af_unix.c 66.2 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 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105
#include <linux/module.h>
#include <linux/kernel.h>
#include <linux/signal.h>
#include <linux/sched.h>
#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>
#include <asm/uaccess.h>
#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>
L
Linus Torvalds 已提交
120

E
Eric Dumazet 已提交
121
struct hlist_head unix_socket_table[2 * UNIX_HASH_SIZE];
122 123 124
EXPORT_SYMBOL_GPL(unix_socket_table);
DEFINE_SPINLOCK(unix_table_lock);
EXPORT_SYMBOL_GPL(unix_table_lock);
E
Eric Dumazet 已提交
125
static atomic_long_t unix_nr_socks;
L
Linus Torvalds 已提交
126 127


E
Eric Dumazet 已提交
128 129 130 131 132 133 134 135 136 137 138
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
139

C
Catherine Zhang 已提交
140
#ifdef CONFIG_SECURITY_NETWORK
141
static void unix_get_secdata(struct scm_cookie *scm, struct sk_buff *skb)
C
Catherine Zhang 已提交
142
{
143
	UNIXCB(skb).secid = scm->secid;
C
Catherine Zhang 已提交
144 145 146 147
}

static inline void unix_set_secdata(struct scm_cookie *scm, struct sk_buff *skb)
{
148 149 150 151 152 153
	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
154 155
}
#else
156
static inline void unix_get_secdata(struct scm_cookie *scm, struct sk_buff *skb)
C
Catherine Zhang 已提交
157 158 159 160
{ }

static inline void unix_set_secdata(struct scm_cookie *scm, struct sk_buff *skb)
{ }
161 162 163 164 165

static inline bool unix_secdata_eq(struct scm_cookie *scm, struct sk_buff *skb)
{
	return true;
}
C
Catherine Zhang 已提交
166 167
#endif /* CONFIG_SECURITY_NETWORK */

L
Linus Torvalds 已提交
168 169
/*
 *  SMP locking strategy:
170
 *    hash table is protected with spinlock unix_table_lock
171
 *    each socket state is protected by separate spin lock.
L
Linus Torvalds 已提交
172 173
 */

174
static inline unsigned int unix_hash_fold(__wsum n)
L
Linus Torvalds 已提交
175
{
176
	unsigned int hash = (__force unsigned int)csum_fold(n);
177

L
Linus Torvalds 已提交
178 179 180 181 182 183 184 185 186 187 188 189 190
	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
191
	return unix_peer(osk) == NULL || unix_our_peer(sk, osk);
L
Linus Torvalds 已提交
192 193
}

194 195 196 197 198
static inline int unix_recvq_full(struct sock const *sk)
{
	return skb_queue_len(&sk->sk_receive_queue) > sk->sk_max_ack_backlog;
}

199
struct sock *unix_peer_get(struct sock *s)
L
Linus Torvalds 已提交
200 201 202
{
	struct sock *peer;

203
	unix_state_lock(s);
L
Linus Torvalds 已提交
204 205 206
	peer = unix_peer(s);
	if (peer)
		sock_hold(peer);
207
	unix_state_unlock(s);
L
Linus Torvalds 已提交
208 209
	return peer;
}
210
EXPORT_SYMBOL_GPL(unix_peer_get);
L
Linus Torvalds 已提交
211 212 213 214 215 216 217 218 219 220 221 222 223

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.
 */
224

225
static int unix_mkname(struct sockaddr_un *sunaddr, int len, unsigned int *hashp)
L
Linus Torvalds 已提交
226 227 228 229 230 231 232 233 234
{
	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
235
		 * sun_path[108] doesn't as such exist.  However in kernel space
L
Linus Torvalds 已提交
236 237 238
		 * we are guaranteed that it is a valid memory location in our
		 * kernel address buffer.
		 */
239
		((char *)sunaddr)[len] = 0;
L
Linus Torvalds 已提交
240 241 242 243
		len = strlen(sunaddr->sun_path)+1+sizeof(short);
		return len;
	}

244
	*hashp = unix_hash_fold(csum_partial(sunaddr, len, 0));
L
Linus Torvalds 已提交
245 246 247 248 249 250 251 252 253 254
	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)
{
255
	WARN_ON(!sk_unhashed(sk));
L
Linus Torvalds 已提交
256 257 258 259 260
	sk_add_node(sk, list);
}

static inline void unix_remove_socket(struct sock *sk)
{
261
	spin_lock(&unix_table_lock);
L
Linus Torvalds 已提交
262
	__unix_remove_socket(sk);
263
	spin_unlock(&unix_table_lock);
L
Linus Torvalds 已提交
264 265 266 267
}

static inline void unix_insert_socket(struct hlist_head *list, struct sock *sk)
{
268
	spin_lock(&unix_table_lock);
L
Linus Torvalds 已提交
269
	__unix_insert_socket(list, sk);
270
	spin_unlock(&unix_table_lock);
L
Linus Torvalds 已提交
271 272
}

273 274
static struct sock *__unix_find_socket_byname(struct net *net,
					      struct sockaddr_un *sunname,
275
					      int len, int type, unsigned int hash)
L
Linus Torvalds 已提交
276 277 278
{
	struct sock *s;

279
	sk_for_each(s, &unix_socket_table[hash ^ type]) {
L
Linus Torvalds 已提交
280 281
		struct unix_sock *u = unix_sk(s);

282
		if (!net_eq(sock_net(s), net))
283 284
			continue;

L
Linus Torvalds 已提交
285 286 287 288 289 290 291 292 293
		if (u->addr->len == len &&
		    !memcmp(u->addr->name, sunname, len))
			goto found;
	}
	s = NULL;
found:
	return s;
}

294 295
static inline struct sock *unix_find_socket_byname(struct net *net,
						   struct sockaddr_un *sunname,
L
Linus Torvalds 已提交
296
						   int len, int type,
297
						   unsigned int hash)
L
Linus Torvalds 已提交
298 299 300
{
	struct sock *s;

301
	spin_lock(&unix_table_lock);
302
	s = __unix_find_socket_byname(net, sunname, len, type, hash);
L
Linus Torvalds 已提交
303 304
	if (s)
		sock_hold(s);
305
	spin_unlock(&unix_table_lock);
L
Linus Torvalds 已提交
306 307 308
	return s;
}

309
static struct sock *unix_find_socket_byinode(struct inode *i)
L
Linus Torvalds 已提交
310 311 312
{
	struct sock *s;

313
	spin_lock(&unix_table_lock);
314
	sk_for_each(s,
L
Linus Torvalds 已提交
315
		    &unix_socket_table[i->i_ino & (UNIX_HASH_SIZE - 1)]) {
A
Al Viro 已提交
316
		struct dentry *dentry = unix_sk(s)->path.dentry;
L
Linus Torvalds 已提交
317

318
		if (dentry && d_backing_inode(dentry) == i) {
L
Linus Torvalds 已提交
319 320 321 322 323 324
			sock_hold(s);
			goto found;
		}
	}
	s = NULL;
found:
325
	spin_unlock(&unix_table_lock);
L
Linus Torvalds 已提交
326 327 328
	return s;
}

329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 439 440
/* 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;
}

441
static int unix_writable(const struct sock *sk)
L
Linus Torvalds 已提交
442
{
443 444
	return sk->sk_state != TCP_LISTEN &&
	       (atomic_read(&sk->sk_wmem_alloc) << 2) <= sk->sk_sndbuf;
L
Linus Torvalds 已提交
445 446 447 448
}

static void unix_write_space(struct sock *sk)
{
449 450 451
	struct socket_wq *wq;

	rcu_read_lock();
L
Linus Torvalds 已提交
452
	if (unix_writable(sk)) {
453 454
		wq = rcu_dereference(sk->sk_wq);
		if (wq_has_sleeper(wq))
E
Eric Dumazet 已提交
455 456
			wake_up_interruptible_sync_poll(&wq->wait,
				POLLOUT | POLLWRNORM | POLLWRBAND);
457
		sk_wake_async(sk, SOCK_WAKE_SPACE, POLL_OUT);
L
Linus Torvalds 已提交
458
	}
459
	rcu_read_unlock();
L
Linus Torvalds 已提交
460 461 462 463 464 465 466 467
}

/* 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)
{
468
	if (!skb_queue_empty(&sk->sk_receive_queue)) {
L
Linus Torvalds 已提交
469 470 471 472 473 474 475 476 477 478 479 480 481 482 483 484 485 486 487 488
		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);

489 490 491
	WARN_ON(atomic_read(&sk->sk_wmem_alloc));
	WARN_ON(!sk_unhashed(sk));
	WARN_ON(sk->sk_socket);
L
Linus Torvalds 已提交
492
	if (!sock_flag(sk, SOCK_DEAD)) {
W
wangweidong 已提交
493
		pr_info("Attempt to release alive unix socket: %p\n", sk);
L
Linus Torvalds 已提交
494 495 496 497 498 499
		return;
	}

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

E
Eric Dumazet 已提交
500
	atomic_long_dec(&unix_nr_socks);
501
	local_bh_disable();
502
	sock_prot_inuse_add(sock_net(sk), sk->sk_prot, -1);
503
	local_bh_enable();
L
Linus Torvalds 已提交
504
#ifdef UNIX_REFCNT_DEBUG
W
wangweidong 已提交
505
	pr_debug("UNIX %p is destroyed, %ld are still alive.\n", sk,
E
Eric Dumazet 已提交
506
		atomic_long_read(&unix_nr_socks));
L
Linus Torvalds 已提交
507 508 509
#endif
}

510
static void unix_release_sock(struct sock *sk, int embrion)
L
Linus Torvalds 已提交
511 512
{
	struct unix_sock *u = unix_sk(sk);
A
Al Viro 已提交
513
	struct path path;
L
Linus Torvalds 已提交
514 515 516 517 518 519 520
	struct sock *skpair;
	struct sk_buff *skb;
	int state;

	unix_remove_socket(sk);

	/* Clear state */
521
	unix_state_lock(sk);
L
Linus Torvalds 已提交
522 523
	sock_orphan(sk);
	sk->sk_shutdown = SHUTDOWN_MASK;
A
Al Viro 已提交
524 525 526
	path	     = u->path;
	u->path.dentry = NULL;
	u->path.mnt = NULL;
L
Linus Torvalds 已提交
527 528
	state = sk->sk_state;
	sk->sk_state = TCP_CLOSE;
529
	unix_state_unlock(sk);
L
Linus Torvalds 已提交
530 531 532

	wake_up_interruptible_all(&u->peer_wait);

533
	skpair = unix_peer(sk);
L
Linus Torvalds 已提交
534

535
	if (skpair != NULL) {
L
Linus Torvalds 已提交
536
		if (sk->sk_type == SOCK_STREAM || sk->sk_type == SOCK_SEQPACKET) {
537
			unix_state_lock(skpair);
L
Linus Torvalds 已提交
538 539 540 541
			/* No more writes */
			skpair->sk_shutdown = SHUTDOWN_MASK;
			if (!skb_queue_empty(&sk->sk_receive_queue) || embrion)
				skpair->sk_err = ECONNRESET;
542
			unix_state_unlock(skpair);
L
Linus Torvalds 已提交
543
			skpair->sk_state_change(skpair);
544
			sk_wake_async(skpair, SOCK_WAKE_WAITD, POLL_HUP);
L
Linus Torvalds 已提交
545
		}
546 547

		unix_dgram_peer_wake_disconnect(sk, skpair);
L
Linus Torvalds 已提交
548 549 550 551 552 553 554
		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) {
555
		if (state == TCP_LISTEN)
L
Linus Torvalds 已提交
556 557
			unix_release_sock(skb->sk, 1);
		/* passed fds are erased in the kfree_skb hook	      */
558
		UNIXCB(skb).consumed = skb->len;
L
Linus Torvalds 已提交
559 560 561
		kfree_skb(skb);
	}

A
Al Viro 已提交
562 563
	if (path.dentry)
		path_put(&path);
L
Linus Torvalds 已提交
564 565 566 567 568 569

	sock_put(sk);

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

	/*
A
Alan Cox 已提交
570
	 * Fixme: BSD difference: In BSD all sockets connected to us get
L
Linus Torvalds 已提交
571 572 573 574 575 576 577 578 579
	 *	  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)
	 */

580
	if (unix_tot_inflight)
581
		unix_gc();		/* Garbage collect fds */
L
Linus Torvalds 已提交
582 583
}

584 585 586 587 588 589 590 591 592 593 594 595 596 597 598 599 600 601
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
602 603 604 605 606
static int unix_listen(struct socket *sock, int backlog)
{
	int err;
	struct sock *sk = sock->sk;
	struct unix_sock *u = unix_sk(sk);
607
	struct pid *old_pid = NULL;
L
Linus Torvalds 已提交
608 609

	err = -EOPNOTSUPP;
E
Eric Dumazet 已提交
610 611
	if (sock->type != SOCK_STREAM && sock->type != SOCK_SEQPACKET)
		goto out;	/* Only stream/seqpacket sockets accept */
L
Linus Torvalds 已提交
612 613
	err = -EINVAL;
	if (!u->addr)
E
Eric Dumazet 已提交
614
		goto out;	/* No listens on an unbound socket */
615
	unix_state_lock(sk);
L
Linus Torvalds 已提交
616 617 618 619 620 621 622
	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 */
623
	init_peercred(sk);
L
Linus Torvalds 已提交
624 625 626
	err = 0;

out_unlock:
627
	unix_state_unlock(sk);
628
	put_pid(old_pid);
L
Linus Torvalds 已提交
629 630 631 632 633 634 635 636 637 638 639 640
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 *);
static int unix_accept(struct socket *, struct socket *, int);
static int unix_getname(struct socket *, struct sockaddr *, int *, int);
static unsigned int unix_poll(struct file *, struct socket *, poll_table *);
641 642
static unsigned int unix_dgram_poll(struct file *, struct socket *,
				    poll_table *);
L
Linus Torvalds 已提交
643 644
static int unix_ioctl(struct socket *, unsigned int, unsigned long);
static int unix_shutdown(struct socket *, int);
645 646
static int unix_stream_sendmsg(struct socket *, struct msghdr *, size_t);
static int unix_stream_recvmsg(struct socket *, struct msghdr *, size_t, int);
647 648
static ssize_t unix_stream_sendpage(struct socket *, struct page *, int offset,
				    size_t size, int flags);
649 650 651
static ssize_t unix_stream_splice_read(struct socket *,  loff_t *ppos,
				       struct pipe_inode_info *, size_t size,
				       unsigned int flags);
652 653
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
654 655
static int unix_dgram_connect(struct socket *, struct sockaddr *,
			      int, int);
656 657 658
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
659

660
static int unix_set_peek_off(struct sock *sk, int val)
661 662 663
{
	struct unix_sock *u = unix_sk(sk);

664 665 666
	if (mutex_lock_interruptible(&u->readlock))
		return -EINTR;

667 668
	sk->sk_peek_off = val;
	mutex_unlock(&u->readlock);
669 670

	return 0;
671 672 673
}


674
static const struct proto_ops unix_stream_ops = {
L
Linus Torvalds 已提交
675 676 677 678 679 680 681 682 683 684 685 686 687 688 689 690 691
	.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,
692
	.sendpage =	unix_stream_sendpage,
693
	.splice_read =	unix_stream_splice_read,
694
	.set_peek_off =	unix_set_peek_off,
L
Linus Torvalds 已提交
695 696
};

697
static const struct proto_ops unix_dgram_ops = {
L
Linus Torvalds 已提交
698 699 700 701 702 703 704 705
	.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,
706
	.poll =		unix_dgram_poll,
L
Linus Torvalds 已提交
707 708 709 710 711 712 713 714 715
	.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,
716
	.set_peek_off =	unix_set_peek_off,
L
Linus Torvalds 已提交
717 718
};

719
static const struct proto_ops unix_seqpacket_ops = {
L
Linus Torvalds 已提交
720 721 722 723 724 725 726 727
	.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,
728
	.poll =		unix_dgram_poll,
L
Linus Torvalds 已提交
729 730 731 732 733 734
	.ioctl =	unix_ioctl,
	.listen =	unix_listen,
	.shutdown =	unix_shutdown,
	.setsockopt =	sock_no_setsockopt,
	.getsockopt =	sock_no_getsockopt,
	.sendmsg =	unix_seqpacket_sendmsg,
735
	.recvmsg =	unix_seqpacket_recvmsg,
L
Linus Torvalds 已提交
736 737
	.mmap =		sock_no_mmap,
	.sendpage =	sock_no_sendpage,
738
	.set_peek_off =	unix_set_peek_off,
L
Linus Torvalds 已提交
739 740 741
};

static struct proto unix_proto = {
742 743 744
	.name			= "UNIX",
	.owner			= THIS_MODULE,
	.obj_size		= sizeof(struct unix_sock),
L
Linus Torvalds 已提交
745 746
};

747 748 749 750 751 752 753 754
/*
 * 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;

755
static struct sock *unix_create1(struct net *net, struct socket *sock, int kern)
L
Linus Torvalds 已提交
756 757 758 759
{
	struct sock *sk = NULL;
	struct unix_sock *u;

E
Eric Dumazet 已提交
760 761
	atomic_long_inc(&unix_nr_socks);
	if (atomic_long_read(&unix_nr_socks) > 2 * get_max_files())
L
Linus Torvalds 已提交
762 763
		goto out;

764
	sk = sk_alloc(net, PF_UNIX, GFP_KERNEL, &unix_proto, kern);
L
Linus Torvalds 已提交
765 766 767
	if (!sk)
		goto out;

E
Eric Dumazet 已提交
768
	sock_init_data(sock, sk);
769 770
	lockdep_set_class(&sk->sk_receive_queue.lock,
				&af_unix_sk_receive_queue_lock_key);
L
Linus Torvalds 已提交
771 772

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

796 797
static int unix_create(struct net *net, struct socket *sock, int protocol,
		       int kern)
L
Linus Torvalds 已提交
798 799 800 801 802 803 804 805 806 807 808 809 810 811 812
{
	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:
813
		sock->type = SOCK_DGRAM;
L
Linus Torvalds 已提交
814 815 816 817 818 819 820 821 822 823
	case SOCK_DGRAM:
		sock->ops = &unix_dgram_ops;
		break;
	case SOCK_SEQPACKET:
		sock->ops = &unix_seqpacket_ops;
		break;
	default:
		return -ESOCKTNOSUPPORT;
	}

824
	return unix_create1(net, sock, kern) ? 0 : -ENOMEM;
L
Linus Torvalds 已提交
825 826 827 828 829 830 831 832 833
}

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

	if (!sk)
		return 0;

834
	unix_release_sock(sk, 0);
L
Linus Torvalds 已提交
835 836
	sock->sk = NULL;

837
	return 0;
L
Linus Torvalds 已提交
838 839 840 841 842
}

static int unix_autobind(struct socket *sock)
{
	struct sock *sk = sock->sk;
843
	struct net *net = sock_net(sk);
L
Linus Torvalds 已提交
844 845
	struct unix_sock *u = unix_sk(sk);
	static u32 ordernum = 1;
E
Eric Dumazet 已提交
846
	struct unix_address *addr;
L
Linus Torvalds 已提交
847
	int err;
848
	unsigned int retries = 0;
L
Linus Torvalds 已提交
849

850 851 852
	err = mutex_lock_interruptible(&u->readlock);
	if (err)
		return err;
L
Linus Torvalds 已提交
853 854 855 856 857 858

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

	err = -ENOMEM;
859
	addr = kzalloc(sizeof(*addr) + sizeof(short) + 16, GFP_KERNEL);
L
Linus Torvalds 已提交
860 861 862 863 864 865 866 867
	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);
868
	addr->hash = unix_hash_fold(csum_partial(addr->name, addr->len, 0));
L
Linus Torvalds 已提交
869

870
	spin_lock(&unix_table_lock);
L
Linus Torvalds 已提交
871 872
	ordernum = (ordernum+1)&0xFFFFF;

873
	if (__unix_find_socket_byname(net, addr->name, addr->len, sock->type,
L
Linus Torvalds 已提交
874
				      addr->hash)) {
875
		spin_unlock(&unix_table_lock);
876 877 878 879 880 881 882 883 884 885 886
		/*
		 * __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
887 888 889 890 891 892 893
		goto retry;
	}
	addr->hash ^= sk->sk_type;

	__unix_remove_socket(sk);
	u->addr = addr;
	__unix_insert_socket(&unix_socket_table[addr->hash], sk);
894
	spin_unlock(&unix_table_lock);
L
Linus Torvalds 已提交
895 896
	err = 0;

I
Ingo Molnar 已提交
897
out:	mutex_unlock(&u->readlock);
L
Linus Torvalds 已提交
898 899 900
	return err;
}

901 902
static struct sock *unix_find_other(struct net *net,
				    struct sockaddr_un *sunname, int len,
903
				    int type, unsigned int hash, int *error)
L
Linus Torvalds 已提交
904 905
{
	struct sock *u;
906
	struct path path;
L
Linus Torvalds 已提交
907
	int err = 0;
908

L
Linus Torvalds 已提交
909
	if (sunname->sun_path[0]) {
910 911
		struct inode *inode;
		err = kern_path(sunname->sun_path, LOOKUP_FOLLOW, &path);
L
Linus Torvalds 已提交
912 913
		if (err)
			goto fail;
914
		inode = d_backing_inode(path.dentry);
915
		err = inode_permission(inode, MAY_WRITE);
L
Linus Torvalds 已提交
916 917 918 919
		if (err)
			goto put_fail;

		err = -ECONNREFUSED;
920
		if (!S_ISSOCK(inode->i_mode))
L
Linus Torvalds 已提交
921
			goto put_fail;
922
		u = unix_find_socket_byinode(inode);
L
Linus Torvalds 已提交
923 924 925 926
		if (!u)
			goto put_fail;

		if (u->sk_type == type)
A
Al Viro 已提交
927
			touch_atime(&path);
L
Linus Torvalds 已提交
928

929
		path_put(&path);
L
Linus Torvalds 已提交
930

931
		err = -EPROTOTYPE;
L
Linus Torvalds 已提交
932 933 934 935 936 937
		if (u->sk_type != type) {
			sock_put(u);
			goto fail;
		}
	} else {
		err = -ECONNREFUSED;
938
		u = unix_find_socket_byname(net, sunname, len, type, hash);
L
Linus Torvalds 已提交
939 940
		if (u) {
			struct dentry *dentry;
A
Al Viro 已提交
941
			dentry = unix_sk(u)->path.dentry;
L
Linus Torvalds 已提交
942
			if (dentry)
A
Al Viro 已提交
943
				touch_atime(&unix_sk(u)->path);
L
Linus Torvalds 已提交
944 945 946 947 948 949
		} else
			goto fail;
	}
	return u;

put_fail:
950
	path_put(&path);
L
Linus Torvalds 已提交
951
fail:
952
	*error = err;
L
Linus Torvalds 已提交
953 954 955
	return NULL;
}

A
Al Viro 已提交
956 957 958 959 960 961 962 963 964 965 966 967 968 969 970 971 972 973 974
static int unix_mknod(const char *sun_path, umode_t mode, struct path *res)
{
	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;

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

static int unix_bind(struct socket *sock, struct sockaddr *uaddr, int addr_len)
{
	struct sock *sk = sock->sk;
988
	struct net *net = sock_net(sk);
L
Linus Torvalds 已提交
989
	struct unix_sock *u = unix_sk(sk);
990
	struct sockaddr_un *sunaddr = (struct sockaddr_un *)uaddr;
991
	char *sun_path = sunaddr->sun_path;
L
Linus Torvalds 已提交
992
	int err;
993
	unsigned int hash;
L
Linus Torvalds 已提交
994 995 996 997 998 999 1000
	struct unix_address *addr;
	struct hlist_head *list;

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

1001
	if (addr_len == sizeof(short)) {
L
Linus Torvalds 已提交
1002 1003 1004 1005 1006 1007 1008 1009 1010
		err = unix_autobind(sock);
		goto out;
	}

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

1011 1012 1013
	err = mutex_lock_interruptible(&u->readlock);
	if (err)
		goto out;
L
Linus Torvalds 已提交
1014 1015 1016 1017 1018 1019 1020 1021 1022 1023 1024 1025 1026 1027 1028

	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);

1029
	if (sun_path[0]) {
A
Al Viro 已提交
1030 1031
		struct path path;
		umode_t mode = S_IFSOCK |
A
Al Viro 已提交
1032
		       (SOCK_INODE(sock)->i_mode & ~current_umask());
A
Al Viro 已提交
1033 1034 1035 1036 1037 1038 1039
		err = unix_mknod(sun_path, mode, &path);
		if (err) {
			if (err == -EEXIST)
				err = -EADDRINUSE;
			unix_release_addr(addr);
			goto out_up;
		}
L
Linus Torvalds 已提交
1040
		addr->hash = UNIX_HASH_SIZE;
1041
		hash = d_backing_inode(path.dentry)->i_ino & (UNIX_HASH_SIZE-1);
A
Al Viro 已提交
1042 1043 1044 1045 1046
		spin_lock(&unix_table_lock);
		u->path = path;
		list = &unix_socket_table[hash];
	} else {
		spin_lock(&unix_table_lock);
L
Linus Torvalds 已提交
1047
		err = -EADDRINUSE;
1048
		if (__unix_find_socket_byname(net, sunaddr, addr_len,
L
Linus Torvalds 已提交
1049 1050 1051 1052 1053 1054 1055 1056 1057 1058 1059 1060 1061 1062
					      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:
1063
	spin_unlock(&unix_table_lock);
L
Linus Torvalds 已提交
1064
out_up:
I
Ingo Molnar 已提交
1065
	mutex_unlock(&u->readlock);
L
Linus Torvalds 已提交
1066 1067 1068 1069
out:
	return err;
}

1070 1071 1072 1073 1074 1075 1076 1077 1078 1079 1080 1081 1082 1083 1084 1085 1086 1087 1088 1089 1090 1091 1092 1093 1094
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
1095 1096 1097 1098
static int unix_dgram_connect(struct socket *sock, struct sockaddr *addr,
			      int alen, int flags)
{
	struct sock *sk = sock->sk;
1099
	struct net *net = sock_net(sk);
1100
	struct sockaddr_un *sunaddr = (struct sockaddr_un *)addr;
L
Linus Torvalds 已提交
1101
	struct sock *other;
1102
	unsigned int hash;
L
Linus Torvalds 已提交
1103 1104 1105 1106 1107 1108 1109 1110 1111 1112 1113 1114
	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;

1115
restart:
1116
		other = unix_find_other(net, sunaddr, alen, sock->type, hash, &err);
L
Linus Torvalds 已提交
1117 1118 1119
		if (!other)
			goto out;

1120 1121 1122 1123 1124 1125 1126 1127
		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
1128 1129 1130 1131 1132 1133 1134 1135 1136 1137 1138 1139 1140 1141

		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;
1142
		unix_state_double_lock(sk, other);
L
Linus Torvalds 已提交
1143 1144 1145 1146 1147 1148 1149
	}

	/*
	 * If it was connected, reconnect.
	 */
	if (unix_peer(sk)) {
		struct sock *old_peer = unix_peer(sk);
1150
		unix_peer(sk) = other;
1151 1152
		unix_dgram_peer_wake_disconnect_wakeup(sk, old_peer);

1153
		unix_state_double_unlock(sk, other);
L
Linus Torvalds 已提交
1154 1155 1156 1157 1158

		if (other != old_peer)
			unix_dgram_disconnected(sk, old_peer);
		sock_put(old_peer);
	} else {
1159
		unix_peer(sk) = other;
1160
		unix_state_double_unlock(sk, other);
L
Linus Torvalds 已提交
1161
	}
1162
	return 0;
L
Linus Torvalds 已提交
1163 1164

out_unlock:
1165
	unix_state_double_unlock(sk, other);
L
Linus Torvalds 已提交
1166 1167 1168 1169 1170 1171 1172 1173 1174 1175 1176 1177 1178 1179 1180
	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) &&
1181
		unix_recvq_full(other);
L
Linus Torvalds 已提交
1182

1183
	unix_state_unlock(other);
L
Linus Torvalds 已提交
1184 1185 1186 1187 1188 1189 1190 1191 1192 1193 1194

	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)
{
1195
	struct sockaddr_un *sunaddr = (struct sockaddr_un *)uaddr;
L
Linus Torvalds 已提交
1196
	struct sock *sk = sock->sk;
1197
	struct net *net = sock_net(sk);
L
Linus Torvalds 已提交
1198 1199 1200 1201
	struct unix_sock *u = unix_sk(sk), *newu, *otheru;
	struct sock *newsk = NULL;
	struct sock *other = NULL;
	struct sk_buff *skb = NULL;
1202
	unsigned int hash;
L
Linus Torvalds 已提交
1203 1204 1205 1206 1207 1208 1209 1210 1211
	int st;
	int err;
	long timeo;

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

1212 1213
	if (test_bit(SOCK_PASSCRED, &sock->flags) && !u->addr &&
	    (err = unix_autobind(sock)) != 0)
L
Linus Torvalds 已提交
1214 1215 1216 1217 1218 1219 1220 1221 1222 1223 1224 1225
		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 */
1226
	newsk = unix_create1(sock_net(sk), NULL, 0);
L
Linus Torvalds 已提交
1227 1228 1229 1230 1231 1232 1233 1234 1235 1236
	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. */
1237
	other = unix_find_other(net, sunaddr, addr_len, sk->sk_type, hash, &err);
L
Linus Torvalds 已提交
1238 1239 1240 1241
	if (!other)
		goto out;

	/* Latch state of peer */
1242
	unix_state_lock(other);
L
Linus Torvalds 已提交
1243 1244 1245

	/* Apparently VFS overslept socket death. Retry. */
	if (sock_flag(other, SOCK_DEAD)) {
1246
		unix_state_unlock(other);
L
Linus Torvalds 已提交
1247 1248 1249 1250 1251 1252 1253
		sock_put(other);
		goto restart;
	}

	err = -ECONNREFUSED;
	if (other->sk_state != TCP_LISTEN)
		goto out_unlock;
1254 1255
	if (other->sk_shutdown & RCV_SHUTDOWN)
		goto out_unlock;
L
Linus Torvalds 已提交
1256

1257
	if (unix_recvq_full(other)) {
L
Linus Torvalds 已提交
1258 1259 1260 1261 1262 1263 1264 1265 1266 1267 1268
		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;
1269
	}
L
Linus Torvalds 已提交
1270 1271 1272

	/* Latch our state.

D
Daniel Baluta 已提交
1273
	   It is tricky place. We need to grab our state lock and cannot
L
Linus Torvalds 已提交
1274 1275 1276 1277 1278 1279 1280 1281 1282 1283 1284 1285 1286 1287 1288 1289 1290 1291 1292 1293 1294 1295 1296
	   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;
	}

1297
	unix_state_lock_nested(sk);
L
Linus Torvalds 已提交
1298 1299

	if (sk->sk_state != st) {
1300 1301
		unix_state_unlock(sk);
		unix_state_unlock(other);
L
Linus Torvalds 已提交
1302 1303 1304 1305
		sock_put(other);
		goto restart;
	}

1306
	err = security_unix_stream_connect(sk, other, newsk);
L
Linus Torvalds 已提交
1307
	if (err) {
1308
		unix_state_unlock(sk);
L
Linus Torvalds 已提交
1309 1310 1311 1312 1313 1314 1315 1316 1317
		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;
1318
	init_peercred(newsk);
L
Linus Torvalds 已提交
1319
	newu = unix_sk(newsk);
1320
	RCU_INIT_POINTER(newsk->sk_wq, &newu->peer_wq);
L
Linus Torvalds 已提交
1321 1322 1323 1324 1325 1326 1327
	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
1328 1329 1330
	if (otheru->path.dentry) {
		path_get(&otheru->path);
		newu->path = otheru->path;
L
Linus Torvalds 已提交
1331 1332 1333
	}

	/* Set credentials */
1334
	copy_peercred(sk, other);
L
Linus Torvalds 已提交
1335 1336 1337

	sock->state	= SS_CONNECTED;
	sk->sk_state	= TCP_ESTABLISHED;
1338 1339
	sock_hold(newsk);

1340
	smp_mb__after_atomic();	/* sock_hold() does an atomic_inc() */
1341
	unix_peer(sk)	= newsk;
L
Linus Torvalds 已提交
1342

1343
	unix_state_unlock(sk);
L
Linus Torvalds 已提交
1344 1345 1346 1347 1348

	/* 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);
1349
	unix_state_unlock(other);
1350
	other->sk_data_ready(other);
L
Linus Torvalds 已提交
1351 1352 1353 1354 1355
	sock_put(other);
	return 0;

out_unlock:
	if (other)
1356
		unix_state_unlock(other);
L
Linus Torvalds 已提交
1357 1358

out:
1359
	kfree_skb(skb);
L
Linus Torvalds 已提交
1360 1361 1362 1363 1364 1365 1366 1367 1368
	if (newsk)
		unix_release_sock(newsk, 0);
	if (other)
		sock_put(other);
	return err;
}

static int unix_socketpair(struct socket *socka, struct socket *sockb)
{
1369
	struct sock *ska = socka->sk, *skb = sockb->sk;
L
Linus Torvalds 已提交
1370 1371 1372 1373

	/* Join our sockets back to back */
	sock_hold(ska);
	sock_hold(skb);
1374 1375
	unix_peer(ska) = skb;
	unix_peer(skb) = ska;
1376 1377
	init_peercred(ska);
	init_peercred(skb);
L
Linus Torvalds 已提交
1378 1379 1380 1381 1382 1383 1384 1385 1386 1387

	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;
}

1388 1389 1390 1391 1392 1393 1394 1395 1396
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);
}

L
Linus Torvalds 已提交
1397 1398 1399 1400 1401 1402 1403 1404
static int unix_accept(struct socket *sock, struct socket *newsock, int flags)
{
	struct sock *sk = sock->sk;
	struct sock *tsk;
	struct sk_buff *skb;
	int err;

	err = -EOPNOTSUPP;
E
Eric Dumazet 已提交
1405
	if (sock->type != SOCK_STREAM && sock->type != SOCK_SEQPACKET)
L
Linus Torvalds 已提交
1406 1407 1408 1409 1410 1411 1412 1413 1414 1415 1416 1417 1418 1419 1420 1421 1422 1423 1424 1425 1426 1427 1428
		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 */
1429
	unix_state_lock(tsk);
L
Linus Torvalds 已提交
1430
	newsock->state = SS_CONNECTED;
1431
	unix_sock_inherit_flags(sock, newsock);
L
Linus Torvalds 已提交
1432
	sock_graft(tsk, newsock);
1433
	unix_state_unlock(tsk);
L
Linus Torvalds 已提交
1434 1435 1436 1437 1438 1439 1440 1441 1442 1443 1444
	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;
1445
	DECLARE_SOCKADDR(struct sockaddr_un *, sunaddr, uaddr);
L
Linus Torvalds 已提交
1446 1447 1448 1449 1450 1451 1452 1453 1454 1455 1456 1457 1458 1459
	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);
1460
	unix_state_lock(sk);
L
Linus Torvalds 已提交
1461 1462 1463 1464 1465 1466 1467 1468 1469 1470
	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);
	}
1471
	unix_state_unlock(sk);
L
Linus Torvalds 已提交
1472 1473 1474 1475 1476 1477 1478 1479 1480 1481 1482 1483
	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
1484
	for (i = scm->fp->count-1; i >= 0; i--)
L
Linus Torvalds 已提交
1485 1486 1487
		unix_notinflight(scm->fp->fp[i]);
}

1488
static void unix_destruct_scm(struct sk_buff *skb)
L
Linus Torvalds 已提交
1489 1490 1491
{
	struct scm_cookie scm;
	memset(&scm, 0, sizeof(scm));
1492 1493 1494
	scm.pid  = UNIXCB(skb).pid;
	if (UNIXCB(skb).fp)
		unix_detach_fds(&scm, skb);
L
Linus Torvalds 已提交
1495 1496 1497 1498 1499 1500 1501

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

E
Eric Dumazet 已提交
1502 1503
#define MAX_RECURSION_LEVEL 4

1504
static int unix_attach_fds(struct scm_cookie *scm, struct sk_buff *skb)
L
Linus Torvalds 已提交
1505 1506
{
	int i;
E
Eric Dumazet 已提交
1507 1508 1509 1510 1511 1512 1513 1514 1515 1516 1517 1518 1519 1520
	unsigned char max_level = 0;
	int unix_sock_count = 0;

	for (i = scm->fp->count - 1; i >= 0; i--) {
		struct sock *sk = unix_get_socket(scm->fp->fp[i]);

		if (sk) {
			unix_sock_count++;
			max_level = max(max_level,
					unix_sk(sk)->recursion_level);
		}
	}
	if (unlikely(max_level > MAX_RECURSION_LEVEL))
		return -ETOOMANYREFS;
1521 1522 1523 1524 1525 1526 1527 1528 1529 1530

	/*
	 * 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;

E
Eric Dumazet 已提交
1531 1532 1533 1534 1535
	if (unix_sock_count) {
		for (i = scm->fp->count - 1; i >= 0; i--)
			unix_inflight(scm->fp->fp[i]);
	}
	return max_level;
L
Linus Torvalds 已提交
1536 1537
}

1538
static int unix_scm_to_skb(struct scm_cookie *scm, struct sk_buff *skb, bool send_fds)
1539 1540
{
	int err = 0;
1541

1542
	UNIXCB(skb).pid  = get_pid(scm->pid);
1543 1544
	UNIXCB(skb).uid = scm->creds.uid;
	UNIXCB(skb).gid = scm->creds.gid;
1545
	UNIXCB(skb).fp = NULL;
1546
	unix_get_secdata(scm, skb);
1547 1548 1549 1550 1551 1552 1553
	if (scm->fp && send_fds)
		err = unix_attach_fds(scm, skb);

	skb->destructor = unix_destruct_scm;
	return err;
}

1554 1555 1556 1557 1558 1559 1560 1561
/*
 * 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)
{
1562
	if (UNIXCB(skb).pid)
1563 1564
		return;
	if (test_bit(SOCK_PASSCRED, &sock->flags) ||
1565 1566
	    !other->sk_socket ||
	    test_bit(SOCK_PASSCRED, &other->sk_socket->flags)) {
1567
		UNIXCB(skb).pid  = get_pid(task_tgid(current));
1568
		current_uid_gid(&UNIXCB(skb).uid, &UNIXCB(skb).gid);
1569 1570 1571
	}
}

L
Linus Torvalds 已提交
1572 1573 1574 1575
/*
 *	Send AF_UNIX data.
 */

1576 1577
static int unix_dgram_sendmsg(struct socket *sock, struct msghdr *msg,
			      size_t len)
L
Linus Torvalds 已提交
1578 1579
{
	struct sock *sk = sock->sk;
1580
	struct net *net = sock_net(sk);
L
Linus Torvalds 已提交
1581
	struct unix_sock *u = unix_sk(sk);
1582
	DECLARE_SOCKADDR(struct sockaddr_un *, sunaddr, msg->msg_name);
L
Linus Torvalds 已提交
1583 1584 1585
	struct sock *other = NULL;
	int namelen = 0; /* fake GCC */
	int err;
1586
	unsigned int hash;
1587
	struct sk_buff *skb;
L
Linus Torvalds 已提交
1588
	long timeo;
C
Christoph Hellwig 已提交
1589
	struct scm_cookie scm;
E
Eric Dumazet 已提交
1590
	int max_level;
1591
	int data_len = 0;
1592
	int sk_locked;
L
Linus Torvalds 已提交
1593

1594
	wait_for_unix_gc();
C
Christoph Hellwig 已提交
1595
	err = scm_send(sock, msg, &scm, false);
L
Linus Torvalds 已提交
1596 1597 1598 1599 1600 1601 1602 1603 1604 1605 1606 1607 1608 1609 1610 1611 1612 1613 1614 1615
	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;
	}

1616 1617
	if (test_bit(SOCK_PASSCRED, &sock->flags) && !u->addr
	    && (err = unix_autobind(sock)) != 0)
L
Linus Torvalds 已提交
1618 1619 1620 1621 1622 1623
		goto out;

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

1624
	if (len > SKB_MAX_ALLOC) {
1625 1626 1627
		data_len = min_t(size_t,
				 len - SKB_MAX_ALLOC,
				 MAX_SKB_FRAGS * PAGE_SIZE);
1628 1629 1630 1631
		data_len = PAGE_ALIGN(data_len);

		BUILD_BUG_ON(SKB_MAX_ALLOC < PAGE_SIZE);
	}
1632 1633

	skb = sock_alloc_send_pskb(sk, len - data_len, data_len,
1634 1635
				   msg->msg_flags & MSG_DONTWAIT, &err,
				   PAGE_ALLOC_COSTLY_ORDER);
1636
	if (skb == NULL)
L
Linus Torvalds 已提交
1637 1638
		goto out;

C
Christoph Hellwig 已提交
1639
	err = unix_scm_to_skb(&scm, skb, true);
E
Eric Dumazet 已提交
1640
	if (err < 0)
1641
		goto out_free;
E
Eric Dumazet 已提交
1642
	max_level = err + 1;
C
Catherine Zhang 已提交
1643

1644 1645 1646
	skb_put(skb, len - data_len);
	skb->data_len = data_len;
	skb->len = len;
A
Al Viro 已提交
1647
	err = skb_copy_datagram_from_iter(skb, 0, &msg->msg_iter, len);
L
Linus Torvalds 已提交
1648 1649 1650 1651 1652 1653 1654 1655 1656 1657 1658
	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;

1659
		other = unix_find_other(net, sunaddr, namelen, sk->sk_type,
L
Linus Torvalds 已提交
1660
					hash, &err);
1661
		if (other == NULL)
L
Linus Torvalds 已提交
1662 1663 1664
			goto out_free;
	}

A
Alban Crequy 已提交
1665 1666 1667 1668 1669 1670
	if (sk_filter(other, skb) < 0) {
		/* Toss the packet but do not return any error to the sender */
		err = len;
		goto out_free;
	}

1671
	sk_locked = 0;
1672
	unix_state_lock(other);
1673
restart_locked:
L
Linus Torvalds 已提交
1674 1675 1676 1677
	err = -EPERM;
	if (!unix_may_send(sk, other))
		goto out_unlock;

1678
	if (unlikely(sock_flag(other, SOCK_DEAD))) {
L
Linus Torvalds 已提交
1679 1680 1681 1682
		/*
		 *	Check with 1003.1g - what should
		 *	datagram error
		 */
1683
		unix_state_unlock(other);
L
Linus Torvalds 已提交
1684 1685
		sock_put(other);

1686 1687 1688
		if (!sk_locked)
			unix_state_lock(sk);

L
Linus Torvalds 已提交
1689 1690
		err = 0;
		if (unix_peer(sk) == other) {
1691
			unix_peer(sk) = NULL;
1692 1693
			unix_dgram_peer_wake_disconnect_wakeup(sk, other);

1694
			unix_state_unlock(sk);
L
Linus Torvalds 已提交
1695 1696 1697 1698 1699

			unix_dgram_disconnected(sk, other);
			sock_put(other);
			err = -ECONNREFUSED;
		} else {
1700
			unix_state_unlock(sk);
L
Linus Torvalds 已提交
1701 1702 1703 1704 1705 1706 1707 1708 1709 1710 1711 1712 1713 1714 1715 1716 1717 1718
		}

		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;
	}

1719 1720 1721 1722 1723 1724 1725 1726 1727
	if (unlikely(unix_peer(other) != sk && unix_recvq_full(other))) {
		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
1728 1729
		}

1730 1731 1732 1733
		if (!sk_locked) {
			unix_state_unlock(other);
			unix_state_double_lock(sk, other);
		}
L
Linus Torvalds 已提交
1734

1735 1736 1737 1738 1739 1740
		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
1741

1742 1743 1744 1745
		if (!sk_locked) {
			sk_locked = 1;
			goto restart_locked;
		}
L
Linus Torvalds 已提交
1746 1747
	}

1748 1749 1750
	if (unlikely(sk_locked))
		unix_state_unlock(sk);

1751 1752
	if (sock_flag(other, SOCK_RCVTSTAMP))
		__net_timestamp(skb);
1753
	maybe_add_creds(skb, sock, other);
L
Linus Torvalds 已提交
1754
	skb_queue_tail(&other->sk_receive_queue, skb);
E
Eric Dumazet 已提交
1755 1756
	if (max_level > unix_sk(other)->recursion_level)
		unix_sk(other)->recursion_level = max_level;
1757
	unix_state_unlock(other);
1758
	other->sk_data_ready(other);
L
Linus Torvalds 已提交
1759
	sock_put(other);
C
Christoph Hellwig 已提交
1760
	scm_destroy(&scm);
L
Linus Torvalds 已提交
1761 1762 1763
	return len;

out_unlock:
1764 1765
	if (sk_locked)
		unix_state_unlock(sk);
1766
	unix_state_unlock(other);
L
Linus Torvalds 已提交
1767 1768 1769 1770 1771
out_free:
	kfree_skb(skb);
out:
	if (other)
		sock_put(other);
C
Christoph Hellwig 已提交
1772
	scm_destroy(&scm);
L
Linus Torvalds 已提交
1773 1774 1775
	return err;
}

1776 1777 1778 1779
/* 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))
1780

1781 1782
static int unix_stream_sendmsg(struct socket *sock, struct msghdr *msg,
			       size_t len)
L
Linus Torvalds 已提交
1783 1784 1785
{
	struct sock *sk = sock->sk;
	struct sock *other = NULL;
E
Eric Dumazet 已提交
1786
	int err, size;
1787
	struct sk_buff *skb;
1788
	int sent = 0;
C
Christoph Hellwig 已提交
1789
	struct scm_cookie scm;
1790
	bool fds_sent = false;
E
Eric Dumazet 已提交
1791
	int max_level;
1792
	int data_len;
L
Linus Torvalds 已提交
1793

1794
	wait_for_unix_gc();
C
Christoph Hellwig 已提交
1795
	err = scm_send(sock, msg, &scm, false);
L
Linus Torvalds 已提交
1796 1797 1798 1799 1800 1801 1802 1803 1804 1805 1806 1807
	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;
1808
		other = unix_peer(sk);
L
Linus Torvalds 已提交
1809 1810 1811 1812 1813 1814 1815
		if (!other)
			goto out_err;
	}

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

E
Eric Dumazet 已提交
1816
	while (sent < len) {
1817
		size = len - sent;
L
Linus Torvalds 已提交
1818 1819

		/* Keep two messages in the pipe so it schedules better */
1820
		size = min_t(int, size, (sk->sk_sndbuf >> 1) - 64);
L
Linus Torvalds 已提交
1821

1822 1823
		/* allow fallback to order-0 allocations */
		size = min_t(int, size, SKB_MAX_HEAD(0) + UNIX_SKB_FRAGS_SZ);
1824

1825
		data_len = max_t(int, 0, size - SKB_MAX_HEAD(0));
L
Linus Torvalds 已提交
1826

1827 1828
		data_len = min_t(size_t, size, PAGE_ALIGN(data_len));

1829
		skb = sock_alloc_send_pskb(sk, size - data_len, data_len,
1830 1831
					   msg->msg_flags & MSG_DONTWAIT, &err,
					   get_order(UNIX_SKB_FRAGS_SZ));
1832
		if (!skb)
L
Linus Torvalds 已提交
1833 1834
			goto out_err;

1835
		/* Only send the fds in the first buffer */
C
Christoph Hellwig 已提交
1836
		err = unix_scm_to_skb(&scm, skb, !fds_sent);
E
Eric Dumazet 已提交
1837
		if (err < 0) {
1838
			kfree_skb(skb);
1839
			goto out_err;
1840
		}
E
Eric Dumazet 已提交
1841
		max_level = err + 1;
1842
		fds_sent = true;
L
Linus Torvalds 已提交
1843

1844 1845 1846
		skb_put(skb, size - data_len);
		skb->data_len = data_len;
		skb->len = size;
A
Al Viro 已提交
1847
		err = skb_copy_datagram_from_iter(skb, 0, &msg->msg_iter, size);
E
Eric Dumazet 已提交
1848
		if (err) {
L
Linus Torvalds 已提交
1849
			kfree_skb(skb);
1850
			goto out_err;
L
Linus Torvalds 已提交
1851 1852
		}

1853
		unix_state_lock(other);
L
Linus Torvalds 已提交
1854 1855 1856 1857 1858

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

1859
		maybe_add_creds(skb, sock, other);
L
Linus Torvalds 已提交
1860
		skb_queue_tail(&other->sk_receive_queue, skb);
E
Eric Dumazet 已提交
1861 1862
		if (max_level > unix_sk(other)->recursion_level)
			unix_sk(other)->recursion_level = max_level;
1863
		unix_state_unlock(other);
1864
		other->sk_data_ready(other);
1865
		sent += size;
L
Linus Torvalds 已提交
1866 1867
	}

C
Christoph Hellwig 已提交
1868
	scm_destroy(&scm);
L
Linus Torvalds 已提交
1869 1870 1871 1872

	return sent;

pipe_err_free:
1873
	unix_state_unlock(other);
L
Linus Torvalds 已提交
1874 1875
	kfree_skb(skb);
pipe_err:
E
Eric Dumazet 已提交
1876 1877
	if (sent == 0 && !(msg->msg_flags&MSG_NOSIGNAL))
		send_sig(SIGPIPE, current, 0);
L
Linus Torvalds 已提交
1878 1879
	err = -EPIPE;
out_err:
C
Christoph Hellwig 已提交
1880
	scm_destroy(&scm);
L
Linus Torvalds 已提交
1881 1882 1883
	return sent ? : err;
}

1884 1885 1886 1887 1888 1889 1890 1891 1892 1893 1894 1895 1896 1897 1898 1899 1900 1901 1902 1903 1904 1905 1906 1907 1908 1909 1910 1911 1912 1913 1914 1915 1916 1917 1918 1919 1920 1921 1922 1923 1924 1925 1926 1927 1928 1929 1930 1931 1932 1933 1934 1935 1936 1937 1938 1939 1940 1941 1942 1943 1944 1945
static ssize_t unix_stream_sendpage(struct socket *socket, struct page *page,
				    int offset, size_t size, int flags)
{
	int err = 0;
	bool send_sigpipe = true;
	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);
		mutex_unlock(&unix_sk(other)->readlock);
		newskb = sock_alloc_send_pskb(sk, 0, 0, flags & MSG_DONTWAIT,
					      &err, 0);
		if (!newskb)
			return err;
	}

	/* we must acquire readlock as we modify already present
	 * skbs in the sk_receive_queue and mess with skb->len
	 */
	err = mutex_lock_interruptible(&unix_sk(other)->readlock);
	if (err) {
		err = flags & MSG_DONTWAIT ? -EAGAIN : -ERESTARTSYS;
		send_sigpipe = false;
		goto err;
	}

	if (sk->sk_shutdown & SEND_SHUTDOWN) {
		err = -EPIPE;
		goto err_unlock;
	}

	unix_state_lock(other);

	if (sock_flag(other, SOCK_DEAD) ||
	    other->sk_shutdown & RCV_SHUTDOWN) {
		err = -EPIPE;
		goto err_state_unlock;
	}

	skb = skb_peek_tail(&other->sk_receive_queue);
	if (tail && tail == skb) {
		skb = newskb;
	} else if (!skb) {
		if (newskb)
			skb = newskb;
		else
			goto alloc_skb;
	} 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);
1946
		newskb = NULL;
1947 1948 1949 1950 1951 1952 1953 1954 1955 1956 1957 1958
	}

	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);

1959 1960
	if (newskb) {
		spin_lock(&other->sk_receive_queue.lock);
1961
		__skb_queue_tail(&other->sk_receive_queue, newskb);
1962 1963
		spin_unlock(&other->sk_receive_queue.lock);
	}
1964 1965 1966 1967 1968 1969 1970 1971 1972 1973 1974 1975 1976 1977 1978 1979 1980 1981 1982

	unix_state_unlock(other);
	mutex_unlock(&unix_sk(other)->readlock);

	other->sk_data_ready(other);

	return size;

err_state_unlock:
	unix_state_unlock(other);
err_unlock:
	mutex_unlock(&unix_sk(other)->readlock);
err:
	kfree_skb(newskb);
	if (send_sigpipe && !(flags & MSG_NOSIGNAL))
		send_sig(SIGPIPE, current, 0);
	return err;
}

1983 1984
static int unix_seqpacket_sendmsg(struct socket *sock, struct msghdr *msg,
				  size_t len)
L
Linus Torvalds 已提交
1985 1986 1987
{
	int err;
	struct sock *sk = sock->sk;
1988

L
Linus Torvalds 已提交
1989 1990 1991 1992 1993 1994 1995 1996 1997 1998
	err = sock_error(sk);
	if (err)
		return err;

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

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

1999
	return unix_dgram_sendmsg(sock, msg, len);
L
Linus Torvalds 已提交
2000
}
2001

2002 2003
static int unix_seqpacket_recvmsg(struct socket *sock, struct msghdr *msg,
				  size_t size, int flags)
2004 2005 2006 2007 2008 2009
{
	struct sock *sk = sock->sk;

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

2010
	return unix_dgram_recvmsg(sock, msg, size, flags);
2011 2012
}

L
Linus Torvalds 已提交
2013 2014 2015 2016 2017 2018 2019 2020 2021 2022
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);
	}
}

2023 2024
static int unix_dgram_recvmsg(struct socket *sock, struct msghdr *msg,
			      size_t size, int flags)
L
Linus Torvalds 已提交
2025
{
C
Christoph Hellwig 已提交
2026
	struct scm_cookie scm;
L
Linus Torvalds 已提交
2027 2028 2029 2030 2031
	struct sock *sk = sock->sk;
	struct unix_sock *u = unix_sk(sk);
	int noblock = flags & MSG_DONTWAIT;
	struct sk_buff *skb;
	int err;
2032
	int peeked, skip;
L
Linus Torvalds 已提交
2033 2034 2035 2036 2037

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

2038
	err = mutex_lock_interruptible(&u->readlock);
2039 2040 2041 2042 2043
	if (unlikely(err)) {
		/* recvmsg() in non blocking mode is supposed to return -EAGAIN
		 * sk_rcvtimeo is not honored by mutex_lock_interruptible()
		 */
		err = noblock ? -EAGAIN : -ERESTARTSYS;
2044 2045
		goto out;
	}
L
Linus Torvalds 已提交
2046

2047 2048 2049
	skip = sk_peek_offset(sk, flags);

	skb = __skb_recv_datagram(sk, flags, &peeked, &skip, &err);
2050 2051 2052 2053 2054 2055 2056
	if (!skb) {
		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);
L
Linus Torvalds 已提交
2057
		goto out_unlock;
2058
	}
L
Linus Torvalds 已提交
2059

E
Eric Dumazet 已提交
2060 2061
	wake_up_interruptible_sync_poll(&u->peer_wait,
					POLLOUT | POLLWRNORM | POLLWRBAND);
L
Linus Torvalds 已提交
2062 2063 2064 2065

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

2066 2067 2068
	if (size > skb->len - skip)
		size = skb->len - skip;
	else if (size < skb->len - skip)
L
Linus Torvalds 已提交
2069 2070
		msg->msg_flags |= MSG_TRUNC;

2071
	err = skb_copy_datagram_msg(skb, skip, msg, size);
L
Linus Torvalds 已提交
2072 2073 2074
	if (err)
		goto out_free;

2075 2076 2077
	if (sock_flag(sk, SOCK_RCVTSTAMP))
		__sock_recv_timestamp(msg, sk, skb);

C
Christoph Hellwig 已提交
2078 2079 2080 2081
	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
2082

E
Eric Dumazet 已提交
2083
	if (!(flags & MSG_PEEK)) {
L
Linus Torvalds 已提交
2084
		if (UNIXCB(skb).fp)
C
Christoph Hellwig 已提交
2085
			unix_detach_fds(&scm, skb);
2086 2087

		sk_peek_offset_bwd(sk, skb->len);
E
Eric Dumazet 已提交
2088
	} else {
L
Linus Torvalds 已提交
2089 2090 2091 2092 2093 2094
		/* 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)
2095 2096 2097 2098 2099

		   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
2100
		*/
2101 2102 2103

		sk_peek_offset_fwd(sk, size);

L
Linus Torvalds 已提交
2104
		if (UNIXCB(skb).fp)
C
Christoph Hellwig 已提交
2105
			scm.fp = scm_fp_dup(UNIXCB(skb).fp);
L
Linus Torvalds 已提交
2106
	}
2107
	err = (flags & MSG_TRUNC) ? skb->len - skip : size;
L
Linus Torvalds 已提交
2108

C
Christoph Hellwig 已提交
2109
	scm_recv(sock, msg, &scm, flags);
L
Linus Torvalds 已提交
2110 2111

out_free:
E
Eric Dumazet 已提交
2112
	skb_free_datagram(sk, skb);
L
Linus Torvalds 已提交
2113
out_unlock:
I
Ingo Molnar 已提交
2114
	mutex_unlock(&u->readlock);
L
Linus Torvalds 已提交
2115 2116 2117 2118 2119
out:
	return err;
}

/*
2120
 *	Sleep until more data has arrived. But check for races..
L
Linus Torvalds 已提交
2121
 */
2122
static long unix_stream_data_wait(struct sock *sk, long timeo,
2123
				  struct sk_buff *last, unsigned int last_len)
L
Linus Torvalds 已提交
2124
{
2125
	struct sk_buff *tail;
L
Linus Torvalds 已提交
2126 2127
	DEFINE_WAIT(wait);

2128
	unix_state_lock(sk);
L
Linus Torvalds 已提交
2129 2130

	for (;;) {
E
Eric Dumazet 已提交
2131
		prepare_to_wait(sk_sleep(sk), &wait, TASK_INTERRUPTIBLE);
L
Linus Torvalds 已提交
2132

2133 2134 2135
		tail = skb_peek_tail(&sk->sk_receive_queue);
		if (tail != last ||
		    (tail && tail->len != last_len) ||
L
Linus Torvalds 已提交
2136 2137 2138 2139 2140 2141 2142
		    sk->sk_err ||
		    (sk->sk_shutdown & RCV_SHUTDOWN) ||
		    signal_pending(current) ||
		    !timeo)
			break;

		set_bit(SOCK_ASYNC_WAITDATA, &sk->sk_socket->flags);
2143
		unix_state_unlock(sk);
2144
		timeo = freezable_schedule_timeout(timeo);
2145
		unix_state_lock(sk);
2146 2147 2148 2149

		if (sock_flag(sk, SOCK_DEAD))
			break;

L
Linus Torvalds 已提交
2150 2151 2152
		clear_bit(SOCK_ASYNC_WAITDATA, &sk->sk_socket->flags);
	}

E
Eric Dumazet 已提交
2153
	finish_wait(sk_sleep(sk), &wait);
2154
	unix_state_unlock(sk);
L
Linus Torvalds 已提交
2155 2156 2157
	return timeo;
}

2158 2159 2160 2161 2162
static unsigned int unix_skb_len(const struct sk_buff *skb)
{
	return skb->len - UNIXCB(skb).consumed;
}

2163 2164 2165 2166 2167 2168 2169 2170 2171 2172 2173 2174
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;
};

static int unix_stream_read_generic(struct unix_stream_read_state *state)
L
Linus Torvalds 已提交
2175
{
C
Christoph Hellwig 已提交
2176
	struct scm_cookie scm;
2177
	struct socket *sock = state->socket;
L
Linus Torvalds 已提交
2178 2179 2180
	struct sock *sk = sock->sk;
	struct unix_sock *u = unix_sk(sk);
	int copied = 0;
2181
	int flags = state->flags;
2182
	int noblock = flags & MSG_DONTWAIT;
2183
	bool check_creds = false;
L
Linus Torvalds 已提交
2184 2185 2186
	int target;
	int err = 0;
	long timeo;
2187
	int skip;
2188 2189
	size_t size = state->size;
	unsigned int last_len;
L
Linus Torvalds 已提交
2190 2191 2192 2193 2194 2195

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

	err = -EOPNOTSUPP;
2196
	if (flags & MSG_OOB)
L
Linus Torvalds 已提交
2197 2198
		goto out;

2199
	target = sock_rcvlowat(sk, flags & MSG_WAITALL, size);
2200
	timeo = sock_rcvtimeo(sk, noblock);
L
Linus Torvalds 已提交
2201

2202 2203
	memset(&scm, 0, sizeof(scm));

L
Linus Torvalds 已提交
2204 2205 2206
	/* Lock the socket to prevent queue disordering
	 * while sleeps in memcpy_tomsg
	 */
2207
	err = mutex_lock_interruptible(&u->readlock);
2208 2209 2210 2211 2212
	if (unlikely(err)) {
		/* recvmsg() in non blocking mode is supposed to return -EAGAIN
		 * sk_rcvtimeo is not honored by mutex_lock_interruptible()
		 */
		err = noblock ? -EAGAIN : -ERESTARTSYS;
2213 2214
		goto out;
	}
L
Linus Torvalds 已提交
2215

2216 2217 2218 2219 2220
	if (flags & MSG_PEEK)
		skip = sk_peek_offset(sk, flags);
	else
		skip = 0;

E
Eric Dumazet 已提交
2221
	do {
L
Linus Torvalds 已提交
2222
		int chunk;
2223
		bool drop_skb;
2224
		struct sk_buff *skb, *last;
L
Linus Torvalds 已提交
2225

2226
		unix_state_lock(sk);
2227 2228 2229 2230
		if (sock_flag(sk, SOCK_DEAD)) {
			err = -ECONNRESET;
			goto unlock;
		}
2231
		last = skb = skb_peek(&sk->sk_receive_queue);
2232
		last_len = last ? last->len : 0;
2233
again:
E
Eric Dumazet 已提交
2234
		if (skb == NULL) {
E
Eric Dumazet 已提交
2235
			unix_sk(sk)->recursion_level = 0;
L
Linus Torvalds 已提交
2236
			if (copied >= target)
2237
				goto unlock;
L
Linus Torvalds 已提交
2238 2239 2240 2241

			/*
			 *	POSIX 1003.1g mandates this order.
			 */
2242

E
Eric Dumazet 已提交
2243 2244
			err = sock_error(sk);
			if (err)
2245
				goto unlock;
L
Linus Torvalds 已提交
2246
			if (sk->sk_shutdown & RCV_SHUTDOWN)
2247 2248 2249
				goto unlock;

			unix_state_unlock(sk);
L
Linus Torvalds 已提交
2250 2251 2252
			err = -EAGAIN;
			if (!timeo)
				break;
I
Ingo Molnar 已提交
2253
			mutex_unlock(&u->readlock);
L
Linus Torvalds 已提交
2254

2255 2256
			timeo = unix_stream_data_wait(sk, timeo, last,
						      last_len);
L
Linus Torvalds 已提交
2257

2258 2259
			if (signal_pending(current) ||
			    mutex_lock_interruptible(&u->readlock)) {
L
Linus Torvalds 已提交
2260 2261 2262
				err = sock_intr_errno(timeo);
				goto out;
			}
2263

L
Linus Torvalds 已提交
2264
			continue;
2265
unlock:
2266 2267
			unix_state_unlock(sk);
			break;
L
Linus Torvalds 已提交
2268
		}
2269

2270 2271
		while (skip >= unix_skb_len(skb)) {
			skip -= unix_skb_len(skb);
2272
			last = skb;
2273
			last_len = skb->len;
2274
			skb = skb_peek_next(skb, &sk->sk_receive_queue);
2275 2276
			if (!skb)
				goto again;
2277 2278
		}

2279
		unix_state_unlock(sk);
L
Linus Torvalds 已提交
2280 2281 2282

		if (check_creds) {
			/* Never glue messages from different writers */
C
Christoph Hellwig 已提交
2283 2284
			if ((UNIXCB(skb).pid  != scm.pid) ||
			    !uid_eq(UNIXCB(skb).uid, scm.creds.uid) ||
2285 2286
			    !gid_eq(UNIXCB(skb).gid, scm.creds.gid) ||
			    !unix_secdata_eq(&scm, skb))
L
Linus Torvalds 已提交
2287
				break;
2288
		} else if (test_bit(SOCK_PASSCRED, &sock->flags)) {
L
Linus Torvalds 已提交
2289
			/* Copy credentials */
C
Christoph Hellwig 已提交
2290
			scm_set_cred(&scm, UNIXCB(skb).pid, UNIXCB(skb).uid, UNIXCB(skb).gid);
2291
			unix_set_secdata(&scm, skb);
2292
			check_creds = true;
L
Linus Torvalds 已提交
2293 2294 2295
		}

		/* Copy address just once */
2296 2297 2298 2299
		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
2300 2301 2302
			sunaddr = NULL;
		}

2303
		chunk = min_t(unsigned int, unix_skb_len(skb) - skip, size);
2304
		skb_get(skb);
2305
		chunk = state->recv_actor(skb, skip, chunk, state);
2306 2307 2308
		drop_skb = !unix_skb_len(skb);
		/* skb is only safe to use if !drop_skb */
		consume_skb(skb);
2309
		if (chunk < 0) {
L
Linus Torvalds 已提交
2310 2311 2312 2313 2314 2315 2316
			if (copied == 0)
				copied = -EFAULT;
			break;
		}
		copied += chunk;
		size -= chunk;

2317 2318 2319 2320 2321 2322 2323 2324 2325 2326 2327 2328
		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
2329
		/* Mark read part of skb as used */
E
Eric Dumazet 已提交
2330
		if (!(flags & MSG_PEEK)) {
2331
			UNIXCB(skb).consumed += chunk;
L
Linus Torvalds 已提交
2332

2333 2334
			sk_peek_offset_bwd(sk, chunk);

L
Linus Torvalds 已提交
2335
			if (UNIXCB(skb).fp)
C
Christoph Hellwig 已提交
2336
				unix_detach_fds(&scm, skb);
L
Linus Torvalds 已提交
2337

2338
			if (unix_skb_len(skb))
L
Linus Torvalds 已提交
2339 2340
				break;

2341
			skb_unlink(skb, &sk->sk_receive_queue);
2342
			consume_skb(skb);
L
Linus Torvalds 已提交
2343

C
Christoph Hellwig 已提交
2344
			if (scm.fp)
L
Linus Torvalds 已提交
2345
				break;
E
Eric Dumazet 已提交
2346
		} else {
L
Linus Torvalds 已提交
2347 2348 2349
			/* It is questionable, see note in unix_dgram_recvmsg.
			 */
			if (UNIXCB(skb).fp)
C
Christoph Hellwig 已提交
2350
				scm.fp = scm_fp_dup(UNIXCB(skb).fp);
L
Linus Torvalds 已提交
2351

2352
			sk_peek_offset_fwd(sk, chunk);
2353

2354 2355 2356
			if (UNIXCB(skb).fp)
				break;

2357
			skip = 0;
2358 2359 2360 2361 2362 2363 2364
			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
2365 2366 2367 2368
			break;
		}
	} while (size);

I
Ingo Molnar 已提交
2369
	mutex_unlock(&u->readlock);
2370 2371 2372 2373
	if (state->msg)
		scm_recv(sock, state->msg, &scm, flags);
	else
		scm_destroy(&scm);
L
Linus Torvalds 已提交
2374 2375 2376 2377
out:
	return copied ? : err;
}

2378 2379 2380 2381 2382 2383 2384 2385 2386 2387 2388 2389 2390 2391 2392 2393 2394 2395 2396 2397 2398 2399 2400 2401 2402 2403 2404 2405 2406 2407 2408 2409 2410 2411 2412 2413 2414 2415 2416 2417 2418 2419 2420 2421 2422 2423 2424 2425 2426 2427 2428 2429 2430 2431 2432 2433 2434 2435 2436 2437 2438 2439 2440 2441 2442 2443 2444 2445 2446 2447 2448
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
	};

	return unix_stream_read_generic(&state);
}

static ssize_t skb_unix_socket_splice(struct sock *sk,
				      struct pipe_inode_info *pipe,
				      struct splice_pipe_desc *spd)
{
	int ret;
	struct unix_sock *u = unix_sk(sk);

	mutex_unlock(&u->readlock);
	ret = splice_to_pipe(pipe, spd);
	mutex_lock(&u->readlock);

	return ret;
}

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,
			       state->pipe, chunk, state->splice_flags,
			       skb_unix_socket_splice);
}

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;

	return unix_stream_read_generic(&state);
}

L
Linus Torvalds 已提交
2449 2450 2451 2452 2453
static int unix_shutdown(struct socket *sock, int mode)
{
	struct sock *sk = sock->sk;
	struct sock *other;

2454 2455 2456 2457 2458 2459 2460 2461
	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;
2462 2463 2464 2465 2466 2467 2468 2469 2470 2471 2472 2473 2474 2475 2476 2477 2478 2479 2480 2481 2482 2483 2484 2485 2486 2487

	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
2488
	}
2489 2490 2491
	if (other)
		sock_put(other);

L
Linus Torvalds 已提交
2492 2493 2494
	return 0;
}

2495 2496 2497 2498 2499 2500 2501 2502 2503 2504 2505 2506
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)
2507
			amount += unix_skb_len(skb);
2508 2509 2510 2511 2512 2513 2514 2515 2516 2517 2518 2519 2520 2521 2522 2523 2524
	} 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);

L
Linus Torvalds 已提交
2525 2526 2527
static int unix_ioctl(struct socket *sock, unsigned int cmd, unsigned long arg)
{
	struct sock *sk = sock->sk;
2528
	long amount = 0;
L
Linus Torvalds 已提交
2529 2530
	int err;

E
Eric Dumazet 已提交
2531 2532
	switch (cmd) {
	case SIOCOUTQ:
2533
		amount = unix_outq_len(sk);
E
Eric Dumazet 已提交
2534 2535 2536
		err = put_user(amount, (int __user *)arg);
		break;
	case SIOCINQ:
2537 2538 2539 2540
		amount = unix_inq_len(sk);
		if (amount < 0)
			err = amount;
		else
L
Linus Torvalds 已提交
2541
			err = put_user(amount, (int __user *)arg);
2542
		break;
E
Eric Dumazet 已提交
2543 2544 2545
	default:
		err = -ENOIOCTLCMD;
		break;
L
Linus Torvalds 已提交
2546 2547 2548 2549
	}
	return err;
}

E
Eric Dumazet 已提交
2550
static unsigned int unix_poll(struct file *file, struct socket *sock, poll_table *wait)
L
Linus Torvalds 已提交
2551 2552 2553 2554
{
	struct sock *sk = sock->sk;
	unsigned int mask;

E
Eric Dumazet 已提交
2555
	sock_poll_wait(file, sk_sleep(sk), wait);
L
Linus Torvalds 已提交
2556 2557 2558 2559 2560 2561 2562
	mask = 0;

	/* exceptional events? */
	if (sk->sk_err)
		mask |= POLLERR;
	if (sk->sk_shutdown == SHUTDOWN_MASK)
		mask |= POLLHUP;
2563
	if (sk->sk_shutdown & RCV_SHUTDOWN)
E
Eric Dumazet 已提交
2564
		mask |= POLLRDHUP | POLLIN | POLLRDNORM;
L
Linus Torvalds 已提交
2565 2566

	/* readable? */
E
Eric Dumazet 已提交
2567
	if (!skb_queue_empty(&sk->sk_receive_queue))
L
Linus Torvalds 已提交
2568 2569 2570
		mask |= POLLIN | POLLRDNORM;

	/* Connection-based need to check for termination and startup */
E
Eric Dumazet 已提交
2571 2572
	if ((sk->sk_type == SOCK_STREAM || sk->sk_type == SOCK_SEQPACKET) &&
	    sk->sk_state == TCP_CLOSE)
L
Linus Torvalds 已提交
2573 2574 2575 2576 2577 2578 2579 2580 2581 2582 2583 2584
		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;
}

2585 2586
static unsigned int unix_dgram_poll(struct file *file, struct socket *sock,
				    poll_table *wait)
2587
{
2588 2589
	struct sock *sk = sock->sk, *other;
	unsigned int mask, writable;
2590

E
Eric Dumazet 已提交
2591
	sock_poll_wait(file, sk_sleep(sk), wait);
2592 2593 2594 2595
	mask = 0;

	/* exceptional events? */
	if (sk->sk_err || !skb_queue_empty(&sk->sk_error_queue))
2596
		mask |= POLLERR |
2597
			(sock_flag(sk, SOCK_SELECT_ERR_QUEUE) ? POLLPRI : 0);
2598

2599
	if (sk->sk_shutdown & RCV_SHUTDOWN)
2600
		mask |= POLLRDHUP | POLLIN | POLLRDNORM;
2601 2602 2603 2604
	if (sk->sk_shutdown == SHUTDOWN_MASK)
		mask |= POLLHUP;

	/* readable? */
2605
	if (!skb_queue_empty(&sk->sk_receive_queue))
2606 2607 2608 2609 2610 2611 2612 2613 2614 2615 2616
		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
2617
	/* No write status requested, avoid expensive OUT tests. */
2618
	if (!(poll_requested_events(wait) & (POLLWRBAND|POLLWRNORM|POLLOUT)))
E
Eric Dumazet 已提交
2619 2620
		return mask;

2621
	writable = unix_writable(sk);
2622 2623 2624 2625 2626 2627 2628 2629 2630 2631
	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);
2632 2633 2634
	}

	if (writable)
2635 2636 2637 2638 2639 2640
		mask |= POLLOUT | POLLWRNORM | POLLWRBAND;
	else
		set_bit(SOCK_ASYNC_NOSPACE, &sk->sk_socket->flags);

	return mask;
}
L
Linus Torvalds 已提交
2641 2642

#ifdef CONFIG_PROC_FS
2643

E
Eric Dumazet 已提交
2644 2645 2646 2647 2648
#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))
2649

E
Eric Dumazet 已提交
2650
static struct sock *unix_from_bucket(struct seq_file *seq, loff_t *pos)
L
Linus Torvalds 已提交
2651
{
E
Eric Dumazet 已提交
2652 2653 2654 2655
	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
2656

E
Eric Dumazet 已提交
2657 2658
	for (sk = sk_head(&unix_socket_table[bucket]); sk; sk = sk_next(sk)) {
		if (sock_net(sk) != seq_file_net(seq))
2659
			continue;
E
Eric Dumazet 已提交
2660 2661 2662 2663 2664 2665 2666 2667 2668 2669 2670 2671 2672 2673 2674 2675 2676 2677 2678
		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
2679
	}
E
Eric Dumazet 已提交
2680 2681 2682 2683 2684 2685 2686 2687 2688 2689 2690

	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
2691 2692 2693 2694
	return NULL;
}

static void *unix_seq_start(struct seq_file *seq, loff_t *pos)
2695
	__acquires(unix_table_lock)
L
Linus Torvalds 已提交
2696
{
2697
	spin_lock(&unix_table_lock);
E
Eric Dumazet 已提交
2698 2699 2700 2701 2702 2703 2704 2705

	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
2706 2707 2708 2709 2710
}

static void *unix_seq_next(struct seq_file *seq, void *v, loff_t *pos)
{
	++*pos;
E
Eric Dumazet 已提交
2711
	return unix_next_socket(seq, v, pos);
L
Linus Torvalds 已提交
2712 2713 2714
}

static void unix_seq_stop(struct seq_file *seq, void *v)
2715
	__releases(unix_table_lock)
L
Linus Torvalds 已提交
2716
{
2717
	spin_unlock(&unix_table_lock);
L
Linus Torvalds 已提交
2718 2719 2720 2721
}

static int unix_seq_show(struct seq_file *seq, void *v)
{
2722

J
Joe Perches 已提交
2723
	if (v == SEQ_START_TOKEN)
L
Linus Torvalds 已提交
2724 2725 2726 2727 2728
		seq_puts(seq, "Num       RefCount Protocol Flags    Type St "
			 "Inode Path\n");
	else {
		struct sock *s = v;
		struct unix_sock *u = unix_sk(s);
2729
		unix_state_lock(s);
L
Linus Torvalds 已提交
2730

D
Dan Rosenberg 已提交
2731
		seq_printf(seq, "%pK: %08X %08X %08X %04X %02X %5lu",
L
Linus Torvalds 已提交
2732 2733 2734 2735 2736 2737 2738 2739 2740 2741 2742 2743 2744 2745 2746 2747 2748 2749 2750 2751 2752 2753 2754 2755 2756
			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++)
				seq_putc(seq, u->addr->name->sun_path[i]);
		}
2757
		unix_state_unlock(s);
L
Linus Torvalds 已提交
2758 2759 2760 2761 2762 2763
		seq_putc(seq, '\n');
	}

	return 0;
}

2764
static const struct seq_operations unix_seq_ops = {
L
Linus Torvalds 已提交
2765 2766 2767 2768 2769 2770 2771 2772
	.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)
{
2773
	return seq_open_net(inode, file, &unix_seq_ops,
E
Eric Dumazet 已提交
2774
			    sizeof(struct seq_net_private));
L
Linus Torvalds 已提交
2775 2776
}

2777
static const struct file_operations unix_seq_fops = {
L
Linus Torvalds 已提交
2778 2779 2780 2781
	.owner		= THIS_MODULE,
	.open		= unix_seq_open,
	.read		= seq_read,
	.llseek		= seq_lseek,
2782
	.release	= seq_release_net,
L
Linus Torvalds 已提交
2783 2784 2785 2786
};

#endif

2787
static const struct net_proto_family unix_family_ops = {
L
Linus Torvalds 已提交
2788 2789 2790 2791 2792
	.family = PF_UNIX,
	.create = unix_create,
	.owner	= THIS_MODULE,
};

2793

2794
static int __net_init unix_net_init(struct net *net)
2795 2796 2797
{
	int error = -ENOMEM;

2798
	net->unx.sysctl_max_dgram_qlen = 10;
2799 2800
	if (unix_sysctl_register(net))
		goto out;
2801

2802
#ifdef CONFIG_PROC_FS
2803
	if (!proc_create("unix", 0, net->proc_net, &unix_seq_fops)) {
2804
		unix_sysctl_unregister(net);
2805
		goto out;
2806
	}
2807 2808 2809
#endif
	error = 0;
out:
2810
	return error;
2811 2812
}

2813
static void __net_exit unix_net_exit(struct net *net)
2814
{
2815
	unix_sysctl_unregister(net);
2816
	remove_proc_entry("unix", net->proc_net);
2817 2818 2819 2820 2821 2822 2823
}

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

L
Linus Torvalds 已提交
2824 2825 2826 2827
static int __init af_unix_init(void)
{
	int rc = -1;

2828
	BUILD_BUG_ON(sizeof(struct unix_skb_parms) > FIELD_SIZEOF(struct sk_buff, cb));
L
Linus Torvalds 已提交
2829 2830

	rc = proto_register(&unix_proto, 1);
2831
	if (rc != 0) {
W
wangweidong 已提交
2832
		pr_crit("%s: Cannot create unix_sock SLAB cache!\n", __func__);
L
Linus Torvalds 已提交
2833 2834 2835 2836
		goto out;
	}

	sock_register(&unix_family_ops);
2837
	register_pernet_subsys(&unix_net_ops);
L
Linus Torvalds 已提交
2838 2839 2840 2841 2842 2843 2844 2845
out:
	return rc;
}

static void __exit af_unix_exit(void)
{
	sock_unregister(PF_UNIX);
	proto_unregister(&unix_proto);
2846
	unregister_pernet_subsys(&unix_net_ops);
L
Linus Torvalds 已提交
2847 2848
}

2849 2850 2851 2852 2853
/* 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
2854 2855 2856 2857
module_exit(af_unix_exit);

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