af_unix.c 55.2 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>

118 119 120 121
struct hlist_head unix_socket_table[UNIX_HASH_SIZE + 1];
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
123 124 125 126 127

#define unix_sockets_unbound	(&unix_socket_table[UNIX_HASH_SIZE])

#define UNIX_ABSTRACT(sk)	(unix_sk(sk)->addr->hash != UNIX_HASH_SIZE)

C
Catherine Zhang 已提交
128
#ifdef CONFIG_SECURITY_NETWORK
129
static void unix_get_secdata(struct scm_cookie *scm, struct sk_buff *skb)
C
Catherine Zhang 已提交
130
{
131
	memcpy(UNIXSID(skb), &scm->secid, sizeof(u32));
C
Catherine Zhang 已提交
132 133 134 135
}

static inline void unix_set_secdata(struct scm_cookie *scm, struct sk_buff *skb)
{
136
	scm->secid = *UNIXSID(skb);
C
Catherine Zhang 已提交
137 138
}
#else
139
static inline void unix_get_secdata(struct scm_cookie *scm, struct sk_buff *skb)
C
Catherine Zhang 已提交
140 141 142 143 144 145
{ }

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

L
Linus Torvalds 已提交
146 147
/*
 *  SMP locking strategy:
148
 *    hash table is protected with spinlock unix_table_lock
149
 *    each socket state is protected by separate spin lock.
L
Linus Torvalds 已提交
150 151
 */

152
static inline unsigned unix_hash_fold(__wsum n)
L
Linus Torvalds 已提交
153
{
154
	unsigned hash = (__force unsigned)n;
L
Linus Torvalds 已提交
155 156 157 158 159 160 161 162 163 164 165 166 167 168
	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
169
	return unix_peer(osk) == NULL || unix_our_peer(sk, osk);
L
Linus Torvalds 已提交
170 171
}

172 173 174 175 176
static inline int unix_recvq_full(struct sock const *sk)
{
	return skb_queue_len(&sk->sk_receive_queue) > sk->sk_max_ack_backlog;
}

177
struct sock *unix_peer_get(struct sock *s)
L
Linus Torvalds 已提交
178 179 180
{
	struct sock *peer;

181
	unix_state_lock(s);
L
Linus Torvalds 已提交
182 183 184
	peer = unix_peer(s);
	if (peer)
		sock_hold(peer);
185
	unix_state_unlock(s);
L
Linus Torvalds 已提交
186 187
	return peer;
}
188
EXPORT_SYMBOL_GPL(unix_peer_get);
L
Linus Torvalds 已提交
189 190 191 192 193 194 195 196 197 198 199 200 201

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.
 */
202

E
Eric Dumazet 已提交
203
static int unix_mkname(struct sockaddr_un *sunaddr, int len, unsigned *hashp)
L
Linus Torvalds 已提交
204 205 206 207 208 209 210 211 212
{
	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
213
		 * sun_path[108] doesn't as such exist.  However in kernel space
L
Linus Torvalds 已提交
214 215 216
		 * we are guaranteed that it is a valid memory location in our
		 * kernel address buffer.
		 */
217
		((char *)sunaddr)[len] = 0;
L
Linus Torvalds 已提交
218 219 220 221
		len = strlen(sunaddr->sun_path)+1+sizeof(short);
		return len;
	}

222
	*hashp = unix_hash_fold(csum_partial(sunaddr, len, 0));
L
Linus Torvalds 已提交
223 224 225 226 227 228 229 230 231 232
	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)
{
233
	WARN_ON(!sk_unhashed(sk));
L
Linus Torvalds 已提交
234 235 236 237 238
	sk_add_node(sk, list);
}

static inline void unix_remove_socket(struct sock *sk)
{
239
	spin_lock(&unix_table_lock);
L
Linus Torvalds 已提交
240
	__unix_remove_socket(sk);
241
	spin_unlock(&unix_table_lock);
L
Linus Torvalds 已提交
242 243 244 245
}

static inline void unix_insert_socket(struct hlist_head *list, struct sock *sk)
{
246
	spin_lock(&unix_table_lock);
L
Linus Torvalds 已提交
247
	__unix_insert_socket(list, sk);
248
	spin_unlock(&unix_table_lock);
L
Linus Torvalds 已提交
249 250
}

251 252
static struct sock *__unix_find_socket_byname(struct net *net,
					      struct sockaddr_un *sunname,
L
Linus Torvalds 已提交
253 254 255 256 257 258 259 260
					      int len, int type, unsigned hash)
{
	struct sock *s;
	struct hlist_node *node;

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

261
		if (!net_eq(sock_net(s), net))
262 263
			continue;

L
Linus Torvalds 已提交
264 265 266 267 268 269 270 271 272
		if (u->addr->len == len &&
		    !memcmp(u->addr->name, sunname, len))
			goto found;
	}
	s = NULL;
found:
	return s;
}

273 274
static inline struct sock *unix_find_socket_byname(struct net *net,
						   struct sockaddr_un *sunname,
L
Linus Torvalds 已提交
275 276 277 278 279
						   int len, int type,
						   unsigned hash)
{
	struct sock *s;

280
	spin_lock(&unix_table_lock);
281
	s = __unix_find_socket_byname(net, sunname, len, type, hash);
L
Linus Torvalds 已提交
282 283
	if (s)
		sock_hold(s);
284
	spin_unlock(&unix_table_lock);
L
Linus Torvalds 已提交
285 286 287
	return s;
}

288
static struct sock *unix_find_socket_byinode(struct inode *i)
L
Linus Torvalds 已提交
289 290 291 292
{
	struct sock *s;
	struct hlist_node *node;

293
	spin_lock(&unix_table_lock);
L
Linus Torvalds 已提交
294 295 296 297
	sk_for_each(s, node,
		    &unix_socket_table[i->i_ino & (UNIX_HASH_SIZE - 1)]) {
		struct dentry *dentry = unix_sk(s)->dentry;

E
Eric Dumazet 已提交
298
		if (dentry && dentry->d_inode == i) {
L
Linus Torvalds 已提交
299 300 301 302 303 304
			sock_hold(s);
			goto found;
		}
	}
	s = NULL;
found:
305
	spin_unlock(&unix_table_lock);
L
Linus Torvalds 已提交
306 307 308 309 310 311 312 313 314 315
	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)
{
316 317 318
	struct socket_wq *wq;

	rcu_read_lock();
L
Linus Torvalds 已提交
319
	if (unix_writable(sk)) {
320 321
		wq = rcu_dereference(sk->sk_wq);
		if (wq_has_sleeper(wq))
E
Eric Dumazet 已提交
322 323
			wake_up_interruptible_sync_poll(&wq->wait,
				POLLOUT | POLLWRNORM | POLLWRBAND);
324
		sk_wake_async(sk, SOCK_WAKE_SPACE, POLL_OUT);
L
Linus Torvalds 已提交
325
	}
326
	rcu_read_unlock();
L
Linus Torvalds 已提交
327 328 329 330 331 332 333 334
}

/* 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)
{
335
	if (!skb_queue_empty(&sk->sk_receive_queue)) {
L
Linus Torvalds 已提交
336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355
		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);

356 357 358
	WARN_ON(atomic_read(&sk->sk_wmem_alloc));
	WARN_ON(!sk_unhashed(sk));
	WARN_ON(sk->sk_socket);
L
Linus Torvalds 已提交
359
	if (!sock_flag(sk, SOCK_DEAD)) {
360
		printk(KERN_INFO "Attempt to release alive unix socket: %p\n", sk);
L
Linus Torvalds 已提交
361 362 363 364 365 366
		return;
	}

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

E
Eric Dumazet 已提交
367
	atomic_long_dec(&unix_nr_socks);
368
	local_bh_disable();
369
	sock_prot_inuse_add(sock_net(sk), sk->sk_prot, -1);
370
	local_bh_enable();
L
Linus Torvalds 已提交
371
#ifdef UNIX_REFCNT_DEBUG
E
Eric Dumazet 已提交
372 373
	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
374 375 376
#endif
}

E
Eric Dumazet 已提交
377
static int unix_release_sock(struct sock *sk, int embrion)
L
Linus Torvalds 已提交
378 379 380 381 382 383 384 385 386 387 388
{
	struct unix_sock *u = unix_sk(sk);
	struct dentry *dentry;
	struct vfsmount *mnt;
	struct sock *skpair;
	struct sk_buff *skb;
	int state;

	unix_remove_socket(sk);

	/* Clear state */
389
	unix_state_lock(sk);
L
Linus Torvalds 已提交
390 391 392 393 394 395 396 397
	sock_orphan(sk);
	sk->sk_shutdown = SHUTDOWN_MASK;
	dentry	     = u->dentry;
	u->dentry    = NULL;
	mnt	     = u->mnt;
	u->mnt	     = NULL;
	state = sk->sk_state;
	sk->sk_state = TCP_CLOSE;
398
	unix_state_unlock(sk);
L
Linus Torvalds 已提交
399 400 401

	wake_up_interruptible_all(&u->peer_wait);

402
	skpair = unix_peer(sk);
L
Linus Torvalds 已提交
403

404
	if (skpair != NULL) {
L
Linus Torvalds 已提交
405
		if (sk->sk_type == SOCK_STREAM || sk->sk_type == SOCK_SEQPACKET) {
406
			unix_state_lock(skpair);
L
Linus Torvalds 已提交
407 408 409 410
			/* No more writes */
			skpair->sk_shutdown = SHUTDOWN_MASK;
			if (!skb_queue_empty(&sk->sk_receive_queue) || embrion)
				skpair->sk_err = ECONNRESET;
411
			unix_state_unlock(skpair);
L
Linus Torvalds 已提交
412
			skpair->sk_state_change(skpair);
413
			sk_wake_async(skpair, SOCK_WAKE_WAITD, POLL_HUP);
L
Linus Torvalds 已提交
414 415 416 417 418 419 420 421
		}
		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) {
422
		if (state == TCP_LISTEN)
L
Linus Torvalds 已提交
423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 439 440 441 442 443 444 445 446 447
			unix_release_sock(skb->sk, 1);
		/* passed fds are erased in the kfree_skb hook	      */
		kfree_skb(skb);
	}

	if (dentry) {
		dput(dentry);
		mntput(mnt);
	}

	sock_put(sk);

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

	/*
	 * Fixme: BSD difference: In BSD all sockets connected to use get
	 *	  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)
	 */

448
	if (unix_tot_inflight)
449
		unix_gc();		/* Garbage collect fds */
L
Linus Torvalds 已提交
450 451 452 453

	return 0;
}

454 455 456 457 458 459 460 461 462 463 464 465 466 467 468 469 470 471
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
472 473 474 475 476
static int unix_listen(struct socket *sock, int backlog)
{
	int err;
	struct sock *sk = sock->sk;
	struct unix_sock *u = unix_sk(sk);
477 478
	struct pid *old_pid = NULL;
	const struct cred *old_cred = NULL;
L
Linus Torvalds 已提交
479 480

	err = -EOPNOTSUPP;
E
Eric Dumazet 已提交
481 482
	if (sock->type != SOCK_STREAM && sock->type != SOCK_SEQPACKET)
		goto out;	/* Only stream/seqpacket sockets accept */
L
Linus Torvalds 已提交
483 484
	err = -EINVAL;
	if (!u->addr)
E
Eric Dumazet 已提交
485
		goto out;	/* No listens on an unbound socket */
486
	unix_state_lock(sk);
L
Linus Torvalds 已提交
487 488 489 490 491 492 493
	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 */
494
	init_peercred(sk);
L
Linus Torvalds 已提交
495 496 497
	err = 0;

out_unlock:
498
	unix_state_unlock(sk);
499 500 501
	put_pid(old_pid);
	if (old_cred)
		put_cred(old_cred);
L
Linus Torvalds 已提交
502 503 504 505 506 507 508 509 510 511 512 513
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 *);
514 515
static unsigned int unix_dgram_poll(struct file *, struct socket *,
				    poll_table *);
L
Linus Torvalds 已提交
516 517 518 519 520 521 522 523 524 525 526 527 528 529
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);
530 531
static int unix_seqpacket_recvmsg(struct kiocb *, struct socket *,
				  struct msghdr *, size_t, int);
L
Linus Torvalds 已提交
532

533
static const struct proto_ops unix_stream_ops = {
L
Linus Torvalds 已提交
534 535 536 537 538 539 540 541 542 543 544 545 546 547 548 549 550 551 552 553
	.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,
};

554
static const struct proto_ops unix_dgram_ops = {
L
Linus Torvalds 已提交
555 556 557 558 559 560 561 562
	.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,
563
	.poll =		unix_dgram_poll,
L
Linus Torvalds 已提交
564 565 566 567 568 569 570 571 572 573 574
	.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,
};

575
static const struct proto_ops unix_seqpacket_ops = {
L
Linus Torvalds 已提交
576 577 578 579 580 581 582 583
	.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,
584
	.poll =		unix_dgram_poll,
L
Linus Torvalds 已提交
585 586 587 588 589 590
	.ioctl =	unix_ioctl,
	.listen =	unix_listen,
	.shutdown =	unix_shutdown,
	.setsockopt =	sock_no_setsockopt,
	.getsockopt =	sock_no_getsockopt,
	.sendmsg =	unix_seqpacket_sendmsg,
591
	.recvmsg =	unix_seqpacket_recvmsg,
L
Linus Torvalds 已提交
592 593 594 595 596
	.mmap =		sock_no_mmap,
	.sendpage =	sock_no_sendpage,
};

static struct proto unix_proto = {
597 598 599
	.name			= "UNIX",
	.owner			= THIS_MODULE,
	.obj_size		= sizeof(struct unix_sock),
L
Linus Torvalds 已提交
600 601
};

602 603 604 605 606 607 608 609
/*
 * 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
610
static struct sock *unix_create1(struct net *net, struct socket *sock)
L
Linus Torvalds 已提交
611 612 613 614
{
	struct sock *sk = NULL;
	struct unix_sock *u;

E
Eric Dumazet 已提交
615 616
	atomic_long_inc(&unix_nr_socks);
	if (atomic_long_read(&unix_nr_socks) > 2 * get_max_files())
L
Linus Torvalds 已提交
617 618
		goto out;

619
	sk = sk_alloc(net, PF_UNIX, GFP_KERNEL, &unix_proto);
L
Linus Torvalds 已提交
620 621 622
	if (!sk)
		goto out;

E
Eric Dumazet 已提交
623
	sock_init_data(sock, sk);
624 625
	lockdep_set_class(&sk->sk_receive_queue.lock,
				&af_unix_sk_receive_queue_lock_key);
L
Linus Torvalds 已提交
626 627

	sk->sk_write_space	= unix_write_space;
628
	sk->sk_max_ack_backlog	= net->unx.sysctl_max_dgram_qlen;
L
Linus Torvalds 已提交
629 630 631 632
	sk->sk_destruct		= unix_sock_destructor;
	u	  = unix_sk(sk);
	u->dentry = NULL;
	u->mnt	  = NULL;
633
	spin_lock_init(&u->lock);
A
Al Viro 已提交
634
	atomic_long_set(&u->inflight, 0);
635
	INIT_LIST_HEAD(&u->link);
I
Ingo Molnar 已提交
636
	mutex_init(&u->readlock); /* single task reading lock */
L
Linus Torvalds 已提交
637 638 639
	init_waitqueue_head(&u->peer_wait);
	unix_insert_socket(unix_sockets_unbound, sk);
out:
640
	if (sk == NULL)
E
Eric Dumazet 已提交
641
		atomic_long_dec(&unix_nr_socks);
642 643
	else {
		local_bh_disable();
644
		sock_prot_inuse_add(sock_net(sk), sk->sk_prot, 1);
645 646
		local_bh_enable();
	}
L
Linus Torvalds 已提交
647 648 649
	return sk;
}

650 651
static int unix_create(struct net *net, struct socket *sock, int protocol,
		       int kern)
L
Linus Torvalds 已提交
652 653 654 655 656 657 658 659 660 661 662 663 664 665 666
{
	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:
667
		sock->type = SOCK_DGRAM;
L
Linus Torvalds 已提交
668 669 670 671 672 673 674 675 676 677
	case SOCK_DGRAM:
		sock->ops = &unix_dgram_ops;
		break;
	case SOCK_SEQPACKET:
		sock->ops = &unix_seqpacket_ops;
		break;
	default:
		return -ESOCKTNOSUPPORT;
	}

678
	return unix_create1(net, sock) ? 0 : -ENOMEM;
L
Linus Torvalds 已提交
679 680 681 682 683 684 685 686 687 688 689
}

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
690
	return unix_release_sock(sk, 0);
L
Linus Torvalds 已提交
691 692 693 694 695
}

static int unix_autobind(struct socket *sock)
{
	struct sock *sk = sock->sk;
696
	struct net *net = sock_net(sk);
L
Linus Torvalds 已提交
697 698
	struct unix_sock *u = unix_sk(sk);
	static u32 ordernum = 1;
E
Eric Dumazet 已提交
699
	struct unix_address *addr;
L
Linus Torvalds 已提交
700
	int err;
701
	unsigned int retries = 0;
L
Linus Torvalds 已提交
702

I
Ingo Molnar 已提交
703
	mutex_lock(&u->readlock);
L
Linus Torvalds 已提交
704 705 706 707 708 709

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

	err = -ENOMEM;
710
	addr = kzalloc(sizeof(*addr) + sizeof(short) + 16, GFP_KERNEL);
L
Linus Torvalds 已提交
711 712 713 714 715 716 717 718
	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);
719
	addr->hash = unix_hash_fold(csum_partial(addr->name, addr->len, 0));
L
Linus Torvalds 已提交
720

721
	spin_lock(&unix_table_lock);
L
Linus Torvalds 已提交
722 723
	ordernum = (ordernum+1)&0xFFFFF;

724
	if (__unix_find_socket_byname(net, addr->name, addr->len, sock->type,
L
Linus Torvalds 已提交
725
				      addr->hash)) {
726
		spin_unlock(&unix_table_lock);
727 728 729 730 731 732 733 734 735 736 737
		/*
		 * __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
738 739 740 741 742 743 744
		goto retry;
	}
	addr->hash ^= sk->sk_type;

	__unix_remove_socket(sk);
	u->addr = addr;
	__unix_insert_socket(&unix_socket_table[addr->hash], sk);
745
	spin_unlock(&unix_table_lock);
L
Linus Torvalds 已提交
746 747
	err = 0;

I
Ingo Molnar 已提交
748
out:	mutex_unlock(&u->readlock);
L
Linus Torvalds 已提交
749 750 751
	return err;
}

752 753
static struct sock *unix_find_other(struct net *net,
				    struct sockaddr_un *sunname, int len,
L
Linus Torvalds 已提交
754 755 756
				    int type, unsigned hash, int *error)
{
	struct sock *u;
757
	struct path path;
L
Linus Torvalds 已提交
758
	int err = 0;
759

L
Linus Torvalds 已提交
760
	if (sunname->sun_path[0]) {
761 762
		struct inode *inode;
		err = kern_path(sunname->sun_path, LOOKUP_FOLLOW, &path);
L
Linus Torvalds 已提交
763 764
		if (err)
			goto fail;
765 766
		inode = path.dentry->d_inode;
		err = inode_permission(inode, MAY_WRITE);
L
Linus Torvalds 已提交
767 768 769 770
		if (err)
			goto put_fail;

		err = -ECONNREFUSED;
771
		if (!S_ISSOCK(inode->i_mode))
L
Linus Torvalds 已提交
772
			goto put_fail;
773
		u = unix_find_socket_byinode(inode);
L
Linus Torvalds 已提交
774 775 776 777
		if (!u)
			goto put_fail;

		if (u->sk_type == type)
778
			touch_atime(path.mnt, path.dentry);
L
Linus Torvalds 已提交
779

780
		path_put(&path);
L
Linus Torvalds 已提交
781

782
		err = -EPROTOTYPE;
L
Linus Torvalds 已提交
783 784 785 786 787 788
		if (u->sk_type != type) {
			sock_put(u);
			goto fail;
		}
	} else {
		err = -ECONNREFUSED;
789
		u = unix_find_socket_byname(net, sunname, len, type, hash);
L
Linus Torvalds 已提交
790 791 792 793 794 795 796 797 798 799 800
		if (u) {
			struct dentry *dentry;
			dentry = unix_sk(u)->dentry;
			if (dentry)
				touch_atime(unix_sk(u)->mnt, dentry);
		} else
			goto fail;
	}
	return u;

put_fail:
801
	path_put(&path);
L
Linus Torvalds 已提交
802
fail:
803
	*error = err;
L
Linus Torvalds 已提交
804 805 806 807 808 809 810
	return NULL;
}


static int unix_bind(struct socket *sock, struct sockaddr *uaddr, int addr_len)
{
	struct sock *sk = sock->sk;
811
	struct net *net = sock_net(sk);
L
Linus Torvalds 已提交
812
	struct unix_sock *u = unix_sk(sk);
813
	struct sockaddr_un *sunaddr = (struct sockaddr_un *)uaddr;
814
	char *sun_path = sunaddr->sun_path;
E
Eric Dumazet 已提交
815
	struct dentry *dentry = NULL;
816
	struct path path;
L
Linus Torvalds 已提交
817 818 819 820 821 822 823 824 825
	int err;
	unsigned hash;
	struct unix_address *addr;
	struct hlist_head *list;

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

826
	if (addr_len == sizeof(short)) {
L
Linus Torvalds 已提交
827 828 829 830 831 832 833 834 835
		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
836
	mutex_lock(&u->readlock);
L
Linus Torvalds 已提交
837 838 839 840 841 842 843 844 845 846 847 848 849 850 851

	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);

852
	if (sun_path[0]) {
A
Al Viro 已提交
853
		umode_t mode;
L
Linus Torvalds 已提交
854 855 856 857 858
		err = 0;
		/*
		 * Get the parent directory, calculate the hash for last
		 * component.
		 */
859
		dentry = kern_path_create(AT_FDCWD, sun_path, &path, 0);
L
Linus Torvalds 已提交
860 861
		err = PTR_ERR(dentry);
		if (IS_ERR(dentry))
862
			goto out_mknod_parent;
863

L
Linus Torvalds 已提交
864 865 866 867
		/*
		 * All right, let's create it.
		 */
		mode = S_IFSOCK |
A
Al Viro 已提交
868
		       (SOCK_INODE(sock)->i_mode & ~current_umask());
869
		err = mnt_want_write(path.mnt);
870 871
		if (err)
			goto out_mknod_dput;
872
		err = security_path_mknod(&path, dentry, mode, 0);
873 874
		if (err)
			goto out_mknod_drop_write;
875
		err = vfs_mknod(path.dentry->d_inode, dentry, mode, 0);
876
out_mknod_drop_write:
877
		mnt_drop_write(path.mnt);
L
Linus Torvalds 已提交
878 879
		if (err)
			goto out_mknod_dput;
880 881 882
		mutex_unlock(&path.dentry->d_inode->i_mutex);
		dput(path.dentry);
		path.dentry = dentry;
L
Linus Torvalds 已提交
883 884 885 886

		addr->hash = UNIX_HASH_SIZE;
	}

887
	spin_lock(&unix_table_lock);
L
Linus Torvalds 已提交
888

889
	if (!sun_path[0]) {
L
Linus Torvalds 已提交
890
		err = -EADDRINUSE;
891
		if (__unix_find_socket_byname(net, sunaddr, addr_len,
L
Linus Torvalds 已提交
892 893 894 895 896 897 898 899
					      sk->sk_type, hash)) {
			unix_release_addr(addr);
			goto out_unlock;
		}

		list = &unix_socket_table[addr->hash];
	} else {
		list = &unix_socket_table[dentry->d_inode->i_ino & (UNIX_HASH_SIZE-1)];
900 901
		u->dentry = path.dentry;
		u->mnt    = path.mnt;
L
Linus Torvalds 已提交
902 903 904 905 906 907 908 909
	}

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

out_unlock:
910
	spin_unlock(&unix_table_lock);
L
Linus Torvalds 已提交
911
out_up:
I
Ingo Molnar 已提交
912
	mutex_unlock(&u->readlock);
L
Linus Torvalds 已提交
913 914 915 916 917
out:
	return err;

out_mknod_dput:
	dput(dentry);
918 919
	mutex_unlock(&path.dentry->d_inode->i_mutex);
	path_put(&path);
L
Linus Torvalds 已提交
920
out_mknod_parent:
921 922
	if (err == -EEXIST)
		err = -EADDRINUSE;
L
Linus Torvalds 已提交
923 924 925 926
	unix_release_addr(addr);
	goto out_up;
}

927 928 929 930 931 932 933 934 935 936 937 938 939 940 941 942 943 944 945 946 947 948 949 950 951
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
952 953 954 955
static int unix_dgram_connect(struct socket *sock, struct sockaddr *addr,
			      int alen, int flags)
{
	struct sock *sk = sock->sk;
956
	struct net *net = sock_net(sk);
957
	struct sockaddr_un *sunaddr = (struct sockaddr_un *)addr;
L
Linus Torvalds 已提交
958 959 960 961 962 963 964 965 966 967 968 969 970 971
	struct sock *other;
	unsigned hash;
	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;

972
restart:
973
		other = unix_find_other(net, sunaddr, alen, sock->type, hash, &err);
L
Linus Torvalds 已提交
974 975 976
		if (!other)
			goto out;

977 978 979 980 981 982 983 984
		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
985 986 987 988 989 990 991 992 993 994 995 996 997 998

		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;
999
		unix_state_double_lock(sk, other);
L
Linus Torvalds 已提交
1000 1001 1002 1003 1004 1005 1006
	}

	/*
	 * If it was connected, reconnect.
	 */
	if (unix_peer(sk)) {
		struct sock *old_peer = unix_peer(sk);
1007
		unix_peer(sk) = other;
1008
		unix_state_double_unlock(sk, other);
L
Linus Torvalds 已提交
1009 1010 1011 1012 1013

		if (other != old_peer)
			unix_dgram_disconnected(sk, old_peer);
		sock_put(old_peer);
	} else {
1014
		unix_peer(sk) = other;
1015
		unix_state_double_unlock(sk, other);
L
Linus Torvalds 已提交
1016
	}
1017
	return 0;
L
Linus Torvalds 已提交
1018 1019

out_unlock:
1020
	unix_state_double_unlock(sk, other);
L
Linus Torvalds 已提交
1021 1022 1023 1024 1025 1026 1027 1028 1029 1030 1031 1032 1033 1034 1035
	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) &&
1036
		unix_recvq_full(other);
L
Linus Torvalds 已提交
1037

1038
	unix_state_unlock(other);
L
Linus Torvalds 已提交
1039 1040 1041 1042 1043 1044 1045 1046 1047 1048 1049

	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)
{
1050
	struct sockaddr_un *sunaddr = (struct sockaddr_un *)uaddr;
L
Linus Torvalds 已提交
1051
	struct sock *sk = sock->sk;
1052
	struct net *net = sock_net(sk);
L
Linus Torvalds 已提交
1053 1054 1055 1056 1057 1058 1059 1060 1061 1062 1063 1064 1065 1066
	struct unix_sock *u = unix_sk(sk), *newu, *otheru;
	struct sock *newsk = NULL;
	struct sock *other = NULL;
	struct sk_buff *skb = NULL;
	unsigned hash;
	int st;
	int err;
	long timeo;

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

1067 1068
	if (test_bit(SOCK_PASSCRED, &sock->flags) && !u->addr &&
	    (err = unix_autobind(sock)) != 0)
L
Linus Torvalds 已提交
1069 1070 1071 1072 1073 1074 1075 1076 1077 1078 1079 1080
		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 */
1081
	newsk = unix_create1(sock_net(sk), NULL);
L
Linus Torvalds 已提交
1082 1083 1084 1085 1086 1087 1088 1089 1090 1091
	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. */
1092
	other = unix_find_other(net, sunaddr, addr_len, sk->sk_type, hash, &err);
L
Linus Torvalds 已提交
1093 1094 1095 1096
	if (!other)
		goto out;

	/* Latch state of peer */
1097
	unix_state_lock(other);
L
Linus Torvalds 已提交
1098 1099 1100

	/* Apparently VFS overslept socket death. Retry. */
	if (sock_flag(other, SOCK_DEAD)) {
1101
		unix_state_unlock(other);
L
Linus Torvalds 已提交
1102 1103 1104 1105 1106 1107 1108
		sock_put(other);
		goto restart;
	}

	err = -ECONNREFUSED;
	if (other->sk_state != TCP_LISTEN)
		goto out_unlock;
1109 1110
	if (other->sk_shutdown & RCV_SHUTDOWN)
		goto out_unlock;
L
Linus Torvalds 已提交
1111

1112
	if (unix_recvq_full(other)) {
L
Linus Torvalds 已提交
1113 1114 1115 1116 1117 1118 1119 1120 1121 1122 1123
		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;
1124
	}
L
Linus Torvalds 已提交
1125 1126 1127

	/* Latch our state.

D
Daniel Baluta 已提交
1128
	   It is tricky place. We need to grab our state lock and cannot
L
Linus Torvalds 已提交
1129 1130 1131 1132 1133 1134 1135 1136 1137 1138 1139 1140 1141 1142 1143 1144 1145 1146 1147 1148 1149 1150 1151
	   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;
	}

1152
	unix_state_lock_nested(sk);
L
Linus Torvalds 已提交
1153 1154

	if (sk->sk_state != st) {
1155 1156
		unix_state_unlock(sk);
		unix_state_unlock(other);
L
Linus Torvalds 已提交
1157 1158 1159 1160
		sock_put(other);
		goto restart;
	}

1161
	err = security_unix_stream_connect(sk, other, newsk);
L
Linus Torvalds 已提交
1162
	if (err) {
1163
		unix_state_unlock(sk);
L
Linus Torvalds 已提交
1164 1165 1166 1167 1168 1169 1170 1171 1172
		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;
1173
	init_peercred(newsk);
L
Linus Torvalds 已提交
1174
	newu = unix_sk(newsk);
1175
	RCU_INIT_POINTER(newsk->sk_wq, &newu->peer_wq);
L
Linus Torvalds 已提交
1176 1177 1178 1179 1180 1181 1182 1183 1184 1185 1186 1187 1188
	otheru = unix_sk(other);

	/* copy address information from listening to new sock*/
	if (otheru->addr) {
		atomic_inc(&otheru->addr->refcnt);
		newu->addr = otheru->addr;
	}
	if (otheru->dentry) {
		newu->dentry	= dget(otheru->dentry);
		newu->mnt	= mntget(otheru->mnt);
	}

	/* Set credentials */
1189
	copy_peercred(sk, other);
L
Linus Torvalds 已提交
1190 1191 1192

	sock->state	= SS_CONNECTED;
	sk->sk_state	= TCP_ESTABLISHED;
1193 1194 1195 1196
	sock_hold(newsk);

	smp_mb__after_atomic_inc();	/* sock_hold() does an atomic_inc() */
	unix_peer(sk)	= newsk;
L
Linus Torvalds 已提交
1197

1198
	unix_state_unlock(sk);
L
Linus Torvalds 已提交
1199 1200 1201 1202 1203

	/* 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);
1204
	unix_state_unlock(other);
L
Linus Torvalds 已提交
1205 1206 1207 1208 1209 1210
	other->sk_data_ready(other, 0);
	sock_put(other);
	return 0;

out_unlock:
	if (other)
1211
		unix_state_unlock(other);
L
Linus Torvalds 已提交
1212 1213

out:
1214
	kfree_skb(skb);
L
Linus Torvalds 已提交
1215 1216 1217 1218 1219 1220 1221 1222 1223
	if (newsk)
		unix_release_sock(newsk, 0);
	if (other)
		sock_put(other);
	return err;
}

static int unix_socketpair(struct socket *socka, struct socket *sockb)
{
1224
	struct sock *ska = socka->sk, *skb = sockb->sk;
L
Linus Torvalds 已提交
1225 1226 1227 1228

	/* Join our sockets back to back */
	sock_hold(ska);
	sock_hold(skb);
1229 1230
	unix_peer(ska) = skb;
	unix_peer(skb) = ska;
1231 1232
	init_peercred(ska);
	init_peercred(skb);
L
Linus Torvalds 已提交
1233 1234 1235 1236 1237 1238 1239 1240 1241 1242 1243 1244 1245 1246 1247 1248 1249 1250

	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
1251
	if (sock->type != SOCK_STREAM && sock->type != SOCK_SEQPACKET)
L
Linus Torvalds 已提交
1252 1253 1254 1255 1256 1257 1258 1259 1260 1261 1262 1263 1264 1265 1266 1267 1268 1269 1270 1271 1272 1273 1274
		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 */
1275
	unix_state_lock(tsk);
L
Linus Torvalds 已提交
1276 1277
	newsock->state = SS_CONNECTED;
	sock_graft(tsk, newsock);
1278
	unix_state_unlock(tsk);
L
Linus Torvalds 已提交
1279 1280 1281 1282 1283 1284 1285 1286 1287 1288 1289
	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;
1290
	DECLARE_SOCKADDR(struct sockaddr_un *, sunaddr, uaddr);
L
Linus Torvalds 已提交
1291 1292 1293 1294 1295 1296 1297 1298 1299 1300 1301 1302 1303 1304
	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);
1305
	unix_state_lock(sk);
L
Linus Torvalds 已提交
1306 1307 1308 1309 1310 1311 1312 1313 1314 1315
	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);
	}
1316
	unix_state_unlock(sk);
L
Linus Torvalds 已提交
1317 1318 1319 1320 1321 1322 1323 1324 1325 1326 1327 1328
	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
1329
	for (i = scm->fp->count-1; i >= 0; i--)
L
Linus Torvalds 已提交
1330 1331 1332
		unix_notinflight(scm->fp->fp[i]);
}

1333
static void unix_destruct_scm(struct sk_buff *skb)
L
Linus Torvalds 已提交
1334 1335 1336
{
	struct scm_cookie scm;
	memset(&scm, 0, sizeof(scm));
1337 1338 1339 1340
	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
1341 1342 1343 1344 1345 1346 1347

	/* 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
1348 1349
#define MAX_RECURSION_LEVEL 4

1350
static int unix_attach_fds(struct scm_cookie *scm, struct sk_buff *skb)
L
Linus Torvalds 已提交
1351 1352
{
	int i;
E
Eric Dumazet 已提交
1353 1354 1355 1356 1357 1358 1359 1360 1361 1362 1363 1364 1365 1366
	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;
1367 1368 1369 1370 1371 1372 1373 1374 1375 1376

	/*
	 * 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
1377 1378 1379 1380 1381
	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
1382 1383
}

1384
static int unix_scm_to_skb(struct scm_cookie *scm, struct sk_buff *skb, bool send_fds)
1385 1386
{
	int err = 0;
1387

1388
	UNIXCB(skb).pid  = get_pid(scm->pid);
1389 1390
	if (scm->cred)
		UNIXCB(skb).cred = get_cred(scm->cred);
1391 1392 1393 1394 1395 1396 1397 1398
	UNIXCB(skb).fp = NULL;
	if (scm->fp && send_fds)
		err = unix_attach_fds(scm, skb);

	skb->destructor = unix_destruct_scm;
	return err;
}

1399 1400 1401 1402 1403 1404 1405 1406 1407 1408 1409 1410 1411 1412 1413 1414 1415 1416
/*
 * 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
1417 1418 1419 1420 1421 1422 1423 1424 1425
/*
 *	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;
1426
	struct net *net = sock_net(sk);
L
Linus Torvalds 已提交
1427
	struct unix_sock *u = unix_sk(sk);
1428
	struct sockaddr_un *sunaddr = msg->msg_name;
L
Linus Torvalds 已提交
1429 1430 1431 1432
	struct sock *other = NULL;
	int namelen = 0; /* fake GCC */
	int err;
	unsigned hash;
1433
	struct sk_buff *skb;
L
Linus Torvalds 已提交
1434 1435
	long timeo;
	struct scm_cookie tmp_scm;
E
Eric Dumazet 已提交
1436
	int max_level;
L
Linus Torvalds 已提交
1437 1438 1439

	if (NULL == siocb->scm)
		siocb->scm = &tmp_scm;
1440
	wait_for_unix_gc();
L
Linus Torvalds 已提交
1441 1442 1443 1444 1445 1446 1447 1448 1449 1450 1451 1452 1453 1454 1455 1456 1457 1458 1459 1460 1461
	err = scm_send(sock, msg, siocb->scm);
	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;
	}

1462 1463
	if (test_bit(SOCK_PASSCRED, &sock->flags) && !u->addr
	    && (err = unix_autobind(sock)) != 0)
L
Linus Torvalds 已提交
1464 1465 1466 1467 1468 1469 1470
		goto out;

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

	skb = sock_alloc_send_skb(sk, len, msg->msg_flags&MSG_DONTWAIT, &err);
1471
	if (skb == NULL)
L
Linus Torvalds 已提交
1472 1473
		goto out;

1474
	err = unix_scm_to_skb(siocb->scm, skb, true);
E
Eric Dumazet 已提交
1475
	if (err < 0)
1476
		goto out_free;
E
Eric Dumazet 已提交
1477
	max_level = err + 1;
1478
	unix_get_secdata(siocb->scm, skb);
C
Catherine Zhang 已提交
1479

1480
	skb_reset_transport_header(skb);
E
Eric Dumazet 已提交
1481
	err = memcpy_fromiovec(skb_put(skb, len), msg->msg_iov, len);
L
Linus Torvalds 已提交
1482 1483 1484 1485 1486 1487 1488 1489 1490 1491 1492
	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;

1493
		other = unix_find_other(net, sunaddr, namelen, sk->sk_type,
L
Linus Torvalds 已提交
1494
					hash, &err);
1495
		if (other == NULL)
L
Linus Torvalds 已提交
1496 1497 1498
			goto out_free;
	}

A
Alban Crequy 已提交
1499 1500 1501 1502 1503 1504
	if (sk_filter(other, skb) < 0) {
		/* Toss the packet but do not return any error to the sender */
		err = len;
		goto out_free;
	}

1505
	unix_state_lock(other);
L
Linus Torvalds 已提交
1506 1507 1508 1509 1510 1511 1512 1513 1514
	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
		 */
1515
		unix_state_unlock(other);
L
Linus Torvalds 已提交
1516 1517 1518
		sock_put(other);

		err = 0;
1519
		unix_state_lock(sk);
L
Linus Torvalds 已提交
1520
		if (unix_peer(sk) == other) {
1521
			unix_peer(sk) = NULL;
1522
			unix_state_unlock(sk);
L
Linus Torvalds 已提交
1523 1524 1525 1526 1527

			unix_dgram_disconnected(sk, other);
			sock_put(other);
			err = -ECONNREFUSED;
		} else {
1528
			unix_state_unlock(sk);
L
Linus Torvalds 已提交
1529 1530 1531 1532 1533 1534 1535 1536 1537 1538 1539 1540 1541 1542 1543 1544 1545 1546
		}

		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;
	}

1547
	if (unix_peer(other) != sk && unix_recvq_full(other)) {
L
Linus Torvalds 已提交
1548 1549 1550 1551 1552 1553 1554 1555 1556 1557 1558 1559 1560 1561
		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;
	}

1562 1563
	if (sock_flag(other, SOCK_RCVTSTAMP))
		__net_timestamp(skb);
1564
	maybe_add_creds(skb, sock, other);
L
Linus Torvalds 已提交
1565
	skb_queue_tail(&other->sk_receive_queue, skb);
E
Eric Dumazet 已提交
1566 1567
	if (max_level > unix_sk(other)->recursion_level)
		unix_sk(other)->recursion_level = max_level;
1568
	unix_state_unlock(other);
L
Linus Torvalds 已提交
1569 1570
	other->sk_data_ready(other, len);
	sock_put(other);
1571
	scm_destroy(siocb->scm);
L
Linus Torvalds 已提交
1572 1573 1574
	return len;

out_unlock:
1575
	unix_state_unlock(other);
L
Linus Torvalds 已提交
1576 1577 1578 1579 1580
out_free:
	kfree_skb(skb);
out:
	if (other)
		sock_put(other);
1581
	scm_destroy(siocb->scm);
L
Linus Torvalds 已提交
1582 1583 1584
	return err;
}

1585

L
Linus Torvalds 已提交
1586 1587 1588 1589 1590 1591
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
1592
	int err, size;
1593
	struct sk_buff *skb;
1594
	int sent = 0;
L
Linus Torvalds 已提交
1595
	struct scm_cookie tmp_scm;
1596
	bool fds_sent = false;
E
Eric Dumazet 已提交
1597
	int max_level;
L
Linus Torvalds 已提交
1598 1599 1600

	if (NULL == siocb->scm)
		siocb->scm = &tmp_scm;
1601
	wait_for_unix_gc();
L
Linus Torvalds 已提交
1602 1603 1604 1605 1606 1607 1608 1609 1610 1611 1612 1613 1614
	err = scm_send(sock, msg, siocb->scm);
	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;
1615
		other = unix_peer(sk);
L
Linus Torvalds 已提交
1616 1617 1618 1619 1620 1621 1622
		if (!other)
			goto out_err;
	}

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

E
Eric Dumazet 已提交
1623
	while (sent < len) {
L
Linus Torvalds 已提交
1624
		/*
1625 1626
		 *	Optimisation for the fact that under 0.01% of X
		 *	messages typically need breaking up.
L
Linus Torvalds 已提交
1627 1628
		 */

1629
		size = len-sent;
L
Linus Torvalds 已提交
1630 1631

		/* Keep two messages in the pipe so it schedules better */
1632 1633
		if (size > ((sk->sk_sndbuf >> 1) - 64))
			size = (sk->sk_sndbuf >> 1) - 64;
L
Linus Torvalds 已提交
1634 1635 1636

		if (size > SKB_MAX_ALLOC)
			size = SKB_MAX_ALLOC;
1637

L
Linus Torvalds 已提交
1638 1639 1640
		/*
		 *	Grab a buffer
		 */
1641

E
Eric Dumazet 已提交
1642 1643
		skb = sock_alloc_send_skb(sk, size, msg->msg_flags&MSG_DONTWAIT,
					  &err);
L
Linus Torvalds 已提交
1644

1645
		if (skb == NULL)
L
Linus Torvalds 已提交
1646 1647 1648 1649 1650 1651 1652 1653 1654 1655 1656
			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));

1657

1658 1659
		/* Only send the fds in the first buffer */
		err = unix_scm_to_skb(siocb->scm, skb, !fds_sent);
E
Eric Dumazet 已提交
1660
		if (err < 0) {
1661
			kfree_skb(skb);
1662
			goto out_err;
1663
		}
E
Eric Dumazet 已提交
1664
		max_level = err + 1;
1665
		fds_sent = true;
L
Linus Torvalds 已提交
1666

E
Eric Dumazet 已提交
1667 1668
		err = memcpy_fromiovec(skb_put(skb, size), msg->msg_iov, size);
		if (err) {
L
Linus Torvalds 已提交
1669
			kfree_skb(skb);
1670
			goto out_err;
L
Linus Torvalds 已提交
1671 1672
		}

1673
		unix_state_lock(other);
L
Linus Torvalds 已提交
1674 1675 1676 1677 1678

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

1679
		maybe_add_creds(skb, sock, other);
L
Linus Torvalds 已提交
1680
		skb_queue_tail(&other->sk_receive_queue, skb);
E
Eric Dumazet 已提交
1681 1682
		if (max_level > unix_sk(other)->recursion_level)
			unix_sk(other)->recursion_level = max_level;
1683
		unix_state_unlock(other);
L
Linus Torvalds 已提交
1684
		other->sk_data_ready(other, size);
1685
		sent += size;
L
Linus Torvalds 已提交
1686 1687
	}

1688
	scm_destroy(siocb->scm);
L
Linus Torvalds 已提交
1689 1690 1691 1692 1693
	siocb->scm = NULL;

	return sent;

pipe_err_free:
1694
	unix_state_unlock(other);
L
Linus Torvalds 已提交
1695 1696
	kfree_skb(skb);
pipe_err:
E
Eric Dumazet 已提交
1697 1698
	if (sent == 0 && !(msg->msg_flags&MSG_NOSIGNAL))
		send_sig(SIGPIPE, current, 0);
L
Linus Torvalds 已提交
1699 1700
	err = -EPIPE;
out_err:
1701
	scm_destroy(siocb->scm);
L
Linus Torvalds 已提交
1702 1703 1704 1705 1706 1707 1708 1709 1710
	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;
1711

L
Linus Torvalds 已提交
1712 1713 1714 1715 1716 1717 1718 1719 1720 1721 1722 1723
	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);
}
1724

1725 1726 1727 1728 1729 1730 1731 1732 1733 1734 1735 1736
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
1737 1738 1739 1740 1741 1742 1743 1744 1745 1746 1747 1748 1749 1750 1751 1752 1753 1754 1755 1756 1757 1758 1759 1760 1761 1762 1763 1764 1765
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;

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

	msg->msg_namelen = 0;

1766 1767 1768 1769 1770
	err = mutex_lock_interruptible(&u->readlock);
	if (err) {
		err = sock_intr_errno(sock_rcvtimeo(sk, noblock));
		goto out;
	}
L
Linus Torvalds 已提交
1771 1772

	skb = skb_recv_datagram(sk, flags, noblock, &err);
1773 1774 1775 1776 1777 1778 1779
	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
1780
		goto out_unlock;
1781
	}
L
Linus Torvalds 已提交
1782

E
Eric Dumazet 已提交
1783 1784
	wake_up_interruptible_sync_poll(&u->peer_wait,
					POLLOUT | POLLWRNORM | POLLWRBAND);
L
Linus Torvalds 已提交
1785 1786 1787 1788 1789 1790 1791 1792 1793 1794 1795 1796 1797

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

	if (size > skb->len)
		size = skb->len;
	else if (size < skb->len)
		msg->msg_flags |= MSG_TRUNC;

	err = skb_copy_datagram_iovec(skb, 0, msg->msg_iov, size);
	if (err)
		goto out_free;

1798 1799 1800
	if (sock_flag(sk, SOCK_RCVTSTAMP))
		__sock_recv_timestamp(msg, sk, skb);

L
Linus Torvalds 已提交
1801 1802 1803 1804
	if (!siocb->scm) {
		siocb->scm = &tmp_scm;
		memset(&tmp_scm, 0, sizeof(tmp_scm));
	}
1805
	scm_set_cred(siocb->scm, UNIXCB(skb).pid, UNIXCB(skb).cred);
C
Catherine Zhang 已提交
1806
	unix_set_secdata(siocb->scm, skb);
L
Linus Torvalds 已提交
1807

E
Eric Dumazet 已提交
1808
	if (!(flags & MSG_PEEK)) {
L
Linus Torvalds 已提交
1809 1810
		if (UNIXCB(skb).fp)
			unix_detach_fds(siocb->scm, skb);
E
Eric Dumazet 已提交
1811
	} else {
L
Linus Torvalds 已提交
1812 1813 1814 1815 1816 1817
		/* 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)
1818 1819 1820 1821 1822

		   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
1823 1824 1825 1826 1827 1828 1829 1830 1831
		*/
		if (UNIXCB(skb).fp)
			siocb->scm->fp = scm_fp_dup(UNIXCB(skb).fp);
	}
	err = size;

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

out_free:
E
Eric Dumazet 已提交
1832
	skb_free_datagram(sk, skb);
L
Linus Torvalds 已提交
1833
out_unlock:
I
Ingo Molnar 已提交
1834
	mutex_unlock(&u->readlock);
L
Linus Torvalds 已提交
1835 1836 1837 1838 1839 1840 1841
out:
	return err;
}

/*
 *	Sleep until data has arrive. But check for races..
 */
1842

E
Eric Dumazet 已提交
1843
static long unix_stream_data_wait(struct sock *sk, long timeo)
L
Linus Torvalds 已提交
1844 1845 1846
{
	DEFINE_WAIT(wait);

1847
	unix_state_lock(sk);
L
Linus Torvalds 已提交
1848 1849

	for (;;) {
E
Eric Dumazet 已提交
1850
		prepare_to_wait(sk_sleep(sk), &wait, TASK_INTERRUPTIBLE);
L
Linus Torvalds 已提交
1851

1852
		if (!skb_queue_empty(&sk->sk_receive_queue) ||
L
Linus Torvalds 已提交
1853 1854 1855 1856 1857 1858 1859
		    sk->sk_err ||
		    (sk->sk_shutdown & RCV_SHUTDOWN) ||
		    signal_pending(current) ||
		    !timeo)
			break;

		set_bit(SOCK_ASYNC_WAITDATA, &sk->sk_socket->flags);
1860
		unix_state_unlock(sk);
L
Linus Torvalds 已提交
1861
		timeo = schedule_timeout(timeo);
1862
		unix_state_lock(sk);
L
Linus Torvalds 已提交
1863 1864 1865
		clear_bit(SOCK_ASYNC_WAITDATA, &sk->sk_socket->flags);
	}

E
Eric Dumazet 已提交
1866
	finish_wait(sk_sleep(sk), &wait);
1867
	unix_state_unlock(sk);
L
Linus Torvalds 已提交
1868 1869 1870 1871 1872 1873 1874 1875 1876 1877 1878 1879 1880
	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);
1881
	struct sockaddr_un *sunaddr = msg->msg_name;
L
Linus Torvalds 已提交
1882 1883 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
	int copied = 0;
	int check_creds = 0;
	int target;
	int err = 0;
	long timeo;

	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));
	}

1910 1911 1912 1913 1914
	err = mutex_lock_interruptible(&u->readlock);
	if (err) {
		err = sock_intr_errno(timeo);
		goto out;
	}
L
Linus Torvalds 已提交
1915

E
Eric Dumazet 已提交
1916
	do {
L
Linus Torvalds 已提交
1917 1918 1919
		int chunk;
		struct sk_buff *skb;

1920
		unix_state_lock(sk);
1921
		skb = skb_peek(&sk->sk_receive_queue);
E
Eric Dumazet 已提交
1922
		if (skb == NULL) {
E
Eric Dumazet 已提交
1923
			unix_sk(sk)->recursion_level = 0;
L
Linus Torvalds 已提交
1924
			if (copied >= target)
1925
				goto unlock;
L
Linus Torvalds 已提交
1926 1927 1928 1929

			/*
			 *	POSIX 1003.1g mandates this order.
			 */
1930

E
Eric Dumazet 已提交
1931 1932
			err = sock_error(sk);
			if (err)
1933
				goto unlock;
L
Linus Torvalds 已提交
1934
			if (sk->sk_shutdown & RCV_SHUTDOWN)
1935 1936 1937
				goto unlock;

			unix_state_unlock(sk);
L
Linus Torvalds 已提交
1938 1939 1940
			err = -EAGAIN;
			if (!timeo)
				break;
I
Ingo Molnar 已提交
1941
			mutex_unlock(&u->readlock);
L
Linus Torvalds 已提交
1942 1943 1944

			timeo = unix_stream_data_wait(sk, timeo);

1945 1946
			if (signal_pending(current)
			    ||  mutex_lock_interruptible(&u->readlock)) {
L
Linus Torvalds 已提交
1947 1948 1949
				err = sock_intr_errno(timeo);
				goto out;
			}
1950

L
Linus Torvalds 已提交
1951
			continue;
1952 1953 1954
 unlock:
			unix_state_unlock(sk);
			break;
L
Linus Torvalds 已提交
1955
		}
1956
		unix_state_unlock(sk);
L
Linus Torvalds 已提交
1957 1958 1959

		if (check_creds) {
			/* Never glue messages from different writers */
1960
			if ((UNIXCB(skb).pid  != siocb->scm->pid) ||
1961
			    (UNIXCB(skb).cred != siocb->scm->cred))
L
Linus Torvalds 已提交
1962 1963 1964
				break;
		} else {
			/* Copy credentials */
1965
			scm_set_cred(siocb->scm, UNIXCB(skb).pid, UNIXCB(skb).cred);
L
Linus Torvalds 已提交
1966 1967 1968 1969
			check_creds = 1;
		}

		/* Copy address just once */
E
Eric Dumazet 已提交
1970
		if (sunaddr) {
L
Linus Torvalds 已提交
1971 1972 1973 1974 1975 1976 1977 1978 1979 1980 1981 1982 1983 1984
			unix_copy_addr(msg, skb->sk);
			sunaddr = NULL;
		}

		chunk = min_t(unsigned int, skb->len, size);
		if (memcpy_toiovec(msg->msg_iov, skb->data, chunk)) {
			if (copied == 0)
				copied = -EFAULT;
			break;
		}
		copied += chunk;
		size -= chunk;

		/* Mark read part of skb as used */
E
Eric Dumazet 已提交
1985
		if (!(flags & MSG_PEEK)) {
L
Linus Torvalds 已提交
1986 1987 1988 1989 1990
			skb_pull(skb, chunk);

			if (UNIXCB(skb).fp)
				unix_detach_fds(siocb->scm, skb);

1991
			if (skb->len)
L
Linus Torvalds 已提交
1992 1993
				break;

1994
			skb_unlink(skb, &sk->sk_receive_queue);
1995
			consume_skb(skb);
L
Linus Torvalds 已提交
1996 1997 1998

			if (siocb->scm->fp)
				break;
E
Eric Dumazet 已提交
1999
		} else {
L
Linus Torvalds 已提交
2000 2001 2002 2003 2004 2005 2006 2007 2008
			/* It is questionable, see note in unix_dgram_recvmsg.
			 */
			if (UNIXCB(skb).fp)
				siocb->scm->fp = scm_fp_dup(UNIXCB(skb).fp);

			break;
		}
	} while (size);

I
Ingo Molnar 已提交
2009
	mutex_unlock(&u->readlock);
L
Linus Torvalds 已提交
2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 2020 2021
	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;

	mode = (mode+1)&(RCV_SHUTDOWN|SEND_SHUTDOWN);

2022 2023 2024 2025 2026 2027 2028 2029 2030 2031 2032 2033 2034 2035 2036 2037 2038 2039 2040 2041 2042 2043 2044 2045 2046 2047 2048 2049
	if (!mode)
		return 0;

	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
2050
	}
2051 2052 2053
	if (other)
		sock_put(other);

L
Linus Torvalds 已提交
2054 2055 2056
	return 0;
}

2057 2058 2059 2060 2061 2062 2063 2064 2065 2066 2067 2068 2069 2070 2071 2072 2073 2074 2075 2076 2077 2078 2079 2080 2081 2082 2083 2084 2085 2086
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
2087 2088 2089
static int unix_ioctl(struct socket *sock, unsigned int cmd, unsigned long arg)
{
	struct sock *sk = sock->sk;
2090
	long amount = 0;
L
Linus Torvalds 已提交
2091 2092
	int err;

E
Eric Dumazet 已提交
2093 2094
	switch (cmd) {
	case SIOCOUTQ:
2095
		amount = unix_outq_len(sk);
E
Eric Dumazet 已提交
2096 2097 2098
		err = put_user(amount, (int __user *)arg);
		break;
	case SIOCINQ:
2099 2100 2101 2102
		amount = unix_inq_len(sk);
		if (amount < 0)
			err = amount;
		else
L
Linus Torvalds 已提交
2103
			err = put_user(amount, (int __user *)arg);
2104
		break;
E
Eric Dumazet 已提交
2105 2106 2107
	default:
		err = -ENOIOCTLCMD;
		break;
L
Linus Torvalds 已提交
2108 2109 2110 2111
	}
	return err;
}

E
Eric Dumazet 已提交
2112
static unsigned int unix_poll(struct file *file, struct socket *sock, poll_table *wait)
L
Linus Torvalds 已提交
2113 2114 2115 2116
{
	struct sock *sk = sock->sk;
	unsigned int mask;

E
Eric Dumazet 已提交
2117
	sock_poll_wait(file, sk_sleep(sk), wait);
L
Linus Torvalds 已提交
2118 2119 2120 2121 2122 2123 2124
	mask = 0;

	/* exceptional events? */
	if (sk->sk_err)
		mask |= POLLERR;
	if (sk->sk_shutdown == SHUTDOWN_MASK)
		mask |= POLLHUP;
2125
	if (sk->sk_shutdown & RCV_SHUTDOWN)
E
Eric Dumazet 已提交
2126
		mask |= POLLRDHUP | POLLIN | POLLRDNORM;
L
Linus Torvalds 已提交
2127 2128

	/* readable? */
E
Eric Dumazet 已提交
2129
	if (!skb_queue_empty(&sk->sk_receive_queue))
L
Linus Torvalds 已提交
2130 2131 2132
		mask |= POLLIN | POLLRDNORM;

	/* Connection-based need to check for termination and startup */
E
Eric Dumazet 已提交
2133 2134
	if ((sk->sk_type == SOCK_STREAM || sk->sk_type == SOCK_SEQPACKET) &&
	    sk->sk_state == TCP_CLOSE)
L
Linus Torvalds 已提交
2135 2136 2137 2138 2139 2140 2141 2142 2143 2144 2145 2146
		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;
}

2147 2148
static unsigned int unix_dgram_poll(struct file *file, struct socket *sock,
				    poll_table *wait)
2149
{
2150 2151
	struct sock *sk = sock->sk, *other;
	unsigned int mask, writable;
2152

E
Eric Dumazet 已提交
2153
	sock_poll_wait(file, sk_sleep(sk), wait);
2154 2155 2156 2157 2158 2159
	mask = 0;

	/* exceptional events? */
	if (sk->sk_err || !skb_queue_empty(&sk->sk_error_queue))
		mask |= POLLERR;
	if (sk->sk_shutdown & RCV_SHUTDOWN)
2160
		mask |= POLLRDHUP | POLLIN | POLLRDNORM;
2161 2162 2163 2164
	if (sk->sk_shutdown == SHUTDOWN_MASK)
		mask |= POLLHUP;

	/* readable? */
2165
	if (!skb_queue_empty(&sk->sk_receive_queue))
2166 2167 2168 2169 2170 2171 2172 2173 2174 2175 2176
		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
2177 2178 2179 2180
	/* No write status requested, avoid expensive OUT tests. */
	if (wait && !(wait->key & (POLLWRBAND | POLLWRNORM | POLLOUT)))
		return mask;

2181
	writable = unix_writable(sk);
2182 2183 2184 2185 2186 2187
	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;
2188
		}
2189
		sock_put(other);
2190 2191 2192
	}

	if (writable)
2193 2194 2195 2196 2197 2198
		mask |= POLLOUT | POLLWRNORM | POLLWRBAND;
	else
		set_bit(SOCK_ASYNC_NOSPACE, &sk->sk_socket->flags);

	return mask;
}
L
Linus Torvalds 已提交
2199 2200

#ifdef CONFIG_PROC_FS
2201 2202 2203 2204 2205 2206 2207 2208 2209 2210 2211 2212 2213 2214 2215 2216 2217 2218 2219 2220 2221 2222 2223
static struct sock *first_unix_socket(int *i)
{
	for (*i = 0; *i <= UNIX_HASH_SIZE; (*i)++) {
		if (!hlist_empty(&unix_socket_table[*i]))
			return __sk_head(&unix_socket_table[*i]);
	}
	return NULL;
}

static struct sock *next_unix_socket(int *i, struct sock *s)
{
	struct sock *next = sk_next(s);
	/* More in this chain? */
	if (next)
		return next;
	/* Look for next non-empty chain. */
	for ((*i)++; *i <= UNIX_HASH_SIZE; (*i)++) {
		if (!hlist_empty(&unix_socket_table[*i]))
			return __sk_head(&unix_socket_table[*i]);
	}
	return NULL;
}

2224
struct unix_iter_state {
2225
	struct seq_net_private p;
2226 2227
	int i;
};
2228

2229
static struct sock *unix_seq_idx(struct seq_file *seq, loff_t pos)
L
Linus Torvalds 已提交
2230
{
2231
	struct unix_iter_state *iter = seq->private;
L
Linus Torvalds 已提交
2232 2233 2234
	loff_t off = 0;
	struct sock *s;

2235
	for (s = first_unix_socket(&iter->i); s; s = next_unix_socket(&iter->i, s)) {
2236
		if (sock_net(s) != seq_file_net(seq))
2237
			continue;
2238
		if (off == pos)
L
Linus Torvalds 已提交
2239 2240 2241 2242 2243 2244 2245
			return s;
		++off;
	}
	return NULL;
}

static void *unix_seq_start(struct seq_file *seq, loff_t *pos)
2246
	__acquires(unix_table_lock)
L
Linus Torvalds 已提交
2247
{
2248
	spin_lock(&unix_table_lock);
J
Joe Perches 已提交
2249
	return *pos ? unix_seq_idx(seq, *pos - 1) : SEQ_START_TOKEN;
L
Linus Torvalds 已提交
2250 2251 2252 2253
}

static void *unix_seq_next(struct seq_file *seq, void *v, loff_t *pos)
{
2254 2255
	struct unix_iter_state *iter = seq->private;
	struct sock *sk = v;
L
Linus Torvalds 已提交
2256 2257
	++*pos;

J
Joe Perches 已提交
2258
	if (v == SEQ_START_TOKEN)
2259 2260 2261
		sk = first_unix_socket(&iter->i);
	else
		sk = next_unix_socket(&iter->i, sk);
2262
	while (sk && (sock_net(sk) != seq_file_net(seq)))
2263 2264
		sk = next_unix_socket(&iter->i, sk);
	return sk;
L
Linus Torvalds 已提交
2265 2266 2267
}

static void unix_seq_stop(struct seq_file *seq, void *v)
2268
	__releases(unix_table_lock)
L
Linus Torvalds 已提交
2269
{
2270
	spin_unlock(&unix_table_lock);
L
Linus Torvalds 已提交
2271 2272 2273 2274
}

static int unix_seq_show(struct seq_file *seq, void *v)
{
2275

J
Joe Perches 已提交
2276
	if (v == SEQ_START_TOKEN)
L
Linus Torvalds 已提交
2277 2278 2279 2280 2281
		seq_puts(seq, "Num       RefCount Protocol Flags    Type St "
			 "Inode Path\n");
	else {
		struct sock *s = v;
		struct unix_sock *u = unix_sk(s);
2282
		unix_state_lock(s);
L
Linus Torvalds 已提交
2283

D
Dan Rosenberg 已提交
2284
		seq_printf(seq, "%pK: %08X %08X %08X %04X %02X %5lu",
L
Linus Torvalds 已提交
2285 2286 2287 2288 2289 2290 2291 2292 2293 2294 2295 2296 2297 2298 2299 2300 2301 2302 2303 2304 2305 2306 2307 2308 2309
			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]);
		}
2310
		unix_state_unlock(s);
L
Linus Torvalds 已提交
2311 2312 2313 2314 2315 2316
		seq_putc(seq, '\n');
	}

	return 0;
}

2317
static const struct seq_operations unix_seq_ops = {
L
Linus Torvalds 已提交
2318 2319 2320 2321 2322 2323 2324 2325
	.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)
{
2326 2327
	return seq_open_net(inode, file, &unix_seq_ops,
			    sizeof(struct unix_iter_state));
L
Linus Torvalds 已提交
2328 2329
}

2330
static const struct file_operations unix_seq_fops = {
L
Linus Torvalds 已提交
2331 2332 2333 2334
	.owner		= THIS_MODULE,
	.open		= unix_seq_open,
	.read		= seq_read,
	.llseek		= seq_lseek,
2335
	.release	= seq_release_net,
L
Linus Torvalds 已提交
2336 2337 2338 2339
};

#endif

2340
static const struct net_proto_family unix_family_ops = {
L
Linus Torvalds 已提交
2341 2342 2343 2344 2345
	.family = PF_UNIX,
	.create = unix_create,
	.owner	= THIS_MODULE,
};

2346

2347
static int __net_init unix_net_init(struct net *net)
2348 2349 2350
{
	int error = -ENOMEM;

2351
	net->unx.sysctl_max_dgram_qlen = 10;
2352 2353
	if (unix_sysctl_register(net))
		goto out;
2354

2355
#ifdef CONFIG_PROC_FS
2356 2357
	if (!proc_net_fops_create(net, "unix", 0, &unix_seq_fops)) {
		unix_sysctl_unregister(net);
2358
		goto out;
2359
	}
2360 2361 2362
#endif
	error = 0;
out:
2363
	return error;
2364 2365
}

2366
static void __net_exit unix_net_exit(struct net *net)
2367
{
2368
	unix_sysctl_unregister(net);
2369 2370 2371 2372 2373 2374 2375 2376
	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
2377 2378 2379 2380 2381
static int __init af_unix_init(void)
{
	int rc = -1;
	struct sk_buff *dummy_skb;

2382
	BUILD_BUG_ON(sizeof(struct unix_skb_parms) > sizeof(dummy_skb->cb));
L
Linus Torvalds 已提交
2383 2384

	rc = proto_register(&unix_proto, 1);
2385 2386
	if (rc != 0) {
		printk(KERN_CRIT "%s: Cannot create unix_sock SLAB cache!\n",
2387
		       __func__);
L
Linus Torvalds 已提交
2388 2389 2390 2391
		goto out;
	}

	sock_register(&unix_family_ops);
2392
	register_pernet_subsys(&unix_net_ops);
L
Linus Torvalds 已提交
2393 2394 2395 2396 2397 2398 2399 2400
out:
	return rc;
}

static void __exit af_unix_exit(void)
{
	sock_unregister(PF_UNIX);
	proto_unregister(&unix_proto);
2401
	unregister_pernet_subsys(&unix_net_ops);
L
Linus Torvalds 已提交
2402 2403
}

2404 2405 2406 2407 2408
/* 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
2409 2410 2411 2412
module_exit(af_unix_exit);

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