i) {
L
Linus Torvalds 已提交
320 321 322 323 324 325
			sock_hold(s);
			goto found;
		}
	}
	s = NULL;
found:
326
	spin_unlock(&unix_table_lock);
L
Linus Torvalds 已提交
327 328 329
	return s;
}

330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 439 440 441
/* Support code for asymmetrically connected dgram sockets
 *
 * If a datagram socket is connected to a socket not itself connected
 * to the first socket (eg, /dev/log), clients may only enqueue more
 * messages if the present receive queue of the server socket is not
 * "too large". This means there's a second writeability condition
 * poll and sendmsg need to test. The dgram recv code will do a wake
 * up on the peer_wait wait queue of a socket upon reception of a
 * datagram which needs to be propagated to sleeping would-be writers
 * since these might not have sent anything so far. This can't be
 * accomplished via poll_wait because the lifetime of the server
 * socket might be less than that of its clients if these break their
 * association with it or if the server socket is closed while clients
 * are still connected to it and there's no way to inform "a polling
 * implementation" that it should let go of a certain wait queue
 *
 * In order to propagate a wake up, a wait_queue_t of the client
 * socket is enqueued on the peer_wait queue of the server socket
 * whose wake function does a wake_up on the ordinary client socket
 * wait queue. This connection is established whenever a write (or
 * poll for write) hit the flow control condition and broken when the
 * association to the server socket is dissolved or after a wake up
 * was relayed.
 */

static int unix_dgram_peer_wake_relay(wait_queue_t *q, unsigned mode, int flags,
				      void *key)
{
	struct unix_sock *u;
	wait_queue_head_t *u_sleep;

	u = container_of(q, struct unix_sock, peer_wake);

	__remove_wait_queue(&unix_sk(u->peer_wake.private)->peer_wait,
			    q);
	u->peer_wake.private = NULL;

	/* relaying can only happen while the wq still exists */
	u_sleep = sk_sleep(&u->sk);
	if (u_sleep)
		wake_up_interruptible_poll(u_sleep, key);

	return 0;
}

static int unix_dgram_peer_wake_connect(struct sock *sk, struct sock *other)
{
	struct unix_sock *u, *u_other;
	int rc;

	u = unix_sk(sk);
	u_other = unix_sk(other);
	rc = 0;
	spin_lock(&u_other->peer_wait.lock);

	if (!u->peer_wake.private) {
		u->peer_wake.private = other;
		__add_wait_queue(&u_other->peer_wait, &u->peer_wake);

		rc = 1;
	}

	spin_unlock(&u_other->peer_wait.lock);
	return rc;
}

static void unix_dgram_peer_wake_disconnect(struct sock *sk,
					    struct sock *other)
{
	struct unix_sock *u, *u_other;

	u = unix_sk(sk);
	u_other = unix_sk(other);
	spin_lock(&u_other->peer_wait.lock);

	if (u->peer_wake.private == other) {
		__remove_wait_queue(&u_other->peer_wait, &u->peer_wake);
		u->peer_wake.private = NULL;
	}

	spin_unlock(&u_other->peer_wait.lock);
}

static void unix_dgram_peer_wake_disconnect_wakeup(struct sock *sk,
						   struct sock *other)
{
	unix_dgram_peer_wake_disconnect(sk, other);
	wake_up_interruptible_poll(sk_sleep(sk),
				   POLLOUT |
				   POLLWRNORM |
				   POLLWRBAND);
}

/* preconditions:
 *	- unix_peer(sk) == other
 *	- association is stable
 */
static int unix_dgram_peer_wake_me(struct sock *sk, struct sock *other)
{
	int connected;

	connected = unix_dgram_peer_wake_connect(sk, other);

	if (unix_recvq_full(other))
		return 1;

	if (connected)
		unix_dgram_peer_wake_disconnect(sk, other);

	return 0;
}

442
static int unix_writable(const struct sock *sk)
L
Linus Torvalds 已提交
443
{
444 445
	return sk->sk_state != TCP_LISTEN &&
	       (atomic_read(&sk->sk_wmem_alloc) << 2) <= sk->sk_sndbuf;
L
Linus Torvalds 已提交
446 447 448 449
}

static void unix_write_space(struct sock *sk)
{
450 451 452
	struct socket_wq *wq;

	rcu_read_lock();
L
Linus Torvalds 已提交
453
	if (unix_writable(sk)) {
454
		wq = rcu_dereference(sk->sk_wq);
H
Herbert Xu 已提交
455
		if (skwq_has_sleeper(wq))
E
Eric Dumazet 已提交
456 457
			wake_up_interruptible_sync_poll(&wq->wait,
				POLLOUT | POLLWRNORM | POLLWRBAND);
458
		sk_wake_async(sk, SOCK_WAKE_SPACE, POLL_OUT);
L
Linus Torvalds 已提交
459
	}
460
	rcu_read_unlock();
L
Linus Torvalds 已提交
461 462 463 464 465 466 467 468
}

/* When dgram socket disconnects (or changes its peer), we clear its receive
 * queue of packets arrived from previous peer. First, it allows to do
 * flow control based only on wmem_alloc; second, sk connected to peer
 * may receive messages only from that peer. */
static void unix_dgram_disconnected(struct sock *sk, struct sock *other)
{
469
	if (!skb_queue_empty(&sk->sk_receive_queue)) {
L
Linus Torvalds 已提交
470 471 472 473 474 475 476 477 478 479 480 481 482 483 484 485 486 487 488 489
		skb_queue_purge(&sk->sk_receive_queue);
		wake_up_interruptible_all(&unix_sk(sk)->peer_wait);

		/* If one link of bidirectional dgram pipe is disconnected,
		 * we signal error. Messages are lost. Do not make this,
		 * when peer was not connected to us.
		 */
		if (!sock_flag(other, SOCK_DEAD) && unix_peer(other) == sk) {
			other->sk_err = ECONNRESET;
			other->sk_error_report(other);
		}
	}
}

static void unix_sock_destructor(struct sock *sk)
{
	struct unix_sock *u = unix_sk(sk);

	skb_queue_purge(&sk->sk_receive_queue);

490 491 492
	WARN_ON(atomic_read(&sk->sk_wmem_alloc));
	WARN_ON(!sk_unhashed(sk));
	WARN_ON(sk->sk_socket);
L
Linus Torvalds 已提交
493
	if (!sock_flag(sk, SOCK_DEAD)) {
W
wangweidong 已提交
494
		pr_info("Attempt to release alive unix socket: %p\n", sk);
L
Linus Torvalds 已提交
495 496 497 498 499 500
		return;
	}

	if (u->addr)
		unix_release_addr(u->addr);

E
Eric Dumazet 已提交
501
	atomic_long_dec(&unix_nr_socks);
502
	local_bh_disable();
503
	sock_prot_inuse_add(sock_net(sk), sk->sk_prot, -1);
504
	local_bh_enable();
L
Linus Torvalds 已提交
505
#ifdef UNIX_REFCNT_DEBUG
W
wangweidong 已提交
506
	pr_debug("UNIX %p is destroyed, %ld are still alive.\n", sk,
E
Eric Dumazet 已提交
507
		atomic_long_read(&unix_nr_socks));
L
Linus Torvalds 已提交
508 509 510
#endif
}

511
static void unix_release_sock(struct sock *sk, int embrion)
L
Linus Torvalds 已提交
512 513
{
	struct unix_sock *u = unix_sk(sk);
A
Al Viro 已提交
514
	struct path path;
L
Linus Torvalds 已提交
515 516 517 518 519 520 521
	struct sock *skpair;
	struct sk_buff *skb;
	int state;

	unix_remove_socket(sk);

	/* Clear state */
522
	unix_state_lock(sk);
L
Linus Torvalds 已提交
523 524
	sock_orphan(sk);
	sk->sk_shutdown = SHUTDOWN_MASK;
A
Al Viro 已提交
525 526 527
	path	     = u->path;
	u->path.dentry = NULL;
	u->path.mnt = NULL;
L
Linus Torvalds 已提交
528 529
	state = sk->sk_state;
	sk->sk_state = TCP_CLOSE;
530
	unix_state_unlock(sk);
L
Linus Torvalds 已提交
531 532 533

	wake_up_interruptible_all(&u->peer_wait);

534
	skpair = unix_peer(sk);
L
Linus Torvalds 已提交
535

536
	if (skpair != NULL) {
L
Linus Torvalds 已提交
537
		if (sk->sk_type == SOCK_STREAM || sk->sk_type == SOCK_SEQPACKET) {
538
			unix_state_lock(skpair);
L
Linus Torvalds 已提交
539 540 541 542
			/* No more writes */
			skpair->sk_shutdown = SHUTDOWN_MASK;
			if (!skb_queue_empty(&sk->sk_receive_queue) || embrion)
				skpair->sk_err = ECONNRESET;
543
			unix_state_unlock(skpair);
L
Linus Torvalds 已提交
544
			skpair->sk_state_change(skpair);
545
			sk_wake_async(skpair, SOCK_WAKE_WAITD, POLL_HUP);
L
Linus Torvalds 已提交
546
		}
547 548

		unix_dgram_peer_wake_disconnect(sk, skpair);
L
Linus Torvalds 已提交
549 550 551 552 553 554 555
		sock_put(skpair); /* It may now die */
		unix_peer(sk) = NULL;
	}

	/* Try to flush out this socket. Throw out buffers at least */

	while ((skb = skb_dequeue(&sk->sk_receive_queue)) != NULL) {
556
		if (state == TCP_LISTEN)
L
Linus Torvalds 已提交
557 558
			unix_release_sock(skb->sk, 1);
		/* passed fds are erased in the kfree_skb hook	      */
559
		UNIXCB(skb).consumed = skb->len;
L
Linus Torvalds 已提交
560 561 562
		kfree_skb(skb);
	}

A
Al Viro 已提交
563 564
	if (path.dentry)
		path_put(&path);
L
Linus Torvalds 已提交
565 566 567 568 569 570

	sock_put(sk);

	/* ---- Socket is dead now and most probably destroyed ---- */

	/*
A
Alan Cox 已提交
571
	 * Fixme: BSD difference: In BSD all sockets connected to us get
L
Linus Torvalds 已提交
572 573 574 575 576 577 578 579 580
	 *	  ECONNRESET and we die on the spot. In Linux we behave
	 *	  like files and pipes do and wait for the last
	 *	  dereference.
	 *
	 * Can't we simply set sock->err?
	 *
	 *	  What the above comment does talk about? --ANK(980817)
	 */

581
	if (unix_tot_inflight)
582
		unix_gc();		/* Garbage collect fds */
L
Linus Torvalds 已提交
583 584
}

585 586 587 588 589 590 591 592 593 594 595 596 597 598 599 600 601 602
static void init_peercred(struct sock *sk)
{
	put_pid(sk->sk_peer_pid);
	if (sk->sk_peer_cred)
		put_cred(sk->sk_peer_cred);
	sk->sk_peer_pid  = get_pid(task_tgid(current));
	sk->sk_peer_cred = get_current_cred();
}

static void copy_peercred(struct sock *sk, struct sock *peersk)
{
	put_pid(sk->sk_peer_pid);
	if (sk->sk_peer_cred)
		put_cred(sk->sk_peer_cred);
	sk->sk_peer_pid  = get_pid(peersk->sk_peer_pid);
	sk->sk_peer_cred = get_cred(peersk->sk_peer_cred);
}

L
Linus Torvalds 已提交
603 604 605 606 607
static int unix_listen(struct socket *sock, int backlog)
{
	int err;
	struct sock *sk = sock->sk;
	struct unix_sock *u = unix_sk(sk);
608
	struct pid *old_pid = NULL;
L
Linus Torvalds 已提交
609 610

	err = -EOPNOTSUPP;
E
Eric Dumazet 已提交
611 612
	if (sock->type != SOCK_STREAM && sock->type != SOCK_SEQPACKET)
		goto out;	/* Only stream/seqpacket sockets accept */
L
Linus Torvalds 已提交
613 614
	err = -EINVAL;
	if (!u->addr)
E
Eric Dumazet 已提交
615
		goto out;	/* No listens on an unbound socket */
616
	unix_state_lock(sk);
L
Linus Torvalds 已提交
617 618 619 620 621 622 623
	if (sk->sk_state != TCP_CLOSE && sk->sk_state != TCP_LISTEN)
		goto out_unlock;
	if (backlog > sk->sk_max_ack_backlog)
		wake_up_interruptible_all(&u->peer_wait);
	sk->sk_max_ack_backlog	= backlog;
	sk->sk_state		= TCP_LISTEN;
	/* set credentials so connect can copy them */
624
	init_peercred(sk);
L
Linus Torvalds 已提交
625 626 627
	err = 0;

out_unlock:
628
	unix_state_unlock(sk);
629
	put_pid(old_pid);
L
Linus Torvalds 已提交
630 631 632 633 634 635 636 637 638
out:
	return err;
}

static int unix_release(struct socket *);
static int unix_bind(struct socket *, struct sockaddr *, int);
static int unix_stream_connect(struct socket *, struct sockaddr *,
			       int addr_len, int flags);
static int unix_socketpair(struct socket *, struct socket *);
639
static int unix_accept(struct socket *, struct socket *, int, bool);
L
Linus Torvalds 已提交
640 641
static int unix_getname(struct socket *, struct sockaddr *, int *, int);
static unsigned int unix_poll(struct file *, struct socket *, poll_table *);
642 643
static unsigned int unix_dgram_poll(struct file *, struct socket *,
				    poll_table *);
L
Linus Torvalds 已提交
644 645
static int unix_ioctl(struct socket *, unsigned int, unsigned long);
static int unix_shutdown(struct socket *, int);
646 647
static int unix_stream_sendmsg(struct socket *, struct msghdr *, size_t);
static int unix_stream_recvmsg(struct socket *, struct msghdr *, size_t, int);
648 649
static ssize_t unix_stream_sendpage(struct socket *, struct page *, int offset,
				    size_t size, int flags);
650 651 652
static ssize_t unix_stream_splice_read(struct socket *,  loff_t *ppos,
				       struct pipe_inode_info *, size_t size,
				       unsigned int flags);
653 654
static int unix_dgram_sendmsg(struct socket *, struct msghdr *, size_t);
static int unix_dgram_recvmsg(struct socket *, struct msghdr *, size_t, int);
L
Linus Torvalds 已提交
655 656
static int unix_dgram_connect(struct socket *, struct sockaddr *,
			      int, int);
657 658 659
static int unix_seqpacket_sendmsg(struct socket *, struct msghdr *, size_t);
static int unix_seqpacket_recvmsg(struct socket *, struct msghdr *, size_t,
				  int);
L
Linus Torvalds 已提交
660

661
static int unix_set_peek_off(struct sock *sk, int val)
662 663 664
{
	struct unix_sock *u = unix_sk(sk);

665
	if (mutex_lock_interruptible(&u->iolock))
666 667
		return -EINTR;

668
	sk->sk_peek_off = val;
669
	mutex_unlock(&u->iolock);
670 671

	return 0;
672 673 674
}


675
static const struct proto_ops unix_stream_ops = {
L
Linus Torvalds 已提交
676 677 678 679 680 681 682 683 684 685 686 687 688 689 690 691 692
	.family =	PF_UNIX,
	.owner =	THIS_MODULE,
	.release =	unix_release,
	.bind =		unix_bind,
	.connect =	unix_stream_connect,
	.socketpair =	unix_socketpair,
	.accept =	unix_accept,
	.getname =	unix_getname,
	.poll =		unix_poll,
	.ioctl =	unix_ioctl,
	.listen =	unix_listen,
	.shutdown =	unix_shutdown,
	.setsockopt =	sock_no_setsockopt,
	.getsockopt =	sock_no_getsockopt,
	.sendmsg =	unix_stream_sendmsg,
	.recvmsg =	unix_stream_recvmsg,
	.mmap =		sock_no_mmap,
693
	.sendpage =	unix_stream_sendpage,
694
	.splice_read =	unix_stream_splice_read,
695
	.set_peek_off =	unix_set_peek_off,
L
Linus Torvalds 已提交
696 697
};

698
static const struct proto_ops unix_dgram_ops = {
L
Linus Torvalds 已提交
699 700 701 702 703 704 705 706
	.family =	PF_UNIX,
	.owner =	THIS_MODULE,
	.release =	unix_release,
	.bind =		unix_bind,
	.connect =	unix_dgram_connect,
	.socketpair =	unix_socketpair,
	.accept =	sock_no_accept,
	.getname =	unix_getname,
707
	.poll =		unix_dgram_poll,
L
Linus Torvalds 已提交
708 709 710 711 712 713 714 715 716
	.ioctl =	unix_ioctl,
	.listen =	sock_no_listen,
	.shutdown =	unix_shutdown,
	.setsockopt =	sock_no_setsockopt,
	.getsockopt =	sock_no_getsockopt,
	.sendmsg =	unix_dgram_sendmsg,
	.recvmsg =	unix_dgram_recvmsg,
	.mmap =		sock_no_mmap,
	.sendpage =	sock_no_sendpage,
717
	.set_peek_off =	unix_set_peek_off,
L
Linus Torvalds 已提交
718 719
};

720
static const struct proto_ops unix_seqpacket_ops = {
L
Linus Torvalds 已提交
721 722 723 724 725 726 727 728
	.family =	PF_UNIX,
	.owner =	THIS_MODULE,
	.release =	unix_release,
	.bind =		unix_bind,
	.connect =	unix_stream_connect,
	.socketpair =	unix_socketpair,
	.accept =	unix_accept,
	.getname =	unix_getname,
729
	.poll =		unix_dgram_poll,
L
Linus Torvalds 已提交
730 731 732 733 734 735
	.ioctl =	unix_ioctl,
	.listen =	unix_listen,
	.shutdown =	unix_shutdown,
	.setsockopt =	sock_no_setsockopt,
	.getsockopt =	sock_no_getsockopt,
	.sendmsg =	unix_seqpacket_sendmsg,
736
	.recvmsg =	unix_seqpacket_recvmsg,
L
Linus Torvalds 已提交
737 738
	.mmap =		sock_no_mmap,
	.sendpage =	sock_no_sendpage,
739
	.set_peek_off =	unix_set_peek_off,
L
Linus Torvalds 已提交
740 741 742
};

static struct proto unix_proto = {
743 744 745
	.name			= "UNIX",
	.owner			= THIS_MODULE,
	.obj_size		= sizeof(struct unix_sock),
L
Linus Torvalds 已提交
746 747
};

748 749 750 751 752 753 754 755
/*
 * AF_UNIX sockets do not interact with hardware, hence they
 * dont trigger interrupts - so it's safe for them to have
 * bh-unsafe locking for their sk_receive_queue.lock. Split off
 * this special lock-class by reinitializing the spinlock key:
 */
static struct lock_class_key af_unix_sk_receive_queue_lock_key;

756
static struct sock *unix_create1(struct net *net, struct socket *sock, int kern)
L
Linus Torvalds 已提交
757 758 759 760
{
	struct sock *sk = NULL;
	struct unix_sock *u;

E
Eric Dumazet 已提交
761 762
	atomic_long_inc(&unix_nr_socks);
	if (atomic_long_read(&unix_nr_socks) > 2 * get_max_files())
L
Linus Torvalds 已提交
763 764
		goto out;

765
	sk = sk_alloc(net, PF_UNIX, GFP_KERNEL, &unix_proto, kern);
L
Linus Torvalds 已提交
766 767 768
	if (!sk)
		goto out;

E
Eric Dumazet 已提交
769
	sock_init_data(sock, sk);
770 771
	lockdep_set_class(&sk->sk_receive_queue.lock,
				&af_unix_sk_receive_queue_lock_key);
L
Linus Torvalds 已提交
772

773
	sk->sk_allocation	= GFP_KERNEL_ACCOUNT;
L
Linus Torvalds 已提交
774
	sk->sk_write_space	= unix_write_space;
775
	sk->sk_max_ack_backlog	= net->unx.sysctl_max_dgram_qlen;
L
Linus Torvalds 已提交
776 777
	sk->sk_destruct		= unix_sock_destructor;
	u	  = unix_sk(sk);
A
Al Viro 已提交
778 779
	u->path.dentry = NULL;
	u->path.mnt = NULL;
780
	spin_lock_init(&u->lock);
A
Al Viro 已提交
781
	atomic_long_set(&u->inflight, 0);
782
	INIT_LIST_HEAD(&u->link);
783 784
	mutex_init(&u->iolock); /* single task reading lock */
	mutex_init(&u->bindlock); /* single task binding lock */
L
Linus Torvalds 已提交
785
	init_waitqueue_head(&u->peer_wait);
786
	init_waitqueue_func_entry(&u->peer_wake, unix_dgram_peer_wake_relay);
E
Eric Dumazet 已提交
787
	unix_insert_socket(unix_sockets_unbound(sk), sk);
L
Linus Torvalds 已提交
788
out:
789
	if (sk == NULL)
E
Eric Dumazet 已提交
790
		atomic_long_dec(&unix_nr_socks);
791 792
	else {
		local_bh_disable();
793
		sock_prot_inuse_add(sock_net(sk), sk->sk_prot, 1);
794 795
		local_bh_enable();
	}
L
Linus Torvalds 已提交
796 797 798
	return sk;
}

799 800
static int unix_create(struct net *net, struct socket *sock, int protocol,
		       int kern)
L
Linus Torvalds 已提交
801 802 803 804 805 806 807 808 809 810 811 812 813 814 815
{
	if (protocol && protocol != PF_UNIX)
		return -EPROTONOSUPPORT;

	sock->state = SS_UNCONNECTED;

	switch (sock->type) {
	case SOCK_STREAM:
		sock->ops = &unix_stream_ops;
		break;
		/*
		 *	Believe it or not BSD has AF_UNIX, SOCK_RAW though
		 *	nothing uses it.
		 */
	case SOCK_RAW:
816
		sock->type = SOCK_DGRAM;
L
Linus Torvalds 已提交
817 818 819 820 821 822 823 824 825 826
	case SOCK_DGRAM:
		sock->ops = &unix_dgram_ops;
		break;
	case SOCK_SEQPACKET:
		sock->ops = &unix_seqpacket_ops;
		break;
	default:
		return -ESOCKTNOSUPPORT;
	}

827
	return unix_create1(net, sock, kern) ? 0 : -ENOMEM;
L
Linus Torvalds 已提交
828 829 830 831 832 833 834 835 836
}

static int unix_release(struct socket *sock)
{
	struct sock *sk = sock->sk;

	if (!sk)
		return 0;

837
	unix_release_sock(sk, 0);
L
Linus Torvalds 已提交
838 839
	sock->sk = NULL;

840
	return 0;
L
Linus Torvalds 已提交
841 842 843 844 845
}

static int unix_autobind(struct socket *sock)
{
	struct sock *sk = sock->sk;
846
	struct net *net = sock_net(sk);
L
Linus Torvalds 已提交
847 848
	struct unix_sock *u = unix_sk(sk);
	static u32 ordernum = 1;
E
Eric Dumazet 已提交
849
	struct unix_address *addr;
L
Linus Torvalds 已提交
850
	int err;
851
	unsigned int retries = 0;
L
Linus Torvalds 已提交
852

853
	err = mutex_lock_interruptible(&u->bindlock);
854 855
	if (err)
		return err;
L
Linus Torvalds 已提交
856 857 858 859 860 861

	err = 0;
	if (u->addr)
		goto out;

	err = -ENOMEM;
862
	addr = kzalloc(sizeof(*addr) + sizeof(short) + 16, GFP_KERNEL);
L
Linus Torvalds 已提交
863 864 865 866 867 868 869 870
	if (!addr)
		goto out;

	addr->name->sun_family = AF_UNIX;
	atomic_set(&addr->refcnt, 1);

retry:
	addr->len = sprintf(addr->name->sun_path+1, "%05x", ordernum) + 1 + sizeof(short);
871
	addr->hash = unix_hash_fold(csum_partial(addr->name, addr->len, 0));
L
Linus Torvalds 已提交
872

873
	spin_lock(&unix_table_lock);
L
Linus Torvalds 已提交
874 875
	ordernum = (ordernum+1)&0xFFFFF;

876
	if (__unix_find_socket_byname(net, addr->name, addr->len, sock->type,
L
Linus Torvalds 已提交
877
				      addr->hash)) {
878
		spin_unlock(&unix_table_lock);
879 880 881 882 883 884 885 886 887 888 889
		/*
		 * __unix_find_socket_byname() may take long time if many names
		 * are already in use.
		 */
		cond_resched();
		/* Give up if all names seems to be in use. */
		if (retries++ == 0xFFFFF) {
			err = -ENOSPC;
			kfree(addr);
			goto out;
		}
L
Linus Torvalds 已提交
890 891 892 893 894 895 896
		goto retry;
	}
	addr->hash ^= sk->sk_type;

	__unix_remove_socket(sk);
	u->addr = addr;
	__unix_insert_socket(&unix_socket_table[addr->hash], sk);
897
	spin_unlock(&unix_table_lock);
L
Linus Torvalds 已提交
898 899
	err = 0;

900
out:	mutex_unlock(&u->bindlock);
L
Linus Torvalds 已提交
901 902 903
	return err;
}

904 905
static struct sock *unix_find_other(struct net *net,
				    struct sockaddr_un *sunname, int len,
906
				    int type, unsigned int hash, int *error)
L
Linus Torvalds 已提交
907 908
{
	struct sock *u;
909
	struct path path;
L
Linus Torvalds 已提交
910
	int err = 0;
911

L
Linus Torvalds 已提交
912
	if (sunname->sun_path[0]) {
913 914
		struct inode *inode;
		err = kern_path(sunname->sun_path, LOOKUP_FOLLOW, &path);
L
Linus Torvalds 已提交
915 916
		if (err)
			goto fail;
917
		inode = d_backing_inode(path.dentry);
918
		err = inode_permission(inode, MAY_WRITE);
L
Linus Torvalds 已提交
919 920 921 922
		if (err)
			goto put_fail;

		err = -ECONNREFUSED;
923
		if (!S_ISSOCK(inode->i_mode))
L
Linus Torvalds 已提交
924
			goto put_fail;
925
		u = unix_find_socket_byinode(inode);
L
Linus Torvalds 已提交
926 927 928 929
		if (!u)
			goto put_fail;

		if (u->sk_type == type)
A
Al Viro 已提交
930
			touch_atime(&path);
L
Linus Torvalds 已提交
931

932
		path_put(&path);
L
Linus Torvalds 已提交
933

934
		err = -EPROTOTYPE;
L
Linus Torvalds 已提交
935 936 937 938 939 940
		if (u->sk_type != type) {
			sock_put(u);
			goto fail;
		}
	} else {
		err = -ECONNREFUSED;
941
		u = unix_find_socket_byname(net, sunname, len, type, hash);
L
Linus Torvalds 已提交
942 943
		if (u) {
			struct dentry *dentry;
A
Al Viro 已提交
944
			dentry = unix_sk(u)->path.dentry;
L
Linus Torvalds 已提交
945
			if (dentry)
A
Al Viro 已提交
946
				touch_atime(&unix_sk(u)->path);
L
Linus Torvalds 已提交
947 948 949 950 951 952
		} else
			goto fail;
	}
	return u;

put_fail:
953
	path_put(&path);
L
Linus Torvalds 已提交
954
fail:
955
	*error = err;
L
Linus Torvalds 已提交
956 957 958
	return NULL;
}

959
static int unix_mknod(const char *sun_path, umode_t mode, struct path *res)
A
Al Viro 已提交
960
{
961 962 963 964 965 966 967 968 969 970 971
	struct dentry *dentry;
	struct path path;
	int err = 0;
	/*
	 * Get the parent directory, calculate the hash for last
	 * component.
	 */
	dentry = kern_path_create(AT_FDCWD, sun_path, &path, 0);
	err = PTR_ERR(dentry);
	if (IS_ERR(dentry))
		return err;
A
Al Viro 已提交
972

973 974 975 976
	/*
	 * All right, let's create it.
	 */
	err = security_path_mknod(&path, dentry, mode, 0);
A
Al Viro 已提交
977
	if (!err) {
978
		err = vfs_mknod(d_inode(path.dentry), dentry, mode, 0);
A
Al Viro 已提交
979
		if (!err) {
980
			res->mnt = mntget(path.mnt);
A
Al Viro 已提交
981 982 983
			res->dentry = dget(dentry);
		}
	}
984
	done_path_create(&path, dentry);
A
Al Viro 已提交
985 986
	return err;
}
L
Linus Torvalds 已提交
987 988 989 990

static int unix_bind(struct socket *sock, struct sockaddr *uaddr, int addr_len)
{
	struct sock *sk = sock->sk;
991
	struct net *net = sock_net(sk);
L
Linus Torvalds 已提交
992
	struct unix_sock *u = unix_sk(sk);
993
	struct sockaddr_un *sunaddr = (struct sockaddr_un *)uaddr;
994
	char *sun_path = sunaddr->sun_path;
995
	int err;
996
	unsigned int hash;
L
Linus Torvalds 已提交
997 998
	struct unix_address *addr;
	struct hlist_head *list;
K
Kees Cook 已提交
999
	struct path path = { };
L
Linus Torvalds 已提交
1000 1001 1002 1003 1004

	err = -EINVAL;
	if (sunaddr->sun_family != AF_UNIX)
		goto out;

1005
	if (addr_len == sizeof(short)) {
L
Linus Torvalds 已提交
1006 1007 1008 1009 1010 1011 1012 1013 1014
		err = unix_autobind(sock);
		goto out;
	}

	err = unix_mkname(sunaddr, addr_len, &hash);
	if (err < 0)
		goto out;
	addr_len = err;

1015 1016 1017 1018 1019 1020 1021 1022 1023 1024 1025
	if (sun_path[0]) {
		umode_t mode = S_IFSOCK |
		       (SOCK_INODE(sock)->i_mode & ~current_umask());
		err = unix_mknod(sun_path, mode, &path);
		if (err) {
			if (err == -EEXIST)
				err = -EADDRINUSE;
			goto out;
		}
	}

1026
	err = mutex_lock_interruptible(&u->bindlock);
1027
	if (err)
1028
		goto out_put;
L
Linus Torvalds 已提交
1029 1030 1031 1032 1033 1034 1035 1036 1037 1038 1039 1040 1041 1042 1043

	err = -EINVAL;
	if (u->addr)
		goto out_up;

	err = -ENOMEM;
	addr = kmalloc(sizeof(*addr)+addr_len, GFP_KERNEL);
	if (!addr)
		goto out_up;

	memcpy(addr->name, sunaddr, addr_len);
	addr->len = addr_len;
	addr->hash = hash ^ sk->sk_type;
	atomic_set(&addr->refcnt, 1);

1044
	if (sun_path[0]) {
L
Linus Torvalds 已提交
1045
		addr->hash = UNIX_HASH_SIZE;
1046
		hash = d_backing_inode(path.dentry)->i_ino & (UNIX_HASH_SIZE - 1);
A
Al Viro 已提交
1047
		spin_lock(&unix_table_lock);
1048
		u->path = path;
A
Al Viro 已提交
1049 1050 1051
		list = &unix_socket_table[hash];
	} else {
		spin_lock(&unix_table_lock);
L
Linus Torvalds 已提交
1052
		err = -EADDRINUSE;
1053
		if (__unix_find_socket_byname(net, sunaddr, addr_len,
L
Linus Torvalds 已提交
1054 1055 1056 1057 1058 1059 1060 1061 1062 1063 1064 1065 1066 1067
					      sk->sk_type, hash)) {
			unix_release_addr(addr);
			goto out_unlock;
		}

		list = &unix_socket_table[addr->hash];
	}

	err = 0;
	__unix_remove_socket(sk);
	u->addr = addr;
	__unix_insert_socket(list, sk);

out_unlock:
1068
	spin_unlock(&unix_table_lock);
L
Linus Torvalds 已提交
1069
out_up:
1070
	mutex_unlock(&u->bindlock);
1071 1072 1073
out_put:
	if (err)
		path_put(&path);
L
Linus Torvalds 已提交
1074 1075 1076 1077
out:
	return err;
}

1078 1079 1080 1081 1082 1083 1084 1085 1086 1087 1088 1089 1090 1091 1092 1093 1094 1095 1096 1097 1098 1099 1100 1101 1102
static void unix_state_double_lock(struct sock *sk1, struct sock *sk2)
{
	if (unlikely(sk1 == sk2) || !sk2) {
		unix_state_lock(sk1);
		return;
	}
	if (sk1 < sk2) {
		unix_state_lock(sk1);
		unix_state_lock_nested(sk2);
	} else {
		unix_state_lock(sk2);
		unix_state_lock_nested(sk1);
	}
}

static void unix_state_double_unlock(struct sock *sk1, struct sock *sk2)
{
	if (unlikely(sk1 == sk2) || !sk2) {
		unix_state_unlock(sk1);
		return;
	}
	unix_state_unlock(sk1);
	unix_state_unlock(sk2);
}

L
Linus Torvalds 已提交
1103 1104 1105 1106
static int unix_dgram_connect(struct socket *sock, struct sockaddr *addr,
			      int alen, int flags)
{
	struct sock *sk = sock->sk;
1107
	struct net *net = sock_net(sk);
1108
	struct sockaddr_un *sunaddr = (struct sockaddr_un *)addr;
L
Linus Torvalds 已提交
1109
	struct sock *other;
1110
	unsigned int hash;
L
Linus Torvalds 已提交
1111 1112 1113 1114 1115 1116 1117 1118 1119 1120 1121 1122
	int err;

	if (addr->sa_family != AF_UNSPEC) {
		err = unix_mkname(sunaddr, alen, &hash);
		if (err < 0)
			goto out;
		alen = err;

		if (test_bit(SOCK_PASSCRED, &sock->flags) &&
		    !unix_sk(sk)->addr && (err = unix_autobind(sock)) != 0)
			goto out;

1123
restart:
1124
		other = unix_find_other(net, sunaddr, alen, sock->type, hash, &err);
L
Linus Torvalds 已提交
1125 1126 1127
		if (!other)
			goto out;

1128 1129 1130 1131 1132 1133 1134 1135
		unix_state_double_lock(sk, other);

		/* Apparently VFS overslept socket death. Retry. */
		if (sock_flag(other, SOCK_DEAD)) {
			unix_state_double_unlock(sk, other);
			sock_put(other);
			goto restart;
		}
L
Linus Torvalds 已提交
1136 1137 1138 1139 1140 1141 1142 1143 1144 1145 1146 1147 1148 1149

		err = -EPERM;
		if (!unix_may_send(sk, other))
			goto out_unlock;

		err = security_unix_may_send(sk->sk_socket, other->sk_socket);
		if (err)
			goto out_unlock;

	} else {
		/*
		 *	1003.1g breaking connected state with AF_UNSPEC
		 */
		other = NULL;
1150
		unix_state_double_lock(sk, other);
L
Linus Torvalds 已提交
1151 1152 1153 1154 1155 1156 1157
	}

	/*
	 * If it was connected, reconnect.
	 */
	if (unix_peer(sk)) {
		struct sock *old_peer = unix_peer(sk);
1158
		unix_peer(sk) = other;
1159 1160
		unix_dgram_peer_wake_disconnect_wakeup(sk, old_peer);

1161
		unix_state_double_unlock(sk, other);
L
Linus Torvalds 已提交
1162 1163 1164 1165 1166

		if (other != old_peer)
			unix_dgram_disconnected(sk, old_peer);
		sock_put(old_peer);
	} else {
1167
		unix_peer(sk) = other;
1168
		unix_state_double_unlock(sk, other);
L
Linus Torvalds 已提交
1169
	}
1170
	return 0;
L
Linus Torvalds 已提交
1171 1172

out_unlock:
1173
	unix_state_double_unlock(sk, other);
L
Linus Torvalds 已提交
1174 1175 1176 1177 1178 1179 1180 1181 1182 1183 1184 1185 1186 1187 1188
	sock_put(other);
out:
	return err;
}

static long unix_wait_for_peer(struct sock *other, long timeo)
{
	struct unix_sock *u = unix_sk(other);
	int sched;
	DEFINE_WAIT(wait);

	prepare_to_wait_exclusive(&u->peer_wait, &wait, TASK_INTERRUPTIBLE);

	sched = !sock_flag(other, SOCK_DEAD) &&
		!(other->sk_shutdown & RCV_SHUTDOWN) &&
1189
		unix_recvq_full(other);
L
Linus Torvalds 已提交
1190

1191
	unix_state_unlock(other);
L
Linus Torvalds 已提交
1192 1193 1194 1195 1196 1197 1198 1199 1200 1201 1202

	if (sched)
		timeo = schedule_timeout(timeo);

	finish_wait(&u->peer_wait, &wait);
	return timeo;
}

static int unix_stream_connect(struct socket *sock, struct sockaddr *uaddr,
			       int addr_len, int flags)
{
1203
	struct sockaddr_un *sunaddr = (struct sockaddr_un *)uaddr;
L
Linus Torvalds 已提交
1204
	struct sock *sk = sock->sk;
1205
	struct net *net = sock_net(sk);
L
Linus Torvalds 已提交
1206 1207 1208 1209
	struct unix_sock *u = unix_sk(sk), *newu, *otheru;
	struct sock *newsk = NULL;
	struct sock *other = NULL;
	struct sk_buff *skb = NULL;
1210
	unsigned int hash;
L
Linus Torvalds 已提交
1211 1212 1213 1214 1215 1216 1217 1218 1219
	int st;
	int err;
	long timeo;

	err = unix_mkname(sunaddr, addr_len, &hash);
	if (err < 0)
		goto out;
	addr_len = err;

1220 1221
	if (test_bit(SOCK_PASSCRED, &sock->flags) && !u->addr &&
	    (err = unix_autobind(sock)) != 0)
L
Linus Torvalds 已提交
1222 1223 1224 1225 1226 1227 1228 1229 1230 1231 1232 1233
		goto out;

	timeo = sock_sndtimeo(sk, flags & O_NONBLOCK);

	/* First of all allocate resources.
	   If we will make it after state is locked,
	   we will have to recheck all again in any case.
	 */

	err = -ENOMEM;

	/* create new sock for complete connection */
1234
	newsk = unix_create1(sock_net(sk), NULL, 0);
L
Linus Torvalds 已提交
1235 1236 1237 1238 1239 1240 1241 1242 1243 1244
	if (newsk == NULL)
		goto out;

	/* Allocate skb for sending to listening sock */
	skb = sock_wmalloc(newsk, 1, 0, GFP_KERNEL);
	if (skb == NULL)
		goto out;

restart:
	/*  Find listening sock. */
1245
	other = unix_find_other(net, sunaddr, addr_len, sk->sk_type, hash, &err);
L
Linus Torvalds 已提交
1246 1247 1248 1249
	if (!other)
		goto out;

	/* Latch state of peer */
1250
	unix_state_lock(other);
L
Linus Torvalds 已提交
1251 1252 1253

	/* Apparently VFS overslept socket death. Retry. */
	if (sock_flag(other, SOCK_DEAD)) {
1254
		unix_state_unlock(other);
L
Linus Torvalds 已提交
1255 1256 1257 1258 1259 1260 1261
		sock_put(other);
		goto restart;
	}

	err = -ECONNREFUSED;
	if (other->sk_state != TCP_LISTEN)
		goto out_unlock;
1262 1263
	if (other->sk_shutdown & RCV_SHUTDOWN)
		goto out_unlock;
L
Linus Torvalds 已提交
1264

1265
	if (unix_recvq_full(other)) {
L
Linus Torvalds 已提交
1266 1267 1268 1269 1270 1271 1272 1273 1274 1275 1276
		err = -EAGAIN;
		if (!timeo)
			goto out_unlock;

		timeo = unix_wait_for_peer(other, timeo);

		err = sock_intr_errno(timeo);
		if (signal_pending(current))
			goto out;
		sock_put(other);
		goto restart;
1277
	}
L
Linus Torvalds 已提交
1278 1279 1280

	/* Latch our state.

D
Daniel Baluta 已提交
1281
	   It is tricky place. We need to grab our state lock and cannot
L
Linus Torvalds 已提交
1282 1283 1284 1285 1286 1287 1288 1289 1290 1291 1292 1293 1294 1295 1296 1297 1298 1299 1300 1301 1302 1303 1304
	   drop lock on peer. It is dangerous because deadlock is
	   possible. Connect to self case and simultaneous
	   attempt to connect are eliminated by checking socket
	   state. other is TCP_LISTEN, if sk is TCP_LISTEN we
	   check this before attempt to grab lock.

	   Well, and we have to recheck the state after socket locked.
	 */
	st = sk->sk_state;

	switch (st) {
	case TCP_CLOSE:
		/* This is ok... continue with connect */
		break;
	case TCP_ESTABLISHED:
		/* Socket is already connected */
		err = -EISCONN;
		goto out_unlock;
	default:
		err = -EINVAL;
		goto out_unlock;
	}

1305
	unix_state_lock_nested(sk);
L
Linus Torvalds 已提交
1306 1307

	if (sk->sk_state != st) {
1308 1309
		unix_state_unlock(sk);
		unix_state_unlock(other);
L
Linus Torvalds 已提交
1310 1311 1312 1313
		sock_put(other);
		goto restart;
	}

1314
	err = security_unix_stream_connect(sk, other, newsk);
L
Linus Torvalds 已提交
1315
	if (err) {
1316
		unix_state_unlock(sk);
L
Linus Torvalds 已提交
1317 1318 1319 1320 1321 1322 1323 1324 1325
		goto out_unlock;
	}

	/* The way is open! Fastly set all the necessary fields... */

	sock_hold(sk);
	unix_peer(newsk)	= sk;
	newsk->sk_state		= TCP_ESTABLISHED;
	newsk->sk_type		= sk->sk_type;
1326
	init_peercred(newsk);
L
Linus Torvalds 已提交
1327
	newu = unix_sk(newsk);
1328
	RCU_INIT_POINTER(newsk->sk_wq, &newu->peer_wq);
L
Linus Torvalds 已提交
1329 1330 1331 1332 1333 1334 1335
	otheru = unix_sk(other);

	/* copy address information from listening to new sock*/
	if (otheru->addr) {
		atomic_inc(&otheru->addr->refcnt);
		newu->addr = otheru->addr;
	}
A
Al Viro 已提交
1336 1337 1338
	if (otheru->path.dentry) {
		path_get(&otheru->path);
		newu->path = otheru->path;
L
Linus Torvalds 已提交
1339 1340 1341
	}

	/* Set credentials */
1342
	copy_peercred(sk, other);
L
Linus Torvalds 已提交
1343 1344 1345

	sock->state	= SS_CONNECTED;
	sk->sk_state	= TCP_ESTABLISHED;
1346 1347
	sock_hold(newsk);

1348
	smp_mb__after_atomic();	/* sock_hold() does an atomic_inc() */
1349
	unix_peer(sk)	= newsk;
L
Linus Torvalds 已提交
1350

1351
	unix_state_unlock(sk);
L
Linus Torvalds 已提交
1352 1353 1354 1355 1356

	/* take ten and and send info to listening sock */
	spin_lock(&other->sk_receive_queue.lock);
	__skb_queue_tail(&other->sk_receive_queue, skb);
	spin_unlock(&other->sk_receive_queue.lock);
1357
	unix_state_unlock(other);
1358
	other->sk_data_ready(other);
L
Linus Torvalds 已提交
1359 1360 1361 1362 1363
	sock_put(other);
	return 0;

out_unlock:
	if (other)
1364
		unix_state_unlock(other);
L
Linus Torvalds 已提交
1365 1366

out:
1367
	kfree_skb(skb);
L
Linus Torvalds 已提交
1368 1369 1370 1371 1372 1373 1374 1375 1376
	if (newsk)
		unix_release_sock(newsk, 0);
	if (other)
		sock_put(other);
	return err;
}

static int unix_socketpair(struct socket *socka, struct socket *sockb)
{
1377
	struct sock *ska = socka->sk, *skb = sockb->sk;
L
Linus Torvalds 已提交
1378 1379 1380 1381

	/* Join our sockets back to back */
	sock_hold(ska);
	sock_hold(skb);
1382 1383
	unix_peer(ska) = skb;
	unix_peer(skb) = ska;
1384 1385
	init_peercred(ska);
	init_peercred(skb);
L
Linus Torvalds 已提交
1386 1387 1388 1389 1390 1391 1392 1393 1394 1395

	if (ska->sk_type != SOCK_DGRAM) {
		ska->sk_state = TCP_ESTABLISHED;
		skb->sk_state = TCP_ESTABLISHED;
		socka->state  = SS_CONNECTED;
		sockb->state  = SS_CONNECTED;
	}
	return 0;
}

1396 1397 1398 1399 1400 1401 1402 1403 1404
static void unix_sock_inherit_flags(const struct socket *old,
				    struct socket *new)
{
	if (test_bit(SOCK_PASSCRED, &old->flags))
		set_bit(SOCK_PASSCRED, &new->flags);
	if (test_bit(SOCK_PASSSEC, &old->flags))
		set_bit(SOCK_PASSSEC, &new->flags);
}

1405 1406
static int unix_accept(struct socket *sock, struct socket *newsock, int flags,
		       bool kern)
L
Linus Torvalds 已提交
1407 1408 1409 1410 1411 1412 1413
{
	struct sock *sk = sock->sk;
	struct sock *tsk;
	struct sk_buff *skb;
	int err;

	err = -EOPNOTSUPP;
E
Eric Dumazet 已提交
1414
	if (sock->type != SOCK_STREAM && sock->type != SOCK_SEQPACKET)
L
Linus Torvalds 已提交
1415 1416 1417 1418 1419 1420 1421 1422 1423 1424 1425 1426 1427 1428 1429 1430 1431 1432 1433 1434 1435 1436 1437
		goto out;

	err = -EINVAL;
	if (sk->sk_state != TCP_LISTEN)
		goto out;

	/* If socket state is TCP_LISTEN it cannot change (for now...),
	 * so that no locks are necessary.
	 */

	skb = skb_recv_datagram(sk, 0, flags&O_NONBLOCK, &err);
	if (!skb) {
		/* This means receive shutdown. */
		if (err == 0)
			err = -EINVAL;
		goto out;
	}

	tsk = skb->sk;
	skb_free_datagram(sk, skb);
	wake_up_interruptible(&unix_sk(sk)->peer_wait);

	/* attach accepted sock to socket */
1438
	unix_state_lock(tsk);
L
Linus Torvalds 已提交
1439
	newsock->state = SS_CONNECTED;
1440
	unix_sock_inherit_flags(sock, newsock);
L
Linus Torvalds 已提交
1441
	sock_graft(tsk, newsock);
1442
	unix_state_unlock(tsk);
L
Linus Torvalds 已提交
1443 1444 1445 1446 1447 1448 1449 1450 1451 1452 1453
	return 0;

out:
	return err;
}


static int unix_getname(struct socket *sock, struct sockaddr *uaddr, int *uaddr_len, int peer)
{
	struct sock *sk = sock->sk;
	struct unix_sock *u;
1454
	DECLARE_SOCKADDR(struct sockaddr_un *, sunaddr, uaddr);
L
Linus Torvalds 已提交
1455 1456 1457 1458 1459 1460 1461 1462 1463 1464 1465 1466 1467 1468
	int err = 0;

	if (peer) {
		sk = unix_peer_get(sk);

		err = -ENOTCONN;
		if (!sk)
			goto out;
		err = 0;
	} else {
		sock_hold(sk);
	}

	u = unix_sk(sk);
1469
	unix_state_lock(sk);
L
Linus Torvalds 已提交
1470 1471 1472 1473 1474 1475 1476 1477 1478 1479
	if (!u->addr) {
		sunaddr->sun_family = AF_UNIX;
		sunaddr->sun_path[0] = 0;
		*uaddr_len = sizeof(short);
	} else {
		struct unix_address *addr = u->addr;

		*uaddr_len = addr->len;
		memcpy(sunaddr, addr->name, *uaddr_len);
	}
1480
	unix_state_unlock(sk);
L
Linus Torvalds 已提交
1481 1482 1483 1484 1485 1486 1487 1488 1489 1490 1491 1492
	sock_put(sk);
out:
	return err;
}

static void unix_detach_fds(struct scm_cookie *scm, struct sk_buff *skb)
{
	int i;

	scm->fp = UNIXCB(skb).fp;
	UNIXCB(skb).fp = NULL;

E
Eric Dumazet 已提交
1493
	for (i = scm->fp->count-1; i >= 0; i--)
1494
		unix_notinflight(scm->fp->user, scm->fp->fp[i]);
L
Linus Torvalds 已提交
1495 1496
}

1497
static void unix_destruct_scm(struct sk_buff *skb)
L
Linus Torvalds 已提交
1498 1499 1500
{
	struct scm_cookie scm;
	memset(&scm, 0, sizeof(scm));
1501 1502 1503
	scm.pid  = UNIXCB(skb).pid;
	if (UNIXCB(skb).fp)
		unix_detach_fds(&scm, skb);
L
Linus Torvalds 已提交
1504 1505 1506 1507 1508 1509 1510

	/* Alas, it calls VFS */
	/* So fscking what? fput() had been SMP-safe since the last Summer */
	scm_destroy(&scm);
	sock_wfree(skb);
}

1511 1512 1513 1514 1515 1516 1517 1518 1519 1520 1521 1522 1523 1524 1525
/*
 * The "user->unix_inflight" variable is protected by the garbage
 * collection lock, and we just read it locklessly here. If you go
 * over the limit, there might be a tiny race in actually noticing
 * it across threads. Tough.
 */
static inline bool too_many_unix_fds(struct task_struct *p)
{
	struct user_struct *user = current_user();

	if (unlikely(user->unix_inflight > task_rlimit(p, RLIMIT_NOFILE)))
		return !capable(CAP_SYS_RESOURCE) && !capable(CAP_SYS_ADMIN);
	return false;
}

E
Eric Dumazet 已提交
1526 1527
#define MAX_RECURSION_LEVEL 4

1528
static int unix_attach_fds(struct scm_cookie *scm, struct sk_buff *skb)
L
Linus Torvalds 已提交
1529 1530
{
	int i;
E
Eric Dumazet 已提交
1531 1532
	unsigned char max_level = 0;

1533 1534 1535
	if (too_many_unix_fds(current))
		return -ETOOMANYREFS;

E
Eric Dumazet 已提交
1536 1537 1538
	for (i = scm->fp->count - 1; i >= 0; i--) {
		struct sock *sk = unix_get_socket(scm->fp->fp[i]);

1539
		if (sk)
E
Eric Dumazet 已提交
1540 1541 1542 1543 1544
			max_level = max(max_level,
					unix_sk(sk)->recursion_level);
	}
	if (unlikely(max_level > MAX_RECURSION_LEVEL))
		return -ETOOMANYREFS;
1545 1546 1547 1548 1549 1550 1551 1552 1553 1554

	/*
	 * Need to duplicate file references for the sake of garbage
	 * collection.  Otherwise a socket in the fps might become a
	 * candidate for GC while the skb is not yet queued.
	 */
	UNIXCB(skb).fp = scm_fp_dup(scm->fp);
	if (!UNIXCB(skb).fp)
		return -ENOMEM;

1555
	for (i = scm->fp->count - 1; i >= 0; i--)
1556
		unix_inflight(scm->fp->user, scm->fp->fp[i]);
E
Eric Dumazet 已提交
1557
	return max_level;
L
Linus Torvalds 已提交
1558 1559
}

1560
static int unix_scm_to_skb(struct scm_cookie *scm, struct sk_buff *skb, bool send_fds)
1561 1562
{
	int err = 0;
1563

1564
	UNIXCB(skb).pid  = get_pid(scm->pid);
1565 1566
	UNIXCB(skb).uid = scm->creds.uid;
	UNIXCB(skb).gid = scm->creds.gid;
1567
	UNIXCB(skb).fp = NULL;
1568
	unix_get_secdata(scm, skb);
1569 1570 1571 1572 1573 1574 1575
	if (scm->fp && send_fds)
		err = unix_attach_fds(scm, skb);

	skb->destructor = unix_destruct_scm;
	return err;
}

1576 1577 1578 1579 1580 1581 1582 1583
static bool unix_passcred_enabled(const struct socket *sock,
				  const struct sock *other)
{
	return test_bit(SOCK_PASSCRED, &sock->flags) ||
	       !other->sk_socket ||
	       test_bit(SOCK_PASSCRED, &other->sk_socket->flags);
}

1584 1585 1586 1587 1588 1589 1590 1591
/*
 * Some apps rely on write() giving SCM_CREDENTIALS
 * We include credentials if source or destination socket
 * asserted SOCK_PASSCRED.
 */
static void maybe_add_creds(struct sk_buff *skb, const struct socket *sock,
			    const struct sock *other)
{
1592
	if (UNIXCB(skb).pid)
1593
		return;
1594
	if (unix_passcred_enabled(sock, other)) {
1595
		UNIXCB(skb).pid  = get_pid(task_tgid(current));
1596
		current_uid_gid(&UNIXCB(skb).uid, &UNIXCB(skb).gid);
1597 1598 1599
	}
}

1600 1601 1602 1603 1604 1605 1606 1607 1608 1609 1610 1611 1612 1613 1614 1615 1616 1617 1618 1619 1620 1621 1622 1623 1624 1625 1626 1627 1628
static int maybe_init_creds(struct scm_cookie *scm,
			    struct socket *socket,
			    const struct sock *other)
{
	int err;
	struct msghdr msg = { .msg_controllen = 0 };

	err = scm_send(socket, &msg, scm, false);
	if (err)
		return err;

	if (unix_passcred_enabled(socket, other)) {
		scm->pid = get_pid(task_tgid(current));
		current_uid_gid(&scm->creds.uid, &scm->creds.gid);
	}
	return err;
}

static bool unix_skb_scm_eq(struct sk_buff *skb,
			    struct scm_cookie *scm)
{
	const struct unix_skb_parms *u = &UNIXCB(skb);

	return u->pid == scm->pid &&
	       uid_eq(u->uid, scm->creds.uid) &&
	       gid_eq(u->gid, scm->creds.gid) &&
	       unix_secdata_eq(scm, skb);
}

L
Linus Torvalds 已提交
1629 1630 1631 1632
/*
 *	Send AF_UNIX data.
 */

1633 1634
static int unix_dgram_sendmsg(struct socket *sock, struct msghdr *msg,
			      size_t len)
L
Linus Torvalds 已提交
1635 1636
{
	struct sock *sk = sock->sk;
1637
	struct net *net = sock_net(sk);
L
Linus Torvalds 已提交
1638
	struct unix_sock *u = unix_sk(sk);
1639
	DECLARE_SOCKADDR(struct sockaddr_un *, sunaddr, msg->msg_name);
L
Linus Torvalds 已提交
1640 1641 1642
	struct sock *other = NULL;
	int namelen = 0; /* fake GCC */
	int err;
1643
	unsigned int hash;
1644
	struct sk_buff *skb;
L
Linus Torvalds 已提交
1645
	long timeo;
C
Christoph Hellwig 已提交
1646
	struct scm_cookie scm;
E
Eric Dumazet 已提交
1647
	int max_level;
1648
	int data_len = 0;
1649
	int sk_locked;
L
Linus Torvalds 已提交
1650

1651
	wait_for_unix_gc();
C
Christoph Hellwig 已提交
1652
	err = scm_send(sock, msg, &scm, false);
L
Linus Torvalds 已提交
1653 1654 1655 1656 1657 1658 1659 1660 1661 1662 1663 1664 1665 1666 1667 1668 1669 1670 1671 1672
	if (err < 0)
		return err;

	err = -EOPNOTSUPP;
	if (msg->msg_flags&MSG_OOB)
		goto out;

	if (msg->msg_namelen) {
		err = unix_mkname(sunaddr, msg->msg_namelen, &hash);
		if (err < 0)
			goto out;
		namelen = err;
	} else {
		sunaddr = NULL;
		err = -ENOTCONN;
		other = unix_peer_get(sk);
		if (!other)
			goto out;
	}

1673 1674
	if (test_bit(SOCK_PASSCRED, &sock->flags) && !u->addr
	    && (err = unix_autobind(sock)) != 0)
L
Linus Torvalds 已提交
1675 1676 1677 1678 1679 1680
		goto out;

	err = -EMSGSIZE;
	if (len > sk->sk_sndbuf - 32)
		goto out;

1681
	if (len > SKB_MAX_ALLOC) {
1682 1683 1684
		data_len = min_t(size_t,
				 len - SKB_MAX_ALLOC,
				 MAX_SKB_FRAGS * PAGE_SIZE);
1685 1686 1687 1688
		data_len = PAGE_ALIGN(data_len);

		BUILD_BUG_ON(SKB_MAX_ALLOC < PAGE_SIZE);
	}
1689 1690

	skb = sock_alloc_send_pskb(sk, len - data_len, data_len,
1691 1692
				   msg->msg_flags & MSG_DONTWAIT, &err,
				   PAGE_ALLOC_COSTLY_ORDER);
1693
	if (skb == NULL)
L
Linus Torvalds 已提交
1694 1695
		goto out;

C
Christoph Hellwig 已提交
1696
	err = unix_scm_to_skb(&scm, skb, true);
E
Eric Dumazet 已提交
1697
	if (err < 0)
1698
		goto out_free;
E
Eric Dumazet 已提交
1699
	max_level = err + 1;
C
Catherine Zhang 已提交
1700

1701 1702 1703
	skb_put(skb, len - data_len);
	skb->data_len = data_len;
	skb->len = len;
A
Al Viro 已提交
1704
	err = skb_copy_datagram_from_iter(skb, 0, &msg->msg_iter, len);
L
Linus Torvalds 已提交
1705 1706 1707 1708 1709 1710 1711 1712 1713 1714 1715
	if (err)
		goto out_free;

	timeo = sock_sndtimeo(sk, msg->msg_flags & MSG_DONTWAIT);

restart:
	if (!other) {
		err = -ECONNRESET;
		if (sunaddr == NULL)
			goto out_free;

1716
		other = unix_find_other(net, sunaddr, namelen, sk->sk_type,
L
Linus Torvalds 已提交
1717
					hash, &err);
1718
		if (other == NULL)
L
Linus Torvalds 已提交
1719 1720 1721
			goto out_free;
	}

A
Alban Crequy 已提交
1722 1723 1724 1725 1726 1727
	if (sk_filter(other, skb) < 0) {
		/* Toss the packet but do not return any error to the sender */
		err = len;
		goto out_free;
	}

1728
	sk_locked = 0;
1729
	unix_state_lock(other);
1730
restart_locked:
L
Linus Torvalds 已提交
1731 1732 1733 1734
	err = -EPERM;
	if (!unix_may_send(sk, other))
		goto out_unlock;

1735
	if (unlikely(sock_flag(other, SOCK_DEAD))) {
L
Linus Torvalds 已提交
1736 1737 1738 1739
		/*
		 *	Check with 1003.1g - what should
		 *	datagram error
		 */
1740
		unix_state_unlock(other);
L
Linus Torvalds 已提交
1741 1742
		sock_put(other);

1743 1744 1745
		if (!sk_locked)
			unix_state_lock(sk);

L
Linus Torvalds 已提交
1746 1747
		err = 0;
		if (unix_peer(sk) == other) {
1748
			unix_peer(sk) = NULL;
1749 1750
			unix_dgram_peer_wake_disconnect_wakeup(sk, other);

1751
			unix_state_unlock(sk);
L
Linus Torvalds 已提交
1752 1753 1754 1755 1756

			unix_dgram_disconnected(sk, other);
			sock_put(other);
			err = -ECONNREFUSED;
		} else {
1757
			unix_state_unlock(sk);
L
Linus Torvalds 已提交
1758 1759 1760 1761 1762 1763 1764 1765 1766 1767 1768 1769 1770 1771 1772 1773 1774 1775
		}

		other = NULL;
		if (err)
			goto out_free;
		goto restart;
	}

	err = -EPIPE;
	if (other->sk_shutdown & RCV_SHUTDOWN)
		goto out_unlock;

	if (sk->sk_type != SOCK_SEQPACKET) {
		err = security_unix_may_send(sk->sk_socket, other->sk_socket);
		if (err)
			goto out_unlock;
	}

1776 1777 1778 1779 1780 1781
	/* other == sk && unix_peer(other) != sk if
	 * - unix_peer(sk) == NULL, destination address bound to sk
	 * - unix_peer(sk) == sk by time of get but disconnected before lock
	 */
	if (other != sk &&
	    unlikely(unix_peer(other) != sk && unix_recvq_full(other))) {
1782 1783 1784 1785 1786 1787 1788 1789
		if (timeo) {
			timeo = unix_wait_for_peer(other, timeo);

			err = sock_intr_errno(timeo);
			if (signal_pending(current))
				goto out_free;

			goto restart;
L
Linus Torvalds 已提交
1790 1791
		}

1792 1793 1794 1795
		if (!sk_locked) {
			unix_state_unlock(other);
			unix_state_double_lock(sk, other);
		}
L
Linus Torvalds 已提交
1796

1797 1798 1799 1800 1801 1802
		if (unix_peer(sk) != other ||
		    unix_dgram_peer_wake_me(sk, other)) {
			err = -EAGAIN;
			sk_locked = 1;
			goto out_unlock;
		}
L
Linus Torvalds 已提交
1803

1804 1805 1806 1807
		if (!sk_locked) {
			sk_locked = 1;
			goto restart_locked;
		}
L
Linus Torvalds 已提交
1808 1809
	}

1810 1811 1812
	if (unlikely(sk_locked))
		unix_state_unlock(sk);

1813 1814
	if (sock_flag(other, SOCK_RCVTSTAMP))
		__net_timestamp(skb);
1815
	maybe_add_creds(skb, sock, other);
L
Linus Torvalds 已提交
1816
	skb_queue_tail(&other->sk_receive_queue, skb);
E
Eric Dumazet 已提交
1817 1818
	if (max_level > unix_sk(other)->recursion_level)
		unix_sk(other)->recursion_level = max_level;
1819
	unix_state_unlock(other);
1820
	other->sk_data_ready(other);
L
Linus Torvalds 已提交
1821
	sock_put(other);
C
Christoph Hellwig 已提交
1822
	scm_destroy(&scm);
L
Linus Torvalds 已提交
1823 1824 1825
	return len;

out_unlock:
1826 1827
	if (sk_locked)
		unix_state_unlock(sk);
1828
	unix_state_unlock(other);
L
Linus Torvalds 已提交
1829 1830 1831 1832 1833
out_free:
	kfree_skb(skb);
out:
	if (other)
		sock_put(other);
C
Christoph Hellwig 已提交
1834
	scm_destroy(&scm);
L
Linus Torvalds 已提交
1835 1836 1837
	return err;
}

1838 1839 1840 1841
/* We use paged skbs for stream sockets, and limit occupancy to 32768
 * bytes, and a minimun of a full page.
 */
#define UNIX_SKB_FRAGS_SZ (PAGE_SIZE << get_order(32768))
1842

1843 1844
static int unix_stream_sendmsg(struct socket *sock, struct msghdr *msg,
			       size_t len)
L
Linus Torvalds 已提交
1845 1846 1847
{
	struct sock *sk = sock->sk;
	struct sock *other = NULL;
E
Eric Dumazet 已提交
1848
	int err, size;
1849
	struct sk_buff *skb;
1850
	int sent = 0;
C
Christoph Hellwig 已提交
1851
	struct scm_cookie scm;
1852
	bool fds_sent = false;
E
Eric Dumazet 已提交
1853
	int max_level;
1854
	int data_len;
L
Linus Torvalds 已提交
1855

1856
	wait_for_unix_gc();
C
Christoph Hellwig 已提交
1857
	err = scm_send(sock, msg, &scm, false);
L
Linus Torvalds 已提交
1858 1859 1860 1861 1862 1863 1864 1865 1866 1867 1868 1869
	if (err < 0)
		return err;

	err = -EOPNOTSUPP;
	if (msg->msg_flags&MSG_OOB)
		goto out_err;

	if (msg->msg_namelen) {
		err = sk->sk_state == TCP_ESTABLISHED ? -EISCONN : -EOPNOTSUPP;
		goto out_err;
	} else {
		err = -ENOTCONN;
1870
		other = unix_peer(sk);
L
Linus Torvalds 已提交
1871 1872 1873 1874 1875 1876 1877
		if (!other)
			goto out_err;
	}

	if (sk->sk_shutdown & SEND_SHUTDOWN)
		goto pipe_err;

E
Eric Dumazet 已提交
1878
	while (sent < len) {
1879
		size = len - sent;
L
Linus Torvalds 已提交
1880 1881

		/* Keep two messages in the pipe so it schedules better */
1882
		size = min_t(int, size, (sk->sk_sndbuf >> 1) - 64);
L
Linus Torvalds 已提交
1883

1884 1885
		/* allow fallback to order-0 allocations */
		size = min_t(int, size, SKB_MAX_HEAD(0) + UNIX_SKB_FRAGS_SZ);
1886

1887
		data_len = max_t(int, 0, size - SKB_MAX_HEAD(0));
L
Linus Torvalds 已提交
1888

1889 1890
		data_len = min_t(size_t, size, PAGE_ALIGN(data_len));

1891
		skb = sock_alloc_send_pskb(sk, size - data_len, data_len,
1892 1893
					   msg->msg_flags & MSG_DONTWAIT, &err,
					   get_order(UNIX_SKB_FRAGS_SZ));
1894
		if (!skb)
L
Linus Torvalds 已提交
1895 1896
			goto out_err;

1897
		/* Only send the fds in the first buffer */
C
Christoph Hellwig 已提交
1898
		err = unix_scm_to_skb(&scm, skb, !fds_sent);
E
Eric Dumazet 已提交
1899
		if (err < 0) {
1900
			kfree_skb(skb);
1901
			goto out_err;
1902
		}
E
Eric Dumazet 已提交
1903
		max_level = err + 1;
1904
		fds_sent = true;
L
Linus Torvalds 已提交
1905

1906 1907 1908
		skb_put(skb, size - data_len);
		skb->data_len = data_len;
		skb->len = size;
A
Al Viro 已提交
1909
		err = skb_copy_datagram_from_iter(skb, 0, &msg->msg_iter, size);
E
Eric Dumazet 已提交
1910
		if (err) {
L
Linus Torvalds 已提交
1911
			kfree_skb(skb);
1912
			goto out_err;
L
Linus Torvalds 已提交
1913 1914
		}

1915
		unix_state_lock(other);
L
Linus Torvalds 已提交
1916 1917 1918 1919 1920

		if (sock_flag(other, SOCK_DEAD) ||
		    (other->sk_shutdown & RCV_SHUTDOWN))
			goto pipe_err_free;

1921
		maybe_add_creds(skb, sock, other);
L
Linus Torvalds 已提交
1922
		skb_queue_tail(&other->sk_receive_queue, skb);
E
Eric Dumazet 已提交
1923 1924
		if (max_level > unix_sk(other)->recursion_level)
			unix_sk(other)->recursion_level = max_level;
1925
		unix_state_unlock(other);
1926
		other->sk_data_ready(other);
1927
		sent += size;
L
Linus Torvalds 已提交
1928 1929
	}

C
Christoph Hellwig 已提交
1930
	scm_destroy(&scm);
L
Linus Torvalds 已提交
1931 1932 1933 1934

	return sent;

pipe_err_free:
1935
	unix_state_unlock(other);
L
Linus Torvalds 已提交
1936 1937
	kfree_skb(skb);
pipe_err:
E
Eric Dumazet 已提交
1938 1939
	if (sent == 0 && !(msg->msg_flags&MSG_NOSIGNAL))
		send_sig(SIGPIPE, current, 0);
L
Linus Torvalds 已提交
1940 1941
	err = -EPIPE;
out_err:
C
Christoph Hellwig 已提交
1942
	scm_destroy(&scm);
L
Linus Torvalds 已提交
1943 1944 1945
	return sent ? : err;
}

1946 1947 1948
static ssize_t unix_stream_sendpage(struct socket *socket, struct page *page,
				    int offset, size_t size, int flags)
{
1949 1950 1951 1952
	int err;
	bool send_sigpipe = false;
	bool init_scm = true;
	struct scm_cookie scm;
1953 1954 1955 1956 1957 1958 1959 1960 1961 1962 1963 1964 1965
	struct sock *other, *sk = socket->sk;
	struct sk_buff *skb, *newskb = NULL, *tail = NULL;

	if (flags & MSG_OOB)
		return -EOPNOTSUPP;

	other = unix_peer(sk);
	if (!other || sk->sk_state != TCP_ESTABLISHED)
		return -ENOTCONN;

	if (false) {
alloc_skb:
		unix_state_unlock(other);
1966
		mutex_unlock(&unix_sk(other)->iolock);
1967 1968 1969
		newskb = sock_alloc_send_pskb(sk, 0, 0, flags & MSG_DONTWAIT,
					      &err, 0);
		if (!newskb)
1970
			goto err;
1971 1972
	}

1973
	/* we must acquire iolock as we modify already present
1974 1975
	 * skbs in the sk_receive_queue and mess with skb->len
	 */
1976
	err = mutex_lock_interruptible(&unix_sk(other)->iolock);
1977 1978 1979 1980 1981 1982 1983
	if (err) {
		err = flags & MSG_DONTWAIT ? -EAGAIN : -ERESTARTSYS;
		goto err;
	}

	if (sk->sk_shutdown & SEND_SHUTDOWN) {
		err = -EPIPE;
1984
		send_sigpipe = true;
1985 1986 1987 1988 1989 1990 1991 1992
		goto err_unlock;
	}

	unix_state_lock(other);

	if (sock_flag(other, SOCK_DEAD) ||
	    other->sk_shutdown & RCV_SHUTDOWN) {
		err = -EPIPE;
1993
		send_sigpipe = true;
1994 1995 1996
		goto err_state_unlock;
	}

1997 1998 1999 2000 2001 2002 2003
	if (init_scm) {
		err = maybe_init_creds(&scm, socket, other);
		if (err)
			goto err_state_unlock;
		init_scm = false;
	}

2004 2005 2006
	skb = skb_peek_tail(&other->sk_receive_queue);
	if (tail && tail == skb) {
		skb = newskb;
2007 2008
	} else if (!skb || !unix_skb_scm_eq(skb, &scm)) {
		if (newskb) {
2009
			skb = newskb;
2010 2011
		} else {
			tail = skb;
2012
			goto alloc_skb;
2013
		}
2014 2015 2016 2017 2018 2019
	} else if (newskb) {
		/* this is fast path, we don't necessarily need to
		 * call to kfree_skb even though with newskb == NULL
		 * this - does no harm
		 */
		consume_skb(newskb);
2020
		newskb = NULL;
2021 2022 2023 2024 2025 2026 2027 2028 2029 2030 2031 2032
	}

	if (skb_append_pagefrags(skb, page, offset, size)) {
		tail = skb;
		goto alloc_skb;
	}

	skb->len += size;
	skb->data_len += size;
	skb->truesize += size;
	atomic_add(size, &sk->sk_wmem_alloc);

2033
	if (newskb) {
2034 2035 2036
		err = unix_scm_to_skb(&scm, skb, false);
		if (err)
			goto err_state_unlock;
2037
		spin_lock(&other->sk_receive_queue.lock);
2038
		__skb_queue_tail(&other->sk_receive_queue, newskb);
2039 2040
		spin_unlock(&other->sk_receive_queue.lock);
	}
2041 2042

	unix_state_unlock(other);
2043
	mutex_unlock(&unix_sk(other)->iolock);
2044 2045

	other->sk_data_ready(other);
2046
	scm_destroy(&scm);
2047 2048 2049 2050 2051
	return size;

err_state_unlock:
	unix_state_unlock(other);
err_unlock:
2052
	mutex_unlock(&unix_sk(other)->iolock);
2053 2054 2055 2056
err:
	kfree_skb(newskb);
	if (send_sigpipe && !(flags & MSG_NOSIGNAL))
		send_sig(SIGPIPE, current, 0);
2057 2058
	if (!init_scm)
		scm_destroy(&scm);
2059 2060 2061
	return err;
}

2062 2063
static int unix_seqpacket_sendmsg(struct socket *sock, struct msghdr *msg,
				  size_t len)
L
Linus Torvalds 已提交
2064 2065 2066
{
	int err;
	struct sock *sk = sock->sk;
2067

L
Linus Torvalds 已提交
2068 2069 2070 2071 2072 2073 2074 2075 2076 2077
	err = sock_error(sk);
	if (err)
		return err;

	if (sk->sk_state != TCP_ESTABLISHED)
		return -ENOTCONN;

	if (msg->msg_namelen)
		msg->msg_namelen = 0;

2078
	return unix_dgram_sendmsg(sock, msg, len);
L
Linus Torvalds 已提交
2079
}
2080

2081 2082
static int unix_seqpacket_recvmsg(struct socket *sock, struct msghdr *msg,
				  size_t size, int flags)
2083 2084 2085 2086 2087 2088
{
	struct sock *sk = sock->sk;

	if (sk->sk_state != TCP_ESTABLISHED)
		return -ENOTCONN;

2089
	return unix_dgram_recvmsg(sock, msg, size, flags);
2090 2091
}

L
Linus Torvalds 已提交
2092 2093 2094 2095 2096 2097 2098 2099 2100 2101
static void unix_copy_addr(struct msghdr *msg, struct sock *sk)
{
	struct unix_sock *u = unix_sk(sk);

	if (u->addr) {
		msg->msg_namelen = u->addr->len;
		memcpy(msg->msg_name, u->addr->name, u->addr->len);
	}
}

2102 2103
static int unix_dgram_recvmsg(struct socket *sock, struct msghdr *msg,
			      size_t size, int flags)
L
Linus Torvalds 已提交
2104
{
C
Christoph Hellwig 已提交
2105
	struct scm_cookie scm;
L
Linus Torvalds 已提交
2106 2107
	struct sock *sk = sock->sk;
	struct unix_sock *u = unix_sk(sk);
2108 2109
	struct sk_buff *skb, *last;
	long timeo;
L
Linus Torvalds 已提交
2110
	int err;
2111
	int peeked, skip;
L
Linus Torvalds 已提交
2112 2113 2114 2115 2116

	err = -EOPNOTSUPP;
	if (flags&MSG_OOB)
		goto out;

2117
	timeo = sock_rcvtimeo(sk, flags & MSG_DONTWAIT);
L
Linus Torvalds 已提交
2118

2119
	do {
2120
		mutex_lock(&u->iolock);
2121

2122
		skip = sk_peek_offset(sk, flags);
2123 2124
		skb = __skb_try_recv_datagram(sk, flags, NULL, &peeked, &skip,
					      &err, &last);
2125 2126 2127
		if (skb)
			break;

2128
		mutex_unlock(&u->iolock);
2129 2130 2131 2132 2133 2134

		if (err != -EAGAIN)
			break;
	} while (timeo &&
		 !__skb_wait_for_more_packets(sk, &err, &timeo, last));

2135
	if (!skb) { /* implies iolock unlocked */
2136 2137 2138 2139 2140 2141
		unix_state_lock(sk);
		/* Signal EOF on disconnected non-blocking SEQPACKET socket. */
		if (sk->sk_type == SOCK_SEQPACKET && err == -EAGAIN &&
		    (sk->sk_shutdown & RCV_SHUTDOWN))
			err = 0;
		unix_state_unlock(sk);
2142
		goto out;
2143
	}
L
Linus Torvalds 已提交
2144

2145 2146 2147 2148
	if (wq_has_sleeper(&u->peer_wait))
		wake_up_interruptible_sync_poll(&u->peer_wait,
						POLLOUT | POLLWRNORM |
						POLLWRBAND);
L
Linus Torvalds 已提交
2149 2150 2151 2152

	if (msg->msg_name)
		unix_copy_addr(msg, skb->sk);

2153 2154 2155
	if (size > skb->len - skip)
		size = skb->len - skip;
	else if (size < skb->len - skip)
L
Linus Torvalds 已提交
2156 2157
		msg->msg_flags |= MSG_TRUNC;

2158
	err = skb_copy_datagram_msg(skb, skip, msg, size);
L
Linus Torvalds 已提交
2159 2160 2161
	if (err)
		goto out_free;

2162 2163 2164
	if (sock_flag(sk, SOCK_RCVTSTAMP))
		__sock_recv_timestamp(msg, sk, skb);

C
Christoph Hellwig 已提交
2165 2166 2167 2168
	memset(&scm, 0, sizeof(scm));

	scm_set_cred(&scm, UNIXCB(skb).pid, UNIXCB(skb).uid, UNIXCB(skb).gid);
	unix_set_secdata(&scm, skb);
L
Linus Torvalds 已提交
2169

E
Eric Dumazet 已提交
2170
	if (!(flags & MSG_PEEK)) {
L
Linus Torvalds 已提交
2171
		if (UNIXCB(skb).fp)
C
Christoph Hellwig 已提交
2172
			unix_detach_fds(&scm, skb);
2173 2174

		sk_peek_offset_bwd(sk, skb->len);
E
Eric Dumazet 已提交
2175
	} else {
L
Linus Torvalds 已提交
2176 2177 2178 2179 2180 2181
		/* It is questionable: on PEEK we could:
		   - do not return fds - good, but too simple 8)
		   - return fds, and do not return them on read (old strategy,
		     apparently wrong)
		   - clone fds (I chose it for now, it is the most universal
		     solution)
2182 2183 2184 2185 2186

		   POSIX 1003.1g does not actually define this clearly
		   at all. POSIX 1003.1g doesn't define a lot of things
		   clearly however!

L
Linus Torvalds 已提交
2187
		*/
2188 2189 2190

		sk_peek_offset_fwd(sk, size);

L
Linus Torvalds 已提交
2191
		if (UNIXCB(skb).fp)
C
Christoph Hellwig 已提交
2192
			scm.fp = scm_fp_dup(UNIXCB(skb).fp);
L
Linus Torvalds 已提交
2193
	}
2194
	err = (flags & MSG_TRUNC) ? skb->len - skip : size;
L
Linus Torvalds 已提交
2195

C
Christoph Hellwig 已提交
2196
	scm_recv(sock, msg, &scm, flags);
L
Linus Torvalds 已提交
2197 2198

out_free:
E
Eric Dumazet 已提交
2199
	skb_free_datagram(sk, skb);
2200
	mutex_unlock(&u->iolock);
L
Linus Torvalds 已提交
2201 2202 2203 2204 2205
out:
	return err;
}

/*
2206
 *	Sleep until more data has arrived. But check for races..
L
Linus Torvalds 已提交
2207
 */
2208
static long unix_stream_data_wait(struct sock *sk, long timeo,
2209 2210
				  struct sk_buff *last, unsigned int last_len,
				  bool freezable)
L
Linus Torvalds 已提交
2211
{
2212
	struct sk_buff *tail;
L
Linus Torvalds 已提交
2213 2214
	DEFINE_WAIT(wait);

2215
	unix_state_lock(sk);
L
Linus Torvalds 已提交
2216 2217

	for (;;) {
E
Eric Dumazet 已提交
2218
		prepare_to_wait(sk_sleep(sk), &wait, TASK_INTERRUPTIBLE);
L
Linus Torvalds 已提交
2219

2220 2221 2222
		tail = skb_peek_tail(&sk->sk_receive_queue);
		if (tail != last ||
		    (tail && tail->len != last_len) ||
L
Linus Torvalds 已提交
2223 2224 2225 2226 2227 2228
		    sk->sk_err ||
		    (sk->sk_shutdown & RCV_SHUTDOWN) ||
		    signal_pending(current) ||
		    !timeo)
			break;

2229
		sk_set_bit(SOCKWQ_ASYNC_WAITDATA, sk);
2230
		unix_state_unlock(sk);
2231 2232 2233 2234
		if (freezable)
			timeo = freezable_schedule_timeout(timeo);
		else
			timeo = schedule_timeout(timeo);
2235
		unix_state_lock(sk);
2236 2237 2238 2239

		if (sock_flag(sk, SOCK_DEAD))
			break;

2240
		sk_clear_bit(SOCKWQ_ASYNC_WAITDATA, sk);
L
Linus Torvalds 已提交
2241 2242
	}

E
Eric Dumazet 已提交
2243
	finish_wait(sk_sleep(sk), &wait);
2244
	unix_state_unlock(sk);
L
Linus Torvalds 已提交
2245 2246 2247
	return timeo;
}

2248 2249 2250 2251 2252
static unsigned int unix_skb_len(const struct sk_buff *skb)
{
	return skb->len - UNIXCB(skb).consumed;
}

2253 2254 2255 2256 2257 2258 2259 2260 2261 2262 2263
struct unix_stream_read_state {
	int (*recv_actor)(struct sk_buff *, int, int,
			  struct unix_stream_read_state *);
	struct socket *socket;
	struct msghdr *msg;
	struct pipe_inode_info *pipe;
	size_t size;
	int flags;
	unsigned int splice_flags;
};

2264 2265
static int unix_stream_read_generic(struct unix_stream_read_state *state,
				    bool freezable)
L
Linus Torvalds 已提交
2266
{
C
Christoph Hellwig 已提交
2267
	struct scm_cookie scm;
2268
	struct socket *sock = state->socket;
L
Linus Torvalds 已提交
2269 2270 2271
	struct sock *sk = sock->sk;
	struct unix_sock *u = unix_sk(sk);
	int copied = 0;
2272
	int flags = state->flags;
2273
	int noblock = flags & MSG_DONTWAIT;
2274
	bool check_creds = false;
L
Linus Torvalds 已提交
2275 2276 2277
	int target;
	int err = 0;
	long timeo;
2278
	int skip;
2279 2280
	size_t size = state->size;
	unsigned int last_len;
L
Linus Torvalds 已提交
2281

2282 2283
	if (unlikely(sk->sk_state != TCP_ESTABLISHED)) {
		err = -EINVAL;
L
Linus Torvalds 已提交
2284
		goto out;
2285
	}
L
Linus Torvalds 已提交
2286

2287 2288
	if (unlikely(flags & MSG_OOB)) {
		err = -EOPNOTSUPP;
L
Linus Torvalds 已提交
2289
		goto out;
2290
	}
L
Linus Torvalds 已提交
2291

2292
	target = sock_rcvlowat(sk, flags & MSG_WAITALL, size);
2293
	timeo = sock_rcvtimeo(sk, noblock);
L
Linus Torvalds 已提交
2294

2295 2296
	memset(&scm, 0, sizeof(scm));

L
Linus Torvalds 已提交
2297 2298 2299
	/* Lock the socket to prevent queue disordering
	 * while sleeps in memcpy_tomsg
	 */
2300
	mutex_lock(&u->iolock);
L
Linus Torvalds 已提交
2301

2302 2303 2304 2305 2306
	if (flags & MSG_PEEK)
		skip = sk_peek_offset(sk, flags);
	else
		skip = 0;

E
Eric Dumazet 已提交
2307
	do {
L
Linus Torvalds 已提交
2308
		int chunk;
2309
		bool drop_skb;
2310
		struct sk_buff *skb, *last;
L
Linus Torvalds 已提交
2311

2312
redo:
2313
		unix_state_lock(sk);
2314 2315 2316 2317
		if (sock_flag(sk, SOCK_DEAD)) {
			err = -ECONNRESET;
			goto unlock;
		}
2318
		last = skb = skb_peek(&sk->sk_receive_queue);
2319
		last_len = last ? last->len : 0;
2320
again:
E
Eric Dumazet 已提交
2321
		if (skb == NULL) {
E
Eric Dumazet 已提交
2322
			unix_sk(sk)->recursion_level = 0;
L
Linus Torvalds 已提交
2323
			if (copied >= target)
2324
				goto unlock;
L
Linus Torvalds 已提交
2325 2326 2327 2328

			/*
			 *	POSIX 1003.1g mandates this order.
			 */
2329

E
Eric Dumazet 已提交
2330 2331
			err = sock_error(sk);
			if (err)
2332
				goto unlock;
L
Linus Torvalds 已提交
2333
			if (sk->sk_shutdown & RCV_SHUTDOWN)
2334 2335 2336
				goto unlock;

			unix_state_unlock(sk);
2337 2338
			if (!timeo) {
				err = -EAGAIN;
L
Linus Torvalds 已提交
2339
				break;
2340 2341
			}

2342
			mutex_unlock(&u->iolock);
L
Linus Torvalds 已提交
2343

2344
			timeo = unix_stream_data_wait(sk, timeo, last,
2345
						      last_len, freezable);
L
Linus Torvalds 已提交
2346

2347
			if (signal_pending(current)) {
L
Linus Torvalds 已提交
2348
				err = sock_intr_errno(timeo);
2349
				scm_destroy(&scm);
L
Linus Torvalds 已提交
2350 2351
				goto out;
			}
2352

2353
			mutex_lock(&u->iolock);
2354
			goto redo;
2355
unlock:
2356 2357
			unix_state_unlock(sk);
			break;
L
Linus Torvalds 已提交
2358
		}
2359

2360 2361
		while (skip >= unix_skb_len(skb)) {
			skip -= unix_skb_len(skb);
2362
			last = skb;
2363
			last_len = skb->len;
2364
			skb = skb_peek_next(skb, &sk->sk_receive_queue);
2365 2366
			if (!skb)
				goto again;
2367 2368
		}

2369
		unix_state_unlock(sk);
L
Linus Torvalds 已提交
2370 2371 2372

		if (check_creds) {
			/* Never glue messages from different writers */
2373
			if (!unix_skb_scm_eq(skb, &scm))
L
Linus Torvalds 已提交
2374
				break;
2375
		} else if (test_bit(SOCK_PASSCRED, &sock->flags)) {
L
Linus Torvalds 已提交
2376
			/* Copy credentials */
C
Christoph Hellwig 已提交
2377
			scm_set_cred(&scm, UNIXCB(skb).pid, UNIXCB(skb).uid, UNIXCB(skb).gid);
2378
			unix_set_secdata(&scm, skb);
2379
			check_creds = true;
L
Linus Torvalds 已提交
2380 2381 2382
		}

		/* Copy address just once */
2383 2384 2385 2386
		if (state->msg && state->msg->msg_name) {
			DECLARE_SOCKADDR(struct sockaddr_un *, sunaddr,
					 state->msg->msg_name);
			unix_copy_addr(state->msg, skb->sk);
L
Linus Torvalds 已提交
2387 2388 2389
			sunaddr = NULL;
		}

2390
		chunk = min_t(unsigned int, unix_skb_len(skb) - skip, size);
2391
		skb_get(skb);
2392
		chunk = state->recv_actor(skb, skip, chunk, state);
2393 2394 2395
		drop_skb = !unix_skb_len(skb);
		/* skb is only safe to use if !drop_skb */
		consume_skb(skb);
2396
		if (chunk < 0) {
L
Linus Torvalds 已提交
2397 2398 2399 2400 2401 2402 2403
			if (copied == 0)
				copied = -EFAULT;
			break;
		}
		copied += chunk;
		size -= chunk;

2404 2405 2406 2407 2408 2409 2410 2411 2412 2413 2414 2415
		if (drop_skb) {
			/* the skb was touched by a concurrent reader;
			 * we should not expect anything from this skb
			 * anymore and assume it invalid - we can be
			 * sure it was dropped from the socket queue
			 *
			 * let's report a short read
			 */
			err = 0;
			break;
		}

L
Linus Torvalds 已提交
2416
		/* Mark read part of skb as used */
E
Eric Dumazet 已提交
2417
		if (!(flags & MSG_PEEK)) {
2418
			UNIXCB(skb).consumed += chunk;
L
Linus Torvalds 已提交
2419

2420 2421
			sk_peek_offset_bwd(sk, chunk);

L
Linus Torvalds 已提交
2422
			if (UNIXCB(skb).fp)
C
Christoph Hellwig 已提交
2423
				unix_detach_fds(&scm, skb);
L
Linus Torvalds 已提交
2424

2425
			if (unix_skb_len(skb))
L
Linus Torvalds 已提交
2426 2427
				break;

2428
			skb_unlink(skb, &sk->sk_receive_queue);
2429
			consume_skb(skb);
L
Linus Torvalds 已提交
2430

C
Christoph Hellwig 已提交
2431
			if (scm.fp)
L
Linus Torvalds 已提交
2432
				break;
E
Eric Dumazet 已提交
2433
		} else {
L
Linus Torvalds 已提交
2434 2435 2436
			/* It is questionable, see note in unix_dgram_recvmsg.
			 */
			if (UNIXCB(skb).fp)
C
Christoph Hellwig 已提交
2437
				scm.fp = scm_fp_dup(UNIXCB(skb).fp);
L
Linus Torvalds 已提交
2438

2439
			sk_peek_offset_fwd(sk, chunk);
2440

2441 2442 2443
			if (UNIXCB(skb).fp)
				break;

2444
			skip = 0;
2445 2446 2447 2448 2449 2450 2451
			last = skb;
			last_len = skb->len;
			unix_state_lock(sk);
			skb = skb_peek_next(skb, &sk->sk_receive_queue);
			if (skb)
				goto again;
			unix_state_unlock(sk);
L
Linus Torvalds 已提交
2452 2453 2454 2455
			break;
		}
	} while (size);

2456
	mutex_unlock(&u->iolock);
2457 2458 2459 2460
	if (state->msg)
		scm_recv(sock, state->msg, &scm, flags);
	else
		scm_destroy(&scm);
L
Linus Torvalds 已提交
2461 2462 2463 2464
out:
	return copied ? : err;
}

2465 2466 2467 2468 2469 2470 2471 2472 2473 2474 2475 2476 2477 2478 2479 2480 2481 2482 2483 2484 2485 2486
static int unix_stream_read_actor(struct sk_buff *skb,
				  int skip, int chunk,
				  struct unix_stream_read_state *state)
{
	int ret;

	ret = skb_copy_datagram_msg(skb, UNIXCB(skb).consumed + skip,
				    state->msg, chunk);
	return ret ?: chunk;
}

static int unix_stream_recvmsg(struct socket *sock, struct msghdr *msg,
			       size_t size, int flags)
{
	struct unix_stream_read_state state = {
		.recv_actor = unix_stream_read_actor,
		.socket = sock,
		.msg = msg,
		.size = size,
		.flags = flags
	};

2487
	return unix_stream_read_generic(&state, true);
2488 2489 2490 2491 2492 2493 2494 2495
}

static int unix_stream_splice_actor(struct sk_buff *skb,
				    int skip, int chunk,
				    struct unix_stream_read_state *state)
{
	return skb_splice_bits(skb, state->socket->sk,
			       UNIXCB(skb).consumed + skip,
A
Al Viro 已提交
2496
			       state->pipe, chunk, state->splice_flags);
2497 2498 2499 2500 2501 2502 2503 2504 2505 2506 2507 2508 2509 2510 2511 2512 2513 2514 2515 2516 2517
}

static ssize_t unix_stream_splice_read(struct socket *sock,  loff_t *ppos,
				       struct pipe_inode_info *pipe,
				       size_t size, unsigned int flags)
{
	struct unix_stream_read_state state = {
		.recv_actor = unix_stream_splice_actor,
		.socket = sock,
		.pipe = pipe,
		.size = size,
		.splice_flags = flags,
	};

	if (unlikely(*ppos))
		return -ESPIPE;

	if (sock->file->f_flags & O_NONBLOCK ||
	    flags & SPLICE_F_NONBLOCK)
		state.flags = MSG_DONTWAIT;

2518
	return unix_stream_read_generic(&state, false);
2519 2520
}

L
Linus Torvalds 已提交
2521 2522 2523 2524 2525
static int unix_shutdown(struct socket *sock, int mode)
{
	struct sock *sk = sock->sk;
	struct sock *other;

2526 2527 2528 2529 2530 2531 2532 2533
	if (mode < SHUT_RD || mode > SHUT_RDWR)
		return -EINVAL;
	/* This maps:
	 * SHUT_RD   (0) -> RCV_SHUTDOWN  (1)
	 * SHUT_WR   (1) -> SEND_SHUTDOWN (2)
	 * SHUT_RDWR (2) -> SHUTDOWN_MASK (3)
	 */
	++mode;
2534 2535 2536 2537 2538 2539 2540 2541 2542 2543 2544 2545 2546 2547 2548 2549 2550 2551 2552 2553 2554 2555 2556 2557 2558 2559

	unix_state_lock(sk);
	sk->sk_shutdown |= mode;
	other = unix_peer(sk);
	if (other)
		sock_hold(other);
	unix_state_unlock(sk);
	sk->sk_state_change(sk);

	if (other &&
		(sk->sk_type == SOCK_STREAM || sk->sk_type == SOCK_SEQPACKET)) {

		int peer_mode = 0;

		if (mode&RCV_SHUTDOWN)
			peer_mode |= SEND_SHUTDOWN;
		if (mode&SEND_SHUTDOWN)
			peer_mode |= RCV_SHUTDOWN;
		unix_state_lock(other);
		other->sk_shutdown |= peer_mode;
		unix_state_unlock(other);
		other->sk_state_change(other);
		if (peer_mode == SHUTDOWN_MASK)
			sk_wake_async(other, SOCK_WAKE_WAITD, POLL_HUP);
		else if (peer_mode & RCV_SHUTDOWN)
			sk_wake_async(other, SOCK_WAKE_WAITD, POLL_IN);
L
Linus Torvalds 已提交
2560
	}
2561 2562 2563
	if (other)
		sock_put(other);

L
Linus Torvalds 已提交
2564 2565 2566
	return 0;
}

2567 2568 2569 2570 2571 2572 2573 2574 2575 2576 2577 2578
long unix_inq_len(struct sock *sk)
{
	struct sk_buff *skb;
	long amount = 0;

	if (sk->sk_state == TCP_LISTEN)
		return -EINVAL;

	spin_lock(&sk->sk_receive_queue.lock);
	if (sk->sk_type == SOCK_STREAM ||
	    sk->sk_type == SOCK_SEQPACKET) {
		skb_queue_walk(&sk->sk_receive_queue, skb)
2579
			amount += unix_skb_len(skb);
2580 2581 2582 2583 2584 2585 2586 2587 2588 2589 2590 2591 2592 2593 2594 2595 2596
	} else {
		skb = skb_peek(&sk->sk_receive_queue);
		if (skb)
			amount = skb->len;
	}
	spin_unlock(&sk->sk_receive_queue.lock);

	return amount;
}
EXPORT_SYMBOL_GPL(unix_inq_len);

long unix_outq_len(struct sock *sk)
{
	return sk_wmem_alloc_get(sk);
}
EXPORT_SYMBOL_GPL(unix_outq_len);

2597 2598 2599 2600 2601 2602 2603 2604 2605 2606 2607 2608 2609 2610 2611 2612 2613 2614 2615 2616 2617 2618 2619 2620 2621 2622 2623 2624 2625 2626 2627 2628 2629 2630 2631 2632 2633
static int unix_open_file(struct sock *sk)
{
	struct path path;
	struct file *f;
	int fd;

	if (!ns_capable(sock_net(sk)->user_ns, CAP_NET_ADMIN))
		return -EPERM;

	unix_state_lock(sk);
	path = unix_sk(sk)->path;
	if (!path.dentry) {
		unix_state_unlock(sk);
		return -ENOENT;
	}

	path_get(&path);
	unix_state_unlock(sk);

	fd = get_unused_fd_flags(O_CLOEXEC);
	if (fd < 0)
		goto out;

	f = dentry_open(&path, O_PATH, current_cred());
	if (IS_ERR(f)) {
		put_unused_fd(fd);
		fd = PTR_ERR(f);
		goto out;
	}

	fd_install(fd, f);
out:
	path_put(&path);

	return fd;
}

L
Linus Torvalds 已提交
2634 2635 2636
static int unix_ioctl(struct socket *sock, unsigned int cmd, unsigned long arg)
{
	struct sock *sk = sock->sk;
2637
	long amount = 0;
L
Linus Torvalds 已提交
2638 2639
	int err;

E
Eric Dumazet 已提交
2640 2641
	switch (cmd) {
	case SIOCOUTQ:
2642
		amount = unix_outq_len(sk);
E
Eric Dumazet 已提交
2643 2644 2645
		err = put_user(amount, (int __user *)arg);
		break;
	case SIOCINQ:
2646 2647 2648 2649
		amount = unix_inq_len(sk);
		if (amount < 0)
			err = amount;
		else
L
Linus Torvalds 已提交
2650
			err = put_user(amount, (int __user *)arg);
2651
		break;
2652 2653 2654
	case SIOCUNIXFILE:
		err = unix_open_file(sk);
		break;
E
Eric Dumazet 已提交
2655 2656 2657
	default:
		err = -ENOIOCTLCMD;
		break;
L
Linus Torvalds 已提交
2658 2659 2660 2661
	}
	return err;
}

E
Eric Dumazet 已提交
2662
static unsigned int unix_poll(struct file *file, struct socket *sock, poll_table *wait)
L
Linus Torvalds 已提交
2663 2664 2665 2666
{
	struct sock *sk = sock->sk;
	unsigned int mask;

E
Eric Dumazet 已提交
2667
	sock_poll_wait(file, sk_sleep(sk), wait);
L
Linus Torvalds 已提交
2668 2669 2670 2671 2672 2673 2674
	mask = 0;

	/* exceptional events? */
	if (sk->sk_err)
		mask |= POLLERR;
	if (sk->sk_shutdown == SHUTDOWN_MASK)
		mask |= POLLHUP;
2675
	if (sk->sk_shutdown & RCV_SHUTDOWN)
E
Eric Dumazet 已提交
2676
		mask |= POLLRDHUP | POLLIN | POLLRDNORM;
L
Linus Torvalds 已提交
2677 2678

	/* readable? */
E
Eric Dumazet 已提交
2679
	if (!skb_queue_empty(&sk->sk_receive_queue))
L
Linus Torvalds 已提交
2680 2681 2682
		mask |= POLLIN | POLLRDNORM;

	/* Connection-based need to check for termination and startup */
E
Eric Dumazet 已提交
2683 2684
	if ((sk->sk_type == SOCK_STREAM || sk->sk_type == SOCK_SEQPACKET) &&
	    sk->sk_state == TCP_CLOSE)
L
Linus Torvalds 已提交
2685 2686 2687 2688 2689 2690 2691 2692 2693 2694 2695 2696
		mask |= POLLHUP;

	/*
	 * we set writable also when the other side has shut down the
	 * connection. This prevents stuck sockets.
	 */
	if (unix_writable(sk))
		mask |= POLLOUT | POLLWRNORM | POLLWRBAND;

	return mask;
}

2697 2698
static unsigned int unix_dgram_poll(struct file *file, struct socket *sock,
				    poll_table *wait)
2699
{
2700 2701
	struct sock *sk = sock->sk, *other;
	unsigned int mask, writable;
2702

E
Eric Dumazet 已提交
2703
	sock_poll_wait(file, sk_sleep(sk), wait);
2704 2705 2706 2707
	mask = 0;

	/* exceptional events? */
	if (sk->sk_err || !skb_queue_empty(&sk->sk_error_queue))
2708
		mask |= POLLERR |
2709
			(sock_flag(sk, SOCK_SELECT_ERR_QUEUE) ? POLLPRI : 0);
2710

2711
	if (sk->sk_shutdown & RCV_SHUTDOWN)
2712
		mask |= POLLRDHUP | POLLIN | POLLRDNORM;
2713 2714 2715 2716
	if (sk->sk_shutdown == SHUTDOWN_MASK)
		mask |= POLLHUP;

	/* readable? */
2717
	if (!skb_queue_empty(&sk->sk_receive_queue))
2718 2719 2720 2721 2722 2723 2724 2725 2726 2727 2728
		mask |= POLLIN | POLLRDNORM;

	/* Connection-based need to check for termination and startup */
	if (sk->sk_type == SOCK_SEQPACKET) {
		if (sk->sk_state == TCP_CLOSE)
			mask |= POLLHUP;
		/* connection hasn't started yet? */
		if (sk->sk_state == TCP_SYN_SENT)
			return mask;
	}

E
Eric Dumazet 已提交
2729
	/* No write status requested, avoid expensive OUT tests. */
2730
	if (!(poll_requested_events(wait) & (POLLWRBAND|POLLWRNORM|POLLOUT)))
E
Eric Dumazet 已提交
2731 2732
		return mask;

2733
	writable = unix_writable(sk);
2734 2735 2736 2737 2738 2739 2740 2741 2742 2743
	if (writable) {
		unix_state_lock(sk);

		other = unix_peer(sk);
		if (other && unix_peer(other) != sk &&
		    unix_recvq_full(other) &&
		    unix_dgram_peer_wake_me(sk, other))
			writable = 0;

		unix_state_unlock(sk);
2744 2745 2746
	}

	if (writable)
2747 2748
		mask |= POLLOUT | POLLWRNORM | POLLWRBAND;
	else
2749
		sk_set_bit(SOCKWQ_ASYNC_NOSPACE, sk);
2750 2751 2752

	return mask;
}
L
Linus Torvalds 已提交
2753 2754

#ifdef CONFIG_PROC_FS
2755

E
Eric Dumazet 已提交
2756 2757 2758 2759 2760
#define BUCKET_SPACE (BITS_PER_LONG - (UNIX_HASH_BITS + 1) - 1)

#define get_bucket(x) ((x) >> BUCKET_SPACE)
#define get_offset(x) ((x) & ((1L << BUCKET_SPACE) - 1))
#define set_bucket_offset(b, o) ((b) << BUCKET_SPACE | (o))
2761

E
Eric Dumazet 已提交
2762
static struct sock *unix_from_bucket(struct seq_file *seq, loff_t *pos)
L
Linus Torvalds 已提交
2763
{
E
Eric Dumazet 已提交
2764 2765 2766 2767
	unsigned long offset = get_offset(*pos);
	unsigned long bucket = get_bucket(*pos);
	struct sock *sk;
	unsigned long count = 0;
L
Linus Torvalds 已提交
2768

E
Eric Dumazet 已提交
2769 2770
	for (sk = sk_head(&unix_socket_table[bucket]); sk; sk = sk_next(sk)) {
		if (sock_net(sk) != seq_file_net(seq))
2771
			continue;
E
Eric Dumazet 已提交
2772 2773 2774 2775 2776 2777 2778 2779 2780 2781 2782 2783 2784 2785 2786 2787 2788 2789 2790
		if (++count == offset)
			break;
	}

	return sk;
}

static struct sock *unix_next_socket(struct seq_file *seq,
				     struct sock *sk,
				     loff_t *pos)
{
	unsigned long bucket;

	while (sk > (struct sock *)SEQ_START_TOKEN) {
		sk = sk_next(sk);
		if (!sk)
			goto next_bucket;
		if (sock_net(sk) == seq_file_net(seq))
			return sk;
L
Linus Torvalds 已提交
2791
	}
E
Eric Dumazet 已提交
2792 2793 2794 2795 2796 2797 2798 2799 2800 2801 2802

	do {
		sk = unix_from_bucket(seq, pos);
		if (sk)
			return sk;

next_bucket:
		bucket = get_bucket(*pos) + 1;
		*pos = set_bucket_offset(bucket, 1);
	} while (bucket < ARRAY_SIZE(unix_socket_table));

L
Linus Torvalds 已提交
2803 2804 2805 2806
	return NULL;
}

static void *unix_seq_start(struct seq_file *seq, loff_t *pos)
2807
	__acquires(unix_table_lock)
L
Linus Torvalds 已提交
2808
{
2809
	spin_lock(&unix_table_lock);
E
Eric Dumazet 已提交
2810 2811 2812 2813 2814 2815 2816 2817

	if (!*pos)
		return SEQ_START_TOKEN;

	if (get_bucket(*pos) >= ARRAY_SIZE(unix_socket_table))
		return NULL;

	return unix_next_socket(seq, NULL, pos);
L
Linus Torvalds 已提交
2818 2819 2820 2821 2822
}

static void *unix_seq_next(struct seq_file *seq, void *v, loff_t *pos)
{
	++*pos;
E
Eric Dumazet 已提交
2823
	return unix_next_socket(seq, v, pos);
L
Linus Torvalds 已提交
2824 2825 2826
}

static void unix_seq_stop(struct seq_file *seq, void *v)
2827
	__releases(unix_table_lock)
L
Linus Torvalds 已提交
2828
{
2829
	spin_unlock(&unix_table_lock);
L
Linus Torvalds 已提交
2830 2831 2832 2833
}

static int unix_seq_show(struct seq_file *seq, void *v)
{
2834

J
Joe Perches 已提交
2835
	if (v == SEQ_START_TOKEN)
L
Linus Torvalds 已提交
2836 2837 2838 2839 2840
		seq_puts(seq, "Num       RefCount Protocol Flags    Type St "
			 "Inode Path\n");
	else {
		struct sock *s = v;
		struct unix_sock *u = unix_sk(s);
2841
		unix_state_lock(s);
L
Linus Torvalds 已提交
2842

D
Dan Rosenberg 已提交
2843
		seq_printf(seq, "%pK: %08X %08X %08X %04X %02X %5lu",
L
Linus Torvalds 已提交
2844 2845 2846 2847 2848 2849 2850 2851 2852 2853 2854 2855 2856 2857 2858 2859 2860 2861 2862 2863 2864 2865 2866
			s,
			atomic_read(&s->sk_refcnt),
			0,
			s->sk_state == TCP_LISTEN ? __SO_ACCEPTCON : 0,
			s->sk_type,
			s->sk_socket ?
			(s->sk_state == TCP_ESTABLISHED ? SS_CONNECTED : SS_UNCONNECTED) :
			(s->sk_state == TCP_ESTABLISHED ? SS_CONNECTING : SS_DISCONNECTING),
			sock_i_ino(s));

		if (u->addr) {
			int i, len;
			seq_putc(seq, ' ');

			i = 0;
			len = u->addr->len - sizeof(short);
			if (!UNIX_ABSTRACT(s))
				len--;
			else {
				seq_putc(seq, '@');
				i++;
			}
			for ( ; i < len; i++)
2867 2868
				seq_putc(seq, u->addr->name->sun_path[i] ?:
					 '@');
L
Linus Torvalds 已提交
2869
		}
2870
		unix_state_unlock(s);
L
Linus Torvalds 已提交
2871 2872 2873 2874 2875 2876
		seq_putc(seq, '\n');
	}

	return 0;
}

2877
static const struct seq_operations unix_seq_ops = {
L
Linus Torvalds 已提交
2878 2879 2880 2881 2882 2883 2884 2885
	.start  = unix_seq_start,
	.next   = unix_seq_next,
	.stop   = unix_seq_stop,
	.show   = unix_seq_show,
};

static int unix_seq_open(struct inode *inode, struct file *file)
{
2886
	return seq_open_net(inode, file, &unix_seq_ops,
E
Eric Dumazet 已提交
2887
			    sizeof(struct seq_net_private));
L
Linus Torvalds 已提交
2888 2889
}

2890
static const struct file_operations unix_seq_fops = {
L
Linus Torvalds 已提交
2891 2892 2893 2894
	.owner		= THIS_MODULE,
	.open		= unix_seq_open,
	.read		= seq_read,
	.llseek		= seq_lseek,
2895
	.release	= seq_release_net,
L
Linus Torvalds 已提交
2896 2897 2898 2899
};

#endif

2900
static const struct net_proto_family unix_family_ops = {
L
Linus Torvalds 已提交
2901 2902 2903 2904 2905
	.family = PF_UNIX,
	.create = unix_create,
	.owner	= THIS_MODULE,
};

2906

2907
static int __net_init unix_net_init(struct net *net)
2908 2909 2910
{
	int error = -ENOMEM;

2911
	net->unx.sysctl_max_dgram_qlen = 10;
2912 2913
	if (unix_sysctl_register(net))
		goto out;
2914

2915
#ifdef CONFIG_PROC_FS
2916
	if (!proc_create("unix", 0, net->proc_net, &unix_seq_fops)) {
2917
		unix_sysctl_unregister(net);
2918
		goto out;
2919
	}
2920 2921 2922
#endif
	error = 0;
out:
2923
	return error;
2924 2925
}

2926
static void __net_exit unix_net_exit(struct net *net)
2927
{
2928
	unix_sysctl_unregister(net);
2929
	remove_proc_entry("unix", net->proc_net);
2930 2931 2932 2933 2934 2935 2936
}

static struct pernet_operations unix_net_ops = {
	.init = unix_net_init,
	.exit = unix_net_exit,
};

L
Linus Torvalds 已提交
2937 2938 2939 2940
static int __init af_unix_init(void)
{
	int rc = -1;

2941
	BUILD_BUG_ON(sizeof(struct unix_skb_parms) > FIELD_SIZEOF(struct sk_buff, cb));
L
Linus Torvalds 已提交
2942 2943

	rc = proto_register(&unix_proto, 1);
2944
	if (rc != 0) {
W
wangweidong 已提交
2945
		pr_crit("%s: Cannot create unix_sock SLAB cache!\n", __func__);
L
Linus Torvalds 已提交
2946 2947 2948 2949
		goto out;
	}

	sock_register(&unix_family_ops);
2950
	register_pernet_subsys(&unix_net_ops);
L
Linus Torvalds 已提交
2951 2952 2953 2954 2955 2956 2957 2958
out:
	return rc;
}

static void __exit af_unix_exit(void)
{
	sock_unregister(PF_UNIX);
	proto_unregister(&unix_proto);
2959
	unregister_pernet_subsys(&unix_net_ops);
L
Linus Torvalds 已提交
2960 2961
}

2962 2963 2964 2965 2966
/* Earlier than device_initcall() so that other drivers invoking
   request_module() don't end up in a loop when modprobe tries
   to use a UNIX socket. But later than subsys_initcall() because
   we depend on stuff initialised there */
fs_initcall(af_unix_init);
L
Linus Torvalds 已提交
2967 2968 2969 2970
module_exit(af_unix_exit);

MODULE_LICENSE("GPL");
MODULE_ALIAS_NETPROTO(PF_UNIX);