socket.c 57.0 KB
Newer Older
L
Linus Torvalds 已提交
1 2 3 4 5 6
/*
 * NET		An implementation of the SOCKET network access protocol.
 *
 * Version:	@(#)socket.c	1.1.93	18/02/95
 *
 * Authors:	Orest Zborowski, <obz@Kodak.COM>
7
 *		Ross Biro
L
Linus Torvalds 已提交
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
 *		Fred N. van Kempen, <waltje@uWalt.NL.Mugnet.ORG>
 *
 * Fixes:
 *		Anonymous	:	NOTSOCK/BADF cleanup. Error fix in
 *					shutdown()
 *		Alan Cox	:	verify_area() fixes
 *		Alan Cox	:	Removed DDI
 *		Jonathan Kamens	:	SOCK_DGRAM reconnect bug
 *		Alan Cox	:	Moved a load of checks to the very
 *					top level.
 *		Alan Cox	:	Move address structures to/from user
 *					mode above the protocol layers.
 *		Rob Janssen	:	Allow 0 length sends.
 *		Alan Cox	:	Asynchronous I/O support (cribbed from the
 *					tty drivers).
 *		Niibe Yutaka	:	Asynchronous I/O for writes (4.4BSD style)
 *		Jeff Uphoff	:	Made max number of sockets command-line
 *					configurable.
 *		Matti Aarnio	:	Made the number of sockets dynamic,
 *					to be allocated when needed, and mr.
 *					Uphoff's max is used as max to be
 *					allowed to allocate.
 *		Linus		:	Argh. removed all the socket allocation
 *					altogether: it's in the inode now.
 *		Alan Cox	:	Made sock_alloc()/sock_release() public
 *					for NetROM and future kernel nfsd type
 *					stuff.
 *		Alan Cox	:	sendmsg/recvmsg basics.
 *		Tom Dyas	:	Export net symbols.
 *		Marcin Dalecki	:	Fixed problems with CONFIG_NET="n".
 *		Alan Cox	:	Added thread locking to sys_* calls
 *					for sockets. May have errors at the
 *					moment.
 *		Kevin Buhr	:	Fixed the dumb errors in the above.
 *		Andi Kleen	:	Some small cleanups, optimizations,
 *					and fixed a copy_from_user() bug.
 *		Tigran Aivazian	:	sys_send(args) calls sys_sendto(args, NULL, 0)
45
 *		Tigran Aivazian	:	Made listen(2) backlog sanity checks
L
Linus Torvalds 已提交
46 47 48 49 50 51 52 53 54 55
 *					protocol-independent
 *
 *
 *		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.
 *
 *
 *	This module is effectively the top level interface to the BSD socket
56
 *	paradigm.
L
Linus Torvalds 已提交
57 58 59 60 61 62 63 64 65
 *
 *	Based upon Swansea University Computer Society NET3.039
 */

#include <linux/mm.h>
#include <linux/socket.h>
#include <linux/file.h>
#include <linux/net.h>
#include <linux/interrupt.h>
U
Ulrich Drepper 已提交
66
#include <linux/thread_info.h>
67
#include <linux/rcupdate.h>
L
Linus Torvalds 已提交
68 69 70
#include <linux/netdevice.h>
#include <linux/proc_fs.h>
#include <linux/seq_file.h>
A
Arjan van de Ven 已提交
71
#include <linux/mutex.h>
L
Linus Torvalds 已提交
72 73
#include <linux/wanrouter.h>
#include <linux/if_bridge.h>
74 75
#include <linux/if_frad.h>
#include <linux/if_vlan.h>
L
Linus Torvalds 已提交
76 77 78 79 80 81 82 83 84 85
#include <linux/init.h>
#include <linux/poll.h>
#include <linux/cache.h>
#include <linux/module.h>
#include <linux/highmem.h>
#include <linux/mount.h>
#include <linux/security.h>
#include <linux/syscalls.h>
#include <linux/compat.h>
#include <linux/kmod.h>
86
#include <linux/audit.h>
87
#include <linux/wireless.h>
88
#include <linux/nsproxy.h>
L
Linus Torvalds 已提交
89 90 91 92 93

#include <asm/uaccess.h>
#include <asm/unistd.h>

#include <net/compat.h>
94
#include <net/wext.h>
L
Linus Torvalds 已提交
95 96 97 98 99

#include <net/sock.h>
#include <linux/netfilter.h>

static int sock_no_open(struct inode *irrelevant, struct file *dontcare);
100 101 102 103
static ssize_t sock_aio_read(struct kiocb *iocb, const struct iovec *iov,
			 unsigned long nr_segs, loff_t pos);
static ssize_t sock_aio_write(struct kiocb *iocb, const struct iovec *iov,
			  unsigned long nr_segs, loff_t pos);
104
static int sock_mmap(struct file *file, struct vm_area_struct *vma);
L
Linus Torvalds 已提交
105 106 107 108

static int sock_close(struct inode *inode, struct file *file);
static unsigned int sock_poll(struct file *file,
			      struct poll_table_struct *wait);
109
static long sock_ioctl(struct file *file, unsigned int cmd, unsigned long arg);
110 111
#ifdef CONFIG_COMPAT
static long compat_sock_ioctl(struct file *file,
112
			      unsigned int cmd, unsigned long arg);
113
#endif
L
Linus Torvalds 已提交
114 115 116
static int sock_fasync(int fd, struct file *filp, int on);
static ssize_t sock_sendpage(struct file *file, struct page *page,
			     int offset, size_t size, loff_t *ppos, int more);
J
Jens Axboe 已提交
117 118 119
static ssize_t sock_splice_read(struct file *file, loff_t *ppos,
			        struct pipe_inode_info *pipe, size_t len,
				unsigned int flags);
L
Linus Torvalds 已提交
120 121 122 123 124 125

/*
 *	Socket files have a set of 'special' operations as well as the generic file ones. These don't appear
 *	in the operation structures but are done directly via the socketcall() multiplexor.
 */

126
static const struct file_operations socket_file_ops = {
L
Linus Torvalds 已提交
127 128 129 130 131 132
	.owner =	THIS_MODULE,
	.llseek =	no_llseek,
	.aio_read =	sock_aio_read,
	.aio_write =	sock_aio_write,
	.poll =		sock_poll,
	.unlocked_ioctl = sock_ioctl,
133 134 135
#ifdef CONFIG_COMPAT
	.compat_ioctl = compat_sock_ioctl,
#endif
L
Linus Torvalds 已提交
136 137 138 139
	.mmap =		sock_mmap,
	.open =		sock_no_open,	/* special open code to disallow open via /proc */
	.release =	sock_close,
	.fasync =	sock_fasync,
140 141
	.sendpage =	sock_sendpage,
	.splice_write = generic_splice_sendpage,
J
Jens Axboe 已提交
142
	.splice_read =	sock_splice_read,
L
Linus Torvalds 已提交
143 144 145 146 147 148 149
};

/*
 *	The protocol list. Each protocol is registered in here.
 */

static DEFINE_SPINLOCK(net_family_lock);
150
static const struct net_proto_family *net_families[NPROTO] __read_mostly;
L
Linus Torvalds 已提交
151 152 153 154 155 156 157 158

/*
 *	Statistics counters of the socket lists
 */

static DEFINE_PER_CPU(int, sockets_in_use) = 0;

/*
159 160 161
 * Support routines.
 * Move socket addresses back and forth across the kernel/user
 * divide and look after the messy bits.
L
Linus Torvalds 已提交
162 163
 */

164
#define MAX_SOCK_ADDR	128		/* 108 for Unix domain -
L
Linus Torvalds 已提交
165 166
					   16 for IP, 16 for IPX,
					   24 for IPv6,
167
					   about 80 for AX.25
L
Linus Torvalds 已提交
168 169
					   must be at least one bigger than
					   the AF_UNIX size (see net/unix/af_unix.c
170
					   :unix_mkname()).
L
Linus Torvalds 已提交
171
					 */
172

L
Linus Torvalds 已提交
173 174 175 176 177 178 179 180 181 182 183
/**
 *	move_addr_to_kernel	-	copy a socket address into kernel space
 *	@uaddr: Address in user space
 *	@kaddr: Address in kernel space
 *	@ulen: Length in user space
 *
 *	The address is copied into kernel space. If the provided address is
 *	too long an error code of -EINVAL is returned. If the copy gives
 *	invalid addresses -EFAULT is returned. On a success 0 is returned.
 */

184
int move_addr_to_kernel(void __user *uaddr, int ulen, struct sockaddr *kaddr)
L
Linus Torvalds 已提交
185
{
186
	if (ulen < 0 || ulen > sizeof(struct sockaddr_storage))
L
Linus Torvalds 已提交
187
		return -EINVAL;
188
	if (ulen == 0)
L
Linus Torvalds 已提交
189
		return 0;
190
	if (copy_from_user(kaddr, uaddr, ulen))
L
Linus Torvalds 已提交
191
		return -EFAULT;
192
	return audit_sockaddr(ulen, kaddr);
L
Linus Torvalds 已提交
193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210
}

/**
 *	move_addr_to_user	-	copy an address to user space
 *	@kaddr: kernel space address
 *	@klen: length of address in kernel
 *	@uaddr: user space address
 *	@ulen: pointer to user length field
 *
 *	The value pointed to by ulen on entry is the buffer length available.
 *	This is overwritten with the buffer space used. -EINVAL is returned
 *	if an overlong buffer is specified or a negative buffer size. -EFAULT
 *	is returned if either the buffer or the length field are not
 *	accessible.
 *	After copying the data up to the limit the user specifies, the true
 *	length of the data is written over the length limit the user
 *	specified. Zero is returned for a success.
 */
211

212
int move_addr_to_user(struct sockaddr *kaddr, int klen, void __user *uaddr,
213
		      int __user *ulen)
L
Linus Torvalds 已提交
214 215 216 217
{
	int err;
	int len;

218 219
	err = get_user(len, ulen);
	if (err)
L
Linus Torvalds 已提交
220
		return err;
221 222
	if (len > klen)
		len = klen;
223
	if (len < 0 || len > sizeof(struct sockaddr_storage))
L
Linus Torvalds 已提交
224
		return -EINVAL;
225
	if (len) {
S
Steve Grubb 已提交
226 227
		if (audit_sockaddr(klen, kaddr))
			return -ENOMEM;
228
		if (copy_to_user(uaddr, kaddr, len))
L
Linus Torvalds 已提交
229 230 231
			return -EFAULT;
	}
	/*
232 233
	 *      "fromlen shall refer to the value before truncation.."
	 *                      1003.1g
L
Linus Torvalds 已提交
234 235 236 237 238 239
	 */
	return __put_user(klen, ulen);
}

#define SOCKFS_MAGIC 0x534F434B

240
static struct kmem_cache *sock_inode_cachep __read_mostly;
L
Linus Torvalds 已提交
241 242 243 244

static struct inode *sock_alloc_inode(struct super_block *sb)
{
	struct socket_alloc *ei;
245

246
	ei = kmem_cache_alloc(sock_inode_cachep, GFP_KERNEL);
L
Linus Torvalds 已提交
247 248 249
	if (!ei)
		return NULL;
	init_waitqueue_head(&ei->socket.wait);
250

L
Linus Torvalds 已提交
251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266
	ei->socket.fasync_list = NULL;
	ei->socket.state = SS_UNCONNECTED;
	ei->socket.flags = 0;
	ei->socket.ops = NULL;
	ei->socket.sk = NULL;
	ei->socket.file = NULL;

	return &ei->vfs_inode;
}

static void sock_destroy_inode(struct inode *inode)
{
	kmem_cache_free(sock_inode_cachep,
			container_of(inode, struct socket_alloc, vfs_inode));
}

267
static void init_once(void *foo)
L
Linus Torvalds 已提交
268
{
269
	struct socket_alloc *ei = (struct socket_alloc *)foo;
L
Linus Torvalds 已提交
270

C
Christoph Lameter 已提交
271
	inode_init_once(&ei->vfs_inode);
L
Linus Torvalds 已提交
272
}
273

L
Linus Torvalds 已提交
274 275 276
static int init_inodecache(void)
{
	sock_inode_cachep = kmem_cache_create("sock_inode_cache",
277 278 279 280 281
					      sizeof(struct socket_alloc),
					      0,
					      (SLAB_HWCACHE_ALIGN |
					       SLAB_RECLAIM_ACCOUNT |
					       SLAB_MEM_SPREAD),
282
					      init_once);
L
Linus Torvalds 已提交
283 284 285 286 287 288 289 290 291 292 293
	if (sock_inode_cachep == NULL)
		return -ENOMEM;
	return 0;
}

static struct super_operations sockfs_ops = {
	.alloc_inode =	sock_alloc_inode,
	.destroy_inode =sock_destroy_inode,
	.statfs =	simple_statfs,
};

294
static int sockfs_get_sb(struct file_system_type *fs_type,
295 296
			 int flags, const char *dev_name, void *data,
			 struct vfsmount *mnt)
L
Linus Torvalds 已提交
297
{
298 299
	return get_sb_pseudo(fs_type, "socket:", &sockfs_ops, SOCKFS_MAGIC,
			     mnt);
L
Linus Torvalds 已提交
300 301
}

302
static struct vfsmount *sock_mnt __read_mostly;
L
Linus Torvalds 已提交
303 304 305 306 307 308

static struct file_system_type sock_fs_type = {
	.name =		"sockfs",
	.get_sb =	sockfs_get_sb,
	.kill_sb =	kill_anon_super,
};
309

L
Linus Torvalds 已提交
310 311
static int sockfs_delete_dentry(struct dentry *dentry)
{
312 313 314 315 316 317 318 319
	/*
	 * At creation time, we pretended this dentry was hashed
	 * (by clearing DCACHE_UNHASHED bit in d_flags)
	 * At delete time, we restore the truth : not hashed.
	 * (so that dput() can proceed correctly)
	 */
	dentry->d_flags |= DCACHE_UNHASHED;
	return 0;
L
Linus Torvalds 已提交
320
}
321 322 323 324 325 326 327 328 329 330

/*
 * sockfs_dname() is called from d_path().
 */
static char *sockfs_dname(struct dentry *dentry, char *buffer, int buflen)
{
	return dynamic_dname(dentry, buffer, buflen, "socket:[%lu]",
				dentry->d_inode->i_ino);
}

L
Linus Torvalds 已提交
331
static struct dentry_operations sockfs_dentry_operations = {
332
	.d_delete = sockfs_delete_dentry,
333
	.d_dname  = sockfs_dname,
L
Linus Torvalds 已提交
334 335 336 337 338
};

/*
 *	Obtains the first available file descriptor and sets it up for use.
 *
339 340
 *	These functions create file structures and maps them to fd space
 *	of the current process. On success it returns file descriptor
L
Linus Torvalds 已提交
341 342 343 344 345 346 347 348 349 350 351 352
 *	and file struct implicitly stored in sock->file.
 *	Note that another thread may close file descriptor before we return
 *	from this function. We use the fact that now we do not refer
 *	to socket after mapping. If one day we will need it, this
 *	function will increment ref. count on file by 1.
 *
 *	In any case returned fd MAY BE not valid!
 *	This race condition is unavoidable
 *	with shared fd spaces, we cannot solve it inside kernel,
 *	but we take care of internal coherence yet.
 */

353
static int sock_alloc_fd(struct file **filep, int flags)
L
Linus Torvalds 已提交
354 355 356
{
	int fd;

357
	fd = get_unused_fd_flags(flags);
358
	if (likely(fd >= 0)) {
L
Linus Torvalds 已提交
359 360
		struct file *file = get_empty_filp();

361 362
		*filep = file;
		if (unlikely(!file)) {
L
Linus Torvalds 已提交
363
			put_unused_fd(fd);
364
			return -ENFILE;
L
Linus Torvalds 已提交
365
		}
366 367 368 369
	} else
		*filep = NULL;
	return fd;
}
L
Linus Torvalds 已提交
370

371
static int sock_attach_fd(struct socket *sock, struct file *file, int flags)
372
{
373
	struct dentry *dentry;
374
	struct qstr name = { .name = "" };
375

376 377
	dentry = d_alloc(sock_mnt->mnt_sb->s_root, &name);
	if (unlikely(!dentry))
378 379
		return -ENOMEM;

380
	dentry->d_op = &sockfs_dentry_operations;
381 382 383 384 385
	/*
	 * We dont want to push this dentry into global dentry hash table.
	 * We pretend dentry is already hashed, by unsetting DCACHE_UNHASHED
	 * This permits a working /proc/$pid/fd/XXX on sockets
	 */
386 387
	dentry->d_flags &= ~DCACHE_UNHASHED;
	d_instantiate(dentry, SOCK_INODE(sock));
388 389

	sock->file = file;
390 391 392
	init_file(file, sock_mnt, dentry, FMODE_READ | FMODE_WRITE,
		  &socket_file_ops);
	SOCK_INODE(sock)->i_fop = &socket_file_ops;
393
	file->f_flags = O_RDWR | (flags & O_NONBLOCK);
394 395
	file->f_pos = 0;
	file->private_data = sock;
L
Linus Torvalds 已提交
396

397 398 399
	return 0;
}

400
int sock_map_fd(struct socket *sock, int flags)
401 402
{
	struct file *newfile;
403
	int fd = sock_alloc_fd(&newfile, flags);
404 405

	if (likely(fd >= 0)) {
406
		int err = sock_attach_fd(sock, newfile, flags);
407 408 409

		if (unlikely(err < 0)) {
			put_filp(newfile);
L
Linus Torvalds 已提交
410
			put_unused_fd(fd);
411
			return err;
L
Linus Torvalds 已提交
412
		}
413
		fd_install(fd, newfile);
L
Linus Torvalds 已提交
414 415 416 417
	}
	return fd;
}

418 419 420 421 422
static struct socket *sock_from_file(struct file *file, int *err)
{
	if (file->f_op == &socket_file_ops)
		return file->private_data;	/* set in sock_map_fd */

E
Eric Dumazet 已提交
423 424
	*err = -ENOTSOCK;
	return NULL;
425 426
}

L
Linus Torvalds 已提交
427 428 429 430 431 432 433 434 435 436 437 438 439 440 441 442 443 444
/**
 *	sockfd_lookup	- 	Go from a file number to its socket slot
 *	@fd: file handle
 *	@err: pointer to an error code return
 *
 *	The file handle passed in is locked and the socket it is bound
 *	too is returned. If an error occurs the err pointer is overwritten
 *	with a negative errno code and NULL is returned. The function checks
 *	for both invalid handles and passing a handle which is not a socket.
 *
 *	On a success the socket object pointer is returned.
 */

struct socket *sockfd_lookup(int fd, int *err)
{
	struct file *file;
	struct socket *sock;

445 446
	file = fget(fd);
	if (!file) {
L
Linus Torvalds 已提交
447 448 449
		*err = -EBADF;
		return NULL;
	}
450

451 452
	sock = sock_from_file(file, err);
	if (!sock)
L
Linus Torvalds 已提交
453
		fput(file);
454 455
	return sock;
}
L
Linus Torvalds 已提交
456

457 458 459 460 461
static struct socket *sockfd_lookup_light(int fd, int *err, int *fput_needed)
{
	struct file *file;
	struct socket *sock;

462
	*err = -EBADF;
463 464 465 466 467 468
	file = fget_light(fd, fput_needed);
	if (file) {
		sock = sock_from_file(file, err);
		if (sock)
			return sock;
		fput_light(file, *fput_needed);
L
Linus Torvalds 已提交
469
	}
470
	return NULL;
L
Linus Torvalds 已提交
471 472 473 474
}

/**
 *	sock_alloc	-	allocate a socket
475
 *
L
Linus Torvalds 已提交
476 477 478 479 480 481 482
 *	Allocate a new inode and socket object. The two are bound together
 *	and initialised. The socket is then returned. If we are out of inodes
 *	NULL is returned.
 */

static struct socket *sock_alloc(void)
{
483 484
	struct inode *inode;
	struct socket *sock;
L
Linus Torvalds 已提交
485 486 487 488 489 490 491

	inode = new_inode(sock_mnt->mnt_sb);
	if (!inode)
		return NULL;

	sock = SOCKET_I(inode);

492
	inode->i_mode = S_IFSOCK | S_IRWXUGO;
493 494
	inode->i_uid = current_fsuid();
	inode->i_gid = current_fsgid();
L
Linus Torvalds 已提交
495 496 497 498 499 500 501 502 503 504 505

	get_cpu_var(sockets_in_use)++;
	put_cpu_var(sockets_in_use);
	return sock;
}

/*
 *	In theory you can't get an open on this inode, but /proc provides
 *	a back door. Remember to keep it shut otherwise you'll let the
 *	creepy crawlies in.
 */
506

L
Linus Torvalds 已提交
507 508 509 510 511
static int sock_no_open(struct inode *irrelevant, struct file *dontcare)
{
	return -ENXIO;
}

512
const struct file_operations bad_sock_fops = {
L
Linus Torvalds 已提交
513 514 515 516 517 518 519 520 521 522
	.owner = THIS_MODULE,
	.open = sock_no_open,
};

/**
 *	sock_release	-	close a socket
 *	@sock: socket to close
 *
 *	The socket is released from the protocol stack if it has a release
 *	callback, and the inode is then released if the socket is bound to
523
 *	an inode not a file.
L
Linus Torvalds 已提交
524
 */
525

L
Linus Torvalds 已提交
526 527 528 529 530 531 532 533 534 535 536 537 538 539 540 541 542 543 544
void sock_release(struct socket *sock)
{
	if (sock->ops) {
		struct module *owner = sock->ops->owner;

		sock->ops->release(sock);
		sock->ops = NULL;
		module_put(owner);
	}

	if (sock->fasync_list)
		printk(KERN_ERR "sock_release: fasync list not empty!\n");

	get_cpu_var(sockets_in_use)--;
	put_cpu_var(sockets_in_use);
	if (!sock->file) {
		iput(SOCK_INODE(sock));
		return;
	}
545
	sock->file = NULL;
L
Linus Torvalds 已提交
546 547
}

548
static inline int __sock_sendmsg(struct kiocb *iocb, struct socket *sock,
L
Linus Torvalds 已提交
549 550 551 552 553 554 555 556 557 558 559 560 561 562 563 564 565 566 567 568 569 570 571 572 573 574 575 576 577 578 579 580 581 582 583 584 585 586 587 588 589 590
				 struct msghdr *msg, size_t size)
{
	struct sock_iocb *si = kiocb_to_siocb(iocb);
	int err;

	si->sock = sock;
	si->scm = NULL;
	si->msg = msg;
	si->size = size;

	err = security_socket_sendmsg(sock, msg, size);
	if (err)
		return err;

	return sock->ops->sendmsg(iocb, sock, msg, size);
}

int sock_sendmsg(struct socket *sock, struct msghdr *msg, size_t size)
{
	struct kiocb iocb;
	struct sock_iocb siocb;
	int ret;

	init_sync_kiocb(&iocb, NULL);
	iocb.private = &siocb;
	ret = __sock_sendmsg(&iocb, sock, msg, size);
	if (-EIOCBQUEUED == ret)
		ret = wait_on_sync_kiocb(&iocb);
	return ret;
}

int kernel_sendmsg(struct socket *sock, struct msghdr *msg,
		   struct kvec *vec, size_t num, size_t size)
{
	mm_segment_t oldfs = get_fs();
	int result;

	set_fs(KERNEL_DS);
	/*
	 * the following is safe, since for compiler definitions of kvec and
	 * iovec are identical, yielding the same in-core layout and alignment
	 */
591
	msg->msg_iov = (struct iovec *)vec;
L
Linus Torvalds 已提交
592 593 594 595 596 597
	msg->msg_iovlen = num;
	result = sock_sendmsg(sock, msg, size);
	set_fs(oldfs);
	return result;
}

598 599 600 601 602 603 604 605 606 607 608 609 610 611 612 613 614 615 616 617 618 619 620 621 622 623 624 625 626
/*
 * called from sock_recv_timestamp() if sock_flag(sk, SOCK_RCVTSTAMP)
 */
void __sock_recv_timestamp(struct msghdr *msg, struct sock *sk,
	struct sk_buff *skb)
{
	ktime_t kt = skb->tstamp;

	if (!sock_flag(sk, SOCK_RCVTSTAMPNS)) {
		struct timeval tv;
		/* Race occurred between timestamp enabling and packet
		   receiving.  Fill in the current time for now. */
		if (kt.tv64 == 0)
			kt = ktime_get_real();
		skb->tstamp = kt;
		tv = ktime_to_timeval(kt);
		put_cmsg(msg, SOL_SOCKET, SCM_TIMESTAMP, sizeof(tv), &tv);
	} else {
		struct timespec ts;
		/* Race occurred between timestamp enabling and packet
		   receiving.  Fill in the current time for now. */
		if (kt.tv64 == 0)
			kt = ktime_get_real();
		skb->tstamp = kt;
		ts = ktime_to_timespec(kt);
		put_cmsg(msg, SOL_SOCKET, SCM_TIMESTAMPNS, sizeof(ts), &ts);
	}
}

627 628
EXPORT_SYMBOL_GPL(__sock_recv_timestamp);

629
static inline int __sock_recvmsg(struct kiocb *iocb, struct socket *sock,
L
Linus Torvalds 已提交
630 631 632 633 634 635 636 637 638 639 640 641 642 643 644 645 646 647
				 struct msghdr *msg, size_t size, int flags)
{
	int err;
	struct sock_iocb *si = kiocb_to_siocb(iocb);

	si->sock = sock;
	si->scm = NULL;
	si->msg = msg;
	si->size = size;
	si->flags = flags;

	err = security_socket_recvmsg(sock, msg, size, flags);
	if (err)
		return err;

	return sock->ops->recvmsg(iocb, sock, msg, size, flags);
}

648
int sock_recvmsg(struct socket *sock, struct msghdr *msg,
L
Linus Torvalds 已提交
649 650 651 652 653 654
		 size_t size, int flags)
{
	struct kiocb iocb;
	struct sock_iocb siocb;
	int ret;

655
	init_sync_kiocb(&iocb, NULL);
L
Linus Torvalds 已提交
656 657 658 659 660 661 662
	iocb.private = &siocb;
	ret = __sock_recvmsg(&iocb, sock, msg, size, flags);
	if (-EIOCBQUEUED == ret)
		ret = wait_on_sync_kiocb(&iocb);
	return ret;
}

663 664
int kernel_recvmsg(struct socket *sock, struct msghdr *msg,
		   struct kvec *vec, size_t num, size_t size, int flags)
L
Linus Torvalds 已提交
665 666 667 668 669 670 671 672 673
{
	mm_segment_t oldfs = get_fs();
	int result;

	set_fs(KERNEL_DS);
	/*
	 * the following is safe, since for compiler definitions of kvec and
	 * iovec are identical, yielding the same in-core layout and alignment
	 */
674
	msg->msg_iov = (struct iovec *)vec, msg->msg_iovlen = num;
L
Linus Torvalds 已提交
675 676 677 678 679 680 681 682 683 684
	result = sock_recvmsg(sock, msg, size, flags);
	set_fs(oldfs);
	return result;
}

static void sock_aio_dtor(struct kiocb *iocb)
{
	kfree(iocb->private);
}

685 686
static ssize_t sock_sendpage(struct file *file, struct page *page,
			     int offset, size_t size, loff_t *ppos, int more)
L
Linus Torvalds 已提交
687 688 689 690
{
	struct socket *sock;
	int flags;

691 692 693 694 695 696 697 698
	sock = file->private_data;

	flags = !(file->f_flags & O_NONBLOCK) ? 0 : MSG_DONTWAIT;
	if (more)
		flags |= MSG_MORE;

	return sock->ops->sendpage(sock, page, offset, size, flags);
}
L
Linus Torvalds 已提交
699

J
Jens Axboe 已提交
700 701 702 703 704 705
static ssize_t sock_splice_read(struct file *file, loff_t *ppos,
			        struct pipe_inode_info *pipe, size_t len,
				unsigned int flags)
{
	struct socket *sock = file->private_data;

706 707 708
	if (unlikely(!sock->ops->splice_read))
		return -EINVAL;

J
Jens Axboe 已提交
709 710 711
	return sock->ops->splice_read(sock, ppos, pipe, len, flags);
}

712
static struct sock_iocb *alloc_sock_iocb(struct kiocb *iocb,
713
					 struct sock_iocb *siocb)
714 715 716 717 718
{
	if (!is_sync_kiocb(iocb)) {
		siocb = kmalloc(sizeof(*siocb), GFP_KERNEL);
		if (!siocb)
			return NULL;
L
Linus Torvalds 已提交
719 720 721
		iocb->ki_dtor = sock_aio_dtor;
	}

722 723 724
	siocb->kiocb = iocb;
	iocb->private = siocb;
	return siocb;
L
Linus Torvalds 已提交
725 726
}

727
static ssize_t do_sock_read(struct msghdr *msg, struct kiocb *iocb,
728 729
		struct file *file, const struct iovec *iov,
		unsigned long nr_segs)
730 731 732 733
{
	struct socket *sock = file->private_data;
	size_t size = 0;
	int i;
L
Linus Torvalds 已提交
734

735 736
	for (i = 0; i < nr_segs; i++)
		size += iov[i].iov_len;
L
Linus Torvalds 已提交
737

738 739 740 741
	msg->msg_name = NULL;
	msg->msg_namelen = 0;
	msg->msg_control = NULL;
	msg->msg_controllen = 0;
742
	msg->msg_iov = (struct iovec *)iov;
743 744 745 746 747 748
	msg->msg_iovlen = nr_segs;
	msg->msg_flags = (file->f_flags & O_NONBLOCK) ? MSG_DONTWAIT : 0;

	return __sock_recvmsg(iocb, sock, msg, size, msg->msg_flags);
}

749 750
static ssize_t sock_aio_read(struct kiocb *iocb, const struct iovec *iov,
				unsigned long nr_segs, loff_t pos)
751 752 753
{
	struct sock_iocb siocb, *x;

L
Linus Torvalds 已提交
754 755
	if (pos != 0)
		return -ESPIPE;
756 757

	if (iocb->ki_left == 0)	/* Match SYS5 behaviour */
L
Linus Torvalds 已提交
758 759
		return 0;

760 761

	x = alloc_sock_iocb(iocb, &siocb);
762 763
	if (!x)
		return -ENOMEM;
764
	return do_sock_read(&x->async_msg, iocb, iocb->ki_filp, iov, nr_segs);
L
Linus Torvalds 已提交
765 766
}

767
static ssize_t do_sock_write(struct msghdr *msg, struct kiocb *iocb,
768 769
			struct file *file, const struct iovec *iov,
			unsigned long nr_segs)
L
Linus Torvalds 已提交
770
{
771 772 773
	struct socket *sock = file->private_data;
	size_t size = 0;
	int i;
L
Linus Torvalds 已提交
774

775 776
	for (i = 0; i < nr_segs; i++)
		size += iov[i].iov_len;
L
Linus Torvalds 已提交
777

778 779 780 781
	msg->msg_name = NULL;
	msg->msg_namelen = 0;
	msg->msg_control = NULL;
	msg->msg_controllen = 0;
782
	msg->msg_iov = (struct iovec *)iov;
783 784 785 786
	msg->msg_iovlen = nr_segs;
	msg->msg_flags = (file->f_flags & O_NONBLOCK) ? MSG_DONTWAIT : 0;
	if (sock->type == SOCK_SEQPACKET)
		msg->msg_flags |= MSG_EOR;
L
Linus Torvalds 已提交
787

788
	return __sock_sendmsg(iocb, sock, msg, size);
L
Linus Torvalds 已提交
789 790
}

791 792
static ssize_t sock_aio_write(struct kiocb *iocb, const struct iovec *iov,
			  unsigned long nr_segs, loff_t pos)
793 794
{
	struct sock_iocb siocb, *x;
L
Linus Torvalds 已提交
795

796 797
	if (pos != 0)
		return -ESPIPE;
798 799

	x = alloc_sock_iocb(iocb, &siocb);
800 801
	if (!x)
		return -ENOMEM;
L
Linus Torvalds 已提交
802

803
	return do_sock_write(&x->async_msg, iocb, iocb->ki_filp, iov, nr_segs);
L
Linus Torvalds 已提交
804 805 806 807 808 809 810
}

/*
 * Atomic setting of ioctl hooks to avoid race
 * with module unload.
 */

A
Arjan van de Ven 已提交
811
static DEFINE_MUTEX(br_ioctl_mutex);
812
static int (*br_ioctl_hook) (struct net *, unsigned int cmd, void __user *arg) = NULL;
L
Linus Torvalds 已提交
813

814
void brioctl_set(int (*hook) (struct net *, unsigned int, void __user *))
L
Linus Torvalds 已提交
815
{
A
Arjan van de Ven 已提交
816
	mutex_lock(&br_ioctl_mutex);
L
Linus Torvalds 已提交
817
	br_ioctl_hook = hook;
A
Arjan van de Ven 已提交
818
	mutex_unlock(&br_ioctl_mutex);
L
Linus Torvalds 已提交
819
}
820

L
Linus Torvalds 已提交
821 822
EXPORT_SYMBOL(brioctl_set);

A
Arjan van de Ven 已提交
823
static DEFINE_MUTEX(vlan_ioctl_mutex);
824
static int (*vlan_ioctl_hook) (struct net *, void __user *arg);
L
Linus Torvalds 已提交
825

826
void vlan_ioctl_set(int (*hook) (struct net *, void __user *))
L
Linus Torvalds 已提交
827
{
A
Arjan van de Ven 已提交
828
	mutex_lock(&vlan_ioctl_mutex);
L
Linus Torvalds 已提交
829
	vlan_ioctl_hook = hook;
A
Arjan van de Ven 已提交
830
	mutex_unlock(&vlan_ioctl_mutex);
L
Linus Torvalds 已提交
831
}
832

L
Linus Torvalds 已提交
833 834
EXPORT_SYMBOL(vlan_ioctl_set);

A
Arjan van de Ven 已提交
835
static DEFINE_MUTEX(dlci_ioctl_mutex);
836
static int (*dlci_ioctl_hook) (unsigned int, void __user *);
L
Linus Torvalds 已提交
837

838
void dlci_ioctl_set(int (*hook) (unsigned int, void __user *))
L
Linus Torvalds 已提交
839
{
A
Arjan van de Ven 已提交
840
	mutex_lock(&dlci_ioctl_mutex);
L
Linus Torvalds 已提交
841
	dlci_ioctl_hook = hook;
A
Arjan van de Ven 已提交
842
	mutex_unlock(&dlci_ioctl_mutex);
L
Linus Torvalds 已提交
843
}
844

L
Linus Torvalds 已提交
845 846 847 848 849 850 851 852 853 854
EXPORT_SYMBOL(dlci_ioctl_set);

/*
 *	With an ioctl, arg may well be a user mode pointer, but we don't know
 *	what to do with it - that's up to the protocol still.
 */

static long sock_ioctl(struct file *file, unsigned cmd, unsigned long arg)
{
	struct socket *sock;
855
	struct sock *sk;
L
Linus Torvalds 已提交
856 857
	void __user *argp = (void __user *)arg;
	int pid, err;
858
	struct net *net;
L
Linus Torvalds 已提交
859

860
	sock = file->private_data;
861
	sk = sock->sk;
862
	net = sock_net(sk);
L
Linus Torvalds 已提交
863
	if (cmd >= SIOCDEVPRIVATE && cmd <= (SIOCDEVPRIVATE + 15)) {
864
		err = dev_ioctl(net, cmd, argp);
L
Linus Torvalds 已提交
865
	} else
866
#ifdef CONFIG_WIRELESS_EXT
L
Linus Torvalds 已提交
867
	if (cmd >= SIOCIWFIRST && cmd <= SIOCIWLAST) {
868
		err = dev_ioctl(net, cmd, argp);
L
Linus Torvalds 已提交
869
	} else
870 871
#endif				/* CONFIG_WIRELESS_EXT */
		switch (cmd) {
L
Linus Torvalds 已提交
872 873 874 875 876 877 878 879 880
		case FIOSETOWN:
		case SIOCSPGRP:
			err = -EFAULT;
			if (get_user(pid, (int __user *)argp))
				break;
			err = f_setown(sock->file, pid, 1);
			break;
		case FIOGETOWN:
		case SIOCGPGRP:
881
			err = put_user(f_getown(sock->file),
882
				       (int __user *)argp);
L
Linus Torvalds 已提交
883 884 885 886 887 888 889 890 891
			break;
		case SIOCGIFBR:
		case SIOCSIFBR:
		case SIOCBRADDBR:
		case SIOCBRDELBR:
			err = -ENOPKG;
			if (!br_ioctl_hook)
				request_module("bridge");

A
Arjan van de Ven 已提交
892
			mutex_lock(&br_ioctl_mutex);
893
			if (br_ioctl_hook)
894
				err = br_ioctl_hook(net, cmd, argp);
A
Arjan van de Ven 已提交
895
			mutex_unlock(&br_ioctl_mutex);
L
Linus Torvalds 已提交
896 897 898 899 900 901 902
			break;
		case SIOCGIFVLAN:
		case SIOCSIFVLAN:
			err = -ENOPKG;
			if (!vlan_ioctl_hook)
				request_module("8021q");

A
Arjan van de Ven 已提交
903
			mutex_lock(&vlan_ioctl_mutex);
L
Linus Torvalds 已提交
904
			if (vlan_ioctl_hook)
905
				err = vlan_ioctl_hook(net, argp);
A
Arjan van de Ven 已提交
906
			mutex_unlock(&vlan_ioctl_mutex);
L
Linus Torvalds 已提交
907 908 909 910 911 912 913
			break;
		case SIOCADDDLCI:
		case SIOCDELDLCI:
			err = -ENOPKG;
			if (!dlci_ioctl_hook)
				request_module("dlci");

914 915
			mutex_lock(&dlci_ioctl_mutex);
			if (dlci_ioctl_hook)
L
Linus Torvalds 已提交
916
				err = dlci_ioctl_hook(cmd, argp);
917
			mutex_unlock(&dlci_ioctl_mutex);
L
Linus Torvalds 已提交
918 919 920
			break;
		default:
			err = sock->ops->ioctl(sock, cmd, arg);
921 922 923 924 925 926

			/*
			 * If this ioctl is unknown try to hand it down
			 * to the NIC driver.
			 */
			if (err == -ENOIOCTLCMD)
927
				err = dev_ioctl(net, cmd, argp);
L
Linus Torvalds 已提交
928
			break;
929
		}
L
Linus Torvalds 已提交
930 931 932 933 934 935 936
	return err;
}

int sock_create_lite(int family, int type, int protocol, struct socket **res)
{
	int err;
	struct socket *sock = NULL;
937

L
Linus Torvalds 已提交
938 939 940 941 942 943 944 945 946 947 948
	err = security_socket_create(family, type, protocol, 1);
	if (err)
		goto out;

	sock = sock_alloc();
	if (!sock) {
		err = -ENOMEM;
		goto out;
	}

	sock->type = type;
V
Venkat Yekkirala 已提交
949 950 951 952
	err = security_socket_post_create(sock, family, type, protocol, 1);
	if (err)
		goto out_release;

L
Linus Torvalds 已提交
953 954 955
out:
	*res = sock;
	return err;
V
Venkat Yekkirala 已提交
956 957 958 959
out_release:
	sock_release(sock);
	sock = NULL;
	goto out;
L
Linus Torvalds 已提交
960 961 962
}

/* No kernel lock held - perfect */
963
static unsigned int sock_poll(struct file *file, poll_table *wait)
L
Linus Torvalds 已提交
964 965 966 967
{
	struct socket *sock;

	/*
968
	 *      We can't return errors to poll, so it's either yes or no.
L
Linus Torvalds 已提交
969
	 */
970
	sock = file->private_data;
L
Linus Torvalds 已提交
971 972 973
	return sock->ops->poll(file, sock, wait);
}

974
static int sock_mmap(struct file *file, struct vm_area_struct *vma)
L
Linus Torvalds 已提交
975
{
976
	struct socket *sock = file->private_data;
L
Linus Torvalds 已提交
977 978 979 980

	return sock->ops->mmap(file, sock, vma);
}

981
static int sock_close(struct inode *inode, struct file *filp)
L
Linus Torvalds 已提交
982 983
{
	/*
984 985
	 *      It was possible the inode is NULL we were
	 *      closing an unfinished socket.
L
Linus Torvalds 已提交
986 987
	 */

988
	if (!inode) {
L
Linus Torvalds 已提交
989 990 991 992 993 994 995 996 997 998 999 1000 1001 1002 1003 1004 1005 1006 1007 1008 1009 1010 1011 1012
		printk(KERN_DEBUG "sock_close: NULL inode\n");
		return 0;
	}
	sock_release(SOCKET_I(inode));
	return 0;
}

/*
 *	Update the socket async list
 *
 *	Fasync_list locking strategy.
 *
 *	1. fasync_list is modified only under process context socket lock
 *	   i.e. under semaphore.
 *	2. fasync_list is used under read_lock(&sk->sk_callback_lock)
 *	   or under socket lock.
 *	3. fasync_list can be used from softirq context, so that
 *	   modification under socket lock have to be enhanced with
 *	   write_lock_bh(&sk->sk_callback_lock).
 *							--ANK (990710)
 */

static int sock_fasync(int fd, struct file *filp, int on)
{
1013
	struct fasync_struct *fa, *fna = NULL, **prev;
L
Linus Torvalds 已提交
1014 1015 1016
	struct socket *sock;
	struct sock *sk;

1017
	if (on) {
1018
		fna = kmalloc(sizeof(struct fasync_struct), GFP_KERNEL);
1019
		if (fna == NULL)
L
Linus Torvalds 已提交
1020 1021 1022
			return -ENOMEM;
	}

1023
	sock = filp->private_data;
L
Linus Torvalds 已提交
1024

1025 1026
	sk = sock->sk;
	if (sk == NULL) {
L
Linus Torvalds 已提交
1027 1028 1029 1030 1031 1032
		kfree(fna);
		return -EINVAL;
	}

	lock_sock(sk);

1033
	prev = &(sock->fasync_list);
L
Linus Torvalds 已提交
1034

1035 1036
	for (fa = *prev; fa != NULL; prev = &fa->fa_next, fa = *prev)
		if (fa->fa_file == filp)
L
Linus Torvalds 已提交
1037 1038
			break;

1039 1040
	if (on) {
		if (fa != NULL) {
L
Linus Torvalds 已提交
1041
			write_lock_bh(&sk->sk_callback_lock);
1042
			fa->fa_fd = fd;
L
Linus Torvalds 已提交
1043 1044 1045 1046 1047
			write_unlock_bh(&sk->sk_callback_lock);

			kfree(fna);
			goto out;
		}
1048 1049 1050 1051
		fna->fa_file = filp;
		fna->fa_fd = fd;
		fna->magic = FASYNC_MAGIC;
		fna->fa_next = sock->fasync_list;
L
Linus Torvalds 已提交
1052
		write_lock_bh(&sk->sk_callback_lock);
1053
		sock->fasync_list = fna;
L
Linus Torvalds 已提交
1054
		write_unlock_bh(&sk->sk_callback_lock);
1055 1056
	} else {
		if (fa != NULL) {
L
Linus Torvalds 已提交
1057
			write_lock_bh(&sk->sk_callback_lock);
1058
			*prev = fa->fa_next;
L
Linus Torvalds 已提交
1059 1060 1061 1062 1063 1064 1065 1066 1067 1068 1069 1070 1071 1072 1073 1074
			write_unlock_bh(&sk->sk_callback_lock);
			kfree(fa);
		}
	}

out:
	release_sock(sock->sk);
	return 0;
}

/* This function may be called only under socket lock or callback_lock */

int sock_wake_async(struct socket *sock, int how, int band)
{
	if (!sock || !sock->fasync_list)
		return -1;
1075
	switch (how) {
1076
	case SOCK_WAKE_WAITD:
L
Linus Torvalds 已提交
1077 1078 1079
		if (test_bit(SOCK_ASYNC_WAITDATA, &sock->flags))
			break;
		goto call_kill;
1080
	case SOCK_WAKE_SPACE:
L
Linus Torvalds 已提交
1081 1082 1083
		if (!test_and_clear_bit(SOCK_ASYNC_NOSPACE, &sock->flags))
			break;
		/* fall through */
1084
	case SOCK_WAKE_IO:
1085
call_kill:
L
Linus Torvalds 已提交
1086 1087
		__kill_fasync(sock->fasync_list, SIGIO, band);
		break;
1088
	case SOCK_WAKE_URG:
L
Linus Torvalds 已提交
1089 1090 1091 1092 1093
		__kill_fasync(sock->fasync_list, SIGURG, band);
	}
	return 0;
}

1094
static int __sock_create(struct net *net, int family, int type, int protocol,
1095
			 struct socket **res, int kern)
L
Linus Torvalds 已提交
1096 1097 1098
{
	int err;
	struct socket *sock;
1099
	const struct net_proto_family *pf;
L
Linus Torvalds 已提交
1100 1101

	/*
1102
	 *      Check protocol is in range
L
Linus Torvalds 已提交
1103 1104 1105 1106 1107 1108 1109 1110 1111 1112 1113 1114
	 */
	if (family < 0 || family >= NPROTO)
		return -EAFNOSUPPORT;
	if (type < 0 || type >= SOCK_MAX)
		return -EINVAL;

	/* Compatibility.

	   This uglymoron is moved from INET layer to here to avoid
	   deadlock in module load.
	 */
	if (family == PF_INET && type == SOCK_PACKET) {
1115
		static int warned;
L
Linus Torvalds 已提交
1116 1117
		if (!warned) {
			warned = 1;
1118 1119
			printk(KERN_INFO "%s uses obsolete (PF_INET,SOCK_PACKET)\n",
			       current->comm);
L
Linus Torvalds 已提交
1120 1121 1122 1123 1124 1125 1126
		}
		family = PF_PACKET;
	}

	err = security_socket_create(family, type, protocol, kern);
	if (err)
		return err;
1127

1128 1129 1130 1131 1132 1133 1134 1135 1136 1137 1138 1139 1140 1141 1142
	/*
	 *	Allocate the socket and allow the family to set things up. if
	 *	the protocol is 0, the family is instructed to select an appropriate
	 *	default.
	 */
	sock = sock_alloc();
	if (!sock) {
		if (net_ratelimit())
			printk(KERN_WARNING "socket: no more sockets\n");
		return -ENFILE;	/* Not exactly a match, but its the
				   closest posix thing */
	}

	sock->type = type;

1143
#ifdef CONFIG_MODULES
1144 1145 1146
	/* Attempt to load a protocol module if the find failed.
	 *
	 * 12/09/1996 Marcin: But! this makes REALLY only sense, if the user
L
Linus Torvalds 已提交
1147 1148 1149
	 * requested real, full-featured networking support upon configuration.
	 * Otherwise module support will break!
	 */
1150
	if (net_families[family] == NULL)
1151
		request_module("net-pf-%d", family);
L
Linus Torvalds 已提交
1152 1153
#endif

1154 1155 1156 1157 1158
	rcu_read_lock();
	pf = rcu_dereference(net_families[family]);
	err = -EAFNOSUPPORT;
	if (!pf)
		goto out_release;
L
Linus Torvalds 已提交
1159 1160 1161 1162 1163

	/*
	 * We will call the ->create function, that possibly is in a loadable
	 * module, so we have to bump that loadable module refcnt first.
	 */
1164
	if (!try_module_get(pf->owner))
L
Linus Torvalds 已提交
1165 1166
		goto out_release;

1167 1168 1169
	/* Now protected by module ref count */
	rcu_read_unlock();

1170
	err = pf->create(net, sock, protocol);
1171
	if (err < 0)
L
Linus Torvalds 已提交
1172
		goto out_module_put;
1173

L
Linus Torvalds 已提交
1174 1175 1176 1177
	/*
	 * Now to bump the refcnt of the [loadable] module that owns this
	 * socket at sock_release time we decrement its refcnt.
	 */
1178 1179 1180
	if (!try_module_get(sock->ops->owner))
		goto out_module_busy;

L
Linus Torvalds 已提交
1181 1182 1183 1184
	/*
	 * Now that we're done with the ->create function, the [loadable]
	 * module can have its refcnt decremented
	 */
1185
	module_put(pf->owner);
V
Venkat Yekkirala 已提交
1186 1187
	err = security_socket_post_create(sock, family, type, protocol, kern);
	if (err)
1188
		goto out_sock_release;
1189
	*res = sock;
L
Linus Torvalds 已提交
1190

1191 1192 1193 1194
	return 0;

out_module_busy:
	err = -EAFNOSUPPORT;
L
Linus Torvalds 已提交
1195
out_module_put:
1196 1197 1198
	sock->ops = NULL;
	module_put(pf->owner);
out_sock_release:
L
Linus Torvalds 已提交
1199
	sock_release(sock);
1200 1201 1202 1203 1204
	return err;

out_release:
	rcu_read_unlock();
	goto out_sock_release;
L
Linus Torvalds 已提交
1205 1206 1207 1208
}

int sock_create(int family, int type, int protocol, struct socket **res)
{
1209
	return __sock_create(current->nsproxy->net_ns, family, type, protocol, res, 0);
L
Linus Torvalds 已提交
1210 1211 1212 1213
}

int sock_create_kern(int family, int type, int protocol, struct socket **res)
{
1214
	return __sock_create(&init_net, family, type, protocol, res, 1);
L
Linus Torvalds 已提交
1215 1216
}

1217
SYSCALL_DEFINE3(socket, int, family, int, type, int, protocol)
L
Linus Torvalds 已提交
1218 1219 1220
{
	int retval;
	struct socket *sock;
1221 1222
	int flags;

1223 1224 1225 1226 1227 1228
	/* Check the SOCK_* constants for consistency.  */
	BUILD_BUG_ON(SOCK_CLOEXEC != O_CLOEXEC);
	BUILD_BUG_ON((SOCK_MAX | SOCK_TYPE_MASK) != SOCK_TYPE_MASK);
	BUILD_BUG_ON(SOCK_CLOEXEC & SOCK_TYPE_MASK);
	BUILD_BUG_ON(SOCK_NONBLOCK & SOCK_TYPE_MASK);

1229
	flags = type & ~SOCK_TYPE_MASK;
1230
	if (flags & ~(SOCK_CLOEXEC | SOCK_NONBLOCK))
1231 1232
		return -EINVAL;
	type &= SOCK_TYPE_MASK;
L
Linus Torvalds 已提交
1233

U
Ulrich Drepper 已提交
1234 1235 1236
	if (SOCK_NONBLOCK != O_NONBLOCK && (flags & SOCK_NONBLOCK))
		flags = (flags & ~SOCK_NONBLOCK) | O_NONBLOCK;

L
Linus Torvalds 已提交
1237 1238 1239 1240
	retval = sock_create(family, type, protocol, &sock);
	if (retval < 0)
		goto out;

1241
	retval = sock_map_fd(sock, flags & (O_CLOEXEC | O_NONBLOCK));
L
Linus Torvalds 已提交
1242 1243 1244 1245 1246 1247 1248 1249 1250 1251 1252 1253 1254 1255 1256 1257
	if (retval < 0)
		goto out_release;

out:
	/* It may be already another descriptor 8) Not kernel problem. */
	return retval;

out_release:
	sock_release(sock);
	return retval;
}

/*
 *	Create a pair of connected sockets.
 */

1258 1259
SYSCALL_DEFINE4(socketpair, int, family, int, type, int, protocol,
		int __user *, usockvec)
L
Linus Torvalds 已提交
1260 1261 1262
{
	struct socket *sock1, *sock2;
	int fd1, fd2, err;
A
Al Viro 已提交
1263
	struct file *newfile1, *newfile2;
1264 1265 1266
	int flags;

	flags = type & ~SOCK_TYPE_MASK;
1267
	if (flags & ~(SOCK_CLOEXEC | SOCK_NONBLOCK))
1268 1269
		return -EINVAL;
	type &= SOCK_TYPE_MASK;
L
Linus Torvalds 已提交
1270

U
Ulrich Drepper 已提交
1271 1272 1273
	if (SOCK_NONBLOCK != O_NONBLOCK && (flags & SOCK_NONBLOCK))
		flags = (flags & ~SOCK_NONBLOCK) | O_NONBLOCK;

L
Linus Torvalds 已提交
1274 1275 1276 1277 1278 1279 1280 1281 1282 1283 1284 1285 1286 1287
	/*
	 * Obtain the first socket and check if the underlying protocol
	 * supports the socketpair call.
	 */

	err = sock_create(family, type, protocol, &sock1);
	if (err < 0)
		goto out;

	err = sock_create(family, type, protocol, &sock2);
	if (err < 0)
		goto out_release_1;

	err = sock1->ops->socketpair(sock1, sock2);
1288
	if (err < 0)
L
Linus Torvalds 已提交
1289 1290
		goto out_release_both;

1291
	fd1 = sock_alloc_fd(&newfile1, flags & O_CLOEXEC);
1292 1293
	if (unlikely(fd1 < 0)) {
		err = fd1;
A
Al Viro 已提交
1294
		goto out_release_both;
1295
	}
L
Linus Torvalds 已提交
1296

1297
	fd2 = sock_alloc_fd(&newfile2, flags & O_CLOEXEC);
A
Al Viro 已提交
1298
	if (unlikely(fd2 < 0)) {
1299
		err = fd2;
A
Al Viro 已提交
1300 1301
		put_filp(newfile1);
		put_unused_fd(fd1);
L
Linus Torvalds 已提交
1302
		goto out_release_both;
A
Al Viro 已提交
1303
	}
L
Linus Torvalds 已提交
1304

1305
	err = sock_attach_fd(sock1, newfile1, flags & O_NONBLOCK);
A
Al Viro 已提交
1306 1307 1308 1309
	if (unlikely(err < 0)) {
		goto out_fd2;
	}

1310
	err = sock_attach_fd(sock2, newfile2, flags & O_NONBLOCK);
A
Al Viro 已提交
1311 1312 1313 1314 1315
	if (unlikely(err < 0)) {
		fput(newfile1);
		goto out_fd1;
	}

A
Al Viro 已提交
1316
	audit_fd_pair(fd1, fd2);
A
Al Viro 已提交
1317 1318
	fd_install(fd1, newfile1);
	fd_install(fd2, newfile2);
L
Linus Torvalds 已提交
1319 1320 1321 1322
	/* fd1 and fd2 may be already another descriptors.
	 * Not kernel problem.
	 */

1323
	err = put_user(fd1, &usockvec[0]);
L
Linus Torvalds 已提交
1324 1325 1326 1327 1328 1329 1330 1331 1332 1333
	if (!err)
		err = put_user(fd2, &usockvec[1]);
	if (!err)
		return 0;

	sys_close(fd2);
	sys_close(fd1);
	return err;

out_release_both:
1334
	sock_release(sock2);
L
Linus Torvalds 已提交
1335
out_release_1:
1336
	sock_release(sock1);
L
Linus Torvalds 已提交
1337 1338
out:
	return err;
A
Al Viro 已提交
1339 1340 1341 1342 1343 1344 1345 1346 1347 1348

out_fd2:
	put_filp(newfile1);
	sock_release(sock1);
out_fd1:
	put_filp(newfile2);
	sock_release(sock2);
	put_unused_fd(fd1);
	put_unused_fd(fd2);
	goto out;
L
Linus Torvalds 已提交
1349 1350 1351 1352 1353 1354 1355 1356 1357 1358
}

/*
 *	Bind a name to a socket. Nothing much to do here since it's
 *	the protocol's responsibility to handle the local address.
 *
 *	We move the socket address to kernel space before we call
 *	the protocol layer (having also checked the address is ok).
 */

1359
SYSCALL_DEFINE3(bind, int, fd, struct sockaddr __user *, umyaddr, int, addrlen)
L
Linus Torvalds 已提交
1360 1361
{
	struct socket *sock;
1362
	struct sockaddr_storage address;
1363
	int err, fput_needed;
L
Linus Torvalds 已提交
1364

1365
	sock = sockfd_lookup_light(fd, &err, &fput_needed);
1366
	if (sock) {
1367
		err = move_addr_to_kernel(umyaddr, addrlen, (struct sockaddr *)&address);
1368 1369
		if (err >= 0) {
			err = security_socket_bind(sock,
1370
						   (struct sockaddr *)&address,
1371
						   addrlen);
1372 1373
			if (!err)
				err = sock->ops->bind(sock,
1374
						      (struct sockaddr *)
1375
						      &address, addrlen);
L
Linus Torvalds 已提交
1376
		}
1377
		fput_light(sock->file, fput_needed);
1378
	}
L
Linus Torvalds 已提交
1379 1380 1381 1382 1383 1384 1385 1386 1387
	return err;
}

/*
 *	Perform a listen. Basically, we allow the protocol to do anything
 *	necessary for a listen, and if that works, we mark the socket as
 *	ready for listening.
 */

1388
SYSCALL_DEFINE2(listen, int, fd, int, backlog)
L
Linus Torvalds 已提交
1389 1390
{
	struct socket *sock;
1391
	int err, fput_needed;
1392
	int somaxconn;
1393 1394 1395

	sock = sockfd_lookup_light(fd, &err, &fput_needed);
	if (sock) {
1396
		somaxconn = sock_net(sock->sk)->core.sysctl_somaxconn;
1397 1398
		if ((unsigned)backlog > somaxconn)
			backlog = somaxconn;
L
Linus Torvalds 已提交
1399 1400

		err = security_socket_listen(sock, backlog);
1401 1402
		if (!err)
			err = sock->ops->listen(sock, backlog);
L
Linus Torvalds 已提交
1403

1404
		fput_light(sock->file, fput_needed);
L
Linus Torvalds 已提交
1405 1406 1407 1408 1409 1410 1411 1412 1413 1414 1415 1416 1417 1418 1419 1420
	}
	return err;
}

/*
 *	For accept, we attempt to create a new socket, set up the link
 *	with the client, wake up the client, then return the new
 *	connected fd. We collect the address of the connector in kernel
 *	space and move it to user at the very end. This is unclean because
 *	we open the socket then return an error.
 *
 *	1003.1g adds the ability to recvmsg() to query connection pending
 *	status to recvmsg. We need to add that support in a way thats
 *	clean when we restucture accept also.
 */

1421 1422
SYSCALL_DEFINE4(accept4, int, fd, struct sockaddr __user *, upeer_sockaddr,
		int __user *, upeer_addrlen, int, flags)
L
Linus Torvalds 已提交
1423 1424
{
	struct socket *sock, *newsock;
1425
	struct file *newfile;
1426
	int err, len, newfd, fput_needed;
1427
	struct sockaddr_storage address;
L
Linus Torvalds 已提交
1428

1429
	if (flags & ~(SOCK_CLOEXEC | SOCK_NONBLOCK))
U
Ulrich Drepper 已提交
1430 1431 1432 1433 1434
		return -EINVAL;

	if (SOCK_NONBLOCK != O_NONBLOCK && (flags & SOCK_NONBLOCK))
		flags = (flags & ~SOCK_NONBLOCK) | O_NONBLOCK;

1435
	sock = sockfd_lookup_light(fd, &err, &fput_needed);
L
Linus Torvalds 已提交
1436 1437 1438 1439
	if (!sock)
		goto out;

	err = -ENFILE;
1440
	if (!(newsock = sock_alloc()))
L
Linus Torvalds 已提交
1441 1442 1443 1444 1445 1446 1447 1448 1449 1450 1451
		goto out_put;

	newsock->type = sock->type;
	newsock->ops = sock->ops;

	/*
	 * We don't need try_module_get here, as the listening socket (sock)
	 * has the protocol module (sock->ops->owner) held.
	 */
	__module_get(newsock->ops->owner);

U
Ulrich Drepper 已提交
1452
	newfd = sock_alloc_fd(&newfile, flags & O_CLOEXEC);
1453 1454
	if (unlikely(newfd < 0)) {
		err = newfd;
1455 1456
		sock_release(newsock);
		goto out_put;
1457 1458
	}

1459
	err = sock_attach_fd(newsock, newfile, flags & O_NONBLOCK);
1460
	if (err < 0)
1461
		goto out_fd_simple;
1462

1463 1464
	err = security_socket_accept(sock, newsock);
	if (err)
1465
		goto out_fd;
1466

L
Linus Torvalds 已提交
1467 1468
	err = sock->ops->accept(sock, newsock, sock->file->f_flags);
	if (err < 0)
1469
		goto out_fd;
L
Linus Torvalds 已提交
1470 1471

	if (upeer_sockaddr) {
1472
		if (newsock->ops->getname(newsock, (struct sockaddr *)&address,
1473
					  &len, 2) < 0) {
L
Linus Torvalds 已提交
1474
			err = -ECONNABORTED;
1475
			goto out_fd;
L
Linus Torvalds 已提交
1476
		}
1477 1478
		err = move_addr_to_user((struct sockaddr *)&address,
					len, upeer_sockaddr, upeer_addrlen);
L
Linus Torvalds 已提交
1479
		if (err < 0)
1480
			goto out_fd;
L
Linus Torvalds 已提交
1481 1482 1483 1484
	}

	/* File flags are not inherited via accept() unlike another OSes. */

1485 1486
	fd_install(newfd, newfile);
	err = newfd;
L
Linus Torvalds 已提交
1487 1488 1489 1490

	security_socket_post_accept(sock, newsock);

out_put:
1491
	fput_light(sock->file, fput_needed);
L
Linus Torvalds 已提交
1492 1493
out:
	return err;
1494 1495 1496 1497 1498
out_fd_simple:
	sock_release(newsock);
	put_filp(newfile);
	put_unused_fd(newfd);
	goto out_put;
1499
out_fd:
1500
	fput(newfile);
1501
	put_unused_fd(newfd);
L
Linus Torvalds 已提交
1502 1503 1504
	goto out_put;
}

1505 1506
SYSCALL_DEFINE3(accept, int, fd, struct sockaddr __user *, upeer_sockaddr,
		int __user *, upeer_addrlen)
U
Ulrich Drepper 已提交
1507
{
U
Ulrich Drepper 已提交
1508
	return sys_accept4(fd, upeer_sockaddr, upeer_addrlen, 0);
U
Ulrich Drepper 已提交
1509 1510
}

L
Linus Torvalds 已提交
1511 1512 1513 1514 1515 1516 1517 1518 1519 1520 1521 1522
/*
 *	Attempt to connect to a socket with the server address.  The address
 *	is in user space so we verify it is OK and move it to kernel space.
 *
 *	For 1003.1g we need to add clean support for a bind to AF_UNSPEC to
 *	break bindings
 *
 *	NOTE: 1003.1g draft 6.3 is broken with respect to AX.25/NetROM and
 *	other SEQPACKET protocols that take time to connect() as it doesn't
 *	include the -EINPROGRESS status for such sockets.
 */

1523 1524
SYSCALL_DEFINE3(connect, int, fd, struct sockaddr __user *, uservaddr,
		int, addrlen)
L
Linus Torvalds 已提交
1525 1526
{
	struct socket *sock;
1527
	struct sockaddr_storage address;
1528
	int err, fput_needed;
L
Linus Torvalds 已提交
1529

1530
	sock = sockfd_lookup_light(fd, &err, &fput_needed);
L
Linus Torvalds 已提交
1531 1532
	if (!sock)
		goto out;
1533
	err = move_addr_to_kernel(uservaddr, addrlen, (struct sockaddr *)&address);
L
Linus Torvalds 已提交
1534 1535 1536
	if (err < 0)
		goto out_put;

1537
	err =
1538
	    security_socket_connect(sock, (struct sockaddr *)&address, addrlen);
L
Linus Torvalds 已提交
1539 1540 1541
	if (err)
		goto out_put;

1542
	err = sock->ops->connect(sock, (struct sockaddr *)&address, addrlen,
L
Linus Torvalds 已提交
1543 1544
				 sock->file->f_flags);
out_put:
1545
	fput_light(sock->file, fput_needed);
L
Linus Torvalds 已提交
1546 1547 1548 1549 1550 1551 1552 1553 1554
out:
	return err;
}

/*
 *	Get the local address ('name') of a socket object. Move the obtained
 *	name to user space.
 */

1555 1556
SYSCALL_DEFINE3(getsockname, int, fd, struct sockaddr __user *, usockaddr,
		int __user *, usockaddr_len)
L
Linus Torvalds 已提交
1557 1558
{
	struct socket *sock;
1559
	struct sockaddr_storage address;
1560
	int len, err, fput_needed;
1561

1562
	sock = sockfd_lookup_light(fd, &err, &fput_needed);
L
Linus Torvalds 已提交
1563 1564 1565 1566 1567 1568 1569
	if (!sock)
		goto out;

	err = security_socket_getsockname(sock);
	if (err)
		goto out_put;

1570
	err = sock->ops->getname(sock, (struct sockaddr *)&address, &len, 0);
L
Linus Torvalds 已提交
1571 1572
	if (err)
		goto out_put;
1573
	err = move_addr_to_user((struct sockaddr *)&address, len, usockaddr, usockaddr_len);
L
Linus Torvalds 已提交
1574 1575

out_put:
1576
	fput_light(sock->file, fput_needed);
L
Linus Torvalds 已提交
1577 1578 1579 1580 1581 1582 1583 1584 1585
out:
	return err;
}

/*
 *	Get the remote address ('name') of a socket object. Move the obtained
 *	name to user space.
 */

1586 1587
SYSCALL_DEFINE3(getpeername, int, fd, struct sockaddr __user *, usockaddr,
		int __user *, usockaddr_len)
L
Linus Torvalds 已提交
1588 1589
{
	struct socket *sock;
1590
	struct sockaddr_storage address;
1591
	int len, err, fput_needed;
L
Linus Torvalds 已提交
1592

1593 1594
	sock = sockfd_lookup_light(fd, &err, &fput_needed);
	if (sock != NULL) {
L
Linus Torvalds 已提交
1595 1596
		err = security_socket_getpeername(sock);
		if (err) {
1597
			fput_light(sock->file, fput_needed);
L
Linus Torvalds 已提交
1598 1599 1600
			return err;
		}

1601
		err =
1602
		    sock->ops->getname(sock, (struct sockaddr *)&address, &len,
1603
				       1);
L
Linus Torvalds 已提交
1604
		if (!err)
1605
			err = move_addr_to_user((struct sockaddr *)&address, len, usockaddr,
1606
						usockaddr_len);
1607
		fput_light(sock->file, fput_needed);
L
Linus Torvalds 已提交
1608 1609 1610 1611 1612 1613 1614 1615 1616 1617
	}
	return err;
}

/*
 *	Send a datagram to a given address. We move the address into kernel
 *	space and check the user space data area is readable before invoking
 *	the protocol.
 */

1618 1619 1620
SYSCALL_DEFINE6(sendto, int, fd, void __user *, buff, size_t, len,
		unsigned, flags, struct sockaddr __user *, addr,
		int, addr_len)
L
Linus Torvalds 已提交
1621 1622
{
	struct socket *sock;
1623
	struct sockaddr_storage address;
L
Linus Torvalds 已提交
1624 1625 1626
	int err;
	struct msghdr msg;
	struct iovec iov;
1627 1628
	int fput_needed;

1629 1630
	sock = sockfd_lookup_light(fd, &err, &fput_needed);
	if (!sock)
1631
		goto out;
1632

1633 1634 1635 1636 1637 1638 1639 1640
	iov.iov_base = buff;
	iov.iov_len = len;
	msg.msg_name = NULL;
	msg.msg_iov = &iov;
	msg.msg_iovlen = 1;
	msg.msg_control = NULL;
	msg.msg_controllen = 0;
	msg.msg_namelen = 0;
1641
	if (addr) {
1642
		err = move_addr_to_kernel(addr, addr_len, (struct sockaddr *)&address);
L
Linus Torvalds 已提交
1643 1644
		if (err < 0)
			goto out_put;
1645
		msg.msg_name = (struct sockaddr *)&address;
1646
		msg.msg_namelen = addr_len;
L
Linus Torvalds 已提交
1647 1648 1649 1650 1651 1652
	}
	if (sock->file->f_flags & O_NONBLOCK)
		flags |= MSG_DONTWAIT;
	msg.msg_flags = flags;
	err = sock_sendmsg(sock, &msg, len);

1653
out_put:
1654
	fput_light(sock->file, fput_needed);
1655
out:
L
Linus Torvalds 已提交
1656 1657 1658 1659
	return err;
}

/*
1660
 *	Send a datagram down a socket.
L
Linus Torvalds 已提交
1661 1662
 */

1663 1664
SYSCALL_DEFINE4(send, int, fd, void __user *, buff, size_t, len,
		unsigned, flags)
L
Linus Torvalds 已提交
1665 1666 1667 1668 1669
{
	return sys_sendto(fd, buff, len, flags, NULL, 0);
}

/*
1670
 *	Receive a frame from the socket and optionally record the address of the
L
Linus Torvalds 已提交
1671 1672 1673 1674
 *	sender. We verify the buffers are writable and if needed move the
 *	sender address from kernel to user space.
 */

1675 1676 1677
SYSCALL_DEFINE6(recvfrom, int, fd, void __user *, ubuf, size_t, size,
		unsigned, flags, struct sockaddr __user *, addr,
		int __user *, addr_len)
L
Linus Torvalds 已提交
1678 1679 1680 1681
{
	struct socket *sock;
	struct iovec iov;
	struct msghdr msg;
1682
	struct sockaddr_storage address;
1683
	int err, err2;
1684 1685
	int fput_needed;

1686
	sock = sockfd_lookup_light(fd, &err, &fput_needed);
L
Linus Torvalds 已提交
1687
	if (!sock)
1688
		goto out;
L
Linus Torvalds 已提交
1689

1690 1691 1692 1693 1694 1695
	msg.msg_control = NULL;
	msg.msg_controllen = 0;
	msg.msg_iovlen = 1;
	msg.msg_iov = &iov;
	iov.iov_len = size;
	iov.iov_base = ubuf;
1696 1697
	msg.msg_name = (struct sockaddr *)&address;
	msg.msg_namelen = sizeof(address);
L
Linus Torvalds 已提交
1698 1699
	if (sock->file->f_flags & O_NONBLOCK)
		flags |= MSG_DONTWAIT;
1700
	err = sock_recvmsg(sock, &msg, size, flags);
L
Linus Torvalds 已提交
1701

1702
	if (err >= 0 && addr != NULL) {
1703 1704
		err2 = move_addr_to_user((struct sockaddr *)&address,
					 msg.msg_namelen, addr, addr_len);
1705 1706
		if (err2 < 0)
			err = err2;
L
Linus Torvalds 已提交
1707
	}
1708 1709

	fput_light(sock->file, fput_needed);
1710
out:
L
Linus Torvalds 已提交
1711 1712 1713 1714
	return err;
}

/*
1715
 *	Receive a datagram from a socket.
L
Linus Torvalds 已提交
1716 1717
 */

1718 1719
asmlinkage long sys_recv(int fd, void __user *ubuf, size_t size,
			 unsigned flags)
L
Linus Torvalds 已提交
1720 1721 1722 1723 1724 1725 1726 1727 1728
{
	return sys_recvfrom(fd, ubuf, size, flags, NULL, NULL);
}

/*
 *	Set a socket option. Because we don't know the option lengths we have
 *	to pass the user mode parameter for the protocols to sort out.
 */

1729 1730
SYSCALL_DEFINE5(setsockopt, int, fd, int, level, int, optname,
		char __user *, optval, int, optlen)
L
Linus Torvalds 已提交
1731
{
1732
	int err, fput_needed;
L
Linus Torvalds 已提交
1733 1734 1735 1736
	struct socket *sock;

	if (optlen < 0)
		return -EINVAL;
1737 1738 1739 1740

	sock = sockfd_lookup_light(fd, &err, &fput_needed);
	if (sock != NULL) {
		err = security_socket_setsockopt(sock, level, optname);
1741 1742
		if (err)
			goto out_put;
L
Linus Torvalds 已提交
1743 1744

		if (level == SOL_SOCKET)
1745 1746 1747
			err =
			    sock_setsockopt(sock, level, optname, optval,
					    optlen);
L
Linus Torvalds 已提交
1748
		else
1749 1750 1751
			err =
			    sock->ops->setsockopt(sock, level, optname, optval,
						  optlen);
1752 1753
out_put:
		fput_light(sock->file, fput_needed);
L
Linus Torvalds 已提交
1754 1755 1756 1757 1758 1759 1760 1761 1762
	}
	return err;
}

/*
 *	Get a socket option. Because we don't know the option lengths we have
 *	to pass a user mode parameter for the protocols to sort out.
 */

1763 1764
SYSCALL_DEFINE5(getsockopt, int, fd, int, level, int, optname,
		char __user *, optval, int __user *, optlen)
L
Linus Torvalds 已提交
1765
{
1766
	int err, fput_needed;
L
Linus Torvalds 已提交
1767 1768
	struct socket *sock;

1769 1770
	sock = sockfd_lookup_light(fd, &err, &fput_needed);
	if (sock != NULL) {
1771 1772 1773
		err = security_socket_getsockopt(sock, level, optname);
		if (err)
			goto out_put;
L
Linus Torvalds 已提交
1774 1775

		if (level == SOL_SOCKET)
1776 1777 1778
			err =
			    sock_getsockopt(sock, level, optname, optval,
					    optlen);
L
Linus Torvalds 已提交
1779
		else
1780 1781 1782
			err =
			    sock->ops->getsockopt(sock, level, optname, optval,
						  optlen);
1783 1784
out_put:
		fput_light(sock->file, fput_needed);
L
Linus Torvalds 已提交
1785 1786 1787 1788 1789 1790 1791 1792
	}
	return err;
}

/*
 *	Shutdown a socket.
 */

1793
SYSCALL_DEFINE2(shutdown, int, fd, int, how)
L
Linus Torvalds 已提交
1794
{
1795
	int err, fput_needed;
L
Linus Torvalds 已提交
1796 1797
	struct socket *sock;

1798 1799
	sock = sockfd_lookup_light(fd, &err, &fput_needed);
	if (sock != NULL) {
L
Linus Torvalds 已提交
1800
		err = security_socket_shutdown(sock, how);
1801 1802 1803
		if (!err)
			err = sock->ops->shutdown(sock, how);
		fput_light(sock->file, fput_needed);
L
Linus Torvalds 已提交
1804 1805 1806 1807
	}
	return err;
}

1808
/* A couple of helpful macros for getting the address of the 32/64 bit
L
Linus Torvalds 已提交
1809 1810 1811 1812 1813 1814 1815 1816 1817 1818
 * fields which are the same type (int / unsigned) on our platforms.
 */
#define COMPAT_MSG(msg, member)	((MSG_CMSG_COMPAT & flags) ? &msg##_compat->member : &msg->member)
#define COMPAT_NAMELEN(msg)	COMPAT_MSG(msg, msg_namelen)
#define COMPAT_FLAGS(msg)	COMPAT_MSG(msg, msg_flags)

/*
 *	BSD sendmsg interface
 */

1819
SYSCALL_DEFINE3(sendmsg, int, fd, struct msghdr __user *, msg, unsigned, flags)
L
Linus Torvalds 已提交
1820
{
1821 1822
	struct compat_msghdr __user *msg_compat =
	    (struct compat_msghdr __user *)msg;
L
Linus Torvalds 已提交
1823
	struct socket *sock;
1824
	struct sockaddr_storage address;
L
Linus Torvalds 已提交
1825
	struct iovec iovstack[UIO_FASTIOV], *iov = iovstack;
1826
	unsigned char ctl[sizeof(struct cmsghdr) + 20]
1827 1828
	    __attribute__ ((aligned(sizeof(__kernel_size_t))));
	/* 20 is size of ipv6_pktinfo */
L
Linus Torvalds 已提交
1829 1830 1831
	unsigned char *ctl_buf = ctl;
	struct msghdr msg_sys;
	int err, ctl_len, iov_size, total_len;
1832
	int fput_needed;
1833

L
Linus Torvalds 已提交
1834 1835 1836 1837
	err = -EFAULT;
	if (MSG_CMSG_COMPAT & flags) {
		if (get_compat_msghdr(&msg_sys, msg_compat))
			return -EFAULT;
1838 1839
	}
	else if (copy_from_user(&msg_sys, msg, sizeof(struct msghdr)))
L
Linus Torvalds 已提交
1840 1841
		return -EFAULT;

1842
	sock = sockfd_lookup_light(fd, &err, &fput_needed);
1843
	if (!sock)
L
Linus Torvalds 已提交
1844 1845 1846 1847 1848 1849 1850
		goto out;

	/* do not move before msg_sys is valid */
	err = -EMSGSIZE;
	if (msg_sys.msg_iovlen > UIO_MAXIOV)
		goto out_put;

1851
	/* Check whether to allocate the iovec area */
L
Linus Torvalds 已提交
1852 1853 1854 1855 1856 1857 1858 1859 1860 1861
	err = -ENOMEM;
	iov_size = msg_sys.msg_iovlen * sizeof(struct iovec);
	if (msg_sys.msg_iovlen > UIO_FASTIOV) {
		iov = sock_kmalloc(sock->sk, iov_size, GFP_KERNEL);
		if (!iov)
			goto out_put;
	}

	/* This will also move the address data into kernel space */
	if (MSG_CMSG_COMPAT & flags) {
1862 1863 1864
		err = verify_compat_iovec(&msg_sys, iov,
					  (struct sockaddr *)&address,
					  VERIFY_READ);
L
Linus Torvalds 已提交
1865
	} else
1866 1867 1868
		err = verify_iovec(&msg_sys, iov,
				   (struct sockaddr *)&address,
				   VERIFY_READ);
1869
	if (err < 0)
L
Linus Torvalds 已提交
1870 1871 1872 1873 1874 1875 1876
		goto out_freeiov;
	total_len = err;

	err = -ENOBUFS;

	if (msg_sys.msg_controllen > INT_MAX)
		goto out_freeiov;
1877
	ctl_len = msg_sys.msg_controllen;
L
Linus Torvalds 已提交
1878
	if ((MSG_CMSG_COMPAT & flags) && ctl_len) {
1879 1880 1881
		err =
		    cmsghdr_from_user_compat_to_kern(&msg_sys, sock->sk, ctl,
						     sizeof(ctl));
L
Linus Torvalds 已提交
1882 1883 1884
		if (err)
			goto out_freeiov;
		ctl_buf = msg_sys.msg_control;
A
Al Viro 已提交
1885
		ctl_len = msg_sys.msg_controllen;
L
Linus Torvalds 已提交
1886
	} else if (ctl_len) {
1887
		if (ctl_len > sizeof(ctl)) {
L
Linus Torvalds 已提交
1888
			ctl_buf = sock_kmalloc(sock->sk, ctl_len, GFP_KERNEL);
1889
			if (ctl_buf == NULL)
L
Linus Torvalds 已提交
1890 1891 1892 1893 1894 1895 1896 1897
				goto out_freeiov;
		}
		err = -EFAULT;
		/*
		 * Careful! Before this, msg_sys.msg_control contains a user pointer.
		 * Afterwards, it will be a kernel pointer. Thus the compiler-assisted
		 * checking falls down on this.
		 */
1898 1899
		if (copy_from_user(ctl_buf, (void __user *)msg_sys.msg_control,
				   ctl_len))
L
Linus Torvalds 已提交
1900 1901 1902 1903 1904 1905 1906 1907 1908 1909
			goto out_freectl;
		msg_sys.msg_control = ctl_buf;
	}
	msg_sys.msg_flags = flags;

	if (sock->file->f_flags & O_NONBLOCK)
		msg_sys.msg_flags |= MSG_DONTWAIT;
	err = sock_sendmsg(sock, &msg_sys, total_len);

out_freectl:
1910
	if (ctl_buf != ctl)
L
Linus Torvalds 已提交
1911 1912 1913 1914 1915
		sock_kfree_s(sock->sk, ctl_buf, ctl_len);
out_freeiov:
	if (iov != iovstack)
		sock_kfree_s(sock->sk, iov, iov_size);
out_put:
1916
	fput_light(sock->file, fput_needed);
1917
out:
L
Linus Torvalds 已提交
1918 1919 1920 1921 1922 1923 1924
	return err;
}

/*
 *	BSD recvmsg interface
 */

1925 1926
SYSCALL_DEFINE3(recvmsg, int, fd, struct msghdr __user *, msg,
		unsigned int, flags)
L
Linus Torvalds 已提交
1927
{
1928 1929
	struct compat_msghdr __user *msg_compat =
	    (struct compat_msghdr __user *)msg;
L
Linus Torvalds 已提交
1930 1931
	struct socket *sock;
	struct iovec iovstack[UIO_FASTIOV];
1932
	struct iovec *iov = iovstack;
L
Linus Torvalds 已提交
1933 1934 1935
	struct msghdr msg_sys;
	unsigned long cmsg_ptr;
	int err, iov_size, total_len, len;
1936
	int fput_needed;
L
Linus Torvalds 已提交
1937 1938

	/* kernel mode address */
1939
	struct sockaddr_storage addr;
L
Linus Torvalds 已提交
1940 1941 1942 1943

	/* user mode address pointers */
	struct sockaddr __user *uaddr;
	int __user *uaddr_len;
1944

L
Linus Torvalds 已提交
1945 1946 1947
	if (MSG_CMSG_COMPAT & flags) {
		if (get_compat_msghdr(&msg_sys, msg_compat))
			return -EFAULT;
1948 1949 1950
	}
	else if (copy_from_user(&msg_sys, msg, sizeof(struct msghdr)))
		return -EFAULT;
L
Linus Torvalds 已提交
1951

1952
	sock = sockfd_lookup_light(fd, &err, &fput_needed);
L
Linus Torvalds 已提交
1953 1954 1955 1956 1957 1958
	if (!sock)
		goto out;

	err = -EMSGSIZE;
	if (msg_sys.msg_iovlen > UIO_MAXIOV)
		goto out_put;
1959 1960

	/* Check whether to allocate the iovec area */
L
Linus Torvalds 已提交
1961 1962 1963 1964 1965 1966 1967 1968 1969
	err = -ENOMEM;
	iov_size = msg_sys.msg_iovlen * sizeof(struct iovec);
	if (msg_sys.msg_iovlen > UIO_FASTIOV) {
		iov = sock_kmalloc(sock->sk, iov_size, GFP_KERNEL);
		if (!iov)
			goto out_put;
	}

	/*
1970 1971
	 *      Save the user-mode address (verify_iovec will change the
	 *      kernel msghdr to use the kernel address space)
L
Linus Torvalds 已提交
1972
	 */
1973

S
Stephen Hemminger 已提交
1974
	uaddr = (__force void __user *)msg_sys.msg_name;
L
Linus Torvalds 已提交
1975 1976
	uaddr_len = COMPAT_NAMELEN(msg);
	if (MSG_CMSG_COMPAT & flags) {
1977 1978 1979
		err = verify_compat_iovec(&msg_sys, iov,
					  (struct sockaddr *)&addr,
					  VERIFY_WRITE);
L
Linus Torvalds 已提交
1980
	} else
1981 1982 1983
		err = verify_iovec(&msg_sys, iov,
				   (struct sockaddr *)&addr,
				   VERIFY_WRITE);
L
Linus Torvalds 已提交
1984 1985
	if (err < 0)
		goto out_freeiov;
1986
	total_len = err;
L
Linus Torvalds 已提交
1987 1988

	cmsg_ptr = (unsigned long)msg_sys.msg_control;
U
Ulrich Drepper 已提交
1989
	msg_sys.msg_flags = flags & (MSG_CMSG_CLOEXEC|MSG_CMSG_COMPAT);
1990

L
Linus Torvalds 已提交
1991 1992 1993 1994 1995 1996 1997 1998
	if (sock->file->f_flags & O_NONBLOCK)
		flags |= MSG_DONTWAIT;
	err = sock_recvmsg(sock, &msg_sys, total_len, flags);
	if (err < 0)
		goto out_freeiov;
	len = err;

	if (uaddr != NULL) {
1999 2000
		err = move_addr_to_user((struct sockaddr *)&addr,
					msg_sys.msg_namelen, uaddr,
2001
					uaddr_len);
L
Linus Torvalds 已提交
2002 2003 2004
		if (err < 0)
			goto out_freeiov;
	}
2005 2006
	err = __put_user((msg_sys.msg_flags & ~MSG_CMSG_COMPAT),
			 COMPAT_FLAGS(msg));
L
Linus Torvalds 已提交
2007 2008 2009
	if (err)
		goto out_freeiov;
	if (MSG_CMSG_COMPAT & flags)
2010
		err = __put_user((unsigned long)msg_sys.msg_control - cmsg_ptr,
L
Linus Torvalds 已提交
2011 2012
				 &msg_compat->msg_controllen);
	else
2013
		err = __put_user((unsigned long)msg_sys.msg_control - cmsg_ptr,
L
Linus Torvalds 已提交
2014 2015 2016 2017 2018 2019 2020 2021 2022
				 &msg->msg_controllen);
	if (err)
		goto out_freeiov;
	err = len;

out_freeiov:
	if (iov != iovstack)
		sock_kfree_s(sock->sk, iov, iov_size);
out_put:
2023
	fput_light(sock->file, fput_needed);
L
Linus Torvalds 已提交
2024 2025 2026 2027 2028 2029 2030 2031
out:
	return err;
}

#ifdef __ARCH_WANT_SYS_SOCKETCALL

/* Argument list sizes for sys_socketcall */
#define AL(x) ((x) * sizeof(unsigned long))
U
Ulrich Drepper 已提交
2032
static const unsigned char nargs[19]={
2033 2034
	AL(0),AL(3),AL(3),AL(3),AL(2),AL(3),
	AL(3),AL(3),AL(4),AL(4),AL(4),AL(6),
U
Ulrich Drepper 已提交
2035
	AL(6),AL(2),AL(5),AL(5),AL(3),AL(3),
U
Ulrich Drepper 已提交
2036
	AL(4)
2037 2038
};

L
Linus Torvalds 已提交
2039 2040 2041
#undef AL

/*
2042
 *	System call vectors.
L
Linus Torvalds 已提交
2043 2044 2045
 *
 *	Argument checking cleaned up. Saved 20% in size.
 *  This function doesn't need to set the kernel lock because
2046
 *  it is set by the callees.
L
Linus Torvalds 已提交
2047 2048
 */

2049
SYSCALL_DEFINE2(socketcall, int, call, unsigned long __user *, args)
L
Linus Torvalds 已提交
2050 2051
{
	unsigned long a[6];
2052
	unsigned long a0, a1;
L
Linus Torvalds 已提交
2053 2054
	int err;

U
Ulrich Drepper 已提交
2055
	if (call < 1 || call > SYS_ACCEPT4)
L
Linus Torvalds 已提交
2056 2057 2058 2059 2060
		return -EINVAL;

	/* copy_from_user should be SMP safe. */
	if (copy_from_user(a, args, nargs[call]))
		return -EFAULT;
2061

A
Al Viro 已提交
2062
	audit_socketcall(nargs[call] / sizeof(unsigned long), a);
2063

2064 2065 2066 2067 2068 2069 2070 2071 2072 2073 2074 2075 2076 2077 2078 2079 2080
	a0 = a[0];
	a1 = a[1];

	switch (call) {
	case SYS_SOCKET:
		err = sys_socket(a0, a1, a[2]);
		break;
	case SYS_BIND:
		err = sys_bind(a0, (struct sockaddr __user *)a1, a[2]);
		break;
	case SYS_CONNECT:
		err = sys_connect(a0, (struct sockaddr __user *)a1, a[2]);
		break;
	case SYS_LISTEN:
		err = sys_listen(a0, a1);
		break;
	case SYS_ACCEPT:
U
Ulrich Drepper 已提交
2081 2082
		err = sys_accept4(a0, (struct sockaddr __user *)a1,
				  (int __user *)a[2], 0);
2083 2084 2085 2086 2087 2088 2089 2090 2091 2092 2093 2094 2095 2096 2097 2098 2099 2100 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
		break;
	case SYS_GETSOCKNAME:
		err =
		    sys_getsockname(a0, (struct sockaddr __user *)a1,
				    (int __user *)a[2]);
		break;
	case SYS_GETPEERNAME:
		err =
		    sys_getpeername(a0, (struct sockaddr __user *)a1,
				    (int __user *)a[2]);
		break;
	case SYS_SOCKETPAIR:
		err = sys_socketpair(a0, a1, a[2], (int __user *)a[3]);
		break;
	case SYS_SEND:
		err = sys_send(a0, (void __user *)a1, a[2], a[3]);
		break;
	case SYS_SENDTO:
		err = sys_sendto(a0, (void __user *)a1, a[2], a[3],
				 (struct sockaddr __user *)a[4], a[5]);
		break;
	case SYS_RECV:
		err = sys_recv(a0, (void __user *)a1, a[2], a[3]);
		break;
	case SYS_RECVFROM:
		err = sys_recvfrom(a0, (void __user *)a1, a[2], a[3],
				   (struct sockaddr __user *)a[4],
				   (int __user *)a[5]);
		break;
	case SYS_SHUTDOWN:
		err = sys_shutdown(a0, a1);
		break;
	case SYS_SETSOCKOPT:
		err = sys_setsockopt(a0, a1, a[2], (char __user *)a[3], a[4]);
		break;
	case SYS_GETSOCKOPT:
		err =
		    sys_getsockopt(a0, a1, a[2], (char __user *)a[3],
				   (int __user *)a[4]);
		break;
	case SYS_SENDMSG:
		err = sys_sendmsg(a0, (struct msghdr __user *)a1, a[2]);
		break;
	case SYS_RECVMSG:
		err = sys_recvmsg(a0, (struct msghdr __user *)a1, a[2]);
		break;
U
Ulrich Drepper 已提交
2129 2130 2131
	case SYS_ACCEPT4:
		err = sys_accept4(a0, (struct sockaddr __user *)a1,
				  (int __user *)a[2], a[3]);
U
Ulrich Drepper 已提交
2132
		break;
2133 2134 2135
	default:
		err = -EINVAL;
		break;
L
Linus Torvalds 已提交
2136 2137 2138 2139
	}
	return err;
}

2140
#endif				/* __ARCH_WANT_SYS_SOCKETCALL */
L
Linus Torvalds 已提交
2141

2142 2143 2144 2145
/**
 *	sock_register - add a socket protocol handler
 *	@ops: description of protocol
 *
L
Linus Torvalds 已提交
2146 2147
 *	This function is called by a protocol handler that wants to
 *	advertise its address family, and have it linked into the
2148 2149
 *	socket interface. The value ops->family coresponds to the
 *	socket system call protocol family.
L
Linus Torvalds 已提交
2150
 */
2151
int sock_register(const struct net_proto_family *ops)
L
Linus Torvalds 已提交
2152 2153 2154 2155
{
	int err;

	if (ops->family >= NPROTO) {
2156 2157
		printk(KERN_CRIT "protocol %d >= NPROTO(%d)\n", ops->family,
		       NPROTO);
L
Linus Torvalds 已提交
2158 2159
		return -ENOBUFS;
	}
2160 2161 2162 2163 2164

	spin_lock(&net_family_lock);
	if (net_families[ops->family])
		err = -EEXIST;
	else {
2165
		net_families[ops->family] = ops;
L
Linus Torvalds 已提交
2166 2167
		err = 0;
	}
2168 2169
	spin_unlock(&net_family_lock);

2170
	printk(KERN_INFO "NET: Registered protocol family %d\n", ops->family);
L
Linus Torvalds 已提交
2171 2172 2173
	return err;
}

2174 2175 2176 2177
/**
 *	sock_unregister - remove a protocol handler
 *	@family: protocol family to remove
 *
L
Linus Torvalds 已提交
2178 2179
 *	This function is called by a protocol handler that wants to
 *	remove its address family, and have it unlinked from the
2180 2181 2182 2183 2184 2185
 *	new socket creation.
 *
 *	If protocol handler is a module, then it can use module reference
 *	counts to protect against new references. If protocol handler is not
 *	a module then it needs to provide its own protection in
 *	the ops->create routine.
L
Linus Torvalds 已提交
2186
 */
2187
void sock_unregister(int family)
L
Linus Torvalds 已提交
2188
{
2189
	BUG_ON(family < 0 || family >= NPROTO);
L
Linus Torvalds 已提交
2190

2191
	spin_lock(&net_family_lock);
2192
	net_families[family] = NULL;
2193 2194 2195 2196
	spin_unlock(&net_family_lock);

	synchronize_rcu();

2197
	printk(KERN_INFO "NET: Unregistered protocol family %d\n", family);
L
Linus Torvalds 已提交
2198 2199
}

2200
static int __init sock_init(void)
L
Linus Torvalds 已提交
2201 2202
{
	/*
2203
	 *      Initialize sock SLAB cache.
L
Linus Torvalds 已提交
2204
	 */
2205

L
Linus Torvalds 已提交
2206 2207 2208
	sk_init();

	/*
2209
	 *      Initialize skbuff SLAB cache
L
Linus Torvalds 已提交
2210 2211 2212 2213
	 */
	skb_init();

	/*
2214
	 *      Initialize the protocols module.
L
Linus Torvalds 已提交
2215 2216 2217 2218 2219
	 */

	init_inodecache();
	register_filesystem(&sock_fs_type);
	sock_mnt = kern_mount(&sock_fs_type);
2220 2221

	/* The real protocol initialization is performed in later initcalls.
L
Linus Torvalds 已提交
2222 2223 2224 2225 2226
	 */

#ifdef CONFIG_NETFILTER
	netfilter_init();
#endif
2227 2228

	return 0;
L
Linus Torvalds 已提交
2229 2230
}

2231 2232
core_initcall(sock_init);	/* early initcall */

L
Linus Torvalds 已提交
2233 2234 2235 2236 2237 2238
#ifdef CONFIG_PROC_FS
void socket_seq_show(struct seq_file *seq)
{
	int cpu;
	int counter = 0;

2239
	for_each_possible_cpu(cpu)
2240
	    counter += per_cpu(sockets_in_use, cpu);
L
Linus Torvalds 已提交
2241 2242 2243 2244 2245 2246 2247

	/* It can be negative, by the way. 8) */
	if (counter < 0)
		counter = 0;

	seq_printf(seq, "sockets: used %d\n", counter);
}
2248
#endif				/* CONFIG_PROC_FS */
L
Linus Torvalds 已提交
2249

2250 2251
#ifdef CONFIG_COMPAT
static long compat_sock_ioctl(struct file *file, unsigned cmd,
2252
			      unsigned long arg)
2253 2254 2255
{
	struct socket *sock = file->private_data;
	int ret = -ENOIOCTLCMD;
2256 2257 2258 2259 2260
	struct sock *sk;
	struct net *net;

	sk = sock->sk;
	net = sock_net(sk);
2261 2262 2263 2264

	if (sock->ops->compat_ioctl)
		ret = sock->ops->compat_ioctl(sock, cmd, arg);

2265 2266 2267 2268
	if (ret == -ENOIOCTLCMD &&
	    (cmd >= SIOCIWFIRST && cmd <= SIOCIWLAST))
		ret = compat_wext_handle_ioctl(net, cmd, arg);

2269 2270 2271 2272
	return ret;
}
#endif

2273 2274 2275 2276 2277 2278 2279 2280 2281 2282 2283 2284 2285 2286 2287 2288 2289 2290 2291 2292 2293 2294 2295
int kernel_bind(struct socket *sock, struct sockaddr *addr, int addrlen)
{
	return sock->ops->bind(sock, addr, addrlen);
}

int kernel_listen(struct socket *sock, int backlog)
{
	return sock->ops->listen(sock, backlog);
}

int kernel_accept(struct socket *sock, struct socket **newsock, int flags)
{
	struct sock *sk = sock->sk;
	int err;

	err = sock_create_lite(sk->sk_family, sk->sk_type, sk->sk_protocol,
			       newsock);
	if (err < 0)
		goto done;

	err = sock->ops->accept(sock, *newsock, flags);
	if (err < 0) {
		sock_release(*newsock);
2296
		*newsock = NULL;
2297 2298 2299 2300
		goto done;
	}

	(*newsock)->ops = sock->ops;
2301
	__module_get((*newsock)->ops->owner);
2302 2303 2304 2305 2306 2307

done:
	return err;
}

int kernel_connect(struct socket *sock, struct sockaddr *addr, int addrlen,
2308
		   int flags)
2309 2310 2311 2312 2313 2314 2315 2316 2317 2318 2319 2320 2321 2322 2323 2324 2325 2326 2327 2328 2329 2330 2331 2332 2333 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 2359 2360 2361 2362 2363 2364 2365 2366 2367 2368 2369 2370 2371 2372 2373 2374 2375 2376 2377
{
	return sock->ops->connect(sock, addr, addrlen, flags);
}

int kernel_getsockname(struct socket *sock, struct sockaddr *addr,
			 int *addrlen)
{
	return sock->ops->getname(sock, addr, addrlen, 0);
}

int kernel_getpeername(struct socket *sock, struct sockaddr *addr,
			 int *addrlen)
{
	return sock->ops->getname(sock, addr, addrlen, 1);
}

int kernel_getsockopt(struct socket *sock, int level, int optname,
			char *optval, int *optlen)
{
	mm_segment_t oldfs = get_fs();
	int err;

	set_fs(KERNEL_DS);
	if (level == SOL_SOCKET)
		err = sock_getsockopt(sock, level, optname, optval, optlen);
	else
		err = sock->ops->getsockopt(sock, level, optname, optval,
					    optlen);
	set_fs(oldfs);
	return err;
}

int kernel_setsockopt(struct socket *sock, int level, int optname,
			char *optval, int optlen)
{
	mm_segment_t oldfs = get_fs();
	int err;

	set_fs(KERNEL_DS);
	if (level == SOL_SOCKET)
		err = sock_setsockopt(sock, level, optname, optval, optlen);
	else
		err = sock->ops->setsockopt(sock, level, optname, optval,
					    optlen);
	set_fs(oldfs);
	return err;
}

int kernel_sendpage(struct socket *sock, struct page *page, int offset,
		    size_t size, int flags)
{
	if (sock->ops->sendpage)
		return sock->ops->sendpage(sock, page, offset, size, flags);

	return sock_no_sendpage(sock, page, offset, size, flags);
}

int kernel_sock_ioctl(struct socket *sock, int cmd, unsigned long arg)
{
	mm_segment_t oldfs = get_fs();
	int err;

	set_fs(KERNEL_DS);
	err = sock->ops->ioctl(sock, cmd, arg);
	set_fs(oldfs);

	return err;
}

2378 2379 2380 2381 2382
int kernel_sock_shutdown(struct socket *sock, enum sock_shutdown_cmd how)
{
	return sock->ops->shutdown(sock, how);
}

L
Linus Torvalds 已提交
2383 2384 2385 2386 2387 2388 2389 2390 2391 2392 2393 2394 2395
EXPORT_SYMBOL(sock_create);
EXPORT_SYMBOL(sock_create_kern);
EXPORT_SYMBOL(sock_create_lite);
EXPORT_SYMBOL(sock_map_fd);
EXPORT_SYMBOL(sock_recvmsg);
EXPORT_SYMBOL(sock_register);
EXPORT_SYMBOL(sock_release);
EXPORT_SYMBOL(sock_sendmsg);
EXPORT_SYMBOL(sock_unregister);
EXPORT_SYMBOL(sock_wake_async);
EXPORT_SYMBOL(sockfd_lookup);
EXPORT_SYMBOL(kernel_sendmsg);
EXPORT_SYMBOL(kernel_recvmsg);
2396 2397 2398 2399 2400 2401 2402 2403 2404 2405
EXPORT_SYMBOL(kernel_bind);
EXPORT_SYMBOL(kernel_listen);
EXPORT_SYMBOL(kernel_accept);
EXPORT_SYMBOL(kernel_connect);
EXPORT_SYMBOL(kernel_getsockname);
EXPORT_SYMBOL(kernel_getpeername);
EXPORT_SYMBOL(kernel_getsockopt);
EXPORT_SYMBOL(kernel_setsockopt);
EXPORT_SYMBOL(kernel_sendpage);
EXPORT_SYMBOL(kernel_sock_ioctl);
2406
EXPORT_SYMBOL(kernel_sock_shutdown);