socket.c 58.1 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>
N
Nick Black 已提交
89
#include <linux/magic.h>
L
Linus Torvalds 已提交
90 91 92 93 94

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

#include <net/compat.h>
95
#include <net/wext.h>
L
Linus Torvalds 已提交
96 97 98 99 100

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

static int sock_no_open(struct inode *irrelevant, struct file *dontcare);
101 102 103 104
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);
105
static int sock_mmap(struct file *file, struct vm_area_struct *vma);
L
Linus Torvalds 已提交
106 107 108 109

static int sock_close(struct inode *inode, struct file *file);
static unsigned int sock_poll(struct file *file,
			      struct poll_table_struct *wait);
110
static long sock_ioctl(struct file *file, unsigned int cmd, unsigned long arg);
111 112
#ifdef CONFIG_COMPAT
static long compat_sock_ioctl(struct file *file,
113
			      unsigned int cmd, unsigned long arg);
114
#endif
L
Linus Torvalds 已提交
115 116 117
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
118 119 120
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
121 122 123 124 125 126

/*
 *	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.
 */

127
static const struct file_operations socket_file_ops = {
L
Linus Torvalds 已提交
128 129 130 131 132 133
	.owner =	THIS_MODULE,
	.llseek =	no_llseek,
	.aio_read =	sock_aio_read,
	.aio_write =	sock_aio_write,
	.poll =		sock_poll,
	.unlocked_ioctl = sock_ioctl,
134 135 136
#ifdef CONFIG_COMPAT
	.compat_ioctl = compat_sock_ioctl,
#endif
L
Linus Torvalds 已提交
137 138 139 140
	.mmap =		sock_mmap,
	.open =		sock_no_open,	/* special open code to disallow open via /proc */
	.release =	sock_close,
	.fasync =	sock_fasync,
141 142
	.sendpage =	sock_sendpage,
	.splice_write = generic_splice_sendpage,
J
Jens Axboe 已提交
143
	.splice_read =	sock_splice_read,
L
Linus Torvalds 已提交
144 145 146 147 148 149 150
};

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

static DEFINE_SPINLOCK(net_family_lock);
151
static const struct net_proto_family *net_families[NPROTO] __read_mostly;
L
Linus Torvalds 已提交
152 153 154 155 156 157 158 159

/*
 *	Statistics counters of the socket lists
 */

static DEFINE_PER_CPU(int, sockets_in_use) = 0;

/*
160 161 162
 * 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
163 164
 */

165
#define MAX_SOCK_ADDR	128		/* 108 for Unix domain -
L
Linus Torvalds 已提交
166 167
					   16 for IP, 16 for IPX,
					   24 for IPv6,
168
					   about 80 for AX.25
L
Linus Torvalds 已提交
169 170
					   must be at least one bigger than
					   the AF_UNIX size (see net/unix/af_unix.c
171
					   :unix_mkname()).
L
Linus Torvalds 已提交
172
					 */
173

L
Linus Torvalds 已提交
174 175 176 177 178 179 180 181 182 183 184
/**
 *	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.
 */

185
int move_addr_to_kernel(void __user *uaddr, int ulen, struct sockaddr *kaddr)
L
Linus Torvalds 已提交
186
{
187
	if (ulen < 0 || ulen > sizeof(struct sockaddr_storage))
L
Linus Torvalds 已提交
188
		return -EINVAL;
189
	if (ulen == 0)
L
Linus Torvalds 已提交
190
		return 0;
191
	if (copy_from_user(kaddr, uaddr, ulen))
L
Linus Torvalds 已提交
192
		return -EFAULT;
193
	return audit_sockaddr(ulen, kaddr);
L
Linus Torvalds 已提交
194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211
}

/**
 *	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.
 */
212

213
int move_addr_to_user(struct sockaddr *kaddr, int klen, void __user *uaddr,
214
		      int __user *ulen)
L
Linus Torvalds 已提交
215 216 217 218
{
	int err;
	int len;

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

239
static struct kmem_cache *sock_inode_cachep __read_mostly;
L
Linus Torvalds 已提交
240 241 242 243

static struct inode *sock_alloc_inode(struct super_block *sb)
{
	struct socket_alloc *ei;
244

245
	ei = kmem_cache_alloc(sock_inode_cachep, GFP_KERNEL);
L
Linus Torvalds 已提交
246 247 248
	if (!ei)
		return NULL;
	init_waitqueue_head(&ei->socket.wait);
249

L
Linus Torvalds 已提交
250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265
	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));
}

266
static void init_once(void *foo)
L
Linus Torvalds 已提交
267
{
268
	struct socket_alloc *ei = (struct socket_alloc *)foo;
L
Linus Torvalds 已提交
269

C
Christoph Lameter 已提交
270
	inode_init_once(&ei->vfs_inode);
L
Linus Torvalds 已提交
271
}
272

L
Linus Torvalds 已提交
273 274 275
static int init_inodecache(void)
{
	sock_inode_cachep = kmem_cache_create("sock_inode_cache",
276 277 278 279 280
					      sizeof(struct socket_alloc),
					      0,
					      (SLAB_HWCACHE_ALIGN |
					       SLAB_RECLAIM_ACCOUNT |
					       SLAB_MEM_SPREAD),
281
					      init_once);
L
Linus Torvalds 已提交
282 283 284 285 286
	if (sock_inode_cachep == NULL)
		return -ENOMEM;
	return 0;
}

287
static const struct super_operations sockfs_ops = {
L
Linus Torvalds 已提交
288 289 290 291 292
	.alloc_inode =	sock_alloc_inode,
	.destroy_inode =sock_destroy_inode,
	.statfs =	simple_statfs,
};

293
static int sockfs_get_sb(struct file_system_type *fs_type,
294 295
			 int flags, const char *dev_name, void *data,
			 struct vfsmount *mnt)
L
Linus Torvalds 已提交
296
{
297 298
	return get_sb_pseudo(fs_type, "socket:", &sockfs_ops, SOCKFS_MAGIC,
			     mnt);
L
Linus Torvalds 已提交
299 300
}

301
static struct vfsmount *sock_mnt __read_mostly;
L
Linus Torvalds 已提交
302 303 304 305 306 307

static struct file_system_type sock_fs_type = {
	.name =		"sockfs",
	.get_sb =	sockfs_get_sb,
	.kill_sb =	kill_anon_super,
};
308

L
Linus Torvalds 已提交
309 310
static int sockfs_delete_dentry(struct dentry *dentry)
{
311 312 313 314 315 316 317 318
	/*
	 * 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
319
}
320 321 322 323 324 325 326 327 328 329

/*
 * 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);
}

A
Al Viro 已提交
330
static const struct dentry_operations sockfs_dentry_operations = {
331
	.d_delete = sockfs_delete_dentry,
332
	.d_dname  = sockfs_dname,
L
Linus Torvalds 已提交
333 334 335 336 337
};

/*
 *	Obtains the first available file descriptor and sets it up for use.
 *
338 339
 *	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
340 341 342 343 344 345 346 347 348 349 350 351
 *	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.
 */

352
static int sock_alloc_fd(struct file **filep, int flags)
L
Linus Torvalds 已提交
353 354 355
{
	int fd;

356
	fd = get_unused_fd_flags(flags);
357
	if (likely(fd >= 0)) {
L
Linus Torvalds 已提交
358 359
		struct file *file = get_empty_filp();

360 361
		*filep = file;
		if (unlikely(!file)) {
L
Linus Torvalds 已提交
362
			put_unused_fd(fd);
363
			return -ENFILE;
L
Linus Torvalds 已提交
364
		}
365 366 367 368
	} else
		*filep = NULL;
	return fd;
}
L
Linus Torvalds 已提交
369

370
static int sock_attach_fd(struct socket *sock, struct file *file, int flags)
371
{
372
	struct dentry *dentry;
373
	struct qstr name = { .name = "" };
374

375 376
	dentry = d_alloc(sock_mnt->mnt_sb->s_root, &name);
	if (unlikely(!dentry))
377 378
		return -ENOMEM;

379
	dentry->d_op = &sockfs_dentry_operations;
380 381 382 383 384
	/*
	 * 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
	 */
385 386
	dentry->d_flags &= ~DCACHE_UNHASHED;
	d_instantiate(dentry, SOCK_INODE(sock));
387 388

	sock->file = file;
389 390 391
	init_file(file, sock_mnt, dentry, FMODE_READ | FMODE_WRITE,
		  &socket_file_ops);
	SOCK_INODE(sock)->i_fop = &socket_file_ops;
392
	file->f_flags = O_RDWR | (flags & O_NONBLOCK);
393 394
	file->f_pos = 0;
	file->private_data = sock;
L
Linus Torvalds 已提交
395

396 397 398
	return 0;
}

399
int sock_map_fd(struct socket *sock, int flags)
400 401
{
	struct file *newfile;
402
	int fd = sock_alloc_fd(&newfile, flags);
403 404

	if (likely(fd >= 0)) {
405
		int err = sock_attach_fd(sock, newfile, flags);
406 407 408

		if (unlikely(err < 0)) {
			put_filp(newfile);
L
Linus Torvalds 已提交
409
			put_unused_fd(fd);
410
			return err;
L
Linus Torvalds 已提交
411
		}
412
		fd_install(fd, newfile);
L
Linus Torvalds 已提交
413 414 415 416
	}
	return fd;
}

417 418 419 420 421
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
422 423
	*err = -ENOTSOCK;
	return NULL;
424 425
}

L
Linus Torvalds 已提交
426 427 428 429 430 431 432 433 434 435 436 437 438 439 440 441 442 443
/**
 *	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;

444 445
	file = fget(fd);
	if (!file) {
L
Linus Torvalds 已提交
446 447 448
		*err = -EBADF;
		return NULL;
	}
449

450 451
	sock = sock_from_file(file, err);
	if (!sock)
L
Linus Torvalds 已提交
452
		fput(file);
453 454
	return sock;
}
L
Linus Torvalds 已提交
455

456 457 458 459 460
static struct socket *sockfd_lookup_light(int fd, int *err, int *fput_needed)
{
	struct file *file;
	struct socket *sock;

461
	*err = -EBADF;
462 463 464 465 466 467
	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
468
	}
469
	return NULL;
L
Linus Torvalds 已提交
470 471 472 473
}

/**
 *	sock_alloc	-	allocate a socket
474
 *
L
Linus Torvalds 已提交
475 476 477 478 479 480 481
 *	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)
{
482 483
	struct inode *inode;
	struct socket *sock;
L
Linus Torvalds 已提交
484 485 486 487 488 489 490

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

	sock = SOCKET_I(inode);

491
	kmemcheck_annotate_bitfield(sock, type);
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
	percpu_add(sockets_in_use, 1);
L
Linus Torvalds 已提交
497 498 499 500 501 502 503 504
	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.
 */
505

L
Linus Torvalds 已提交
506 507 508 509 510
static int sock_no_open(struct inode *irrelevant, struct file *dontcare)
{
	return -ENXIO;
}

511
const struct file_operations bad_sock_fops = {
L
Linus Torvalds 已提交
512 513 514 515 516 517 518 519 520 521
	.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
522
 *	an inode not a file.
L
Linus Torvalds 已提交
523
 */
524

L
Linus Torvalds 已提交
525 526 527 528 529 530 531 532 533 534 535 536 537
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");

538
	percpu_sub(sockets_in_use, 1);
L
Linus Torvalds 已提交
539 540 541 542
	if (!sock->file) {
		iput(SOCK_INODE(sock));
		return;
	}
543
	sock->file = NULL;
L
Linus Torvalds 已提交
544 545
}

546 547 548 549 550 551 552 553 554 555 556 557
int sock_tx_timestamp(struct msghdr *msg, struct sock *sk,
		      union skb_shared_tx *shtx)
{
	shtx->flags = 0;
	if (sock_flag(sk, SOCK_TIMESTAMPING_TX_HARDWARE))
		shtx->hardware = 1;
	if (sock_flag(sk, SOCK_TIMESTAMPING_TX_SOFTWARE))
		shtx->software = 1;
	return 0;
}
EXPORT_SYMBOL(sock_tx_timestamp);

558
static inline int __sock_sendmsg(struct kiocb *iocb, struct socket *sock,
L
Linus Torvalds 已提交
559 560 561 562 563 564 565 566 567 568 569 570 571 572 573 574 575 576 577 578 579 580 581 582 583 584 585 586 587 588 589 590 591 592 593 594 595 596 597 598 599 600
				 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
	 */
601
	msg->msg_iov = (struct iovec *)vec;
L
Linus Torvalds 已提交
602 603 604 605 606 607
	msg->msg_iovlen = num;
	result = sock_sendmsg(sock, msg, size);
	set_fs(oldfs);
	return result;
}

608 609 610 611 612 613 614 615 616 617
static int ktime2ts(ktime_t kt, struct timespec *ts)
{
	if (kt.tv64) {
		*ts = ktime_to_timespec(kt);
		return 1;
	} else {
		return 0;
	}
}

618 619 620 621 622 623
/*
 * 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)
{
624 625 626 627 628 629 630 631 632 633 634 635 636 637 638 639 640 641 642 643 644 645 646 647 648 649 650 651 652 653 654 655 656 657 658 659 660 661 662
	int need_software_tstamp = sock_flag(sk, SOCK_RCVTSTAMP);
	struct timespec ts[3];
	int empty = 1;
	struct skb_shared_hwtstamps *shhwtstamps =
		skb_hwtstamps(skb);

	/* Race occurred between timestamp enabling and packet
	   receiving.  Fill in the current time for now. */
	if (need_software_tstamp && skb->tstamp.tv64 == 0)
		__net_timestamp(skb);

	if (need_software_tstamp) {
		if (!sock_flag(sk, SOCK_RCVTSTAMPNS)) {
			struct timeval tv;
			skb_get_timestamp(skb, &tv);
			put_cmsg(msg, SOL_SOCKET, SCM_TIMESTAMP,
				 sizeof(tv), &tv);
		} else {
			struct timespec ts;
			skb_get_timestampns(skb, &ts);
			put_cmsg(msg, SOL_SOCKET, SCM_TIMESTAMPNS,
				 sizeof(ts), &ts);
		}
	}


	memset(ts, 0, sizeof(ts));
	if (skb->tstamp.tv64 &&
	    sock_flag(sk, SOCK_TIMESTAMPING_SOFTWARE)) {
		skb_get_timestampns(skb, ts + 0);
		empty = 0;
	}
	if (shhwtstamps) {
		if (sock_flag(sk, SOCK_TIMESTAMPING_SYS_HARDWARE) &&
		    ktime2ts(shhwtstamps->syststamp, ts + 1))
			empty = 0;
		if (sock_flag(sk, SOCK_TIMESTAMPING_RAW_HARDWARE) &&
		    ktime2ts(shhwtstamps->hwtstamp, ts + 2))
			empty = 0;
663
	}
664 665 666
	if (!empty)
		put_cmsg(msg, SOL_SOCKET,
			 SCM_TIMESTAMPING, sizeof(ts), &ts);
667 668
}

669 670
EXPORT_SYMBOL_GPL(__sock_recv_timestamp);

671
static inline int __sock_recvmsg(struct kiocb *iocb, struct socket *sock,
L
Linus Torvalds 已提交
672 673 674 675 676 677 678 679 680 681 682 683 684 685 686 687 688 689
				 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);
}

690
int sock_recvmsg(struct socket *sock, struct msghdr *msg,
L
Linus Torvalds 已提交
691 692 693 694 695 696
		 size_t size, int flags)
{
	struct kiocb iocb;
	struct sock_iocb siocb;
	int ret;

697
	init_sync_kiocb(&iocb, NULL);
L
Linus Torvalds 已提交
698 699 700 701 702 703 704
	iocb.private = &siocb;
	ret = __sock_recvmsg(&iocb, sock, msg, size, flags);
	if (-EIOCBQUEUED == ret)
		ret = wait_on_sync_kiocb(&iocb);
	return ret;
}

705 706
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
707 708 709 710 711 712 713 714 715
{
	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
	 */
716
	msg->msg_iov = (struct iovec *)vec, msg->msg_iovlen = num;
L
Linus Torvalds 已提交
717 718 719 720 721 722 723 724 725 726
	result = sock_recvmsg(sock, msg, size, flags);
	set_fs(oldfs);
	return result;
}

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

727 728
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
729 730 731 732
{
	struct socket *sock;
	int flags;

733 734 735 736 737 738
	sock = file->private_data;

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

739
	return kernel_sendpage(sock, page, offset, size, flags);
740
}
L
Linus Torvalds 已提交
741

J
Jens Axboe 已提交
742 743 744 745 746 747
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;

748 749 750
	if (unlikely(!sock->ops->splice_read))
		return -EINVAL;

J
Jens Axboe 已提交
751 752 753
	return sock->ops->splice_read(sock, ppos, pipe, len, flags);
}

754
static struct sock_iocb *alloc_sock_iocb(struct kiocb *iocb,
755
					 struct sock_iocb *siocb)
756 757 758 759 760
{
	if (!is_sync_kiocb(iocb)) {
		siocb = kmalloc(sizeof(*siocb), GFP_KERNEL);
		if (!siocb)
			return NULL;
L
Linus Torvalds 已提交
761 762 763
		iocb->ki_dtor = sock_aio_dtor;
	}

764 765 766
	siocb->kiocb = iocb;
	iocb->private = siocb;
	return siocb;
L
Linus Torvalds 已提交
767 768
}

769
static ssize_t do_sock_read(struct msghdr *msg, struct kiocb *iocb,
770 771
		struct file *file, const struct iovec *iov,
		unsigned long nr_segs)
772 773 774 775
{
	struct socket *sock = file->private_data;
	size_t size = 0;
	int i;
L
Linus Torvalds 已提交
776

777 778
	for (i = 0; i < nr_segs; i++)
		size += iov[i].iov_len;
L
Linus Torvalds 已提交
779

780 781 782 783
	msg->msg_name = NULL;
	msg->msg_namelen = 0;
	msg->msg_control = NULL;
	msg->msg_controllen = 0;
784
	msg->msg_iov = (struct iovec *)iov;
785 786 787 788 789 790
	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);
}

791 792
static ssize_t sock_aio_read(struct kiocb *iocb, const struct iovec *iov,
				unsigned long nr_segs, loff_t pos)
793 794 795
{
	struct sock_iocb siocb, *x;

L
Linus Torvalds 已提交
796 797
	if (pos != 0)
		return -ESPIPE;
798 799

	if (iocb->ki_left == 0)	/* Match SYS5 behaviour */
L
Linus Torvalds 已提交
800 801
		return 0;

802 803

	x = alloc_sock_iocb(iocb, &siocb);
804 805
	if (!x)
		return -ENOMEM;
806
	return do_sock_read(&x->async_msg, iocb, iocb->ki_filp, iov, nr_segs);
L
Linus Torvalds 已提交
807 808
}

809
static ssize_t do_sock_write(struct msghdr *msg, struct kiocb *iocb,
810 811
			struct file *file, const struct iovec *iov,
			unsigned long nr_segs)
L
Linus Torvalds 已提交
812
{
813 814 815
	struct socket *sock = file->private_data;
	size_t size = 0;
	int i;
L
Linus Torvalds 已提交
816

817 818
	for (i = 0; i < nr_segs; i++)
		size += iov[i].iov_len;
L
Linus Torvalds 已提交
819

820 821 822 823
	msg->msg_name = NULL;
	msg->msg_namelen = 0;
	msg->msg_control = NULL;
	msg->msg_controllen = 0;
824
	msg->msg_iov = (struct iovec *)iov;
825 826 827 828
	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
829

830
	return __sock_sendmsg(iocb, sock, msg, size);
L
Linus Torvalds 已提交
831 832
}

833 834
static ssize_t sock_aio_write(struct kiocb *iocb, const struct iovec *iov,
			  unsigned long nr_segs, loff_t pos)
835 836
{
	struct sock_iocb siocb, *x;
L
Linus Torvalds 已提交
837

838 839
	if (pos != 0)
		return -ESPIPE;
840 841

	x = alloc_sock_iocb(iocb, &siocb);
842 843
	if (!x)
		return -ENOMEM;
L
Linus Torvalds 已提交
844

845
	return do_sock_write(&x->async_msg, iocb, iocb->ki_filp, iov, nr_segs);
L
Linus Torvalds 已提交
846 847 848 849 850 851 852
}

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

A
Arjan van de Ven 已提交
853
static DEFINE_MUTEX(br_ioctl_mutex);
854
static int (*br_ioctl_hook) (struct net *, unsigned int cmd, void __user *arg) = NULL;
L
Linus Torvalds 已提交
855

856
void brioctl_set(int (*hook) (struct net *, unsigned int, void __user *))
L
Linus Torvalds 已提交
857
{
A
Arjan van de Ven 已提交
858
	mutex_lock(&br_ioctl_mutex);
L
Linus Torvalds 已提交
859
	br_ioctl_hook = hook;
A
Arjan van de Ven 已提交
860
	mutex_unlock(&br_ioctl_mutex);
L
Linus Torvalds 已提交
861
}
862

L
Linus Torvalds 已提交
863 864
EXPORT_SYMBOL(brioctl_set);

A
Arjan van de Ven 已提交
865
static DEFINE_MUTEX(vlan_ioctl_mutex);
866
static int (*vlan_ioctl_hook) (struct net *, void __user *arg);
L
Linus Torvalds 已提交
867

868
void vlan_ioctl_set(int (*hook) (struct net *, void __user *))
L
Linus Torvalds 已提交
869
{
A
Arjan van de Ven 已提交
870
	mutex_lock(&vlan_ioctl_mutex);
L
Linus Torvalds 已提交
871
	vlan_ioctl_hook = hook;
A
Arjan van de Ven 已提交
872
	mutex_unlock(&vlan_ioctl_mutex);
L
Linus Torvalds 已提交
873
}
874

L
Linus Torvalds 已提交
875 876
EXPORT_SYMBOL(vlan_ioctl_set);

A
Arjan van de Ven 已提交
877
static DEFINE_MUTEX(dlci_ioctl_mutex);
878
static int (*dlci_ioctl_hook) (unsigned int, void __user *);
L
Linus Torvalds 已提交
879

880
void dlci_ioctl_set(int (*hook) (unsigned int, void __user *))
L
Linus Torvalds 已提交
881
{
A
Arjan van de Ven 已提交
882
	mutex_lock(&dlci_ioctl_mutex);
L
Linus Torvalds 已提交
883
	dlci_ioctl_hook = hook;
A
Arjan van de Ven 已提交
884
	mutex_unlock(&dlci_ioctl_mutex);
L
Linus Torvalds 已提交
885
}
886

L
Linus Torvalds 已提交
887 888 889 890 891 892 893 894 895 896
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;
897
	struct sock *sk;
L
Linus Torvalds 已提交
898 899
	void __user *argp = (void __user *)arg;
	int pid, err;
900
	struct net *net;
L
Linus Torvalds 已提交
901

902
	sock = file->private_data;
903
	sk = sock->sk;
904
	net = sock_net(sk);
L
Linus Torvalds 已提交
905
	if (cmd >= SIOCDEVPRIVATE && cmd <= (SIOCDEVPRIVATE + 15)) {
906
		err = dev_ioctl(net, cmd, argp);
L
Linus Torvalds 已提交
907
	} else
908
#ifdef CONFIG_WIRELESS_EXT
L
Linus Torvalds 已提交
909
	if (cmd >= SIOCIWFIRST && cmd <= SIOCIWLAST) {
910
		err = dev_ioctl(net, cmd, argp);
L
Linus Torvalds 已提交
911
	} else
912 913
#endif				/* CONFIG_WIRELESS_EXT */
		switch (cmd) {
L
Linus Torvalds 已提交
914 915 916 917 918 919 920 921 922
		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:
923
			err = put_user(f_getown(sock->file),
924
				       (int __user *)argp);
L
Linus Torvalds 已提交
925 926 927 928 929 930 931 932 933
			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
934
			mutex_lock(&br_ioctl_mutex);
935
			if (br_ioctl_hook)
936
				err = br_ioctl_hook(net, cmd, argp);
A
Arjan van de Ven 已提交
937
			mutex_unlock(&br_ioctl_mutex);
L
Linus Torvalds 已提交
938 939 940 941 942 943 944
			break;
		case SIOCGIFVLAN:
		case SIOCSIFVLAN:
			err = -ENOPKG;
			if (!vlan_ioctl_hook)
				request_module("8021q");

A
Arjan van de Ven 已提交
945
			mutex_lock(&vlan_ioctl_mutex);
L
Linus Torvalds 已提交
946
			if (vlan_ioctl_hook)
947
				err = vlan_ioctl_hook(net, argp);
A
Arjan van de Ven 已提交
948
			mutex_unlock(&vlan_ioctl_mutex);
L
Linus Torvalds 已提交
949 950 951 952 953 954 955
			break;
		case SIOCADDDLCI:
		case SIOCDELDLCI:
			err = -ENOPKG;
			if (!dlci_ioctl_hook)
				request_module("dlci");

956 957
			mutex_lock(&dlci_ioctl_mutex);
			if (dlci_ioctl_hook)
L
Linus Torvalds 已提交
958
				err = dlci_ioctl_hook(cmd, argp);
959
			mutex_unlock(&dlci_ioctl_mutex);
L
Linus Torvalds 已提交
960 961 962
			break;
		default:
			err = sock->ops->ioctl(sock, cmd, arg);
963 964 965 966 967 968

			/*
			 * If this ioctl is unknown try to hand it down
			 * to the NIC driver.
			 */
			if (err == -ENOIOCTLCMD)
969
				err = dev_ioctl(net, cmd, argp);
L
Linus Torvalds 已提交
970
			break;
971
		}
L
Linus Torvalds 已提交
972 973 974 975 976 977 978
	return err;
}

int sock_create_lite(int family, int type, int protocol, struct socket **res)
{
	int err;
	struct socket *sock = NULL;
979

L
Linus Torvalds 已提交
980 981 982 983 984 985 986 987 988 989 990
	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
991 992 993 994
	err = security_socket_post_create(sock, family, type, protocol, 1);
	if (err)
		goto out_release;

L
Linus Torvalds 已提交
995 996 997
out:
	*res = sock;
	return err;
V
Venkat Yekkirala 已提交
998 999 1000 1001
out_release:
	sock_release(sock);
	sock = NULL;
	goto out;
L
Linus Torvalds 已提交
1002 1003 1004
}

/* No kernel lock held - perfect */
1005
static unsigned int sock_poll(struct file *file, poll_table *wait)
L
Linus Torvalds 已提交
1006 1007 1008 1009
{
	struct socket *sock;

	/*
1010
	 *      We can't return errors to poll, so it's either yes or no.
L
Linus Torvalds 已提交
1011
	 */
1012
	sock = file->private_data;
L
Linus Torvalds 已提交
1013 1014 1015
	return sock->ops->poll(file, sock, wait);
}

1016
static int sock_mmap(struct file *file, struct vm_area_struct *vma)
L
Linus Torvalds 已提交
1017
{
1018
	struct socket *sock = file->private_data;
L
Linus Torvalds 已提交
1019 1020 1021 1022

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

1023
static int sock_close(struct inode *inode, struct file *filp)
L
Linus Torvalds 已提交
1024 1025
{
	/*
1026 1027
	 *      It was possible the inode is NULL we were
	 *      closing an unfinished socket.
L
Linus Torvalds 已提交
1028 1029
	 */

1030
	if (!inode) {
L
Linus Torvalds 已提交
1031 1032 1033 1034 1035 1036 1037 1038 1039 1040 1041 1042 1043 1044 1045 1046 1047 1048 1049 1050 1051 1052 1053 1054
		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)
{
1055
	struct fasync_struct *fa, *fna = NULL, **prev;
L
Linus Torvalds 已提交
1056 1057 1058
	struct socket *sock;
	struct sock *sk;

1059
	if (on) {
1060
		fna = kmalloc(sizeof(struct fasync_struct), GFP_KERNEL);
1061
		if (fna == NULL)
L
Linus Torvalds 已提交
1062 1063 1064
			return -ENOMEM;
	}

1065
	sock = filp->private_data;
L
Linus Torvalds 已提交
1066

1067 1068
	sk = sock->sk;
	if (sk == NULL) {
L
Linus Torvalds 已提交
1069 1070 1071 1072 1073 1074
		kfree(fna);
		return -EINVAL;
	}

	lock_sock(sk);

1075 1076 1077 1078 1079 1080 1081
	spin_lock(&filp->f_lock);
	if (on)
		filp->f_flags |= FASYNC;
	else
		filp->f_flags &= ~FASYNC;
	spin_unlock(&filp->f_lock);

1082
	prev = &(sock->fasync_list);
L
Linus Torvalds 已提交
1083

1084 1085
	for (fa = *prev; fa != NULL; prev = &fa->fa_next, fa = *prev)
		if (fa->fa_file == filp)
L
Linus Torvalds 已提交
1086 1087
			break;

1088 1089
	if (on) {
		if (fa != NULL) {
L
Linus Torvalds 已提交
1090
			write_lock_bh(&sk->sk_callback_lock);
1091
			fa->fa_fd = fd;
L
Linus Torvalds 已提交
1092 1093 1094 1095 1096
			write_unlock_bh(&sk->sk_callback_lock);

			kfree(fna);
			goto out;
		}
1097 1098 1099 1100
		fna->fa_file = filp;
		fna->fa_fd = fd;
		fna->magic = FASYNC_MAGIC;
		fna->fa_next = sock->fasync_list;
L
Linus Torvalds 已提交
1101
		write_lock_bh(&sk->sk_callback_lock);
1102
		sock->fasync_list = fna;
L
Linus Torvalds 已提交
1103
		write_unlock_bh(&sk->sk_callback_lock);
1104 1105
	} else {
		if (fa != NULL) {
L
Linus Torvalds 已提交
1106
			write_lock_bh(&sk->sk_callback_lock);
1107
			*prev = fa->fa_next;
L
Linus Torvalds 已提交
1108 1109 1110 1111 1112 1113 1114 1115 1116 1117 1118 1119 1120 1121 1122 1123
			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;
1124
	switch (how) {
1125
	case SOCK_WAKE_WAITD:
L
Linus Torvalds 已提交
1126 1127 1128
		if (test_bit(SOCK_ASYNC_WAITDATA, &sock->flags))
			break;
		goto call_kill;
1129
	case SOCK_WAKE_SPACE:
L
Linus Torvalds 已提交
1130 1131 1132
		if (!test_and_clear_bit(SOCK_ASYNC_NOSPACE, &sock->flags))
			break;
		/* fall through */
1133
	case SOCK_WAKE_IO:
1134
call_kill:
L
Linus Torvalds 已提交
1135 1136
		__kill_fasync(sock->fasync_list, SIGIO, band);
		break;
1137
	case SOCK_WAKE_URG:
L
Linus Torvalds 已提交
1138 1139 1140 1141 1142
		__kill_fasync(sock->fasync_list, SIGURG, band);
	}
	return 0;
}

1143
static int __sock_create(struct net *net, int family, int type, int protocol,
1144
			 struct socket **res, int kern)
L
Linus Torvalds 已提交
1145 1146 1147
{
	int err;
	struct socket *sock;
1148
	const struct net_proto_family *pf;
L
Linus Torvalds 已提交
1149 1150

	/*
1151
	 *      Check protocol is in range
L
Linus Torvalds 已提交
1152 1153 1154 1155 1156 1157 1158 1159 1160 1161 1162 1163
	 */
	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) {
1164
		static int warned;
L
Linus Torvalds 已提交
1165 1166
		if (!warned) {
			warned = 1;
1167 1168
			printk(KERN_INFO "%s uses obsolete (PF_INET,SOCK_PACKET)\n",
			       current->comm);
L
Linus Torvalds 已提交
1169 1170 1171 1172 1173 1174 1175
		}
		family = PF_PACKET;
	}

	err = security_socket_create(family, type, protocol, kern);
	if (err)
		return err;
1176

1177 1178 1179 1180 1181 1182 1183 1184 1185 1186 1187 1188 1189 1190 1191
	/*
	 *	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;

1192
#ifdef CONFIG_MODULES
1193 1194 1195
	/* 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
1196 1197 1198
	 * requested real, full-featured networking support upon configuration.
	 * Otherwise module support will break!
	 */
1199
	if (net_families[family] == NULL)
1200
		request_module("net-pf-%d", family);
L
Linus Torvalds 已提交
1201 1202
#endif

1203 1204 1205 1206 1207
	rcu_read_lock();
	pf = rcu_dereference(net_families[family]);
	err = -EAFNOSUPPORT;
	if (!pf)
		goto out_release;
L
Linus Torvalds 已提交
1208 1209 1210 1211 1212

	/*
	 * We will call the ->create function, that possibly is in a loadable
	 * module, so we have to bump that loadable module refcnt first.
	 */
1213
	if (!try_module_get(pf->owner))
L
Linus Torvalds 已提交
1214 1215
		goto out_release;

1216 1217 1218
	/* Now protected by module ref count */
	rcu_read_unlock();

1219
	err = pf->create(net, sock, protocol);
1220
	if (err < 0)
L
Linus Torvalds 已提交
1221
		goto out_module_put;
1222

L
Linus Torvalds 已提交
1223 1224 1225 1226
	/*
	 * Now to bump the refcnt of the [loadable] module that owns this
	 * socket at sock_release time we decrement its refcnt.
	 */
1227 1228 1229
	if (!try_module_get(sock->ops->owner))
		goto out_module_busy;

L
Linus Torvalds 已提交
1230 1231 1232 1233
	/*
	 * Now that we're done with the ->create function, the [loadable]
	 * module can have its refcnt decremented
	 */
1234
	module_put(pf->owner);
V
Venkat Yekkirala 已提交
1235 1236
	err = security_socket_post_create(sock, family, type, protocol, kern);
	if (err)
1237
		goto out_sock_release;
1238
	*res = sock;
L
Linus Torvalds 已提交
1239

1240 1241 1242 1243
	return 0;

out_module_busy:
	err = -EAFNOSUPPORT;
L
Linus Torvalds 已提交
1244
out_module_put:
1245 1246 1247
	sock->ops = NULL;
	module_put(pf->owner);
out_sock_release:
L
Linus Torvalds 已提交
1248
	sock_release(sock);
1249 1250 1251 1252 1253
	return err;

out_release:
	rcu_read_unlock();
	goto out_sock_release;
L
Linus Torvalds 已提交
1254 1255 1256 1257
}

int sock_create(int family, int type, int protocol, struct socket **res)
{
1258
	return __sock_create(current->nsproxy->net_ns, family, type, protocol, res, 0);
L
Linus Torvalds 已提交
1259 1260 1261 1262
}

int sock_create_kern(int family, int type, int protocol, struct socket **res)
{
1263
	return __sock_create(&init_net, family, type, protocol, res, 1);
L
Linus Torvalds 已提交
1264 1265
}

1266
SYSCALL_DEFINE3(socket, int, family, int, type, int, protocol)
L
Linus Torvalds 已提交
1267 1268 1269
{
	int retval;
	struct socket *sock;
1270 1271
	int flags;

1272 1273 1274 1275 1276 1277
	/* 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);

1278
	flags = type & ~SOCK_TYPE_MASK;
1279
	if (flags & ~(SOCK_CLOEXEC | SOCK_NONBLOCK))
1280 1281
		return -EINVAL;
	type &= SOCK_TYPE_MASK;
L
Linus Torvalds 已提交
1282

U
Ulrich Drepper 已提交
1283 1284 1285
	if (SOCK_NONBLOCK != O_NONBLOCK && (flags & SOCK_NONBLOCK))
		flags = (flags & ~SOCK_NONBLOCK) | O_NONBLOCK;

L
Linus Torvalds 已提交
1286 1287 1288 1289
	retval = sock_create(family, type, protocol, &sock);
	if (retval < 0)
		goto out;

1290
	retval = sock_map_fd(sock, flags & (O_CLOEXEC | O_NONBLOCK));
L
Linus Torvalds 已提交
1291 1292 1293 1294 1295 1296 1297 1298 1299 1300 1301 1302 1303 1304 1305 1306
	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.
 */

1307 1308
SYSCALL_DEFINE4(socketpair, int, family, int, type, int, protocol,
		int __user *, usockvec)
L
Linus Torvalds 已提交
1309 1310 1311
{
	struct socket *sock1, *sock2;
	int fd1, fd2, err;
A
Al Viro 已提交
1312
	struct file *newfile1, *newfile2;
1313 1314 1315
	int flags;

	flags = type & ~SOCK_TYPE_MASK;
1316
	if (flags & ~(SOCK_CLOEXEC | SOCK_NONBLOCK))
1317 1318
		return -EINVAL;
	type &= SOCK_TYPE_MASK;
L
Linus Torvalds 已提交
1319

U
Ulrich Drepper 已提交
1320 1321 1322
	if (SOCK_NONBLOCK != O_NONBLOCK && (flags & SOCK_NONBLOCK))
		flags = (flags & ~SOCK_NONBLOCK) | O_NONBLOCK;

L
Linus Torvalds 已提交
1323 1324 1325 1326 1327 1328 1329 1330 1331 1332 1333 1334 1335 1336
	/*
	 * 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);
1337
	if (err < 0)
L
Linus Torvalds 已提交
1338 1339
		goto out_release_both;

1340
	fd1 = sock_alloc_fd(&newfile1, flags & O_CLOEXEC);
1341 1342
	if (unlikely(fd1 < 0)) {
		err = fd1;
A
Al Viro 已提交
1343
		goto out_release_both;
1344
	}
L
Linus Torvalds 已提交
1345

1346
	fd2 = sock_alloc_fd(&newfile2, flags & O_CLOEXEC);
A
Al Viro 已提交
1347
	if (unlikely(fd2 < 0)) {
1348
		err = fd2;
A
Al Viro 已提交
1349 1350
		put_filp(newfile1);
		put_unused_fd(fd1);
L
Linus Torvalds 已提交
1351
		goto out_release_both;
A
Al Viro 已提交
1352
	}
L
Linus Torvalds 已提交
1353

1354
	err = sock_attach_fd(sock1, newfile1, flags & O_NONBLOCK);
A
Al Viro 已提交
1355 1356 1357 1358
	if (unlikely(err < 0)) {
		goto out_fd2;
	}

1359
	err = sock_attach_fd(sock2, newfile2, flags & O_NONBLOCK);
A
Al Viro 已提交
1360 1361 1362 1363 1364
	if (unlikely(err < 0)) {
		fput(newfile1);
		goto out_fd1;
	}

A
Al Viro 已提交
1365
	audit_fd_pair(fd1, fd2);
A
Al Viro 已提交
1366 1367
	fd_install(fd1, newfile1);
	fd_install(fd2, newfile2);
L
Linus Torvalds 已提交
1368 1369 1370 1371
	/* fd1 and fd2 may be already another descriptors.
	 * Not kernel problem.
	 */

1372
	err = put_user(fd1, &usockvec[0]);
L
Linus Torvalds 已提交
1373 1374 1375 1376 1377 1378 1379 1380 1381 1382
	if (!err)
		err = put_user(fd2, &usockvec[1]);
	if (!err)
		return 0;

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

out_release_both:
1383
	sock_release(sock2);
L
Linus Torvalds 已提交
1384
out_release_1:
1385
	sock_release(sock1);
L
Linus Torvalds 已提交
1386 1387
out:
	return err;
A
Al Viro 已提交
1388 1389 1390 1391 1392 1393 1394 1395 1396 1397

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
1398 1399 1400 1401 1402 1403 1404 1405 1406 1407
}

/*
 *	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).
 */

1408
SYSCALL_DEFINE3(bind, int, fd, struct sockaddr __user *, umyaddr, int, addrlen)
L
Linus Torvalds 已提交
1409 1410
{
	struct socket *sock;
1411
	struct sockaddr_storage address;
1412
	int err, fput_needed;
L
Linus Torvalds 已提交
1413

1414
	sock = sockfd_lookup_light(fd, &err, &fput_needed);
1415
	if (sock) {
1416
		err = move_addr_to_kernel(umyaddr, addrlen, (struct sockaddr *)&address);
1417 1418
		if (err >= 0) {
			err = security_socket_bind(sock,
1419
						   (struct sockaddr *)&address,
1420
						   addrlen);
1421 1422
			if (!err)
				err = sock->ops->bind(sock,
1423
						      (struct sockaddr *)
1424
						      &address, addrlen);
L
Linus Torvalds 已提交
1425
		}
1426
		fput_light(sock->file, fput_needed);
1427
	}
L
Linus Torvalds 已提交
1428 1429 1430 1431 1432 1433 1434 1435 1436
	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.
 */

1437
SYSCALL_DEFINE2(listen, int, fd, int, backlog)
L
Linus Torvalds 已提交
1438 1439
{
	struct socket *sock;
1440
	int err, fput_needed;
1441
	int somaxconn;
1442 1443 1444

	sock = sockfd_lookup_light(fd, &err, &fput_needed);
	if (sock) {
1445
		somaxconn = sock_net(sock->sk)->core.sysctl_somaxconn;
1446 1447
		if ((unsigned)backlog > somaxconn)
			backlog = somaxconn;
L
Linus Torvalds 已提交
1448 1449

		err = security_socket_listen(sock, backlog);
1450 1451
		if (!err)
			err = sock->ops->listen(sock, backlog);
L
Linus Torvalds 已提交
1452

1453
		fput_light(sock->file, fput_needed);
L
Linus Torvalds 已提交
1454 1455 1456 1457 1458 1459 1460 1461 1462 1463 1464 1465 1466 1467 1468 1469
	}
	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.
 */

1470 1471
SYSCALL_DEFINE4(accept4, int, fd, struct sockaddr __user *, upeer_sockaddr,
		int __user *, upeer_addrlen, int, flags)
L
Linus Torvalds 已提交
1472 1473
{
	struct socket *sock, *newsock;
1474
	struct file *newfile;
1475
	int err, len, newfd, fput_needed;
1476
	struct sockaddr_storage address;
L
Linus Torvalds 已提交
1477

1478
	if (flags & ~(SOCK_CLOEXEC | SOCK_NONBLOCK))
U
Ulrich Drepper 已提交
1479 1480 1481 1482 1483
		return -EINVAL;

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

1484
	sock = sockfd_lookup_light(fd, &err, &fput_needed);
L
Linus Torvalds 已提交
1485 1486 1487 1488
	if (!sock)
		goto out;

	err = -ENFILE;
1489
	if (!(newsock = sock_alloc()))
L
Linus Torvalds 已提交
1490 1491 1492 1493 1494 1495 1496 1497 1498 1499 1500
		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
1501
	newfd = sock_alloc_fd(&newfile, flags & O_CLOEXEC);
1502 1503
	if (unlikely(newfd < 0)) {
		err = newfd;
1504 1505
		sock_release(newsock);
		goto out_put;
1506 1507
	}

1508
	err = sock_attach_fd(newsock, newfile, flags & O_NONBLOCK);
1509
	if (err < 0)
1510
		goto out_fd_simple;
1511

1512 1513
	err = security_socket_accept(sock, newsock);
	if (err)
1514
		goto out_fd;
1515

L
Linus Torvalds 已提交
1516 1517
	err = sock->ops->accept(sock, newsock, sock->file->f_flags);
	if (err < 0)
1518
		goto out_fd;
L
Linus Torvalds 已提交
1519 1520

	if (upeer_sockaddr) {
1521
		if (newsock->ops->getname(newsock, (struct sockaddr *)&address,
1522
					  &len, 2) < 0) {
L
Linus Torvalds 已提交
1523
			err = -ECONNABORTED;
1524
			goto out_fd;
L
Linus Torvalds 已提交
1525
		}
1526 1527
		err = move_addr_to_user((struct sockaddr *)&address,
					len, upeer_sockaddr, upeer_addrlen);
L
Linus Torvalds 已提交
1528
		if (err < 0)
1529
			goto out_fd;
L
Linus Torvalds 已提交
1530 1531 1532 1533
	}

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

1534 1535
	fd_install(newfd, newfile);
	err = newfd;
L
Linus Torvalds 已提交
1536 1537

out_put:
1538
	fput_light(sock->file, fput_needed);
L
Linus Torvalds 已提交
1539 1540
out:
	return err;
1541 1542 1543 1544 1545
out_fd_simple:
	sock_release(newsock);
	put_filp(newfile);
	put_unused_fd(newfd);
	goto out_put;
1546
out_fd:
1547
	fput(newfile);
1548
	put_unused_fd(newfd);
L
Linus Torvalds 已提交
1549 1550 1551
	goto out_put;
}

1552 1553
SYSCALL_DEFINE3(accept, int, fd, struct sockaddr __user *, upeer_sockaddr,
		int __user *, upeer_addrlen)
U
Ulrich Drepper 已提交
1554
{
U
Ulrich Drepper 已提交
1555
	return sys_accept4(fd, upeer_sockaddr, upeer_addrlen, 0);
U
Ulrich Drepper 已提交
1556 1557
}

L
Linus Torvalds 已提交
1558 1559 1560 1561 1562 1563 1564 1565 1566 1567 1568 1569
/*
 *	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.
 */

1570 1571
SYSCALL_DEFINE3(connect, int, fd, struct sockaddr __user *, uservaddr,
		int, addrlen)
L
Linus Torvalds 已提交
1572 1573
{
	struct socket *sock;
1574
	struct sockaddr_storage address;
1575
	int err, fput_needed;
L
Linus Torvalds 已提交
1576

1577
	sock = sockfd_lookup_light(fd, &err, &fput_needed);
L
Linus Torvalds 已提交
1578 1579
	if (!sock)
		goto out;
1580
	err = move_addr_to_kernel(uservaddr, addrlen, (struct sockaddr *)&address);
L
Linus Torvalds 已提交
1581 1582 1583
	if (err < 0)
		goto out_put;

1584
	err =
1585
	    security_socket_connect(sock, (struct sockaddr *)&address, addrlen);
L
Linus Torvalds 已提交
1586 1587 1588
	if (err)
		goto out_put;

1589
	err = sock->ops->connect(sock, (struct sockaddr *)&address, addrlen,
L
Linus Torvalds 已提交
1590 1591
				 sock->file->f_flags);
out_put:
1592
	fput_light(sock->file, fput_needed);
L
Linus Torvalds 已提交
1593 1594 1595 1596 1597 1598 1599 1600 1601
out:
	return err;
}

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

1602 1603
SYSCALL_DEFINE3(getsockname, int, fd, struct sockaddr __user *, usockaddr,
		int __user *, usockaddr_len)
L
Linus Torvalds 已提交
1604 1605
{
	struct socket *sock;
1606
	struct sockaddr_storage address;
1607
	int len, err, fput_needed;
1608

1609
	sock = sockfd_lookup_light(fd, &err, &fput_needed);
L
Linus Torvalds 已提交
1610 1611 1612 1613 1614 1615 1616
	if (!sock)
		goto out;

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

1617
	err = sock->ops->getname(sock, (struct sockaddr *)&address, &len, 0);
L
Linus Torvalds 已提交
1618 1619
	if (err)
		goto out_put;
1620
	err = move_addr_to_user((struct sockaddr *)&address, len, usockaddr, usockaddr_len);
L
Linus Torvalds 已提交
1621 1622

out_put:
1623
	fput_light(sock->file, fput_needed);
L
Linus Torvalds 已提交
1624 1625 1626 1627 1628 1629 1630 1631 1632
out:
	return err;
}

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

1633 1634
SYSCALL_DEFINE3(getpeername, int, fd, struct sockaddr __user *, usockaddr,
		int __user *, usockaddr_len)
L
Linus Torvalds 已提交
1635 1636
{
	struct socket *sock;
1637
	struct sockaddr_storage address;
1638
	int len, err, fput_needed;
L
Linus Torvalds 已提交
1639

1640 1641
	sock = sockfd_lookup_light(fd, &err, &fput_needed);
	if (sock != NULL) {
L
Linus Torvalds 已提交
1642 1643
		err = security_socket_getpeername(sock);
		if (err) {
1644
			fput_light(sock->file, fput_needed);
L
Linus Torvalds 已提交
1645 1646 1647
			return err;
		}

1648
		err =
1649
		    sock->ops->getname(sock, (struct sockaddr *)&address, &len,
1650
				       1);
L
Linus Torvalds 已提交
1651
		if (!err)
1652
			err = move_addr_to_user((struct sockaddr *)&address, len, usockaddr,
1653
						usockaddr_len);
1654
		fput_light(sock->file, fput_needed);
L
Linus Torvalds 已提交
1655 1656 1657 1658 1659 1660 1661 1662 1663 1664
	}
	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.
 */

1665 1666 1667
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
1668 1669
{
	struct socket *sock;
1670
	struct sockaddr_storage address;
L
Linus Torvalds 已提交
1671 1672 1673
	int err;
	struct msghdr msg;
	struct iovec iov;
1674 1675
	int fput_needed;

1676 1677
	sock = sockfd_lookup_light(fd, &err, &fput_needed);
	if (!sock)
1678
		goto out;
1679

1680 1681 1682 1683 1684 1685 1686 1687
	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;
1688
	if (addr) {
1689
		err = move_addr_to_kernel(addr, addr_len, (struct sockaddr *)&address);
L
Linus Torvalds 已提交
1690 1691
		if (err < 0)
			goto out_put;
1692
		msg.msg_name = (struct sockaddr *)&address;
1693
		msg.msg_namelen = addr_len;
L
Linus Torvalds 已提交
1694 1695 1696 1697 1698 1699
	}
	if (sock->file->f_flags & O_NONBLOCK)
		flags |= MSG_DONTWAIT;
	msg.msg_flags = flags;
	err = sock_sendmsg(sock, &msg, len);

1700
out_put:
1701
	fput_light(sock->file, fput_needed);
1702
out:
L
Linus Torvalds 已提交
1703 1704 1705 1706
	return err;
}

/*
1707
 *	Send a datagram down a socket.
L
Linus Torvalds 已提交
1708 1709
 */

1710 1711
SYSCALL_DEFINE4(send, int, fd, void __user *, buff, size_t, len,
		unsigned, flags)
L
Linus Torvalds 已提交
1712 1713 1714 1715 1716
{
	return sys_sendto(fd, buff, len, flags, NULL, 0);
}

/*
1717
 *	Receive a frame from the socket and optionally record the address of the
L
Linus Torvalds 已提交
1718 1719 1720 1721
 *	sender. We verify the buffers are writable and if needed move the
 *	sender address from kernel to user space.
 */

1722 1723 1724
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
1725 1726 1727 1728
{
	struct socket *sock;
	struct iovec iov;
	struct msghdr msg;
1729
	struct sockaddr_storage address;
1730
	int err, err2;
1731 1732
	int fput_needed;

1733
	sock = sockfd_lookup_light(fd, &err, &fput_needed);
L
Linus Torvalds 已提交
1734
	if (!sock)
1735
		goto out;
L
Linus Torvalds 已提交
1736

1737 1738 1739 1740 1741 1742
	msg.msg_control = NULL;
	msg.msg_controllen = 0;
	msg.msg_iovlen = 1;
	msg.msg_iov = &iov;
	iov.iov_len = size;
	iov.iov_base = ubuf;
1743 1744
	msg.msg_name = (struct sockaddr *)&address;
	msg.msg_namelen = sizeof(address);
L
Linus Torvalds 已提交
1745 1746
	if (sock->file->f_flags & O_NONBLOCK)
		flags |= MSG_DONTWAIT;
1747
	err = sock_recvmsg(sock, &msg, size, flags);
L
Linus Torvalds 已提交
1748

1749
	if (err >= 0 && addr != NULL) {
1750 1751
		err2 = move_addr_to_user((struct sockaddr *)&address,
					 msg.msg_namelen, addr, addr_len);
1752 1753
		if (err2 < 0)
			err = err2;
L
Linus Torvalds 已提交
1754
	}
1755 1756

	fput_light(sock->file, fput_needed);
1757
out:
L
Linus Torvalds 已提交
1758 1759 1760 1761
	return err;
}

/*
1762
 *	Receive a datagram from a socket.
L
Linus Torvalds 已提交
1763 1764
 */

1765 1766
asmlinkage long sys_recv(int fd, void __user *ubuf, size_t size,
			 unsigned flags)
L
Linus Torvalds 已提交
1767 1768 1769 1770 1771 1772 1773 1774 1775
{
	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.
 */

1776 1777
SYSCALL_DEFINE5(setsockopt, int, fd, int, level, int, optname,
		char __user *, optval, int, optlen)
L
Linus Torvalds 已提交
1778
{
1779
	int err, fput_needed;
L
Linus Torvalds 已提交
1780 1781 1782 1783
	struct socket *sock;

	if (optlen < 0)
		return -EINVAL;
1784 1785 1786 1787

	sock = sockfd_lookup_light(fd, &err, &fput_needed);
	if (sock != NULL) {
		err = security_socket_setsockopt(sock, level, optname);
1788 1789
		if (err)
			goto out_put;
L
Linus Torvalds 已提交
1790 1791

		if (level == SOL_SOCKET)
1792 1793 1794
			err =
			    sock_setsockopt(sock, level, optname, optval,
					    optlen);
L
Linus Torvalds 已提交
1795
		else
1796 1797 1798
			err =
			    sock->ops->setsockopt(sock, level, optname, optval,
						  optlen);
1799 1800
out_put:
		fput_light(sock->file, fput_needed);
L
Linus Torvalds 已提交
1801 1802 1803 1804 1805 1806 1807 1808 1809
	}
	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.
 */

1810 1811
SYSCALL_DEFINE5(getsockopt, int, fd, int, level, int, optname,
		char __user *, optval, int __user *, optlen)
L
Linus Torvalds 已提交
1812
{
1813
	int err, fput_needed;
L
Linus Torvalds 已提交
1814 1815
	struct socket *sock;

1816 1817
	sock = sockfd_lookup_light(fd, &err, &fput_needed);
	if (sock != NULL) {
1818 1819 1820
		err = security_socket_getsockopt(sock, level, optname);
		if (err)
			goto out_put;
L
Linus Torvalds 已提交
1821 1822

		if (level == SOL_SOCKET)
1823 1824 1825
			err =
			    sock_getsockopt(sock, level, optname, optval,
					    optlen);
L
Linus Torvalds 已提交
1826
		else
1827 1828 1829
			err =
			    sock->ops->getsockopt(sock, level, optname, optval,
						  optlen);
1830 1831
out_put:
		fput_light(sock->file, fput_needed);
L
Linus Torvalds 已提交
1832 1833 1834 1835 1836 1837 1838 1839
	}
	return err;
}

/*
 *	Shutdown a socket.
 */

1840
SYSCALL_DEFINE2(shutdown, int, fd, int, how)
L
Linus Torvalds 已提交
1841
{
1842
	int err, fput_needed;
L
Linus Torvalds 已提交
1843 1844
	struct socket *sock;

1845 1846
	sock = sockfd_lookup_light(fd, &err, &fput_needed);
	if (sock != NULL) {
L
Linus Torvalds 已提交
1847
		err = security_socket_shutdown(sock, how);
1848 1849 1850
		if (!err)
			err = sock->ops->shutdown(sock, how);
		fput_light(sock->file, fput_needed);
L
Linus Torvalds 已提交
1851 1852 1853 1854
	}
	return err;
}

1855
/* A couple of helpful macros for getting the address of the 32/64 bit
L
Linus Torvalds 已提交
1856 1857 1858 1859 1860 1861 1862 1863 1864 1865
 * 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
 */

1866
SYSCALL_DEFINE3(sendmsg, int, fd, struct msghdr __user *, msg, unsigned, flags)
L
Linus Torvalds 已提交
1867
{
1868 1869
	struct compat_msghdr __user *msg_compat =
	    (struct compat_msghdr __user *)msg;
L
Linus Torvalds 已提交
1870
	struct socket *sock;
1871
	struct sockaddr_storage address;
L
Linus Torvalds 已提交
1872
	struct iovec iovstack[UIO_FASTIOV], *iov = iovstack;
1873
	unsigned char ctl[sizeof(struct cmsghdr) + 20]
1874 1875
	    __attribute__ ((aligned(sizeof(__kernel_size_t))));
	/* 20 is size of ipv6_pktinfo */
L
Linus Torvalds 已提交
1876 1877 1878
	unsigned char *ctl_buf = ctl;
	struct msghdr msg_sys;
	int err, ctl_len, iov_size, total_len;
1879
	int fput_needed;
1880

L
Linus Torvalds 已提交
1881 1882 1883 1884
	err = -EFAULT;
	if (MSG_CMSG_COMPAT & flags) {
		if (get_compat_msghdr(&msg_sys, msg_compat))
			return -EFAULT;
1885 1886
	}
	else if (copy_from_user(&msg_sys, msg, sizeof(struct msghdr)))
L
Linus Torvalds 已提交
1887 1888
		return -EFAULT;

1889
	sock = sockfd_lookup_light(fd, &err, &fput_needed);
1890
	if (!sock)
L
Linus Torvalds 已提交
1891 1892 1893 1894 1895 1896 1897
		goto out;

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

1898
	/* Check whether to allocate the iovec area */
L
Linus Torvalds 已提交
1899 1900 1901 1902 1903 1904 1905 1906 1907 1908
	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) {
1909 1910 1911
		err = verify_compat_iovec(&msg_sys, iov,
					  (struct sockaddr *)&address,
					  VERIFY_READ);
L
Linus Torvalds 已提交
1912
	} else
1913 1914 1915
		err = verify_iovec(&msg_sys, iov,
				   (struct sockaddr *)&address,
				   VERIFY_READ);
1916
	if (err < 0)
L
Linus Torvalds 已提交
1917 1918 1919 1920 1921 1922 1923
		goto out_freeiov;
	total_len = err;

	err = -ENOBUFS;

	if (msg_sys.msg_controllen > INT_MAX)
		goto out_freeiov;
1924
	ctl_len = msg_sys.msg_controllen;
L
Linus Torvalds 已提交
1925
	if ((MSG_CMSG_COMPAT & flags) && ctl_len) {
1926 1927 1928
		err =
		    cmsghdr_from_user_compat_to_kern(&msg_sys, sock->sk, ctl,
						     sizeof(ctl));
L
Linus Torvalds 已提交
1929 1930 1931
		if (err)
			goto out_freeiov;
		ctl_buf = msg_sys.msg_control;
A
Al Viro 已提交
1932
		ctl_len = msg_sys.msg_controllen;
L
Linus Torvalds 已提交
1933
	} else if (ctl_len) {
1934
		if (ctl_len > sizeof(ctl)) {
L
Linus Torvalds 已提交
1935
			ctl_buf = sock_kmalloc(sock->sk, ctl_len, GFP_KERNEL);
1936
			if (ctl_buf == NULL)
L
Linus Torvalds 已提交
1937 1938 1939 1940 1941 1942 1943 1944
				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.
		 */
1945 1946
		if (copy_from_user(ctl_buf, (void __user *)msg_sys.msg_control,
				   ctl_len))
L
Linus Torvalds 已提交
1947 1948 1949 1950 1951 1952 1953 1954 1955 1956
			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:
1957
	if (ctl_buf != ctl)
L
Linus Torvalds 已提交
1958 1959 1960 1961 1962
		sock_kfree_s(sock->sk, ctl_buf, ctl_len);
out_freeiov:
	if (iov != iovstack)
		sock_kfree_s(sock->sk, iov, iov_size);
out_put:
1963
	fput_light(sock->file, fput_needed);
1964
out:
L
Linus Torvalds 已提交
1965 1966 1967 1968 1969 1970 1971
	return err;
}

/*
 *	BSD recvmsg interface
 */

1972 1973
SYSCALL_DEFINE3(recvmsg, int, fd, struct msghdr __user *, msg,
		unsigned int, flags)
L
Linus Torvalds 已提交
1974
{
1975 1976
	struct compat_msghdr __user *msg_compat =
	    (struct compat_msghdr __user *)msg;
L
Linus Torvalds 已提交
1977 1978
	struct socket *sock;
	struct iovec iovstack[UIO_FASTIOV];
1979
	struct iovec *iov = iovstack;
L
Linus Torvalds 已提交
1980 1981 1982
	struct msghdr msg_sys;
	unsigned long cmsg_ptr;
	int err, iov_size, total_len, len;
1983
	int fput_needed;
L
Linus Torvalds 已提交
1984 1985

	/* kernel mode address */
1986
	struct sockaddr_storage addr;
L
Linus Torvalds 已提交
1987 1988 1989 1990

	/* user mode address pointers */
	struct sockaddr __user *uaddr;
	int __user *uaddr_len;
1991

L
Linus Torvalds 已提交
1992 1993 1994
	if (MSG_CMSG_COMPAT & flags) {
		if (get_compat_msghdr(&msg_sys, msg_compat))
			return -EFAULT;
1995 1996 1997
	}
	else if (copy_from_user(&msg_sys, msg, sizeof(struct msghdr)))
		return -EFAULT;
L
Linus Torvalds 已提交
1998

1999
	sock = sockfd_lookup_light(fd, &err, &fput_needed);
L
Linus Torvalds 已提交
2000 2001 2002 2003 2004 2005
	if (!sock)
		goto out;

	err = -EMSGSIZE;
	if (msg_sys.msg_iovlen > UIO_MAXIOV)
		goto out_put;
2006 2007

	/* Check whether to allocate the iovec area */
L
Linus Torvalds 已提交
2008 2009 2010 2011 2012 2013 2014 2015 2016
	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;
	}

	/*
2017 2018
	 *      Save the user-mode address (verify_iovec will change the
	 *      kernel msghdr to use the kernel address space)
L
Linus Torvalds 已提交
2019
	 */
2020

S
Stephen Hemminger 已提交
2021
	uaddr = (__force void __user *)msg_sys.msg_name;
L
Linus Torvalds 已提交
2022 2023
	uaddr_len = COMPAT_NAMELEN(msg);
	if (MSG_CMSG_COMPAT & flags) {
2024 2025 2026
		err = verify_compat_iovec(&msg_sys, iov,
					  (struct sockaddr *)&addr,
					  VERIFY_WRITE);
L
Linus Torvalds 已提交
2027
	} else
2028 2029 2030
		err = verify_iovec(&msg_sys, iov,
				   (struct sockaddr *)&addr,
				   VERIFY_WRITE);
L
Linus Torvalds 已提交
2031 2032
	if (err < 0)
		goto out_freeiov;
2033
	total_len = err;
L
Linus Torvalds 已提交
2034 2035

	cmsg_ptr = (unsigned long)msg_sys.msg_control;
U
Ulrich Drepper 已提交
2036
	msg_sys.msg_flags = flags & (MSG_CMSG_CLOEXEC|MSG_CMSG_COMPAT);
2037

L
Linus Torvalds 已提交
2038 2039 2040 2041 2042 2043 2044 2045
	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) {
2046 2047
		err = move_addr_to_user((struct sockaddr *)&addr,
					msg_sys.msg_namelen, uaddr,
2048
					uaddr_len);
L
Linus Torvalds 已提交
2049 2050 2051
		if (err < 0)
			goto out_freeiov;
	}
2052 2053
	err = __put_user((msg_sys.msg_flags & ~MSG_CMSG_COMPAT),
			 COMPAT_FLAGS(msg));
L
Linus Torvalds 已提交
2054 2055 2056
	if (err)
		goto out_freeiov;
	if (MSG_CMSG_COMPAT & flags)
2057
		err = __put_user((unsigned long)msg_sys.msg_control - cmsg_ptr,
L
Linus Torvalds 已提交
2058 2059
				 &msg_compat->msg_controllen);
	else
2060
		err = __put_user((unsigned long)msg_sys.msg_control - cmsg_ptr,
L
Linus Torvalds 已提交
2061 2062 2063 2064 2065 2066 2067 2068 2069
				 &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:
2070
	fput_light(sock->file, fput_needed);
L
Linus Torvalds 已提交
2071 2072 2073 2074 2075 2076 2077 2078
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
2079
static const unsigned char nargs[19]={
2080 2081
	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
2082
	AL(6),AL(2),AL(5),AL(5),AL(3),AL(3),
U
Ulrich Drepper 已提交
2083
	AL(4)
2084 2085
};

L
Linus Torvalds 已提交
2086 2087 2088
#undef AL

/*
2089
 *	System call vectors.
L
Linus Torvalds 已提交
2090 2091 2092
 *
 *	Argument checking cleaned up. Saved 20% in size.
 *  This function doesn't need to set the kernel lock because
2093
 *  it is set by the callees.
L
Linus Torvalds 已提交
2094 2095
 */

2096
SYSCALL_DEFINE2(socketcall, int, call, unsigned long __user *, args)
L
Linus Torvalds 已提交
2097 2098
{
	unsigned long a[6];
2099
	unsigned long a0, a1;
L
Linus Torvalds 已提交
2100
	int err;
2101
	unsigned int len;
L
Linus Torvalds 已提交
2102

U
Ulrich Drepper 已提交
2103
	if (call < 1 || call > SYS_ACCEPT4)
L
Linus Torvalds 已提交
2104 2105
		return -EINVAL;

2106 2107 2108 2109
	len = nargs[call];
	if (len > sizeof(a))
		return -EINVAL;

L
Linus Torvalds 已提交
2110
	/* copy_from_user should be SMP safe. */
2111
	if (copy_from_user(a, args, len))
L
Linus Torvalds 已提交
2112
		return -EFAULT;
2113

A
Al Viro 已提交
2114
	audit_socketcall(nargs[call] / sizeof(unsigned long), a);
2115

2116 2117 2118 2119 2120 2121 2122 2123 2124 2125 2126 2127 2128 2129 2130 2131 2132
	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
2133 2134
		err = sys_accept4(a0, (struct sockaddr __user *)a1,
				  (int __user *)a[2], 0);
2135 2136 2137 2138 2139 2140 2141 2142 2143 2144 2145 2146 2147 2148 2149 2150 2151 2152 2153 2154 2155 2156 2157 2158 2159 2160 2161 2162 2163 2164 2165 2166 2167 2168 2169 2170 2171 2172 2173 2174 2175 2176 2177 2178 2179 2180
		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
2181 2182 2183
	case SYS_ACCEPT4:
		err = sys_accept4(a0, (struct sockaddr __user *)a1,
				  (int __user *)a[2], a[3]);
U
Ulrich Drepper 已提交
2184
		break;
2185 2186 2187
	default:
		err = -EINVAL;
		break;
L
Linus Torvalds 已提交
2188 2189 2190 2191
	}
	return err;
}

2192
#endif				/* __ARCH_WANT_SYS_SOCKETCALL */
L
Linus Torvalds 已提交
2193

2194 2195 2196 2197
/**
 *	sock_register - add a socket protocol handler
 *	@ops: description of protocol
 *
L
Linus Torvalds 已提交
2198 2199
 *	This function is called by a protocol handler that wants to
 *	advertise its address family, and have it linked into the
2200 2201
 *	socket interface. The value ops->family coresponds to the
 *	socket system call protocol family.
L
Linus Torvalds 已提交
2202
 */
2203
int sock_register(const struct net_proto_family *ops)
L
Linus Torvalds 已提交
2204 2205 2206 2207
{
	int err;

	if (ops->family >= NPROTO) {
2208 2209
		printk(KERN_CRIT "protocol %d >= NPROTO(%d)\n", ops->family,
		       NPROTO);
L
Linus Torvalds 已提交
2210 2211
		return -ENOBUFS;
	}
2212 2213 2214 2215 2216

	spin_lock(&net_family_lock);
	if (net_families[ops->family])
		err = -EEXIST;
	else {
2217
		net_families[ops->family] = ops;
L
Linus Torvalds 已提交
2218 2219
		err = 0;
	}
2220 2221
	spin_unlock(&net_family_lock);

2222
	printk(KERN_INFO "NET: Registered protocol family %d\n", ops->family);
L
Linus Torvalds 已提交
2223 2224 2225
	return err;
}

2226 2227 2228 2229
/**
 *	sock_unregister - remove a protocol handler
 *	@family: protocol family to remove
 *
L
Linus Torvalds 已提交
2230 2231
 *	This function is called by a protocol handler that wants to
 *	remove its address family, and have it unlinked from the
2232 2233 2234 2235 2236 2237
 *	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
2238
 */
2239
void sock_unregister(int family)
L
Linus Torvalds 已提交
2240
{
2241
	BUG_ON(family < 0 || family >= NPROTO);
L
Linus Torvalds 已提交
2242

2243
	spin_lock(&net_family_lock);
2244
	net_families[family] = NULL;
2245 2246 2247 2248
	spin_unlock(&net_family_lock);

	synchronize_rcu();

2249
	printk(KERN_INFO "NET: Unregistered protocol family %d\n", family);
L
Linus Torvalds 已提交
2250 2251
}

2252
static int __init sock_init(void)
L
Linus Torvalds 已提交
2253 2254
{
	/*
2255
	 *      Initialize sock SLAB cache.
L
Linus Torvalds 已提交
2256
	 */
2257

L
Linus Torvalds 已提交
2258 2259 2260
	sk_init();

	/*
2261
	 *      Initialize skbuff SLAB cache
L
Linus Torvalds 已提交
2262 2263 2264 2265
	 */
	skb_init();

	/*
2266
	 *      Initialize the protocols module.
L
Linus Torvalds 已提交
2267 2268 2269 2270 2271
	 */

	init_inodecache();
	register_filesystem(&sock_fs_type);
	sock_mnt = kern_mount(&sock_fs_type);
2272 2273

	/* The real protocol initialization is performed in later initcalls.
L
Linus Torvalds 已提交
2274 2275 2276 2277 2278
	 */

#ifdef CONFIG_NETFILTER
	netfilter_init();
#endif
2279 2280

	return 0;
L
Linus Torvalds 已提交
2281 2282
}

2283 2284
core_initcall(sock_init);	/* early initcall */

L
Linus Torvalds 已提交
2285 2286 2287 2288 2289 2290
#ifdef CONFIG_PROC_FS
void socket_seq_show(struct seq_file *seq)
{
	int cpu;
	int counter = 0;

2291
	for_each_possible_cpu(cpu)
2292
	    counter += per_cpu(sockets_in_use, cpu);
L
Linus Torvalds 已提交
2293 2294 2295 2296 2297 2298 2299

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

	seq_printf(seq, "sockets: used %d\n", counter);
}
2300
#endif				/* CONFIG_PROC_FS */
L
Linus Torvalds 已提交
2301

2302 2303
#ifdef CONFIG_COMPAT
static long compat_sock_ioctl(struct file *file, unsigned cmd,
2304
			      unsigned long arg)
2305 2306 2307
{
	struct socket *sock = file->private_data;
	int ret = -ENOIOCTLCMD;
2308 2309 2310 2311 2312
	struct sock *sk;
	struct net *net;

	sk = sock->sk;
	net = sock_net(sk);
2313 2314 2315 2316

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

2317 2318 2319 2320
	if (ret == -ENOIOCTLCMD &&
	    (cmd >= SIOCIWFIRST && cmd <= SIOCIWLAST))
		ret = compat_wext_handle_ioctl(net, cmd, arg);

2321 2322 2323 2324
	return ret;
}
#endif

2325 2326 2327 2328 2329 2330 2331 2332 2333 2334 2335 2336 2337 2338 2339 2340 2341 2342 2343 2344 2345 2346 2347
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);
2348
		*newsock = NULL;
2349 2350 2351 2352
		goto done;
	}

	(*newsock)->ops = sock->ops;
2353
	__module_get((*newsock)->ops->owner);
2354 2355 2356 2357 2358 2359

done:
	return err;
}

int kernel_connect(struct socket *sock, struct sockaddr *addr, int addrlen,
2360
		   int flags)
2361 2362 2363 2364 2365 2366 2367 2368 2369 2370 2371 2372 2373 2374 2375 2376 2377 2378 2379 2380 2381 2382 2383 2384 2385 2386 2387 2388 2389 2390 2391 2392 2393
{
	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,
2394
			char *optval, unsigned int optlen)
2395 2396 2397 2398 2399 2400 2401 2402 2403 2404 2405 2406 2407 2408 2409 2410 2411 2412 2413 2414 2415 2416 2417 2418 2419 2420 2421 2422 2423 2424 2425 2426 2427 2428 2429
{
	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;
}

2430 2431 2432 2433 2434
int kernel_sock_shutdown(struct socket *sock, enum sock_shutdown_cmd how)
{
	return sock->ops->shutdown(sock, how);
}

L
Linus Torvalds 已提交
2435 2436 2437 2438 2439 2440 2441 2442 2443 2444 2445 2446 2447
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);
2448 2449 2450 2451 2452 2453 2454 2455 2456 2457
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);
2458
EXPORT_SYMBOL(kernel_sock_shutdown);