socket.c 84.0 KB
Newer Older
L
Linus Torvalds 已提交
1 2 3 4 5 6
/*
 * NET		An implementation of the SOCKET network access protocol.
 *
 * Version:	@(#)socket.c	1.1.93	18/02/95
 *
 * Authors:	Orest Zborowski, <obz@Kodak.COM>
7
 *		Ross Biro
L
Linus Torvalds 已提交
8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44
 *		Fred N. van Kempen, <waltje@uWalt.NL.Mugnet.ORG>
 *
 * Fixes:
 *		Anonymous	:	NOTSOCK/BADF cleanup. Error fix in
 *					shutdown()
 *		Alan Cox	:	verify_area() fixes
 *		Alan Cox	:	Removed DDI
 *		Jonathan Kamens	:	SOCK_DGRAM reconnect bug
 *		Alan Cox	:	Moved a load of checks to the very
 *					top level.
 *		Alan Cox	:	Move address structures to/from user
 *					mode above the protocol layers.
 *		Rob Janssen	:	Allow 0 length sends.
 *		Alan Cox	:	Asynchronous I/O support (cribbed from the
 *					tty drivers).
 *		Niibe Yutaka	:	Asynchronous I/O for writes (4.4BSD style)
 *		Jeff Uphoff	:	Made max number of sockets command-line
 *					configurable.
 *		Matti Aarnio	:	Made the number of sockets dynamic,
 *					to be allocated when needed, and mr.
 *					Uphoff's max is used as max to be
 *					allowed to allocate.
 *		Linus		:	Argh. removed all the socket allocation
 *					altogether: it's in the inode now.
 *		Alan Cox	:	Made sock_alloc()/sock_release() public
 *					for NetROM and future kernel nfsd type
 *					stuff.
 *		Alan Cox	:	sendmsg/recvmsg basics.
 *		Tom Dyas	:	Export net symbols.
 *		Marcin Dalecki	:	Fixed problems with CONFIG_NET="n".
 *		Alan Cox	:	Added thread locking to sys_* calls
 *					for sockets. May have errors at the
 *					moment.
 *		Kevin Buhr	:	Fixed the dumb errors in the above.
 *		Andi Kleen	:	Some small cleanups, optimizations,
 *					and fixed a copy_from_user() bug.
 *		Tigran Aivazian	:	sys_send(args) calls sys_sendto(args, NULL, 0)
45
 *		Tigran Aivazian	:	Made listen(2) backlog sanity checks
L
Linus Torvalds 已提交
46 47 48 49 50 51 52 53 54 55
 *					protocol-independent
 *
 *
 *		This program is free software; you can redistribute it and/or
 *		modify it under the terms of the GNU General Public License
 *		as published by the Free Software Foundation; either version
 *		2 of the License, or (at your option) any later version.
 *
 *
 *	This module is effectively the top level interface to the BSD socket
56
 *	paradigm.
L
Linus Torvalds 已提交
57 58 59 60 61 62 63 64 65
 *
 *	Based upon Swansea University Computer Society NET3.039
 */

#include <linux/mm.h>
#include <linux/socket.h>
#include <linux/file.h>
#include <linux/net.h>
#include <linux/interrupt.h>
U
Ulrich Drepper 已提交
66
#include <linux/thread_info.h>
67
#include <linux/rcupdate.h>
L
Linus Torvalds 已提交
68 69 70
#include <linux/netdevice.h>
#include <linux/proc_fs.h>
#include <linux/seq_file.h>
A
Arjan van de Ven 已提交
71
#include <linux/mutex.h>
L
Linus Torvalds 已提交
72
#include <linux/if_bridge.h>
73 74
#include <linux/if_frad.h>
#include <linux/if_vlan.h>
L
Linus Torvalds 已提交
75 76 77 78 79 80 81 82 83 84
#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>
85
#include <linux/audit.h>
86
#include <linux/wireless.h>
87
#include <linux/nsproxy.h>
N
Nick Black 已提交
88
#include <linux/magic.h>
89
#include <linux/slab.h>
90
#include <linux/xattr.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
#include <linux/if_tun.h>
#include <linux/ipv6_route.h>
#include <linux/route.h>
#include <linux/sockios.h>
#include <linux/atalk.h>
107
#include <net/busy_poll.h>
E
Eliezer Tamir 已提交
108 109

#ifdef CONFIG_NET_LL_RX_POLL
110 111
unsigned int sysctl_net_busy_read __read_mostly;
unsigned int sysctl_net_busy_poll __read_mostly;
E
Eliezer Tamir 已提交
112
#endif
113

L
Linus Torvalds 已提交
114
static int sock_no_open(struct inode *irrelevant, struct file *dontcare);
115 116 117 118
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);
119
static int sock_mmap(struct file *file, struct vm_area_struct *vma);
L
Linus Torvalds 已提交
120 121 122 123

static int sock_close(struct inode *inode, struct file *file);
static unsigned int sock_poll(struct file *file,
			      struct poll_table_struct *wait);
124
static long sock_ioctl(struct file *file, unsigned int cmd, unsigned long arg);
125 126
#ifdef CONFIG_COMPAT
static long compat_sock_ioctl(struct file *file,
127
			      unsigned int cmd, unsigned long arg);
128
#endif
L
Linus Torvalds 已提交
129 130 131
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
132
static ssize_t sock_splice_read(struct file *file, loff_t *ppos,
133
				struct pipe_inode_info *pipe, size_t len,
J
Jens Axboe 已提交
134
				unsigned int flags);
L
Linus Torvalds 已提交
135 136 137 138 139 140

/*
 *	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.
 */

141
static const struct file_operations socket_file_ops = {
L
Linus Torvalds 已提交
142 143 144 145 146 147
	.owner =	THIS_MODULE,
	.llseek =	no_llseek,
	.aio_read =	sock_aio_read,
	.aio_write =	sock_aio_write,
	.poll =		sock_poll,
	.unlocked_ioctl = sock_ioctl,
148 149 150
#ifdef CONFIG_COMPAT
	.compat_ioctl = compat_sock_ioctl,
#endif
L
Linus Torvalds 已提交
151 152 153 154
	.mmap =		sock_mmap,
	.open =		sock_no_open,	/* special open code to disallow open via /proc */
	.release =	sock_close,
	.fasync =	sock_fasync,
155 156
	.sendpage =	sock_sendpage,
	.splice_write = generic_splice_sendpage,
J
Jens Axboe 已提交
157
	.splice_read =	sock_splice_read,
L
Linus Torvalds 已提交
158 159 160 161 162 163 164
};

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

static DEFINE_SPINLOCK(net_family_lock);
E
Eric Dumazet 已提交
165
static const struct net_proto_family __rcu *net_families[NPROTO] __read_mostly;
L
Linus Torvalds 已提交
166 167 168 169 170

/*
 *	Statistics counters of the socket lists
 */

171
static DEFINE_PER_CPU(int, sockets_in_use);
L
Linus Torvalds 已提交
172 173

/*
174 175 176
 * 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
177 178 179 180 181 182 183 184 185 186 187 188 189
 */

/**
 *	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.
 */

190
int move_addr_to_kernel(void __user *uaddr, int ulen, struct sockaddr_storage *kaddr)
L
Linus Torvalds 已提交
191
{
192
	if (ulen < 0 || ulen > sizeof(struct sockaddr_storage))
L
Linus Torvalds 已提交
193
		return -EINVAL;
194
	if (ulen == 0)
L
Linus Torvalds 已提交
195
		return 0;
196
	if (copy_from_user(kaddr, uaddr, ulen))
L
Linus Torvalds 已提交
197
		return -EFAULT;
198
	return audit_sockaddr(ulen, kaddr);
L
Linus Torvalds 已提交
199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216
}

/**
 *	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.
 */
217

218
static int move_addr_to_user(struct sockaddr_storage *kaddr, int klen,
S
stephen hemminger 已提交
219
			     void __user *uaddr, int __user *ulen)
L
Linus Torvalds 已提交
220 221 222 223
{
	int err;
	int len;

224 225
	err = get_user(len, ulen);
	if (err)
L
Linus Torvalds 已提交
226
		return err;
227 228
	if (len > klen)
		len = klen;
229
	if (len < 0 || len > sizeof(struct sockaddr_storage))
L
Linus Torvalds 已提交
230
		return -EINVAL;
231
	if (len) {
S
Steve Grubb 已提交
232 233
		if (audit_sockaddr(klen, kaddr))
			return -ENOMEM;
234
		if (copy_to_user(uaddr, kaddr, len))
L
Linus Torvalds 已提交
235 236 237
			return -EFAULT;
	}
	/*
238 239
	 *      "fromlen shall refer to the value before truncation.."
	 *                      1003.1g
L
Linus Torvalds 已提交
240 241 242 243
	 */
	return __put_user(klen, ulen);
}

244
static struct kmem_cache *sock_inode_cachep __read_mostly;
L
Linus Torvalds 已提交
245 246 247 248

static struct inode *sock_alloc_inode(struct super_block *sb)
{
	struct socket_alloc *ei;
249
	struct socket_wq *wq;
250

251
	ei = kmem_cache_alloc(sock_inode_cachep, GFP_KERNEL);
L
Linus Torvalds 已提交
252 253
	if (!ei)
		return NULL;
254 255
	wq = kmalloc(sizeof(*wq), GFP_KERNEL);
	if (!wq) {
256 257 258
		kmem_cache_free(sock_inode_cachep, ei);
		return NULL;
	}
259 260 261
	init_waitqueue_head(&wq->wait);
	wq->fasync_list = NULL;
	RCU_INIT_POINTER(ei->socket.wq, wq);
262

L
Linus Torvalds 已提交
263 264 265 266 267 268 269 270 271 272 273
	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)
{
274
	struct socket_alloc *ei;
275
	struct socket_wq *wq;
276 277

	ei = container_of(inode, struct socket_alloc, vfs_inode);
278
	wq = rcu_dereference_protected(ei->socket.wq, 1);
279
	kfree_rcu(wq, rcu);
280
	kmem_cache_free(sock_inode_cachep, ei);
L
Linus Torvalds 已提交
281 282
}

283
static void init_once(void *foo)
L
Linus Torvalds 已提交
284
{
285
	struct socket_alloc *ei = (struct socket_alloc *)foo;
L
Linus Torvalds 已提交
286

C
Christoph Lameter 已提交
287
	inode_init_once(&ei->vfs_inode);
L
Linus Torvalds 已提交
288
}
289

L
Linus Torvalds 已提交
290 291 292
static int init_inodecache(void)
{
	sock_inode_cachep = kmem_cache_create("sock_inode_cache",
293 294 295 296 297
					      sizeof(struct socket_alloc),
					      0,
					      (SLAB_HWCACHE_ALIGN |
					       SLAB_RECLAIM_ACCOUNT |
					       SLAB_MEM_SPREAD),
298
					      init_once);
L
Linus Torvalds 已提交
299 300 301 302 303
	if (sock_inode_cachep == NULL)
		return -ENOMEM;
	return 0;
}

304
static const struct super_operations sockfs_ops = {
305 306 307
	.alloc_inode	= sock_alloc_inode,
	.destroy_inode	= sock_destroy_inode,
	.statfs		= simple_statfs,
L
Linus Torvalds 已提交
308 309
};

310 311 312 313 314 315 316 317 318
/*
 * 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
319
static const struct dentry_operations sockfs_dentry_operations = {
320
	.d_dname  = sockfs_dname,
L
Linus Torvalds 已提交
321 322
};

323 324 325 326 327 328 329 330 331 332 333 334 335 336 337
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
338 339 340
/*
 *	Obtains the first available file descriptor and sets it up for use.
 *
341 342
 *	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
343 344 345 346 347 348 349 350 351 352 353 354
 *	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.
 */

355
struct file *sock_alloc_file(struct socket *sock, int flags, const char *dname)
L
Linus Torvalds 已提交
356
{
357
	struct qstr name = { .name = "" };
358
	struct path path;
359
	struct file *file;
L
Linus Torvalds 已提交
360

361 362 363 364 365 366 367
	if (dname) {
		name.name = dname;
		name.len = strlen(name.name);
	} else if (sock->sk) {
		name.name = sock->sk->sk_prot_creator->name;
		name.len = strlen(name.name);
	}
368
	path.dentry = d_alloc_pseudo(sock_mnt->mnt_sb, &name);
369 370
	if (unlikely(!path.dentry))
		return ERR_PTR(-ENOMEM);
371
	path.mnt = mntget(sock_mnt);
372

373
	d_instantiate(path.dentry, SOCK_INODE(sock));
A
Al Viro 已提交
374
	SOCK_INODE(sock)->i_fop = &socket_file_ops;
375

376
	file = alloc_file(&path, FMODE_READ | FMODE_WRITE,
377
		  &socket_file_ops);
378
	if (unlikely(IS_ERR(file))) {
A
Al Viro 已提交
379
		/* drop dentry, keep inode */
A
Al Viro 已提交
380
		ihold(path.dentry->d_inode);
381
		path_put(&path);
382
		return file;
A
Al Viro 已提交
383 384 385
	}

	sock->file = file;
386
	file->f_flags = O_RDWR | (flags & O_NONBLOCK);
387
	file->private_data = sock;
388
	return file;
389
}
390
EXPORT_SYMBOL(sock_alloc_file);
391

392
static int sock_map_fd(struct socket *sock, int flags)
393 394
{
	struct file *newfile;
395 396 397
	int fd = get_unused_fd_flags(flags);
	if (unlikely(fd < 0))
		return fd;
398

399
	newfile = sock_alloc_file(sock, flags, NULL);
400
	if (likely(!IS_ERR(newfile))) {
401
		fd_install(fd, newfile);
402 403
		return fd;
	}
404

405 406
	put_unused_fd(fd);
	return PTR_ERR(newfile);
L
Linus Torvalds 已提交
407 408
}

409
struct socket *sock_from_file(struct file *file, int *err)
410 411 412 413
{
	if (file->f_op == &socket_file_ops)
		return file->private_data;	/* set in sock_map_fd */

E
Eric Dumazet 已提交
414 415
	*err = -ENOTSOCK;
	return NULL;
416
}
417
EXPORT_SYMBOL(sock_from_file);
418

L
Linus Torvalds 已提交
419
/**
420
 *	sockfd_lookup - Go from a file number to its socket slot
L
Linus Torvalds 已提交
421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436
 *	@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;

437 438
	file = fget(fd);
	if (!file) {
L
Linus Torvalds 已提交
439 440 441
		*err = -EBADF;
		return NULL;
	}
442

443 444
	sock = sock_from_file(file, err);
	if (!sock)
L
Linus Torvalds 已提交
445
		fput(file);
446 447
	return sock;
}
448
EXPORT_SYMBOL(sockfd_lookup);
L
Linus Torvalds 已提交
449

450 451 452 453 454
static struct socket *sockfd_lookup_light(int fd, int *err, int *fput_needed)
{
	struct file *file;
	struct socket *sock;

455
	*err = -EBADF;
456 457 458 459 460 461
	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
462
	}
463
	return NULL;
L
Linus Torvalds 已提交
464 465
}

466 467 468 469 470 471 472 473 474 475 476 477 478 479 480 481 482 483 484 485 486 487 488 489 490 491 492 493 494 495 496 497 498 499 500 501 502 503 504 505 506 507 508 509 510 511 512 513 514 515 516 517 518 519 520 521 522 523 524 525 526 527
#define XATTR_SOCKPROTONAME_SUFFIX "sockprotoname"
#define XATTR_NAME_SOCKPROTONAME (XATTR_SYSTEM_PREFIX XATTR_SOCKPROTONAME_SUFFIX)
#define XATTR_NAME_SOCKPROTONAME_LEN (sizeof(XATTR_NAME_SOCKPROTONAME)-1)
static ssize_t sockfs_getxattr(struct dentry *dentry,
			       const char *name, void *value, size_t size)
{
	const char *proto_name;
	size_t proto_size;
	int error;

	error = -ENODATA;
	if (!strncmp(name, XATTR_NAME_SOCKPROTONAME, XATTR_NAME_SOCKPROTONAME_LEN)) {
		proto_name = dentry->d_name.name;
		proto_size = strlen(proto_name);

		if (value) {
			error = -ERANGE;
			if (proto_size + 1 > size)
				goto out;

			strncpy(value, proto_name, proto_size + 1);
		}
		error = proto_size + 1;
	}

out:
	return error;
}

static ssize_t sockfs_listxattr(struct dentry *dentry, char *buffer,
				size_t size)
{
	ssize_t len;
	ssize_t used = 0;

	len = security_inode_listsecurity(dentry->d_inode, buffer, size);
	if (len < 0)
		return len;
	used += len;
	if (buffer) {
		if (size < used)
			return -ERANGE;
		buffer += len;
	}

	len = (XATTR_NAME_SOCKPROTONAME_LEN + 1);
	used += len;
	if (buffer) {
		if (size < used)
			return -ERANGE;
		memcpy(buffer, XATTR_NAME_SOCKPROTONAME, len);
		buffer += len;
	}

	return used;
}

static const struct inode_operations sockfs_inode_ops = {
	.getxattr = sockfs_getxattr,
	.listxattr = sockfs_listxattr,
};

L
Linus Torvalds 已提交
528 529
/**
 *	sock_alloc	-	allocate a socket
530
 *
L
Linus Torvalds 已提交
531 532 533 534 535 536 537
 *	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)
{
538 539
	struct inode *inode;
	struct socket *sock;
L
Linus Torvalds 已提交
540

541
	inode = new_inode_pseudo(sock_mnt->mnt_sb);
L
Linus Torvalds 已提交
542 543 544 545 546
	if (!inode)
		return NULL;

	sock = SOCKET_I(inode);

547
	kmemcheck_annotate_bitfield(sock, type);
548
	inode->i_ino = get_next_ino();
549
	inode->i_mode = S_IFSOCK | S_IRWXUGO;
550 551
	inode->i_uid = current_fsuid();
	inode->i_gid = current_fsgid();
552
	inode->i_op = &sockfs_inode_ops;
L
Linus Torvalds 已提交
553

554
	this_cpu_add(sockets_in_use, 1);
L
Linus Torvalds 已提交
555 556 557 558 559 560 561 562
	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.
 */
563

L
Linus Torvalds 已提交
564 565 566 567 568
static int sock_no_open(struct inode *irrelevant, struct file *dontcare)
{
	return -ENXIO;
}

569
const struct file_operations bad_sock_fops = {
L
Linus Torvalds 已提交
570 571
	.owner = THIS_MODULE,
	.open = sock_no_open,
572
	.llseek = noop_llseek,
L
Linus Torvalds 已提交
573 574 575 576 577 578 579 580
};

/**
 *	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
581
 *	an inode not a file.
L
Linus Torvalds 已提交
582
 */
583

L
Linus Torvalds 已提交
584 585 586 587 588 589 590 591 592 593
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);
	}

594
	if (rcu_dereference_protected(sock->wq, 1)->fasync_list)
L
Linus Torvalds 已提交
595 596
		printk(KERN_ERR "sock_release: fasync list not empty!\n");

597 598 599
	if (test_bit(SOCK_EXTERNALLY_ALLOCATED, &sock->flags))
		return;

600
	this_cpu_sub(sockets_in_use, 1);
L
Linus Torvalds 已提交
601 602 603 604
	if (!sock->file) {
		iput(SOCK_INODE(sock));
		return;
	}
605
	sock->file = NULL;
L
Linus Torvalds 已提交
606
}
607
EXPORT_SYMBOL(sock_release);
L
Linus Torvalds 已提交
608

609
void sock_tx_timestamp(struct sock *sk, __u8 *tx_flags)
610
{
611
	*tx_flags = 0;
612
	if (sock_flag(sk, SOCK_TIMESTAMPING_TX_HARDWARE))
613
		*tx_flags |= SKBTX_HW_TSTAMP;
614
	if (sock_flag(sk, SOCK_TIMESTAMPING_TX_SOFTWARE))
615
		*tx_flags |= SKBTX_SW_TSTAMP;
616 617
	if (sock_flag(sk, SOCK_WIFI_STATUS))
		*tx_flags |= SKBTX_WIFI_STATUS;
618 619 620
}
EXPORT_SYMBOL(sock_tx_timestamp);

621 622
static inline int __sock_sendmsg_nosec(struct kiocb *iocb, struct socket *sock,
				       struct msghdr *msg, size_t size)
L
Linus Torvalds 已提交
623 624 625 626 627 628 629 630 631 632 633
{
	struct sock_iocb *si = kiocb_to_siocb(iocb);

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

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

634 635 636 637 638 639 640 641
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
642 643 644 645 646 647 648 649 650 651 652 653 654
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;
}
655
EXPORT_SYMBOL(sock_sendmsg);
L
Linus Torvalds 已提交
656

657
static int sock_sendmsg_nosec(struct socket *sock, struct msghdr *msg, size_t size)
658 659 660 661 662 663 664 665 666 667 668 669 670
{
	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
671 672 673 674 675 676 677 678 679 680 681
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
	 */
682
	msg->msg_iov = (struct iovec *)vec;
L
Linus Torvalds 已提交
683 684 685 686 687
	msg->msg_iovlen = num;
	result = sock_sendmsg(sock, msg, size);
	set_fs(oldfs);
	return result;
}
688
EXPORT_SYMBOL(kernel_sendmsg);
L
Linus Torvalds 已提交
689

690 691 692 693 694 695
/*
 * 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)
{
696 697 698 699 700 701 702 703 704 705 706 707 708 709 710 711 712 713
	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 {
714
			skb_get_timestampns(skb, &ts[0]);
715
			put_cmsg(msg, SOL_SOCKET, SCM_TIMESTAMPNS,
716
				 sizeof(ts[0]), &ts[0]);
717 718 719 720 721
		}
	}


	memset(ts, 0, sizeof(ts));
722 723
	if (sock_flag(sk, SOCK_TIMESTAMPING_SOFTWARE) &&
	    ktime_to_timespec_cond(skb->tstamp, ts + 0))
724 725 726
		empty = 0;
	if (shhwtstamps) {
		if (sock_flag(sk, SOCK_TIMESTAMPING_SYS_HARDWARE) &&
727
		    ktime_to_timespec_cond(shhwtstamps->syststamp, ts + 1))
728 729
			empty = 0;
		if (sock_flag(sk, SOCK_TIMESTAMPING_RAW_HARDWARE) &&
730
		    ktime_to_timespec_cond(shhwtstamps->hwtstamp, ts + 2))
731
			empty = 0;
732
	}
733 734 735
	if (!empty)
		put_cmsg(msg, SOL_SOCKET,
			 SCM_TIMESTAMPING, sizeof(ts), &ts);
736
}
737 738
EXPORT_SYMBOL_GPL(__sock_recv_timestamp);

739 740 741 742 743 744 745 746 747 748 749 750 751 752 753 754
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
755 756
static inline void sock_recv_drops(struct msghdr *msg, struct sock *sk,
				   struct sk_buff *skb)
757 758 759 760 761 762
{
	if (sock_flag(sk, SOCK_RXQ_OVFL) && skb && skb->dropcount)
		put_cmsg(msg, SOL_SOCKET, SO_RXQ_OVFL,
			sizeof(__u32), &skb->dropcount);
}

763
void __sock_recv_ts_and_drops(struct msghdr *msg, struct sock *sk,
764 765 766 767 768
	struct sk_buff *skb)
{
	sock_recv_timestamp(msg, sk, skb);
	sock_recv_drops(msg, sk, skb);
}
769
EXPORT_SYMBOL_GPL(__sock_recv_ts_and_drops);
770

771 772
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
773 774 775 776 777 778 779 780 781 782 783 784
{
	struct sock_iocb *si = kiocb_to_siocb(iocb);

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

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

785 786 787 788 789 790 791 792
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);
}

793
int sock_recvmsg(struct socket *sock, struct msghdr *msg,
L
Linus Torvalds 已提交
794 795 796 797 798 799
		 size_t size, int flags)
{
	struct kiocb iocb;
	struct sock_iocb siocb;
	int ret;

800
	init_sync_kiocb(&iocb, NULL);
L
Linus Torvalds 已提交
801 802 803 804 805 806
	iocb.private = &siocb;
	ret = __sock_recvmsg(&iocb, sock, msg, size, flags);
	if (-EIOCBQUEUED == ret)
		ret = wait_on_sync_kiocb(&iocb);
	return ret;
}
807
EXPORT_SYMBOL(sock_recvmsg);
L
Linus Torvalds 已提交
808

809 810 811 812 813 814 815 816 817 818 819 820 821 822 823
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;
}

824 825 826 827 828 829 830 831 832 833 834 835 836 837 838
/**
 * 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.
 */
839 840
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
841 842 843 844 845 846 847 848 849
{
	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
	 */
850
	msg->msg_iov = (struct iovec *)vec, msg->msg_iovlen = num;
L
Linus Torvalds 已提交
851 852 853 854
	result = sock_recvmsg(sock, msg, size, flags);
	set_fs(oldfs);
	return result;
}
855
EXPORT_SYMBOL(kernel_recvmsg);
L
Linus Torvalds 已提交
856 857 858 859 860 861

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

862 863
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
864 865 866 867
{
	struct socket *sock;
	int flags;

868 869
	sock = file->private_data;

870 871 872
	flags = (file->f_flags & O_NONBLOCK) ? MSG_DONTWAIT : 0;
	/* more is a combination of MSG_MORE and MSG_SENDPAGE_NOTLAST */
	flags |= more;
873

874
	return kernel_sendpage(sock, page, offset, size, flags);
875
}
L
Linus Torvalds 已提交
876

J
Jens Axboe 已提交
877
static ssize_t sock_splice_read(struct file *file, loff_t *ppos,
878
				struct pipe_inode_info *pipe, size_t len,
J
Jens Axboe 已提交
879 880 881 882
				unsigned int flags)
{
	struct socket *sock = file->private_data;

883 884 885
	if (unlikely(!sock->ops->splice_read))
		return -EINVAL;

J
Jens Axboe 已提交
886 887 888
	return sock->ops->splice_read(sock, ppos, pipe, len, flags);
}

889
static struct sock_iocb *alloc_sock_iocb(struct kiocb *iocb,
890
					 struct sock_iocb *siocb)
891 892 893 894 895
{
	if (!is_sync_kiocb(iocb)) {
		siocb = kmalloc(sizeof(*siocb), GFP_KERNEL);
		if (!siocb)
			return NULL;
L
Linus Torvalds 已提交
896 897 898
		iocb->ki_dtor = sock_aio_dtor;
	}

899 900 901
	siocb->kiocb = iocb;
	iocb->private = siocb;
	return siocb;
L
Linus Torvalds 已提交
902 903
}

904
static ssize_t do_sock_read(struct msghdr *msg, struct kiocb *iocb,
905 906
		struct file *file, const struct iovec *iov,
		unsigned long nr_segs)
907 908 909 910
{
	struct socket *sock = file->private_data;
	size_t size = 0;
	int i;
L
Linus Torvalds 已提交
911

912 913
	for (i = 0; i < nr_segs; i++)
		size += iov[i].iov_len;
L
Linus Torvalds 已提交
914

915 916 917 918
	msg->msg_name = NULL;
	msg->msg_namelen = 0;
	msg->msg_control = NULL;
	msg->msg_controllen = 0;
919
	msg->msg_iov = (struct iovec *)iov;
920 921 922 923 924 925
	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);
}

926 927
static ssize_t sock_aio_read(struct kiocb *iocb, const struct iovec *iov,
				unsigned long nr_segs, loff_t pos)
928 929 930
{
	struct sock_iocb siocb, *x;

L
Linus Torvalds 已提交
931 932
	if (pos != 0)
		return -ESPIPE;
933

K
Kent Overstreet 已提交
934
	if (iocb->ki_nbytes == 0)	/* Match SYS5 behaviour */
L
Linus Torvalds 已提交
935 936
		return 0;

937 938

	x = alloc_sock_iocb(iocb, &siocb);
939 940
	if (!x)
		return -ENOMEM;
941
	return do_sock_read(&x->async_msg, iocb, iocb->ki_filp, iov, nr_segs);
L
Linus Torvalds 已提交
942 943
}

944
static ssize_t do_sock_write(struct msghdr *msg, struct kiocb *iocb,
945 946
			struct file *file, const struct iovec *iov,
			unsigned long nr_segs)
L
Linus Torvalds 已提交
947
{
948 949 950
	struct socket *sock = file->private_data;
	size_t size = 0;
	int i;
L
Linus Torvalds 已提交
951

952 953
	for (i = 0; i < nr_segs; i++)
		size += iov[i].iov_len;
L
Linus Torvalds 已提交
954

955 956 957 958
	msg->msg_name = NULL;
	msg->msg_namelen = 0;
	msg->msg_control = NULL;
	msg->msg_controllen = 0;
959
	msg->msg_iov = (struct iovec *)iov;
960 961 962 963
	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
964

965
	return __sock_sendmsg(iocb, sock, msg, size);
L
Linus Torvalds 已提交
966 967
}

968 969
static ssize_t sock_aio_write(struct kiocb *iocb, const struct iovec *iov,
			  unsigned long nr_segs, loff_t pos)
970 971
{
	struct sock_iocb siocb, *x;
L
Linus Torvalds 已提交
972

973 974
	if (pos != 0)
		return -ESPIPE;
975 976

	x = alloc_sock_iocb(iocb, &siocb);
977 978
	if (!x)
		return -ENOMEM;
L
Linus Torvalds 已提交
979

980
	return do_sock_write(&x->async_msg, iocb, iocb->ki_filp, iov, nr_segs);
L
Linus Torvalds 已提交
981 982 983 984 985 986 987
}

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

A
Arjan van de Ven 已提交
988
static DEFINE_MUTEX(br_ioctl_mutex);
989
static int (*br_ioctl_hook) (struct net *, unsigned int cmd, void __user *arg);
L
Linus Torvalds 已提交
990

991
void brioctl_set(int (*hook) (struct net *, unsigned int, void __user *))
L
Linus Torvalds 已提交
992
{
A
Arjan van de Ven 已提交
993
	mutex_lock(&br_ioctl_mutex);
L
Linus Torvalds 已提交
994
	br_ioctl_hook = hook;
A
Arjan van de Ven 已提交
995
	mutex_unlock(&br_ioctl_mutex);
L
Linus Torvalds 已提交
996 997 998
}
EXPORT_SYMBOL(brioctl_set);

A
Arjan van de Ven 已提交
999
static DEFINE_MUTEX(vlan_ioctl_mutex);
1000
static int (*vlan_ioctl_hook) (struct net *, void __user *arg);
L
Linus Torvalds 已提交
1001

1002
void vlan_ioctl_set(int (*hook) (struct net *, void __user *))
L
Linus Torvalds 已提交
1003
{
A
Arjan van de Ven 已提交
1004
	mutex_lock(&vlan_ioctl_mutex);
L
Linus Torvalds 已提交
1005
	vlan_ioctl_hook = hook;
A
Arjan van de Ven 已提交
1006
	mutex_unlock(&vlan_ioctl_mutex);
L
Linus Torvalds 已提交
1007 1008 1009
}
EXPORT_SYMBOL(vlan_ioctl_set);

A
Arjan van de Ven 已提交
1010
static DEFINE_MUTEX(dlci_ioctl_mutex);
1011
static int (*dlci_ioctl_hook) (unsigned int, void __user *);
L
Linus Torvalds 已提交
1012

1013
void dlci_ioctl_set(int (*hook) (unsigned int, void __user *))
L
Linus Torvalds 已提交
1014
{
A
Arjan van de Ven 已提交
1015
	mutex_lock(&dlci_ioctl_mutex);
L
Linus Torvalds 已提交
1016
	dlci_ioctl_hook = hook;
A
Arjan van de Ven 已提交
1017
	mutex_unlock(&dlci_ioctl_mutex);
L
Linus Torvalds 已提交
1018 1019 1020
}
EXPORT_SYMBOL(dlci_ioctl_set);

1021 1022 1023 1024 1025 1026 1027 1028 1029 1030 1031 1032 1033 1034 1035 1036 1037 1038
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
1039 1040 1041 1042 1043 1044 1045 1046
/*
 *	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;
1047
	struct sock *sk;
L
Linus Torvalds 已提交
1048 1049
	void __user *argp = (void __user *)arg;
	int pid, err;
1050
	struct net *net;
L
Linus Torvalds 已提交
1051

1052
	sock = file->private_data;
1053
	sk = sock->sk;
1054
	net = sock_net(sk);
L
Linus Torvalds 已提交
1055
	if (cmd >= SIOCDEVPRIVATE && cmd <= (SIOCDEVPRIVATE + 15)) {
1056
		err = dev_ioctl(net, cmd, argp);
L
Linus Torvalds 已提交
1057
	} else
J
Johannes Berg 已提交
1058
#ifdef CONFIG_WEXT_CORE
L
Linus Torvalds 已提交
1059
	if (cmd >= SIOCIWFIRST && cmd <= SIOCIWLAST) {
1060
		err = dev_ioctl(net, cmd, argp);
L
Linus Torvalds 已提交
1061
	} else
J
Johannes Berg 已提交
1062
#endif
1063
		switch (cmd) {
L
Linus Torvalds 已提交
1064 1065 1066 1067 1068 1069 1070 1071 1072
		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:
1073
			err = put_user(f_getown(sock->file),
1074
				       (int __user *)argp);
L
Linus Torvalds 已提交
1075 1076 1077 1078 1079 1080 1081 1082 1083
			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
1084
			mutex_lock(&br_ioctl_mutex);
1085
			if (br_ioctl_hook)
1086
				err = br_ioctl_hook(net, cmd, argp);
A
Arjan van de Ven 已提交
1087
			mutex_unlock(&br_ioctl_mutex);
L
Linus Torvalds 已提交
1088 1089 1090 1091 1092 1093 1094
			break;
		case SIOCGIFVLAN:
		case SIOCSIFVLAN:
			err = -ENOPKG;
			if (!vlan_ioctl_hook)
				request_module("8021q");

A
Arjan van de Ven 已提交
1095
			mutex_lock(&vlan_ioctl_mutex);
L
Linus Torvalds 已提交
1096
			if (vlan_ioctl_hook)
1097
				err = vlan_ioctl_hook(net, argp);
A
Arjan van de Ven 已提交
1098
			mutex_unlock(&vlan_ioctl_mutex);
L
Linus Torvalds 已提交
1099 1100 1101 1102 1103 1104 1105
			break;
		case SIOCADDDLCI:
		case SIOCDELDLCI:
			err = -ENOPKG;
			if (!dlci_ioctl_hook)
				request_module("dlci");

1106 1107
			mutex_lock(&dlci_ioctl_mutex);
			if (dlci_ioctl_hook)
L
Linus Torvalds 已提交
1108
				err = dlci_ioctl_hook(cmd, argp);
1109
			mutex_unlock(&dlci_ioctl_mutex);
L
Linus Torvalds 已提交
1110 1111
			break;
		default:
1112
			err = sock_do_ioctl(net, sock, cmd, arg);
L
Linus Torvalds 已提交
1113
			break;
1114
		}
L
Linus Torvalds 已提交
1115 1116 1117 1118 1119 1120 1121
	return err;
}

int sock_create_lite(int family, int type, int protocol, struct socket **res)
{
	int err;
	struct socket *sock = NULL;
1122

L
Linus Torvalds 已提交
1123 1124 1125 1126 1127 1128 1129 1130 1131 1132 1133
	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
1134 1135 1136 1137
	err = security_socket_post_create(sock, family, type, protocol, 1);
	if (err)
		goto out_release;

L
Linus Torvalds 已提交
1138 1139 1140
out:
	*res = sock;
	return err;
V
Venkat Yekkirala 已提交
1141 1142 1143 1144
out_release:
	sock_release(sock);
	sock = NULL;
	goto out;
L
Linus Torvalds 已提交
1145
}
1146
EXPORT_SYMBOL(sock_create_lite);
L
Linus Torvalds 已提交
1147 1148

/* No kernel lock held - perfect */
1149
static unsigned int sock_poll(struct file *file, poll_table *wait)
L
Linus Torvalds 已提交
1150
{
1151
	unsigned int busy_flag = 0;
L
Linus Torvalds 已提交
1152 1153 1154
	struct socket *sock;

	/*
1155
	 *      We can't return errors to poll, so it's either yes or no.
L
Linus Torvalds 已提交
1156
	 */
1157
	sock = file->private_data;
1158

1159
	if (sk_can_busy_loop(sock->sk)) {
1160
		/* this socket can poll_ll so tell the system call */
1161
		busy_flag = POLL_BUSY_LOOP;
1162 1163

		/* once, only if requested by syscall */
1164 1165
		if (wait && (wait->_key & POLL_BUSY_LOOP))
			sk_busy_loop(sock->sk, 1);
1166 1167
	}

1168
	return busy_flag | sock->ops->poll(file, sock, wait);
L
Linus Torvalds 已提交
1169 1170
}

1171
static int sock_mmap(struct file *file, struct vm_area_struct *vma)
L
Linus Torvalds 已提交
1172
{
1173
	struct socket *sock = file->private_data;
L
Linus Torvalds 已提交
1174 1175 1176 1177

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

1178
static int sock_close(struct inode *inode, struct file *filp)
L
Linus Torvalds 已提交
1179 1180 1181 1182 1183 1184 1185 1186 1187 1188 1189 1190 1191
{
	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)
1192
 *	   or under socket lock
L
Linus Torvalds 已提交
1193 1194 1195 1196
 */

static int sock_fasync(int fd, struct file *filp, int on)
{
1197 1198
	struct socket *sock = filp->private_data;
	struct sock *sk = sock->sk;
1199
	struct socket_wq *wq;
L
Linus Torvalds 已提交
1200

1201
	if (sk == NULL)
L
Linus Torvalds 已提交
1202 1203 1204
		return -EINVAL;

	lock_sock(sk);
1205 1206
	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
1207

1208
	if (!wq->fasync_list)
1209 1210
		sock_reset_flag(sk, SOCK_FASYNC);
	else
E
Eric Dumazet 已提交
1211
		sock_set_flag(sk, SOCK_FASYNC);
L
Linus Torvalds 已提交
1212

1213
	release_sock(sk);
L
Linus Torvalds 已提交
1214 1215 1216
	return 0;
}

1217
/* This function may be called only under socket lock or callback_lock or rcu_lock */
L
Linus Torvalds 已提交
1218 1219 1220

int sock_wake_async(struct socket *sock, int how, int band)
{
1221 1222 1223 1224 1225 1226 1227 1228
	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
1229
		return -1;
1230
	}
1231
	switch (how) {
1232
	case SOCK_WAKE_WAITD:
L
Linus Torvalds 已提交
1233 1234 1235
		if (test_bit(SOCK_ASYNC_WAITDATA, &sock->flags))
			break;
		goto call_kill;
1236
	case SOCK_WAKE_SPACE:
L
Linus Torvalds 已提交
1237 1238 1239
		if (!test_and_clear_bit(SOCK_ASYNC_NOSPACE, &sock->flags))
			break;
		/* fall through */
1240
	case SOCK_WAKE_IO:
1241
call_kill:
1242
		kill_fasync(&wq->fasync_list, SIGIO, band);
L
Linus Torvalds 已提交
1243
		break;
1244
	case SOCK_WAKE_URG:
1245
		kill_fasync(&wq->fasync_list, SIGURG, band);
L
Linus Torvalds 已提交
1246
	}
1247
	rcu_read_unlock();
L
Linus Torvalds 已提交
1248 1249
	return 0;
}
1250
EXPORT_SYMBOL(sock_wake_async);
L
Linus Torvalds 已提交
1251

P
Pavel Emelyanov 已提交
1252
int __sock_create(struct net *net, int family, int type, int protocol,
1253
			 struct socket **res, int kern)
L
Linus Torvalds 已提交
1254 1255 1256
{
	int err;
	struct socket *sock;
1257
	const struct net_proto_family *pf;
L
Linus Torvalds 已提交
1258 1259

	/*
1260
	 *      Check protocol is in range
L
Linus Torvalds 已提交
1261 1262 1263 1264 1265 1266 1267 1268 1269 1270 1271 1272
	 */
	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) {
1273
		static int warned;
L
Linus Torvalds 已提交
1274 1275
		if (!warned) {
			warned = 1;
1276 1277
			printk(KERN_INFO "%s uses obsolete (PF_INET,SOCK_PACKET)\n",
			       current->comm);
L
Linus Torvalds 已提交
1278 1279 1280 1281 1282 1283 1284
		}
		family = PF_PACKET;
	}

	err = security_socket_create(family, type, protocol, kern);
	if (err)
		return err;
1285

1286 1287 1288 1289 1290 1291 1292
	/*
	 *	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) {
1293
		net_warn_ratelimited("socket: no more sockets\n");
1294 1295 1296 1297 1298 1299
		return -ENFILE;	/* Not exactly a match, but its the
				   closest posix thing */
	}

	sock->type = type;

1300
#ifdef CONFIG_MODULES
1301 1302 1303
	/* 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
1304 1305 1306
	 * requested real, full-featured networking support upon configuration.
	 * Otherwise module support will break!
	 */
E
Eric Dumazet 已提交
1307
	if (rcu_access_pointer(net_families[family]) == NULL)
1308
		request_module("net-pf-%d", family);
L
Linus Torvalds 已提交
1309 1310
#endif

1311 1312 1313 1314 1315
	rcu_read_lock();
	pf = rcu_dereference(net_families[family]);
	err = -EAFNOSUPPORT;
	if (!pf)
		goto out_release;
L
Linus Torvalds 已提交
1316 1317 1318 1319 1320

	/*
	 * We will call the ->create function, that possibly is in a loadable
	 * module, so we have to bump that loadable module refcnt first.
	 */
1321
	if (!try_module_get(pf->owner))
L
Linus Torvalds 已提交
1322 1323
		goto out_release;

1324 1325 1326
	/* Now protected by module ref count */
	rcu_read_unlock();

1327
	err = pf->create(net, sock, protocol, kern);
1328
	if (err < 0)
L
Linus Torvalds 已提交
1329
		goto out_module_put;
1330

L
Linus Torvalds 已提交
1331 1332 1333 1334
	/*
	 * Now to bump the refcnt of the [loadable] module that owns this
	 * socket at sock_release time we decrement its refcnt.
	 */
1335 1336 1337
	if (!try_module_get(sock->ops->owner))
		goto out_module_busy;

L
Linus Torvalds 已提交
1338 1339 1340 1341
	/*
	 * Now that we're done with the ->create function, the [loadable]
	 * module can have its refcnt decremented
	 */
1342
	module_put(pf->owner);
V
Venkat Yekkirala 已提交
1343 1344
	err = security_socket_post_create(sock, family, type, protocol, kern);
	if (err)
1345
		goto out_sock_release;
1346
	*res = sock;
L
Linus Torvalds 已提交
1347

1348 1349 1350 1351
	return 0;

out_module_busy:
	err = -EAFNOSUPPORT;
L
Linus Torvalds 已提交
1352
out_module_put:
1353 1354 1355
	sock->ops = NULL;
	module_put(pf->owner);
out_sock_release:
L
Linus Torvalds 已提交
1356
	sock_release(sock);
1357 1358 1359 1360 1361
	return err;

out_release:
	rcu_read_unlock();
	goto out_sock_release;
L
Linus Torvalds 已提交
1362
}
P
Pavel Emelyanov 已提交
1363
EXPORT_SYMBOL(__sock_create);
L
Linus Torvalds 已提交
1364 1365 1366

int sock_create(int family, int type, int protocol, struct socket **res)
{
1367
	return __sock_create(current->nsproxy->net_ns, family, type, protocol, res, 0);
L
Linus Torvalds 已提交
1368
}
1369
EXPORT_SYMBOL(sock_create);
L
Linus Torvalds 已提交
1370 1371 1372

int sock_create_kern(int family, int type, int protocol, struct socket **res)
{
1373
	return __sock_create(&init_net, family, type, protocol, res, 1);
L
Linus Torvalds 已提交
1374
}
1375
EXPORT_SYMBOL(sock_create_kern);
L
Linus Torvalds 已提交
1376

1377
SYSCALL_DEFINE3(socket, int, family, int, type, int, protocol)
L
Linus Torvalds 已提交
1378 1379 1380
{
	int retval;
	struct socket *sock;
1381 1382
	int flags;

1383 1384 1385 1386 1387 1388
	/* 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);

1389
	flags = type & ~SOCK_TYPE_MASK;
1390
	if (flags & ~(SOCK_CLOEXEC | SOCK_NONBLOCK))
1391 1392
		return -EINVAL;
	type &= SOCK_TYPE_MASK;
L
Linus Torvalds 已提交
1393

U
Ulrich Drepper 已提交
1394 1395 1396
	if (SOCK_NONBLOCK != O_NONBLOCK && (flags & SOCK_NONBLOCK))
		flags = (flags & ~SOCK_NONBLOCK) | O_NONBLOCK;

L
Linus Torvalds 已提交
1397 1398 1399 1400
	retval = sock_create(family, type, protocol, &sock);
	if (retval < 0)
		goto out;

1401
	retval = sock_map_fd(sock, flags & (O_CLOEXEC | O_NONBLOCK));
L
Linus Torvalds 已提交
1402 1403 1404 1405 1406 1407 1408 1409 1410 1411 1412 1413 1414 1415 1416 1417
	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.
 */

1418 1419
SYSCALL_DEFINE4(socketpair, int, family, int, type, int, protocol,
		int __user *, usockvec)
L
Linus Torvalds 已提交
1420 1421 1422
{
	struct socket *sock1, *sock2;
	int fd1, fd2, err;
A
Al Viro 已提交
1423
	struct file *newfile1, *newfile2;
1424 1425 1426
	int flags;

	flags = type & ~SOCK_TYPE_MASK;
1427
	if (flags & ~(SOCK_CLOEXEC | SOCK_NONBLOCK))
1428 1429
		return -EINVAL;
	type &= SOCK_TYPE_MASK;
L
Linus Torvalds 已提交
1430

U
Ulrich Drepper 已提交
1431 1432 1433
	if (SOCK_NONBLOCK != O_NONBLOCK && (flags & SOCK_NONBLOCK))
		flags = (flags & ~SOCK_NONBLOCK) | O_NONBLOCK;

L
Linus Torvalds 已提交
1434 1435 1436 1437 1438 1439 1440 1441 1442 1443 1444 1445 1446 1447
	/*
	 * 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);
1448
	if (err < 0)
L
Linus Torvalds 已提交
1449 1450
		goto out_release_both;

1451
	fd1 = get_unused_fd_flags(flags);
1452 1453
	if (unlikely(fd1 < 0)) {
		err = fd1;
A
Al Viro 已提交
1454
		goto out_release_both;
1455
	}
1456
	fd2 = get_unused_fd_flags(flags);
1457 1458
	if (unlikely(fd2 < 0)) {
		err = fd2;
1459 1460 1461 1462
		put_unused_fd(fd1);
		goto out_release_both;
	}

1463
	newfile1 = sock_alloc_file(sock1, flags, NULL);
1464 1465 1466 1467 1468 1469 1470
	if (unlikely(IS_ERR(newfile1))) {
		err = PTR_ERR(newfile1);
		put_unused_fd(fd1);
		put_unused_fd(fd2);
		goto out_release_both;
	}

1471
	newfile2 = sock_alloc_file(sock2, flags, NULL);
1472 1473
	if (IS_ERR(newfile2)) {
		err = PTR_ERR(newfile2);
1474 1475
		fput(newfile1);
		put_unused_fd(fd1);
1476
		put_unused_fd(fd2);
1477 1478
		sock_release(sock2);
		goto out;
A
Al Viro 已提交
1479 1480
	}

A
Al Viro 已提交
1481
	audit_fd_pair(fd1, fd2);
A
Al Viro 已提交
1482 1483
	fd_install(fd1, newfile1);
	fd_install(fd2, newfile2);
L
Linus Torvalds 已提交
1484 1485 1486 1487
	/* fd1 and fd2 may be already another descriptors.
	 * Not kernel problem.
	 */

1488
	err = put_user(fd1, &usockvec[0]);
L
Linus Torvalds 已提交
1489 1490 1491 1492 1493 1494 1495 1496 1497 1498
	if (!err)
		err = put_user(fd2, &usockvec[1]);
	if (!err)
		return 0;

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

out_release_both:
1499
	sock_release(sock2);
L
Linus Torvalds 已提交
1500
out_release_1:
1501
	sock_release(sock1);
L
Linus Torvalds 已提交
1502 1503 1504 1505 1506 1507 1508 1509 1510 1511 1512 1513
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).
 */

1514
SYSCALL_DEFINE3(bind, int, fd, struct sockaddr __user *, umyaddr, int, addrlen)
L
Linus Torvalds 已提交
1515 1516
{
	struct socket *sock;
1517
	struct sockaddr_storage address;
1518
	int err, fput_needed;
L
Linus Torvalds 已提交
1519

1520
	sock = sockfd_lookup_light(fd, &err, &fput_needed);
1521
	if (sock) {
1522
		err = move_addr_to_kernel(umyaddr, addrlen, &address);
1523 1524
		if (err >= 0) {
			err = security_socket_bind(sock,
1525
						   (struct sockaddr *)&address,
1526
						   addrlen);
1527 1528
			if (!err)
				err = sock->ops->bind(sock,
1529
						      (struct sockaddr *)
1530
						      &address, addrlen);
L
Linus Torvalds 已提交
1531
		}
1532
		fput_light(sock->file, fput_needed);
1533
	}
L
Linus Torvalds 已提交
1534 1535 1536 1537 1538 1539 1540 1541 1542
	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.
 */

1543
SYSCALL_DEFINE2(listen, int, fd, int, backlog)
L
Linus Torvalds 已提交
1544 1545
{
	struct socket *sock;
1546
	int err, fput_needed;
1547
	int somaxconn;
1548 1549 1550

	sock = sockfd_lookup_light(fd, &err, &fput_needed);
	if (sock) {
1551
		somaxconn = sock_net(sock->sk)->core.sysctl_somaxconn;
1552
		if ((unsigned int)backlog > somaxconn)
1553
			backlog = somaxconn;
L
Linus Torvalds 已提交
1554 1555

		err = security_socket_listen(sock, backlog);
1556 1557
		if (!err)
			err = sock->ops->listen(sock, backlog);
L
Linus Torvalds 已提交
1558

1559
		fput_light(sock->file, fput_needed);
L
Linus Torvalds 已提交
1560 1561 1562 1563 1564 1565 1566 1567 1568 1569 1570 1571 1572 1573 1574 1575
	}
	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.
 */

1576 1577
SYSCALL_DEFINE4(accept4, int, fd, struct sockaddr __user *, upeer_sockaddr,
		int __user *, upeer_addrlen, int, flags)
L
Linus Torvalds 已提交
1578 1579
{
	struct socket *sock, *newsock;
1580
	struct file *newfile;
1581
	int err, len, newfd, fput_needed;
1582
	struct sockaddr_storage address;
L
Linus Torvalds 已提交
1583

1584
	if (flags & ~(SOCK_CLOEXEC | SOCK_NONBLOCK))
U
Ulrich Drepper 已提交
1585 1586 1587 1588 1589
		return -EINVAL;

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

1590
	sock = sockfd_lookup_light(fd, &err, &fput_needed);
L
Linus Torvalds 已提交
1591 1592 1593 1594
	if (!sock)
		goto out;

	err = -ENFILE;
1595 1596
	newsock = sock_alloc();
	if (!newsock)
L
Linus Torvalds 已提交
1597 1598 1599 1600 1601 1602 1603 1604 1605 1606 1607
		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);

1608
	newfd = get_unused_fd_flags(flags);
1609 1610
	if (unlikely(newfd < 0)) {
		err = newfd;
1611 1612
		sock_release(newsock);
		goto out_put;
1613
	}
1614
	newfile = sock_alloc_file(newsock, flags, sock->sk->sk_prot_creator->name);
1615 1616 1617 1618 1619 1620
	if (unlikely(IS_ERR(newfile))) {
		err = PTR_ERR(newfile);
		put_unused_fd(newfd);
		sock_release(newsock);
		goto out_put;
	}
1621

1622 1623
	err = security_socket_accept(sock, newsock);
	if (err)
1624
		goto out_fd;
1625

L
Linus Torvalds 已提交
1626 1627
	err = sock->ops->accept(sock, newsock, sock->file->f_flags);
	if (err < 0)
1628
		goto out_fd;
L
Linus Torvalds 已提交
1629 1630

	if (upeer_sockaddr) {
1631
		if (newsock->ops->getname(newsock, (struct sockaddr *)&address,
1632
					  &len, 2) < 0) {
L
Linus Torvalds 已提交
1633
			err = -ECONNABORTED;
1634
			goto out_fd;
L
Linus Torvalds 已提交
1635
		}
1636
		err = move_addr_to_user(&address,
1637
					len, upeer_sockaddr, upeer_addrlen);
L
Linus Torvalds 已提交
1638
		if (err < 0)
1639
			goto out_fd;
L
Linus Torvalds 已提交
1640 1641 1642 1643
	}

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

1644 1645
	fd_install(newfd, newfile);
	err = newfd;
L
Linus Torvalds 已提交
1646 1647

out_put:
1648
	fput_light(sock->file, fput_needed);
L
Linus Torvalds 已提交
1649 1650
out:
	return err;
1651
out_fd:
1652
	fput(newfile);
1653
	put_unused_fd(newfd);
L
Linus Torvalds 已提交
1654 1655 1656
	goto out_put;
}

1657 1658
SYSCALL_DEFINE3(accept, int, fd, struct sockaddr __user *, upeer_sockaddr,
		int __user *, upeer_addrlen)
U
Ulrich Drepper 已提交
1659
{
U
Ulrich Drepper 已提交
1660
	return sys_accept4(fd, upeer_sockaddr, upeer_addrlen, 0);
U
Ulrich Drepper 已提交
1661 1662
}

L
Linus Torvalds 已提交
1663 1664 1665 1666 1667 1668 1669 1670 1671 1672 1673 1674
/*
 *	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.
 */

1675 1676
SYSCALL_DEFINE3(connect, int, fd, struct sockaddr __user *, uservaddr,
		int, addrlen)
L
Linus Torvalds 已提交
1677 1678
{
	struct socket *sock;
1679
	struct sockaddr_storage address;
1680
	int err, fput_needed;
L
Linus Torvalds 已提交
1681

1682
	sock = sockfd_lookup_light(fd, &err, &fput_needed);
L
Linus Torvalds 已提交
1683 1684
	if (!sock)
		goto out;
1685
	err = move_addr_to_kernel(uservaddr, addrlen, &address);
L
Linus Torvalds 已提交
1686 1687 1688
	if (err < 0)
		goto out_put;

1689
	err =
1690
	    security_socket_connect(sock, (struct sockaddr *)&address, addrlen);
L
Linus Torvalds 已提交
1691 1692 1693
	if (err)
		goto out_put;

1694
	err = sock->ops->connect(sock, (struct sockaddr *)&address, addrlen,
L
Linus Torvalds 已提交
1695 1696
				 sock->file->f_flags);
out_put:
1697
	fput_light(sock->file, fput_needed);
L
Linus Torvalds 已提交
1698 1699 1700 1701 1702 1703 1704 1705 1706
out:
	return err;
}

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

1707 1708
SYSCALL_DEFINE3(getsockname, int, fd, struct sockaddr __user *, usockaddr,
		int __user *, usockaddr_len)
L
Linus Torvalds 已提交
1709 1710
{
	struct socket *sock;
1711
	struct sockaddr_storage address;
1712
	int len, err, fput_needed;
1713

1714
	sock = sockfd_lookup_light(fd, &err, &fput_needed);
L
Linus Torvalds 已提交
1715 1716 1717 1718 1719 1720 1721
	if (!sock)
		goto out;

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

1722
	err = sock->ops->getname(sock, (struct sockaddr *)&address, &len, 0);
L
Linus Torvalds 已提交
1723 1724
	if (err)
		goto out_put;
1725
	err = move_addr_to_user(&address, len, usockaddr, usockaddr_len);
L
Linus Torvalds 已提交
1726 1727

out_put:
1728
	fput_light(sock->file, fput_needed);
L
Linus Torvalds 已提交
1729 1730 1731 1732 1733 1734 1735 1736 1737
out:
	return err;
}

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

1738 1739
SYSCALL_DEFINE3(getpeername, int, fd, struct sockaddr __user *, usockaddr,
		int __user *, usockaddr_len)
L
Linus Torvalds 已提交
1740 1741
{
	struct socket *sock;
1742
	struct sockaddr_storage address;
1743
	int len, err, fput_needed;
L
Linus Torvalds 已提交
1744

1745 1746
	sock = sockfd_lookup_light(fd, &err, &fput_needed);
	if (sock != NULL) {
L
Linus Torvalds 已提交
1747 1748
		err = security_socket_getpeername(sock);
		if (err) {
1749
			fput_light(sock->file, fput_needed);
L
Linus Torvalds 已提交
1750 1751 1752
			return err;
		}

1753
		err =
1754
		    sock->ops->getname(sock, (struct sockaddr *)&address, &len,
1755
				       1);
L
Linus Torvalds 已提交
1756
		if (!err)
1757
			err = move_addr_to_user(&address, len, usockaddr,
1758
						usockaddr_len);
1759
		fput_light(sock->file, fput_needed);
L
Linus Torvalds 已提交
1760 1761 1762 1763 1764 1765 1766 1767 1768 1769
	}
	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.
 */

1770
SYSCALL_DEFINE6(sendto, int, fd, void __user *, buff, size_t, len,
1771
		unsigned int, flags, struct sockaddr __user *, addr,
1772
		int, addr_len)
L
Linus Torvalds 已提交
1773 1774
{
	struct socket *sock;
1775
	struct sockaddr_storage address;
L
Linus Torvalds 已提交
1776 1777 1778
	int err;
	struct msghdr msg;
	struct iovec iov;
1779 1780
	int fput_needed;

1781 1782
	if (len > INT_MAX)
		len = INT_MAX;
1783 1784
	sock = sockfd_lookup_light(fd, &err, &fput_needed);
	if (!sock)
1785
		goto out;
1786

1787 1788 1789 1790 1791 1792 1793 1794
	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;
1795
	if (addr) {
1796
		err = move_addr_to_kernel(addr, addr_len, &address);
L
Linus Torvalds 已提交
1797 1798
		if (err < 0)
			goto out_put;
1799
		msg.msg_name = (struct sockaddr *)&address;
1800
		msg.msg_namelen = addr_len;
L
Linus Torvalds 已提交
1801 1802 1803 1804 1805 1806
	}
	if (sock->file->f_flags & O_NONBLOCK)
		flags |= MSG_DONTWAIT;
	msg.msg_flags = flags;
	err = sock_sendmsg(sock, &msg, len);

1807
out_put:
1808
	fput_light(sock->file, fput_needed);
1809
out:
L
Linus Torvalds 已提交
1810 1811 1812 1813
	return err;
}

/*
1814
 *	Send a datagram down a socket.
L
Linus Torvalds 已提交
1815 1816
 */

1817
SYSCALL_DEFINE4(send, int, fd, void __user *, buff, size_t, len,
1818
		unsigned int, flags)
L
Linus Torvalds 已提交
1819 1820 1821 1822 1823
{
	return sys_sendto(fd, buff, len, flags, NULL, 0);
}

/*
1824
 *	Receive a frame from the socket and optionally record the address of the
L
Linus Torvalds 已提交
1825 1826 1827 1828
 *	sender. We verify the buffers are writable and if needed move the
 *	sender address from kernel to user space.
 */

1829
SYSCALL_DEFINE6(recvfrom, int, fd, void __user *, ubuf, size_t, size,
1830
		unsigned int, flags, struct sockaddr __user *, addr,
1831
		int __user *, addr_len)
L
Linus Torvalds 已提交
1832 1833 1834 1835
{
	struct socket *sock;
	struct iovec iov;
	struct msghdr msg;
1836
	struct sockaddr_storage address;
1837
	int err, err2;
1838 1839
	int fput_needed;

1840 1841
	if (size > INT_MAX)
		size = INT_MAX;
1842
	sock = sockfd_lookup_light(fd, &err, &fput_needed);
L
Linus Torvalds 已提交
1843
	if (!sock)
1844
		goto out;
L
Linus Torvalds 已提交
1845

1846 1847 1848 1849 1850 1851
	msg.msg_control = NULL;
	msg.msg_controllen = 0;
	msg.msg_iovlen = 1;
	msg.msg_iov = &iov;
	iov.iov_len = size;
	iov.iov_base = ubuf;
1852 1853
	msg.msg_name = (struct sockaddr *)&address;
	msg.msg_namelen = sizeof(address);
L
Linus Torvalds 已提交
1854 1855
	if (sock->file->f_flags & O_NONBLOCK)
		flags |= MSG_DONTWAIT;
1856
	err = sock_recvmsg(sock, &msg, size, flags);
L
Linus Torvalds 已提交
1857

1858
	if (err >= 0 && addr != NULL) {
1859
		err2 = move_addr_to_user(&address,
1860
					 msg.msg_namelen, addr, addr_len);
1861 1862
		if (err2 < 0)
			err = err2;
L
Linus Torvalds 已提交
1863
	}
1864 1865

	fput_light(sock->file, fput_needed);
1866
out:
L
Linus Torvalds 已提交
1867 1868 1869 1870
	return err;
}

/*
1871
 *	Receive a datagram from a socket.
L
Linus Torvalds 已提交
1872 1873
 */

1874
asmlinkage long sys_recv(int fd, void __user *ubuf, size_t size,
1875
			 unsigned int flags)
L
Linus Torvalds 已提交
1876 1877 1878 1879 1880 1881 1882 1883 1884
{
	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.
 */

1885 1886
SYSCALL_DEFINE5(setsockopt, int, fd, int, level, int, optname,
		char __user *, optval, int, optlen)
L
Linus Torvalds 已提交
1887
{
1888
	int err, fput_needed;
L
Linus Torvalds 已提交
1889 1890 1891 1892
	struct socket *sock;

	if (optlen < 0)
		return -EINVAL;
1893 1894 1895 1896

	sock = sockfd_lookup_light(fd, &err, &fput_needed);
	if (sock != NULL) {
		err = security_socket_setsockopt(sock, level, optname);
1897 1898
		if (err)
			goto out_put;
L
Linus Torvalds 已提交
1899 1900

		if (level == SOL_SOCKET)
1901 1902 1903
			err =
			    sock_setsockopt(sock, level, optname, optval,
					    optlen);
L
Linus Torvalds 已提交
1904
		else
1905 1906 1907
			err =
			    sock->ops->setsockopt(sock, level, optname, optval,
						  optlen);
1908 1909
out_put:
		fput_light(sock->file, fput_needed);
L
Linus Torvalds 已提交
1910 1911 1912 1913 1914 1915 1916 1917 1918
	}
	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.
 */

1919 1920
SYSCALL_DEFINE5(getsockopt, int, fd, int, level, int, optname,
		char __user *, optval, int __user *, optlen)
L
Linus Torvalds 已提交
1921
{
1922
	int err, fput_needed;
L
Linus Torvalds 已提交
1923 1924
	struct socket *sock;

1925 1926
	sock = sockfd_lookup_light(fd, &err, &fput_needed);
	if (sock != NULL) {
1927 1928 1929
		err = security_socket_getsockopt(sock, level, optname);
		if (err)
			goto out_put;
L
Linus Torvalds 已提交
1930 1931

		if (level == SOL_SOCKET)
1932 1933 1934
			err =
			    sock_getsockopt(sock, level, optname, optval,
					    optlen);
L
Linus Torvalds 已提交
1935
		else
1936 1937 1938
			err =
			    sock->ops->getsockopt(sock, level, optname, optval,
						  optlen);
1939 1940
out_put:
		fput_light(sock->file, fput_needed);
L
Linus Torvalds 已提交
1941 1942 1943 1944 1945 1946 1947 1948
	}
	return err;
}

/*
 *	Shutdown a socket.
 */

1949
SYSCALL_DEFINE2(shutdown, int, fd, int, how)
L
Linus Torvalds 已提交
1950
{
1951
	int err, fput_needed;
L
Linus Torvalds 已提交
1952 1953
	struct socket *sock;

1954 1955
	sock = sockfd_lookup_light(fd, &err, &fput_needed);
	if (sock != NULL) {
L
Linus Torvalds 已提交
1956
		err = security_socket_shutdown(sock, how);
1957 1958 1959
		if (!err)
			err = sock->ops->shutdown(sock, how);
		fput_light(sock->file, fput_needed);
L
Linus Torvalds 已提交
1960 1961 1962 1963
	}
	return err;
}

1964
/* A couple of helpful macros for getting the address of the 32/64 bit
L
Linus Torvalds 已提交
1965 1966 1967 1968 1969 1970
 * 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)

1971 1972 1973 1974 1975
struct used_address {
	struct sockaddr_storage name;
	unsigned int name_len;
};

1976
static int ___sys_sendmsg(struct socket *sock, struct msghdr __user *msg,
1977
			 struct msghdr *msg_sys, unsigned int flags,
1978
			 struct used_address *used_address)
L
Linus Torvalds 已提交
1979
{
1980 1981
	struct compat_msghdr __user *msg_compat =
	    (struct compat_msghdr __user *)msg;
1982
	struct sockaddr_storage address;
L
Linus Torvalds 已提交
1983
	struct iovec iovstack[UIO_FASTIOV], *iov = iovstack;
1984
	unsigned char ctl[sizeof(struct cmsghdr) + 20]
1985 1986
	    __attribute__ ((aligned(sizeof(__kernel_size_t))));
	/* 20 is size of ipv6_pktinfo */
L
Linus Torvalds 已提交
1987
	unsigned char *ctl_buf = ctl;
E
Eric Dumazet 已提交
1988
	int err, ctl_len, total_len;
1989

L
Linus Torvalds 已提交
1990 1991
	err = -EFAULT;
	if (MSG_CMSG_COMPAT & flags) {
1992
		if (get_compat_msghdr(msg_sys, msg_compat))
L
Linus Torvalds 已提交
1993
			return -EFAULT;
1994
	} else if (copy_from_user(msg_sys, msg, sizeof(struct msghdr)))
L
Linus Torvalds 已提交
1995 1996
		return -EFAULT;

1997
	if (msg_sys->msg_iovlen > UIO_FASTIOV) {
E
Eric Dumazet 已提交
1998 1999 2000 2001 2002 2003
		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
2004
		if (!iov)
2005
			goto out;
L
Linus Torvalds 已提交
2006 2007 2008 2009
	}

	/* This will also move the address data into kernel space */
	if (MSG_CMSG_COMPAT & flags) {
2010
		err = verify_compat_iovec(msg_sys, iov, &address, VERIFY_READ);
L
Linus Torvalds 已提交
2011
	} else
2012
		err = verify_iovec(msg_sys, iov, &address, VERIFY_READ);
2013
	if (err < 0)
L
Linus Torvalds 已提交
2014 2015 2016 2017 2018
		goto out_freeiov;
	total_len = err;

	err = -ENOBUFS;

2019
	if (msg_sys->msg_controllen > INT_MAX)
L
Linus Torvalds 已提交
2020
		goto out_freeiov;
2021
	ctl_len = msg_sys->msg_controllen;
L
Linus Torvalds 已提交
2022
	if ((MSG_CMSG_COMPAT & flags) && ctl_len) {
2023
		err =
2024
		    cmsghdr_from_user_compat_to_kern(msg_sys, sock->sk, ctl,
2025
						     sizeof(ctl));
L
Linus Torvalds 已提交
2026 2027
		if (err)
			goto out_freeiov;
2028 2029
		ctl_buf = msg_sys->msg_control;
		ctl_len = msg_sys->msg_controllen;
L
Linus Torvalds 已提交
2030
	} else if (ctl_len) {
2031
		if (ctl_len > sizeof(ctl)) {
L
Linus Torvalds 已提交
2032
			ctl_buf = sock_kmalloc(sock->sk, ctl_len, GFP_KERNEL);
2033
			if (ctl_buf == NULL)
L
Linus Torvalds 已提交
2034 2035 2036 2037
				goto out_freeiov;
		}
		err = -EFAULT;
		/*
2038
		 * Careful! Before this, msg_sys->msg_control contains a user pointer.
L
Linus Torvalds 已提交
2039 2040 2041
		 * Afterwards, it will be a kernel pointer. Thus the compiler-assisted
		 * checking falls down on this.
		 */
2042
		if (copy_from_user(ctl_buf,
2043
				   (void __user __force *)msg_sys->msg_control,
2044
				   ctl_len))
L
Linus Torvalds 已提交
2045
			goto out_freectl;
2046
		msg_sys->msg_control = ctl_buf;
L
Linus Torvalds 已提交
2047
	}
2048
	msg_sys->msg_flags = flags;
L
Linus Torvalds 已提交
2049 2050

	if (sock->file->f_flags & O_NONBLOCK)
2051
		msg_sys->msg_flags |= MSG_DONTWAIT;
2052 2053 2054 2055 2056 2057
	/*
	 * 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.
	 */
2058 2059 2060
	if (used_address && msg_sys->msg_name &&
	    used_address->name_len == msg_sys->msg_namelen &&
	    !memcmp(&used_address->name, msg_sys->msg_name,
2061 2062 2063 2064 2065 2066 2067 2068 2069 2070 2071
		    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;
2072 2073 2074
		if (msg_sys->msg_name)
			memcpy(&used_address->name, msg_sys->msg_name,
			       used_address->name_len);
2075
	}
L
Linus Torvalds 已提交
2076 2077

out_freectl:
2078
	if (ctl_buf != ctl)
L
Linus Torvalds 已提交
2079 2080 2081
		sock_kfree_s(sock->sk, ctl_buf, ctl_len);
out_freeiov:
	if (iov != iovstack)
E
Eric Dumazet 已提交
2082
		kfree(iov);
2083 2084 2085 2086 2087 2088 2089 2090
out:
	return err;
}

/*
 *	BSD sendmsg interface
 */

2091
long __sys_sendmsg(int fd, struct msghdr __user *msg, unsigned flags)
2092 2093 2094
{
	int fput_needed, err;
	struct msghdr msg_sys;
2095 2096 2097
	struct socket *sock;

	sock = sockfd_lookup_light(fd, &err, &fput_needed);
2098 2099 2100
	if (!sock)
		goto out;

2101
	err = ___sys_sendmsg(sock, msg, &msg_sys, flags, NULL);
2102

2103
	fput_light(sock->file, fput_needed);
2104
out:
L
Linus Torvalds 已提交
2105 2106 2107
	return err;
}

2108 2109 2110 2111 2112 2113 2114
SYSCALL_DEFINE3(sendmsg, int, fd, struct msghdr __user *, msg, unsigned int, flags)
{
	if (flags & MSG_CMSG_COMPAT)
		return -EINVAL;
	return __sys_sendmsg(fd, msg, flags);
}

2115 2116 2117 2118 2119 2120 2121 2122 2123 2124 2125 2126
/*
 *	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;
2127
	struct used_address used_address;
2128

2129 2130
	if (vlen > UIO_MAXIOV)
		vlen = UIO_MAXIOV;
2131 2132 2133 2134 2135 2136 2137

	datagrams = 0;

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

2138
	used_address.name_len = UINT_MAX;
2139 2140
	entry = mmsg;
	compat_entry = (struct compat_mmsghdr __user *)mmsg;
2141
	err = 0;
2142 2143 2144

	while (datagrams < vlen) {
		if (MSG_CMSG_COMPAT & flags) {
2145 2146
			err = ___sys_sendmsg(sock, (struct msghdr __user *)compat_entry,
					     &msg_sys, flags, &used_address);
2147 2148 2149 2150 2151
			if (err < 0)
				break;
			err = __put_user(err, &compat_entry->msg_len);
			++compat_entry;
		} else {
2152 2153 2154
			err = ___sys_sendmsg(sock,
					     (struct msghdr __user *)entry,
					     &msg_sys, flags, &used_address);
2155 2156 2157 2158 2159 2160 2161 2162 2163 2164 2165 2166 2167
			if (err < 0)
				break;
			err = put_user(err, &entry->msg_len);
			++entry;
		}

		if (err)
			break;
		++datagrams;
	}

	fput_light(sock->file, fput_needed);

2168 2169
	/* We only return an error if no datagrams were able to be sent */
	if (datagrams != 0)
2170 2171 2172 2173 2174 2175 2176 2177
		return datagrams;

	return err;
}

SYSCALL_DEFINE4(sendmmsg, int, fd, struct mmsghdr __user *, mmsg,
		unsigned int, vlen, unsigned int, flags)
{
2178 2179
	if (flags & MSG_CMSG_COMPAT)
		return -EINVAL;
2180 2181 2182
	return __sys_sendmmsg(fd, mmsg, vlen, flags);
}

2183
static int ___sys_recvmsg(struct socket *sock, struct msghdr __user *msg,
2184
			 struct msghdr *msg_sys, unsigned int flags, int nosec)
L
Linus Torvalds 已提交
2185
{
2186 2187
	struct compat_msghdr __user *msg_compat =
	    (struct compat_msghdr __user *)msg;
L
Linus Torvalds 已提交
2188
	struct iovec iovstack[UIO_FASTIOV];
2189
	struct iovec *iov = iovstack;
L
Linus Torvalds 已提交
2190
	unsigned long cmsg_ptr;
E
Eric Dumazet 已提交
2191
	int err, total_len, len;
L
Linus Torvalds 已提交
2192 2193

	/* kernel mode address */
2194
	struct sockaddr_storage addr;
L
Linus Torvalds 已提交
2195 2196 2197 2198

	/* user mode address pointers */
	struct sockaddr __user *uaddr;
	int __user *uaddr_len;
2199

L
Linus Torvalds 已提交
2200
	if (MSG_CMSG_COMPAT & flags) {
2201
		if (get_compat_msghdr(msg_sys, msg_compat))
L
Linus Torvalds 已提交
2202
			return -EFAULT;
2203
	} else if (copy_from_user(msg_sys, msg, sizeof(struct msghdr)))
2204
		return -EFAULT;
L
Linus Torvalds 已提交
2205

2206
	if (msg_sys->msg_iovlen > UIO_FASTIOV) {
E
Eric Dumazet 已提交
2207 2208 2209 2210 2211 2212
		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
2213
		if (!iov)
2214
			goto out;
L
Linus Torvalds 已提交
2215 2216 2217
	}

	/*
2218 2219
	 *      Save the user-mode address (verify_iovec will change the
	 *      kernel msghdr to use the kernel address space)
L
Linus Torvalds 已提交
2220
	 */
2221

2222
	uaddr = (__force void __user *)msg_sys->msg_name;
L
Linus Torvalds 已提交
2223 2224
	uaddr_len = COMPAT_NAMELEN(msg);
	if (MSG_CMSG_COMPAT & flags) {
2225
		err = verify_compat_iovec(msg_sys, iov, &addr, VERIFY_WRITE);
L
Linus Torvalds 已提交
2226
	} else
2227
		err = verify_iovec(msg_sys, iov, &addr, VERIFY_WRITE);
L
Linus Torvalds 已提交
2228 2229
	if (err < 0)
		goto out_freeiov;
2230
	total_len = err;
L
Linus Torvalds 已提交
2231

2232 2233
	cmsg_ptr = (unsigned long)msg_sys->msg_control;
	msg_sys->msg_flags = flags & (MSG_CMSG_CLOEXEC|MSG_CMSG_COMPAT);
2234

L
Linus Torvalds 已提交
2235 2236
	if (sock->file->f_flags & O_NONBLOCK)
		flags |= MSG_DONTWAIT;
2237 2238
	err = (nosec ? sock_recvmsg_nosec : sock_recvmsg)(sock, msg_sys,
							  total_len, flags);
L
Linus Torvalds 已提交
2239 2240 2241 2242 2243
	if (err < 0)
		goto out_freeiov;
	len = err;

	if (uaddr != NULL) {
2244
		err = move_addr_to_user(&addr,
2245
					msg_sys->msg_namelen, uaddr,
2246
					uaddr_len);
L
Linus Torvalds 已提交
2247 2248 2249
		if (err < 0)
			goto out_freeiov;
	}
2250
	err = __put_user((msg_sys->msg_flags & ~MSG_CMSG_COMPAT),
2251
			 COMPAT_FLAGS(msg));
L
Linus Torvalds 已提交
2252 2253 2254
	if (err)
		goto out_freeiov;
	if (MSG_CMSG_COMPAT & flags)
2255
		err = __put_user((unsigned long)msg_sys->msg_control - cmsg_ptr,
L
Linus Torvalds 已提交
2256 2257
				 &msg_compat->msg_controllen);
	else
2258
		err = __put_user((unsigned long)msg_sys->msg_control - cmsg_ptr,
L
Linus Torvalds 已提交
2259 2260 2261 2262 2263 2264 2265
				 &msg->msg_controllen);
	if (err)
		goto out_freeiov;
	err = len;

out_freeiov:
	if (iov != iovstack)
E
Eric Dumazet 已提交
2266
		kfree(iov);
2267 2268 2269 2270 2271 2272 2273 2274
out:
	return err;
}

/*
 *	BSD recvmsg interface
 */

2275
long __sys_recvmsg(int fd, struct msghdr __user *msg, unsigned flags)
2276 2277 2278
{
	int fput_needed, err;
	struct msghdr msg_sys;
2279 2280 2281
	struct socket *sock;

	sock = sockfd_lookup_light(fd, &err, &fput_needed);
2282 2283 2284
	if (!sock)
		goto out;

2285
	err = ___sys_recvmsg(sock, msg, &msg_sys, flags, 0);
2286

2287
	fput_light(sock->file, fput_needed);
L
Linus Torvalds 已提交
2288 2289 2290 2291
out:
	return err;
}

2292 2293 2294 2295 2296 2297 2298 2299
SYSCALL_DEFINE3(recvmsg, int, fd, struct msghdr __user *, msg,
		unsigned int, flags)
{
	if (flags & MSG_CMSG_COMPAT)
		return -EINVAL;
	return __sys_recvmsg(fd, msg, flags);
}

2300 2301 2302 2303 2304 2305 2306 2307 2308 2309
/*
 *     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;
2310
	struct compat_mmsghdr __user *compat_entry;
2311 2312 2313 2314 2315 2316 2317 2318 2319 2320 2321 2322 2323 2324 2325 2326 2327 2328 2329
	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;
2330
	compat_entry = (struct compat_mmsghdr __user *)mmsg;
2331 2332 2333 2334 2335

	while (datagrams < vlen) {
		/*
		 * No need to ask LSM for more than the first datagram.
		 */
2336
		if (MSG_CMSG_COMPAT & flags) {
2337 2338 2339
			err = ___sys_recvmsg(sock, (struct msghdr __user *)compat_entry,
					     &msg_sys, flags & ~MSG_WAITFORONE,
					     datagrams);
2340 2341 2342 2343 2344
			if (err < 0)
				break;
			err = __put_user(err, &compat_entry->msg_len);
			++compat_entry;
		} else {
2345 2346 2347 2348
			err = ___sys_recvmsg(sock,
					     (struct msghdr __user *)entry,
					     &msg_sys, flags & ~MSG_WAITFORONE,
					     datagrams);
2349 2350 2351 2352 2353 2354
			if (err < 0)
				break;
			err = put_user(err, &entry->msg_len);
			++entry;
		}

2355 2356 2357 2358
		if (err)
			break;
		++datagrams;

2359 2360 2361 2362
		/* MSG_WAITFORONE turns on MSG_DONTWAIT after one packet */
		if (flags & MSG_WAITFORONE)
			flags |= MSG_DONTWAIT;

2363 2364 2365 2366 2367 2368 2369 2370 2371 2372 2373 2374 2375 2376 2377 2378 2379 2380 2381 2382
		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
2383

2384 2385 2386 2387 2388 2389 2390 2391 2392 2393 2394 2395 2396 2397 2398 2399 2400 2401 2402 2403 2404 2405 2406 2407 2408 2409 2410 2411 2412 2413 2414
	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;

2415 2416 2417
	if (flags & MSG_CMSG_COMPAT)
		return -EINVAL;

2418 2419 2420 2421 2422 2423 2424 2425 2426 2427 2428 2429 2430 2431 2432 2433
	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
2434 2435
/* Argument list sizes for sys_socketcall */
#define AL(x) ((x) * sizeof(unsigned long))
2436
static const unsigned char nargs[21] = {
2437 2438 2439
	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),
2440
	AL(4), AL(5), AL(4)
2441 2442
};

L
Linus Torvalds 已提交
2443 2444 2445
#undef AL

/*
2446
 *	System call vectors.
L
Linus Torvalds 已提交
2447 2448 2449
 *
 *	Argument checking cleaned up. Saved 20% in size.
 *  This function doesn't need to set the kernel lock because
2450
 *  it is set by the callees.
L
Linus Torvalds 已提交
2451 2452
 */

2453
SYSCALL_DEFINE2(socketcall, int, call, unsigned long __user *, args)
L
Linus Torvalds 已提交
2454
{
2455
	unsigned long a[AUDITSC_ARGS];
2456
	unsigned long a0, a1;
L
Linus Torvalds 已提交
2457
	int err;
2458
	unsigned int len;
L
Linus Torvalds 已提交
2459

2460
	if (call < 1 || call > SYS_SENDMMSG)
L
Linus Torvalds 已提交
2461 2462
		return -EINVAL;

2463 2464 2465 2466
	len = nargs[call];
	if (len > sizeof(a))
		return -EINVAL;

L
Linus Torvalds 已提交
2467
	/* copy_from_user should be SMP safe. */
2468
	if (copy_from_user(a, args, len))
L
Linus Torvalds 已提交
2469
		return -EFAULT;
2470

2471 2472 2473
	err = audit_socketcall(nargs[call] / sizeof(unsigned long), a);
	if (err)
		return err;
2474

2475 2476 2477 2478 2479 2480 2481 2482 2483 2484 2485 2486 2487 2488 2489 2490 2491
	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
2492 2493
		err = sys_accept4(a0, (struct sockaddr __user *)a1,
				  (int __user *)a[2], 0);
2494 2495 2496 2497 2498 2499 2500 2501 2502 2503 2504 2505 2506 2507 2508 2509 2510 2511 2512 2513 2514 2515 2516 2517 2518 2519 2520 2521 2522 2523 2524 2525 2526 2527 2528 2529 2530 2531 2532 2533 2534 2535 2536
		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;
2537 2538 2539
	case SYS_SENDMMSG:
		err = sys_sendmmsg(a0, (struct mmsghdr __user *)a1, a[2], a[3]);
		break;
2540 2541 2542
	case SYS_RECVMSG:
		err = sys_recvmsg(a0, (struct msghdr __user *)a1, a[2]);
		break;
2543 2544 2545 2546
	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
2547 2548 2549
	case SYS_ACCEPT4:
		err = sys_accept4(a0, (struct sockaddr __user *)a1,
				  (int __user *)a[2], a[3]);
U
Ulrich Drepper 已提交
2550
		break;
2551 2552 2553
	default:
		err = -EINVAL;
		break;
L
Linus Torvalds 已提交
2554 2555 2556 2557
	}
	return err;
}

2558
#endif				/* __ARCH_WANT_SYS_SOCKETCALL */
L
Linus Torvalds 已提交
2559

2560 2561 2562 2563
/**
 *	sock_register - add a socket protocol handler
 *	@ops: description of protocol
 *
L
Linus Torvalds 已提交
2564 2565
 *	This function is called by a protocol handler that wants to
 *	advertise its address family, and have it linked into the
2566 2567
 *	socket interface. The value ops->family coresponds to the
 *	socket system call protocol family.
L
Linus Torvalds 已提交
2568
 */
2569
int sock_register(const struct net_proto_family *ops)
L
Linus Torvalds 已提交
2570 2571 2572 2573
{
	int err;

	if (ops->family >= NPROTO) {
2574 2575
		printk(KERN_CRIT "protocol %d >= NPROTO(%d)\n", ops->family,
		       NPROTO);
L
Linus Torvalds 已提交
2576 2577
		return -ENOBUFS;
	}
2578 2579

	spin_lock(&net_family_lock);
E
Eric Dumazet 已提交
2580 2581
	if (rcu_dereference_protected(net_families[ops->family],
				      lockdep_is_held(&net_family_lock)))
2582 2583
		err = -EEXIST;
	else {
2584
		rcu_assign_pointer(net_families[ops->family], ops);
L
Linus Torvalds 已提交
2585 2586
		err = 0;
	}
2587 2588
	spin_unlock(&net_family_lock);

2589
	printk(KERN_INFO "NET: Registered protocol family %d\n", ops->family);
L
Linus Torvalds 已提交
2590 2591
	return err;
}
2592
EXPORT_SYMBOL(sock_register);
L
Linus Torvalds 已提交
2593

2594 2595 2596 2597
/**
 *	sock_unregister - remove a protocol handler
 *	@family: protocol family to remove
 *
L
Linus Torvalds 已提交
2598 2599
 *	This function is called by a protocol handler that wants to
 *	remove its address family, and have it unlinked from the
2600 2601 2602 2603 2604 2605
 *	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
2606
 */
2607
void sock_unregister(int family)
L
Linus Torvalds 已提交
2608
{
2609
	BUG_ON(family < 0 || family >= NPROTO);
L
Linus Torvalds 已提交
2610

2611
	spin_lock(&net_family_lock);
2612
	RCU_INIT_POINTER(net_families[family], NULL);
2613 2614 2615 2616
	spin_unlock(&net_family_lock);

	synchronize_rcu();

2617
	printk(KERN_INFO "NET: Unregistered protocol family %d\n", family);
L
Linus Torvalds 已提交
2618
}
2619
EXPORT_SYMBOL(sock_unregister);
L
Linus Torvalds 已提交
2620

2621
static int __init sock_init(void)
L
Linus Torvalds 已提交
2622
{
N
Nick Piggin 已提交
2623
	int err;
2624 2625 2626 2627 2628 2629
	/*
	 *      Initialize the network sysctl infrastructure.
	 */
	err = net_sysctl_init();
	if (err)
		goto out;
N
Nick Piggin 已提交
2630

L
Linus Torvalds 已提交
2631
	/*
2632
	 *      Initialize skbuff SLAB cache
L
Linus Torvalds 已提交
2633 2634 2635 2636
	 */
	skb_init();

	/*
2637
	 *      Initialize the protocols module.
L
Linus Torvalds 已提交
2638 2639 2640
	 */

	init_inodecache();
N
Nick Piggin 已提交
2641 2642 2643 2644

	err = register_filesystem(&sock_fs_type);
	if (err)
		goto out_fs;
L
Linus Torvalds 已提交
2645
	sock_mnt = kern_mount(&sock_fs_type);
N
Nick Piggin 已提交
2646 2647 2648 2649
	if (IS_ERR(sock_mnt)) {
		err = PTR_ERR(sock_mnt);
		goto out_mount;
	}
2650 2651

	/* The real protocol initialization is performed in later initcalls.
L
Linus Torvalds 已提交
2652 2653 2654
	 */

#ifdef CONFIG_NETFILTER
2655 2656 2657
	err = netfilter_init();
	if (err)
		goto out;
L
Linus Torvalds 已提交
2658
#endif
2659

2660 2661 2662 2663
#ifdef CONFIG_NETWORK_PHY_TIMESTAMPING
	skb_timestamping_init();
#endif

N
Nick Piggin 已提交
2664 2665 2666 2667 2668 2669 2670
out:
	return err;

out_mount:
	unregister_filesystem(&sock_fs_type);
out_fs:
	goto out;
L
Linus Torvalds 已提交
2671 2672
}

2673 2674
core_initcall(sock_init);	/* early initcall */

L
Linus Torvalds 已提交
2675 2676 2677 2678 2679 2680
#ifdef CONFIG_PROC_FS
void socket_seq_show(struct seq_file *seq)
{
	int cpu;
	int counter = 0;

2681
	for_each_possible_cpu(cpu)
2682
	    counter += per_cpu(sockets_in_use, cpu);
L
Linus Torvalds 已提交
2683 2684 2685 2686 2687 2688 2689

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

	seq_printf(seq, "sockets: used %d\n", counter);
}
2690
#endif				/* CONFIG_PROC_FS */
L
Linus Torvalds 已提交
2691

2692
#ifdef CONFIG_COMPAT
2693
static int do_siocgstamp(struct net *net, struct socket *sock,
2694
			 unsigned int cmd, void __user *up)
2695 2696 2697 2698 2699 2700
{
	mm_segment_t old_fs = get_fs();
	struct timeval ktv;
	int err;

	set_fs(KERNEL_DS);
2701
	err = sock_do_ioctl(net, sock, cmd, (unsigned long)&ktv);
2702
	set_fs(old_fs);
2703
	if (!err)
2704
		err = compat_put_timeval(&ktv, up);
2705

2706 2707 2708
	return err;
}

2709
static int do_siocgstampns(struct net *net, struct socket *sock,
2710
			   unsigned int cmd, void __user *up)
2711 2712 2713 2714 2715 2716
{
	mm_segment_t old_fs = get_fs();
	struct timespec kts;
	int err;

	set_fs(KERNEL_DS);
2717
	err = sock_do_ioctl(net, sock, cmd, (unsigned long)&kts);
2718
	set_fs(old_fs);
2719
	if (!err)
2720
		err = compat_put_timespec(&kts, up);
2721

2722 2723 2724
	return err;
}

2725
static int dev_ifname32(struct net *net, struct compat_ifreq __user *uifr32)
2726 2727 2728 2729 2730
{
	struct ifreq __user *uifr;
	int err;

	uifr = compat_alloc_user_space(sizeof(struct ifreq));
2731
	if (copy_in_user(uifr, uifr32, sizeof(struct compat_ifreq)))
2732 2733
		return -EFAULT;

2734
	err = dev_ioctl(net, SIOCGIFNAME, uifr);
2735 2736 2737
	if (err)
		return err;

2738
	if (copy_in_user(uifr32, uifr, sizeof(struct compat_ifreq)))
2739 2740 2741 2742 2743
		return -EFAULT;

	return 0;
}

2744
static int dev_ifconf(struct net *net, struct compat_ifconf __user *uifc32)
2745
{
2746
	struct compat_ifconf ifc32;
2747 2748
	struct ifconf ifc;
	struct ifconf __user *uifc;
2749
	struct compat_ifreq __user *ifr32;
2750 2751 2752 2753
	struct ifreq __user *ifr;
	unsigned int i, j;
	int err;

2754
	if (copy_from_user(&ifc32, uifc32, sizeof(struct compat_ifconf)))
2755 2756
		return -EFAULT;

2757
	memset(&ifc, 0, sizeof(ifc));
2758 2759 2760 2761 2762 2763
	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 {
2764 2765
		size_t len = ((ifc32.ifc_len / sizeof(struct compat_ifreq)) + 1) *
			sizeof(struct ifreq);
2766 2767 2768 2769
		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);
2770
		for (i = 0; i < ifc32.ifc_len; i += sizeof(struct compat_ifreq)) {
2771
			if (copy_in_user(ifr, ifr32, sizeof(struct compat_ifreq)))
2772 2773 2774 2775 2776 2777 2778 2779
				return -EFAULT;
			ifr++;
			ifr32++;
		}
	}
	if (copy_to_user(uifc, &ifc, sizeof(struct ifconf)))
		return -EFAULT;

2780
	err = dev_ioctl(net, SIOCGIFCONF, uifc);
2781 2782 2783 2784 2785 2786 2787 2788 2789
	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;
2790 2791 2792
	     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)))
2793 2794 2795 2796 2797 2798 2799 2800 2801 2802
			return -EFAULT;
		ifr32++;
		ifr++;
	}

	if (ifc32.ifcbuf == 0) {
		/* Translate from 64-bit structure multiple to
		 * a 32-bit one.
		 */
		i = ifc.ifc_len;
2803
		i = ((i / sizeof(struct ifreq)) * sizeof(struct compat_ifreq));
2804 2805 2806 2807
		ifc32.ifc_len = i;
	} else {
		ifc32.ifc_len = i;
	}
2808
	if (copy_to_user(uifc32, &ifc32, sizeof(struct compat_ifconf)))
2809 2810 2811 2812 2813
		return -EFAULT;

	return 0;
}

2814
static int ethtool_ioctl(struct net *net, struct compat_ifreq __user *ifr32)
2815
{
2816 2817 2818 2819
	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;
2820
	struct ifreq __user *ifr;
2821 2822
	u32 rule_cnt = 0, actual_rule_cnt;
	u32 ethcmd;
2823
	u32 data;
2824
	int ret;
2825

2826 2827
	if (get_user(data, &ifr32->ifr_ifru.ifru_data))
		return -EFAULT;
2828

2829 2830 2831
	compat_rxnfc = compat_ptr(data);

	if (get_user(ethcmd, &compat_rxnfc->cmd))
2832 2833
		return -EFAULT;

2834 2835 2836 2837 2838 2839 2840 2841 2842 2843 2844 2845 2846 2847 2848 2849 2850
	/* 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:
2851
	case ETHTOOL_SRXCLSRLINS:
2852 2853 2854 2855 2856 2857 2858 2859 2860
		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);
S
Stephen Hemminger 已提交
2861
	rxnfc = (void __user *)ifr + ALIGN(sizeof(struct ifreq), 8);
2862 2863

	if (copy_in_user(&ifr->ifr_name, &ifr32->ifr_name, IFNAMSIZ))
2864 2865
		return -EFAULT;

2866 2867
	if (put_user(convert_in ? rxnfc : compat_ptr(data),
		     &ifr->ifr_ifru.ifru_data))
2868 2869
		return -EFAULT;

2870
	if (convert_in) {
2871
		/* We expect there to be holes between fs.m_ext and
2872 2873
		 * fs.ring_cookie and at the end of fs, but nowhere else.
		 */
2874 2875 2876 2877
		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));
2878 2879 2880 2881 2882 2883 2884
		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,
S
Stephen Hemminger 已提交
2885 2886
				 (void __user *)(&rxnfc->fs.m_ext + 1) -
				 (void __user *)rxnfc) ||
2887 2888
		    copy_in_user(&rxnfc->fs.ring_cookie,
				 &compat_rxnfc->fs.ring_cookie,
S
Stephen Hemminger 已提交
2889 2890
				 (void __user *)(&rxnfc->fs.location + 1) -
				 (void __user *)&rxnfc->fs.ring_cookie) ||
2891 2892 2893 2894 2895 2896 2897 2898 2899 2900 2901
		    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,
S
Stephen Hemminger 已提交
2902 2903
				 (const void __user *)(&rxnfc->fs.m_ext + 1) -
				 (const void __user *)rxnfc) ||
2904 2905
		    copy_in_user(&compat_rxnfc->fs.ring_cookie,
				 &rxnfc->fs.ring_cookie,
S
Stephen Hemminger 已提交
2906 2907
				 (const void __user *)(&rxnfc->fs.location + 1) -
				 (const void __user *)&rxnfc->fs.ring_cookie) ||
2908 2909 2910 2911 2912 2913 2914 2915 2916 2917 2918 2919 2920 2921 2922 2923 2924 2925 2926 2927 2928 2929 2930 2931 2932
		    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;
2933 2934
}

2935 2936 2937 2938 2939 2940
static int compat_siocwandev(struct net *net, struct compat_ifreq __user *uifr32)
{
	void __user *uptr;
	compat_uptr_t uptr32;
	struct ifreq __user *uifr;

2941
	uifr = compat_alloc_user_space(sizeof(*uifr));
2942 2943 2944 2945 2946 2947 2948 2949 2950 2951 2952 2953 2954 2955
	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);
}

2956 2957
static int bond_ioctl(struct net *net, unsigned int cmd,
			 struct compat_ifreq __user *ifr32)
2958 2959 2960 2961 2962 2963 2964 2965 2966 2967 2968 2969 2970
{
	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:
2971
		if (copy_from_user(&kifr, ifr32, sizeof(struct compat_ifreq)))
2972 2973 2974
			return -EFAULT;

		old_fs = get_fs();
2975
		set_fs(KERNEL_DS);
2976 2977
		err = dev_ioctl(net, cmd,
				(struct ifreq __user __force *) &kifr);
2978
		set_fs(old_fs);
2979 2980 2981 2982 2983 2984 2985 2986 2987 2988 2989 2990 2991 2992 2993

		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;

2994
		return dev_ioctl(net, cmd, uifr);
2995
	default:
2996
		return -ENOIOCTLCMD;
2997
	}
2998 2999
}

3000 3001
static int siocdevprivate_ioctl(struct net *net, unsigned int cmd,
				 struct compat_ifreq __user *u_ifreq32)
3002 3003 3004 3005 3006 3007 3008 3009 3010 3011 3012 3013 3014 3015 3016 3017 3018 3019 3020 3021 3022 3023 3024 3025
{
	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;

3026
	return dev_ioctl(net, cmd, u_ifreq64);
3027 3028
}

3029 3030
static int dev_ifsioc(struct net *net, struct socket *sock,
			 unsigned int cmd, struct compat_ifreq __user *uifr32)
3031
{
3032
	struct ifreq __user *uifr;
3033 3034
	int err;

3035 3036 3037 3038 3039 3040
	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);

3041 3042 3043 3044 3045 3046 3047 3048 3049 3050 3051 3052
	if (!err) {
		switch (cmd) {
		case SIOCGIFFLAGS:
		case SIOCGIFMETRIC:
		case SIOCGIFMTU:
		case SIOCGIFMEM:
		case SIOCGIFHWADDR:
		case SIOCGIFINDEX:
		case SIOCGIFADDR:
		case SIOCGIFBRDADDR:
		case SIOCGIFDSTADDR:
		case SIOCGIFNETMASK:
3053
		case SIOCGIFPFLAGS:
3054
		case SIOCGIFTXQLEN:
3055 3056
		case SIOCGMIIPHY:
		case SIOCGMIIREG:
3057
			if (copy_in_user(uifr32, uifr, sizeof(*uifr32)))
3058 3059 3060 3061 3062 3063 3064
				err = -EFAULT;
			break;
		}
	}
	return err;
}

3065 3066 3067 3068 3069 3070 3071 3072 3073 3074 3075 3076 3077 3078 3079 3080 3081 3082 3083 3084
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();
3085
	set_fs(KERNEL_DS);
3086
	err = dev_ioctl(net, cmd, (void  __user __force *)&ifr);
3087
	set_fs(old_fs);
3088 3089 3090 3091 3092 3093 3094 3095 3096 3097 3098 3099 3100 3101 3102 3103 3104 3105 3106 3107 3108

	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;

3109
	uifr = compat_alloc_user_space(sizeof(*uifr));
3110 3111 3112 3113 3114 3115 3116 3117 3118 3119 3120 3121 3122 3123
	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);
}

3124
struct rtentry32 {
3125
	u32		rt_pad1;
3126 3127 3128
	struct sockaddr rt_dst;         /* target address               */
	struct sockaddr rt_gateway;     /* gateway addr (RTF_GATEWAY)   */
	struct sockaddr rt_genmask;     /* target network mask (IP)     */
3129 3130 3131 3132 3133 3134 3135
	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! */
3136
	/* char * */ u32 rt_dev;        /* forcing the device at add    */
3137 3138
	u32		rt_mtu;         /* per route MTU/Window         */
	u32		rt_window;      /* Window clamping              */
3139 3140 3141 3142 3143 3144 3145 3146 3147 3148 3149 3150 3151 3152 3153 3154
	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;
};

3155 3156
static int routing_ioctl(struct net *net, struct socket *sock,
			 unsigned int cmd, void __user *argp)
3157 3158 3159 3160 3161 3162 3163 3164 3165
{
	int ret;
	void *r = NULL;
	struct in6_rtmsg r6;
	struct rtentry r4;
	char devname[16];
	u32 rtdev;
	mm_segment_t old_fs = get_fs();

3166 3167
	if (sock && sock->sk && sock->sk->sk_family == AF_INET6) { /* ipv6 */
		struct in6_rtmsg32 __user *ur6 = argp;
3168
		ret = copy_from_user(&r6.rtmsg_dst, &(ur6->rtmsg_dst),
3169
			3 * sizeof(struct in6_addr));
3170 3171 3172 3173 3174 3175 3176
		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));
3177 3178 3179

		r = (void *) &r6;
	} else { /* ipv4 */
3180
		struct rtentry32 __user *ur4 = argp;
3181
		ret = copy_from_user(&r4.rt_dst, &(ur4->rt_dst),
3182
					3 * sizeof(struct sockaddr));
3183 3184 3185 3186 3187 3188
		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));
3189
		if (rtdev) {
3190
			ret |= copy_from_user(devname, compat_ptr(rtdev), 15);
3191 3192
			r4.rt_dev = (char __user __force *)devname;
			devname[15] = 0;
3193 3194 3195 3196 3197 3198 3199 3200 3201 3202 3203
		} else
			r4.rt_dev = NULL;

		r = (void *) &r4;
	}

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

3204
	set_fs(KERNEL_DS);
3205
	ret = sock_do_ioctl(net, sock, cmd, (unsigned long) r);
3206
	set_fs(old_fs);
3207 3208 3209 3210 3211 3212 3213

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
3214
 * use compatible ioctls
3215
 */
3216
static int old_bridge_ioctl(compat_ulong_t __user *argp)
3217
{
3218
	compat_ulong_t tmp;
3219

3220
	if (get_user(tmp, argp))
3221 3222 3223 3224 3225 3226
		return -EFAULT;
	if (tmp == BRCTL_GET_VERSION)
		return BRCTL_VERSION + 1;
	return -EINVAL;
}

3227 3228 3229 3230 3231 3232
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);
3233

3234 3235 3236 3237 3238 3239 3240 3241 3242 3243 3244 3245 3246
	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);
3247 3248
	case SIOCWANDEV:
		return compat_siocwandev(net, argp);
3249 3250 3251
	case SIOCGIFMAP:
	case SIOCSIFMAP:
		return compat_sioc_ifmap(net, cmd, argp);
3252 3253 3254 3255 3256 3257 3258 3259 3260 3261 3262 3263 3264 3265
	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);
3266 3267
	case SIOCSHWTSTAMP:
		return compat_siocshwtstamp(net, argp);
3268 3269 3270 3271 3272 3273 3274 3275 3276 3277 3278 3279 3280 3281 3282 3283 3284 3285 3286 3287 3288 3289 3290 3291 3292 3293 3294 3295 3296 3297 3298 3299 3300 3301 3302 3303 3304 3305 3306 3307 3308 3309

	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:
3310 3311 3312 3313
	case SIOCSIFNAME:
	case SIOCGMIIPHY:
	case SIOCGMIIREG:
	case SIOCSMIIREG:
3314
		return dev_ifsioc(net, sock, cmd, argp);
3315

3316 3317 3318 3319
	case SIOCSARP:
	case SIOCGARP:
	case SIOCDARP:
	case SIOCATMARK:
3320 3321 3322
		return sock_do_ioctl(net, sock, cmd, arg);
	}

3323 3324
	return -ENOIOCTLCMD;
}
3325

3326
static long compat_sock_ioctl(struct file *file, unsigned int cmd,
3327
			      unsigned long arg)
3328 3329 3330
{
	struct socket *sock = file->private_data;
	int ret = -ENOIOCTLCMD;
3331 3332 3333 3334 3335
	struct sock *sk;
	struct net *net;

	sk = sock->sk;
	net = sock_net(sk);
3336 3337 3338 3339

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

3340 3341 3342 3343
	if (ret == -ENOIOCTLCMD &&
	    (cmd >= SIOCIWFIRST && cmd <= SIOCIWLAST))
		ret = compat_wext_handle_ioctl(net, cmd, arg);

3344 3345 3346
	if (ret == -ENOIOCTLCMD)
		ret = compat_sock_ioctl_trans(file, sock, cmd, arg);

3347 3348 3349 3350
	return ret;
}
#endif

3351 3352 3353 3354
int kernel_bind(struct socket *sock, struct sockaddr *addr, int addrlen)
{
	return sock->ops->bind(sock, addr, addrlen);
}
3355
EXPORT_SYMBOL(kernel_bind);
3356 3357 3358 3359 3360

int kernel_listen(struct socket *sock, int backlog)
{
	return sock->ops->listen(sock, backlog);
}
3361
EXPORT_SYMBOL(kernel_listen);
3362 3363 3364 3365 3366 3367 3368 3369 3370 3371 3372 3373 3374 3375

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);
3376
		*newsock = NULL;
3377 3378 3379 3380
		goto done;
	}

	(*newsock)->ops = sock->ops;
3381
	__module_get((*newsock)->ops->owner);
3382 3383 3384 3385

done:
	return err;
}
3386
EXPORT_SYMBOL(kernel_accept);
3387 3388

int kernel_connect(struct socket *sock, struct sockaddr *addr, int addrlen,
3389
		   int flags)
3390 3391 3392
{
	return sock->ops->connect(sock, addr, addrlen, flags);
}
3393
EXPORT_SYMBOL(kernel_connect);
3394 3395 3396 3397 3398 3399

int kernel_getsockname(struct socket *sock, struct sockaddr *addr,
			 int *addrlen)
{
	return sock->ops->getname(sock, addr, addrlen, 0);
}
3400
EXPORT_SYMBOL(kernel_getsockname);
3401 3402 3403 3404 3405 3406

int kernel_getpeername(struct socket *sock, struct sockaddr *addr,
			 int *addrlen)
{
	return sock->ops->getname(sock, addr, addrlen, 1);
}
3407
EXPORT_SYMBOL(kernel_getpeername);
3408 3409 3410 3411 3412

int kernel_getsockopt(struct socket *sock, int level, int optname,
			char *optval, int *optlen)
{
	mm_segment_t oldfs = get_fs();
3413 3414
	char __user *uoptval;
	int __user *uoptlen;
3415 3416
	int err;

3417 3418 3419
	uoptval = (char __user __force *) optval;
	uoptlen = (int __user __force *) optlen;

3420 3421
	set_fs(KERNEL_DS);
	if (level == SOL_SOCKET)
3422
		err = sock_getsockopt(sock, level, optname, uoptval, uoptlen);
3423
	else
3424 3425
		err = sock->ops->getsockopt(sock, level, optname, uoptval,
					    uoptlen);
3426 3427 3428
	set_fs(oldfs);
	return err;
}
3429
EXPORT_SYMBOL(kernel_getsockopt);
3430 3431

int kernel_setsockopt(struct socket *sock, int level, int optname,
3432
			char *optval, unsigned int optlen)
3433 3434
{
	mm_segment_t oldfs = get_fs();
3435
	char __user *uoptval;
3436 3437
	int err;

3438 3439
	uoptval = (char __user __force *) optval;

3440 3441
	set_fs(KERNEL_DS);
	if (level == SOL_SOCKET)
3442
		err = sock_setsockopt(sock, level, optname, uoptval, optlen);
3443
	else
3444
		err = sock->ops->setsockopt(sock, level, optname, uoptval,
3445 3446 3447 3448
					    optlen);
	set_fs(oldfs);
	return err;
}
3449
EXPORT_SYMBOL(kernel_setsockopt);
3450 3451 3452 3453 3454 3455 3456 3457 3458

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

	return sock_no_sendpage(sock, page, offset, size, flags);
}
3459
EXPORT_SYMBOL(kernel_sendpage);
3460 3461 3462 3463 3464 3465 3466 3467 3468 3469 3470 3471

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;
}
3472
EXPORT_SYMBOL(kernel_sock_ioctl);
3473

3474 3475 3476 3477 3478
int kernel_sock_shutdown(struct socket *sock, enum sock_shutdown_cmd how)
{
	return sock->ops->shutdown(sock, how);
}
EXPORT_SYMBOL(kernel_sock_shutdown);