socket.c 80.8 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>
75
#include <linux/ptp_classify.h>
L
Linus Torvalds 已提交
76 77 78 79 80 81 82 83 84 85
#include <linux/init.h>
#include <linux/poll.h>
#include <linux/cache.h>
#include <linux/module.h>
#include <linux/highmem.h>
#include <linux/mount.h>
#include <linux/security.h>
#include <linux/syscalls.h>
#include <linux/compat.h>
#include <linux/kmod.h>
86
#include <linux/audit.h>
87
#include <linux/wireless.h>
88
#include <linux/nsproxy.h>
N
Nick Black 已提交
89
#include <linux/magic.h>
90
#include <linux/slab.h>
91
#include <linux/xattr.h>
L
Linus Torvalds 已提交
92 93 94 95 96

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

#include <net/compat.h>
97
#include <net/wext.h>
98
#include <net/cls_cgroup.h>
L
Linus Torvalds 已提交
99 100 101 102

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

103 104 105 106 107
#include <linux/if_tun.h>
#include <linux/ipv6_route.h>
#include <linux/route.h>
#include <linux/sockios.h>
#include <linux/atalk.h>
108
#include <net/busy_poll.h>
109
#include <linux/errqueue.h>
E
Eliezer Tamir 已提交
110

111
#ifdef CONFIG_NET_RX_BUSY_POLL
112 113
unsigned int sysctl_net_busy_read __read_mostly;
unsigned int sysctl_net_busy_poll __read_mostly;
E
Eliezer Tamir 已提交
114
#endif
115

116 117
static ssize_t sock_read_iter(struct kiocb *iocb, struct iov_iter *to);
static ssize_t sock_write_iter(struct kiocb *iocb, struct iov_iter *from);
118
static int sock_mmap(struct file *file, struct vm_area_struct *vma);
L
Linus Torvalds 已提交
119 120 121 122

static int sock_close(struct inode *inode, struct file *file);
static unsigned int sock_poll(struct file *file,
			      struct poll_table_struct *wait);
123
static long sock_ioctl(struct file *file, unsigned int cmd, unsigned long arg);
124 125
#ifdef CONFIG_COMPAT
static long compat_sock_ioctl(struct file *file,
126
			      unsigned int cmd, unsigned long arg);
127
#endif
L
Linus Torvalds 已提交
128 129 130
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
131
static ssize_t sock_splice_read(struct file *file, loff_t *ppos,
132
				struct pipe_inode_info *pipe, size_t len,
J
Jens Axboe 已提交
133
				unsigned int flags);
L
Linus Torvalds 已提交
134 135 136 137 138 139

/*
 *	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.
 */

140
static const struct file_operations socket_file_ops = {
L
Linus Torvalds 已提交
141 142
	.owner =	THIS_MODULE,
	.llseek =	no_llseek,
143 144
	.read_iter =	sock_read_iter,
	.write_iter =	sock_write_iter,
L
Linus Torvalds 已提交
145 146
	.poll =		sock_poll,
	.unlocked_ioctl = sock_ioctl,
147 148 149
#ifdef CONFIG_COMPAT
	.compat_ioctl = compat_sock_ioctl,
#endif
L
Linus Torvalds 已提交
150 151 152
	.mmap =		sock_mmap,
	.release =	sock_close,
	.fasync =	sock_fasync,
153 154
	.sendpage =	sock_sendpage,
	.splice_write = generic_splice_sendpage,
J
Jens Axboe 已提交
155
	.splice_read =	sock_splice_read,
L
Linus Torvalds 已提交
156 157 158 159 160 161 162
};

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

static DEFINE_SPINLOCK(net_family_lock);
E
Eric Dumazet 已提交
163
static const struct net_proto_family __rcu *net_families[NPROTO] __read_mostly;
L
Linus Torvalds 已提交
164 165 166 167 168

/*
 *	Statistics counters of the socket lists
 */

169
static DEFINE_PER_CPU(int, sockets_in_use);
L
Linus Torvalds 已提交
170 171

/*
172 173 174
 * 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
175 176 177 178 179 180 181 182 183 184 185 186 187
 */

/**
 *	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.
 */

188
int move_addr_to_kernel(void __user *uaddr, int ulen, struct sockaddr_storage *kaddr)
L
Linus Torvalds 已提交
189
{
190
	if (ulen < 0 || ulen > sizeof(struct sockaddr_storage))
L
Linus Torvalds 已提交
191
		return -EINVAL;
192
	if (ulen == 0)
L
Linus Torvalds 已提交
193
		return 0;
194
	if (copy_from_user(kaddr, uaddr, ulen))
L
Linus Torvalds 已提交
195
		return -EFAULT;
196
	return audit_sockaddr(ulen, kaddr);
L
Linus Torvalds 已提交
197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214
}

/**
 *	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.
 */
215

216
static int move_addr_to_user(struct sockaddr_storage *kaddr, int klen,
S
stephen hemminger 已提交
217
			     void __user *uaddr, int __user *ulen)
L
Linus Torvalds 已提交
218 219 220 221
{
	int err;
	int len;

222
	BUG_ON(klen > sizeof(struct sockaddr_storage));
223 224
	err = get_user(len, ulen);
	if (err)
L
Linus Torvalds 已提交
225
		return err;
226 227
	if (len > klen)
		len = klen;
228
	if (len < 0)
L
Linus Torvalds 已提交
229
		return -EINVAL;
230
	if (len) {
S
Steve Grubb 已提交
231 232
		if (audit_sockaddr(klen, kaddr))
			return -ENOMEM;
233
		if (copy_to_user(uaddr, kaddr, len))
L
Linus Torvalds 已提交
234 235 236
			return -EFAULT;
	}
	/*
237 238
	 *      "fromlen shall refer to the value before truncation.."
	 *                      1003.1g
L
Linus Torvalds 已提交
239 240 241 242
	 */
	return __put_user(klen, ulen);
}

243
static struct kmem_cache *sock_inode_cachep __read_mostly;
L
Linus Torvalds 已提交
244 245 246 247

static struct inode *sock_alloc_inode(struct super_block *sb)
{
	struct socket_alloc *ei;
248
	struct socket_wq *wq;
249

250
	ei = kmem_cache_alloc(sock_inode_cachep, GFP_KERNEL);
L
Linus Torvalds 已提交
251 252
	if (!ei)
		return NULL;
253 254
	wq = kmalloc(sizeof(*wq), GFP_KERNEL);
	if (!wq) {
255 256 257
		kmem_cache_free(sock_inode_cachep, ei);
		return NULL;
	}
258 259
	init_waitqueue_head(&wq->wait);
	wq->fasync_list = NULL;
260
	wq->flags = 0;
261
	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
					      sizeof(struct socket_alloc),
					      0,
					      (SLAB_HWCACHE_ALIGN |
					       SLAB_RECLAIM_ACCOUNT |
297
					       SLAB_MEM_SPREAD | SLAB_ACCOUNT),
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
/*
 * 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]",
D
David Howells 已提交
316
				d_inode(dentry)->i_ino);
317 318
}

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
};

323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343
static int sockfs_xattr_get(const struct xattr_handler *handler,
			    struct dentry *dentry, struct inode *inode,
			    const char *suffix, void *value, size_t size)
{
	if (value) {
		if (dentry->d_name.len + 1 > size)
			return -ERANGE;
		memcpy(value, dentry->d_name.name, dentry->d_name.len + 1);
	}
	return dentry->d_name.len + 1;
}

#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 const struct xattr_handler sockfs_xattr_handler = {
	.name = XATTR_NAME_SOCKPROTONAME,
	.get = sockfs_xattr_get,
};

344 345 346 347 348 349 350 351 352 353 354 355 356 357
static int sockfs_security_xattr_set(const struct xattr_handler *handler,
				     struct dentry *dentry, struct inode *inode,
				     const char *suffix, const void *value,
				     size_t size, int flags)
{
	/* Handled by LSM. */
	return -EAGAIN;
}

static const struct xattr_handler sockfs_security_xattr_handler = {
	.prefix = XATTR_SECURITY_PREFIX,
	.set = sockfs_security_xattr_set,
};

358 359
static const struct xattr_handler *sockfs_xattr_handlers[] = {
	&sockfs_xattr_handler,
360
	&sockfs_security_xattr_handler,
361 362 363
	NULL
};

364 365 366
static struct dentry *sockfs_mount(struct file_system_type *fs_type,
			 int flags, const char *dev_name, void *data)
{
367 368 369
	return mount_pseudo_xattr(fs_type, "socket:", &sockfs_ops,
				  sockfs_xattr_handlers,
				  &sockfs_dentry_operations, SOCKFS_MAGIC);
370 371 372 373 374 375 376 377 378 379
}

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
380 381 382
/*
 *	Obtains the first available file descriptor and sets it up for use.
 *
383 384
 *	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
385 386 387 388 389 390 391 392 393 394 395 396
 *	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.
 */

397
struct file *sock_alloc_file(struct socket *sock, int flags, const char *dname)
L
Linus Torvalds 已提交
398
{
399
	struct qstr name = { .name = "" };
400
	struct path path;
401
	struct file *file;
L
Linus Torvalds 已提交
402

403 404 405 406 407 408 409
	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);
	}
410
	path.dentry = d_alloc_pseudo(sock_mnt->mnt_sb, &name);
411 412
	if (unlikely(!path.dentry))
		return ERR_PTR(-ENOMEM);
413
	path.mnt = mntget(sock_mnt);
414

415
	d_instantiate(path.dentry, SOCK_INODE(sock));
416

417
	file = alloc_file(&path, FMODE_READ | FMODE_WRITE,
418
		  &socket_file_ops);
419
	if (IS_ERR(file)) {
A
Al Viro 已提交
420
		/* drop dentry, keep inode */
D
David Howells 已提交
421
		ihold(d_inode(path.dentry));
422
		path_put(&path);
423
		return file;
A
Al Viro 已提交
424 425 426
	}

	sock->file = file;
427
	file->f_flags = O_RDWR | (flags & O_NONBLOCK);
428
	file->private_data = sock;
429
	return file;
430
}
431
EXPORT_SYMBOL(sock_alloc_file);
432

433
static int sock_map_fd(struct socket *sock, int flags)
434 435
{
	struct file *newfile;
436 437 438
	int fd = get_unused_fd_flags(flags);
	if (unlikely(fd < 0))
		return fd;
439

440
	newfile = sock_alloc_file(sock, flags, NULL);
441
	if (likely(!IS_ERR(newfile))) {
442
		fd_install(fd, newfile);
443 444
		return fd;
	}
445

446 447
	put_unused_fd(fd);
	return PTR_ERR(newfile);
L
Linus Torvalds 已提交
448 449
}

450
struct socket *sock_from_file(struct file *file, int *err)
451 452 453 454
{
	if (file->f_op == &socket_file_ops)
		return file->private_data;	/* set in sock_map_fd */

E
Eric Dumazet 已提交
455 456
	*err = -ENOTSOCK;
	return NULL;
457
}
458
EXPORT_SYMBOL(sock_from_file);
459

L
Linus Torvalds 已提交
460
/**
461
 *	sockfd_lookup - Go from a file number to its socket slot
L
Linus Torvalds 已提交
462 463 464 465 466 467 468 469 470 471 472 473 474 475 476 477
 *	@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;

478 479
	file = fget(fd);
	if (!file) {
L
Linus Torvalds 已提交
480 481 482
		*err = -EBADF;
		return NULL;
	}
483

484 485
	sock = sock_from_file(file, err);
	if (!sock)
L
Linus Torvalds 已提交
486
		fput(file);
487 488
	return sock;
}
489
EXPORT_SYMBOL(sockfd_lookup);
L
Linus Torvalds 已提交
490

491 492
static struct socket *sockfd_lookup_light(int fd, int *err, int *fput_needed)
{
493
	struct fd f = fdget(fd);
494 495
	struct socket *sock;

496
	*err = -EBADF;
497 498 499 500
	if (f.file) {
		sock = sock_from_file(f.file, err);
		if (likely(sock)) {
			*fput_needed = f.flags;
501
			return sock;
502 503
		}
		fdput(f);
L
Linus Torvalds 已提交
504
	}
505
	return NULL;
L
Linus Torvalds 已提交
506 507
}

508 509 510 511 512 513
static ssize_t sockfs_listxattr(struct dentry *dentry, char *buffer,
				size_t size)
{
	ssize_t len;
	ssize_t used = 0;

D
David Howells 已提交
514
	len = security_inode_listsecurity(d_inode(dentry), buffer, size);
515 516 517 518 519 520 521 522 523 524 525 526 527 528 529 530 531 532 533 534 535
	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;
}

536 537 538 539 540 541 542 543 544 545 546 547 548
int sockfs_setattr(struct dentry *dentry, struct iattr *iattr)
{
	int err = simple_setattr(dentry, iattr);

	if (!err) {
		struct socket *sock = SOCKET_I(d_inode(dentry));

		sock->sk->sk_uid = iattr->ia_uid;
	}

	return err;
}

549 550
static const struct inode_operations sockfs_inode_ops = {
	.listxattr = sockfs_listxattr,
551
	.setattr = sockfs_setattr,
552 553
};

L
Linus Torvalds 已提交
554 555
/**
 *	sock_alloc	-	allocate a socket
556
 *
L
Linus Torvalds 已提交
557 558 559 560 561
 *	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.
 */

T
Tom Herbert 已提交
562
struct socket *sock_alloc(void)
L
Linus Torvalds 已提交
563
{
564 565
	struct inode *inode;
	struct socket *sock;
L
Linus Torvalds 已提交
566

567
	inode = new_inode_pseudo(sock_mnt->mnt_sb);
L
Linus Torvalds 已提交
568 569 570 571 572
	if (!inode)
		return NULL;

	sock = SOCKET_I(inode);

573
	kmemcheck_annotate_bitfield(sock, type);
574
	inode->i_ino = get_next_ino();
575
	inode->i_mode = S_IFSOCK | S_IRWXUGO;
576 577
	inode->i_uid = current_fsuid();
	inode->i_gid = current_fsgid();
578
	inode->i_op = &sockfs_inode_ops;
L
Linus Torvalds 已提交
579

580
	this_cpu_add(sockets_in_use, 1);
L
Linus Torvalds 已提交
581 582
	return sock;
}
T
Tom Herbert 已提交
583
EXPORT_SYMBOL(sock_alloc);
L
Linus Torvalds 已提交
584 585 586 587 588 589 590

/**
 *	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
591
 *	an inode not a file.
L
Linus Torvalds 已提交
592
 */
593

L
Linus Torvalds 已提交
594 595 596 597 598 599 600 601 602 603
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);
	}

604
	if (rcu_dereference_protected(sock->wq, 1)->fasync_list)
605
		pr_err("%s: fasync list not empty!\n", __func__);
L
Linus Torvalds 已提交
606

607
	this_cpu_sub(sockets_in_use, 1);
L
Linus Torvalds 已提交
608 609 610 611
	if (!sock->file) {
		iput(SOCK_INODE(sock));
		return;
	}
612
	sock->file = NULL;
L
Linus Torvalds 已提交
613
}
614
EXPORT_SYMBOL(sock_release);
L
Linus Torvalds 已提交
615

616
void __sock_tx_timestamp(__u16 tsflags, __u8 *tx_flags)
617
{
618 619
	u8 flags = *tx_flags;

620
	if (tsflags & SOF_TIMESTAMPING_TX_HARDWARE)
621 622
		flags |= SKBTX_HW_TSTAMP;

623
	if (tsflags & SOF_TIMESTAMPING_TX_SOFTWARE)
624 625
		flags |= SKBTX_SW_TSTAMP;

626
	if (tsflags & SOF_TIMESTAMPING_TX_SCHED)
627 628 629
		flags |= SKBTX_SCHED_TSTAMP;

	*tx_flags = flags;
630
}
631
EXPORT_SYMBOL(__sock_tx_timestamp);
632

633
static inline int sock_sendmsg_nosec(struct socket *sock, struct msghdr *msg)
L
Linus Torvalds 已提交
634
{
A
Al Viro 已提交
635
	int ret = sock->ops->sendmsg(sock, msg, msg_data_left(msg));
636 637
	BUG_ON(ret == -EIOCBQUEUED);
	return ret;
L
Linus Torvalds 已提交
638 639
}

640
int sock_sendmsg(struct socket *sock, struct msghdr *msg)
641
{
642
	int err = security_socket_sendmsg(sock, msg,
A
Al Viro 已提交
643
					  msg_data_left(msg));
644

645
	return err ?: sock_sendmsg_nosec(sock, msg);
646
}
647
EXPORT_SYMBOL(sock_sendmsg);
L
Linus Torvalds 已提交
648 649 650 651

int kernel_sendmsg(struct socket *sock, struct msghdr *msg,
		   struct kvec *vec, size_t num, size_t size)
{
652
	iov_iter_kvec(&msg->msg_iter, WRITE | ITER_KVEC, vec, num, size);
653
	return sock_sendmsg(sock, msg);
L
Linus Torvalds 已提交
654
}
655
EXPORT_SYMBOL(kernel_sendmsg);
L
Linus Torvalds 已提交
656

657 658 659 660 661 662
/*
 * 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)
{
663
	int need_software_tstamp = sock_flag(sk, SOCK_RCVTSTAMP);
664
	struct scm_timestamping tss;
665 666 667 668 669 670 671 672 673 674 675 676 677 678 679 680
	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 {
681 682
			struct timespec ts;
			skb_get_timestampns(skb, &ts);
683
			put_cmsg(msg, SOL_SOCKET, SCM_TIMESTAMPNS,
684
				 sizeof(ts), &ts);
685 686 687
		}
	}

688
	memset(&tss, 0, sizeof(tss));
689
	if ((sk->sk_tsflags & SOF_TIMESTAMPING_SOFTWARE) &&
690
	    ktime_to_timespec_cond(skb->tstamp, tss.ts + 0))
691
		empty = 0;
692
	if (shhwtstamps &&
693
	    (sk->sk_tsflags & SOF_TIMESTAMPING_RAW_HARDWARE) &&
694
	    ktime_to_timespec_cond(shhwtstamps->hwtstamp, tss.ts + 2))
695
		empty = 0;
696
	if (!empty) {
697
		put_cmsg(msg, SOL_SOCKET,
698
			 SCM_TIMESTAMPING, sizeof(tss), &tss);
699 700 701 702 703

		if (skb->len && (sk->sk_tsflags & SOF_TIMESTAMPING_OPT_STATS))
			put_cmsg(msg, SOL_SOCKET, SCM_TIMESTAMPING_OPT_STATS,
				 skb->len, skb->data);
	}
704
}
705 706
EXPORT_SYMBOL_GPL(__sock_recv_timestamp);

707 708 709 710 711 712 713 714 715 716 717 718 719 720 721 722
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
723 724
static inline void sock_recv_drops(struct msghdr *msg, struct sock *sk,
				   struct sk_buff *skb)
725
{
726
	if (sock_flag(sk, SOCK_RXQ_OVFL) && skb && SOCK_SKB_CB(skb)->dropcount)
727
		put_cmsg(msg, SOL_SOCKET, SO_RXQ_OVFL,
728
			sizeof(__u32), &SOCK_SKB_CB(skb)->dropcount);
729 730
}

731
void __sock_recv_ts_and_drops(struct msghdr *msg, struct sock *sk,
732 733 734 735 736
	struct sk_buff *skb)
{
	sock_recv_timestamp(msg, sk, skb);
	sock_recv_drops(msg, sk, skb);
}
737
EXPORT_SYMBOL_GPL(__sock_recv_ts_and_drops);
738

739
static inline int sock_recvmsg_nosec(struct socket *sock, struct msghdr *msg,
740
				     int flags)
L
Linus Torvalds 已提交
741
{
742
	return sock->ops->recvmsg(sock, msg, msg_data_left(msg), flags);
L
Linus Torvalds 已提交
743 744
}

745
int sock_recvmsg(struct socket *sock, struct msghdr *msg, int flags)
746
{
747
	int err = security_socket_recvmsg(sock, msg, msg_data_left(msg), flags);
748

749
	return err ?: sock_recvmsg_nosec(sock, msg, flags);
L
Linus Torvalds 已提交
750
}
751
EXPORT_SYMBOL(sock_recvmsg);
L
Linus Torvalds 已提交
752

753 754 755 756 757 758 759 760 761 762 763 764 765 766 767
/**
 * 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.
 */
768 769
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
770 771 772 773
{
	mm_segment_t oldfs = get_fs();
	int result;

774
	iov_iter_kvec(&msg->msg_iter, READ | ITER_KVEC, vec, num, size);
L
Linus Torvalds 已提交
775
	set_fs(KERNEL_DS);
776
	result = sock_recvmsg(sock, msg, flags);
L
Linus Torvalds 已提交
777 778 779
	set_fs(oldfs);
	return result;
}
780
EXPORT_SYMBOL(kernel_recvmsg);
L
Linus Torvalds 已提交
781

782 783
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
784 785 786 787
{
	struct socket *sock;
	int flags;

788 789
	sock = file->private_data;

790 791 792
	flags = (file->f_flags & O_NONBLOCK) ? MSG_DONTWAIT : 0;
	/* more is a combination of MSG_MORE and MSG_SENDPAGE_NOTLAST */
	flags |= more;
793

794
	return kernel_sendpage(sock, page, offset, size, flags);
795
}
L
Linus Torvalds 已提交
796

J
Jens Axboe 已提交
797
static ssize_t sock_splice_read(struct file *file, loff_t *ppos,
798
				struct pipe_inode_info *pipe, size_t len,
J
Jens Axboe 已提交
799 800 801 802
				unsigned int flags)
{
	struct socket *sock = file->private_data;

803 804 805
	if (unlikely(!sock->ops->splice_read))
		return -EINVAL;

J
Jens Axboe 已提交
806 807 808
	return sock->ops->splice_read(sock, ppos, pipe, len, flags);
}

809
static ssize_t sock_read_iter(struct kiocb *iocb, struct iov_iter *to)
810
{
811 812
	struct file *file = iocb->ki_filp;
	struct socket *sock = file->private_data;
813 814
	struct msghdr msg = {.msg_iter = *to,
			     .msg_iocb = iocb};
815
	ssize_t res;
816

817 818 819 820
	if (file->f_flags & O_NONBLOCK)
		msg.msg_flags = MSG_DONTWAIT;

	if (iocb->ki_pos != 0)
L
Linus Torvalds 已提交
821
		return -ESPIPE;
822

C
Christoph Hellwig 已提交
823
	if (!iov_iter_count(to))	/* Match SYS5 behaviour */
L
Linus Torvalds 已提交
824 825
		return 0;

826
	res = sock_recvmsg(sock, &msg, msg.msg_flags);
827 828
	*to = msg.msg_iter;
	return res;
L
Linus Torvalds 已提交
829 830
}

831
static ssize_t sock_write_iter(struct kiocb *iocb, struct iov_iter *from)
832
{
833 834
	struct file *file = iocb->ki_filp;
	struct socket *sock = file->private_data;
835 836
	struct msghdr msg = {.msg_iter = *from,
			     .msg_iocb = iocb};
837
	ssize_t res;
L
Linus Torvalds 已提交
838

839
	if (iocb->ki_pos != 0)
840
		return -ESPIPE;
841

842 843 844
	if (file->f_flags & O_NONBLOCK)
		msg.msg_flags = MSG_DONTWAIT;

845 846 847
	if (sock->type == SOCK_SEQPACKET)
		msg.msg_flags |= MSG_EOR;

848
	res = sock_sendmsg(sock, &msg);
849 850
	*from = msg.msg_iter;
	return res;
L
Linus Torvalds 已提交
851 852 853 854 855 856 857
}

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

A
Arjan van de Ven 已提交
858
static DEFINE_MUTEX(br_ioctl_mutex);
859
static int (*br_ioctl_hook) (struct net *, unsigned int cmd, void __user *arg);
L
Linus Torvalds 已提交
860

861
void brioctl_set(int (*hook) (struct net *, unsigned int, void __user *))
L
Linus Torvalds 已提交
862
{
A
Arjan van de Ven 已提交
863
	mutex_lock(&br_ioctl_mutex);
L
Linus Torvalds 已提交
864
	br_ioctl_hook = hook;
A
Arjan van de Ven 已提交
865
	mutex_unlock(&br_ioctl_mutex);
L
Linus Torvalds 已提交
866 867 868
}
EXPORT_SYMBOL(brioctl_set);

A
Arjan van de Ven 已提交
869
static DEFINE_MUTEX(vlan_ioctl_mutex);
870
static int (*vlan_ioctl_hook) (struct net *, void __user *arg);
L
Linus Torvalds 已提交
871

872
void vlan_ioctl_set(int (*hook) (struct net *, void __user *))
L
Linus Torvalds 已提交
873
{
A
Arjan van de Ven 已提交
874
	mutex_lock(&vlan_ioctl_mutex);
L
Linus Torvalds 已提交
875
	vlan_ioctl_hook = hook;
A
Arjan van de Ven 已提交
876
	mutex_unlock(&vlan_ioctl_mutex);
L
Linus Torvalds 已提交
877 878 879
}
EXPORT_SYMBOL(vlan_ioctl_set);

A
Arjan van de Ven 已提交
880
static DEFINE_MUTEX(dlci_ioctl_mutex);
881
static int (*dlci_ioctl_hook) (unsigned int, void __user *);
L
Linus Torvalds 已提交
882

883
void dlci_ioctl_set(int (*hook) (unsigned int, void __user *))
L
Linus Torvalds 已提交
884
{
A
Arjan van de Ven 已提交
885
	mutex_lock(&dlci_ioctl_mutex);
L
Linus Torvalds 已提交
886
	dlci_ioctl_hook = hook;
A
Arjan van de Ven 已提交
887
	mutex_unlock(&dlci_ioctl_mutex);
L
Linus Torvalds 已提交
888 889 890
}
EXPORT_SYMBOL(dlci_ioctl_set);

891 892 893 894 895 896 897 898 899 900 901 902 903 904 905 906 907 908
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
909 910 911 912 913
/*
 *	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.
 */

914 915 916 917 918
static struct ns_common *get_net_ns(struct ns_common *ns)
{
	return &get_net(container_of(ns, struct net, ns))->ns;
}

L
Linus Torvalds 已提交
919 920 921
static long sock_ioctl(struct file *file, unsigned cmd, unsigned long arg)
{
	struct socket *sock;
922
	struct sock *sk;
L
Linus Torvalds 已提交
923 924
	void __user *argp = (void __user *)arg;
	int pid, err;
925
	struct net *net;
L
Linus Torvalds 已提交
926

927
	sock = file->private_data;
928
	sk = sock->sk;
929
	net = sock_net(sk);
L
Linus Torvalds 已提交
930
	if (cmd >= SIOCDEVPRIVATE && cmd <= (SIOCDEVPRIVATE + 15)) {
931
		err = dev_ioctl(net, cmd, argp);
L
Linus Torvalds 已提交
932
	} else
J
Johannes Berg 已提交
933
#ifdef CONFIG_WEXT_CORE
L
Linus Torvalds 已提交
934
	if (cmd >= SIOCIWFIRST && cmd <= SIOCIWLAST) {
935
		err = dev_ioctl(net, cmd, argp);
L
Linus Torvalds 已提交
936
	} else
J
Johannes Berg 已提交
937
#endif
938
		switch (cmd) {
L
Linus Torvalds 已提交
939 940 941 942 943
		case FIOSETOWN:
		case SIOCSPGRP:
			err = -EFAULT;
			if (get_user(pid, (int __user *)argp))
				break;
944 945
			f_setown(sock->file, pid, 1);
			err = 0;
L
Linus Torvalds 已提交
946 947 948
			break;
		case FIOGETOWN:
		case SIOCGPGRP:
949
			err = put_user(f_getown(sock->file),
950
				       (int __user *)argp);
L
Linus Torvalds 已提交
951 952 953 954 955 956 957 958 959
			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
960
			mutex_lock(&br_ioctl_mutex);
961
			if (br_ioctl_hook)
962
				err = br_ioctl_hook(net, cmd, argp);
A
Arjan van de Ven 已提交
963
			mutex_unlock(&br_ioctl_mutex);
L
Linus Torvalds 已提交
964 965 966 967 968 969 970
			break;
		case SIOCGIFVLAN:
		case SIOCSIFVLAN:
			err = -ENOPKG;
			if (!vlan_ioctl_hook)
				request_module("8021q");

A
Arjan van de Ven 已提交
971
			mutex_lock(&vlan_ioctl_mutex);
L
Linus Torvalds 已提交
972
			if (vlan_ioctl_hook)
973
				err = vlan_ioctl_hook(net, argp);
A
Arjan van de Ven 已提交
974
			mutex_unlock(&vlan_ioctl_mutex);
L
Linus Torvalds 已提交
975 976 977 978 979 980 981
			break;
		case SIOCADDDLCI:
		case SIOCDELDLCI:
			err = -ENOPKG;
			if (!dlci_ioctl_hook)
				request_module("dlci");

982 983
			mutex_lock(&dlci_ioctl_mutex);
			if (dlci_ioctl_hook)
L
Linus Torvalds 已提交
984
				err = dlci_ioctl_hook(cmd, argp);
985
			mutex_unlock(&dlci_ioctl_mutex);
L
Linus Torvalds 已提交
986
			break;
987 988 989 990 991 992 993
		case SIOCGSKNS:
			err = -EPERM;
			if (!ns_capable(net->user_ns, CAP_NET_ADMIN))
				break;

			err = open_related_ns(&net->ns, get_net_ns);
			break;
L
Linus Torvalds 已提交
994
		default:
995
			err = sock_do_ioctl(net, sock, cmd, arg);
L
Linus Torvalds 已提交
996
			break;
997
		}
L
Linus Torvalds 已提交
998 999 1000 1001 1002 1003 1004
	return err;
}

int sock_create_lite(int family, int type, int protocol, struct socket **res)
{
	int err;
	struct socket *sock = NULL;
1005

L
Linus Torvalds 已提交
1006 1007 1008 1009 1010 1011 1012 1013 1014 1015 1016
	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
1017 1018 1019 1020
	err = security_socket_post_create(sock, family, type, protocol, 1);
	if (err)
		goto out_release;

L
Linus Torvalds 已提交
1021 1022 1023
out:
	*res = sock;
	return err;
V
Venkat Yekkirala 已提交
1024 1025 1026 1027
out_release:
	sock_release(sock);
	sock = NULL;
	goto out;
L
Linus Torvalds 已提交
1028
}
1029
EXPORT_SYMBOL(sock_create_lite);
L
Linus Torvalds 已提交
1030 1031

/* No kernel lock held - perfect */
1032
static unsigned int sock_poll(struct file *file, poll_table *wait)
L
Linus Torvalds 已提交
1033
{
1034
	unsigned int busy_flag = 0;
L
Linus Torvalds 已提交
1035 1036 1037
	struct socket *sock;

	/*
1038
	 *      We can't return errors to poll, so it's either yes or no.
L
Linus Torvalds 已提交
1039
	 */
1040
	sock = file->private_data;
1041

1042
	if (sk_can_busy_loop(sock->sk)) {
1043
		/* this socket can poll_ll so tell the system call */
1044
		busy_flag = POLL_BUSY_LOOP;
1045 1046

		/* once, only if requested by syscall */
1047 1048
		if (wait && (wait->_key & POLL_BUSY_LOOP))
			sk_busy_loop(sock->sk, 1);
1049 1050
	}

1051
	return busy_flag | sock->ops->poll(file, sock, wait);
L
Linus Torvalds 已提交
1052 1053
}

1054
static int sock_mmap(struct file *file, struct vm_area_struct *vma)
L
Linus Torvalds 已提交
1055
{
1056
	struct socket *sock = file->private_data;
L
Linus Torvalds 已提交
1057 1058 1059 1060

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

1061
static int sock_close(struct inode *inode, struct file *filp)
L
Linus Torvalds 已提交
1062 1063 1064 1065 1066 1067 1068 1069 1070 1071 1072 1073 1074
{
	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)
1075
 *	   or under socket lock
L
Linus Torvalds 已提交
1076 1077 1078 1079
 */

static int sock_fasync(int fd, struct file *filp, int on)
{
1080 1081
	struct socket *sock = filp->private_data;
	struct sock *sk = sock->sk;
1082
	struct socket_wq *wq;
L
Linus Torvalds 已提交
1083

1084
	if (sk == NULL)
L
Linus Torvalds 已提交
1085 1086 1087
		return -EINVAL;

	lock_sock(sk);
1088
	wq = rcu_dereference_protected(sock->wq, lockdep_sock_is_held(sk));
1089
	fasync_helper(fd, filp, on, &wq->fasync_list);
L
Linus Torvalds 已提交
1090

1091
	if (!wq->fasync_list)
1092 1093
		sock_reset_flag(sk, SOCK_FASYNC);
	else
E
Eric Dumazet 已提交
1094
		sock_set_flag(sk, SOCK_FASYNC);
L
Linus Torvalds 已提交
1095

1096
	release_sock(sk);
L
Linus Torvalds 已提交
1097 1098 1099
	return 0;
}

1100
/* This function may be called only under rcu_lock */
L
Linus Torvalds 已提交
1101

1102
int sock_wake_async(struct socket_wq *wq, int how, int band)
L
Linus Torvalds 已提交
1103
{
1104
	if (!wq || !wq->fasync_list)
L
Linus Torvalds 已提交
1105
		return -1;
1106

1107
	switch (how) {
1108
	case SOCK_WAKE_WAITD:
1109
		if (test_bit(SOCKWQ_ASYNC_WAITDATA, &wq->flags))
L
Linus Torvalds 已提交
1110 1111
			break;
		goto call_kill;
1112
	case SOCK_WAKE_SPACE:
1113
		if (!test_and_clear_bit(SOCKWQ_ASYNC_NOSPACE, &wq->flags))
L
Linus Torvalds 已提交
1114 1115
			break;
		/* fall through */
1116
	case SOCK_WAKE_IO:
1117
call_kill:
1118
		kill_fasync(&wq->fasync_list, SIGIO, band);
L
Linus Torvalds 已提交
1119
		break;
1120
	case SOCK_WAKE_URG:
1121
		kill_fasync(&wq->fasync_list, SIGURG, band);
L
Linus Torvalds 已提交
1122
	}
1123

L
Linus Torvalds 已提交
1124 1125
	return 0;
}
1126
EXPORT_SYMBOL(sock_wake_async);
L
Linus Torvalds 已提交
1127

P
Pavel Emelyanov 已提交
1128
int __sock_create(struct net *net, int family, int type, int protocol,
1129
			 struct socket **res, int kern)
L
Linus Torvalds 已提交
1130 1131 1132
{
	int err;
	struct socket *sock;
1133
	const struct net_proto_family *pf;
L
Linus Torvalds 已提交
1134 1135

	/*
1136
	 *      Check protocol is in range
L
Linus Torvalds 已提交
1137 1138 1139 1140 1141 1142 1143 1144 1145 1146 1147 1148
	 */
	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) {
1149 1150
		pr_info_once("%s uses obsolete (PF_INET,SOCK_PACKET)\n",
			     current->comm);
L
Linus Torvalds 已提交
1151 1152 1153 1154 1155 1156
		family = PF_PACKET;
	}

	err = security_socket_create(family, type, protocol, kern);
	if (err)
		return err;
1157

1158 1159 1160 1161 1162 1163 1164
	/*
	 *	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) {
1165
		net_warn_ratelimited("socket: no more sockets\n");
1166 1167 1168 1169 1170 1171
		return -ENFILE;	/* Not exactly a match, but its the
				   closest posix thing */
	}

	sock->type = type;

1172
#ifdef CONFIG_MODULES
1173 1174 1175
	/* 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
1176 1177 1178
	 * requested real, full-featured networking support upon configuration.
	 * Otherwise module support will break!
	 */
E
Eric Dumazet 已提交
1179
	if (rcu_access_pointer(net_families[family]) == NULL)
1180
		request_module("net-pf-%d", family);
L
Linus Torvalds 已提交
1181 1182
#endif

1183 1184 1185 1186 1187
	rcu_read_lock();
	pf = rcu_dereference(net_families[family]);
	err = -EAFNOSUPPORT;
	if (!pf)
		goto out_release;
L
Linus Torvalds 已提交
1188 1189 1190 1191 1192

	/*
	 * We will call the ->create function, that possibly is in a loadable
	 * module, so we have to bump that loadable module refcnt first.
	 */
1193
	if (!try_module_get(pf->owner))
L
Linus Torvalds 已提交
1194 1195
		goto out_release;

1196 1197 1198
	/* Now protected by module ref count */
	rcu_read_unlock();

1199
	err = pf->create(net, sock, protocol, kern);
1200
	if (err < 0)
L
Linus Torvalds 已提交
1201
		goto out_module_put;
1202

L
Linus Torvalds 已提交
1203 1204 1205 1206
	/*
	 * Now to bump the refcnt of the [loadable] module that owns this
	 * socket at sock_release time we decrement its refcnt.
	 */
1207 1208 1209
	if (!try_module_get(sock->ops->owner))
		goto out_module_busy;

L
Linus Torvalds 已提交
1210 1211 1212 1213
	/*
	 * Now that we're done with the ->create function, the [loadable]
	 * module can have its refcnt decremented
	 */
1214
	module_put(pf->owner);
V
Venkat Yekkirala 已提交
1215 1216
	err = security_socket_post_create(sock, family, type, protocol, kern);
	if (err)
1217
		goto out_sock_release;
1218
	*res = sock;
L
Linus Torvalds 已提交
1219

1220 1221 1222 1223
	return 0;

out_module_busy:
	err = -EAFNOSUPPORT;
L
Linus Torvalds 已提交
1224
out_module_put:
1225 1226 1227
	sock->ops = NULL;
	module_put(pf->owner);
out_sock_release:
L
Linus Torvalds 已提交
1228
	sock_release(sock);
1229 1230 1231 1232 1233
	return err;

out_release:
	rcu_read_unlock();
	goto out_sock_release;
L
Linus Torvalds 已提交
1234
}
P
Pavel Emelyanov 已提交
1235
EXPORT_SYMBOL(__sock_create);
L
Linus Torvalds 已提交
1236 1237 1238

int sock_create(int family, int type, int protocol, struct socket **res)
{
1239
	return __sock_create(current->nsproxy->net_ns, family, type, protocol, res, 0);
L
Linus Torvalds 已提交
1240
}
1241
EXPORT_SYMBOL(sock_create);
L
Linus Torvalds 已提交
1242

1243
int sock_create_kern(struct net *net, int family, int type, int protocol, struct socket **res)
L
Linus Torvalds 已提交
1244
{
1245
	return __sock_create(net, family, type, protocol, res, 1);
L
Linus Torvalds 已提交
1246
}
1247
EXPORT_SYMBOL(sock_create_kern);
L
Linus Torvalds 已提交
1248

1249
SYSCALL_DEFINE3(socket, int, family, int, type, int, protocol)
L
Linus Torvalds 已提交
1250 1251 1252
{
	int retval;
	struct socket *sock;
1253 1254
	int flags;

1255 1256 1257 1258 1259 1260
	/* 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);

1261
	flags = type & ~SOCK_TYPE_MASK;
1262
	if (flags & ~(SOCK_CLOEXEC | SOCK_NONBLOCK))
1263 1264
		return -EINVAL;
	type &= SOCK_TYPE_MASK;
L
Linus Torvalds 已提交
1265

U
Ulrich Drepper 已提交
1266 1267 1268
	if (SOCK_NONBLOCK != O_NONBLOCK && (flags & SOCK_NONBLOCK))
		flags = (flags & ~SOCK_NONBLOCK) | O_NONBLOCK;

L
Linus Torvalds 已提交
1269 1270 1271 1272
	retval = sock_create(family, type, protocol, &sock);
	if (retval < 0)
		goto out;

1273
	retval = sock_map_fd(sock, flags & (O_CLOEXEC | O_NONBLOCK));
L
Linus Torvalds 已提交
1274 1275 1276 1277 1278 1279 1280 1281 1282 1283 1284 1285 1286 1287 1288 1289
	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.
 */

1290 1291
SYSCALL_DEFINE4(socketpair, int, family, int, type, int, protocol,
		int __user *, usockvec)
L
Linus Torvalds 已提交
1292 1293 1294
{
	struct socket *sock1, *sock2;
	int fd1, fd2, err;
A
Al Viro 已提交
1295
	struct file *newfile1, *newfile2;
1296 1297 1298
	int flags;

	flags = type & ~SOCK_TYPE_MASK;
1299
	if (flags & ~(SOCK_CLOEXEC | SOCK_NONBLOCK))
1300 1301
		return -EINVAL;
	type &= SOCK_TYPE_MASK;
L
Linus Torvalds 已提交
1302

U
Ulrich Drepper 已提交
1303 1304 1305
	if (SOCK_NONBLOCK != O_NONBLOCK && (flags & SOCK_NONBLOCK))
		flags = (flags & ~SOCK_NONBLOCK) | O_NONBLOCK;

L
Linus Torvalds 已提交
1306 1307 1308 1309 1310 1311 1312 1313 1314 1315 1316 1317 1318 1319
	/*
	 * 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);
1320
	if (err < 0)
L
Linus Torvalds 已提交
1321 1322
		goto out_release_both;

1323
	fd1 = get_unused_fd_flags(flags);
1324 1325
	if (unlikely(fd1 < 0)) {
		err = fd1;
A
Al Viro 已提交
1326
		goto out_release_both;
1327
	}
1328

1329
	fd2 = get_unused_fd_flags(flags);
1330 1331
	if (unlikely(fd2 < 0)) {
		err = fd2;
1332
		goto out_put_unused_1;
1333 1334
	}

1335
	newfile1 = sock_alloc_file(sock1, flags, NULL);
1336
	if (IS_ERR(newfile1)) {
1337
		err = PTR_ERR(newfile1);
1338
		goto out_put_unused_both;
1339 1340
	}

1341
	newfile2 = sock_alloc_file(sock2, flags, NULL);
1342 1343
	if (IS_ERR(newfile2)) {
		err = PTR_ERR(newfile2);
1344
		goto out_fput_1;
A
Al Viro 已提交
1345 1346
	}

1347 1348 1349 1350 1351 1352 1353 1354
	err = put_user(fd1, &usockvec[0]);
	if (err)
		goto out_fput_both;

	err = put_user(fd2, &usockvec[1]);
	if (err)
		goto out_fput_both;

A
Al Viro 已提交
1355
	audit_fd_pair(fd1, fd2);
1356

A
Al Viro 已提交
1357 1358
	fd_install(fd1, newfile1);
	fd_install(fd2, newfile2);
L
Linus Torvalds 已提交
1359 1360 1361 1362
	/* fd1 and fd2 may be already another descriptors.
	 * Not kernel problem.
	 */

1363
	return 0;
L
Linus Torvalds 已提交
1364

1365 1366 1367 1368 1369 1370 1371 1372 1373 1374 1375 1376 1377
out_fput_both:
	fput(newfile2);
	fput(newfile1);
	put_unused_fd(fd2);
	put_unused_fd(fd1);
	goto out;

out_fput_1:
	fput(newfile1);
	put_unused_fd(fd2);
	put_unused_fd(fd1);
	sock_release(sock2);
	goto out;
L
Linus Torvalds 已提交
1378

1379 1380 1381 1382
out_put_unused_both:
	put_unused_fd(fd2);
out_put_unused_1:
	put_unused_fd(fd1);
L
Linus Torvalds 已提交
1383
out_release_both:
1384
	sock_release(sock2);
L
Linus Torvalds 已提交
1385
out_release_1:
1386
	sock_release(sock1);
L
Linus Torvalds 已提交
1387 1388 1389 1390 1391 1392 1393 1394 1395 1396 1397 1398
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).
 */

1399
SYSCALL_DEFINE3(bind, int, fd, struct sockaddr __user *, umyaddr, int, addrlen)
L
Linus Torvalds 已提交
1400 1401
{
	struct socket *sock;
1402
	struct sockaddr_storage address;
1403
	int err, fput_needed;
L
Linus Torvalds 已提交
1404

1405
	sock = sockfd_lookup_light(fd, &err, &fput_needed);
1406
	if (sock) {
1407
		err = move_addr_to_kernel(umyaddr, addrlen, &address);
1408 1409
		if (err >= 0) {
			err = security_socket_bind(sock,
1410
						   (struct sockaddr *)&address,
1411
						   addrlen);
1412 1413
			if (!err)
				err = sock->ops->bind(sock,
1414
						      (struct sockaddr *)
1415
						      &address, addrlen);
L
Linus Torvalds 已提交
1416
		}
1417
		fput_light(sock->file, fput_needed);
1418
	}
L
Linus Torvalds 已提交
1419 1420 1421 1422 1423 1424 1425 1426 1427
	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.
 */

1428
SYSCALL_DEFINE2(listen, int, fd, int, backlog)
L
Linus Torvalds 已提交
1429 1430
{
	struct socket *sock;
1431
	int err, fput_needed;
1432
	int somaxconn;
1433 1434 1435

	sock = sockfd_lookup_light(fd, &err, &fput_needed);
	if (sock) {
1436
		somaxconn = sock_net(sock->sk)->core.sysctl_somaxconn;
1437
		if ((unsigned int)backlog > somaxconn)
1438
			backlog = somaxconn;
L
Linus Torvalds 已提交
1439 1440

		err = security_socket_listen(sock, backlog);
1441 1442
		if (!err)
			err = sock->ops->listen(sock, backlog);
L
Linus Torvalds 已提交
1443

1444
		fput_light(sock->file, fput_needed);
L
Linus Torvalds 已提交
1445 1446 1447 1448 1449 1450 1451 1452 1453 1454 1455 1456 1457 1458 1459 1460
	}
	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.
 */

1461 1462
SYSCALL_DEFINE4(accept4, int, fd, struct sockaddr __user *, upeer_sockaddr,
		int __user *, upeer_addrlen, int, flags)
L
Linus Torvalds 已提交
1463 1464
{
	struct socket *sock, *newsock;
1465
	struct file *newfile;
1466
	int err, len, newfd, fput_needed;
1467
	struct sockaddr_storage address;
L
Linus Torvalds 已提交
1468

1469
	if (flags & ~(SOCK_CLOEXEC | SOCK_NONBLOCK))
U
Ulrich Drepper 已提交
1470 1471 1472 1473 1474
		return -EINVAL;

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

1475
	sock = sockfd_lookup_light(fd, &err, &fput_needed);
L
Linus Torvalds 已提交
1476 1477 1478 1479
	if (!sock)
		goto out;

	err = -ENFILE;
1480 1481
	newsock = sock_alloc();
	if (!newsock)
L
Linus Torvalds 已提交
1482 1483 1484 1485 1486 1487 1488 1489 1490 1491 1492
		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);

1493
	newfd = get_unused_fd_flags(flags);
1494 1495
	if (unlikely(newfd < 0)) {
		err = newfd;
1496 1497
		sock_release(newsock);
		goto out_put;
1498
	}
1499
	newfile = sock_alloc_file(newsock, flags, sock->sk->sk_prot_creator->name);
1500
	if (IS_ERR(newfile)) {
1501 1502 1503 1504 1505
		err = PTR_ERR(newfile);
		put_unused_fd(newfd);
		sock_release(newsock);
		goto out_put;
	}
1506

1507 1508
	err = security_socket_accept(sock, newsock);
	if (err)
1509
		goto out_fd;
1510

L
Linus Torvalds 已提交
1511 1512
	err = sock->ops->accept(sock, newsock, sock->file->f_flags);
	if (err < 0)
1513
		goto out_fd;
L
Linus Torvalds 已提交
1514 1515

	if (upeer_sockaddr) {
1516
		if (newsock->ops->getname(newsock, (struct sockaddr *)&address,
1517
					  &len, 2) < 0) {
L
Linus Torvalds 已提交
1518
			err = -ECONNABORTED;
1519
			goto out_fd;
L
Linus Torvalds 已提交
1520
		}
1521
		err = move_addr_to_user(&address,
1522
					len, upeer_sockaddr, upeer_addrlen);
L
Linus Torvalds 已提交
1523
		if (err < 0)
1524
			goto out_fd;
L
Linus Torvalds 已提交
1525 1526 1527 1528
	}

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

1529 1530
	fd_install(newfd, newfile);
	err = newfd;
L
Linus Torvalds 已提交
1531 1532

out_put:
1533
	fput_light(sock->file, fput_needed);
L
Linus Torvalds 已提交
1534 1535
out:
	return err;
1536
out_fd:
1537
	fput(newfile);
1538
	put_unused_fd(newfd);
L
Linus Torvalds 已提交
1539 1540 1541
	goto out_put;
}

1542 1543
SYSCALL_DEFINE3(accept, int, fd, struct sockaddr __user *, upeer_sockaddr,
		int __user *, upeer_addrlen)
U
Ulrich Drepper 已提交
1544
{
U
Ulrich Drepper 已提交
1545
	return sys_accept4(fd, upeer_sockaddr, upeer_addrlen, 0);
U
Ulrich Drepper 已提交
1546 1547
}

L
Linus Torvalds 已提交
1548 1549 1550 1551 1552 1553 1554 1555 1556 1557 1558 1559
/*
 *	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.
 */

1560 1561
SYSCALL_DEFINE3(connect, int, fd, struct sockaddr __user *, uservaddr,
		int, addrlen)
L
Linus Torvalds 已提交
1562 1563
{
	struct socket *sock;
1564
	struct sockaddr_storage address;
1565
	int err, fput_needed;
L
Linus Torvalds 已提交
1566

1567
	sock = sockfd_lookup_light(fd, &err, &fput_needed);
L
Linus Torvalds 已提交
1568 1569
	if (!sock)
		goto out;
1570
	err = move_addr_to_kernel(uservaddr, addrlen, &address);
L
Linus Torvalds 已提交
1571 1572 1573
	if (err < 0)
		goto out_put;

1574
	err =
1575
	    security_socket_connect(sock, (struct sockaddr *)&address, addrlen);
L
Linus Torvalds 已提交
1576 1577 1578
	if (err)
		goto out_put;

1579
	err = sock->ops->connect(sock, (struct sockaddr *)&address, addrlen,
L
Linus Torvalds 已提交
1580 1581
				 sock->file->f_flags);
out_put:
1582
	fput_light(sock->file, fput_needed);
L
Linus Torvalds 已提交
1583 1584 1585 1586 1587 1588 1589 1590 1591
out:
	return err;
}

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

1592 1593
SYSCALL_DEFINE3(getsockname, int, fd, struct sockaddr __user *, usockaddr,
		int __user *, usockaddr_len)
L
Linus Torvalds 已提交
1594 1595
{
	struct socket *sock;
1596
	struct sockaddr_storage address;
1597
	int len, err, fput_needed;
1598

1599
	sock = sockfd_lookup_light(fd, &err, &fput_needed);
L
Linus Torvalds 已提交
1600 1601 1602 1603 1604 1605 1606
	if (!sock)
		goto out;

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

1607
	err = sock->ops->getname(sock, (struct sockaddr *)&address, &len, 0);
L
Linus Torvalds 已提交
1608 1609
	if (err)
		goto out_put;
1610
	err = move_addr_to_user(&address, len, usockaddr, usockaddr_len);
L
Linus Torvalds 已提交
1611 1612

out_put:
1613
	fput_light(sock->file, fput_needed);
L
Linus Torvalds 已提交
1614 1615 1616 1617 1618 1619 1620 1621 1622
out:
	return err;
}

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

1623 1624
SYSCALL_DEFINE3(getpeername, int, fd, struct sockaddr __user *, usockaddr,
		int __user *, usockaddr_len)
L
Linus Torvalds 已提交
1625 1626
{
	struct socket *sock;
1627
	struct sockaddr_storage address;
1628
	int len, err, fput_needed;
L
Linus Torvalds 已提交
1629

1630 1631
	sock = sockfd_lookup_light(fd, &err, &fput_needed);
	if (sock != NULL) {
L
Linus Torvalds 已提交
1632 1633
		err = security_socket_getpeername(sock);
		if (err) {
1634
			fput_light(sock->file, fput_needed);
L
Linus Torvalds 已提交
1635 1636 1637
			return err;
		}

1638
		err =
1639
		    sock->ops->getname(sock, (struct sockaddr *)&address, &len,
1640
				       1);
L
Linus Torvalds 已提交
1641
		if (!err)
1642
			err = move_addr_to_user(&address, len, usockaddr,
1643
						usockaddr_len);
1644
		fput_light(sock->file, fput_needed);
L
Linus Torvalds 已提交
1645 1646 1647 1648 1649 1650 1651 1652 1653 1654
	}
	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.
 */

1655
SYSCALL_DEFINE6(sendto, int, fd, void __user *, buff, size_t, len,
1656
		unsigned int, flags, struct sockaddr __user *, addr,
1657
		int, addr_len)
L
Linus Torvalds 已提交
1658 1659
{
	struct socket *sock;
1660
	struct sockaddr_storage address;
L
Linus Torvalds 已提交
1661 1662 1663
	int err;
	struct msghdr msg;
	struct iovec iov;
1664 1665
	int fput_needed;

1666 1667 1668
	err = import_single_range(WRITE, buff, len, &iov, &msg.msg_iter);
	if (unlikely(err))
		return err;
1669 1670
	sock = sockfd_lookup_light(fd, &err, &fput_needed);
	if (!sock)
1671
		goto out;
1672

1673 1674 1675 1676
	msg.msg_name = NULL;
	msg.msg_control = NULL;
	msg.msg_controllen = 0;
	msg.msg_namelen = 0;
1677
	if (addr) {
1678
		err = move_addr_to_kernel(addr, addr_len, &address);
L
Linus Torvalds 已提交
1679 1680
		if (err < 0)
			goto out_put;
1681
		msg.msg_name = (struct sockaddr *)&address;
1682
		msg.msg_namelen = addr_len;
L
Linus Torvalds 已提交
1683 1684 1685 1686
	}
	if (sock->file->f_flags & O_NONBLOCK)
		flags |= MSG_DONTWAIT;
	msg.msg_flags = flags;
1687
	err = sock_sendmsg(sock, &msg);
L
Linus Torvalds 已提交
1688

1689
out_put:
1690
	fput_light(sock->file, fput_needed);
1691
out:
L
Linus Torvalds 已提交
1692 1693 1694 1695
	return err;
}

/*
1696
 *	Send a datagram down a socket.
L
Linus Torvalds 已提交
1697 1698
 */

1699
SYSCALL_DEFINE4(send, int, fd, void __user *, buff, size_t, len,
1700
		unsigned int, flags)
L
Linus Torvalds 已提交
1701 1702 1703 1704 1705
{
	return sys_sendto(fd, buff, len, flags, NULL, 0);
}

/*
1706
 *	Receive a frame from the socket and optionally record the address of the
L
Linus Torvalds 已提交
1707 1708 1709 1710
 *	sender. We verify the buffers are writable and if needed move the
 *	sender address from kernel to user space.
 */

1711
SYSCALL_DEFINE6(recvfrom, int, fd, void __user *, ubuf, size_t, size,
1712
		unsigned int, flags, struct sockaddr __user *, addr,
1713
		int __user *, addr_len)
L
Linus Torvalds 已提交
1714 1715 1716 1717
{
	struct socket *sock;
	struct iovec iov;
	struct msghdr msg;
1718
	struct sockaddr_storage address;
1719
	int err, err2;
1720 1721
	int fput_needed;

1722 1723 1724
	err = import_single_range(READ, ubuf, size, &iov, &msg.msg_iter);
	if (unlikely(err))
		return err;
1725
	sock = sockfd_lookup_light(fd, &err, &fput_needed);
L
Linus Torvalds 已提交
1726
	if (!sock)
1727
		goto out;
L
Linus Torvalds 已提交
1728

1729 1730
	msg.msg_control = NULL;
	msg.msg_controllen = 0;
1731 1732 1733 1734
	/* Save some cycles and don't copy the address if not needed */
	msg.msg_name = addr ? (struct sockaddr *)&address : NULL;
	/* We assume all kernel code knows the size of sockaddr_storage */
	msg.msg_namelen = 0;
1735
	msg.msg_iocb = NULL;
L
Linus Torvalds 已提交
1736 1737
	if (sock->file->f_flags & O_NONBLOCK)
		flags |= MSG_DONTWAIT;
1738
	err = sock_recvmsg(sock, &msg, flags);
L
Linus Torvalds 已提交
1739

1740
	if (err >= 0 && addr != NULL) {
1741
		err2 = move_addr_to_user(&address,
1742
					 msg.msg_namelen, addr, addr_len);
1743 1744
		if (err2 < 0)
			err = err2;
L
Linus Torvalds 已提交
1745
	}
1746 1747

	fput_light(sock->file, fput_needed);
1748
out:
L
Linus Torvalds 已提交
1749 1750 1751 1752
	return err;
}

/*
1753
 *	Receive a datagram from a socket.
L
Linus Torvalds 已提交
1754 1755
 */

1756 1757
SYSCALL_DEFINE4(recv, int, fd, void __user *, ubuf, size_t, size,
		unsigned int, flags)
L
Linus Torvalds 已提交
1758 1759 1760 1761 1762 1763 1764 1765 1766
{
	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.
 */

1767 1768
SYSCALL_DEFINE5(setsockopt, int, fd, int, level, int, optname,
		char __user *, optval, int, optlen)
L
Linus Torvalds 已提交
1769
{
1770
	int err, fput_needed;
L
Linus Torvalds 已提交
1771 1772 1773 1774
	struct socket *sock;

	if (optlen < 0)
		return -EINVAL;
1775 1776 1777 1778

	sock = sockfd_lookup_light(fd, &err, &fput_needed);
	if (sock != NULL) {
		err = security_socket_setsockopt(sock, level, optname);
1779 1780
		if (err)
			goto out_put;
L
Linus Torvalds 已提交
1781 1782

		if (level == SOL_SOCKET)
1783 1784 1785
			err =
			    sock_setsockopt(sock, level, optname, optval,
					    optlen);
L
Linus Torvalds 已提交
1786
		else
1787 1788 1789
			err =
			    sock->ops->setsockopt(sock, level, optname, optval,
						  optlen);
1790 1791
out_put:
		fput_light(sock->file, fput_needed);
L
Linus Torvalds 已提交
1792 1793 1794 1795 1796 1797 1798 1799 1800
	}
	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.
 */

1801 1802
SYSCALL_DEFINE5(getsockopt, int, fd, int, level, int, optname,
		char __user *, optval, int __user *, optlen)
L
Linus Torvalds 已提交
1803
{
1804
	int err, fput_needed;
L
Linus Torvalds 已提交
1805 1806
	struct socket *sock;

1807 1808
	sock = sockfd_lookup_light(fd, &err, &fput_needed);
	if (sock != NULL) {
1809 1810 1811
		err = security_socket_getsockopt(sock, level, optname);
		if (err)
			goto out_put;
L
Linus Torvalds 已提交
1812 1813

		if (level == SOL_SOCKET)
1814 1815 1816
			err =
			    sock_getsockopt(sock, level, optname, optval,
					    optlen);
L
Linus Torvalds 已提交
1817
		else
1818 1819 1820
			err =
			    sock->ops->getsockopt(sock, level, optname, optval,
						  optlen);
1821 1822
out_put:
		fput_light(sock->file, fput_needed);
L
Linus Torvalds 已提交
1823 1824 1825 1826 1827 1828 1829 1830
	}
	return err;
}

/*
 *	Shutdown a socket.
 */

1831
SYSCALL_DEFINE2(shutdown, int, fd, int, how)
L
Linus Torvalds 已提交
1832
{
1833
	int err, fput_needed;
L
Linus Torvalds 已提交
1834 1835
	struct socket *sock;

1836 1837
	sock = sockfd_lookup_light(fd, &err, &fput_needed);
	if (sock != NULL) {
L
Linus Torvalds 已提交
1838
		err = security_socket_shutdown(sock, how);
1839 1840 1841
		if (!err)
			err = sock->ops->shutdown(sock, how);
		fput_light(sock->file, fput_needed);
L
Linus Torvalds 已提交
1842 1843 1844 1845
	}
	return err;
}

1846
/* A couple of helpful macros for getting the address of the 32/64 bit
L
Linus Torvalds 已提交
1847 1848 1849 1850 1851 1852
 * 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)

1853 1854 1855 1856 1857
struct used_address {
	struct sockaddr_storage name;
	unsigned int name_len;
};

1858 1859 1860 1861
static int copy_msghdr_from_user(struct msghdr *kmsg,
				 struct user_msghdr __user *umsg,
				 struct sockaddr __user **save_addr,
				 struct iovec **iov)
1862
{
1863 1864
	struct sockaddr __user *uaddr;
	struct iovec __user *uiov;
A
Al Viro 已提交
1865
	size_t nr_segs;
1866 1867 1868 1869 1870 1871
	ssize_t err;

	if (!access_ok(VERIFY_READ, umsg, sizeof(*umsg)) ||
	    __get_user(uaddr, &umsg->msg_name) ||
	    __get_user(kmsg->msg_namelen, &umsg->msg_namelen) ||
	    __get_user(uiov, &umsg->msg_iov) ||
A
Al Viro 已提交
1872
	    __get_user(nr_segs, &umsg->msg_iovlen) ||
1873 1874 1875
	    __get_user(kmsg->msg_control, &umsg->msg_control) ||
	    __get_user(kmsg->msg_controllen, &umsg->msg_controllen) ||
	    __get_user(kmsg->msg_flags, &umsg->msg_flags))
1876
		return -EFAULT;
1877

1878
	if (!uaddr)
1879 1880
		kmsg->msg_namelen = 0;

1881 1882 1883
	if (kmsg->msg_namelen < 0)
		return -EINVAL;

1884
	if (kmsg->msg_namelen > sizeof(struct sockaddr_storage))
1885
		kmsg->msg_namelen = sizeof(struct sockaddr_storage);
1886 1887 1888 1889 1890 1891 1892 1893 1894 1895 1896 1897 1898 1899 1900 1901

	if (save_addr)
		*save_addr = uaddr;

	if (uaddr && kmsg->msg_namelen) {
		if (!save_addr) {
			err = move_addr_to_kernel(uaddr, kmsg->msg_namelen,
						  kmsg->msg_name);
			if (err < 0)
				return err;
		}
	} else {
		kmsg->msg_name = NULL;
		kmsg->msg_namelen = 0;
	}

A
Al Viro 已提交
1902
	if (nr_segs > UIO_MAXIOV)
1903 1904
		return -EMSGSIZE;

1905 1906
	kmsg->msg_iocb = NULL;

1907 1908
	return import_iovec(save_addr ? READ : WRITE, uiov, nr_segs,
			    UIO_FASTIOV, iov, &kmsg->msg_iter);
1909 1910
}

1911
static int ___sys_sendmsg(struct socket *sock, struct user_msghdr __user *msg,
1912
			 struct msghdr *msg_sys, unsigned int flags,
1913 1914
			 struct used_address *used_address,
			 unsigned int allowed_msghdr_flags)
L
Linus Torvalds 已提交
1915
{
1916 1917
	struct compat_msghdr __user *msg_compat =
	    (struct compat_msghdr __user *)msg;
1918
	struct sockaddr_storage address;
L
Linus Torvalds 已提交
1919
	struct iovec iovstack[UIO_FASTIOV], *iov = iovstack;
1920
	unsigned char ctl[sizeof(struct cmsghdr) + 20]
1921 1922
	    __attribute__ ((aligned(sizeof(__kernel_size_t))));
	/* 20 is size of ipv6_pktinfo */
L
Linus Torvalds 已提交
1923
	unsigned char *ctl_buf = ctl;
1924
	int ctl_len;
1925
	ssize_t err;
1926

1927
	msg_sys->msg_name = &address;
L
Linus Torvalds 已提交
1928

1929
	if (MSG_CMSG_COMPAT & flags)
1930
		err = get_compat_msghdr(msg_sys, msg_compat, NULL, &iov);
1931
	else
1932
		err = copy_msghdr_from_user(msg_sys, msg, NULL, &iov);
1933
	if (err < 0)
1934
		return err;
L
Linus Torvalds 已提交
1935 1936 1937

	err = -ENOBUFS;

1938
	if (msg_sys->msg_controllen > INT_MAX)
L
Linus Torvalds 已提交
1939
		goto out_freeiov;
1940
	flags |= (msg_sys->msg_flags & allowed_msghdr_flags);
1941
	ctl_len = msg_sys->msg_controllen;
L
Linus Torvalds 已提交
1942
	if ((MSG_CMSG_COMPAT & flags) && ctl_len) {
1943
		err =
1944
		    cmsghdr_from_user_compat_to_kern(msg_sys, sock->sk, ctl,
1945
						     sizeof(ctl));
L
Linus Torvalds 已提交
1946 1947
		if (err)
			goto out_freeiov;
1948 1949
		ctl_buf = msg_sys->msg_control;
		ctl_len = msg_sys->msg_controllen;
L
Linus Torvalds 已提交
1950
	} else if (ctl_len) {
1951
		if (ctl_len > sizeof(ctl)) {
L
Linus Torvalds 已提交
1952
			ctl_buf = sock_kmalloc(sock->sk, ctl_len, GFP_KERNEL);
1953
			if (ctl_buf == NULL)
L
Linus Torvalds 已提交
1954 1955 1956 1957
				goto out_freeiov;
		}
		err = -EFAULT;
		/*
1958
		 * Careful! Before this, msg_sys->msg_control contains a user pointer.
L
Linus Torvalds 已提交
1959 1960 1961
		 * Afterwards, it will be a kernel pointer. Thus the compiler-assisted
		 * checking falls down on this.
		 */
1962
		if (copy_from_user(ctl_buf,
1963
				   (void __user __force *)msg_sys->msg_control,
1964
				   ctl_len))
L
Linus Torvalds 已提交
1965
			goto out_freectl;
1966
		msg_sys->msg_control = ctl_buf;
L
Linus Torvalds 已提交
1967
	}
1968
	msg_sys->msg_flags = flags;
L
Linus Torvalds 已提交
1969 1970

	if (sock->file->f_flags & O_NONBLOCK)
1971
		msg_sys->msg_flags |= MSG_DONTWAIT;
1972 1973 1974 1975 1976 1977
	/*
	 * 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.
	 */
1978 1979 1980
	if (used_address && msg_sys->msg_name &&
	    used_address->name_len == msg_sys->msg_namelen &&
	    !memcmp(&used_address->name, msg_sys->msg_name,
1981
		    used_address->name_len)) {
1982
		err = sock_sendmsg_nosec(sock, msg_sys);
1983 1984
		goto out_freectl;
	}
1985
	err = sock_sendmsg(sock, msg_sys);
1986 1987 1988 1989 1990 1991
	/*
	 * 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;
1992 1993 1994
		if (msg_sys->msg_name)
			memcpy(&used_address->name, msg_sys->msg_name,
			       used_address->name_len);
1995
	}
L
Linus Torvalds 已提交
1996 1997

out_freectl:
1998
	if (ctl_buf != ctl)
L
Linus Torvalds 已提交
1999 2000
		sock_kfree_s(sock->sk, ctl_buf, ctl_len);
out_freeiov:
2001
	kfree(iov);
2002 2003 2004 2005 2006 2007 2008
	return err;
}

/*
 *	BSD sendmsg interface
 */

2009
long __sys_sendmsg(int fd, struct user_msghdr __user *msg, unsigned flags)
2010 2011 2012
{
	int fput_needed, err;
	struct msghdr msg_sys;
2013 2014 2015
	struct socket *sock;

	sock = sockfd_lookup_light(fd, &err, &fput_needed);
2016 2017 2018
	if (!sock)
		goto out;

2019
	err = ___sys_sendmsg(sock, msg, &msg_sys, flags, NULL, 0);
2020

2021
	fput_light(sock->file, fput_needed);
2022
out:
L
Linus Torvalds 已提交
2023 2024 2025
	return err;
}

2026
SYSCALL_DEFINE3(sendmsg, int, fd, struct user_msghdr __user *, msg, unsigned int, flags)
2027 2028 2029 2030 2031 2032
{
	if (flags & MSG_CMSG_COMPAT)
		return -EINVAL;
	return __sys_sendmsg(fd, msg, flags);
}

2033 2034 2035 2036 2037 2038 2039 2040 2041 2042 2043 2044
/*
 *	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;
2045
	struct used_address used_address;
T
Tom Herbert 已提交
2046
	unsigned int oflags = flags;
2047

2048 2049
	if (vlen > UIO_MAXIOV)
		vlen = UIO_MAXIOV;
2050 2051 2052 2053 2054 2055 2056

	datagrams = 0;

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

2057
	used_address.name_len = UINT_MAX;
2058 2059
	entry = mmsg;
	compat_entry = (struct compat_mmsghdr __user *)mmsg;
2060
	err = 0;
T
Tom Herbert 已提交
2061
	flags |= MSG_BATCH;
2062 2063

	while (datagrams < vlen) {
T
Tom Herbert 已提交
2064 2065 2066
		if (datagrams == vlen - 1)
			flags = oflags;

2067
		if (MSG_CMSG_COMPAT & flags) {
2068
			err = ___sys_sendmsg(sock, (struct user_msghdr __user *)compat_entry,
2069
					     &msg_sys, flags, &used_address, MSG_EOR);
2070 2071 2072 2073 2074
			if (err < 0)
				break;
			err = __put_user(err, &compat_entry->msg_len);
			++compat_entry;
		} else {
2075
			err = ___sys_sendmsg(sock,
2076
					     (struct user_msghdr __user *)entry,
2077
					     &msg_sys, flags, &used_address, MSG_EOR);
2078 2079 2080 2081 2082 2083 2084 2085 2086
			if (err < 0)
				break;
			err = put_user(err, &entry->msg_len);
			++entry;
		}

		if (err)
			break;
		++datagrams;
2087 2088
		if (msg_data_left(&msg_sys))
			break;
2089
		cond_resched();
2090 2091 2092 2093
	}

	fput_light(sock->file, fput_needed);

2094 2095
	/* We only return an error if no datagrams were able to be sent */
	if (datagrams != 0)
2096 2097 2098 2099 2100 2101 2102 2103
		return datagrams;

	return err;
}

SYSCALL_DEFINE4(sendmmsg, int, fd, struct mmsghdr __user *, mmsg,
		unsigned int, vlen, unsigned int, flags)
{
2104 2105
	if (flags & MSG_CMSG_COMPAT)
		return -EINVAL;
2106 2107 2108
	return __sys_sendmmsg(fd, mmsg, vlen, flags);
}

2109
static int ___sys_recvmsg(struct socket *sock, struct user_msghdr __user *msg,
2110
			 struct msghdr *msg_sys, unsigned int flags, int nosec)
L
Linus Torvalds 已提交
2111
{
2112 2113
	struct compat_msghdr __user *msg_compat =
	    (struct compat_msghdr __user *)msg;
L
Linus Torvalds 已提交
2114
	struct iovec iovstack[UIO_FASTIOV];
2115
	struct iovec *iov = iovstack;
L
Linus Torvalds 已提交
2116
	unsigned long cmsg_ptr;
2117
	int len;
2118
	ssize_t err;
L
Linus Torvalds 已提交
2119 2120

	/* kernel mode address */
2121
	struct sockaddr_storage addr;
L
Linus Torvalds 已提交
2122 2123 2124

	/* user mode address pointers */
	struct sockaddr __user *uaddr;
2125
	int __user *uaddr_len = COMPAT_NAMELEN(msg);
2126

2127
	msg_sys->msg_name = &addr;
L
Linus Torvalds 已提交
2128

2129
	if (MSG_CMSG_COMPAT & flags)
2130
		err = get_compat_msghdr(msg_sys, msg_compat, &uaddr, &iov);
2131
	else
2132
		err = copy_msghdr_from_user(msg_sys, msg, &uaddr, &iov);
L
Linus Torvalds 已提交
2133
	if (err < 0)
2134
		return err;
L
Linus Torvalds 已提交
2135

2136 2137
	cmsg_ptr = (unsigned long)msg_sys->msg_control;
	msg_sys->msg_flags = flags & (MSG_CMSG_CLOEXEC|MSG_CMSG_COMPAT);
2138

2139 2140 2141
	/* We assume all kernel code knows the size of sockaddr_storage */
	msg_sys->msg_namelen = 0;

L
Linus Torvalds 已提交
2142 2143
	if (sock->file->f_flags & O_NONBLOCK)
		flags |= MSG_DONTWAIT;
2144
	err = (nosec ? sock_recvmsg_nosec : sock_recvmsg)(sock, msg_sys, flags);
L
Linus Torvalds 已提交
2145 2146 2147 2148 2149
	if (err < 0)
		goto out_freeiov;
	len = err;

	if (uaddr != NULL) {
2150
		err = move_addr_to_user(&addr,
2151
					msg_sys->msg_namelen, uaddr,
2152
					uaddr_len);
L
Linus Torvalds 已提交
2153 2154 2155
		if (err < 0)
			goto out_freeiov;
	}
2156
	err = __put_user((msg_sys->msg_flags & ~MSG_CMSG_COMPAT),
2157
			 COMPAT_FLAGS(msg));
L
Linus Torvalds 已提交
2158 2159 2160
	if (err)
		goto out_freeiov;
	if (MSG_CMSG_COMPAT & flags)
2161
		err = __put_user((unsigned long)msg_sys->msg_control - cmsg_ptr,
L
Linus Torvalds 已提交
2162 2163
				 &msg_compat->msg_controllen);
	else
2164
		err = __put_user((unsigned long)msg_sys->msg_control - cmsg_ptr,
L
Linus Torvalds 已提交
2165 2166 2167 2168 2169 2170
				 &msg->msg_controllen);
	if (err)
		goto out_freeiov;
	err = len;

out_freeiov:
2171
	kfree(iov);
2172 2173 2174 2175 2176 2177 2178
	return err;
}

/*
 *	BSD recvmsg interface
 */

2179
long __sys_recvmsg(int fd, struct user_msghdr __user *msg, unsigned flags)
2180 2181 2182
{
	int fput_needed, err;
	struct msghdr msg_sys;
2183 2184 2185
	struct socket *sock;

	sock = sockfd_lookup_light(fd, &err, &fput_needed);
2186 2187 2188
	if (!sock)
		goto out;

2189
	err = ___sys_recvmsg(sock, msg, &msg_sys, flags, 0);
2190

2191
	fput_light(sock->file, fput_needed);
L
Linus Torvalds 已提交
2192 2193 2194 2195
out:
	return err;
}

2196
SYSCALL_DEFINE3(recvmsg, int, fd, struct user_msghdr __user *, msg,
2197 2198 2199 2200 2201 2202 2203
		unsigned int, flags)
{
	if (flags & MSG_CMSG_COMPAT)
		return -EINVAL;
	return __sys_recvmsg(fd, msg, flags);
}

2204 2205 2206 2207 2208 2209 2210 2211 2212 2213
/*
 *     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;
2214
	struct compat_mmsghdr __user *compat_entry;
2215
	struct msghdr msg_sys;
2216 2217
	struct timespec64 end_time;
	struct timespec64 timeout64;
2218 2219 2220 2221 2222 2223 2224 2225 2226 2227 2228 2229 2230 2231 2232 2233 2234

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

	datagrams = 0;

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

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

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

	while (datagrams < vlen) {
		/*
		 * No need to ask LSM for more than the first datagram.
		 */
2241
		if (MSG_CMSG_COMPAT & flags) {
2242
			err = ___sys_recvmsg(sock, (struct user_msghdr __user *)compat_entry,
2243 2244
					     &msg_sys, flags & ~MSG_WAITFORONE,
					     datagrams);
2245 2246 2247 2248 2249
			if (err < 0)
				break;
			err = __put_user(err, &compat_entry->msg_len);
			++compat_entry;
		} else {
2250
			err = ___sys_recvmsg(sock,
2251
					     (struct user_msghdr __user *)entry,
2252 2253
					     &msg_sys, flags & ~MSG_WAITFORONE,
					     datagrams);
2254 2255 2256 2257 2258 2259
			if (err < 0)
				break;
			err = put_user(err, &entry->msg_len);
			++entry;
		}

2260 2261 2262 2263
		if (err)
			break;
		++datagrams;

2264 2265 2266 2267
		/* MSG_WAITFORONE turns on MSG_DONTWAIT after one packet */
		if (flags & MSG_WAITFORONE)
			flags |= MSG_DONTWAIT;

2268
		if (timeout) {
2269 2270 2271
			ktime_get_ts64(&timeout64);
			*timeout = timespec64_to_timespec(
					timespec64_sub(end_time, timeout64));
2272 2273 2274 2275 2276 2277 2278 2279 2280 2281 2282 2283 2284
			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;
2285
		cond_resched();
2286 2287 2288
	}

	if (err == 0)
2289 2290 2291 2292 2293 2294
		goto out_put;

	if (datagrams == 0) {
		datagrams = err;
		goto out_put;
	}
2295

2296 2297 2298 2299 2300
	/*
	 * We may return less entries than requested (vlen) if the
	 * sock is non block and there aren't enough datagrams...
	 */
	if (err != -EAGAIN) {
2301
		/*
2302 2303 2304 2305
		 * ... 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).
2306
		 */
2307
		sock->sk->sk_err = -err;
2308
	}
2309 2310
out_put:
	fput_light(sock->file, fput_needed);
2311

2312
	return datagrams;
2313 2314 2315 2316 2317 2318 2319 2320 2321
}

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;

2322 2323 2324
	if (flags & MSG_CMSG_COMPAT)
		return -EINVAL;

2325 2326 2327 2328 2329 2330 2331 2332 2333 2334 2335 2336 2337 2338 2339 2340
	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
2341 2342
/* Argument list sizes for sys_socketcall */
#define AL(x) ((x) * sizeof(unsigned long))
2343
static const unsigned char nargs[21] = {
2344 2345 2346
	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),
2347
	AL(4), AL(5), AL(4)
2348 2349
};

L
Linus Torvalds 已提交
2350 2351 2352
#undef AL

/*
2353
 *	System call vectors.
L
Linus Torvalds 已提交
2354 2355 2356
 *
 *	Argument checking cleaned up. Saved 20% in size.
 *  This function doesn't need to set the kernel lock because
2357
 *  it is set by the callees.
L
Linus Torvalds 已提交
2358 2359
 */

2360
SYSCALL_DEFINE2(socketcall, int, call, unsigned long __user *, args)
L
Linus Torvalds 已提交
2361
{
2362
	unsigned long a[AUDITSC_ARGS];
2363
	unsigned long a0, a1;
L
Linus Torvalds 已提交
2364
	int err;
2365
	unsigned int len;
L
Linus Torvalds 已提交
2366

2367
	if (call < 1 || call > SYS_SENDMMSG)
L
Linus Torvalds 已提交
2368 2369
		return -EINVAL;

2370 2371 2372 2373
	len = nargs[call];
	if (len > sizeof(a))
		return -EINVAL;

L
Linus Torvalds 已提交
2374
	/* copy_from_user should be SMP safe. */
2375
	if (copy_from_user(a, args, len))
L
Linus Torvalds 已提交
2376
		return -EFAULT;
2377

2378 2379 2380
	err = audit_socketcall(nargs[call] / sizeof(unsigned long), a);
	if (err)
		return err;
2381

2382 2383 2384 2385 2386 2387 2388 2389 2390 2391 2392 2393 2394 2395 2396 2397 2398
	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
2399 2400
		err = sys_accept4(a0, (struct sockaddr __user *)a1,
				  (int __user *)a[2], 0);
2401 2402 2403 2404 2405 2406 2407 2408 2409 2410 2411 2412 2413 2414 2415 2416 2417 2418 2419 2420 2421 2422 2423 2424 2425 2426 2427 2428 2429 2430 2431 2432 2433 2434 2435 2436 2437 2438 2439 2440 2441
		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:
2442
		err = sys_sendmsg(a0, (struct user_msghdr __user *)a1, a[2]);
2443
		break;
2444 2445 2446
	case SYS_SENDMMSG:
		err = sys_sendmmsg(a0, (struct mmsghdr __user *)a1, a[2], a[3]);
		break;
2447
	case SYS_RECVMSG:
2448
		err = sys_recvmsg(a0, (struct user_msghdr __user *)a1, a[2]);
2449
		break;
2450 2451 2452 2453
	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
2454 2455 2456
	case SYS_ACCEPT4:
		err = sys_accept4(a0, (struct sockaddr __user *)a1,
				  (int __user *)a[2], a[3]);
U
Ulrich Drepper 已提交
2457
		break;
2458 2459 2460
	default:
		err = -EINVAL;
		break;
L
Linus Torvalds 已提交
2461 2462 2463 2464
	}
	return err;
}

2465
#endif				/* __ARCH_WANT_SYS_SOCKETCALL */
L
Linus Torvalds 已提交
2466

2467 2468 2469 2470
/**
 *	sock_register - add a socket protocol handler
 *	@ops: description of protocol
 *
L
Linus Torvalds 已提交
2471 2472
 *	This function is called by a protocol handler that wants to
 *	advertise its address family, and have it linked into the
2473
 *	socket interface. The value ops->family corresponds to the
2474
 *	socket system call protocol family.
L
Linus Torvalds 已提交
2475
 */
2476
int sock_register(const struct net_proto_family *ops)
L
Linus Torvalds 已提交
2477 2478 2479 2480
{
	int err;

	if (ops->family >= NPROTO) {
2481
		pr_crit("protocol %d >= NPROTO(%d)\n", ops->family, NPROTO);
L
Linus Torvalds 已提交
2482 2483
		return -ENOBUFS;
	}
2484 2485

	spin_lock(&net_family_lock);
E
Eric Dumazet 已提交
2486 2487
	if (rcu_dereference_protected(net_families[ops->family],
				      lockdep_is_held(&net_family_lock)))
2488 2489
		err = -EEXIST;
	else {
2490
		rcu_assign_pointer(net_families[ops->family], ops);
L
Linus Torvalds 已提交
2491 2492
		err = 0;
	}
2493 2494
	spin_unlock(&net_family_lock);

2495
	pr_info("NET: Registered protocol family %d\n", ops->family);
L
Linus Torvalds 已提交
2496 2497
	return err;
}
2498
EXPORT_SYMBOL(sock_register);
L
Linus Torvalds 已提交
2499

2500 2501 2502 2503
/**
 *	sock_unregister - remove a protocol handler
 *	@family: protocol family to remove
 *
L
Linus Torvalds 已提交
2504 2505
 *	This function is called by a protocol handler that wants to
 *	remove its address family, and have it unlinked from the
2506 2507 2508 2509 2510 2511
 *	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
2512
 */
2513
void sock_unregister(int family)
L
Linus Torvalds 已提交
2514
{
2515
	BUG_ON(family < 0 || family >= NPROTO);
L
Linus Torvalds 已提交
2516

2517
	spin_lock(&net_family_lock);
2518
	RCU_INIT_POINTER(net_families[family], NULL);
2519 2520 2521 2522
	spin_unlock(&net_family_lock);

	synchronize_rcu();

2523
	pr_info("NET: Unregistered protocol family %d\n", family);
L
Linus Torvalds 已提交
2524
}
2525
EXPORT_SYMBOL(sock_unregister);
L
Linus Torvalds 已提交
2526

2527
static int __init sock_init(void)
L
Linus Torvalds 已提交
2528
{
N
Nick Piggin 已提交
2529
	int err;
2530 2531 2532 2533 2534 2535
	/*
	 *      Initialize the network sysctl infrastructure.
	 */
	err = net_sysctl_init();
	if (err)
		goto out;
N
Nick Piggin 已提交
2536

L
Linus Torvalds 已提交
2537
	/*
2538
	 *      Initialize skbuff SLAB cache
L
Linus Torvalds 已提交
2539 2540 2541 2542
	 */
	skb_init();

	/*
2543
	 *      Initialize the protocols module.
L
Linus Torvalds 已提交
2544 2545 2546
	 */

	init_inodecache();
N
Nick Piggin 已提交
2547 2548 2549 2550

	err = register_filesystem(&sock_fs_type);
	if (err)
		goto out_fs;
L
Linus Torvalds 已提交
2551
	sock_mnt = kern_mount(&sock_fs_type);
N
Nick Piggin 已提交
2552 2553 2554 2555
	if (IS_ERR(sock_mnt)) {
		err = PTR_ERR(sock_mnt);
		goto out_mount;
	}
2556 2557

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

#ifdef CONFIG_NETFILTER
2561 2562 2563
	err = netfilter_init();
	if (err)
		goto out;
L
Linus Torvalds 已提交
2564
#endif
2565

2566
	ptp_classifier_init();
2567

N
Nick Piggin 已提交
2568 2569 2570 2571 2572 2573 2574
out:
	return err;

out_mount:
	unregister_filesystem(&sock_fs_type);
out_fs:
	goto out;
L
Linus Torvalds 已提交
2575 2576
}

2577 2578
core_initcall(sock_init);	/* early initcall */

L
Linus Torvalds 已提交
2579 2580 2581 2582 2583 2584
#ifdef CONFIG_PROC_FS
void socket_seq_show(struct seq_file *seq)
{
	int cpu;
	int counter = 0;

2585
	for_each_possible_cpu(cpu)
2586
	    counter += per_cpu(sockets_in_use, cpu);
L
Linus Torvalds 已提交
2587 2588 2589 2590 2591 2592 2593

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

	seq_printf(seq, "sockets: used %d\n", counter);
}
2594
#endif				/* CONFIG_PROC_FS */
L
Linus Torvalds 已提交
2595

2596
#ifdef CONFIG_COMPAT
2597
static int do_siocgstamp(struct net *net, struct socket *sock,
2598
			 unsigned int cmd, void __user *up)
2599 2600 2601 2602 2603 2604
{
	mm_segment_t old_fs = get_fs();
	struct timeval ktv;
	int err;

	set_fs(KERNEL_DS);
2605
	err = sock_do_ioctl(net, sock, cmd, (unsigned long)&ktv);
2606
	set_fs(old_fs);
2607
	if (!err)
2608
		err = compat_put_timeval(&ktv, up);
2609

2610 2611 2612
	return err;
}

2613
static int do_siocgstampns(struct net *net, struct socket *sock,
2614
			   unsigned int cmd, void __user *up)
2615 2616 2617 2618 2619 2620
{
	mm_segment_t old_fs = get_fs();
	struct timespec kts;
	int err;

	set_fs(KERNEL_DS);
2621
	err = sock_do_ioctl(net, sock, cmd, (unsigned long)&kts);
2622
	set_fs(old_fs);
2623
	if (!err)
2624
		err = compat_put_timespec(&kts, up);
2625

2626 2627 2628
	return err;
}

2629
static int dev_ifname32(struct net *net, struct compat_ifreq __user *uifr32)
2630 2631 2632 2633 2634
{
	struct ifreq __user *uifr;
	int err;

	uifr = compat_alloc_user_space(sizeof(struct ifreq));
2635
	if (copy_in_user(uifr, uifr32, sizeof(struct compat_ifreq)))
2636 2637
		return -EFAULT;

2638
	err = dev_ioctl(net, SIOCGIFNAME, uifr);
2639 2640 2641
	if (err)
		return err;

2642
	if (copy_in_user(uifr32, uifr, sizeof(struct compat_ifreq)))
2643 2644 2645 2646 2647
		return -EFAULT;

	return 0;
}

2648
static int dev_ifconf(struct net *net, struct compat_ifconf __user *uifc32)
2649
{
2650
	struct compat_ifconf ifc32;
2651 2652
	struct ifconf ifc;
	struct ifconf __user *uifc;
2653
	struct compat_ifreq __user *ifr32;
2654 2655 2656 2657
	struct ifreq __user *ifr;
	unsigned int i, j;
	int err;

2658
	if (copy_from_user(&ifc32, uifc32, sizeof(struct compat_ifconf)))
2659 2660
		return -EFAULT;

2661
	memset(&ifc, 0, sizeof(ifc));
2662 2663 2664 2665 2666 2667
	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 {
2668 2669
		size_t len = ((ifc32.ifc_len / sizeof(struct compat_ifreq)) + 1) *
			sizeof(struct ifreq);
2670 2671 2672 2673
		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);
2674
		for (i = 0; i < ifc32.ifc_len; i += sizeof(struct compat_ifreq)) {
2675
			if (copy_in_user(ifr, ifr32, sizeof(struct compat_ifreq)))
2676 2677 2678 2679 2680 2681 2682 2683
				return -EFAULT;
			ifr++;
			ifr32++;
		}
	}
	if (copy_to_user(uifc, &ifc, sizeof(struct ifconf)))
		return -EFAULT;

2684
	err = dev_ioctl(net, SIOCGIFCONF, uifc);
2685 2686 2687 2688 2689 2690 2691 2692 2693
	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;
2694 2695 2696
	     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)))
2697 2698 2699 2700 2701 2702 2703 2704 2705 2706
			return -EFAULT;
		ifr32++;
		ifr++;
	}

	if (ifc32.ifcbuf == 0) {
		/* Translate from 64-bit structure multiple to
		 * a 32-bit one.
		 */
		i = ifc.ifc_len;
2707
		i = ((i / sizeof(struct ifreq)) * sizeof(struct compat_ifreq));
2708 2709 2710 2711
		ifc32.ifc_len = i;
	} else {
		ifc32.ifc_len = i;
	}
2712
	if (copy_to_user(uifc32, &ifc32, sizeof(struct compat_ifconf)))
2713 2714 2715 2716 2717
		return -EFAULT;

	return 0;
}

2718
static int ethtool_ioctl(struct net *net, struct compat_ifreq __user *ifr32)
2719
{
2720 2721 2722 2723
	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;
2724
	struct ifreq __user *ifr;
2725 2726
	u32 rule_cnt = 0, actual_rule_cnt;
	u32 ethcmd;
2727
	u32 data;
2728
	int ret;
2729

2730 2731
	if (get_user(data, &ifr32->ifr_ifru.ifru_data))
		return -EFAULT;
2732

2733 2734 2735
	compat_rxnfc = compat_ptr(data);

	if (get_user(ethcmd, &compat_rxnfc->cmd))
2736 2737
		return -EFAULT;

2738 2739 2740 2741 2742 2743 2744 2745 2746 2747 2748 2749 2750 2751 2752 2753 2754
	/* 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:
2755
	case ETHTOOL_SRXCLSRLINS:
2756 2757 2758 2759 2760 2761 2762 2763 2764
		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
2765
	rxnfc = (void __user *)ifr + ALIGN(sizeof(struct ifreq), 8);
2766 2767

	if (copy_in_user(&ifr->ifr_name, &ifr32->ifr_name, IFNAMSIZ))
2768 2769
		return -EFAULT;

2770 2771
	if (put_user(convert_in ? rxnfc : compat_ptr(data),
		     &ifr->ifr_ifru.ifru_data))
2772 2773
		return -EFAULT;

2774
	if (convert_in) {
2775
		/* We expect there to be holes between fs.m_ext and
2776 2777
		 * fs.ring_cookie and at the end of fs, but nowhere else.
		 */
2778 2779 2780 2781
		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));
2782 2783 2784 2785 2786 2787 2788
		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
2789 2790
				 (void __user *)(&rxnfc->fs.m_ext + 1) -
				 (void __user *)rxnfc) ||
2791 2792
		    copy_in_user(&rxnfc->fs.ring_cookie,
				 &compat_rxnfc->fs.ring_cookie,
S
Stephen Hemminger 已提交
2793 2794
				 (void __user *)(&rxnfc->fs.location + 1) -
				 (void __user *)&rxnfc->fs.ring_cookie) ||
2795 2796 2797 2798 2799 2800 2801 2802 2803 2804 2805
		    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
2806 2807
				 (const void __user *)(&rxnfc->fs.m_ext + 1) -
				 (const void __user *)rxnfc) ||
2808 2809
		    copy_in_user(&compat_rxnfc->fs.ring_cookie,
				 &rxnfc->fs.ring_cookie,
S
Stephen Hemminger 已提交
2810 2811
				 (const void __user *)(&rxnfc->fs.location + 1) -
				 (const void __user *)&rxnfc->fs.ring_cookie) ||
2812 2813 2814 2815 2816 2817 2818 2819 2820 2821 2822 2823 2824 2825 2826 2827 2828 2829 2830 2831 2832 2833 2834 2835 2836
		    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;
2837 2838
}

2839 2840 2841 2842 2843 2844
static int compat_siocwandev(struct net *net, struct compat_ifreq __user *uifr32)
{
	void __user *uptr;
	compat_uptr_t uptr32;
	struct ifreq __user *uifr;

2845
	uifr = compat_alloc_user_space(sizeof(*uifr));
2846 2847 2848 2849 2850 2851 2852 2853 2854 2855 2856 2857 2858 2859
	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);
}

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

	switch (cmd) {
	case SIOCBONDENSLAVE:
	case SIOCBONDRELEASE:
	case SIOCBONDSETHWADDR:
	case SIOCBONDCHANGEACTIVE:
2872
		if (copy_from_user(&kifr, ifr32, sizeof(struct compat_ifreq)))
2873 2874 2875
			return -EFAULT;

		old_fs = get_fs();
2876
		set_fs(KERNEL_DS);
2877 2878
		err = dev_ioctl(net, cmd,
				(struct ifreq __user __force *) &kifr);
2879
		set_fs(old_fs);
2880 2881 2882

		return err;
	default:
2883
		return -ENOIOCTLCMD;
2884
	}
2885 2886
}

2887 2888
/* Handle ioctls that use ifreq::ifr_data and just need struct ifreq converted */
static int compat_ifr_data_ioctl(struct net *net, unsigned int cmd,
2889
				 struct compat_ifreq __user *u_ifreq32)
2890 2891 2892 2893 2894 2895 2896 2897 2898
{
	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;
2899
	if (get_user(data32, &u_ifreq32->ifr_ifru.ifru_data))
2900 2901 2902 2903 2904 2905 2906 2907
		return -EFAULT;
	data64 = compat_ptr(data32);

	u_ifreq64 = compat_alloc_user_space(sizeof(*u_ifreq64));

	if (copy_to_user(&u_ifreq64->ifr_ifrn.ifrn_name[0], &tmp_buf[0],
			 IFNAMSIZ))
		return -EFAULT;
2908
	if (put_user(data64, &u_ifreq64->ifr_ifru.ifru_data))
2909 2910
		return -EFAULT;

2911
	return dev_ioctl(net, cmd, u_ifreq64);
2912 2913
}

2914 2915
static int dev_ifsioc(struct net *net, struct socket *sock,
			 unsigned int cmd, struct compat_ifreq __user *uifr32)
2916
{
2917
	struct ifreq __user *uifr;
2918 2919
	int err;

2920 2921 2922 2923 2924 2925
	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);

2926 2927 2928 2929 2930 2931 2932 2933 2934 2935 2936 2937
	if (!err) {
		switch (cmd) {
		case SIOCGIFFLAGS:
		case SIOCGIFMETRIC:
		case SIOCGIFMTU:
		case SIOCGIFMEM:
		case SIOCGIFHWADDR:
		case SIOCGIFINDEX:
		case SIOCGIFADDR:
		case SIOCGIFBRDADDR:
		case SIOCGIFDSTADDR:
		case SIOCGIFNETMASK:
2938
		case SIOCGIFPFLAGS:
2939
		case SIOCGIFTXQLEN:
2940 2941
		case SIOCGMIIPHY:
		case SIOCGMIIREG:
2942
			if (copy_in_user(uifr32, uifr, sizeof(*uifr32)))
2943 2944 2945 2946 2947 2948 2949
				err = -EFAULT;
			break;
		}
	}
	return err;
}

2950 2951 2952 2953 2954 2955 2956 2957 2958 2959
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));
2960 2961 2962 2963 2964 2965
	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);
2966 2967 2968 2969
	if (err)
		return -EFAULT;

	old_fs = get_fs();
2970
	set_fs(KERNEL_DS);
2971
	err = dev_ioctl(net, cmd, (void  __user __force *)&ifr);
2972
	set_fs(old_fs);
2973 2974 2975

	if (cmd == SIOCGIFMAP && !err) {
		err = copy_to_user(uifr32, &ifr, sizeof(ifr.ifr_name));
2976 2977 2978 2979 2980 2981
		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);
2982 2983 2984 2985 2986 2987
		if (err)
			err = -EFAULT;
	}
	return err;
}

2988
struct rtentry32 {
2989
	u32		rt_pad1;
2990 2991 2992
	struct sockaddr rt_dst;         /* target address               */
	struct sockaddr rt_gateway;     /* gateway addr (RTF_GATEWAY)   */
	struct sockaddr rt_genmask;     /* target network mask (IP)     */
2993 2994 2995 2996 2997 2998 2999
	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! */
3000
	/* char * */ u32 rt_dev;        /* forcing the device at add    */
3001 3002
	u32		rt_mtu;         /* per route MTU/Window         */
	u32		rt_window;      /* Window clamping              */
3003 3004 3005 3006 3007 3008 3009 3010 3011 3012 3013 3014 3015 3016 3017 3018
	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;
};

3019 3020
static int routing_ioctl(struct net *net, struct socket *sock,
			 unsigned int cmd, void __user *argp)
3021 3022 3023 3024 3025 3026 3027 3028 3029
{
	int ret;
	void *r = NULL;
	struct in6_rtmsg r6;
	struct rtentry r4;
	char devname[16];
	u32 rtdev;
	mm_segment_t old_fs = get_fs();

3030 3031
	if (sock && sock->sk && sock->sk->sk_family == AF_INET6) { /* ipv6 */
		struct in6_rtmsg32 __user *ur6 = argp;
3032
		ret = copy_from_user(&r6.rtmsg_dst, &(ur6->rtmsg_dst),
3033
			3 * sizeof(struct in6_addr));
3034 3035 3036 3037 3038 3039 3040
		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));
3041 3042 3043

		r = (void *) &r6;
	} else { /* ipv4 */
3044
		struct rtentry32 __user *ur4 = argp;
3045
		ret = copy_from_user(&r4.rt_dst, &(ur4->rt_dst),
3046
					3 * sizeof(struct sockaddr));
3047 3048 3049 3050 3051 3052
		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));
3053
		if (rtdev) {
3054
			ret |= copy_from_user(devname, compat_ptr(rtdev), 15);
3055 3056
			r4.rt_dev = (char __user __force *)devname;
			devname[15] = 0;
3057 3058 3059 3060 3061 3062 3063 3064 3065 3066 3067
		} else
			r4.rt_dev = NULL;

		r = (void *) &r4;
	}

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

3068
	set_fs(KERNEL_DS);
3069
	ret = sock_do_ioctl(net, sock, cmd, (unsigned long) r);
3070
	set_fs(old_fs);
3071 3072 3073 3074 3075 3076 3077

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
3078
 * use compatible ioctls
3079
 */
3080
static int old_bridge_ioctl(compat_ulong_t __user *argp)
3081
{
3082
	compat_ulong_t tmp;
3083

3084
	if (get_user(tmp, argp))
3085 3086 3087 3088 3089 3090
		return -EFAULT;
	if (tmp == BRCTL_GET_VERSION)
		return BRCTL_VERSION + 1;
	return -EINVAL;
}

3091 3092 3093 3094 3095 3096
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);
3097

3098
	if (cmd >= SIOCDEVPRIVATE && cmd <= (SIOCDEVPRIVATE + 15))
3099
		return compat_ifr_data_ioctl(net, cmd, argp);
3100 3101 3102 3103 3104 3105 3106 3107 3108 3109 3110

	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);
3111 3112
	case SIOCWANDEV:
		return compat_siocwandev(net, argp);
3113 3114 3115
	case SIOCGIFMAP:
	case SIOCSIFMAP:
		return compat_sioc_ifmap(net, cmd, argp);
3116 3117 3118 3119 3120 3121 3122 3123 3124 3125 3126 3127
	case SIOCBONDENSLAVE:
	case SIOCBONDRELEASE:
	case SIOCBONDSETHWADDR:
	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);
3128 3129
	case SIOCBONDSLAVEINFOQUERY:
	case SIOCBONDINFOQUERY:
3130
	case SIOCSHWTSTAMP:
3131
	case SIOCGHWTSTAMP:
3132
		return compat_ifr_data_ioctl(net, cmd, argp);
3133 3134 3135 3136 3137 3138 3139 3140 3141 3142 3143

	case FIOSETOWN:
	case SIOCSPGRP:
	case FIOGETOWN:
	case SIOCGPGRP:
	case SIOCBRADDBR:
	case SIOCBRDELBR:
	case SIOCGIFVLAN:
	case SIOCSIFVLAN:
	case SIOCADDDLCI:
	case SIOCDELDLCI:
3144
	case SIOCGSKNS:
3145 3146 3147 3148 3149 3150 3151 3152 3153 3154 3155 3156 3157 3158 3159 3160 3161 3162 3163 3164 3165 3166 3167 3168 3169 3170 3171 3172 3173 3174 3175
		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:
3176 3177 3178 3179
	case SIOCSIFNAME:
	case SIOCGMIIPHY:
	case SIOCGMIIREG:
	case SIOCSMIIREG:
3180
		return dev_ifsioc(net, sock, cmd, argp);
3181

3182 3183 3184 3185
	case SIOCSARP:
	case SIOCGARP:
	case SIOCDARP:
	case SIOCATMARK:
3186 3187 3188
		return sock_do_ioctl(net, sock, cmd, arg);
	}

3189 3190
	return -ENOIOCTLCMD;
}
3191

3192
static long compat_sock_ioctl(struct file *file, unsigned int cmd,
3193
			      unsigned long arg)
3194 3195 3196
{
	struct socket *sock = file->private_data;
	int ret = -ENOIOCTLCMD;
3197 3198 3199 3200 3201
	struct sock *sk;
	struct net *net;

	sk = sock->sk;
	net = sock_net(sk);
3202 3203 3204 3205

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

3206 3207 3208 3209
	if (ret == -ENOIOCTLCMD &&
	    (cmd >= SIOCIWFIRST && cmd <= SIOCIWLAST))
		ret = compat_wext_handle_ioctl(net, cmd, arg);

3210 3211 3212
	if (ret == -ENOIOCTLCMD)
		ret = compat_sock_ioctl_trans(file, sock, cmd, arg);

3213 3214 3215 3216
	return ret;
}
#endif

3217 3218 3219 3220
int kernel_bind(struct socket *sock, struct sockaddr *addr, int addrlen)
{
	return sock->ops->bind(sock, addr, addrlen);
}
3221
EXPORT_SYMBOL(kernel_bind);
3222 3223 3224 3225 3226

int kernel_listen(struct socket *sock, int backlog)
{
	return sock->ops->listen(sock, backlog);
}
3227
EXPORT_SYMBOL(kernel_listen);
3228 3229 3230 3231 3232 3233 3234 3235 3236 3237 3238 3239 3240 3241

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);
3242
		*newsock = NULL;
3243 3244 3245 3246
		goto done;
	}

	(*newsock)->ops = sock->ops;
3247
	__module_get((*newsock)->ops->owner);
3248 3249 3250 3251

done:
	return err;
}
3252
EXPORT_SYMBOL(kernel_accept);
3253 3254

int kernel_connect(struct socket *sock, struct sockaddr *addr, int addrlen,
3255
		   int flags)
3256 3257 3258
{
	return sock->ops->connect(sock, addr, addrlen, flags);
}
3259
EXPORT_SYMBOL(kernel_connect);
3260 3261 3262 3263 3264 3265

int kernel_getsockname(struct socket *sock, struct sockaddr *addr,
			 int *addrlen)
{
	return sock->ops->getname(sock, addr, addrlen, 0);
}
3266
EXPORT_SYMBOL(kernel_getsockname);
3267 3268 3269 3270 3271 3272

int kernel_getpeername(struct socket *sock, struct sockaddr *addr,
			 int *addrlen)
{
	return sock->ops->getname(sock, addr, addrlen, 1);
}
3273
EXPORT_SYMBOL(kernel_getpeername);
3274 3275 3276 3277 3278

int kernel_getsockopt(struct socket *sock, int level, int optname,
			char *optval, int *optlen)
{
	mm_segment_t oldfs = get_fs();
3279 3280
	char __user *uoptval;
	int __user *uoptlen;
3281 3282
	int err;

3283 3284 3285
	uoptval = (char __user __force *) optval;
	uoptlen = (int __user __force *) optlen;

3286 3287
	set_fs(KERNEL_DS);
	if (level == SOL_SOCKET)
3288
		err = sock_getsockopt(sock, level, optname, uoptval, uoptlen);
3289
	else
3290 3291
		err = sock->ops->getsockopt(sock, level, optname, uoptval,
					    uoptlen);
3292 3293 3294
	set_fs(oldfs);
	return err;
}
3295
EXPORT_SYMBOL(kernel_getsockopt);
3296 3297

int kernel_setsockopt(struct socket *sock, int level, int optname,
3298
			char *optval, unsigned int optlen)
3299 3300
{
	mm_segment_t oldfs = get_fs();
3301
	char __user *uoptval;
3302 3303
	int err;

3304 3305
	uoptval = (char __user __force *) optval;

3306 3307
	set_fs(KERNEL_DS);
	if (level == SOL_SOCKET)
3308
		err = sock_setsockopt(sock, level, optname, uoptval, optlen);
3309
	else
3310
		err = sock->ops->setsockopt(sock, level, optname, uoptval,
3311 3312 3313 3314
					    optlen);
	set_fs(oldfs);
	return err;
}
3315
EXPORT_SYMBOL(kernel_setsockopt);
3316 3317 3318 3319 3320 3321 3322 3323 3324

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);
}
3325
EXPORT_SYMBOL(kernel_sendpage);
3326 3327 3328 3329 3330 3331 3332 3333 3334 3335 3336 3337

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;
}
3338
EXPORT_SYMBOL(kernel_sock_ioctl);
3339

3340 3341 3342 3343 3344
int kernel_sock_shutdown(struct socket *sock, enum sock_shutdown_cmd how)
{
	return sock->ops->shutdown(sock, how);
}
EXPORT_SYMBOL(kernel_sock_shutdown);