socket.c 81.4 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>
90
#include <linux/slab.h>
L
Linus Torvalds 已提交
91 92 93 94 95

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

#include <net/compat.h>
96
#include <net/wext.h>
97
#include <net/cls_cgroup.h>
L
Linus Torvalds 已提交
98 99 100 101

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

102 103 104 105 106 107
#include <linux/if_tun.h>
#include <linux/ipv6_route.h>
#include <linux/route.h>
#include <linux/sockios.h>
#include <linux/atalk.h>

L
Linus Torvalds 已提交
108
static int sock_no_open(struct inode *irrelevant, struct file *dontcare);
109 110 111 112
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);
113
static int sock_mmap(struct file *file, struct vm_area_struct *vma);
L
Linus Torvalds 已提交
114 115 116 117

static int sock_close(struct inode *inode, struct file *file);
static unsigned int sock_poll(struct file *file,
			      struct poll_table_struct *wait);
118
static long sock_ioctl(struct file *file, unsigned int cmd, unsigned long arg);
119 120
#ifdef CONFIG_COMPAT
static long compat_sock_ioctl(struct file *file,
121
			      unsigned int cmd, unsigned long arg);
122
#endif
L
Linus Torvalds 已提交
123 124 125
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
126
static ssize_t sock_splice_read(struct file *file, loff_t *ppos,
127
				struct pipe_inode_info *pipe, size_t len,
J
Jens Axboe 已提交
128
				unsigned int flags);
L
Linus Torvalds 已提交
129 130 131 132 133 134

/*
 *	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.
 */

135
static const struct file_operations socket_file_ops = {
L
Linus Torvalds 已提交
136 137 138 139 140 141
	.owner =	THIS_MODULE,
	.llseek =	no_llseek,
	.aio_read =	sock_aio_read,
	.aio_write =	sock_aio_write,
	.poll =		sock_poll,
	.unlocked_ioctl = sock_ioctl,
142 143 144
#ifdef CONFIG_COMPAT
	.compat_ioctl = compat_sock_ioctl,
#endif
L
Linus Torvalds 已提交
145 146 147 148
	.mmap =		sock_mmap,
	.open =		sock_no_open,	/* special open code to disallow open via /proc */
	.release =	sock_close,
	.fasync =	sock_fasync,
149 150
	.sendpage =	sock_sendpage,
	.splice_write = generic_splice_sendpage,
J
Jens Axboe 已提交
151
	.splice_read =	sock_splice_read,
L
Linus Torvalds 已提交
152 153 154 155 156 157 158
};

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

static DEFINE_SPINLOCK(net_family_lock);
E
Eric Dumazet 已提交
159
static const struct net_proto_family __rcu *net_families[NPROTO] __read_mostly;
L
Linus Torvalds 已提交
160 161 162 163 164

/*
 *	Statistics counters of the socket lists
 */

165
static DEFINE_PER_CPU(int, sockets_in_use);
L
Linus Torvalds 已提交
166 167

/*
168 169 170
 * 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
171 172 173 174 175 176 177 178 179 180 181 182 183
 */

/**
 *	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_storage *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
static int move_addr_to_user(struct sockaddr_storage *kaddr, int klen,
S
stephen hemminger 已提交
213
			     void __user *uaddr, 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
	 */
	return __put_user(klen, ulen);
}

238
static struct kmem_cache *sock_inode_cachep __read_mostly;
L
Linus Torvalds 已提交
239 240 241 242

static struct inode *sock_alloc_inode(struct super_block *sb)
{
	struct socket_alloc *ei;
243
	struct socket_wq *wq;
244

245
	ei = kmem_cache_alloc(sock_inode_cachep, GFP_KERNEL);
L
Linus Torvalds 已提交
246 247
	if (!ei)
		return NULL;
248 249
	wq = kmalloc(sizeof(*wq), GFP_KERNEL);
	if (!wq) {
250 251 252
		kmem_cache_free(sock_inode_cachep, ei);
		return NULL;
	}
253 254 255
	init_waitqueue_head(&wq->wait);
	wq->fasync_list = NULL;
	RCU_INIT_POINTER(ei->socket.wq, wq);
256

L
Linus Torvalds 已提交
257 258 259 260 261 262 263 264 265 266 267
	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)
{
268
	struct socket_alloc *ei;
269
	struct socket_wq *wq;
270 271

	ei = container_of(inode, struct socket_alloc, vfs_inode);
272
	wq = rcu_dereference_protected(ei->socket.wq, 1);
273
	kfree_rcu(wq, rcu);
274
	kmem_cache_free(sock_inode_cachep, ei);
L
Linus Torvalds 已提交
275 276
}

277
static void init_once(void *foo)
L
Linus Torvalds 已提交
278
{
279
	struct socket_alloc *ei = (struct socket_alloc *)foo;
L
Linus Torvalds 已提交
280

C
Christoph Lameter 已提交
281
	inode_init_once(&ei->vfs_inode);
L
Linus Torvalds 已提交
282
}
283

L
Linus Torvalds 已提交
284 285 286
static int init_inodecache(void)
{
	sock_inode_cachep = kmem_cache_create("sock_inode_cache",
287 288 289 290 291
					      sizeof(struct socket_alloc),
					      0,
					      (SLAB_HWCACHE_ALIGN |
					       SLAB_RECLAIM_ACCOUNT |
					       SLAB_MEM_SPREAD),
292
					      init_once);
L
Linus Torvalds 已提交
293 294 295 296 297
	if (sock_inode_cachep == NULL)
		return -ENOMEM;
	return 0;
}

298
static const struct super_operations sockfs_ops = {
299 300 301
	.alloc_inode	= sock_alloc_inode,
	.destroy_inode	= sock_destroy_inode,
	.statfs		= simple_statfs,
L
Linus Torvalds 已提交
302 303
};

304 305 306 307 308 309 310 311 312
/*
 * 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
313
static const struct dentry_operations sockfs_dentry_operations = {
314
	.d_dname  = sockfs_dname,
L
Linus Torvalds 已提交
315 316
};

317 318 319 320 321 322 323 324 325 326 327 328 329 330 331
static struct dentry *sockfs_mount(struct file_system_type *fs_type,
			 int flags, const char *dev_name, void *data)
{
	return mount_pseudo(fs_type, "socket:", &sockfs_ops,
		&sockfs_dentry_operations, SOCKFS_MAGIC);
}

static struct vfsmount *sock_mnt __read_mostly;

static struct file_system_type sock_fs_type = {
	.name =		"sockfs",
	.mount =	sockfs_mount,
	.kill_sb =	kill_anon_super,
};

L
Linus Torvalds 已提交
332 333 334
/*
 *	Obtains the first available file descriptor and sets it up for use.
 *
335 336
 *	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
337 338 339 340 341 342 343 344 345 346 347 348
 *	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.
 */

349
static int sock_alloc_file(struct socket *sock, struct file **f, int flags)
L
Linus Torvalds 已提交
350
{
351
	struct qstr name = { .name = "" };
352
	struct path path;
353
	struct file *file;
L
Linus Torvalds 已提交
354 355
	int fd;

356
	fd = get_unused_fd_flags(flags);
357 358
	if (unlikely(fd < 0))
		return fd;
L
Linus Torvalds 已提交
359

360
	path.dentry = d_alloc_pseudo(sock_mnt->mnt_sb, &name);
361
	if (unlikely(!path.dentry)) {
362
		put_unused_fd(fd);
363
		return -ENOMEM;
364
	}
365
	path.mnt = mntget(sock_mnt);
366

367
	d_instantiate(path.dentry, SOCK_INODE(sock));
A
Al Viro 已提交
368
	SOCK_INODE(sock)->i_fop = &socket_file_ops;
369

370
	file = alloc_file(&path, FMODE_READ | FMODE_WRITE,
371
		  &socket_file_ops);
A
Al Viro 已提交
372 373
	if (unlikely(!file)) {
		/* drop dentry, keep inode */
A
Al Viro 已提交
374
		ihold(path.dentry->d_inode);
375
		path_put(&path);
A
Al Viro 已提交
376 377 378 379 380
		put_unused_fd(fd);
		return -ENFILE;
	}

	sock->file = file;
381
	file->f_flags = O_RDWR | (flags & O_NONBLOCK);
382 383
	file->f_pos = 0;
	file->private_data = sock;
L
Linus Torvalds 已提交
384

385 386
	*f = file;
	return fd;
387 388
}

389
int sock_map_fd(struct socket *sock, int flags)
390 391
{
	struct file *newfile;
392
	int fd = sock_alloc_file(sock, &newfile, flags);
393

394
	if (likely(fd >= 0))
395
		fd_install(fd, newfile);
396

L
Linus Torvalds 已提交
397 398
	return fd;
}
399
EXPORT_SYMBOL(sock_map_fd);
L
Linus Torvalds 已提交
400

401
struct socket *sock_from_file(struct file *file, int *err)
402 403 404 405
{
	if (file->f_op == &socket_file_ops)
		return file->private_data;	/* set in sock_map_fd */

E
Eric Dumazet 已提交
406 407
	*err = -ENOTSOCK;
	return NULL;
408
}
409
EXPORT_SYMBOL(sock_from_file);
410

L
Linus Torvalds 已提交
411
/**
412
 *	sockfd_lookup - Go from a file number to its socket slot
L
Linus Torvalds 已提交
413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428
 *	@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;

429 430
	file = fget(fd);
	if (!file) {
L
Linus Torvalds 已提交
431 432 433
		*err = -EBADF;
		return NULL;
	}
434

435 436
	sock = sock_from_file(file, err);
	if (!sock)
L
Linus Torvalds 已提交
437
		fput(file);
438 439
	return sock;
}
440
EXPORT_SYMBOL(sockfd_lookup);
L
Linus Torvalds 已提交
441

442 443 444 445 446
static struct socket *sockfd_lookup_light(int fd, int *err, int *fput_needed)
{
	struct file *file;
	struct socket *sock;

447
	*err = -EBADF;
448 449 450 451 452 453
	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
454
	}
455
	return NULL;
L
Linus Torvalds 已提交
456 457 458 459
}

/**
 *	sock_alloc	-	allocate a socket
460
 *
L
Linus Torvalds 已提交
461 462 463 464 465 466 467
 *	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)
{
468 469
	struct inode *inode;
	struct socket *sock;
L
Linus Torvalds 已提交
470

471
	inode = new_inode_pseudo(sock_mnt->mnt_sb);
L
Linus Torvalds 已提交
472 473 474 475 476
	if (!inode)
		return NULL;

	sock = SOCKET_I(inode);

477
	kmemcheck_annotate_bitfield(sock, type);
478
	inode->i_ino = get_next_ino();
479
	inode->i_mode = S_IFSOCK | S_IRWXUGO;
480 481
	inode->i_uid = current_fsuid();
	inode->i_gid = current_fsgid();
L
Linus Torvalds 已提交
482

483
	this_cpu_add(sockets_in_use, 1);
L
Linus Torvalds 已提交
484 485 486 487 488 489 490 491
	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.
 */
492

L
Linus Torvalds 已提交
493 494 495 496 497
static int sock_no_open(struct inode *irrelevant, struct file *dontcare)
{
	return -ENXIO;
}

498
const struct file_operations bad_sock_fops = {
L
Linus Torvalds 已提交
499 500
	.owner = THIS_MODULE,
	.open = sock_no_open,
501
	.llseek = noop_llseek,
L
Linus Torvalds 已提交
502 503 504 505 506 507 508 509
};

/**
 *	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
510
 *	an inode not a file.
L
Linus Torvalds 已提交
511
 */
512

L
Linus Torvalds 已提交
513 514 515 516 517 518 519 520 521 522
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);
	}

523
	if (rcu_dereference_protected(sock->wq, 1)->fasync_list)
L
Linus Torvalds 已提交
524 525
		printk(KERN_ERR "sock_release: fasync list not empty!\n");

526 527 528
	if (test_bit(SOCK_EXTERNALLY_ALLOCATED, &sock->flags))
		return;

529
	this_cpu_sub(sockets_in_use, 1);
L
Linus Torvalds 已提交
530 531 532 533
	if (!sock->file) {
		iput(SOCK_INODE(sock));
		return;
	}
534
	sock->file = NULL;
L
Linus Torvalds 已提交
535
}
536
EXPORT_SYMBOL(sock_release);
L
Linus Torvalds 已提交
537

538
int sock_tx_timestamp(struct sock *sk, __u8 *tx_flags)
539
{
540
	*tx_flags = 0;
541
	if (sock_flag(sk, SOCK_TIMESTAMPING_TX_HARDWARE))
542
		*tx_flags |= SKBTX_HW_TSTAMP;
543
	if (sock_flag(sk, SOCK_TIMESTAMPING_TX_SOFTWARE))
544
		*tx_flags |= SKBTX_SW_TSTAMP;
545 546
	if (sock_flag(sk, SOCK_WIFI_STATUS))
		*tx_flags |= SKBTX_WIFI_STATUS;
547 548 549 550
	return 0;
}
EXPORT_SYMBOL(sock_tx_timestamp);

551 552
static inline int __sock_sendmsg_nosec(struct kiocb *iocb, struct socket *sock,
				       struct msghdr *msg, size_t size)
L
Linus Torvalds 已提交
553 554 555
{
	struct sock_iocb *si = kiocb_to_siocb(iocb);

556 557
	sock_update_classid(sock->sk);

L
Linus Torvalds 已提交
558 559 560 561 562 563 564 565
	si->sock = sock;
	si->scm = NULL;
	si->msg = msg;
	si->size = size;

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

566 567 568 569 570 571 572 573
static inline int __sock_sendmsg(struct kiocb *iocb, struct socket *sock,
				 struct msghdr *msg, size_t size)
{
	int err = security_socket_sendmsg(sock, msg, size);

	return err ?: __sock_sendmsg_nosec(iocb, sock, msg, size);
}

L
Linus Torvalds 已提交
574 575 576 577 578 579 580 581 582 583 584 585 586
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;
}
587
EXPORT_SYMBOL(sock_sendmsg);
L
Linus Torvalds 已提交
588

589
static int sock_sendmsg_nosec(struct socket *sock, struct msghdr *msg, size_t size)
590 591 592 593 594 595 596 597 598 599 600 601 602
{
	struct kiocb iocb;
	struct sock_iocb siocb;
	int ret;

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

L
Linus Torvalds 已提交
603 604 605 606 607 608 609 610 611 612 613
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
	 */
614
	msg->msg_iov = (struct iovec *)vec;
L
Linus Torvalds 已提交
615 616 617 618 619
	msg->msg_iovlen = num;
	result = sock_sendmsg(sock, msg, size);
	set_fs(oldfs);
	return result;
}
620
EXPORT_SYMBOL(kernel_sendmsg);
L
Linus Torvalds 已提交
621

622 623 624 625 626 627 628 629 630 631
static int ktime2ts(ktime_t kt, struct timespec *ts)
{
	if (kt.tv64) {
		*ts = ktime_to_timespec(kt);
		return 1;
	} else {
		return 0;
	}
}

632 633 634 635 636 637
/*
 * 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)
{
638 639 640 641 642 643 644 645 646 647 648 649 650 651 652 653 654 655
	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 {
656
			skb_get_timestampns(skb, &ts[0]);
657
			put_cmsg(msg, SOL_SOCKET, SCM_TIMESTAMPNS,
658
				 sizeof(ts[0]), &ts[0]);
659 660 661 662 663 664 665 666 667 668 669 670 671 672 673 674 675
		}
	}


	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;
676
	}
677 678 679
	if (!empty)
		put_cmsg(msg, SOL_SOCKET,
			 SCM_TIMESTAMPING, sizeof(ts), &ts);
680
}
681 682
EXPORT_SYMBOL_GPL(__sock_recv_timestamp);

683 684 685 686 687 688 689 690 691 692 693 694 695 696 697 698
void __sock_recv_wifi_status(struct msghdr *msg, struct sock *sk,
	struct sk_buff *skb)
{
	int ack;

	if (!sock_flag(sk, SOCK_WIFI_STATUS))
		return;
	if (!skb->wifi_acked_valid)
		return;

	ack = skb->wifi_acked;

	put_cmsg(msg, SOL_SOCKET, SCM_WIFI_STATUS, sizeof(ack), &ack);
}
EXPORT_SYMBOL_GPL(__sock_recv_wifi_status);

S
stephen hemminger 已提交
699 700
static inline void sock_recv_drops(struct msghdr *msg, struct sock *sk,
				   struct sk_buff *skb)
701 702 703 704 705 706
{
	if (sock_flag(sk, SOCK_RXQ_OVFL) && skb && skb->dropcount)
		put_cmsg(msg, SOL_SOCKET, SO_RXQ_OVFL,
			sizeof(__u32), &skb->dropcount);
}

707
void __sock_recv_ts_and_drops(struct msghdr *msg, struct sock *sk,
708 709 710 711 712
	struct sk_buff *skb)
{
	sock_recv_timestamp(msg, sk, skb);
	sock_recv_drops(msg, sk, skb);
}
713
EXPORT_SYMBOL_GPL(__sock_recv_ts_and_drops);
714

715 716
static inline int __sock_recvmsg_nosec(struct kiocb *iocb, struct socket *sock,
				       struct msghdr *msg, size_t size, int flags)
L
Linus Torvalds 已提交
717 718 719
{
	struct sock_iocb *si = kiocb_to_siocb(iocb);

720 721
	sock_update_classid(sock->sk);

L
Linus Torvalds 已提交
722 723 724 725 726 727 728 729 730
	si->sock = sock;
	si->scm = NULL;
	si->msg = msg;
	si->size = size;
	si->flags = flags;

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

731 732 733 734 735 736 737 738
static inline int __sock_recvmsg(struct kiocb *iocb, struct socket *sock,
				 struct msghdr *msg, size_t size, int flags)
{
	int err = security_socket_recvmsg(sock, msg, size, flags);

	return err ?: __sock_recvmsg_nosec(iocb, sock, msg, size, flags);
}

739
int sock_recvmsg(struct socket *sock, struct msghdr *msg,
L
Linus Torvalds 已提交
740 741 742 743 744 745
		 size_t size, int flags)
{
	struct kiocb iocb;
	struct sock_iocb siocb;
	int ret;

746
	init_sync_kiocb(&iocb, NULL);
L
Linus Torvalds 已提交
747 748 749 750 751 752
	iocb.private = &siocb;
	ret = __sock_recvmsg(&iocb, sock, msg, size, flags);
	if (-EIOCBQUEUED == ret)
		ret = wait_on_sync_kiocb(&iocb);
	return ret;
}
753
EXPORT_SYMBOL(sock_recvmsg);
L
Linus Torvalds 已提交
754

755 756 757 758 759 760 761 762 763 764 765 766 767 768 769
static int sock_recvmsg_nosec(struct socket *sock, struct msghdr *msg,
			      size_t size, int flags)
{
	struct kiocb iocb;
	struct sock_iocb siocb;
	int ret;

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

770 771 772 773 774 775 776 777 778 779 780 781 782 783 784
/**
 * kernel_recvmsg - Receive a message from a socket (kernel space)
 * @sock:       The socket to receive the message from
 * @msg:        Received message
 * @vec:        Input s/g array for message data
 * @num:        Size of input s/g array
 * @size:       Number of bytes to read
 * @flags:      Message flags (MSG_DONTWAIT, etc...)
 *
 * On return the msg structure contains the scatter/gather array passed in the
 * vec argument. The array is modified so that it consists of the unfilled
 * portion of the original array.
 *
 * The returned value is the total number of bytes received, or an error.
 */
785 786
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
787 788 789 790 791 792 793 794 795
{
	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
	 */
796
	msg->msg_iov = (struct iovec *)vec, msg->msg_iovlen = num;
L
Linus Torvalds 已提交
797 798 799 800
	result = sock_recvmsg(sock, msg, size, flags);
	set_fs(oldfs);
	return result;
}
801
EXPORT_SYMBOL(kernel_recvmsg);
L
Linus Torvalds 已提交
802 803 804 805 806 807

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

808 809
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
810 811 812 813
{
	struct socket *sock;
	int flags;

814 815
	sock = file->private_data;

816 817 818
	flags = (file->f_flags & O_NONBLOCK) ? MSG_DONTWAIT : 0;
	/* more is a combination of MSG_MORE and MSG_SENDPAGE_NOTLAST */
	flags |= more;
819

820
	return kernel_sendpage(sock, page, offset, size, flags);
821
}
L
Linus Torvalds 已提交
822

J
Jens Axboe 已提交
823
static ssize_t sock_splice_read(struct file *file, loff_t *ppos,
824
				struct pipe_inode_info *pipe, size_t len,
J
Jens Axboe 已提交
825 826 827 828
				unsigned int flags)
{
	struct socket *sock = file->private_data;

829 830 831
	if (unlikely(!sock->ops->splice_read))
		return -EINVAL;

832 833
	sock_update_classid(sock->sk);

J
Jens Axboe 已提交
834 835 836
	return sock->ops->splice_read(sock, ppos, pipe, len, flags);
}

837
static struct sock_iocb *alloc_sock_iocb(struct kiocb *iocb,
838
					 struct sock_iocb *siocb)
839 840 841 842 843
{
	if (!is_sync_kiocb(iocb)) {
		siocb = kmalloc(sizeof(*siocb), GFP_KERNEL);
		if (!siocb)
			return NULL;
L
Linus Torvalds 已提交
844 845 846
		iocb->ki_dtor = sock_aio_dtor;
	}

847 848 849
	siocb->kiocb = iocb;
	iocb->private = siocb;
	return siocb;
L
Linus Torvalds 已提交
850 851
}

852
static ssize_t do_sock_read(struct msghdr *msg, struct kiocb *iocb,
853 854
		struct file *file, const struct iovec *iov,
		unsigned long nr_segs)
855 856 857 858
{
	struct socket *sock = file->private_data;
	size_t size = 0;
	int i;
L
Linus Torvalds 已提交
859

860 861
	for (i = 0; i < nr_segs; i++)
		size += iov[i].iov_len;
L
Linus Torvalds 已提交
862

863 864 865 866
	msg->msg_name = NULL;
	msg->msg_namelen = 0;
	msg->msg_control = NULL;
	msg->msg_controllen = 0;
867
	msg->msg_iov = (struct iovec *)iov;
868 869 870 871 872 873
	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);
}

874 875
static ssize_t sock_aio_read(struct kiocb *iocb, const struct iovec *iov,
				unsigned long nr_segs, loff_t pos)
876 877 878
{
	struct sock_iocb siocb, *x;

L
Linus Torvalds 已提交
879 880
	if (pos != 0)
		return -ESPIPE;
881 882

	if (iocb->ki_left == 0)	/* Match SYS5 behaviour */
L
Linus Torvalds 已提交
883 884
		return 0;

885 886

	x = alloc_sock_iocb(iocb, &siocb);
887 888
	if (!x)
		return -ENOMEM;
889
	return do_sock_read(&x->async_msg, iocb, iocb->ki_filp, iov, nr_segs);
L
Linus Torvalds 已提交
890 891
}

892
static ssize_t do_sock_write(struct msghdr *msg, struct kiocb *iocb,
893 894
			struct file *file, const struct iovec *iov,
			unsigned long nr_segs)
L
Linus Torvalds 已提交
895
{
896 897 898
	struct socket *sock = file->private_data;
	size_t size = 0;
	int i;
L
Linus Torvalds 已提交
899

900 901
	for (i = 0; i < nr_segs; i++)
		size += iov[i].iov_len;
L
Linus Torvalds 已提交
902

903 904 905 906
	msg->msg_name = NULL;
	msg->msg_namelen = 0;
	msg->msg_control = NULL;
	msg->msg_controllen = 0;
907
	msg->msg_iov = (struct iovec *)iov;
908 909 910 911
	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
912

913
	return __sock_sendmsg(iocb, sock, msg, size);
L
Linus Torvalds 已提交
914 915
}

916 917
static ssize_t sock_aio_write(struct kiocb *iocb, const struct iovec *iov,
			  unsigned long nr_segs, loff_t pos)
918 919
{
	struct sock_iocb siocb, *x;
L
Linus Torvalds 已提交
920

921 922
	if (pos != 0)
		return -ESPIPE;
923 924

	x = alloc_sock_iocb(iocb, &siocb);
925 926
	if (!x)
		return -ENOMEM;
L
Linus Torvalds 已提交
927

928
	return do_sock_write(&x->async_msg, iocb, iocb->ki_filp, iov, nr_segs);
L
Linus Torvalds 已提交
929 930 931 932 933 934 935
}

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

A
Arjan van de Ven 已提交
936
static DEFINE_MUTEX(br_ioctl_mutex);
937
static int (*br_ioctl_hook) (struct net *, unsigned int cmd, void __user *arg);
L
Linus Torvalds 已提交
938

939
void brioctl_set(int (*hook) (struct net *, unsigned int, void __user *))
L
Linus Torvalds 已提交
940
{
A
Arjan van de Ven 已提交
941
	mutex_lock(&br_ioctl_mutex);
L
Linus Torvalds 已提交
942
	br_ioctl_hook = hook;
A
Arjan van de Ven 已提交
943
	mutex_unlock(&br_ioctl_mutex);
L
Linus Torvalds 已提交
944 945 946
}
EXPORT_SYMBOL(brioctl_set);

A
Arjan van de Ven 已提交
947
static DEFINE_MUTEX(vlan_ioctl_mutex);
948
static int (*vlan_ioctl_hook) (struct net *, void __user *arg);
L
Linus Torvalds 已提交
949

950
void vlan_ioctl_set(int (*hook) (struct net *, void __user *))
L
Linus Torvalds 已提交
951
{
A
Arjan van de Ven 已提交
952
	mutex_lock(&vlan_ioctl_mutex);
L
Linus Torvalds 已提交
953
	vlan_ioctl_hook = hook;
A
Arjan van de Ven 已提交
954
	mutex_unlock(&vlan_ioctl_mutex);
L
Linus Torvalds 已提交
955 956 957
}
EXPORT_SYMBOL(vlan_ioctl_set);

A
Arjan van de Ven 已提交
958
static DEFINE_MUTEX(dlci_ioctl_mutex);
959
static int (*dlci_ioctl_hook) (unsigned int, void __user *);
L
Linus Torvalds 已提交
960

961
void dlci_ioctl_set(int (*hook) (unsigned int, void __user *))
L
Linus Torvalds 已提交
962
{
A
Arjan van de Ven 已提交
963
	mutex_lock(&dlci_ioctl_mutex);
L
Linus Torvalds 已提交
964
	dlci_ioctl_hook = hook;
A
Arjan van de Ven 已提交
965
	mutex_unlock(&dlci_ioctl_mutex);
L
Linus Torvalds 已提交
966 967 968
}
EXPORT_SYMBOL(dlci_ioctl_set);

969 970 971 972 973 974 975 976 977 978 979 980 981 982 983 984 985 986
static long sock_do_ioctl(struct net *net, struct socket *sock,
				 unsigned int cmd, unsigned long arg)
{
	int err;
	void __user *argp = (void __user *)arg;

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

	/*
	 * If this ioctl is unknown try to hand it down
	 * to the NIC driver.
	 */
	if (err == -ENOIOCTLCMD)
		err = dev_ioctl(net, cmd, argp);

	return err;
}

L
Linus Torvalds 已提交
987 988 989 990 991 992 993 994
/*
 *	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;
995
	struct sock *sk;
L
Linus Torvalds 已提交
996 997
	void __user *argp = (void __user *)arg;
	int pid, err;
998
	struct net *net;
L
Linus Torvalds 已提交
999

1000
	sock = file->private_data;
1001
	sk = sock->sk;
1002
	net = sock_net(sk);
L
Linus Torvalds 已提交
1003
	if (cmd >= SIOCDEVPRIVATE && cmd <= (SIOCDEVPRIVATE + 15)) {
1004
		err = dev_ioctl(net, cmd, argp);
L
Linus Torvalds 已提交
1005
	} else
J
Johannes Berg 已提交
1006
#ifdef CONFIG_WEXT_CORE
L
Linus Torvalds 已提交
1007
	if (cmd >= SIOCIWFIRST && cmd <= SIOCIWLAST) {
1008
		err = dev_ioctl(net, cmd, argp);
L
Linus Torvalds 已提交
1009
	} else
J
Johannes Berg 已提交
1010
#endif
1011
		switch (cmd) {
L
Linus Torvalds 已提交
1012 1013 1014 1015 1016 1017 1018 1019 1020
		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:
1021
			err = put_user(f_getown(sock->file),
1022
				       (int __user *)argp);
L
Linus Torvalds 已提交
1023 1024 1025 1026 1027 1028 1029 1030 1031
			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
1032
			mutex_lock(&br_ioctl_mutex);
1033
			if (br_ioctl_hook)
1034
				err = br_ioctl_hook(net, cmd, argp);
A
Arjan van de Ven 已提交
1035
			mutex_unlock(&br_ioctl_mutex);
L
Linus Torvalds 已提交
1036 1037 1038 1039 1040 1041 1042
			break;
		case SIOCGIFVLAN:
		case SIOCSIFVLAN:
			err = -ENOPKG;
			if (!vlan_ioctl_hook)
				request_module("8021q");

A
Arjan van de Ven 已提交
1043
			mutex_lock(&vlan_ioctl_mutex);
L
Linus Torvalds 已提交
1044
			if (vlan_ioctl_hook)
1045
				err = vlan_ioctl_hook(net, argp);
A
Arjan van de Ven 已提交
1046
			mutex_unlock(&vlan_ioctl_mutex);
L
Linus Torvalds 已提交
1047 1048 1049 1050 1051 1052 1053
			break;
		case SIOCADDDLCI:
		case SIOCDELDLCI:
			err = -ENOPKG;
			if (!dlci_ioctl_hook)
				request_module("dlci");

1054 1055
			mutex_lock(&dlci_ioctl_mutex);
			if (dlci_ioctl_hook)
L
Linus Torvalds 已提交
1056
				err = dlci_ioctl_hook(cmd, argp);
1057
			mutex_unlock(&dlci_ioctl_mutex);
L
Linus Torvalds 已提交
1058 1059
			break;
		default:
1060
			err = sock_do_ioctl(net, sock, cmd, arg);
L
Linus Torvalds 已提交
1061
			break;
1062
		}
L
Linus Torvalds 已提交
1063 1064 1065 1066 1067 1068 1069
	return err;
}

int sock_create_lite(int family, int type, int protocol, struct socket **res)
{
	int err;
	struct socket *sock = NULL;
1070

L
Linus Torvalds 已提交
1071 1072 1073 1074 1075 1076 1077 1078 1079 1080 1081
	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
1082 1083 1084 1085
	err = security_socket_post_create(sock, family, type, protocol, 1);
	if (err)
		goto out_release;

L
Linus Torvalds 已提交
1086 1087 1088
out:
	*res = sock;
	return err;
V
Venkat Yekkirala 已提交
1089 1090 1091 1092
out_release:
	sock_release(sock);
	sock = NULL;
	goto out;
L
Linus Torvalds 已提交
1093
}
1094
EXPORT_SYMBOL(sock_create_lite);
L
Linus Torvalds 已提交
1095 1096

/* No kernel lock held - perfect */
1097
static unsigned int sock_poll(struct file *file, poll_table *wait)
L
Linus Torvalds 已提交
1098 1099 1100 1101
{
	struct socket *sock;

	/*
1102
	 *      We can't return errors to poll, so it's either yes or no.
L
Linus Torvalds 已提交
1103
	 */
1104
	sock = file->private_data;
L
Linus Torvalds 已提交
1105 1106 1107
	return sock->ops->poll(file, sock, wait);
}

1108
static int sock_mmap(struct file *file, struct vm_area_struct *vma)
L
Linus Torvalds 已提交
1109
{
1110
	struct socket *sock = file->private_data;
L
Linus Torvalds 已提交
1111 1112 1113 1114

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

1115
static int sock_close(struct inode *inode, struct file *filp)
L
Linus Torvalds 已提交
1116 1117
{
	/*
1118 1119
	 *      It was possible the inode is NULL we were
	 *      closing an unfinished socket.
L
Linus Torvalds 已提交
1120 1121
	 */

1122
	if (!inode) {
L
Linus Torvalds 已提交
1123 1124 1125 1126 1127 1128 1129 1130 1131 1132 1133 1134 1135 1136 1137
		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)
1138
 *	   or under socket lock
L
Linus Torvalds 已提交
1139 1140 1141 1142
 */

static int sock_fasync(int fd, struct file *filp, int on)
{
1143 1144
	struct socket *sock = filp->private_data;
	struct sock *sk = sock->sk;
1145
	struct socket_wq *wq;
L
Linus Torvalds 已提交
1146

1147
	if (sk == NULL)
L
Linus Torvalds 已提交
1148 1149 1150
		return -EINVAL;

	lock_sock(sk);
1151 1152
	wq = rcu_dereference_protected(sock->wq, sock_owned_by_user(sk));
	fasync_helper(fd, filp, on, &wq->fasync_list);
L
Linus Torvalds 已提交
1153

1154
	if (!wq->fasync_list)
1155 1156
		sock_reset_flag(sk, SOCK_FASYNC);
	else
E
Eric Dumazet 已提交
1157
		sock_set_flag(sk, SOCK_FASYNC);
L
Linus Torvalds 已提交
1158

1159
	release_sock(sk);
L
Linus Torvalds 已提交
1160 1161 1162
	return 0;
}

1163
/* This function may be called only under socket lock or callback_lock or rcu_lock */
L
Linus Torvalds 已提交
1164 1165 1166

int sock_wake_async(struct socket *sock, int how, int band)
{
1167 1168 1169 1170 1171 1172 1173 1174
	struct socket_wq *wq;

	if (!sock)
		return -1;
	rcu_read_lock();
	wq = rcu_dereference(sock->wq);
	if (!wq || !wq->fasync_list) {
		rcu_read_unlock();
L
Linus Torvalds 已提交
1175
		return -1;
1176
	}
1177
	switch (how) {
1178
	case SOCK_WAKE_WAITD:
L
Linus Torvalds 已提交
1179 1180 1181
		if (test_bit(SOCK_ASYNC_WAITDATA, &sock->flags))
			break;
		goto call_kill;
1182
	case SOCK_WAKE_SPACE:
L
Linus Torvalds 已提交
1183 1184 1185
		if (!test_and_clear_bit(SOCK_ASYNC_NOSPACE, &sock->flags))
			break;
		/* fall through */
1186
	case SOCK_WAKE_IO:
1187
call_kill:
1188
		kill_fasync(&wq->fasync_list, SIGIO, band);
L
Linus Torvalds 已提交
1189
		break;
1190
	case SOCK_WAKE_URG:
1191
		kill_fasync(&wq->fasync_list, SIGURG, band);
L
Linus Torvalds 已提交
1192
	}
1193
	rcu_read_unlock();
L
Linus Torvalds 已提交
1194 1195
	return 0;
}
1196
EXPORT_SYMBOL(sock_wake_async);
L
Linus Torvalds 已提交
1197

P
Pavel Emelyanov 已提交
1198
int __sock_create(struct net *net, int family, int type, int protocol,
1199
			 struct socket **res, int kern)
L
Linus Torvalds 已提交
1200 1201 1202
{
	int err;
	struct socket *sock;
1203
	const struct net_proto_family *pf;
L
Linus Torvalds 已提交
1204 1205

	/*
1206
	 *      Check protocol is in range
L
Linus Torvalds 已提交
1207 1208 1209 1210 1211 1212 1213 1214 1215 1216 1217 1218
	 */
	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) {
1219
		static int warned;
L
Linus Torvalds 已提交
1220 1221
		if (!warned) {
			warned = 1;
1222 1223
			printk(KERN_INFO "%s uses obsolete (PF_INET,SOCK_PACKET)\n",
			       current->comm);
L
Linus Torvalds 已提交
1224 1225 1226 1227 1228 1229 1230
		}
		family = PF_PACKET;
	}

	err = security_socket_create(family, type, protocol, kern);
	if (err)
		return err;
1231

1232 1233 1234 1235 1236 1237 1238
	/*
	 *	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) {
1239
		net_warn_ratelimited("socket: no more sockets\n");
1240 1241 1242 1243 1244 1245
		return -ENFILE;	/* Not exactly a match, but its the
				   closest posix thing */
	}

	sock->type = type;

1246
#ifdef CONFIG_MODULES
1247 1248 1249
	/* 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
1250 1251 1252
	 * requested real, full-featured networking support upon configuration.
	 * Otherwise module support will break!
	 */
E
Eric Dumazet 已提交
1253
	if (rcu_access_pointer(net_families[family]) == NULL)
1254
		request_module("net-pf-%d", family);
L
Linus Torvalds 已提交
1255 1256
#endif

1257 1258 1259 1260 1261
	rcu_read_lock();
	pf = rcu_dereference(net_families[family]);
	err = -EAFNOSUPPORT;
	if (!pf)
		goto out_release;
L
Linus Torvalds 已提交
1262 1263 1264 1265 1266

	/*
	 * We will call the ->create function, that possibly is in a loadable
	 * module, so we have to bump that loadable module refcnt first.
	 */
1267
	if (!try_module_get(pf->owner))
L
Linus Torvalds 已提交
1268 1269
		goto out_release;

1270 1271 1272
	/* Now protected by module ref count */
	rcu_read_unlock();

1273
	err = pf->create(net, sock, protocol, kern);
1274
	if (err < 0)
L
Linus Torvalds 已提交
1275
		goto out_module_put;
1276

L
Linus Torvalds 已提交
1277 1278 1279 1280
	/*
	 * Now to bump the refcnt of the [loadable] module that owns this
	 * socket at sock_release time we decrement its refcnt.
	 */
1281 1282 1283
	if (!try_module_get(sock->ops->owner))
		goto out_module_busy;

L
Linus Torvalds 已提交
1284 1285 1286 1287
	/*
	 * Now that we're done with the ->create function, the [loadable]
	 * module can have its refcnt decremented
	 */
1288
	module_put(pf->owner);
V
Venkat Yekkirala 已提交
1289 1290
	err = security_socket_post_create(sock, family, type, protocol, kern);
	if (err)
1291
		goto out_sock_release;
1292
	*res = sock;
L
Linus Torvalds 已提交
1293

1294 1295 1296 1297
	return 0;

out_module_busy:
	err = -EAFNOSUPPORT;
L
Linus Torvalds 已提交
1298
out_module_put:
1299 1300 1301
	sock->ops = NULL;
	module_put(pf->owner);
out_sock_release:
L
Linus Torvalds 已提交
1302
	sock_release(sock);
1303 1304 1305 1306 1307
	return err;

out_release:
	rcu_read_unlock();
	goto out_sock_release;
L
Linus Torvalds 已提交
1308
}
P
Pavel Emelyanov 已提交
1309
EXPORT_SYMBOL(__sock_create);
L
Linus Torvalds 已提交
1310 1311 1312

int sock_create(int family, int type, int protocol, struct socket **res)
{
1313
	return __sock_create(current->nsproxy->net_ns, family, type, protocol, res, 0);
L
Linus Torvalds 已提交
1314
}
1315
EXPORT_SYMBOL(sock_create);
L
Linus Torvalds 已提交
1316 1317 1318

int sock_create_kern(int family, int type, int protocol, struct socket **res)
{
1319
	return __sock_create(&init_net, family, type, protocol, res, 1);
L
Linus Torvalds 已提交
1320
}
1321
EXPORT_SYMBOL(sock_create_kern);
L
Linus Torvalds 已提交
1322

1323
SYSCALL_DEFINE3(socket, int, family, int, type, int, protocol)
L
Linus Torvalds 已提交
1324 1325 1326
{
	int retval;
	struct socket *sock;
1327 1328
	int flags;

1329 1330 1331 1332 1333 1334
	/* 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);

1335
	flags = type & ~SOCK_TYPE_MASK;
1336
	if (flags & ~(SOCK_CLOEXEC | SOCK_NONBLOCK))
1337 1338
		return -EINVAL;
	type &= SOCK_TYPE_MASK;
L
Linus Torvalds 已提交
1339

U
Ulrich Drepper 已提交
1340 1341 1342
	if (SOCK_NONBLOCK != O_NONBLOCK && (flags & SOCK_NONBLOCK))
		flags = (flags & ~SOCK_NONBLOCK) | O_NONBLOCK;

L
Linus Torvalds 已提交
1343 1344 1345 1346
	retval = sock_create(family, type, protocol, &sock);
	if (retval < 0)
		goto out;

1347
	retval = sock_map_fd(sock, flags & (O_CLOEXEC | O_NONBLOCK));
L
Linus Torvalds 已提交
1348 1349 1350 1351 1352 1353 1354 1355 1356 1357 1358 1359 1360 1361 1362 1363
	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.
 */

1364 1365
SYSCALL_DEFINE4(socketpair, int, family, int, type, int, protocol,
		int __user *, usockvec)
L
Linus Torvalds 已提交
1366 1367 1368
{
	struct socket *sock1, *sock2;
	int fd1, fd2, err;
A
Al Viro 已提交
1369
	struct file *newfile1, *newfile2;
1370 1371 1372
	int flags;

	flags = type & ~SOCK_TYPE_MASK;
1373
	if (flags & ~(SOCK_CLOEXEC | SOCK_NONBLOCK))
1374 1375
		return -EINVAL;
	type &= SOCK_TYPE_MASK;
L
Linus Torvalds 已提交
1376

U
Ulrich Drepper 已提交
1377 1378 1379
	if (SOCK_NONBLOCK != O_NONBLOCK && (flags & SOCK_NONBLOCK))
		flags = (flags & ~SOCK_NONBLOCK) | O_NONBLOCK;

L
Linus Torvalds 已提交
1380 1381 1382 1383 1384 1385 1386 1387 1388 1389 1390 1391 1392 1393
	/*
	 * 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);
1394
	if (err < 0)
L
Linus Torvalds 已提交
1395 1396
		goto out_release_both;

1397
	fd1 = sock_alloc_file(sock1, &newfile1, flags);
1398 1399
	if (unlikely(fd1 < 0)) {
		err = fd1;
A
Al Viro 已提交
1400
		goto out_release_both;
1401
	}
L
Linus Torvalds 已提交
1402

1403
	fd2 = sock_alloc_file(sock2, &newfile2, flags);
1404 1405 1406 1407 1408 1409
	if (unlikely(fd2 < 0)) {
		err = fd2;
		fput(newfile1);
		put_unused_fd(fd1);
		sock_release(sock2);
		goto out;
A
Al Viro 已提交
1410 1411
	}

A
Al Viro 已提交
1412
	audit_fd_pair(fd1, fd2);
A
Al Viro 已提交
1413 1414
	fd_install(fd1, newfile1);
	fd_install(fd2, newfile2);
L
Linus Torvalds 已提交
1415 1416 1417 1418
	/* fd1 and fd2 may be already another descriptors.
	 * Not kernel problem.
	 */

1419
	err = put_user(fd1, &usockvec[0]);
L
Linus Torvalds 已提交
1420 1421 1422 1423 1424 1425 1426 1427 1428 1429
	if (!err)
		err = put_user(fd2, &usockvec[1]);
	if (!err)
		return 0;

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

out_release_both:
1430
	sock_release(sock2);
L
Linus Torvalds 已提交
1431
out_release_1:
1432
	sock_release(sock1);
L
Linus Torvalds 已提交
1433 1434 1435 1436 1437 1438 1439 1440 1441 1442 1443 1444
out:
	return err;
}

/*
 *	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).
 */

1445
SYSCALL_DEFINE3(bind, int, fd, struct sockaddr __user *, umyaddr, int, addrlen)
L
Linus Torvalds 已提交
1446 1447
{
	struct socket *sock;
1448
	struct sockaddr_storage address;
1449
	int err, fput_needed;
L
Linus Torvalds 已提交
1450

1451
	sock = sockfd_lookup_light(fd, &err, &fput_needed);
1452
	if (sock) {
1453
		err = move_addr_to_kernel(umyaddr, addrlen, &address);
1454 1455
		if (err >= 0) {
			err = security_socket_bind(sock,
1456
						   (struct sockaddr *)&address,
1457
						   addrlen);
1458 1459
			if (!err)
				err = sock->ops->bind(sock,
1460
						      (struct sockaddr *)
1461
						      &address, addrlen);
L
Linus Torvalds 已提交
1462
		}
1463
		fput_light(sock->file, fput_needed);
1464
	}
L
Linus Torvalds 已提交
1465 1466 1467 1468 1469 1470 1471 1472 1473
	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.
 */

1474
SYSCALL_DEFINE2(listen, int, fd, int, backlog)
L
Linus Torvalds 已提交
1475 1476
{
	struct socket *sock;
1477
	int err, fput_needed;
1478
	int somaxconn;
1479 1480 1481

	sock = sockfd_lookup_light(fd, &err, &fput_needed);
	if (sock) {
1482
		somaxconn = sock_net(sock->sk)->core.sysctl_somaxconn;
1483
		if ((unsigned int)backlog > somaxconn)
1484
			backlog = somaxconn;
L
Linus Torvalds 已提交
1485 1486

		err = security_socket_listen(sock, backlog);
1487 1488
		if (!err)
			err = sock->ops->listen(sock, backlog);
L
Linus Torvalds 已提交
1489

1490
		fput_light(sock->file, fput_needed);
L
Linus Torvalds 已提交
1491 1492 1493 1494 1495 1496 1497 1498 1499 1500 1501 1502 1503 1504 1505 1506
	}
	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.
 */

1507 1508
SYSCALL_DEFINE4(accept4, int, fd, struct sockaddr __user *, upeer_sockaddr,
		int __user *, upeer_addrlen, int, flags)
L
Linus Torvalds 已提交
1509 1510
{
	struct socket *sock, *newsock;
1511
	struct file *newfile;
1512
	int err, len, newfd, fput_needed;
1513
	struct sockaddr_storage address;
L
Linus Torvalds 已提交
1514

1515
	if (flags & ~(SOCK_CLOEXEC | SOCK_NONBLOCK))
U
Ulrich Drepper 已提交
1516 1517 1518 1519 1520
		return -EINVAL;

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

1521
	sock = sockfd_lookup_light(fd, &err, &fput_needed);
L
Linus Torvalds 已提交
1522 1523 1524 1525
	if (!sock)
		goto out;

	err = -ENFILE;
1526 1527
	newsock = sock_alloc();
	if (!newsock)
L
Linus Torvalds 已提交
1528 1529 1530 1531 1532 1533 1534 1535 1536 1537 1538
		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);

1539
	newfd = sock_alloc_file(newsock, &newfile, flags);
1540 1541
	if (unlikely(newfd < 0)) {
		err = newfd;
1542 1543
		sock_release(newsock);
		goto out_put;
1544 1545
	}

1546 1547
	err = security_socket_accept(sock, newsock);
	if (err)
1548
		goto out_fd;
1549

L
Linus Torvalds 已提交
1550 1551
	err = sock->ops->accept(sock, newsock, sock->file->f_flags);
	if (err < 0)
1552
		goto out_fd;
L
Linus Torvalds 已提交
1553 1554

	if (upeer_sockaddr) {
1555
		if (newsock->ops->getname(newsock, (struct sockaddr *)&address,
1556
					  &len, 2) < 0) {
L
Linus Torvalds 已提交
1557
			err = -ECONNABORTED;
1558
			goto out_fd;
L
Linus Torvalds 已提交
1559
		}
1560
		err = move_addr_to_user(&address,
1561
					len, upeer_sockaddr, upeer_addrlen);
L
Linus Torvalds 已提交
1562
		if (err < 0)
1563
			goto out_fd;
L
Linus Torvalds 已提交
1564 1565 1566 1567
	}

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

1568 1569
	fd_install(newfd, newfile);
	err = newfd;
L
Linus Torvalds 已提交
1570 1571

out_put:
1572
	fput_light(sock->file, fput_needed);
L
Linus Torvalds 已提交
1573 1574
out:
	return err;
1575
out_fd:
1576
	fput(newfile);
1577
	put_unused_fd(newfd);
L
Linus Torvalds 已提交
1578 1579 1580
	goto out_put;
}

1581 1582
SYSCALL_DEFINE3(accept, int, fd, struct sockaddr __user *, upeer_sockaddr,
		int __user *, upeer_addrlen)
U
Ulrich Drepper 已提交
1583
{
U
Ulrich Drepper 已提交
1584
	return sys_accept4(fd, upeer_sockaddr, upeer_addrlen, 0);
U
Ulrich Drepper 已提交
1585 1586
}

L
Linus Torvalds 已提交
1587 1588 1589 1590 1591 1592 1593 1594 1595 1596 1597 1598
/*
 *	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.
 */

1599 1600
SYSCALL_DEFINE3(connect, int, fd, struct sockaddr __user *, uservaddr,
		int, addrlen)
L
Linus Torvalds 已提交
1601 1602
{
	struct socket *sock;
1603
	struct sockaddr_storage address;
1604
	int err, fput_needed;
L
Linus Torvalds 已提交
1605

1606
	sock = sockfd_lookup_light(fd, &err, &fput_needed);
L
Linus Torvalds 已提交
1607 1608
	if (!sock)
		goto out;
1609
	err = move_addr_to_kernel(uservaddr, addrlen, &address);
L
Linus Torvalds 已提交
1610 1611 1612
	if (err < 0)
		goto out_put;

1613
	err =
1614
	    security_socket_connect(sock, (struct sockaddr *)&address, addrlen);
L
Linus Torvalds 已提交
1615 1616 1617
	if (err)
		goto out_put;

1618
	err = sock->ops->connect(sock, (struct sockaddr *)&address, addrlen,
L
Linus Torvalds 已提交
1619 1620
				 sock->file->f_flags);
out_put:
1621
	fput_light(sock->file, fput_needed);
L
Linus Torvalds 已提交
1622 1623 1624 1625 1626 1627 1628 1629 1630
out:
	return err;
}

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

1631 1632
SYSCALL_DEFINE3(getsockname, int, fd, struct sockaddr __user *, usockaddr,
		int __user *, usockaddr_len)
L
Linus Torvalds 已提交
1633 1634
{
	struct socket *sock;
1635
	struct sockaddr_storage address;
1636
	int len, err, fput_needed;
1637

1638
	sock = sockfd_lookup_light(fd, &err, &fput_needed);
L
Linus Torvalds 已提交
1639 1640 1641 1642 1643 1644 1645
	if (!sock)
		goto out;

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

1646
	err = sock->ops->getname(sock, (struct sockaddr *)&address, &len, 0);
L
Linus Torvalds 已提交
1647 1648
	if (err)
		goto out_put;
1649
	err = move_addr_to_user(&address, len, usockaddr, usockaddr_len);
L
Linus Torvalds 已提交
1650 1651

out_put:
1652
	fput_light(sock->file, fput_needed);
L
Linus Torvalds 已提交
1653 1654 1655 1656 1657 1658 1659 1660 1661
out:
	return err;
}

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

1662 1663
SYSCALL_DEFINE3(getpeername, int, fd, struct sockaddr __user *, usockaddr,
		int __user *, usockaddr_len)
L
Linus Torvalds 已提交
1664 1665
{
	struct socket *sock;
1666
	struct sockaddr_storage address;
1667
	int len, err, fput_needed;
L
Linus Torvalds 已提交
1668

1669 1670
	sock = sockfd_lookup_light(fd, &err, &fput_needed);
	if (sock != NULL) {
L
Linus Torvalds 已提交
1671 1672
		err = security_socket_getpeername(sock);
		if (err) {
1673
			fput_light(sock->file, fput_needed);
L
Linus Torvalds 已提交
1674 1675 1676
			return err;
		}

1677
		err =
1678
		    sock->ops->getname(sock, (struct sockaddr *)&address, &len,
1679
				       1);
L
Linus Torvalds 已提交
1680
		if (!err)
1681
			err = move_addr_to_user(&address, len, usockaddr,
1682
						usockaddr_len);
1683
		fput_light(sock->file, fput_needed);
L
Linus Torvalds 已提交
1684 1685 1686 1687 1688 1689 1690 1691 1692 1693
	}
	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.
 */

1694
SYSCALL_DEFINE6(sendto, int, fd, void __user *, buff, size_t, len,
1695
		unsigned int, flags, struct sockaddr __user *, addr,
1696
		int, addr_len)
L
Linus Torvalds 已提交
1697 1698
{
	struct socket *sock;
1699
	struct sockaddr_storage address;
L
Linus Torvalds 已提交
1700 1701 1702
	int err;
	struct msghdr msg;
	struct iovec iov;
1703 1704
	int fput_needed;

1705 1706
	if (len > INT_MAX)
		len = INT_MAX;
1707 1708
	sock = sockfd_lookup_light(fd, &err, &fput_needed);
	if (!sock)
1709
		goto out;
1710

1711 1712 1713 1714 1715 1716 1717 1718
	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;
1719
	if (addr) {
1720
		err = move_addr_to_kernel(addr, addr_len, &address);
L
Linus Torvalds 已提交
1721 1722
		if (err < 0)
			goto out_put;
1723
		msg.msg_name = (struct sockaddr *)&address;
1724
		msg.msg_namelen = addr_len;
L
Linus Torvalds 已提交
1725 1726 1727 1728 1729 1730
	}
	if (sock->file->f_flags & O_NONBLOCK)
		flags |= MSG_DONTWAIT;
	msg.msg_flags = flags;
	err = sock_sendmsg(sock, &msg, len);

1731
out_put:
1732
	fput_light(sock->file, fput_needed);
1733
out:
L
Linus Torvalds 已提交
1734 1735 1736 1737
	return err;
}

/*
1738
 *	Send a datagram down a socket.
L
Linus Torvalds 已提交
1739 1740
 */

1741
SYSCALL_DEFINE4(send, int, fd, void __user *, buff, size_t, len,
1742
		unsigned int, flags)
L
Linus Torvalds 已提交
1743 1744 1745 1746 1747
{
	return sys_sendto(fd, buff, len, flags, NULL, 0);
}

/*
1748
 *	Receive a frame from the socket and optionally record the address of the
L
Linus Torvalds 已提交
1749 1750 1751 1752
 *	sender. We verify the buffers are writable and if needed move the
 *	sender address from kernel to user space.
 */

1753
SYSCALL_DEFINE6(recvfrom, int, fd, void __user *, ubuf, size_t, size,
1754
		unsigned int, flags, struct sockaddr __user *, addr,
1755
		int __user *, addr_len)
L
Linus Torvalds 已提交
1756 1757 1758 1759
{
	struct socket *sock;
	struct iovec iov;
	struct msghdr msg;
1760
	struct sockaddr_storage address;
1761
	int err, err2;
1762 1763
	int fput_needed;

1764 1765
	if (size > INT_MAX)
		size = INT_MAX;
1766
	sock = sockfd_lookup_light(fd, &err, &fput_needed);
L
Linus Torvalds 已提交
1767
	if (!sock)
1768
		goto out;
L
Linus Torvalds 已提交
1769

1770 1771 1772 1773 1774 1775
	msg.msg_control = NULL;
	msg.msg_controllen = 0;
	msg.msg_iovlen = 1;
	msg.msg_iov = &iov;
	iov.iov_len = size;
	iov.iov_base = ubuf;
1776 1777
	msg.msg_name = (struct sockaddr *)&address;
	msg.msg_namelen = sizeof(address);
L
Linus Torvalds 已提交
1778 1779
	if (sock->file->f_flags & O_NONBLOCK)
		flags |= MSG_DONTWAIT;
1780
	err = sock_recvmsg(sock, &msg, size, flags);
L
Linus Torvalds 已提交
1781

1782
	if (err >= 0 && addr != NULL) {
1783
		err2 = move_addr_to_user(&address,
1784
					 msg.msg_namelen, addr, addr_len);
1785 1786
		if (err2 < 0)
			err = err2;
L
Linus Torvalds 已提交
1787
	}
1788 1789

	fput_light(sock->file, fput_needed);
1790
out:
L
Linus Torvalds 已提交
1791 1792 1793 1794
	return err;
}

/*
1795
 *	Receive a datagram from a socket.
L
Linus Torvalds 已提交
1796 1797
 */

1798
asmlinkage long sys_recv(int fd, void __user *ubuf, size_t size,
1799
			 unsigned int flags)
L
Linus Torvalds 已提交
1800 1801 1802 1803 1804 1805 1806 1807 1808
{
	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.
 */

1809 1810
SYSCALL_DEFINE5(setsockopt, int, fd, int, level, int, optname,
		char __user *, optval, int, optlen)
L
Linus Torvalds 已提交
1811
{
1812
	int err, fput_needed;
L
Linus Torvalds 已提交
1813 1814 1815 1816
	struct socket *sock;

	if (optlen < 0)
		return -EINVAL;
1817 1818 1819 1820

	sock = sockfd_lookup_light(fd, &err, &fput_needed);
	if (sock != NULL) {
		err = security_socket_setsockopt(sock, level, optname);
1821 1822
		if (err)
			goto out_put;
L
Linus Torvalds 已提交
1823 1824

		if (level == SOL_SOCKET)
1825 1826 1827
			err =
			    sock_setsockopt(sock, level, optname, optval,
					    optlen);
L
Linus Torvalds 已提交
1828
		else
1829 1830 1831
			err =
			    sock->ops->setsockopt(sock, level, optname, optval,
						  optlen);
1832 1833
out_put:
		fput_light(sock->file, fput_needed);
L
Linus Torvalds 已提交
1834 1835 1836 1837 1838 1839 1840 1841 1842
	}
	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.
 */

1843 1844
SYSCALL_DEFINE5(getsockopt, int, fd, int, level, int, optname,
		char __user *, optval, int __user *, optlen)
L
Linus Torvalds 已提交
1845
{
1846
	int err, fput_needed;
L
Linus Torvalds 已提交
1847 1848
	struct socket *sock;

1849 1850
	sock = sockfd_lookup_light(fd, &err, &fput_needed);
	if (sock != NULL) {
1851 1852 1853
		err = security_socket_getsockopt(sock, level, optname);
		if (err)
			goto out_put;
L
Linus Torvalds 已提交
1854 1855

		if (level == SOL_SOCKET)
1856 1857 1858
			err =
			    sock_getsockopt(sock, level, optname, optval,
					    optlen);
L
Linus Torvalds 已提交
1859
		else
1860 1861 1862
			err =
			    sock->ops->getsockopt(sock, level, optname, optval,
						  optlen);
1863 1864
out_put:
		fput_light(sock->file, fput_needed);
L
Linus Torvalds 已提交
1865 1866 1867 1868 1869 1870 1871 1872
	}
	return err;
}

/*
 *	Shutdown a socket.
 */

1873
SYSCALL_DEFINE2(shutdown, int, fd, int, how)
L
Linus Torvalds 已提交
1874
{
1875
	int err, fput_needed;
L
Linus Torvalds 已提交
1876 1877
	struct socket *sock;

1878 1879
	sock = sockfd_lookup_light(fd, &err, &fput_needed);
	if (sock != NULL) {
L
Linus Torvalds 已提交
1880
		err = security_socket_shutdown(sock, how);
1881 1882 1883
		if (!err)
			err = sock->ops->shutdown(sock, how);
		fput_light(sock->file, fput_needed);
L
Linus Torvalds 已提交
1884 1885 1886 1887
	}
	return err;
}

1888
/* A couple of helpful macros for getting the address of the 32/64 bit
L
Linus Torvalds 已提交
1889 1890 1891 1892 1893 1894
 * 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)

1895 1896 1897 1898 1899
struct used_address {
	struct sockaddr_storage name;
	unsigned int name_len;
};

1900
static int __sys_sendmsg(struct socket *sock, struct msghdr __user *msg,
1901
			 struct msghdr *msg_sys, unsigned int flags,
1902
			 struct used_address *used_address)
L
Linus Torvalds 已提交
1903
{
1904 1905
	struct compat_msghdr __user *msg_compat =
	    (struct compat_msghdr __user *)msg;
1906
	struct sockaddr_storage address;
L
Linus Torvalds 已提交
1907
	struct iovec iovstack[UIO_FASTIOV], *iov = iovstack;
1908
	unsigned char ctl[sizeof(struct cmsghdr) + 20]
1909 1910
	    __attribute__ ((aligned(sizeof(__kernel_size_t))));
	/* 20 is size of ipv6_pktinfo */
L
Linus Torvalds 已提交
1911
	unsigned char *ctl_buf = ctl;
E
Eric Dumazet 已提交
1912
	int err, ctl_len, total_len;
1913

L
Linus Torvalds 已提交
1914 1915
	err = -EFAULT;
	if (MSG_CMSG_COMPAT & flags) {
1916
		if (get_compat_msghdr(msg_sys, msg_compat))
L
Linus Torvalds 已提交
1917
			return -EFAULT;
1918
	} else if (copy_from_user(msg_sys, msg, sizeof(struct msghdr)))
L
Linus Torvalds 已提交
1919 1920
		return -EFAULT;

1921
	if (msg_sys->msg_iovlen > UIO_FASTIOV) {
E
Eric Dumazet 已提交
1922 1923 1924 1925 1926 1927
		err = -EMSGSIZE;
		if (msg_sys->msg_iovlen > UIO_MAXIOV)
			goto out;
		err = -ENOMEM;
		iov = kmalloc(msg_sys->msg_iovlen * sizeof(struct iovec),
			      GFP_KERNEL);
L
Linus Torvalds 已提交
1928
		if (!iov)
1929
			goto out;
L
Linus Torvalds 已提交
1930 1931 1932 1933
	}

	/* This will also move the address data into kernel space */
	if (MSG_CMSG_COMPAT & flags) {
1934
		err = verify_compat_iovec(msg_sys, iov, &address, VERIFY_READ);
L
Linus Torvalds 已提交
1935
	} else
1936
		err = verify_iovec(msg_sys, iov, &address, VERIFY_READ);
1937
	if (err < 0)
L
Linus Torvalds 已提交
1938 1939 1940 1941 1942
		goto out_freeiov;
	total_len = err;

	err = -ENOBUFS;

1943
	if (msg_sys->msg_controllen > INT_MAX)
L
Linus Torvalds 已提交
1944
		goto out_freeiov;
1945
	ctl_len = msg_sys->msg_controllen;
L
Linus Torvalds 已提交
1946
	if ((MSG_CMSG_COMPAT & flags) && ctl_len) {
1947
		err =
1948
		    cmsghdr_from_user_compat_to_kern(msg_sys, sock->sk, ctl,
1949
						     sizeof(ctl));
L
Linus Torvalds 已提交
1950 1951
		if (err)
			goto out_freeiov;
1952 1953
		ctl_buf = msg_sys->msg_control;
		ctl_len = msg_sys->msg_controllen;
L
Linus Torvalds 已提交
1954
	} else if (ctl_len) {
1955
		if (ctl_len > sizeof(ctl)) {
L
Linus Torvalds 已提交
1956
			ctl_buf = sock_kmalloc(sock->sk, ctl_len, GFP_KERNEL);
1957
			if (ctl_buf == NULL)
L
Linus Torvalds 已提交
1958 1959 1960 1961
				goto out_freeiov;
		}
		err = -EFAULT;
		/*
1962
		 * Careful! Before this, msg_sys->msg_control contains a user pointer.
L
Linus Torvalds 已提交
1963 1964 1965
		 * Afterwards, it will be a kernel pointer. Thus the compiler-assisted
		 * checking falls down on this.
		 */
1966
		if (copy_from_user(ctl_buf,
1967
				   (void __user __force *)msg_sys->msg_control,
1968
				   ctl_len))
L
Linus Torvalds 已提交
1969
			goto out_freectl;
1970
		msg_sys->msg_control = ctl_buf;
L
Linus Torvalds 已提交
1971
	}
1972
	msg_sys->msg_flags = flags;
L
Linus Torvalds 已提交
1973 1974

	if (sock->file->f_flags & O_NONBLOCK)
1975
		msg_sys->msg_flags |= MSG_DONTWAIT;
1976 1977 1978 1979 1980 1981
	/*
	 * If this is sendmmsg() and current destination address is same as
	 * previously succeeded address, omit asking LSM's decision.
	 * used_address->name_len is initialized to UINT_MAX so that the first
	 * destination address never matches.
	 */
1982 1983 1984
	if (used_address && msg_sys->msg_name &&
	    used_address->name_len == msg_sys->msg_namelen &&
	    !memcmp(&used_address->name, msg_sys->msg_name,
1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995
		    used_address->name_len)) {
		err = sock_sendmsg_nosec(sock, msg_sys, total_len);
		goto out_freectl;
	}
	err = sock_sendmsg(sock, msg_sys, total_len);
	/*
	 * If this is sendmmsg() and sending to current destination address was
	 * successful, remember it.
	 */
	if (used_address && err >= 0) {
		used_address->name_len = msg_sys->msg_namelen;
1996 1997 1998
		if (msg_sys->msg_name)
			memcpy(&used_address->name, msg_sys->msg_name,
			       used_address->name_len);
1999
	}
L
Linus Torvalds 已提交
2000 2001

out_freectl:
2002
	if (ctl_buf != ctl)
L
Linus Torvalds 已提交
2003 2004 2005
		sock_kfree_s(sock->sk, ctl_buf, ctl_len);
out_freeiov:
	if (iov != iovstack)
E
Eric Dumazet 已提交
2006
		kfree(iov);
2007 2008 2009 2010 2011 2012 2013 2014
out:
	return err;
}

/*
 *	BSD sendmsg interface
 */

2015
SYSCALL_DEFINE3(sendmsg, int, fd, struct msghdr __user *, msg, unsigned int, flags)
2016 2017 2018 2019 2020 2021 2022 2023
{
	int fput_needed, err;
	struct msghdr msg_sys;
	struct socket *sock = sockfd_lookup_light(fd, &err, &fput_needed);

	if (!sock)
		goto out;

2024
	err = __sys_sendmsg(sock, msg, &msg_sys, flags, NULL);
2025

2026
	fput_light(sock->file, fput_needed);
2027
out:
L
Linus Torvalds 已提交
2028 2029 2030
	return err;
}

2031 2032 2033 2034 2035 2036 2037 2038 2039 2040 2041 2042
/*
 *	Linux sendmmsg interface
 */

int __sys_sendmmsg(int fd, struct mmsghdr __user *mmsg, unsigned int vlen,
		   unsigned int flags)
{
	int fput_needed, err, datagrams;
	struct socket *sock;
	struct mmsghdr __user *entry;
	struct compat_mmsghdr __user *compat_entry;
	struct msghdr msg_sys;
2043
	struct used_address used_address;
2044

2045 2046
	if (vlen > UIO_MAXIOV)
		vlen = UIO_MAXIOV;
2047 2048 2049 2050 2051 2052 2053

	datagrams = 0;

	sock = sockfd_lookup_light(fd, &err, &fput_needed);
	if (!sock)
		return err;

2054
	used_address.name_len = UINT_MAX;
2055 2056
	entry = mmsg;
	compat_entry = (struct compat_mmsghdr __user *)mmsg;
2057
	err = 0;
2058 2059 2060 2061

	while (datagrams < vlen) {
		if (MSG_CMSG_COMPAT & flags) {
			err = __sys_sendmsg(sock, (struct msghdr __user *)compat_entry,
2062
					    &msg_sys, flags, &used_address);
2063 2064 2065 2066 2067 2068
			if (err < 0)
				break;
			err = __put_user(err, &compat_entry->msg_len);
			++compat_entry;
		} else {
			err = __sys_sendmsg(sock, (struct msghdr __user *)entry,
2069
					    &msg_sys, flags, &used_address);
2070 2071 2072 2073 2074 2075 2076 2077 2078 2079 2080 2081 2082
			if (err < 0)
				break;
			err = put_user(err, &entry->msg_len);
			++entry;
		}

		if (err)
			break;
		++datagrams;
	}

	fput_light(sock->file, fput_needed);

2083 2084
	/* We only return an error if no datagrams were able to be sent */
	if (datagrams != 0)
2085 2086 2087 2088 2089 2090 2091 2092 2093 2094 2095
		return datagrams;

	return err;
}

SYSCALL_DEFINE4(sendmmsg, int, fd, struct mmsghdr __user *, mmsg,
		unsigned int, vlen, unsigned int, flags)
{
	return __sys_sendmmsg(fd, mmsg, vlen, flags);
}

2096
static int __sys_recvmsg(struct socket *sock, struct msghdr __user *msg,
2097
			 struct msghdr *msg_sys, unsigned int flags, int nosec)
L
Linus Torvalds 已提交
2098
{
2099 2100
	struct compat_msghdr __user *msg_compat =
	    (struct compat_msghdr __user *)msg;
L
Linus Torvalds 已提交
2101
	struct iovec iovstack[UIO_FASTIOV];
2102
	struct iovec *iov = iovstack;
L
Linus Torvalds 已提交
2103
	unsigned long cmsg_ptr;
E
Eric Dumazet 已提交
2104
	int err, total_len, len;
L
Linus Torvalds 已提交
2105 2106

	/* kernel mode address */
2107
	struct sockaddr_storage addr;
L
Linus Torvalds 已提交
2108 2109 2110 2111

	/* user mode address pointers */
	struct sockaddr __user *uaddr;
	int __user *uaddr_len;
2112

L
Linus Torvalds 已提交
2113
	if (MSG_CMSG_COMPAT & flags) {
2114
		if (get_compat_msghdr(msg_sys, msg_compat))
L
Linus Torvalds 已提交
2115
			return -EFAULT;
2116
	} else if (copy_from_user(msg_sys, msg, sizeof(struct msghdr)))
2117
		return -EFAULT;
L
Linus Torvalds 已提交
2118

2119
	if (msg_sys->msg_iovlen > UIO_FASTIOV) {
E
Eric Dumazet 已提交
2120 2121 2122 2123 2124 2125
		err = -EMSGSIZE;
		if (msg_sys->msg_iovlen > UIO_MAXIOV)
			goto out;
		err = -ENOMEM;
		iov = kmalloc(msg_sys->msg_iovlen * sizeof(struct iovec),
			      GFP_KERNEL);
L
Linus Torvalds 已提交
2126
		if (!iov)
2127
			goto out;
L
Linus Torvalds 已提交
2128 2129 2130
	}

	/*
2131 2132
	 *      Save the user-mode address (verify_iovec will change the
	 *      kernel msghdr to use the kernel address space)
L
Linus Torvalds 已提交
2133
	 */
2134

2135
	uaddr = (__force void __user *)msg_sys->msg_name;
L
Linus Torvalds 已提交
2136 2137
	uaddr_len = COMPAT_NAMELEN(msg);
	if (MSG_CMSG_COMPAT & flags) {
2138
		err = verify_compat_iovec(msg_sys, iov, &addr, VERIFY_WRITE);
L
Linus Torvalds 已提交
2139
	} else
2140
		err = verify_iovec(msg_sys, iov, &addr, VERIFY_WRITE);
L
Linus Torvalds 已提交
2141 2142
	if (err < 0)
		goto out_freeiov;
2143
	total_len = err;
L
Linus Torvalds 已提交
2144

2145 2146
	cmsg_ptr = (unsigned long)msg_sys->msg_control;
	msg_sys->msg_flags = flags & (MSG_CMSG_CLOEXEC|MSG_CMSG_COMPAT);
2147

L
Linus Torvalds 已提交
2148 2149
	if (sock->file->f_flags & O_NONBLOCK)
		flags |= MSG_DONTWAIT;
2150 2151
	err = (nosec ? sock_recvmsg_nosec : sock_recvmsg)(sock, msg_sys,
							  total_len, flags);
L
Linus Torvalds 已提交
2152 2153 2154 2155 2156
	if (err < 0)
		goto out_freeiov;
	len = err;

	if (uaddr != NULL) {
2157
		err = move_addr_to_user(&addr,
2158
					msg_sys->msg_namelen, uaddr,
2159
					uaddr_len);
L
Linus Torvalds 已提交
2160 2161 2162
		if (err < 0)
			goto out_freeiov;
	}
2163
	err = __put_user((msg_sys->msg_flags & ~MSG_CMSG_COMPAT),
2164
			 COMPAT_FLAGS(msg));
L
Linus Torvalds 已提交
2165 2166 2167
	if (err)
		goto out_freeiov;
	if (MSG_CMSG_COMPAT & flags)
2168
		err = __put_user((unsigned long)msg_sys->msg_control - cmsg_ptr,
L
Linus Torvalds 已提交
2169 2170
				 &msg_compat->msg_controllen);
	else
2171
		err = __put_user((unsigned long)msg_sys->msg_control - cmsg_ptr,
L
Linus Torvalds 已提交
2172 2173 2174 2175 2176 2177 2178
				 &msg->msg_controllen);
	if (err)
		goto out_freeiov;
	err = len;

out_freeiov:
	if (iov != iovstack)
E
Eric Dumazet 已提交
2179
		kfree(iov);
2180 2181 2182 2183 2184 2185 2186 2187 2188 2189 2190 2191 2192 2193 2194 2195 2196 2197 2198 2199
out:
	return err;
}

/*
 *	BSD recvmsg interface
 */

SYSCALL_DEFINE3(recvmsg, int, fd, struct msghdr __user *, msg,
		unsigned int, flags)
{
	int fput_needed, err;
	struct msghdr msg_sys;
	struct socket *sock = sockfd_lookup_light(fd, &err, &fput_needed);

	if (!sock)
		goto out;

	err = __sys_recvmsg(sock, msg, &msg_sys, flags, 0);

2200
	fput_light(sock->file, fput_needed);
L
Linus Torvalds 已提交
2201 2202 2203 2204
out:
	return err;
}

2205 2206 2207 2208 2209 2210 2211 2212 2213 2214
/*
 *     Linux recvmmsg interface
 */

int __sys_recvmmsg(int fd, struct mmsghdr __user *mmsg, unsigned int vlen,
		   unsigned int flags, struct timespec *timeout)
{
	int fput_needed, err, datagrams;
	struct socket *sock;
	struct mmsghdr __user *entry;
2215
	struct compat_mmsghdr __user *compat_entry;
2216 2217 2218 2219 2220 2221 2222 2223 2224 2225 2226 2227 2228 2229 2230 2231 2232 2233 2234
	struct msghdr msg_sys;
	struct timespec end_time;

	if (timeout &&
	    poll_select_set_timeout(&end_time, timeout->tv_sec,
				    timeout->tv_nsec))
		return -EINVAL;

	datagrams = 0;

	sock = sockfd_lookup_light(fd, &err, &fput_needed);
	if (!sock)
		return err;

	err = sock_error(sock->sk);
	if (err)
		goto out_put;

	entry = mmsg;
2235
	compat_entry = (struct compat_mmsghdr __user *)mmsg;
2236 2237 2238 2239 2240

	while (datagrams < vlen) {
		/*
		 * No need to ask LSM for more than the first datagram.
		 */
2241 2242
		if (MSG_CMSG_COMPAT & flags) {
			err = __sys_recvmsg(sock, (struct msghdr __user *)compat_entry,
2243 2244
					    &msg_sys, flags & ~MSG_WAITFORONE,
					    datagrams);
2245 2246 2247 2248 2249 2250
			if (err < 0)
				break;
			err = __put_user(err, &compat_entry->msg_len);
			++compat_entry;
		} else {
			err = __sys_recvmsg(sock, (struct msghdr __user *)entry,
2251 2252
					    &msg_sys, flags & ~MSG_WAITFORONE,
					    datagrams);
2253 2254 2255 2256 2257 2258
			if (err < 0)
				break;
			err = put_user(err, &entry->msg_len);
			++entry;
		}

2259 2260 2261 2262
		if (err)
			break;
		++datagrams;

2263 2264 2265 2266
		/* MSG_WAITFORONE turns on MSG_DONTWAIT after one packet */
		if (flags & MSG_WAITFORONE)
			flags |= MSG_DONTWAIT;

2267 2268 2269 2270 2271 2272 2273 2274 2275 2276 2277 2278 2279 2280 2281 2282 2283 2284 2285 2286
		if (timeout) {
			ktime_get_ts(timeout);
			*timeout = timespec_sub(end_time, *timeout);
			if (timeout->tv_sec < 0) {
				timeout->tv_sec = timeout->tv_nsec = 0;
				break;
			}

			/* Timeout, return less than vlen datagrams */
			if (timeout->tv_nsec == 0 && timeout->tv_sec == 0)
				break;
		}

		/* Out of band data, return right away */
		if (msg_sys.msg_flags & MSG_OOB)
			break;
	}

out_put:
	fput_light(sock->file, fput_needed);
L
Linus Torvalds 已提交
2287

2288 2289 2290 2291 2292 2293 2294 2295 2296 2297 2298 2299 2300 2301 2302 2303 2304 2305 2306 2307 2308 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
	if (err == 0)
		return datagrams;

	if (datagrams != 0) {
		/*
		 * We may return less entries than requested (vlen) if the
		 * sock is non block and there aren't enough datagrams...
		 */
		if (err != -EAGAIN) {
			/*
			 * ... or  if recvmsg returns an error after we
			 * received some datagrams, where we record the
			 * error to return on the next call or if the
			 * app asks about it using getsockopt(SO_ERROR).
			 */
			sock->sk->sk_err = -err;
		}

		return datagrams;
	}

	return err;
}

SYSCALL_DEFINE5(recvmmsg, int, fd, struct mmsghdr __user *, mmsg,
		unsigned int, vlen, unsigned int, flags,
		struct timespec __user *, timeout)
{
	int datagrams;
	struct timespec timeout_sys;

	if (!timeout)
		return __sys_recvmmsg(fd, mmsg, vlen, flags, NULL);

	if (copy_from_user(&timeout_sys, timeout, sizeof(timeout_sys)))
		return -EFAULT;

	datagrams = __sys_recvmmsg(fd, mmsg, vlen, flags, &timeout_sys);

	if (datagrams > 0 &&
	    copy_to_user(timeout, &timeout_sys, sizeof(timeout_sys)))
		datagrams = -EFAULT;

	return datagrams;
}

#ifdef __ARCH_WANT_SYS_SOCKETCALL
L
Linus Torvalds 已提交
2335 2336
/* Argument list sizes for sys_socketcall */
#define AL(x) ((x) * sizeof(unsigned long))
2337
static const unsigned char nargs[21] = {
2338 2339 2340
	AL(0), AL(3), AL(3), AL(3), AL(2), AL(3),
	AL(3), AL(3), AL(4), AL(4), AL(4), AL(6),
	AL(6), AL(2), AL(5), AL(5), AL(3), AL(3),
2341
	AL(4), AL(5), AL(4)
2342 2343
};

L
Linus Torvalds 已提交
2344 2345 2346
#undef AL

/*
2347
 *	System call vectors.
L
Linus Torvalds 已提交
2348 2349 2350
 *
 *	Argument checking cleaned up. Saved 20% in size.
 *  This function doesn't need to set the kernel lock because
2351
 *  it is set by the callees.
L
Linus Torvalds 已提交
2352 2353
 */

2354
SYSCALL_DEFINE2(socketcall, int, call, unsigned long __user *, args)
L
Linus Torvalds 已提交
2355 2356
{
	unsigned long a[6];
2357
	unsigned long a0, a1;
L
Linus Torvalds 已提交
2358
	int err;
2359
	unsigned int len;
L
Linus Torvalds 已提交
2360

2361
	if (call < 1 || call > SYS_SENDMMSG)
L
Linus Torvalds 已提交
2362 2363
		return -EINVAL;

2364 2365 2366 2367
	len = nargs[call];
	if (len > sizeof(a))
		return -EINVAL;

L
Linus Torvalds 已提交
2368
	/* copy_from_user should be SMP safe. */
2369
	if (copy_from_user(a, args, len))
L
Linus Torvalds 已提交
2370
		return -EFAULT;
2371

A
Al Viro 已提交
2372
	audit_socketcall(nargs[call] / sizeof(unsigned long), a);
2373

2374 2375 2376 2377 2378 2379 2380 2381 2382 2383 2384 2385 2386 2387 2388 2389 2390
	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
2391 2392
		err = sys_accept4(a0, (struct sockaddr __user *)a1,
				  (int __user *)a[2], 0);
2393 2394 2395 2396 2397 2398 2399 2400 2401 2402 2403 2404 2405 2406 2407 2408 2409 2410 2411 2412 2413 2414 2415 2416 2417 2418 2419 2420 2421 2422 2423 2424 2425 2426 2427 2428 2429 2430 2431 2432 2433 2434 2435
		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;
2436 2437 2438
	case SYS_SENDMMSG:
		err = sys_sendmmsg(a0, (struct mmsghdr __user *)a1, a[2], a[3]);
		break;
2439 2440 2441
	case SYS_RECVMSG:
		err = sys_recvmsg(a0, (struct msghdr __user *)a1, a[2]);
		break;
2442 2443 2444 2445
	case SYS_RECVMMSG:
		err = sys_recvmmsg(a0, (struct mmsghdr __user *)a1, a[2], a[3],
				   (struct timespec __user *)a[4]);
		break;
U
Ulrich Drepper 已提交
2446 2447 2448
	case SYS_ACCEPT4:
		err = sys_accept4(a0, (struct sockaddr __user *)a1,
				  (int __user *)a[2], a[3]);
U
Ulrich Drepper 已提交
2449
		break;
2450 2451 2452
	default:
		err = -EINVAL;
		break;
L
Linus Torvalds 已提交
2453 2454 2455 2456
	}
	return err;
}

2457
#endif				/* __ARCH_WANT_SYS_SOCKETCALL */
L
Linus Torvalds 已提交
2458

2459 2460 2461 2462
/**
 *	sock_register - add a socket protocol handler
 *	@ops: description of protocol
 *
L
Linus Torvalds 已提交
2463 2464
 *	This function is called by a protocol handler that wants to
 *	advertise its address family, and have it linked into the
2465 2466
 *	socket interface. The value ops->family coresponds to the
 *	socket system call protocol family.
L
Linus Torvalds 已提交
2467
 */
2468
int sock_register(const struct net_proto_family *ops)
L
Linus Torvalds 已提交
2469 2470 2471 2472
{
	int err;

	if (ops->family >= NPROTO) {
2473 2474
		printk(KERN_CRIT "protocol %d >= NPROTO(%d)\n", ops->family,
		       NPROTO);
L
Linus Torvalds 已提交
2475 2476
		return -ENOBUFS;
	}
2477 2478

	spin_lock(&net_family_lock);
E
Eric Dumazet 已提交
2479 2480
	if (rcu_dereference_protected(net_families[ops->family],
				      lockdep_is_held(&net_family_lock)))
2481 2482
		err = -EEXIST;
	else {
2483
		rcu_assign_pointer(net_families[ops->family], ops);
L
Linus Torvalds 已提交
2484 2485
		err = 0;
	}
2486 2487
	spin_unlock(&net_family_lock);

2488
	printk(KERN_INFO "NET: Registered protocol family %d\n", ops->family);
L
Linus Torvalds 已提交
2489 2490
	return err;
}
2491
EXPORT_SYMBOL(sock_register);
L
Linus Torvalds 已提交
2492

2493 2494 2495 2496
/**
 *	sock_unregister - remove a protocol handler
 *	@family: protocol family to remove
 *
L
Linus Torvalds 已提交
2497 2498
 *	This function is called by a protocol handler that wants to
 *	remove its address family, and have it unlinked from the
2499 2500 2501 2502 2503 2504
 *	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
2505
 */
2506
void sock_unregister(int family)
L
Linus Torvalds 已提交
2507
{
2508
	BUG_ON(family < 0 || family >= NPROTO);
L
Linus Torvalds 已提交
2509

2510
	spin_lock(&net_family_lock);
2511
	RCU_INIT_POINTER(net_families[family], NULL);
2512 2513 2514 2515
	spin_unlock(&net_family_lock);

	synchronize_rcu();

2516
	printk(KERN_INFO "NET: Unregistered protocol family %d\n", family);
L
Linus Torvalds 已提交
2517
}
2518
EXPORT_SYMBOL(sock_unregister);
L
Linus Torvalds 已提交
2519

2520
static int __init sock_init(void)
L
Linus Torvalds 已提交
2521
{
N
Nick Piggin 已提交
2522
	int err;
2523 2524 2525 2526 2527 2528
	/*
	 *      Initialize the network sysctl infrastructure.
	 */
	err = net_sysctl_init();
	if (err)
		goto out;
N
Nick Piggin 已提交
2529

L
Linus Torvalds 已提交
2530
	/*
2531
	 *      Initialize sock SLAB cache.
L
Linus Torvalds 已提交
2532
	 */
2533

L
Linus Torvalds 已提交
2534 2535 2536
	sk_init();

	/*
2537
	 *      Initialize skbuff SLAB cache
L
Linus Torvalds 已提交
2538 2539 2540 2541
	 */
	skb_init();

	/*
2542
	 *      Initialize the protocols module.
L
Linus Torvalds 已提交
2543 2544 2545
	 */

	init_inodecache();
N
Nick Piggin 已提交
2546 2547 2548 2549

	err = register_filesystem(&sock_fs_type);
	if (err)
		goto out_fs;
L
Linus Torvalds 已提交
2550
	sock_mnt = kern_mount(&sock_fs_type);
N
Nick Piggin 已提交
2551 2552 2553 2554
	if (IS_ERR(sock_mnt)) {
		err = PTR_ERR(sock_mnt);
		goto out_mount;
	}
2555 2556

	/* The real protocol initialization is performed in later initcalls.
L
Linus Torvalds 已提交
2557 2558 2559 2560 2561
	 */

#ifdef CONFIG_NETFILTER
	netfilter_init();
#endif
2562

2563 2564 2565 2566
#ifdef CONFIG_NETWORK_PHY_TIMESTAMPING
	skb_timestamping_init();
#endif

N
Nick Piggin 已提交
2567 2568 2569 2570 2571 2572 2573
out:
	return err;

out_mount:
	unregister_filesystem(&sock_fs_type);
out_fs:
	goto out;
L
Linus Torvalds 已提交
2574 2575
}

2576 2577
core_initcall(sock_init);	/* early initcall */

L
Linus Torvalds 已提交
2578 2579 2580 2581 2582 2583
#ifdef CONFIG_PROC_FS
void socket_seq_show(struct seq_file *seq)
{
	int cpu;
	int counter = 0;

2584
	for_each_possible_cpu(cpu)
2585
	    counter += per_cpu(sockets_in_use, cpu);
L
Linus Torvalds 已提交
2586 2587 2588 2589 2590 2591 2592

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

	seq_printf(seq, "sockets: used %d\n", counter);
}
2593
#endif				/* CONFIG_PROC_FS */
L
Linus Torvalds 已提交
2594

2595
#ifdef CONFIG_COMPAT
2596
static int do_siocgstamp(struct net *net, struct socket *sock,
2597
			 unsigned int cmd, void __user *up)
2598 2599 2600 2601 2602 2603
{
	mm_segment_t old_fs = get_fs();
	struct timeval ktv;
	int err;

	set_fs(KERNEL_DS);
2604
	err = sock_do_ioctl(net, sock, cmd, (unsigned long)&ktv);
2605
	set_fs(old_fs);
2606
	if (!err)
2607
		err = compat_put_timeval(&ktv, up);
2608

2609 2610 2611
	return err;
}

2612
static int do_siocgstampns(struct net *net, struct socket *sock,
2613
			   unsigned int cmd, void __user *up)
2614 2615 2616 2617 2618 2619
{
	mm_segment_t old_fs = get_fs();
	struct timespec kts;
	int err;

	set_fs(KERNEL_DS);
2620
	err = sock_do_ioctl(net, sock, cmd, (unsigned long)&kts);
2621
	set_fs(old_fs);
2622
	if (!err)
2623
		err = compat_put_timespec(&kts, up);
2624

2625 2626 2627
	return err;
}

2628
static int dev_ifname32(struct net *net, struct compat_ifreq __user *uifr32)
2629 2630 2631 2632 2633
{
	struct ifreq __user *uifr;
	int err;

	uifr = compat_alloc_user_space(sizeof(struct ifreq));
2634
	if (copy_in_user(uifr, uifr32, sizeof(struct compat_ifreq)))
2635 2636
		return -EFAULT;

2637
	err = dev_ioctl(net, SIOCGIFNAME, uifr);
2638 2639 2640
	if (err)
		return err;

2641
	if (copy_in_user(uifr32, uifr, sizeof(struct compat_ifreq)))
2642 2643 2644 2645 2646
		return -EFAULT;

	return 0;
}

2647
static int dev_ifconf(struct net *net, struct compat_ifconf __user *uifc32)
2648
{
2649
	struct compat_ifconf ifc32;
2650 2651
	struct ifconf ifc;
	struct ifconf __user *uifc;
2652
	struct compat_ifreq __user *ifr32;
2653 2654 2655 2656
	struct ifreq __user *ifr;
	unsigned int i, j;
	int err;

2657
	if (copy_from_user(&ifc32, uifc32, sizeof(struct compat_ifconf)))
2658 2659
		return -EFAULT;

2660
	memset(&ifc, 0, sizeof(ifc));
2661 2662 2663 2664 2665 2666
	if (ifc32.ifcbuf == 0) {
		ifc32.ifc_len = 0;
		ifc.ifc_len = 0;
		ifc.ifc_req = NULL;
		uifc = compat_alloc_user_space(sizeof(struct ifconf));
	} else {
2667 2668
		size_t len = ((ifc32.ifc_len / sizeof(struct compat_ifreq)) + 1) *
			sizeof(struct ifreq);
2669 2670 2671 2672
		uifc = compat_alloc_user_space(sizeof(struct ifconf) + len);
		ifc.ifc_len = len;
		ifr = ifc.ifc_req = (void __user *)(uifc + 1);
		ifr32 = compat_ptr(ifc32.ifcbuf);
2673
		for (i = 0; i < ifc32.ifc_len; i += sizeof(struct compat_ifreq)) {
2674
			if (copy_in_user(ifr, ifr32, sizeof(struct compat_ifreq)))
2675 2676 2677 2678 2679 2680 2681 2682
				return -EFAULT;
			ifr++;
			ifr32++;
		}
	}
	if (copy_to_user(uifc, &ifc, sizeof(struct ifconf)))
		return -EFAULT;

2683
	err = dev_ioctl(net, SIOCGIFCONF, uifc);
2684 2685 2686 2687 2688 2689 2690 2691 2692
	if (err)
		return err;

	if (copy_from_user(&ifc, uifc, sizeof(struct ifconf)))
		return -EFAULT;

	ifr = ifc.ifc_req;
	ifr32 = compat_ptr(ifc32.ifcbuf);
	for (i = 0, j = 0;
2693 2694 2695
	     i + sizeof(struct compat_ifreq) <= ifc32.ifc_len && j < ifc.ifc_len;
	     i += sizeof(struct compat_ifreq), j += sizeof(struct ifreq)) {
		if (copy_in_user(ifr32, ifr, sizeof(struct compat_ifreq)))
2696 2697 2698 2699 2700 2701 2702 2703 2704 2705
			return -EFAULT;
		ifr32++;
		ifr++;
	}

	if (ifc32.ifcbuf == 0) {
		/* Translate from 64-bit structure multiple to
		 * a 32-bit one.
		 */
		i = ifc.ifc_len;
2706
		i = ((i / sizeof(struct ifreq)) * sizeof(struct compat_ifreq));
2707 2708 2709 2710
		ifc32.ifc_len = i;
	} else {
		ifc32.ifc_len = i;
	}
2711
	if (copy_to_user(uifc32, &ifc32, sizeof(struct compat_ifconf)))
2712 2713 2714 2715 2716
		return -EFAULT;

	return 0;
}

2717
static int ethtool_ioctl(struct net *net, struct compat_ifreq __user *ifr32)
2718
{
2719 2720 2721 2722
	struct compat_ethtool_rxnfc __user *compat_rxnfc;
	bool convert_in = false, convert_out = false;
	size_t buf_size = ALIGN(sizeof(struct ifreq), 8);
	struct ethtool_rxnfc __user *rxnfc;
2723
	struct ifreq __user *ifr;
2724 2725
	u32 rule_cnt = 0, actual_rule_cnt;
	u32 ethcmd;
2726
	u32 data;
2727
	int ret;
2728

2729 2730
	if (get_user(data, &ifr32->ifr_ifru.ifru_data))
		return -EFAULT;
2731

2732 2733 2734
	compat_rxnfc = compat_ptr(data);

	if (get_user(ethcmd, &compat_rxnfc->cmd))
2735 2736
		return -EFAULT;

2737 2738 2739 2740 2741 2742 2743 2744 2745 2746 2747 2748 2749 2750 2751 2752 2753
	/* Most ethtool structures are defined without padding.
	 * Unfortunately struct ethtool_rxnfc is an exception.
	 */
	switch (ethcmd) {
	default:
		break;
	case ETHTOOL_GRXCLSRLALL:
		/* Buffer size is variable */
		if (get_user(rule_cnt, &compat_rxnfc->rule_cnt))
			return -EFAULT;
		if (rule_cnt > KMALLOC_MAX_SIZE / sizeof(u32))
			return -ENOMEM;
		buf_size += rule_cnt * sizeof(u32);
		/* fall through */
	case ETHTOOL_GRXRINGS:
	case ETHTOOL_GRXCLSRLCNT:
	case ETHTOOL_GRXCLSRULE:
2754
	case ETHTOOL_SRXCLSRLINS:
2755 2756 2757 2758 2759 2760 2761 2762 2763 2764 2765 2766
		convert_out = true;
		/* fall through */
	case ETHTOOL_SRXCLSRLDEL:
		buf_size += sizeof(struct ethtool_rxnfc);
		convert_in = true;
		break;
	}

	ifr = compat_alloc_user_space(buf_size);
	rxnfc = (void *)ifr + ALIGN(sizeof(struct ifreq), 8);

	if (copy_in_user(&ifr->ifr_name, &ifr32->ifr_name, IFNAMSIZ))
2767 2768
		return -EFAULT;

2769 2770
	if (put_user(convert_in ? rxnfc : compat_ptr(data),
		     &ifr->ifr_ifru.ifru_data))
2771 2772
		return -EFAULT;

2773
	if (convert_in) {
2774
		/* We expect there to be holes between fs.m_ext and
2775 2776
		 * fs.ring_cookie and at the end of fs, but nowhere else.
		 */
2777 2778 2779 2780
		BUILD_BUG_ON(offsetof(struct compat_ethtool_rxnfc, fs.m_ext) +
			     sizeof(compat_rxnfc->fs.m_ext) !=
			     offsetof(struct ethtool_rxnfc, fs.m_ext) +
			     sizeof(rxnfc->fs.m_ext));
2781 2782 2783 2784 2785 2786 2787
		BUILD_BUG_ON(
			offsetof(struct compat_ethtool_rxnfc, fs.location) -
			offsetof(struct compat_ethtool_rxnfc, fs.ring_cookie) !=
			offsetof(struct ethtool_rxnfc, fs.location) -
			offsetof(struct ethtool_rxnfc, fs.ring_cookie));

		if (copy_in_user(rxnfc, compat_rxnfc,
2788
				 (void *)(&rxnfc->fs.m_ext + 1) -
2789 2790 2791 2792 2793 2794 2795 2796 2797 2798 2799 2800 2801 2802 2803 2804
				 (void *)rxnfc) ||
		    copy_in_user(&rxnfc->fs.ring_cookie,
				 &compat_rxnfc->fs.ring_cookie,
				 (void *)(&rxnfc->fs.location + 1) -
				 (void *)&rxnfc->fs.ring_cookie) ||
		    copy_in_user(&rxnfc->rule_cnt, &compat_rxnfc->rule_cnt,
				 sizeof(rxnfc->rule_cnt)))
			return -EFAULT;
	}

	ret = dev_ioctl(net, SIOCETHTOOL, ifr);
	if (ret)
		return ret;

	if (convert_out) {
		if (copy_in_user(compat_rxnfc, rxnfc,
2805
				 (const void *)(&rxnfc->fs.m_ext + 1) -
2806 2807 2808 2809 2810 2811 2812 2813 2814 2815 2816 2817 2818 2819 2820 2821 2822 2823 2824 2825 2826 2827 2828 2829 2830 2831 2832 2833 2834 2835
				 (const void *)rxnfc) ||
		    copy_in_user(&compat_rxnfc->fs.ring_cookie,
				 &rxnfc->fs.ring_cookie,
				 (const void *)(&rxnfc->fs.location + 1) -
				 (const void *)&rxnfc->fs.ring_cookie) ||
		    copy_in_user(&compat_rxnfc->rule_cnt, &rxnfc->rule_cnt,
				 sizeof(rxnfc->rule_cnt)))
			return -EFAULT;

		if (ethcmd == ETHTOOL_GRXCLSRLALL) {
			/* As an optimisation, we only copy the actual
			 * number of rules that the underlying
			 * function returned.  Since Mallory might
			 * change the rule count in user memory, we
			 * check that it is less than the rule count
			 * originally given (as the user buffer size),
			 * which has been range-checked.
			 */
			if (get_user(actual_rule_cnt, &rxnfc->rule_cnt))
				return -EFAULT;
			if (actual_rule_cnt < rule_cnt)
				rule_cnt = actual_rule_cnt;
			if (copy_in_user(&compat_rxnfc->rule_locs[0],
					 &rxnfc->rule_locs[0],
					 rule_cnt * sizeof(u32)))
				return -EFAULT;
		}
	}

	return 0;
2836 2837
}

2838 2839 2840 2841 2842 2843
static int compat_siocwandev(struct net *net, struct compat_ifreq __user *uifr32)
{
	void __user *uptr;
	compat_uptr_t uptr32;
	struct ifreq __user *uifr;

2844
	uifr = compat_alloc_user_space(sizeof(*uifr));
2845 2846 2847 2848 2849 2850 2851 2852 2853 2854 2855 2856 2857 2858
	if (copy_in_user(uifr, uifr32, sizeof(struct compat_ifreq)))
		return -EFAULT;

	if (get_user(uptr32, &uifr32->ifr_settings.ifs_ifsu))
		return -EFAULT;

	uptr = compat_ptr(uptr32);

	if (put_user(uptr, &uifr->ifr_settings.ifs_ifsu.raw_hdlc))
		return -EFAULT;

	return dev_ioctl(net, SIOCWANDEV, uifr);
}

2859 2860
static int bond_ioctl(struct net *net, unsigned int cmd,
			 struct compat_ifreq __user *ifr32)
2861 2862 2863 2864 2865 2866 2867 2868 2869 2870 2871 2872 2873
{
	struct ifreq kifr;
	struct ifreq __user *uifr;
	mm_segment_t old_fs;
	int err;
	u32 data;
	void __user *datap;

	switch (cmd) {
	case SIOCBONDENSLAVE:
	case SIOCBONDRELEASE:
	case SIOCBONDSETHWADDR:
	case SIOCBONDCHANGEACTIVE:
2874
		if (copy_from_user(&kifr, ifr32, sizeof(struct compat_ifreq)))
2875 2876 2877
			return -EFAULT;

		old_fs = get_fs();
2878
		set_fs(KERNEL_DS);
2879 2880
		err = dev_ioctl(net, cmd,
				(struct ifreq __user __force *) &kifr);
2881
		set_fs(old_fs);
2882 2883 2884 2885 2886 2887 2888 2889 2890 2891 2892 2893 2894 2895 2896

		return err;
	case SIOCBONDSLAVEINFOQUERY:
	case SIOCBONDINFOQUERY:
		uifr = compat_alloc_user_space(sizeof(*uifr));
		if (copy_in_user(&uifr->ifr_name, &ifr32->ifr_name, IFNAMSIZ))
			return -EFAULT;

		if (get_user(data, &ifr32->ifr_ifru.ifru_data))
			return -EFAULT;

		datap = compat_ptr(data);
		if (put_user(datap, &uifr->ifr_ifru.ifru_data))
			return -EFAULT;

2897
		return dev_ioctl(net, cmd, uifr);
2898
	default:
2899
		return -ENOIOCTLCMD;
2900
	}
2901 2902
}

2903 2904
static int siocdevprivate_ioctl(struct net *net, unsigned int cmd,
				 struct compat_ifreq __user *u_ifreq32)
2905 2906 2907 2908 2909 2910 2911 2912 2913 2914 2915 2916 2917 2918 2919 2920 2921 2922 2923 2924 2925 2926 2927 2928
{
	struct ifreq __user *u_ifreq64;
	char tmp_buf[IFNAMSIZ];
	void __user *data64;
	u32 data32;

	if (copy_from_user(&tmp_buf[0], &(u_ifreq32->ifr_ifrn.ifrn_name[0]),
			   IFNAMSIZ))
		return -EFAULT;
	if (__get_user(data32, &u_ifreq32->ifr_ifru.ifru_data))
		return -EFAULT;
	data64 = compat_ptr(data32);

	u_ifreq64 = compat_alloc_user_space(sizeof(*u_ifreq64));

	/* Don't check these user accesses, just let that get trapped
	 * in the ioctl handler instead.
	 */
	if (copy_to_user(&u_ifreq64->ifr_ifrn.ifrn_name[0], &tmp_buf[0],
			 IFNAMSIZ))
		return -EFAULT;
	if (__put_user(data64, &u_ifreq64->ifr_ifru.ifru_data))
		return -EFAULT;

2929
	return dev_ioctl(net, cmd, u_ifreq64);
2930 2931
}

2932 2933
static int dev_ifsioc(struct net *net, struct socket *sock,
			 unsigned int cmd, struct compat_ifreq __user *uifr32)
2934
{
2935
	struct ifreq __user *uifr;
2936 2937
	int err;

2938 2939 2940 2941 2942 2943
	uifr = compat_alloc_user_space(sizeof(*uifr));
	if (copy_in_user(uifr, uifr32, sizeof(*uifr32)))
		return -EFAULT;

	err = sock_do_ioctl(net, sock, cmd, (unsigned long)uifr);

2944 2945 2946 2947 2948 2949 2950 2951 2952 2953 2954 2955
	if (!err) {
		switch (cmd) {
		case SIOCGIFFLAGS:
		case SIOCGIFMETRIC:
		case SIOCGIFMTU:
		case SIOCGIFMEM:
		case SIOCGIFHWADDR:
		case SIOCGIFINDEX:
		case SIOCGIFADDR:
		case SIOCGIFBRDADDR:
		case SIOCGIFDSTADDR:
		case SIOCGIFNETMASK:
2956
		case SIOCGIFPFLAGS:
2957
		case SIOCGIFTXQLEN:
2958 2959
		case SIOCGMIIPHY:
		case SIOCGMIIREG:
2960
			if (copy_in_user(uifr32, uifr, sizeof(*uifr32)))
2961 2962 2963 2964 2965 2966 2967
				err = -EFAULT;
			break;
		}
	}
	return err;
}

2968 2969 2970 2971 2972 2973 2974 2975 2976 2977 2978 2979 2980 2981 2982 2983 2984 2985 2986 2987
static int compat_sioc_ifmap(struct net *net, unsigned int cmd,
			struct compat_ifreq __user *uifr32)
{
	struct ifreq ifr;
	struct compat_ifmap __user *uifmap32;
	mm_segment_t old_fs;
	int err;

	uifmap32 = &uifr32->ifr_ifru.ifru_map;
	err = copy_from_user(&ifr, uifr32, sizeof(ifr.ifr_name));
	err |= __get_user(ifr.ifr_map.mem_start, &uifmap32->mem_start);
	err |= __get_user(ifr.ifr_map.mem_end, &uifmap32->mem_end);
	err |= __get_user(ifr.ifr_map.base_addr, &uifmap32->base_addr);
	err |= __get_user(ifr.ifr_map.irq, &uifmap32->irq);
	err |= __get_user(ifr.ifr_map.dma, &uifmap32->dma);
	err |= __get_user(ifr.ifr_map.port, &uifmap32->port);
	if (err)
		return -EFAULT;

	old_fs = get_fs();
2988
	set_fs(KERNEL_DS);
2989
	err = dev_ioctl(net, cmd, (void  __user __force *)&ifr);
2990
	set_fs(old_fs);
2991 2992 2993 2994 2995 2996 2997 2998 2999 3000 3001 3002 3003 3004 3005 3006 3007 3008 3009 3010 3011

	if (cmd == SIOCGIFMAP && !err) {
		err = copy_to_user(uifr32, &ifr, sizeof(ifr.ifr_name));
		err |= __put_user(ifr.ifr_map.mem_start, &uifmap32->mem_start);
		err |= __put_user(ifr.ifr_map.mem_end, &uifmap32->mem_end);
		err |= __put_user(ifr.ifr_map.base_addr, &uifmap32->base_addr);
		err |= __put_user(ifr.ifr_map.irq, &uifmap32->irq);
		err |= __put_user(ifr.ifr_map.dma, &uifmap32->dma);
		err |= __put_user(ifr.ifr_map.port, &uifmap32->port);
		if (err)
			err = -EFAULT;
	}
	return err;
}

static int compat_siocshwtstamp(struct net *net, struct compat_ifreq __user *uifr32)
{
	void __user *uptr;
	compat_uptr_t uptr32;
	struct ifreq __user *uifr;

3012
	uifr = compat_alloc_user_space(sizeof(*uifr));
3013 3014 3015 3016 3017 3018 3019 3020 3021 3022 3023 3024 3025 3026
	if (copy_in_user(uifr, uifr32, sizeof(struct compat_ifreq)))
		return -EFAULT;

	if (get_user(uptr32, &uifr32->ifr_data))
		return -EFAULT;

	uptr = compat_ptr(uptr32);

	if (put_user(uptr, &uifr->ifr_data))
		return -EFAULT;

	return dev_ioctl(net, SIOCSHWTSTAMP, uifr);
}

3027
struct rtentry32 {
3028
	u32		rt_pad1;
3029 3030 3031
	struct sockaddr rt_dst;         /* target address               */
	struct sockaddr rt_gateway;     /* gateway addr (RTF_GATEWAY)   */
	struct sockaddr rt_genmask;     /* target network mask (IP)     */
3032 3033 3034 3035 3036 3037 3038
	unsigned short	rt_flags;
	short		rt_pad2;
	u32		rt_pad3;
	unsigned char	rt_tos;
	unsigned char	rt_class;
	short		rt_pad4;
	short		rt_metric;      /* +1 for binary compatibility! */
3039
	/* char * */ u32 rt_dev;        /* forcing the device at add    */
3040 3041
	u32		rt_mtu;         /* per route MTU/Window         */
	u32		rt_window;      /* Window clamping              */
3042 3043 3044 3045 3046 3047 3048 3049 3050 3051 3052 3053 3054 3055 3056 3057
	unsigned short  rt_irtt;        /* Initial RTT                  */
};

struct in6_rtmsg32 {
	struct in6_addr		rtmsg_dst;
	struct in6_addr		rtmsg_src;
	struct in6_addr		rtmsg_gateway;
	u32			rtmsg_type;
	u16			rtmsg_dst_len;
	u16			rtmsg_src_len;
	u32			rtmsg_metric;
	u32			rtmsg_info;
	u32			rtmsg_flags;
	s32			rtmsg_ifindex;
};

3058 3059
static int routing_ioctl(struct net *net, struct socket *sock,
			 unsigned int cmd, void __user *argp)
3060 3061 3062 3063 3064 3065 3066 3067 3068
{
	int ret;
	void *r = NULL;
	struct in6_rtmsg r6;
	struct rtentry r4;
	char devname[16];
	u32 rtdev;
	mm_segment_t old_fs = get_fs();

3069 3070
	if (sock && sock->sk && sock->sk->sk_family == AF_INET6) { /* ipv6 */
		struct in6_rtmsg32 __user *ur6 = argp;
3071
		ret = copy_from_user(&r6.rtmsg_dst, &(ur6->rtmsg_dst),
3072
			3 * sizeof(struct in6_addr));
3073 3074 3075 3076 3077 3078 3079
		ret |= __get_user(r6.rtmsg_type, &(ur6->rtmsg_type));
		ret |= __get_user(r6.rtmsg_dst_len, &(ur6->rtmsg_dst_len));
		ret |= __get_user(r6.rtmsg_src_len, &(ur6->rtmsg_src_len));
		ret |= __get_user(r6.rtmsg_metric, &(ur6->rtmsg_metric));
		ret |= __get_user(r6.rtmsg_info, &(ur6->rtmsg_info));
		ret |= __get_user(r6.rtmsg_flags, &(ur6->rtmsg_flags));
		ret |= __get_user(r6.rtmsg_ifindex, &(ur6->rtmsg_ifindex));
3080 3081 3082

		r = (void *) &r6;
	} else { /* ipv4 */
3083
		struct rtentry32 __user *ur4 = argp;
3084
		ret = copy_from_user(&r4.rt_dst, &(ur4->rt_dst),
3085
					3 * sizeof(struct sockaddr));
3086 3087 3088 3089 3090 3091
		ret |= __get_user(r4.rt_flags, &(ur4->rt_flags));
		ret |= __get_user(r4.rt_metric, &(ur4->rt_metric));
		ret |= __get_user(r4.rt_mtu, &(ur4->rt_mtu));
		ret |= __get_user(r4.rt_window, &(ur4->rt_window));
		ret |= __get_user(r4.rt_irtt, &(ur4->rt_irtt));
		ret |= __get_user(rtdev, &(ur4->rt_dev));
3092
		if (rtdev) {
3093
			ret |= copy_from_user(devname, compat_ptr(rtdev), 15);
3094 3095
			r4.rt_dev = (char __user __force *)devname;
			devname[15] = 0;
3096 3097 3098 3099 3100 3101 3102 3103 3104 3105 3106
		} else
			r4.rt_dev = NULL;

		r = (void *) &r4;
	}

	if (ret) {
		ret = -EFAULT;
		goto out;
	}

3107
	set_fs(KERNEL_DS);
3108
	ret = sock_do_ioctl(net, sock, cmd, (unsigned long) r);
3109
	set_fs(old_fs);
3110 3111 3112 3113 3114 3115 3116

out:
	return ret;
}

/* Since old style bridge ioctl's endup using SIOCDEVPRIVATE
 * for some operations; this forces use of the newer bridge-utils that
L
Lucas De Marchi 已提交
3117
 * use compatible ioctls
3118
 */
3119
static int old_bridge_ioctl(compat_ulong_t __user *argp)
3120
{
3121
	compat_ulong_t tmp;
3122

3123
	if (get_user(tmp, argp))
3124 3125 3126 3127 3128 3129
		return -EFAULT;
	if (tmp == BRCTL_GET_VERSION)
		return BRCTL_VERSION + 1;
	return -EINVAL;
}

3130 3131 3132 3133 3134 3135
static int compat_sock_ioctl_trans(struct file *file, struct socket *sock,
			 unsigned int cmd, unsigned long arg)
{
	void __user *argp = compat_ptr(arg);
	struct sock *sk = sock->sk;
	struct net *net = sock_net(sk);
3136

3137 3138 3139 3140 3141 3142 3143 3144 3145 3146 3147 3148 3149
	if (cmd >= SIOCDEVPRIVATE && cmd <= (SIOCDEVPRIVATE + 15))
		return siocdevprivate_ioctl(net, cmd, argp);

	switch (cmd) {
	case SIOCSIFBR:
	case SIOCGIFBR:
		return old_bridge_ioctl(argp);
	case SIOCGIFNAME:
		return dev_ifname32(net, argp);
	case SIOCGIFCONF:
		return dev_ifconf(net, argp);
	case SIOCETHTOOL:
		return ethtool_ioctl(net, argp);
3150 3151
	case SIOCWANDEV:
		return compat_siocwandev(net, argp);
3152 3153 3154
	case SIOCGIFMAP:
	case SIOCSIFMAP:
		return compat_sioc_ifmap(net, cmd, argp);
3155 3156 3157 3158 3159 3160 3161 3162 3163 3164 3165 3166 3167 3168
	case SIOCBONDENSLAVE:
	case SIOCBONDRELEASE:
	case SIOCBONDSETHWADDR:
	case SIOCBONDSLAVEINFOQUERY:
	case SIOCBONDINFOQUERY:
	case SIOCBONDCHANGEACTIVE:
		return bond_ioctl(net, cmd, argp);
	case SIOCADDRT:
	case SIOCDELRT:
		return routing_ioctl(net, sock, cmd, argp);
	case SIOCGSTAMP:
		return do_siocgstamp(net, sock, cmd, argp);
	case SIOCGSTAMPNS:
		return do_siocgstampns(net, sock, cmd, argp);
3169 3170
	case SIOCSHWTSTAMP:
		return compat_siocshwtstamp(net, argp);
3171 3172 3173 3174 3175 3176 3177 3178 3179 3180 3181 3182 3183 3184 3185 3186 3187 3188 3189 3190 3191 3192 3193 3194 3195 3196 3197 3198 3199 3200 3201 3202 3203 3204 3205 3206 3207 3208 3209 3210 3211 3212

	case FIOSETOWN:
	case SIOCSPGRP:
	case FIOGETOWN:
	case SIOCGPGRP:
	case SIOCBRADDBR:
	case SIOCBRDELBR:
	case SIOCGIFVLAN:
	case SIOCSIFVLAN:
	case SIOCADDDLCI:
	case SIOCDELDLCI:
		return sock_ioctl(file, cmd, arg);

	case SIOCGIFFLAGS:
	case SIOCSIFFLAGS:
	case SIOCGIFMETRIC:
	case SIOCSIFMETRIC:
	case SIOCGIFMTU:
	case SIOCSIFMTU:
	case SIOCGIFMEM:
	case SIOCSIFMEM:
	case SIOCGIFHWADDR:
	case SIOCSIFHWADDR:
	case SIOCADDMULTI:
	case SIOCDELMULTI:
	case SIOCGIFINDEX:
	case SIOCGIFADDR:
	case SIOCSIFADDR:
	case SIOCSIFHWBROADCAST:
	case SIOCDIFADDR:
	case SIOCGIFBRDADDR:
	case SIOCSIFBRDADDR:
	case SIOCGIFDSTADDR:
	case SIOCSIFDSTADDR:
	case SIOCGIFNETMASK:
	case SIOCSIFNETMASK:
	case SIOCSIFPFLAGS:
	case SIOCGIFPFLAGS:
	case SIOCGIFTXQLEN:
	case SIOCSIFTXQLEN:
	case SIOCBRADDIF:
	case SIOCBRDELIF:
3213 3214 3215 3216
	case SIOCSIFNAME:
	case SIOCGMIIPHY:
	case SIOCGMIIREG:
	case SIOCSMIIREG:
3217
		return dev_ifsioc(net, sock, cmd, argp);
3218

3219 3220 3221 3222
	case SIOCSARP:
	case SIOCGARP:
	case SIOCDARP:
	case SIOCATMARK:
3223 3224 3225
		return sock_do_ioctl(net, sock, cmd, arg);
	}

3226 3227
	return -ENOIOCTLCMD;
}
3228

3229
static long compat_sock_ioctl(struct file *file, unsigned int cmd,
3230
			      unsigned long arg)
3231 3232 3233
{
	struct socket *sock = file->private_data;
	int ret = -ENOIOCTLCMD;
3234 3235 3236 3237 3238
	struct sock *sk;
	struct net *net;

	sk = sock->sk;
	net = sock_net(sk);
3239 3240 3241 3242

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

3243 3244 3245 3246
	if (ret == -ENOIOCTLCMD &&
	    (cmd >= SIOCIWFIRST && cmd <= SIOCIWLAST))
		ret = compat_wext_handle_ioctl(net, cmd, arg);

3247 3248 3249
	if (ret == -ENOIOCTLCMD)
		ret = compat_sock_ioctl_trans(file, sock, cmd, arg);

3250 3251 3252 3253
	return ret;
}
#endif

3254 3255 3256 3257
int kernel_bind(struct socket *sock, struct sockaddr *addr, int addrlen)
{
	return sock->ops->bind(sock, addr, addrlen);
}
3258
EXPORT_SYMBOL(kernel_bind);
3259 3260 3261 3262 3263

int kernel_listen(struct socket *sock, int backlog)
{
	return sock->ops->listen(sock, backlog);
}
3264
EXPORT_SYMBOL(kernel_listen);
3265 3266 3267 3268 3269 3270 3271 3272 3273 3274 3275 3276 3277 3278

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);
3279
		*newsock = NULL;
3280 3281 3282 3283
		goto done;
	}

	(*newsock)->ops = sock->ops;
3284
	__module_get((*newsock)->ops->owner);
3285 3286 3287 3288

done:
	return err;
}
3289
EXPORT_SYMBOL(kernel_accept);
3290 3291

int kernel_connect(struct socket *sock, struct sockaddr *addr, int addrlen,
3292
		   int flags)
3293 3294 3295
{
	return sock->ops->connect(sock, addr, addrlen, flags);
}
3296
EXPORT_SYMBOL(kernel_connect);
3297 3298 3299 3300 3301 3302

int kernel_getsockname(struct socket *sock, struct sockaddr *addr,
			 int *addrlen)
{
	return sock->ops->getname(sock, addr, addrlen, 0);
}
3303
EXPORT_SYMBOL(kernel_getsockname);
3304 3305 3306 3307 3308 3309

int kernel_getpeername(struct socket *sock, struct sockaddr *addr,
			 int *addrlen)
{
	return sock->ops->getname(sock, addr, addrlen, 1);
}
3310
EXPORT_SYMBOL(kernel_getpeername);
3311 3312 3313 3314 3315

int kernel_getsockopt(struct socket *sock, int level, int optname,
			char *optval, int *optlen)
{
	mm_segment_t oldfs = get_fs();
3316 3317
	char __user *uoptval;
	int __user *uoptlen;
3318 3319
	int err;

3320 3321 3322
	uoptval = (char __user __force *) optval;
	uoptlen = (int __user __force *) optlen;

3323 3324
	set_fs(KERNEL_DS);
	if (level == SOL_SOCKET)
3325
		err = sock_getsockopt(sock, level, optname, uoptval, uoptlen);
3326
	else
3327 3328
		err = sock->ops->getsockopt(sock, level, optname, uoptval,
					    uoptlen);
3329 3330 3331
	set_fs(oldfs);
	return err;
}
3332
EXPORT_SYMBOL(kernel_getsockopt);
3333 3334

int kernel_setsockopt(struct socket *sock, int level, int optname,
3335
			char *optval, unsigned int optlen)
3336 3337
{
	mm_segment_t oldfs = get_fs();
3338
	char __user *uoptval;
3339 3340
	int err;

3341 3342
	uoptval = (char __user __force *) optval;

3343 3344
	set_fs(KERNEL_DS);
	if (level == SOL_SOCKET)
3345
		err = sock_setsockopt(sock, level, optname, uoptval, optlen);
3346
	else
3347
		err = sock->ops->setsockopt(sock, level, optname, uoptval,
3348 3349 3350 3351
					    optlen);
	set_fs(oldfs);
	return err;
}
3352
EXPORT_SYMBOL(kernel_setsockopt);
3353 3354 3355 3356

int kernel_sendpage(struct socket *sock, struct page *page, int offset,
		    size_t size, int flags)
{
3357 3358
	sock_update_classid(sock->sk);

3359 3360 3361 3362 3363
	if (sock->ops->sendpage)
		return sock->ops->sendpage(sock, page, offset, size, flags);

	return sock_no_sendpage(sock, page, offset, size, flags);
}
3364
EXPORT_SYMBOL(kernel_sendpage);
3365 3366 3367 3368 3369 3370 3371 3372 3373 3374 3375 3376

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;
}
3377
EXPORT_SYMBOL(kernel_sock_ioctl);
3378

3379 3380 3381 3382 3383
int kernel_sock_shutdown(struct socket *sock, enum sock_shutdown_cmd how)
{
	return sock->ops->shutdown(sock, how);
}
EXPORT_SYMBOL(kernel_sock_shutdown);