af_unix.c 56.2 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 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103
 *
 *		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.
 */

#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>
104
#include <net/net_namespace.h>
L
Linus Torvalds 已提交
105
#include <net/sock.h>
106
#include <net/tcp_states.h>
L
Linus Torvalds 已提交
107 108 109 110 111 112 113 114 115 116 117
#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>

E
Eric Dumazet 已提交
118
struct hlist_head unix_socket_table[2 * UNIX_HASH_SIZE];
119 120 121
EXPORT_SYMBOL_GPL(unix_socket_table);
DEFINE_SPINLOCK(unix_table_lock);
EXPORT_SYMBOL_GPL(unix_table_lock);
E
Eric Dumazet 已提交
122
static atomic_long_t unix_nr_socks;
L
Linus Torvalds 已提交
123 124


E
Eric Dumazet 已提交
125 126 127 128 129 130 131 132 133 134 135
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
136

C
Catherine Zhang 已提交
137
#ifdef CONFIG_SECURITY_NETWORK
138
static void unix_get_secdata(struct scm_cookie *scm, struct sk_buff *skb)
C
Catherine Zhang 已提交
139
{
140
	memcpy(UNIXSID(skb), &scm->secid, sizeof(u32));
C
Catherine Zhang 已提交
141 142 143 144
}

static inline void unix_set_secdata(struct scm_cookie *scm, struct sk_buff *skb)
{
145
	scm->secid = *UNIXSID(skb);
C
Catherine Zhang 已提交
146 147
}
#else
148
static inline void unix_get_secdata(struct scm_cookie *scm, struct sk_buff *skb)
C
Catherine Zhang 已提交
149 150 151 152 153 154
{ }

static inline void unix_set_secdata(struct scm_cookie *scm, struct sk_buff *skb)
{ }
#endif /* CONFIG_SECURITY_NETWORK */

L
Linus Torvalds 已提交
155 156
/*
 *  SMP locking strategy:
157
 *    hash table is protected with spinlock unix_table_lock
158
 *    each socket state is protected by separate spin lock.
L
Linus Torvalds 已提交
159 160
 */

161
static inline unsigned int unix_hash_fold(__wsum n)
L
Linus Torvalds 已提交
162
{
163 164
	unsigned int hash = (__force unsigned int)n;

L
Linus Torvalds 已提交
165 166 167 168 169 170 171 172 173 174 175 176 177 178
	hash ^= hash>>16;
	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
179
	return unix_peer(osk) == NULL || unix_our_peer(sk, osk);
L
Linus Torvalds 已提交
180 181
}

182 183 184 185 186
static inline int unix_recvq_full(struct sock const *sk)
{
	return skb_queue_len(&sk->sk_receive_queue) > sk->sk_max_ack_backlog;
}

187
struct sock *unix_peer_get(struct sock *s)
L
Linus Torvalds 已提交
188 189 190
{
	struct sock *peer;

191
	unix_state_lock(s);
L
Linus Torvalds 已提交
192 193 194
	peer = unix_peer(s);
	if (peer)
		sock_hold(peer);
195
	unix_state_unlock(s);
L
Linus Torvalds 已提交
196 197
	return peer;
}
198
EXPORT_SYMBOL_GPL(unix_peer_get);
L
Linus Torvalds 已提交
199 200 201 202 203 204 205 206 207 208 209 210 211

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.
 */
212

213
static int unix_mkname(struct sockaddr_un *sunaddr, int len, unsigned int *hashp)
L
Linus Torvalds 已提交
214 215 216 217 218 219 220 221 222
{
	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
223
		 * sun_path[108] doesn't as such exist.  However in kernel space
L
Linus Torvalds 已提交
224 225 226
		 * we are guaranteed that it is a valid memory location in our
		 * kernel address buffer.
		 */
227
		((char *)sunaddr)[len] = 0;
L
Linus Torvalds 已提交
228 229 230 231
		len = strlen(sunaddr->sun_path)+1+sizeof(short);
		return len;
	}

232
	*hashp = unix_hash_fold(csum_partial(sunaddr, len, 0));
L
Linus Torvalds 已提交
233 234 235 236 237 238 239 240 241 242
	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)
{
243
	WARN_ON(!sk_unhashed(sk));
L
Linus Torvalds 已提交
244 245 246 247 248
	sk_add_node(sk, list);
}

static inline void unix_remove_socket(struct sock *sk)
{
249
	spin_lock(&unix_table_lock);
L
Linus Torvalds 已提交
250
	__unix_remove_socket(sk);
251
	spin_unlock(&unix_table_lock);
L
Linus Torvalds 已提交
252 253 254 255
}

static inline void unix_insert_socket(struct hlist_head *list, struct sock *sk)
{
256
	spin_lock(&unix_table_lock);
L
Linus Torvalds 已提交
257
	__unix_insert_socket(list, sk);
258
	spin_unlock(&unix_table_lock);
L
Linus Torvalds 已提交
259 260
}

261 262
static struct sock *__unix_find_socket_byname(struct net *net,
					      struct sockaddr_un *sunname,
263
					      int len, int type, unsigned int hash)
L
Linus Torvalds 已提交
264 265 266 267 268 269 270
{
	struct sock *s;
	struct hlist_node *node;

	sk_for_each(s, node, &unix_socket_table[hash ^ type]) {
		struct unix_sock *u = unix_sk(s);

271
		if (!net_eq(sock_net(s), net))
272 273
			continue;

L
Linus Torvalds 已提交
274 275 276 277 278 279 280 281 282
		if (u->addr->len == len &&
		    !memcmp(u->addr->name, sunname, len))
			goto found;
	}
	s = NULL;
found:
	return s;
}

283 284
static inline struct sock *unix_find_socket_byname(struct net *net,
						   struct sockaddr_un *sunname,
L
Linus Torvalds 已提交
285
						   int len, int type,
286
						   unsigned int hash)
L
Linus Torvalds 已提交
287 288 289
{
	struct sock *s;

290
	spin_lock(&unix_table_lock);
291
	s = __unix_find_socket_byname(net, sunname, len, type, hash);
L
Linus Torvalds 已提交
292 293
	if (s)
		sock_hold(s);
294
	spin_unlock(&unix_table_lock);
L
Linus Torvalds 已提交
295 296 297
	return s;
}

298
static struct sock *unix_find_socket_byinode(struct inode *i)
L
Linus Torvalds 已提交
299 300 301 302
{
	struct sock *s;
	struct hlist_node *node;

303
	spin_lock(&unix_table_lock);
L
Linus Torvalds 已提交
304 305
	sk_for_each(s, node,
		    &unix_socket_table[i->i_ino & (UNIX_HASH_SIZE - 1)]) {
A
Al Viro 已提交
306
		struct dentry *dentry = unix_sk(s)->path.dentry;
L
Linus Torvalds 已提交
307

E
Eric Dumazet 已提交
308
		if (dentry && dentry->d_inode == i) {
L
Linus Torvalds 已提交
309 310 311 312 313 314
			sock_hold(s);
			goto found;
		}
	}
	s = NULL;
found:
315
	spin_unlock(&unix_table_lock);
L
Linus Torvalds 已提交
316 317 318 319 320 321 322 323 324 325
	return s;
}

static inline int unix_writable(struct sock *sk)
{
	return (atomic_read(&sk->sk_wmem_alloc) << 2) <= sk->sk_sndbuf;
}

static void unix_write_space(struct sock *sk)
{
326 327 328
	struct socket_wq *wq;

	rcu_read_lock();
L
Linus Torvalds 已提交
329
	if (unix_writable(sk)) {
330 331
		wq = rcu_dereference(sk->sk_wq);
		if (wq_has_sleeper(wq))
E
Eric Dumazet 已提交
332 333
			wake_up_interruptible_sync_poll(&wq->wait,
				POLLOUT | POLLWRNORM | POLLWRBAND);
334
		sk_wake_async(sk, SOCK_WAKE_SPACE, POLL_OUT);
L
Linus Torvalds 已提交
335
	}
336
	rcu_read_unlock();
L
Linus Torvalds 已提交
337 338 339 340 341 342 343 344
}

/* 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)
{
345
	if (!skb_queue_empty(&sk->sk_receive_queue)) {
L
Linus Torvalds 已提交
346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365
		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);

366 367 368
	WARN_ON(atomic_read(&sk->sk_wmem_alloc));
	WARN_ON(!sk_unhashed(sk));
	WARN_ON(sk->sk_socket);
L
Linus Torvalds 已提交
369
	if (!sock_flag(sk, SOCK_DEAD)) {
370
		printk(KERN_INFO "Attempt to release alive unix socket: %p\n", sk);
L
Linus Torvalds 已提交
371 372 373 374 375 376
		return;
	}

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

E
Eric Dumazet 已提交
377
	atomic_long_dec(&unix_nr_socks);
378
	local_bh_disable();
379
	sock_prot_inuse_add(sock_net(sk), sk->sk_prot, -1);
380
	local_bh_enable();
L
Linus Torvalds 已提交
381
#ifdef UNIX_REFCNT_DEBUG
E
Eric Dumazet 已提交
382 383
	printk(KERN_DEBUG "UNIX %p is destroyed, %ld are still alive.\n", sk,
		atomic_long_read(&unix_nr_socks));
L
Linus Torvalds 已提交
384 385 386
#endif
}

E
Eric Dumazet 已提交
387
static int unix_release_sock(struct sock *sk, int embrion)
L
Linus Torvalds 已提交
388 389
{
	struct unix_sock *u = unix_sk(sk);
A
Al Viro 已提交
390
	struct path path;
L
Linus Torvalds 已提交
391 392 393 394 395 396 397
	struct sock *skpair;
	struct sk_buff *skb;
	int state;

	unix_remove_socket(sk);

	/* Clear state */
398
	unix_state_lock(sk);
L
Linus Torvalds 已提交
399 400
	sock_orphan(sk);
	sk->sk_shutdown = SHUTDOWN_MASK;
A
Al Viro 已提交
401 402 403
	path	     = u->path;
	u->path.dentry = NULL;
	u->path.mnt = NULL;
L
Linus Torvalds 已提交
404 405
	state = sk->sk_state;
	sk->sk_state = TCP_CLOSE;
406
	unix_state_unlock(sk);
L
Linus Torvalds 已提交
407 408 409

	wake_up_interruptible_all(&u->peer_wait);

410
	skpair = unix_peer(sk);
L
Linus Torvalds 已提交
411

412
	if (skpair != NULL) {
L
Linus Torvalds 已提交
413
		if (sk->sk_type == SOCK_STREAM || sk->sk_type == SOCK_SEQPACKET) {
414
			unix_state_lock(skpair);
L
Linus Torvalds 已提交
415 416 417 418
			/* No more writes */
			skpair->sk_shutdown = SHUTDOWN_MASK;
			if (!skb_queue_empty(&sk->sk_receive_queue) || embrion)
				skpair->sk_err = ECONNRESET;
419
			unix_state_unlock(skpair);
L
Linus Torvalds 已提交
420
			skpair->sk_state_change(skpair);
421
			sk_wake_async(skpair, SOCK_WAKE_WAITD, POLL_HUP);
L
Linus Torvalds 已提交
422 423 424 425 426 427 428 429
		}
		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) {
430
		if (state == TCP_LISTEN)
L
Linus Torvalds 已提交
431 432 433 434 435
			unix_release_sock(skb->sk, 1);
		/* passed fds are erased in the kfree_skb hook	      */
		kfree_skb(skb);
	}

A
Al Viro 已提交
436 437
	if (path.dentry)
		path_put(&path);
L
Linus Torvalds 已提交
438 439 440 441 442 443

	sock_put(sk);

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

	/*
A
Alan Cox 已提交
444
	 * Fixme: BSD difference: In BSD all sockets connected to us get
L
Linus Torvalds 已提交
445 446 447 448 449 450 451 452 453
	 *	  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)
	 */

454
	if (unix_tot_inflight)
455
		unix_gc();		/* Garbage collect fds */
L
Linus Torvalds 已提交
456 457 458 459

	return 0;
}

460 461 462 463 464 465 466 467 468 469 470 471 472 473 474 475 476 477
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
478 479 480 481 482
static int unix_listen(struct socket *sock, int backlog)
{
	int err;
	struct sock *sk = sock->sk;
	struct unix_sock *u = unix_sk(sk);
483
	struct pid *old_pid = NULL;
L
Linus Torvalds 已提交
484 485

	err = -EOPNOTSUPP;
E
Eric Dumazet 已提交
486 487
	if (sock->type != SOCK_STREAM && sock->type != SOCK_SEQPACKET)
		goto out;	/* Only stream/seqpacket sockets accept */
L
Linus Torvalds 已提交
488 489
	err = -EINVAL;
	if (!u->addr)
E
Eric Dumazet 已提交
490
		goto out;	/* No listens on an unbound socket */
491
	unix_state_lock(sk);
L
Linus Torvalds 已提交
492 493 494 495 496 497 498
	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 */
499
	init_peercred(sk);
L
Linus Torvalds 已提交
500 501 502
	err = 0;

out_unlock:
503
	unix_state_unlock(sk);
504
	put_pid(old_pid);
L
Linus Torvalds 已提交
505 506 507 508 509 510 511 512 513 514 515 516
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 *);
517 518
static unsigned int unix_dgram_poll(struct file *, struct socket *,
				    poll_table *);
L
Linus Torvalds 已提交
519 520 521 522 523 524 525 526 527 528 529 530 531 532
static int unix_ioctl(struct socket *, unsigned int, unsigned long);
static int unix_shutdown(struct socket *, int);
static int unix_stream_sendmsg(struct kiocb *, struct socket *,
			       struct msghdr *, size_t);
static int unix_stream_recvmsg(struct kiocb *, struct socket *,
			       struct msghdr *, size_t, int);
static int unix_dgram_sendmsg(struct kiocb *, struct socket *,
			      struct msghdr *, size_t);
static int unix_dgram_recvmsg(struct kiocb *, struct socket *,
			      struct msghdr *, size_t, int);
static int unix_dgram_connect(struct socket *, struct sockaddr *,
			      int, int);
static int unix_seqpacket_sendmsg(struct kiocb *, struct socket *,
				  struct msghdr *, size_t);
533 534
static int unix_seqpacket_recvmsg(struct kiocb *, struct socket *,
				  struct msghdr *, size_t, int);
L
Linus Torvalds 已提交
535

536 537 538 539 540 541 542 543 544 545
static void unix_set_peek_off(struct sock *sk, int val)
{
	struct unix_sock *u = unix_sk(sk);

	mutex_lock(&u->readlock);
	sk->sk_peek_off = val;
	mutex_unlock(&u->readlock);
}


546
static const struct proto_ops unix_stream_ops = {
L
Linus Torvalds 已提交
547 548 549 550 551 552 553 554 555 556 557 558 559 560 561 562 563 564
	.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,
	.sendpage =	sock_no_sendpage,
565
	.set_peek_off =	unix_set_peek_off,
L
Linus Torvalds 已提交
566 567
};

568
static const struct proto_ops unix_dgram_ops = {
L
Linus Torvalds 已提交
569 570 571 572 573 574 575 576
	.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,
577
	.poll =		unix_dgram_poll,
L
Linus Torvalds 已提交
578 579 580 581 582 583 584 585 586
	.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,
587
	.set_peek_off =	unix_set_peek_off,
L
Linus Torvalds 已提交
588 589
};

590
static const struct proto_ops unix_seqpacket_ops = {
L
Linus Torvalds 已提交
591 592 593 594 595 596 597 598
	.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,
599
	.poll =		unix_dgram_poll,
L
Linus Torvalds 已提交
600 601 602 603 604 605
	.ioctl =	unix_ioctl,
	.listen =	unix_listen,
	.shutdown =	unix_shutdown,
	.setsockopt =	sock_no_setsockopt,
	.getsockopt =	sock_no_getsockopt,
	.sendmsg =	unix_seqpacket_sendmsg,
606
	.recvmsg =	unix_seqpacket_recvmsg,
L
Linus Torvalds 已提交
607 608
	.mmap =		sock_no_mmap,
	.sendpage =	sock_no_sendpage,
609
	.set_peek_off =	unix_set_peek_off,
L
Linus Torvalds 已提交
610 611 612
};

static struct proto unix_proto = {
613 614 615
	.name			= "UNIX",
	.owner			= THIS_MODULE,
	.obj_size		= sizeof(struct unix_sock),
L
Linus Torvalds 已提交
616 617
};

618 619 620 621 622 623 624 625
/*
 * 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;

E
Eric Dumazet 已提交
626
static struct sock *unix_create1(struct net *net, struct socket *sock)
L
Linus Torvalds 已提交
627 628 629 630
{
	struct sock *sk = NULL;
	struct unix_sock *u;

E
Eric Dumazet 已提交
631 632
	atomic_long_inc(&unix_nr_socks);
	if (atomic_long_read(&unix_nr_socks) > 2 * get_max_files())
L
Linus Torvalds 已提交
633 634
		goto out;

635
	sk = sk_alloc(net, PF_UNIX, GFP_KERNEL, &unix_proto);
L
Linus Torvalds 已提交
636 637 638
	if (!sk)
		goto out;

E
Eric Dumazet 已提交
639
	sock_init_data(sock, sk);
640 641
	lockdep_set_class(&sk->sk_receive_queue.lock,
				&af_unix_sk_receive_queue_lock_key);
L
Linus Torvalds 已提交
642 643

	sk->sk_write_space	= unix_write_space;
644
	sk->sk_max_ack_backlog	= net->unx.sysctl_max_dgram_qlen;
L
Linus Torvalds 已提交
645 646
	sk->sk_destruct		= unix_sock_destructor;
	u	  = unix_sk(sk);
A
Al Viro 已提交
647 648
	u->path.dentry = NULL;
	u->path.mnt = NULL;
649
	spin_lock_init(&u->lock);
A
Al Viro 已提交
650
	atomic_long_set(&u->inflight, 0);
651
	INIT_LIST_HEAD(&u->link);
I
Ingo Molnar 已提交
652
	mutex_init(&u->readlock); /* single task reading lock */
L
Linus Torvalds 已提交
653
	init_waitqueue_head(&u->peer_wait);
E
Eric Dumazet 已提交
654
	unix_insert_socket(unix_sockets_unbound(sk), sk);
L
Linus Torvalds 已提交
655
out:
656
	if (sk == NULL)
E
Eric Dumazet 已提交
657
		atomic_long_dec(&unix_nr_socks);
658 659
	else {
		local_bh_disable();
660
		sock_prot_inuse_add(sock_net(sk), sk->sk_prot, 1);
661 662
		local_bh_enable();
	}
L
Linus Torvalds 已提交
663 664 665
	return sk;
}

666 667
static int unix_create(struct net *net, struct socket *sock, int protocol,
		       int kern)
L
Linus Torvalds 已提交
668 669 670 671 672 673 674 675 676 677 678 679 680 681 682
{
	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:
683
		sock->type = SOCK_DGRAM;
L
Linus Torvalds 已提交
684 685 686 687 688 689 690 691 692 693
	case SOCK_DGRAM:
		sock->ops = &unix_dgram_ops;
		break;
	case SOCK_SEQPACKET:
		sock->ops = &unix_seqpacket_ops;
		break;
	default:
		return -ESOCKTNOSUPPORT;
	}

694
	return unix_create1(net, sock) ? 0 : -ENOMEM;
L
Linus Torvalds 已提交
695 696 697 698 699 700 701 702 703 704 705
}

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

	if (!sk)
		return 0;

	sock->sk = NULL;

E
Eric Dumazet 已提交
706
	return unix_release_sock(sk, 0);
L
Linus Torvalds 已提交
707 708 709 710 711
}

static int unix_autobind(struct socket *sock)
{
	struct sock *sk = sock->sk;
712
	struct net *net = sock_net(sk);
L
Linus Torvalds 已提交
713 714
	struct unix_sock *u = unix_sk(sk);
	static u32 ordernum = 1;
E
Eric Dumazet 已提交
715
	struct unix_address *addr;
L
Linus Torvalds 已提交
716
	int err;
717
	unsigned int retries = 0;
L
Linus Torvalds 已提交
718

I
Ingo Molnar 已提交
719
	mutex_lock(&u->readlock);
L
Linus Torvalds 已提交
720 721 722 723 724 725

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

	err = -ENOMEM;
726
	addr = kzalloc(sizeof(*addr) + sizeof(short) + 16, GFP_KERNEL);
L
Linus Torvalds 已提交
727 728 729 730 731 732 733 734
	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);
735
	addr->hash = unix_hash_fold(csum_partial(addr->name, addr->len, 0));
L
Linus Torvalds 已提交
736

737
	spin_lock(&unix_table_lock);
L
Linus Torvalds 已提交
738 739
	ordernum = (ordernum+1)&0xFFFFF;

740
	if (__unix_find_socket_byname(net, addr->name, addr->len, sock->type,
L
Linus Torvalds 已提交
741
				      addr->hash)) {
742
		spin_unlock(&unix_table_lock);
743 744 745 746 747 748 749 750 751 752 753
		/*
		 * __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
754 755 756 757 758 759 760
		goto retry;
	}
	addr->hash ^= sk->sk_type;

	__unix_remove_socket(sk);
	u->addr = addr;
	__unix_insert_socket(&unix_socket_table[addr->hash], sk);
761
	spin_unlock(&unix_table_lock);
L
Linus Torvalds 已提交
762 763
	err = 0;

I
Ingo Molnar 已提交
764
out:	mutex_unlock(&u->readlock);
L
Linus Torvalds 已提交
765 766 767
	return err;
}

768 769
static struct sock *unix_find_other(struct net *net,
				    struct sockaddr_un *sunname, int len,
770
				    int type, unsigned int hash, int *error)
L
Linus Torvalds 已提交
771 772
{
	struct sock *u;
773
	struct path path;
L
Linus Torvalds 已提交
774
	int err = 0;
775

L
Linus Torvalds 已提交
776
	if (sunname->sun_path[0]) {
777 778
		struct inode *inode;
		err = kern_path(sunname->sun_path, LOOKUP_FOLLOW, &path);
L
Linus Torvalds 已提交
779 780
		if (err)
			goto fail;
781 782
		inode = path.dentry->d_inode;
		err = inode_permission(inode, MAY_WRITE);
L
Linus Torvalds 已提交
783 784 785 786
		if (err)
			goto put_fail;

		err = -ECONNREFUSED;
787
		if (!S_ISSOCK(inode->i_mode))
L
Linus Torvalds 已提交
788
			goto put_fail;
789
		u = unix_find_socket_byinode(inode);
L
Linus Torvalds 已提交
790 791 792 793
		if (!u)
			goto put_fail;

		if (u->sk_type == type)
A
Al Viro 已提交
794
			touch_atime(&path);
L
Linus Torvalds 已提交
795

796
		path_put(&path);
L
Linus Torvalds 已提交
797

798
		err = -EPROTOTYPE;
L
Linus Torvalds 已提交
799 800 801 802 803 804
		if (u->sk_type != type) {
			sock_put(u);
			goto fail;
		}
	} else {
		err = -ECONNREFUSED;
805
		u = unix_find_socket_byname(net, sunname, len, type, hash);
L
Linus Torvalds 已提交
806 807
		if (u) {
			struct dentry *dentry;
A
Al Viro 已提交
808
			dentry = unix_sk(u)->path.dentry;
L
Linus Torvalds 已提交
809
			if (dentry)
A
Al Viro 已提交
810
				touch_atime(&unix_sk(u)->path);
L
Linus Torvalds 已提交
811 812 813 814 815 816
		} else
			goto fail;
	}
	return u;

put_fail:
817
	path_put(&path);
L
Linus Torvalds 已提交
818
fail:
819
	*error = err;
L
Linus Torvalds 已提交
820 821 822
	return NULL;
}

A
Al Viro 已提交
823 824 825 826 827 828 829 830 831 832 833 834 835 836 837 838 839 840 841 842 843 844 845 846 847 848 849 850
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) {
		err = vfs_mknod(path.dentry->d_inode, dentry, mode, 0);
		if (!err) {
			res->mnt = mntget(path.mnt);
			res->dentry = dget(dentry);
		}
	}
	done_path_create(&path, dentry);
	return err;
}
L
Linus Torvalds 已提交
851 852 853 854

static int unix_bind(struct socket *sock, struct sockaddr *uaddr, int addr_len)
{
	struct sock *sk = sock->sk;
855
	struct net *net = sock_net(sk);
L
Linus Torvalds 已提交
856
	struct unix_sock *u = unix_sk(sk);
857
	struct sockaddr_un *sunaddr = (struct sockaddr_un *)uaddr;
858
	char *sun_path = sunaddr->sun_path;
L
Linus Torvalds 已提交
859
	int err;
860
	unsigned int hash;
L
Linus Torvalds 已提交
861 862 863 864 865 866 867
	struct unix_address *addr;
	struct hlist_head *list;

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

868
	if (addr_len == sizeof(short)) {
L
Linus Torvalds 已提交
869 870 871 872 873 874 875 876 877
		err = unix_autobind(sock);
		goto out;
	}

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

I
Ingo Molnar 已提交
878
	mutex_lock(&u->readlock);
L
Linus Torvalds 已提交
879 880 881 882 883 884 885 886 887 888 889 890 891 892 893

	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);

894
	if (sun_path[0]) {
A
Al Viro 已提交
895 896
		struct path path;
		umode_t mode = S_IFSOCK |
A
Al Viro 已提交
897
		       (SOCK_INODE(sock)->i_mode & ~current_umask());
A
Al Viro 已提交
898 899 900 901 902 903 904
		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
905
		addr->hash = UNIX_HASH_SIZE;
A
Al Viro 已提交
906 907 908 909 910 911
		hash = path.dentry->d_inode->i_ino & (UNIX_HASH_SIZE-1);
		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
912
		err = -EADDRINUSE;
913
		if (__unix_find_socket_byname(net, sunaddr, addr_len,
L
Linus Torvalds 已提交
914 915 916 917 918 919 920 921 922 923 924 925 926 927
					      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:
928
	spin_unlock(&unix_table_lock);
L
Linus Torvalds 已提交
929
out_up:
I
Ingo Molnar 已提交
930
	mutex_unlock(&u->readlock);
L
Linus Torvalds 已提交
931 932 933 934
out:
	return err;
}

935 936 937 938 939 940 941 942 943 944 945 946 947 948 949 950 951 952 953 954 955 956 957 958 959
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
960 961 962 963
static int unix_dgram_connect(struct socket *sock, struct sockaddr *addr,
			      int alen, int flags)
{
	struct sock *sk = sock->sk;
964
	struct net *net = sock_net(sk);
965
	struct sockaddr_un *sunaddr = (struct sockaddr_un *)addr;
L
Linus Torvalds 已提交
966
	struct sock *other;
967
	unsigned int hash;
L
Linus Torvalds 已提交
968 969 970 971 972 973 974 975 976 977 978 979
	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;

980
restart:
981
		other = unix_find_other(net, sunaddr, alen, sock->type, hash, &err);
L
Linus Torvalds 已提交
982 983 984
		if (!other)
			goto out;

985 986 987 988 989 990 991 992
		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
993 994 995 996 997 998 999 1000 1001 1002 1003 1004 1005 1006

		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;
1007
		unix_state_double_lock(sk, other);
L
Linus Torvalds 已提交
1008 1009 1010 1011 1012 1013 1014
	}

	/*
	 * If it was connected, reconnect.
	 */
	if (unix_peer(sk)) {
		struct sock *old_peer = unix_peer(sk);
1015
		unix_peer(sk) = other;
1016
		unix_state_double_unlock(sk, other);
L
Linus Torvalds 已提交
1017 1018 1019 1020 1021

		if (other != old_peer)
			unix_dgram_disconnected(sk, old_peer);
		sock_put(old_peer);
	} else {
1022
		unix_peer(sk) = other;
1023
		unix_state_double_unlock(sk, other);
L
Linus Torvalds 已提交
1024
	}
1025
	return 0;
L
Linus Torvalds 已提交
1026 1027

out_unlock:
1028
	unix_state_double_unlock(sk, other);
L
Linus Torvalds 已提交
1029 1030 1031 1032 1033 1034 1035 1036 1037 1038 1039 1040 1041 1042 1043
	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) &&
1044
		unix_recvq_full(other);
L
Linus Torvalds 已提交
1045

1046
	unix_state_unlock(other);
L
Linus Torvalds 已提交
1047 1048 1049 1050 1051 1052 1053 1054 1055 1056 1057

	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)
{
1058
	struct sockaddr_un *sunaddr = (struct sockaddr_un *)uaddr;
L
Linus Torvalds 已提交
1059
	struct sock *sk = sock->sk;
1060
	struct net *net = sock_net(sk);
L
Linus Torvalds 已提交
1061 1062 1063 1064
	struct unix_sock *u = unix_sk(sk), *newu, *otheru;
	struct sock *newsk = NULL;
	struct sock *other = NULL;
	struct sk_buff *skb = NULL;
1065
	unsigned int hash;
L
Linus Torvalds 已提交
1066 1067 1068 1069 1070 1071 1072 1073 1074
	int st;
	int err;
	long timeo;

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

1075 1076
	if (test_bit(SOCK_PASSCRED, &sock->flags) && !u->addr &&
	    (err = unix_autobind(sock)) != 0)
L
Linus Torvalds 已提交
1077 1078 1079 1080 1081 1082 1083 1084 1085 1086 1087 1088
		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 */
1089
	newsk = unix_create1(sock_net(sk), NULL);
L
Linus Torvalds 已提交
1090 1091 1092 1093 1094 1095 1096 1097 1098 1099
	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. */
1100
	other = unix_find_other(net, sunaddr, addr_len, sk->sk_type, hash, &err);
L
Linus Torvalds 已提交
1101 1102 1103 1104
	if (!other)
		goto out;

	/* Latch state of peer */
1105
	unix_state_lock(other);
L
Linus Torvalds 已提交
1106 1107 1108

	/* Apparently VFS overslept socket death. Retry. */
	if (sock_flag(other, SOCK_DEAD)) {
1109
		unix_state_unlock(other);
L
Linus Torvalds 已提交
1110 1111 1112 1113 1114 1115 1116
		sock_put(other);
		goto restart;
	}

	err = -ECONNREFUSED;
	if (other->sk_state != TCP_LISTEN)
		goto out_unlock;
1117 1118
	if (other->sk_shutdown & RCV_SHUTDOWN)
		goto out_unlock;
L
Linus Torvalds 已提交
1119

1120
	if (unix_recvq_full(other)) {
L
Linus Torvalds 已提交
1121 1122 1123 1124 1125 1126 1127 1128 1129 1130 1131
		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;
1132
	}
L
Linus Torvalds 已提交
1133 1134 1135

	/* Latch our state.

D
Daniel Baluta 已提交
1136
	   It is tricky place. We need to grab our state lock and cannot
L
Linus Torvalds 已提交
1137 1138 1139 1140 1141 1142 1143 1144 1145 1146 1147 1148 1149 1150 1151 1152 1153 1154 1155 1156 1157 1158 1159
	   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;
	}

1160
	unix_state_lock_nested(sk);
L
Linus Torvalds 已提交
1161 1162

	if (sk->sk_state != st) {
1163 1164
		unix_state_unlock(sk);
		unix_state_unlock(other);
L
Linus Torvalds 已提交
1165 1166 1167 1168
		sock_put(other);
		goto restart;
	}

1169
	err = security_unix_stream_connect(sk, other, newsk);
L
Linus Torvalds 已提交
1170
	if (err) {
1171
		unix_state_unlock(sk);
L
Linus Torvalds 已提交
1172 1173 1174 1175 1176 1177 1178 1179 1180
		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;
1181
	init_peercred(newsk);
L
Linus Torvalds 已提交
1182
	newu = unix_sk(newsk);
1183
	RCU_INIT_POINTER(newsk->sk_wq, &newu->peer_wq);
L
Linus Torvalds 已提交
1184 1185 1186 1187 1188 1189 1190
	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
1191 1192 1193
	if (otheru->path.dentry) {
		path_get(&otheru->path);
		newu->path = otheru->path;
L
Linus Torvalds 已提交
1194 1195 1196
	}

	/* Set credentials */
1197
	copy_peercred(sk, other);
L
Linus Torvalds 已提交
1198 1199 1200

	sock->state	= SS_CONNECTED;
	sk->sk_state	= TCP_ESTABLISHED;
1201 1202 1203 1204
	sock_hold(newsk);

	smp_mb__after_atomic_inc();	/* sock_hold() does an atomic_inc() */
	unix_peer(sk)	= newsk;
L
Linus Torvalds 已提交
1205

1206
	unix_state_unlock(sk);
L
Linus Torvalds 已提交
1207 1208 1209 1210 1211

	/* 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);
1212
	unix_state_unlock(other);
L
Linus Torvalds 已提交
1213 1214 1215 1216 1217 1218
	other->sk_data_ready(other, 0);
	sock_put(other);
	return 0;

out_unlock:
	if (other)
1219
		unix_state_unlock(other);
L
Linus Torvalds 已提交
1220 1221

out:
1222
	kfree_skb(skb);
L
Linus Torvalds 已提交
1223 1224 1225 1226 1227 1228 1229 1230 1231
	if (newsk)
		unix_release_sock(newsk, 0);
	if (other)
		sock_put(other);
	return err;
}

static int unix_socketpair(struct socket *socka, struct socket *sockb)
{
1232
	struct sock *ska = socka->sk, *skb = sockb->sk;
L
Linus Torvalds 已提交
1233 1234 1235 1236

	/* Join our sockets back to back */
	sock_hold(ska);
	sock_hold(skb);
1237 1238
	unix_peer(ska) = skb;
	unix_peer(skb) = ska;
1239 1240
	init_peercred(ska);
	init_peercred(skb);
L
Linus Torvalds 已提交
1241 1242 1243 1244 1245 1246 1247 1248 1249 1250 1251 1252 1253 1254 1255 1256 1257 1258

	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;
}

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
1259
	if (sock->type != SOCK_STREAM && sock->type != SOCK_SEQPACKET)
L
Linus Torvalds 已提交
1260 1261 1262 1263 1264 1265 1266 1267 1268 1269 1270 1271 1272 1273 1274 1275 1276 1277 1278 1279 1280 1281 1282
		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 */
1283
	unix_state_lock(tsk);
L
Linus Torvalds 已提交
1284 1285
	newsock->state = SS_CONNECTED;
	sock_graft(tsk, newsock);
1286
	unix_state_unlock(tsk);
L
Linus Torvalds 已提交
1287 1288 1289 1290 1291 1292 1293 1294 1295 1296 1297
	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;
1298
	DECLARE_SOCKADDR(struct sockaddr_un *, sunaddr, uaddr);
L
Linus Torvalds 已提交
1299 1300 1301 1302 1303 1304 1305 1306 1307 1308 1309 1310 1311 1312
	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);
1313
	unix_state_lock(sk);
L
Linus Torvalds 已提交
1314 1315 1316 1317 1318 1319 1320 1321 1322 1323
	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);
	}
1324
	unix_state_unlock(sk);
L
Linus Torvalds 已提交
1325 1326 1327 1328 1329 1330 1331 1332 1333 1334 1335 1336
	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
1337
	for (i = scm->fp->count-1; i >= 0; i--)
L
Linus Torvalds 已提交
1338 1339 1340
		unix_notinflight(scm->fp->fp[i]);
}

1341
static void unix_destruct_scm(struct sk_buff *skb)
L
Linus Torvalds 已提交
1342 1343 1344
{
	struct scm_cookie scm;
	memset(&scm, 0, sizeof(scm));
1345 1346 1347 1348
	scm.pid  = UNIXCB(skb).pid;
	scm.cred = UNIXCB(skb).cred;
	if (UNIXCB(skb).fp)
		unix_detach_fds(&scm, skb);
L
Linus Torvalds 已提交
1349 1350 1351 1352 1353 1354 1355

	/* 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
1356 1357
#define MAX_RECURSION_LEVEL 4

1358
static int unix_attach_fds(struct scm_cookie *scm, struct sk_buff *skb)
L
Linus Torvalds 已提交
1359 1360
{
	int i;
E
Eric Dumazet 已提交
1361 1362 1363 1364 1365 1366 1367 1368 1369 1370 1371 1372 1373 1374
	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;
1375 1376 1377 1378 1379 1380 1381 1382 1383 1384

	/*
	 * 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
1385 1386 1387 1388 1389
	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
1390 1391
}

1392
static int unix_scm_to_skb(struct scm_cookie *scm, struct sk_buff *skb, bool send_fds)
1393 1394
{
	int err = 0;
1395

1396
	UNIXCB(skb).pid  = get_pid(scm->pid);
1397 1398
	if (scm->cred)
		UNIXCB(skb).cred = get_cred(scm->cred);
1399 1400 1401 1402 1403 1404 1405 1406
	UNIXCB(skb).fp = NULL;
	if (scm->fp && send_fds)
		err = unix_attach_fds(scm, skb);

	skb->destructor = unix_destruct_scm;
	return err;
}

1407 1408 1409 1410 1411 1412 1413 1414 1415 1416 1417 1418 1419 1420 1421 1422 1423 1424
/*
 * 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)
{
	if (UNIXCB(skb).cred)
		return;
	if (test_bit(SOCK_PASSCRED, &sock->flags) ||
	    !other->sk_socket ||
	    test_bit(SOCK_PASSCRED, &other->sk_socket->flags)) {
		UNIXCB(skb).pid  = get_pid(task_tgid(current));
		UNIXCB(skb).cred = get_current_cred();
	}
}

L
Linus Torvalds 已提交
1425 1426 1427 1428 1429 1430 1431 1432 1433
/*
 *	Send AF_UNIX data.
 */

static int unix_dgram_sendmsg(struct kiocb *kiocb, struct socket *sock,
			      struct msghdr *msg, size_t len)
{
	struct sock_iocb *siocb = kiocb_to_siocb(kiocb);
	struct sock *sk = sock->sk;
1434
	struct net *net = sock_net(sk);
L
Linus Torvalds 已提交
1435
	struct unix_sock *u = unix_sk(sk);
1436
	struct sockaddr_un *sunaddr = msg->msg_name;
L
Linus Torvalds 已提交
1437 1438 1439
	struct sock *other = NULL;
	int namelen = 0; /* fake GCC */
	int err;
1440
	unsigned int hash;
1441
	struct sk_buff *skb;
L
Linus Torvalds 已提交
1442 1443
	long timeo;
	struct scm_cookie tmp_scm;
E
Eric Dumazet 已提交
1444
	int max_level;
1445
	int data_len = 0;
L
Linus Torvalds 已提交
1446 1447 1448

	if (NULL == siocb->scm)
		siocb->scm = &tmp_scm;
1449
	wait_for_unix_gc();
1450
	err = scm_send(sock, msg, siocb->scm, false);
L
Linus Torvalds 已提交
1451 1452 1453 1454 1455 1456 1457 1458 1459 1460 1461 1462 1463 1464 1465 1466 1467 1468 1469 1470
	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;
	}

1471 1472
	if (test_bit(SOCK_PASSCRED, &sock->flags) && !u->addr
	    && (err = unix_autobind(sock)) != 0)
L
Linus Torvalds 已提交
1473 1474 1475 1476 1477 1478
		goto out;

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

1479 1480 1481 1482 1483 1484 1485
	if (len > SKB_MAX_ALLOC)
		data_len = min_t(size_t,
				 len - SKB_MAX_ALLOC,
				 MAX_SKB_FRAGS * PAGE_SIZE);

	skb = sock_alloc_send_pskb(sk, len - data_len, data_len,
				   msg->msg_flags & MSG_DONTWAIT, &err);
1486
	if (skb == NULL)
L
Linus Torvalds 已提交
1487 1488
		goto out;

1489
	err = unix_scm_to_skb(siocb->scm, skb, true);
E
Eric Dumazet 已提交
1490
	if (err < 0)
1491
		goto out_free;
E
Eric Dumazet 已提交
1492
	max_level = err + 1;
1493
	unix_get_secdata(siocb->scm, skb);
C
Catherine Zhang 已提交
1494

1495 1496 1497 1498
	skb_put(skb, len - data_len);
	skb->data_len = data_len;
	skb->len = len;
	err = skb_copy_datagram_from_iovec(skb, 0, msg->msg_iov, 0, len);
L
Linus Torvalds 已提交
1499 1500 1501 1502 1503 1504 1505 1506 1507 1508 1509
	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;

1510
		other = unix_find_other(net, sunaddr, namelen, sk->sk_type,
L
Linus Torvalds 已提交
1511
					hash, &err);
1512
		if (other == NULL)
L
Linus Torvalds 已提交
1513 1514 1515
			goto out_free;
	}

A
Alban Crequy 已提交
1516 1517 1518 1519 1520 1521
	if (sk_filter(other, skb) < 0) {
		/* Toss the packet but do not return any error to the sender */
		err = len;
		goto out_free;
	}

1522
	unix_state_lock(other);
L
Linus Torvalds 已提交
1523 1524 1525 1526 1527 1528 1529 1530 1531
	err = -EPERM;
	if (!unix_may_send(sk, other))
		goto out_unlock;

	if (sock_flag(other, SOCK_DEAD)) {
		/*
		 *	Check with 1003.1g - what should
		 *	datagram error
		 */
1532
		unix_state_unlock(other);
L
Linus Torvalds 已提交
1533 1534 1535
		sock_put(other);

		err = 0;
1536
		unix_state_lock(sk);
L
Linus Torvalds 已提交
1537
		if (unix_peer(sk) == other) {
1538
			unix_peer(sk) = NULL;
1539
			unix_state_unlock(sk);
L
Linus Torvalds 已提交
1540 1541 1542 1543 1544

			unix_dgram_disconnected(sk, other);
			sock_put(other);
			err = -ECONNREFUSED;
		} else {
1545
			unix_state_unlock(sk);
L
Linus Torvalds 已提交
1546 1547 1548 1549 1550 1551 1552 1553 1554 1555 1556 1557 1558 1559 1560 1561 1562 1563
		}

		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;
	}

1564
	if (unix_peer(other) != sk && unix_recvq_full(other)) {
L
Linus Torvalds 已提交
1565 1566 1567 1568 1569 1570 1571 1572 1573 1574 1575 1576 1577 1578
		if (!timeo) {
			err = -EAGAIN;
			goto out_unlock;
		}

		timeo = unix_wait_for_peer(other, timeo);

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

		goto restart;
	}

1579 1580
	if (sock_flag(other, SOCK_RCVTSTAMP))
		__net_timestamp(skb);
1581
	maybe_add_creds(skb, sock, other);
L
Linus Torvalds 已提交
1582
	skb_queue_tail(&other->sk_receive_queue, skb);
E
Eric Dumazet 已提交
1583 1584
	if (max_level > unix_sk(other)->recursion_level)
		unix_sk(other)->recursion_level = max_level;
1585
	unix_state_unlock(other);
L
Linus Torvalds 已提交
1586 1587
	other->sk_data_ready(other, len);
	sock_put(other);
1588
	scm_destroy(siocb->scm);
L
Linus Torvalds 已提交
1589 1590 1591
	return len;

out_unlock:
1592
	unix_state_unlock(other);
L
Linus Torvalds 已提交
1593 1594 1595 1596 1597
out_free:
	kfree_skb(skb);
out:
	if (other)
		sock_put(other);
1598
	scm_destroy(siocb->scm);
L
Linus Torvalds 已提交
1599 1600 1601
	return err;
}

1602

L
Linus Torvalds 已提交
1603 1604 1605 1606 1607 1608
static int unix_stream_sendmsg(struct kiocb *kiocb, struct socket *sock,
			       struct msghdr *msg, size_t len)
{
	struct sock_iocb *siocb = kiocb_to_siocb(kiocb);
	struct sock *sk = sock->sk;
	struct sock *other = NULL;
E
Eric Dumazet 已提交
1609
	int err, size;
1610
	struct sk_buff *skb;
1611
	int sent = 0;
L
Linus Torvalds 已提交
1612
	struct scm_cookie tmp_scm;
1613
	bool fds_sent = false;
E
Eric Dumazet 已提交
1614
	int max_level;
L
Linus Torvalds 已提交
1615 1616 1617

	if (NULL == siocb->scm)
		siocb->scm = &tmp_scm;
1618
	wait_for_unix_gc();
1619
	err = scm_send(sock, msg, siocb->scm, false);
L
Linus Torvalds 已提交
1620 1621 1622 1623 1624 1625 1626 1627 1628 1629 1630 1631
	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;
1632
		other = unix_peer(sk);
L
Linus Torvalds 已提交
1633 1634 1635 1636 1637 1638 1639
		if (!other)
			goto out_err;
	}

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

E
Eric Dumazet 已提交
1640
	while (sent < len) {
L
Linus Torvalds 已提交
1641
		/*
1642 1643
		 *	Optimisation for the fact that under 0.01% of X
		 *	messages typically need breaking up.
L
Linus Torvalds 已提交
1644 1645
		 */

1646
		size = len-sent;
L
Linus Torvalds 已提交
1647 1648

		/* Keep two messages in the pipe so it schedules better */
1649 1650
		if (size > ((sk->sk_sndbuf >> 1) - 64))
			size = (sk->sk_sndbuf >> 1) - 64;
L
Linus Torvalds 已提交
1651 1652 1653

		if (size > SKB_MAX_ALLOC)
			size = SKB_MAX_ALLOC;
1654

L
Linus Torvalds 已提交
1655 1656 1657
		/*
		 *	Grab a buffer
		 */
1658

E
Eric Dumazet 已提交
1659 1660
		skb = sock_alloc_send_skb(sk, size, msg->msg_flags&MSG_DONTWAIT,
					  &err);
L
Linus Torvalds 已提交
1661

1662
		if (skb == NULL)
L
Linus Torvalds 已提交
1663 1664 1665 1666 1667 1668 1669 1670 1671 1672 1673
			goto out_err;

		/*
		 *	If you pass two values to the sock_alloc_send_skb
		 *	it tries to grab the large buffer with GFP_NOFS
		 *	(which can fail easily), and if it fails grab the
		 *	fallback size buffer which is under a page and will
		 *	succeed. [Alan]
		 */
		size = min_t(int, size, skb_tailroom(skb));

1674

1675 1676
		/* Only send the fds in the first buffer */
		err = unix_scm_to_skb(siocb->scm, skb, !fds_sent);
E
Eric Dumazet 已提交
1677
		if (err < 0) {
1678
			kfree_skb(skb);
1679
			goto out_err;
1680
		}
E
Eric Dumazet 已提交
1681
		max_level = err + 1;
1682
		fds_sent = true;
L
Linus Torvalds 已提交
1683

E
Eric Dumazet 已提交
1684 1685
		err = memcpy_fromiovec(skb_put(skb, size), msg->msg_iov, size);
		if (err) {
L
Linus Torvalds 已提交
1686
			kfree_skb(skb);
1687
			goto out_err;
L
Linus Torvalds 已提交
1688 1689
		}

1690
		unix_state_lock(other);
L
Linus Torvalds 已提交
1691 1692 1693 1694 1695

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

1696
		maybe_add_creds(skb, sock, other);
L
Linus Torvalds 已提交
1697
		skb_queue_tail(&other->sk_receive_queue, skb);
E
Eric Dumazet 已提交
1698 1699
		if (max_level > unix_sk(other)->recursion_level)
			unix_sk(other)->recursion_level = max_level;
1700
		unix_state_unlock(other);
L
Linus Torvalds 已提交
1701
		other->sk_data_ready(other, size);
1702
		sent += size;
L
Linus Torvalds 已提交
1703 1704
	}

1705
	scm_destroy(siocb->scm);
L
Linus Torvalds 已提交
1706 1707 1708 1709 1710
	siocb->scm = NULL;

	return sent;

pipe_err_free:
1711
	unix_state_unlock(other);
L
Linus Torvalds 已提交
1712 1713
	kfree_skb(skb);
pipe_err:
E
Eric Dumazet 已提交
1714 1715
	if (sent == 0 && !(msg->msg_flags&MSG_NOSIGNAL))
		send_sig(SIGPIPE, current, 0);
L
Linus Torvalds 已提交
1716 1717
	err = -EPIPE;
out_err:
1718
	scm_destroy(siocb->scm);
L
Linus Torvalds 已提交
1719 1720 1721 1722 1723 1724 1725 1726 1727
	siocb->scm = NULL;
	return sent ? : err;
}

static int unix_seqpacket_sendmsg(struct kiocb *kiocb, struct socket *sock,
				  struct msghdr *msg, size_t len)
{
	int err;
	struct sock *sk = sock->sk;
1728

L
Linus Torvalds 已提交
1729 1730 1731 1732 1733 1734 1735 1736 1737 1738 1739 1740
	err = sock_error(sk);
	if (err)
		return err;

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

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

	return unix_dgram_sendmsg(kiocb, sock, msg, len);
}
1741

1742 1743 1744 1745 1746 1747 1748 1749 1750 1751 1752 1753
static int unix_seqpacket_recvmsg(struct kiocb *iocb, struct socket *sock,
			      struct msghdr *msg, size_t size,
			      int flags)
{
	struct sock *sk = sock->sk;

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

	return unix_dgram_recvmsg(iocb, sock, msg, size, flags);
}

L
Linus Torvalds 已提交
1754 1755 1756 1757 1758 1759 1760 1761 1762 1763 1764 1765 1766 1767 1768 1769 1770 1771 1772 1773 1774 1775
static void unix_copy_addr(struct msghdr *msg, struct sock *sk)
{
	struct unix_sock *u = unix_sk(sk);

	msg->msg_namelen = 0;
	if (u->addr) {
		msg->msg_namelen = u->addr->len;
		memcpy(msg->msg_name, u->addr->name, u->addr->len);
	}
}

static int unix_dgram_recvmsg(struct kiocb *iocb, struct socket *sock,
			      struct msghdr *msg, size_t size,
			      int flags)
{
	struct sock_iocb *siocb = kiocb_to_siocb(iocb);
	struct scm_cookie tmp_scm;
	struct sock *sk = sock->sk;
	struct unix_sock *u = unix_sk(sk);
	int noblock = flags & MSG_DONTWAIT;
	struct sk_buff *skb;
	int err;
1776
	int peeked, skip;
L
Linus Torvalds 已提交
1777 1778 1779 1780 1781 1782 1783

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

	msg->msg_namelen = 0;

1784 1785 1786 1787 1788
	err = mutex_lock_interruptible(&u->readlock);
	if (err) {
		err = sock_intr_errno(sock_rcvtimeo(sk, noblock));
		goto out;
	}
L
Linus Torvalds 已提交
1789

1790 1791 1792
	skip = sk_peek_offset(sk, flags);

	skb = __skb_recv_datagram(sk, flags, &peeked, &skip, &err);
1793 1794 1795 1796 1797 1798 1799
	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
1800
		goto out_unlock;
1801
	}
L
Linus Torvalds 已提交
1802

E
Eric Dumazet 已提交
1803 1804
	wake_up_interruptible_sync_poll(&u->peer_wait,
					POLLOUT | POLLWRNORM | POLLWRBAND);
L
Linus Torvalds 已提交
1805 1806 1807 1808

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

1809 1810 1811
	if (size > skb->len - skip)
		size = skb->len - skip;
	else if (size < skb->len - skip)
L
Linus Torvalds 已提交
1812 1813
		msg->msg_flags |= MSG_TRUNC;

1814
	err = skb_copy_datagram_iovec(skb, skip, msg->msg_iov, size);
L
Linus Torvalds 已提交
1815 1816 1817
	if (err)
		goto out_free;

1818 1819 1820
	if (sock_flag(sk, SOCK_RCVTSTAMP))
		__sock_recv_timestamp(msg, sk, skb);

L
Linus Torvalds 已提交
1821 1822 1823 1824
	if (!siocb->scm) {
		siocb->scm = &tmp_scm;
		memset(&tmp_scm, 0, sizeof(tmp_scm));
	}
1825
	scm_set_cred(siocb->scm, UNIXCB(skb).pid, UNIXCB(skb).cred);
C
Catherine Zhang 已提交
1826
	unix_set_secdata(siocb->scm, skb);
L
Linus Torvalds 已提交
1827

E
Eric Dumazet 已提交
1828
	if (!(flags & MSG_PEEK)) {
L
Linus Torvalds 已提交
1829 1830
		if (UNIXCB(skb).fp)
			unix_detach_fds(siocb->scm, skb);
1831 1832

		sk_peek_offset_bwd(sk, skb->len);
E
Eric Dumazet 已提交
1833
	} else {
L
Linus Torvalds 已提交
1834 1835 1836 1837 1838 1839
		/* 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)
1840 1841 1842 1843 1844

		   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
1845
		*/
1846 1847 1848

		sk_peek_offset_fwd(sk, size);

L
Linus Torvalds 已提交
1849 1850 1851
		if (UNIXCB(skb).fp)
			siocb->scm->fp = scm_fp_dup(UNIXCB(skb).fp);
	}
1852
	err = (flags & MSG_TRUNC) ? skb->len - skip : size;
L
Linus Torvalds 已提交
1853 1854 1855 1856

	scm_recv(sock, msg, siocb->scm, flags);

out_free:
E
Eric Dumazet 已提交
1857
	skb_free_datagram(sk, skb);
L
Linus Torvalds 已提交
1858
out_unlock:
I
Ingo Molnar 已提交
1859
	mutex_unlock(&u->readlock);
L
Linus Torvalds 已提交
1860 1861 1862 1863 1864 1865 1866
out:
	return err;
}

/*
 *	Sleep until data has arrive. But check for races..
 */
1867

E
Eric Dumazet 已提交
1868
static long unix_stream_data_wait(struct sock *sk, long timeo)
L
Linus Torvalds 已提交
1869 1870 1871
{
	DEFINE_WAIT(wait);

1872
	unix_state_lock(sk);
L
Linus Torvalds 已提交
1873 1874

	for (;;) {
E
Eric Dumazet 已提交
1875
		prepare_to_wait(sk_sleep(sk), &wait, TASK_INTERRUPTIBLE);
L
Linus Torvalds 已提交
1876

1877
		if (!skb_queue_empty(&sk->sk_receive_queue) ||
L
Linus Torvalds 已提交
1878 1879 1880 1881 1882 1883 1884
		    sk->sk_err ||
		    (sk->sk_shutdown & RCV_SHUTDOWN) ||
		    signal_pending(current) ||
		    !timeo)
			break;

		set_bit(SOCK_ASYNC_WAITDATA, &sk->sk_socket->flags);
1885
		unix_state_unlock(sk);
L
Linus Torvalds 已提交
1886
		timeo = schedule_timeout(timeo);
1887
		unix_state_lock(sk);
L
Linus Torvalds 已提交
1888 1889 1890
		clear_bit(SOCK_ASYNC_WAITDATA, &sk->sk_socket->flags);
	}

E
Eric Dumazet 已提交
1891
	finish_wait(sk_sleep(sk), &wait);
1892
	unix_state_unlock(sk);
L
Linus Torvalds 已提交
1893 1894 1895 1896 1897 1898 1899 1900 1901 1902 1903 1904 1905
	return timeo;
}



static int unix_stream_recvmsg(struct kiocb *iocb, struct socket *sock,
			       struct msghdr *msg, size_t size,
			       int flags)
{
	struct sock_iocb *siocb = kiocb_to_siocb(iocb);
	struct scm_cookie tmp_scm;
	struct sock *sk = sock->sk;
	struct unix_sock *u = unix_sk(sk);
1906
	struct sockaddr_un *sunaddr = msg->msg_name;
L
Linus Torvalds 已提交
1907 1908 1909 1910 1911
	int copied = 0;
	int check_creds = 0;
	int target;
	int err = 0;
	long timeo;
1912
	int skip;
L
Linus Torvalds 已提交
1913 1914 1915 1916 1917 1918 1919 1920 1921 1922 1923 1924 1925 1926 1927 1928 1929 1930 1931 1932 1933 1934 1935

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

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

	target = sock_rcvlowat(sk, flags&MSG_WAITALL, size);
	timeo = sock_rcvtimeo(sk, flags&MSG_DONTWAIT);

	msg->msg_namelen = 0;

	/* Lock the socket to prevent queue disordering
	 * while sleeps in memcpy_tomsg
	 */

	if (!siocb->scm) {
		siocb->scm = &tmp_scm;
		memset(&tmp_scm, 0, sizeof(tmp_scm));
	}

1936 1937 1938 1939 1940
	err = mutex_lock_interruptible(&u->readlock);
	if (err) {
		err = sock_intr_errno(timeo);
		goto out;
	}
L
Linus Torvalds 已提交
1941

1942 1943
	skip = sk_peek_offset(sk, flags);

E
Eric Dumazet 已提交
1944
	do {
L
Linus Torvalds 已提交
1945 1946 1947
		int chunk;
		struct sk_buff *skb;

1948
		unix_state_lock(sk);
1949
		skb = skb_peek(&sk->sk_receive_queue);
1950
again:
E
Eric Dumazet 已提交
1951
		if (skb == NULL) {
E
Eric Dumazet 已提交
1952
			unix_sk(sk)->recursion_level = 0;
L
Linus Torvalds 已提交
1953
			if (copied >= target)
1954
				goto unlock;
L
Linus Torvalds 已提交
1955 1956 1957 1958

			/*
			 *	POSIX 1003.1g mandates this order.
			 */
1959

E
Eric Dumazet 已提交
1960 1961
			err = sock_error(sk);
			if (err)
1962
				goto unlock;
L
Linus Torvalds 已提交
1963
			if (sk->sk_shutdown & RCV_SHUTDOWN)
1964 1965 1966
				goto unlock;

			unix_state_unlock(sk);
L
Linus Torvalds 已提交
1967 1968 1969
			err = -EAGAIN;
			if (!timeo)
				break;
I
Ingo Molnar 已提交
1970
			mutex_unlock(&u->readlock);
L
Linus Torvalds 已提交
1971 1972 1973

			timeo = unix_stream_data_wait(sk, timeo);

1974 1975
			if (signal_pending(current)
			    ||  mutex_lock_interruptible(&u->readlock)) {
L
Linus Torvalds 已提交
1976 1977 1978
				err = sock_intr_errno(timeo);
				goto out;
			}
1979

L
Linus Torvalds 已提交
1980
			continue;
1981 1982 1983
 unlock:
			unix_state_unlock(sk);
			break;
L
Linus Torvalds 已提交
1984
		}
1985 1986 1987 1988 1989 1990 1991

		if (skip >= skb->len) {
			skip -= skb->len;
			skb = skb_peek_next(skb, &sk->sk_receive_queue);
			goto again;
		}

1992
		unix_state_unlock(sk);
L
Linus Torvalds 已提交
1993 1994 1995

		if (check_creds) {
			/* Never glue messages from different writers */
1996
			if ((UNIXCB(skb).pid  != siocb->scm->pid) ||
1997
			    (UNIXCB(skb).cred != siocb->scm->cred))
L
Linus Torvalds 已提交
1998 1999 2000
				break;
		} else {
			/* Copy credentials */
2001
			scm_set_cred(siocb->scm, UNIXCB(skb).pid, UNIXCB(skb).cred);
L
Linus Torvalds 已提交
2002 2003 2004 2005
			check_creds = 1;
		}

		/* Copy address just once */
E
Eric Dumazet 已提交
2006
		if (sunaddr) {
L
Linus Torvalds 已提交
2007 2008 2009 2010
			unix_copy_addr(msg, skb->sk);
			sunaddr = NULL;
		}

2011 2012
		chunk = min_t(unsigned int, skb->len - skip, size);
		if (memcpy_toiovec(msg->msg_iov, skb->data + skip, chunk)) {
L
Linus Torvalds 已提交
2013 2014 2015 2016 2017 2018 2019 2020
			if (copied == 0)
				copied = -EFAULT;
			break;
		}
		copied += chunk;
		size -= chunk;

		/* Mark read part of skb as used */
E
Eric Dumazet 已提交
2021
		if (!(flags & MSG_PEEK)) {
L
Linus Torvalds 已提交
2022 2023
			skb_pull(skb, chunk);

2024 2025
			sk_peek_offset_bwd(sk, chunk);

L
Linus Torvalds 已提交
2026 2027 2028
			if (UNIXCB(skb).fp)
				unix_detach_fds(siocb->scm, skb);

2029
			if (skb->len)
L
Linus Torvalds 已提交
2030 2031
				break;

2032
			skb_unlink(skb, &sk->sk_receive_queue);
2033
			consume_skb(skb);
L
Linus Torvalds 已提交
2034 2035 2036

			if (siocb->scm->fp)
				break;
E
Eric Dumazet 已提交
2037
		} else {
L
Linus Torvalds 已提交
2038 2039 2040 2041 2042
			/* It is questionable, see note in unix_dgram_recvmsg.
			 */
			if (UNIXCB(skb).fp)
				siocb->scm->fp = scm_fp_dup(UNIXCB(skb).fp);

2043 2044
			sk_peek_offset_fwd(sk, chunk);

L
Linus Torvalds 已提交
2045 2046 2047 2048
			break;
		}
	} while (size);

I
Ingo Molnar 已提交
2049
	mutex_unlock(&u->readlock);
L
Linus Torvalds 已提交
2050 2051 2052 2053 2054 2055 2056 2057 2058 2059
	scm_recv(sock, msg, siocb->scm, flags);
out:
	return copied ? : err;
}

static int unix_shutdown(struct socket *sock, int mode)
{
	struct sock *sk = sock->sk;
	struct sock *other;

2060 2061 2062 2063 2064 2065 2066 2067
	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;
2068 2069 2070 2071 2072 2073 2074 2075 2076 2077 2078 2079 2080 2081 2082 2083 2084 2085 2086 2087 2088 2089 2090 2091 2092 2093

	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
2094
	}
2095 2096 2097
	if (other)
		sock_put(other);

L
Linus Torvalds 已提交
2098 2099 2100
	return 0;
}

2101 2102 2103 2104 2105 2106 2107 2108 2109 2110 2111 2112 2113 2114 2115 2116 2117 2118 2119 2120 2121 2122 2123 2124 2125 2126 2127 2128 2129 2130
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)
			amount += skb->len;
	} 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
2131 2132 2133
static int unix_ioctl(struct socket *sock, unsigned int cmd, unsigned long arg)
{
	struct sock *sk = sock->sk;
2134
	long amount = 0;
L
Linus Torvalds 已提交
2135 2136
	int err;

E
Eric Dumazet 已提交
2137 2138
	switch (cmd) {
	case SIOCOUTQ:
2139
		amount = unix_outq_len(sk);
E
Eric Dumazet 已提交
2140 2141 2142
		err = put_user(amount, (int __user *)arg);
		break;
	case SIOCINQ:
2143 2144 2145 2146
		amount = unix_inq_len(sk);
		if (amount < 0)
			err = amount;
		else
L
Linus Torvalds 已提交
2147
			err = put_user(amount, (int __user *)arg);
2148
		break;
E
Eric Dumazet 已提交
2149 2150 2151
	default:
		err = -ENOIOCTLCMD;
		break;
L
Linus Torvalds 已提交
2152 2153 2154 2155
	}
	return err;
}

E
Eric Dumazet 已提交
2156
static unsigned int unix_poll(struct file *file, struct socket *sock, poll_table *wait)
L
Linus Torvalds 已提交
2157 2158 2159 2160
{
	struct sock *sk = sock->sk;
	unsigned int mask;

E
Eric Dumazet 已提交
2161
	sock_poll_wait(file, sk_sleep(sk), wait);
L
Linus Torvalds 已提交
2162 2163 2164 2165 2166 2167 2168
	mask = 0;

	/* exceptional events? */
	if (sk->sk_err)
		mask |= POLLERR;
	if (sk->sk_shutdown == SHUTDOWN_MASK)
		mask |= POLLHUP;
2169
	if (sk->sk_shutdown & RCV_SHUTDOWN)
E
Eric Dumazet 已提交
2170
		mask |= POLLRDHUP | POLLIN | POLLRDNORM;
L
Linus Torvalds 已提交
2171 2172

	/* readable? */
E
Eric Dumazet 已提交
2173
	if (!skb_queue_empty(&sk->sk_receive_queue))
L
Linus Torvalds 已提交
2174 2175 2176
		mask |= POLLIN | POLLRDNORM;

	/* Connection-based need to check for termination and startup */
E
Eric Dumazet 已提交
2177 2178
	if ((sk->sk_type == SOCK_STREAM || sk->sk_type == SOCK_SEQPACKET) &&
	    sk->sk_state == TCP_CLOSE)
L
Linus Torvalds 已提交
2179 2180 2181 2182 2183 2184 2185 2186 2187 2188 2189 2190
		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;
}

2191 2192
static unsigned int unix_dgram_poll(struct file *file, struct socket *sock,
				    poll_table *wait)
2193
{
2194 2195
	struct sock *sk = sock->sk, *other;
	unsigned int mask, writable;
2196

E
Eric Dumazet 已提交
2197
	sock_poll_wait(file, sk_sleep(sk), wait);
2198 2199 2200 2201 2202 2203
	mask = 0;

	/* exceptional events? */
	if (sk->sk_err || !skb_queue_empty(&sk->sk_error_queue))
		mask |= POLLERR;
	if (sk->sk_shutdown & RCV_SHUTDOWN)
2204
		mask |= POLLRDHUP | POLLIN | POLLRDNORM;
2205 2206 2207 2208
	if (sk->sk_shutdown == SHUTDOWN_MASK)
		mask |= POLLHUP;

	/* readable? */
2209
	if (!skb_queue_empty(&sk->sk_receive_queue))
2210 2211 2212 2213 2214 2215 2216 2217 2218 2219 2220
		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
2221
	/* No write status requested, avoid expensive OUT tests. */
2222
	if (!(poll_requested_events(wait) & (POLLWRBAND|POLLWRNORM|POLLOUT)))
E
Eric Dumazet 已提交
2223 2224
		return mask;

2225
	writable = unix_writable(sk);
2226 2227 2228 2229 2230 2231
	other = unix_peer_get(sk);
	if (other) {
		if (unix_peer(other) != sk) {
			sock_poll_wait(file, &unix_sk(other)->peer_wait, wait);
			if (unix_recvq_full(other))
				writable = 0;
2232
		}
2233
		sock_put(other);
2234 2235 2236
	}

	if (writable)
2237 2238 2239 2240 2241 2242
		mask |= POLLOUT | POLLWRNORM | POLLWRBAND;
	else
		set_bit(SOCK_ASYNC_NOSPACE, &sk->sk_socket->flags);

	return mask;
}
L
Linus Torvalds 已提交
2243 2244

#ifdef CONFIG_PROC_FS
2245

E
Eric Dumazet 已提交
2246 2247 2248 2249 2250
#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))
2251

E
Eric Dumazet 已提交
2252
static struct sock *unix_from_bucket(struct seq_file *seq, loff_t *pos)
L
Linus Torvalds 已提交
2253
{
E
Eric Dumazet 已提交
2254 2255 2256 2257
	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
2258

E
Eric Dumazet 已提交
2259 2260
	for (sk = sk_head(&unix_socket_table[bucket]); sk; sk = sk_next(sk)) {
		if (sock_net(sk) != seq_file_net(seq))
2261
			continue;
E
Eric Dumazet 已提交
2262 2263 2264 2265 2266 2267 2268 2269 2270 2271 2272 2273 2274 2275 2276 2277 2278 2279 2280
		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
2281
	}
E
Eric Dumazet 已提交
2282 2283 2284 2285 2286 2287 2288 2289 2290 2291 2292

	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
2293 2294 2295 2296
	return NULL;
}

static void *unix_seq_start(struct seq_file *seq, loff_t *pos)
2297
	__acquires(unix_table_lock)
L
Linus Torvalds 已提交
2298
{
2299
	spin_lock(&unix_table_lock);
E
Eric Dumazet 已提交
2300 2301 2302 2303 2304 2305 2306 2307

	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
2308 2309 2310 2311 2312
}

static void *unix_seq_next(struct seq_file *seq, void *v, loff_t *pos)
{
	++*pos;
E
Eric Dumazet 已提交
2313
	return unix_next_socket(seq, v, pos);
L
Linus Torvalds 已提交
2314 2315 2316
}

static void unix_seq_stop(struct seq_file *seq, void *v)
2317
	__releases(unix_table_lock)
L
Linus Torvalds 已提交
2318
{
2319
	spin_unlock(&unix_table_lock);
L
Linus Torvalds 已提交
2320 2321 2322 2323
}

static int unix_seq_show(struct seq_file *seq, void *v)
{
2324

J
Joe Perches 已提交
2325
	if (v == SEQ_START_TOKEN)
L
Linus Torvalds 已提交
2326 2327 2328 2329 2330
		seq_puts(seq, "Num       RefCount Protocol Flags    Type St "
			 "Inode Path\n");
	else {
		struct sock *s = v;
		struct unix_sock *u = unix_sk(s);
2331
		unix_state_lock(s);
L
Linus Torvalds 已提交
2332

D
Dan Rosenberg 已提交
2333
		seq_printf(seq, "%pK: %08X %08X %08X %04X %02X %5lu",
L
Linus Torvalds 已提交
2334 2335 2336 2337 2338 2339 2340 2341 2342 2343 2344 2345 2346 2347 2348 2349 2350 2351 2352 2353 2354 2355 2356 2357 2358
			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]);
		}
2359
		unix_state_unlock(s);
L
Linus Torvalds 已提交
2360 2361 2362 2363 2364 2365
		seq_putc(seq, '\n');
	}

	return 0;
}

2366
static const struct seq_operations unix_seq_ops = {
L
Linus Torvalds 已提交
2367 2368 2369 2370 2371 2372 2373 2374
	.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)
{
2375
	return seq_open_net(inode, file, &unix_seq_ops,
E
Eric Dumazet 已提交
2376
			    sizeof(struct seq_net_private));
L
Linus Torvalds 已提交
2377 2378
}

2379
static const struct file_operations unix_seq_fops = {
L
Linus Torvalds 已提交
2380 2381 2382 2383
	.owner		= THIS_MODULE,
	.open		= unix_seq_open,
	.read		= seq_read,
	.llseek		= seq_lseek,
2384
	.release	= seq_release_net,
L
Linus Torvalds 已提交
2385 2386 2387 2388
};

#endif

2389
static const struct net_proto_family unix_family_ops = {
L
Linus Torvalds 已提交
2390 2391 2392 2393 2394
	.family = PF_UNIX,
	.create = unix_create,
	.owner	= THIS_MODULE,
};

2395

2396
static int __net_init unix_net_init(struct net *net)
2397 2398 2399
{
	int error = -ENOMEM;

2400
	net->unx.sysctl_max_dgram_qlen = 10;
2401 2402
	if (unix_sysctl_register(net))
		goto out;
2403

2404
#ifdef CONFIG_PROC_FS
2405 2406
	if (!proc_net_fops_create(net, "unix", 0, &unix_seq_fops)) {
		unix_sysctl_unregister(net);
2407
		goto out;
2408
	}
2409 2410 2411
#endif
	error = 0;
out:
2412
	return error;
2413 2414
}

2415
static void __net_exit unix_net_exit(struct net *net)
2416
{
2417
	unix_sysctl_unregister(net);
2418 2419 2420 2421 2422 2423 2424 2425
	proc_net_remove(net, "unix");
}

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

L
Linus Torvalds 已提交
2426 2427 2428 2429 2430
static int __init af_unix_init(void)
{
	int rc = -1;
	struct sk_buff *dummy_skb;

2431
	BUILD_BUG_ON(sizeof(struct unix_skb_parms) > sizeof(dummy_skb->cb));
L
Linus Torvalds 已提交
2432 2433

	rc = proto_register(&unix_proto, 1);
2434 2435
	if (rc != 0) {
		printk(KERN_CRIT "%s: Cannot create unix_sock SLAB cache!\n",
2436
		       __func__);
L
Linus Torvalds 已提交
2437 2438 2439 2440
		goto out;
	}

	sock_register(&unix_family_ops);
2441
	register_pernet_subsys(&unix_net_ops);
L
Linus Torvalds 已提交
2442 2443 2444 2445 2446 2447 2448 2449
out:
	return rc;
}

static void __exit af_unix_exit(void)
{
	sock_unregister(PF_UNIX);
	proto_unregister(&unix_proto);
2450
	unregister_pernet_subsys(&unix_net_ops);
L
Linus Torvalds 已提交
2451 2452
}

2453 2454 2455 2456 2457
/* 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
2458 2459 2460 2461
module_exit(af_unix_exit);

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