socket.c 82.5 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
#include <linux/uaccess.h>
L
Linus Torvalds 已提交
94 95 96
#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

290
static void init_inodecache(void)
L
Linus Torvalds 已提交
291 292
{
	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);
299
	BUG_ON(sock_inode_cachep == NULL);
L
Linus Torvalds 已提交
300 301
}

302
static const struct super_operations sockfs_ops = {
303 304 305
	.alloc_inode	= sock_alloc_inode,
	.destroy_inode	= sock_destroy_inode,
	.statfs		= simple_statfs,
L
Linus Torvalds 已提交
306 307
};

308 309 310 311 312 313
/*
 * 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
314
				d_inode(dentry)->i_ino);
315 316
}

A
Al Viro 已提交
317
static const struct dentry_operations sockfs_dentry_operations = {
318
	.d_dname  = sockfs_dname,
L
Linus Torvalds 已提交
319 320
};

321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341
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,
};

342 343 344 345 346 347 348 349 350 351 352 353 354 355
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,
};

356 357
static const struct xattr_handler *sockfs_xattr_handlers[] = {
	&sockfs_xattr_handler,
358
	&sockfs_security_xattr_handler,
359 360 361
	NULL
};

362 363 364
static struct dentry *sockfs_mount(struct file_system_type *fs_type,
			 int flags, const char *dev_name, void *data)
{
365 366 367
	return mount_pseudo_xattr(fs_type, "socket:", &sockfs_ops,
				  sockfs_xattr_handlers,
				  &sockfs_dentry_operations, SOCKFS_MAGIC);
368 369 370 371 372 373 374 375 376 377
}

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
378 379 380
/*
 *	Obtains the first available file descriptor and sets it up for use.
 *
381 382
 *	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
383 384 385 386 387 388 389 390 391 392 393 394
 *	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.
 */

395
struct file *sock_alloc_file(struct socket *sock, int flags, const char *dname)
L
Linus Torvalds 已提交
396
{
397
	struct qstr name = { .name = "" };
398
	struct path path;
399
	struct file *file;
L
Linus Torvalds 已提交
400

401 402 403 404 405 406 407
	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);
	}
408
	path.dentry = d_alloc_pseudo(sock_mnt->mnt_sb, &name);
409 410
	if (unlikely(!path.dentry))
		return ERR_PTR(-ENOMEM);
411
	path.mnt = mntget(sock_mnt);
412

413
	d_instantiate(path.dentry, SOCK_INODE(sock));
414

415
	file = alloc_file(&path, FMODE_READ | FMODE_WRITE,
416
		  &socket_file_ops);
417
	if (IS_ERR(file)) {
A
Al Viro 已提交
418
		/* drop dentry, keep inode */
D
David Howells 已提交
419
		ihold(d_inode(path.dentry));
420
		path_put(&path);
421
		return file;
A
Al Viro 已提交
422 423 424
	}

	sock->file = file;
425
	file->f_flags = O_RDWR | (flags & O_NONBLOCK);
426
	file->private_data = sock;
427
	return file;
428
}
429
EXPORT_SYMBOL(sock_alloc_file);
430

431
static int sock_map_fd(struct socket *sock, int flags)
432 433
{
	struct file *newfile;
434 435 436
	int fd = get_unused_fd_flags(flags);
	if (unlikely(fd < 0))
		return fd;
437

438
	newfile = sock_alloc_file(sock, flags, NULL);
439
	if (likely(!IS_ERR(newfile))) {
440
		fd_install(fd, newfile);
441 442
		return fd;
	}
443

444 445
	put_unused_fd(fd);
	return PTR_ERR(newfile);
L
Linus Torvalds 已提交
446 447
}

448
struct socket *sock_from_file(struct file *file, int *err)
449 450 451 452
{
	if (file->f_op == &socket_file_ops)
		return file->private_data;	/* set in sock_map_fd */

E
Eric Dumazet 已提交
453 454
	*err = -ENOTSOCK;
	return NULL;
455
}
456
EXPORT_SYMBOL(sock_from_file);
457

L
Linus Torvalds 已提交
458
/**
459
 *	sockfd_lookup - Go from a file number to its socket slot
L
Linus Torvalds 已提交
460 461 462 463 464 465 466 467 468 469 470 471 472 473 474 475
 *	@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;

476 477
	file = fget(fd);
	if (!file) {
L
Linus Torvalds 已提交
478 479 480
		*err = -EBADF;
		return NULL;
	}
481

482 483
	sock = sock_from_file(file, err);
	if (!sock)
L
Linus Torvalds 已提交
484
		fput(file);
485 486
	return sock;
}
487
EXPORT_SYMBOL(sockfd_lookup);
L
Linus Torvalds 已提交
488

489 490
static struct socket *sockfd_lookup_light(int fd, int *err, int *fput_needed)
{
491
	struct fd f = fdget(fd);
492 493
	struct socket *sock;

494
	*err = -EBADF;
495 496 497 498
	if (f.file) {
		sock = sock_from_file(f.file, err);
		if (likely(sock)) {
			*fput_needed = f.flags;
499
			return sock;
500 501
		}
		fdput(f);
L
Linus Torvalds 已提交
502
	}
503
	return NULL;
L
Linus Torvalds 已提交
504 505
}

506 507 508 509 510 511
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
512
	len = security_inode_listsecurity(d_inode(dentry), buffer, size);
513 514 515 516 517 518 519 520 521 522 523 524 525 526 527 528 529 530 531 532 533
	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;
}

534
static int sockfs_setattr(struct dentry *dentry, struct iattr *iattr)
535 536 537
{
	int err = simple_setattr(dentry, iattr);

538
	if (!err && (iattr->ia_valid & ATTR_UID)) {
539 540 541 542 543 544 545 546
		struct socket *sock = SOCKET_I(d_inode(dentry));

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

	return err;
}

547 548
static const struct inode_operations sockfs_inode_ops = {
	.listxattr = sockfs_listxattr,
549
	.setattr = sockfs_setattr,
550 551
};

L
Linus Torvalds 已提交
552 553
/**
 *	sock_alloc	-	allocate a socket
554
 *
L
Linus Torvalds 已提交
555 556 557 558 559
 *	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
560
struct socket *sock_alloc(void)
L
Linus Torvalds 已提交
561
{
562 563
	struct inode *inode;
	struct socket *sock;
L
Linus Torvalds 已提交
564

565
	inode = new_inode_pseudo(sock_mnt->mnt_sb);
L
Linus Torvalds 已提交
566 567 568 569 570
	if (!inode)
		return NULL;

	sock = SOCKET_I(inode);

571
	kmemcheck_annotate_bitfield(sock, type);
572
	inode->i_ino = get_next_ino();
573
	inode->i_mode = S_IFSOCK | S_IRWXUGO;
574 575
	inode->i_uid = current_fsuid();
	inode->i_gid = current_fsgid();
576
	inode->i_op = &sockfs_inode_ops;
L
Linus Torvalds 已提交
577

578
	this_cpu_add(sockets_in_use, 1);
L
Linus Torvalds 已提交
579 580
	return sock;
}
T
Tom Herbert 已提交
581
EXPORT_SYMBOL(sock_alloc);
L
Linus Torvalds 已提交
582 583 584 585 586 587 588

/**
 *	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
589
 *	an inode not a file.
L
Linus Torvalds 已提交
590
 */
591

L
Linus Torvalds 已提交
592 593 594 595 596 597 598 599 600 601
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);
	}

602
	if (rcu_dereference_protected(sock->wq, 1)->fasync_list)
603
		pr_err("%s: fasync list not empty!\n", __func__);
L
Linus Torvalds 已提交
604

605
	this_cpu_sub(sockets_in_use, 1);
L
Linus Torvalds 已提交
606 607 608 609
	if (!sock->file) {
		iput(SOCK_INODE(sock));
		return;
	}
610
	sock->file = NULL;
L
Linus Torvalds 已提交
611
}
612
EXPORT_SYMBOL(sock_release);
L
Linus Torvalds 已提交
613

614
void __sock_tx_timestamp(__u16 tsflags, __u8 *tx_flags)
615
{
616 617
	u8 flags = *tx_flags;

618
	if (tsflags & SOF_TIMESTAMPING_TX_HARDWARE)
619 620
		flags |= SKBTX_HW_TSTAMP;

621
	if (tsflags & SOF_TIMESTAMPING_TX_SOFTWARE)
622 623
		flags |= SKBTX_SW_TSTAMP;

624
	if (tsflags & SOF_TIMESTAMPING_TX_SCHED)
625 626 627
		flags |= SKBTX_SCHED_TSTAMP;

	*tx_flags = flags;
628
}
629
EXPORT_SYMBOL(__sock_tx_timestamp);
630

631
static inline int sock_sendmsg_nosec(struct socket *sock, struct msghdr *msg)
L
Linus Torvalds 已提交
632
{
A
Al Viro 已提交
633
	int ret = sock->ops->sendmsg(sock, msg, msg_data_left(msg));
634 635
	BUG_ON(ret == -EIOCBQUEUED);
	return ret;
L
Linus Torvalds 已提交
636 637
}

638
int sock_sendmsg(struct socket *sock, struct msghdr *msg)
639
{
640
	int err = security_socket_sendmsg(sock, msg,
A
Al Viro 已提交
641
					  msg_data_left(msg));
642

643
	return err ?: sock_sendmsg_nosec(sock, msg);
644
}
645
EXPORT_SYMBOL(sock_sendmsg);
L
Linus Torvalds 已提交
646 647 648 649

int kernel_sendmsg(struct socket *sock, struct msghdr *msg,
		   struct kvec *vec, size_t num, size_t size)
{
650
	iov_iter_kvec(&msg->msg_iter, WRITE | ITER_KVEC, vec, num, size);
651
	return sock_sendmsg(sock, msg);
L
Linus Torvalds 已提交
652
}
653
EXPORT_SYMBOL(kernel_sendmsg);
L
Linus Torvalds 已提交
654

655 656 657 658 659 660 661 662 663 664
static bool skb_is_err_queue(const struct sk_buff *skb)
{
	/* pkt_type of skbs enqueued on the error queue are set to
	 * PACKET_OUTGOING in skb_set_err_queue(). This is only safe to do
	 * in recvmsg, since skbs received on a local socket will never
	 * have a pkt_type of PACKET_OUTGOING.
	 */
	return skb->pkt_type == PACKET_OUTGOING;
}

665 666 667 668 669 670
/*
 * 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)
{
671
	int need_software_tstamp = sock_flag(sk, SOCK_RCVTSTAMP);
672
	struct scm_timestamping tss;
673 674 675 676 677 678
	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. */
T
Thomas Gleixner 已提交
679
	if (need_software_tstamp && skb->tstamp == 0)
680 681 682 683 684 685 686 687 688
		__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 {
689 690
			struct timespec ts;
			skb_get_timestampns(skb, &ts);
691
			put_cmsg(msg, SOL_SOCKET, SCM_TIMESTAMPNS,
692
				 sizeof(ts), &ts);
693 694 695
		}
	}

696
	memset(&tss, 0, sizeof(tss));
697
	if ((sk->sk_tsflags & SOF_TIMESTAMPING_SOFTWARE) &&
698
	    ktime_to_timespec_cond(skb->tstamp, tss.ts + 0))
699
		empty = 0;
700
	if (shhwtstamps &&
701
	    (sk->sk_tsflags & SOF_TIMESTAMPING_RAW_HARDWARE) &&
702
	    ktime_to_timespec_cond(shhwtstamps->hwtstamp, tss.ts + 2))
703
		empty = 0;
704
	if (!empty) {
705
		put_cmsg(msg, SOL_SOCKET,
706
			 SCM_TIMESTAMPING, sizeof(tss), &tss);
707

708
		if (skb_is_err_queue(skb) && skb->len &&
709
		    SKB_EXT_ERR(skb)->opt_stats)
710 711 712
			put_cmsg(msg, SOL_SOCKET, SCM_TIMESTAMPING_OPT_STATS,
				 skb->len, skb->data);
	}
713
}
714 715
EXPORT_SYMBOL_GPL(__sock_recv_timestamp);

716 717 718 719 720 721 722 723 724 725 726 727 728 729 730 731
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
732 733
static inline void sock_recv_drops(struct msghdr *msg, struct sock *sk,
				   struct sk_buff *skb)
734
{
735
	if (sock_flag(sk, SOCK_RXQ_OVFL) && skb && SOCK_SKB_CB(skb)->dropcount)
736
		put_cmsg(msg, SOL_SOCKET, SO_RXQ_OVFL,
737
			sizeof(__u32), &SOCK_SKB_CB(skb)->dropcount);
738 739
}

740
void __sock_recv_ts_and_drops(struct msghdr *msg, struct sock *sk,
741 742 743 744 745
	struct sk_buff *skb)
{
	sock_recv_timestamp(msg, sk, skb);
	sock_recv_drops(msg, sk, skb);
}
746
EXPORT_SYMBOL_GPL(__sock_recv_ts_and_drops);
747

748
static inline int sock_recvmsg_nosec(struct socket *sock, struct msghdr *msg,
749
				     int flags)
L
Linus Torvalds 已提交
750
{
751
	return sock->ops->recvmsg(sock, msg, msg_data_left(msg), flags);
L
Linus Torvalds 已提交
752 753
}

754
int sock_recvmsg(struct socket *sock, struct msghdr *msg, int flags)
755
{
756
	int err = security_socket_recvmsg(sock, msg, msg_data_left(msg), flags);
757

758
	return err ?: sock_recvmsg_nosec(sock, msg, flags);
L
Linus Torvalds 已提交
759
}
760
EXPORT_SYMBOL(sock_recvmsg);
L
Linus Torvalds 已提交
761

762 763 764 765 766 767 768 769 770 771 772 773 774 775 776
/**
 * 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.
 */
777 778
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
779 780 781 782
{
	mm_segment_t oldfs = get_fs();
	int result;

783
	iov_iter_kvec(&msg->msg_iter, READ | ITER_KVEC, vec, num, size);
L
Linus Torvalds 已提交
784
	set_fs(KERNEL_DS);
785
	result = sock_recvmsg(sock, msg, flags);
L
Linus Torvalds 已提交
786 787 788
	set_fs(oldfs);
	return result;
}
789
EXPORT_SYMBOL(kernel_recvmsg);
L
Linus Torvalds 已提交
790

791 792
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
793 794 795 796
{
	struct socket *sock;
	int flags;

797 798
	sock = file->private_data;

799 800 801
	flags = (file->f_flags & O_NONBLOCK) ? MSG_DONTWAIT : 0;
	/* more is a combination of MSG_MORE and MSG_SENDPAGE_NOTLAST */
	flags |= more;
802

803
	return kernel_sendpage(sock, page, offset, size, flags);
804
}
L
Linus Torvalds 已提交
805

J
Jens Axboe 已提交
806
static ssize_t sock_splice_read(struct file *file, loff_t *ppos,
807
				struct pipe_inode_info *pipe, size_t len,
J
Jens Axboe 已提交
808 809 810 811
				unsigned int flags)
{
	struct socket *sock = file->private_data;

812 813 814
	if (unlikely(!sock->ops->splice_read))
		return -EINVAL;

J
Jens Axboe 已提交
815 816 817
	return sock->ops->splice_read(sock, ppos, pipe, len, flags);
}

818
static ssize_t sock_read_iter(struct kiocb *iocb, struct iov_iter *to)
819
{
820 821
	struct file *file = iocb->ki_filp;
	struct socket *sock = file->private_data;
822 823
	struct msghdr msg = {.msg_iter = *to,
			     .msg_iocb = iocb};
824
	ssize_t res;
825

826 827 828 829
	if (file->f_flags & O_NONBLOCK)
		msg.msg_flags = MSG_DONTWAIT;

	if (iocb->ki_pos != 0)
L
Linus Torvalds 已提交
830
		return -ESPIPE;
831

C
Christoph Hellwig 已提交
832
	if (!iov_iter_count(to))	/* Match SYS5 behaviour */
L
Linus Torvalds 已提交
833 834
		return 0;

835
	res = sock_recvmsg(sock, &msg, msg.msg_flags);
836 837
	*to = msg.msg_iter;
	return res;
L
Linus Torvalds 已提交
838 839
}

840
static ssize_t sock_write_iter(struct kiocb *iocb, struct iov_iter *from)
841
{
842 843
	struct file *file = iocb->ki_filp;
	struct socket *sock = file->private_data;
844 845
	struct msghdr msg = {.msg_iter = *from,
			     .msg_iocb = iocb};
846
	ssize_t res;
L
Linus Torvalds 已提交
847

848
	if (iocb->ki_pos != 0)
849
		return -ESPIPE;
850

851 852 853
	if (file->f_flags & O_NONBLOCK)
		msg.msg_flags = MSG_DONTWAIT;

854 855 856
	if (sock->type == SOCK_SEQPACKET)
		msg.msg_flags |= MSG_EOR;

857
	res = sock_sendmsg(sock, &msg);
858 859
	*from = msg.msg_iter;
	return res;
L
Linus Torvalds 已提交
860 861 862 863 864 865 866
}

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

A
Arjan van de Ven 已提交
867
static DEFINE_MUTEX(br_ioctl_mutex);
868
static int (*br_ioctl_hook) (struct net *, unsigned int cmd, void __user *arg);
L
Linus Torvalds 已提交
869

870
void brioctl_set(int (*hook) (struct net *, unsigned int, void __user *))
L
Linus Torvalds 已提交
871
{
A
Arjan van de Ven 已提交
872
	mutex_lock(&br_ioctl_mutex);
L
Linus Torvalds 已提交
873
	br_ioctl_hook = hook;
A
Arjan van de Ven 已提交
874
	mutex_unlock(&br_ioctl_mutex);
L
Linus Torvalds 已提交
875 876 877
}
EXPORT_SYMBOL(brioctl_set);

A
Arjan van de Ven 已提交
878
static DEFINE_MUTEX(vlan_ioctl_mutex);
879
static int (*vlan_ioctl_hook) (struct net *, void __user *arg);
L
Linus Torvalds 已提交
880

881
void vlan_ioctl_set(int (*hook) (struct net *, void __user *))
L
Linus Torvalds 已提交
882
{
A
Arjan van de Ven 已提交
883
	mutex_lock(&vlan_ioctl_mutex);
L
Linus Torvalds 已提交
884
	vlan_ioctl_hook = hook;
A
Arjan van de Ven 已提交
885
	mutex_unlock(&vlan_ioctl_mutex);
L
Linus Torvalds 已提交
886 887 888
}
EXPORT_SYMBOL(vlan_ioctl_set);

A
Arjan van de Ven 已提交
889
static DEFINE_MUTEX(dlci_ioctl_mutex);
890
static int (*dlci_ioctl_hook) (unsigned int, void __user *);
L
Linus Torvalds 已提交
891

892
void dlci_ioctl_set(int (*hook) (unsigned int, void __user *))
L
Linus Torvalds 已提交
893
{
A
Arjan van de Ven 已提交
894
	mutex_lock(&dlci_ioctl_mutex);
L
Linus Torvalds 已提交
895
	dlci_ioctl_hook = hook;
A
Arjan van de Ven 已提交
896
	mutex_unlock(&dlci_ioctl_mutex);
L
Linus Torvalds 已提交
897 898 899
}
EXPORT_SYMBOL(dlci_ioctl_set);

900 901 902 903 904 905 906 907 908 909 910 911 912 913 914 915 916 917
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
918 919 920 921 922
/*
 *	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.
 */

923 924 925 926 927
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
928 929 930
static long sock_ioctl(struct file *file, unsigned cmd, unsigned long arg)
{
	struct socket *sock;
931
	struct sock *sk;
L
Linus Torvalds 已提交
932 933
	void __user *argp = (void __user *)arg;
	int pid, err;
934
	struct net *net;
L
Linus Torvalds 已提交
935

936
	sock = file->private_data;
937
	sk = sock->sk;
938
	net = sock_net(sk);
L
Linus Torvalds 已提交
939
	if (cmd >= SIOCDEVPRIVATE && cmd <= (SIOCDEVPRIVATE + 15)) {
940
		err = dev_ioctl(net, cmd, argp);
L
Linus Torvalds 已提交
941
	} else
J
Johannes Berg 已提交
942
#ifdef CONFIG_WEXT_CORE
L
Linus Torvalds 已提交
943
	if (cmd >= SIOCIWFIRST && cmd <= SIOCIWLAST) {
944
		err = dev_ioctl(net, cmd, argp);
L
Linus Torvalds 已提交
945
	} else
J
Johannes Berg 已提交
946
#endif
947
		switch (cmd) {
L
Linus Torvalds 已提交
948 949 950 951 952
		case FIOSETOWN:
		case SIOCSPGRP:
			err = -EFAULT;
			if (get_user(pid, (int __user *)argp))
				break;
953 954
			f_setown(sock->file, pid, 1);
			err = 0;
L
Linus Torvalds 已提交
955 956 957
			break;
		case FIOGETOWN:
		case SIOCGPGRP:
958
			err = put_user(f_getown(sock->file),
959
				       (int __user *)argp);
L
Linus Torvalds 已提交
960 961 962 963 964 965 966 967 968
			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
969
			mutex_lock(&br_ioctl_mutex);
970
			if (br_ioctl_hook)
971
				err = br_ioctl_hook(net, cmd, argp);
A
Arjan van de Ven 已提交
972
			mutex_unlock(&br_ioctl_mutex);
L
Linus Torvalds 已提交
973 974 975 976 977 978 979
			break;
		case SIOCGIFVLAN:
		case SIOCSIFVLAN:
			err = -ENOPKG;
			if (!vlan_ioctl_hook)
				request_module("8021q");

A
Arjan van de Ven 已提交
980
			mutex_lock(&vlan_ioctl_mutex);
L
Linus Torvalds 已提交
981
			if (vlan_ioctl_hook)
982
				err = vlan_ioctl_hook(net, argp);
A
Arjan van de Ven 已提交
983
			mutex_unlock(&vlan_ioctl_mutex);
L
Linus Torvalds 已提交
984 985 986 987 988 989 990
			break;
		case SIOCADDDLCI:
		case SIOCDELDLCI:
			err = -ENOPKG;
			if (!dlci_ioctl_hook)
				request_module("dlci");

991 992
			mutex_lock(&dlci_ioctl_mutex);
			if (dlci_ioctl_hook)
L
Linus Torvalds 已提交
993
				err = dlci_ioctl_hook(cmd, argp);
994
			mutex_unlock(&dlci_ioctl_mutex);
L
Linus Torvalds 已提交
995
			break;
996 997 998 999 1000 1001 1002
		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
1003
		default:
1004
			err = sock_do_ioctl(net, sock, cmd, arg);
L
Linus Torvalds 已提交
1005
			break;
1006
		}
L
Linus Torvalds 已提交
1007 1008 1009 1010 1011 1012 1013
	return err;
}

int sock_create_lite(int family, int type, int protocol, struct socket **res)
{
	int err;
	struct socket *sock = NULL;
1014

L
Linus Torvalds 已提交
1015 1016 1017 1018 1019 1020 1021 1022 1023 1024 1025
	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
1026 1027 1028 1029
	err = security_socket_post_create(sock, family, type, protocol, 1);
	if (err)
		goto out_release;

L
Linus Torvalds 已提交
1030 1031 1032
out:
	*res = sock;
	return err;
V
Venkat Yekkirala 已提交
1033 1034 1035 1036
out_release:
	sock_release(sock);
	sock = NULL;
	goto out;
L
Linus Torvalds 已提交
1037
}
1038
EXPORT_SYMBOL(sock_create_lite);
L
Linus Torvalds 已提交
1039 1040

/* No kernel lock held - perfect */
1041
static unsigned int sock_poll(struct file *file, poll_table *wait)
L
Linus Torvalds 已提交
1042
{
1043
	unsigned int busy_flag = 0;
L
Linus Torvalds 已提交
1044 1045 1046
	struct socket *sock;

	/*
1047
	 *      We can't return errors to poll, so it's either yes or no.
L
Linus Torvalds 已提交
1048
	 */
1049
	sock = file->private_data;
1050

1051
	if (sk_can_busy_loop(sock->sk)) {
1052
		/* this socket can poll_ll so tell the system call */
1053
		busy_flag = POLL_BUSY_LOOP;
1054 1055

		/* once, only if requested by syscall */
1056 1057
		if (wait && (wait->_key & POLL_BUSY_LOOP))
			sk_busy_loop(sock->sk, 1);
1058 1059
	}

1060
	return busy_flag | sock->ops->poll(file, sock, wait);
L
Linus Torvalds 已提交
1061 1062
}

1063
static int sock_mmap(struct file *file, struct vm_area_struct *vma)
L
Linus Torvalds 已提交
1064
{
1065
	struct socket *sock = file->private_data;
L
Linus Torvalds 已提交
1066 1067 1068 1069

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

1070
static int sock_close(struct inode *inode, struct file *filp)
L
Linus Torvalds 已提交
1071 1072 1073 1074 1075 1076 1077 1078 1079 1080 1081 1082 1083
{
	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)
1084
 *	   or under socket lock
L
Linus Torvalds 已提交
1085 1086 1087 1088
 */

static int sock_fasync(int fd, struct file *filp, int on)
{
1089 1090
	struct socket *sock = filp->private_data;
	struct sock *sk = sock->sk;
1091
	struct socket_wq *wq;
L
Linus Torvalds 已提交
1092

1093
	if (sk == NULL)
L
Linus Torvalds 已提交
1094 1095 1096
		return -EINVAL;

	lock_sock(sk);
1097
	wq = rcu_dereference_protected(sock->wq, lockdep_sock_is_held(sk));
1098
	fasync_helper(fd, filp, on, &wq->fasync_list);
L
Linus Torvalds 已提交
1099

1100
	if (!wq->fasync_list)
1101 1102
		sock_reset_flag(sk, SOCK_FASYNC);
	else
E
Eric Dumazet 已提交
1103
		sock_set_flag(sk, SOCK_FASYNC);
L
Linus Torvalds 已提交
1104

1105
	release_sock(sk);
L
Linus Torvalds 已提交
1106 1107 1108
	return 0;
}

1109
/* This function may be called only under rcu_lock */
L
Linus Torvalds 已提交
1110

1111
int sock_wake_async(struct socket_wq *wq, int how, int band)
L
Linus Torvalds 已提交
1112
{
1113
	if (!wq || !wq->fasync_list)
L
Linus Torvalds 已提交
1114
		return -1;
1115

1116
	switch (how) {
1117
	case SOCK_WAKE_WAITD:
1118
		if (test_bit(SOCKWQ_ASYNC_WAITDATA, &wq->flags))
L
Linus Torvalds 已提交
1119 1120
			break;
		goto call_kill;
1121
	case SOCK_WAKE_SPACE:
1122
		if (!test_and_clear_bit(SOCKWQ_ASYNC_NOSPACE, &wq->flags))
L
Linus Torvalds 已提交
1123 1124
			break;
		/* fall through */
1125
	case SOCK_WAKE_IO:
1126
call_kill:
1127
		kill_fasync(&wq->fasync_list, SIGIO, band);
L
Linus Torvalds 已提交
1128
		break;
1129
	case SOCK_WAKE_URG:
1130
		kill_fasync(&wq->fasync_list, SIGURG, band);
L
Linus Torvalds 已提交
1131
	}
1132

L
Linus Torvalds 已提交
1133 1134
	return 0;
}
1135
EXPORT_SYMBOL(sock_wake_async);
L
Linus Torvalds 已提交
1136

P
Pavel Emelyanov 已提交
1137
int __sock_create(struct net *net, int family, int type, int protocol,
1138
			 struct socket **res, int kern)
L
Linus Torvalds 已提交
1139 1140 1141
{
	int err;
	struct socket *sock;
1142
	const struct net_proto_family *pf;
L
Linus Torvalds 已提交
1143 1144

	/*
1145
	 *      Check protocol is in range
L
Linus Torvalds 已提交
1146 1147 1148 1149 1150 1151 1152 1153 1154 1155 1156 1157
	 */
	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) {
1158 1159
		pr_info_once("%s uses obsolete (PF_INET,SOCK_PACKET)\n",
			     current->comm);
L
Linus Torvalds 已提交
1160 1161 1162 1163 1164 1165
		family = PF_PACKET;
	}

	err = security_socket_create(family, type, protocol, kern);
	if (err)
		return err;
1166

1167 1168 1169 1170 1171 1172 1173
	/*
	 *	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) {
1174
		net_warn_ratelimited("socket: no more sockets\n");
1175 1176 1177 1178 1179 1180
		return -ENFILE;	/* Not exactly a match, but its the
				   closest posix thing */
	}

	sock->type = type;

1181
#ifdef CONFIG_MODULES
1182 1183 1184
	/* 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
1185 1186 1187
	 * requested real, full-featured networking support upon configuration.
	 * Otherwise module support will break!
	 */
E
Eric Dumazet 已提交
1188
	if (rcu_access_pointer(net_families[family]) == NULL)
1189
		request_module("net-pf-%d", family);
L
Linus Torvalds 已提交
1190 1191
#endif

1192 1193 1194 1195 1196
	rcu_read_lock();
	pf = rcu_dereference(net_families[family]);
	err = -EAFNOSUPPORT;
	if (!pf)
		goto out_release;
L
Linus Torvalds 已提交
1197 1198 1199 1200 1201

	/*
	 * We will call the ->create function, that possibly is in a loadable
	 * module, so we have to bump that loadable module refcnt first.
	 */
1202
	if (!try_module_get(pf->owner))
L
Linus Torvalds 已提交
1203 1204
		goto out_release;

1205 1206 1207
	/* Now protected by module ref count */
	rcu_read_unlock();

1208
	err = pf->create(net, sock, protocol, kern);
1209
	if (err < 0)
L
Linus Torvalds 已提交
1210
		goto out_module_put;
1211

L
Linus Torvalds 已提交
1212 1213 1214 1215
	/*
	 * Now to bump the refcnt of the [loadable] module that owns this
	 * socket at sock_release time we decrement its refcnt.
	 */
1216 1217 1218
	if (!try_module_get(sock->ops->owner))
		goto out_module_busy;

L
Linus Torvalds 已提交
1219 1220 1221 1222
	/*
	 * Now that we're done with the ->create function, the [loadable]
	 * module can have its refcnt decremented
	 */
1223
	module_put(pf->owner);
V
Venkat Yekkirala 已提交
1224 1225
	err = security_socket_post_create(sock, family, type, protocol, kern);
	if (err)
1226
		goto out_sock_release;
1227
	*res = sock;
L
Linus Torvalds 已提交
1228

1229 1230 1231 1232
	return 0;

out_module_busy:
	err = -EAFNOSUPPORT;
L
Linus Torvalds 已提交
1233
out_module_put:
1234 1235 1236
	sock->ops = NULL;
	module_put(pf->owner);
out_sock_release:
L
Linus Torvalds 已提交
1237
	sock_release(sock);
1238 1239 1240 1241 1242
	return err;

out_release:
	rcu_read_unlock();
	goto out_sock_release;
L
Linus Torvalds 已提交
1243
}
P
Pavel Emelyanov 已提交
1244
EXPORT_SYMBOL(__sock_create);
L
Linus Torvalds 已提交
1245 1246 1247

int sock_create(int family, int type, int protocol, struct socket **res)
{
1248
	return __sock_create(current->nsproxy->net_ns, family, type, protocol, res, 0);
L
Linus Torvalds 已提交
1249
}
1250
EXPORT_SYMBOL(sock_create);
L
Linus Torvalds 已提交
1251

1252
int sock_create_kern(struct net *net, int family, int type, int protocol, struct socket **res)
L
Linus Torvalds 已提交
1253
{
1254
	return __sock_create(net, family, type, protocol, res, 1);
L
Linus Torvalds 已提交
1255
}
1256
EXPORT_SYMBOL(sock_create_kern);
L
Linus Torvalds 已提交
1257

1258
SYSCALL_DEFINE3(socket, int, family, int, type, int, protocol)
L
Linus Torvalds 已提交
1259 1260 1261
{
	int retval;
	struct socket *sock;
1262 1263
	int flags;

1264 1265 1266 1267 1268 1269
	/* 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);

1270
	flags = type & ~SOCK_TYPE_MASK;
1271
	if (flags & ~(SOCK_CLOEXEC | SOCK_NONBLOCK))
1272 1273
		return -EINVAL;
	type &= SOCK_TYPE_MASK;
L
Linus Torvalds 已提交
1274

U
Ulrich Drepper 已提交
1275 1276 1277
	if (SOCK_NONBLOCK != O_NONBLOCK && (flags & SOCK_NONBLOCK))
		flags = (flags & ~SOCK_NONBLOCK) | O_NONBLOCK;

L
Linus Torvalds 已提交
1278 1279 1280 1281
	retval = sock_create(family, type, protocol, &sock);
	if (retval < 0)
		goto out;

1282
	retval = sock_map_fd(sock, flags & (O_CLOEXEC | O_NONBLOCK));
L
Linus Torvalds 已提交
1283 1284 1285 1286 1287 1288 1289 1290 1291 1292 1293 1294 1295 1296 1297 1298
	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.
 */

1299 1300
SYSCALL_DEFINE4(socketpair, int, family, int, type, int, protocol,
		int __user *, usockvec)
L
Linus Torvalds 已提交
1301 1302 1303
{
	struct socket *sock1, *sock2;
	int fd1, fd2, err;
A
Al Viro 已提交
1304
	struct file *newfile1, *newfile2;
1305 1306 1307
	int flags;

	flags = type & ~SOCK_TYPE_MASK;
1308
	if (flags & ~(SOCK_CLOEXEC | SOCK_NONBLOCK))
1309 1310
		return -EINVAL;
	type &= SOCK_TYPE_MASK;
L
Linus Torvalds 已提交
1311

U
Ulrich Drepper 已提交
1312 1313 1314
	if (SOCK_NONBLOCK != O_NONBLOCK && (flags & SOCK_NONBLOCK))
		flags = (flags & ~SOCK_NONBLOCK) | O_NONBLOCK;

L
Linus Torvalds 已提交
1315 1316 1317 1318 1319 1320 1321 1322 1323 1324 1325 1326 1327 1328
	/*
	 * 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);
1329
	if (err < 0)
L
Linus Torvalds 已提交
1330 1331
		goto out_release_both;

1332
	fd1 = get_unused_fd_flags(flags);
1333 1334
	if (unlikely(fd1 < 0)) {
		err = fd1;
A
Al Viro 已提交
1335
		goto out_release_both;
1336
	}
1337

1338
	fd2 = get_unused_fd_flags(flags);
1339 1340
	if (unlikely(fd2 < 0)) {
		err = fd2;
1341
		goto out_put_unused_1;
1342 1343
	}

1344
	newfile1 = sock_alloc_file(sock1, flags, NULL);
1345
	if (IS_ERR(newfile1)) {
1346
		err = PTR_ERR(newfile1);
1347
		goto out_put_unused_both;
1348 1349
	}

1350
	newfile2 = sock_alloc_file(sock2, flags, NULL);
1351 1352
	if (IS_ERR(newfile2)) {
		err = PTR_ERR(newfile2);
1353
		goto out_fput_1;
A
Al Viro 已提交
1354 1355
	}

1356 1357 1358 1359 1360 1361 1362 1363
	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
1364
	audit_fd_pair(fd1, fd2);
1365

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

1372
	return 0;
L
Linus Torvalds 已提交
1373

1374 1375 1376 1377 1378 1379 1380 1381 1382 1383 1384 1385 1386
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
1387

1388 1389 1390 1391
out_put_unused_both:
	put_unused_fd(fd2);
out_put_unused_1:
	put_unused_fd(fd1);
L
Linus Torvalds 已提交
1392
out_release_both:
1393
	sock_release(sock2);
L
Linus Torvalds 已提交
1394
out_release_1:
1395
	sock_release(sock1);
L
Linus Torvalds 已提交
1396 1397 1398 1399 1400 1401 1402 1403 1404 1405 1406 1407
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).
 */

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

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

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

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

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

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

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

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

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

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

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

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

	err = -ENFILE;
1489 1490
	newsock = sock_alloc();
	if (!newsock)
L
Linus Torvalds 已提交
1491 1492 1493 1494 1495 1496 1497 1498 1499 1500 1501
		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);

1502
	newfd = get_unused_fd_flags(flags);
1503 1504
	if (unlikely(newfd < 0)) {
		err = newfd;
1505 1506
		sock_release(newsock);
		goto out_put;
1507
	}
1508
	newfile = sock_alloc_file(newsock, flags, sock->sk->sk_prot_creator->name);
1509
	if (IS_ERR(newfile)) {
1510 1511 1512 1513 1514
		err = PTR_ERR(newfile);
		put_unused_fd(newfd);
		sock_release(newsock);
		goto out_put;
	}
1515

1516 1517
	err = security_socket_accept(sock, newsock);
	if (err)
1518
		goto out_fd;
1519

1520
	err = sock->ops->accept(sock, newsock, sock->file->f_flags, false);
L
Linus Torvalds 已提交
1521
	if (err < 0)
1522
		goto out_fd;
L
Linus Torvalds 已提交
1523 1524

	if (upeer_sockaddr) {
1525
		if (newsock->ops->getname(newsock, (struct sockaddr *)&address,
1526
					  &len, 2) < 0) {
L
Linus Torvalds 已提交
1527
			err = -ECONNABORTED;
1528
			goto out_fd;
L
Linus Torvalds 已提交
1529
		}
1530
		err = move_addr_to_user(&address,
1531
					len, upeer_sockaddr, upeer_addrlen);
L
Linus Torvalds 已提交
1532
		if (err < 0)
1533
			goto out_fd;
L
Linus Torvalds 已提交
1534 1535 1536 1537
	}

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

1538 1539
	fd_install(newfd, newfile);
	err = newfd;
L
Linus Torvalds 已提交
1540 1541

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

1551 1552
SYSCALL_DEFINE3(accept, int, fd, struct sockaddr __user *, upeer_sockaddr,
		int __user *, upeer_addrlen)
U
Ulrich Drepper 已提交
1553
{
U
Ulrich Drepper 已提交
1554
	return sys_accept4(fd, upeer_sockaddr, upeer_addrlen, 0);
U
Ulrich Drepper 已提交
1555 1556
}

L
Linus Torvalds 已提交
1557 1558 1559 1560 1561 1562 1563 1564 1565 1566 1567 1568
/*
 *	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.
 */

1569 1570
SYSCALL_DEFINE3(connect, int, fd, struct sockaddr __user *, uservaddr,
		int, addrlen)
L
Linus Torvalds 已提交
1571 1572
{
	struct socket *sock;
1573
	struct sockaddr_storage address;
1574
	int err, fput_needed;
L
Linus Torvalds 已提交
1575

1576
	sock = sockfd_lookup_light(fd, &err, &fput_needed);
L
Linus Torvalds 已提交
1577 1578
	if (!sock)
		goto out;
1579
	err = move_addr_to_kernel(uservaddr, addrlen, &address);
L
Linus Torvalds 已提交
1580 1581 1582
	if (err < 0)
		goto out_put;

1583
	err =
1584
	    security_socket_connect(sock, (struct sockaddr *)&address, addrlen);
L
Linus Torvalds 已提交
1585 1586 1587
	if (err)
		goto out_put;

1588
	err = sock->ops->connect(sock, (struct sockaddr *)&address, addrlen,
L
Linus Torvalds 已提交
1589 1590
				 sock->file->f_flags);
out_put:
1591
	fput_light(sock->file, fput_needed);
L
Linus Torvalds 已提交
1592 1593 1594 1595 1596 1597 1598 1599 1600
out:
	return err;
}

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

1601 1602
SYSCALL_DEFINE3(getsockname, int, fd, struct sockaddr __user *, usockaddr,
		int __user *, usockaddr_len)
L
Linus Torvalds 已提交
1603 1604
{
	struct socket *sock;
1605
	struct sockaddr_storage address;
1606
	int len, err, fput_needed;
1607

1608
	sock = sockfd_lookup_light(fd, &err, &fput_needed);
L
Linus Torvalds 已提交
1609 1610 1611 1612 1613 1614 1615
	if (!sock)
		goto out;

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

1616
	err = sock->ops->getname(sock, (struct sockaddr *)&address, &len, 0);
L
Linus Torvalds 已提交
1617 1618
	if (err)
		goto out_put;
1619
	err = move_addr_to_user(&address, len, usockaddr, usockaddr_len);
L
Linus Torvalds 已提交
1620 1621

out_put:
1622
	fput_light(sock->file, fput_needed);
L
Linus Torvalds 已提交
1623 1624 1625 1626 1627 1628 1629 1630 1631
out:
	return err;
}

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

1632 1633
SYSCALL_DEFINE3(getpeername, int, fd, struct sockaddr __user *, usockaddr,
		int __user *, usockaddr_len)
L
Linus Torvalds 已提交
1634 1635
{
	struct socket *sock;
1636
	struct sockaddr_storage address;
1637
	int len, err, fput_needed;
L
Linus Torvalds 已提交
1638

1639 1640
	sock = sockfd_lookup_light(fd, &err, &fput_needed);
	if (sock != NULL) {
L
Linus Torvalds 已提交
1641 1642
		err = security_socket_getpeername(sock);
		if (err) {
1643
			fput_light(sock->file, fput_needed);
L
Linus Torvalds 已提交
1644 1645 1646
			return err;
		}

1647
		err =
1648
		    sock->ops->getname(sock, (struct sockaddr *)&address, &len,
1649
				       1);
L
Linus Torvalds 已提交
1650
		if (!err)
1651
			err = move_addr_to_user(&address, len, usockaddr,
1652
						usockaddr_len);
1653
		fput_light(sock->file, fput_needed);
L
Linus Torvalds 已提交
1654 1655 1656 1657 1658 1659 1660 1661 1662 1663
	}
	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.
 */

1664
SYSCALL_DEFINE6(sendto, int, fd, void __user *, buff, size_t, len,
1665
		unsigned int, flags, struct sockaddr __user *, addr,
1666
		int, addr_len)
L
Linus Torvalds 已提交
1667 1668
{
	struct socket *sock;
1669
	struct sockaddr_storage address;
L
Linus Torvalds 已提交
1670 1671 1672
	int err;
	struct msghdr msg;
	struct iovec iov;
1673 1674
	int fput_needed;

1675 1676 1677
	err = import_single_range(WRITE, buff, len, &iov, &msg.msg_iter);
	if (unlikely(err))
		return err;
1678 1679
	sock = sockfd_lookup_light(fd, &err, &fput_needed);
	if (!sock)
1680
		goto out;
1681

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

1698
out_put:
1699
	fput_light(sock->file, fput_needed);
1700
out:
L
Linus Torvalds 已提交
1701 1702 1703 1704
	return err;
}

/*
1705
 *	Send a datagram down a socket.
L
Linus Torvalds 已提交
1706 1707
 */

1708
SYSCALL_DEFINE4(send, int, fd, void __user *, buff, size_t, len,
1709
		unsigned int, flags)
L
Linus Torvalds 已提交
1710 1711 1712 1713 1714
{
	return sys_sendto(fd, buff, len, flags, NULL, 0);
}

/*
1715
 *	Receive a frame from the socket and optionally record the address of the
L
Linus Torvalds 已提交
1716 1717 1718 1719
 *	sender. We verify the buffers are writable and if needed move the
 *	sender address from kernel to user space.
 */

1720
SYSCALL_DEFINE6(recvfrom, int, fd, void __user *, ubuf, size_t, size,
1721
		unsigned int, flags, struct sockaddr __user *, addr,
1722
		int __user *, addr_len)
L
Linus Torvalds 已提交
1723 1724 1725 1726
{
	struct socket *sock;
	struct iovec iov;
	struct msghdr msg;
1727
	struct sockaddr_storage address;
1728
	int err, err2;
1729 1730
	int fput_needed;

1731 1732 1733
	err = import_single_range(READ, ubuf, size, &iov, &msg.msg_iter);
	if (unlikely(err))
		return err;
1734
	sock = sockfd_lookup_light(fd, &err, &fput_needed);
L
Linus Torvalds 已提交
1735
	if (!sock)
1736
		goto out;
L
Linus Torvalds 已提交
1737

1738 1739
	msg.msg_control = NULL;
	msg.msg_controllen = 0;
1740 1741 1742 1743
	/* 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;
1744
	msg.msg_iocb = NULL;
1745
	msg.msg_flags = 0;
L
Linus Torvalds 已提交
1746 1747
	if (sock->file->f_flags & O_NONBLOCK)
		flags |= MSG_DONTWAIT;
1748
	err = sock_recvmsg(sock, &msg, flags);
L
Linus Torvalds 已提交
1749

1750
	if (err >= 0 && addr != NULL) {
1751
		err2 = move_addr_to_user(&address,
1752
					 msg.msg_namelen, addr, addr_len);
1753 1754
		if (err2 < 0)
			err = err2;
L
Linus Torvalds 已提交
1755
	}
1756 1757

	fput_light(sock->file, fput_needed);
1758
out:
L
Linus Torvalds 已提交
1759 1760 1761 1762
	return err;
}

/*
1763
 *	Receive a datagram from a socket.
L
Linus Torvalds 已提交
1764 1765
 */

1766 1767
SYSCALL_DEFINE4(recv, int, fd, void __user *, ubuf, size_t, size,
		unsigned int, flags)
L
Linus Torvalds 已提交
1768 1769 1770 1771 1772 1773 1774 1775 1776
{
	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.
 */

1777 1778
SYSCALL_DEFINE5(setsockopt, int, fd, int, level, int, optname,
		char __user *, optval, int, optlen)
L
Linus Torvalds 已提交
1779
{
1780
	int err, fput_needed;
L
Linus Torvalds 已提交
1781 1782 1783 1784
	struct socket *sock;

	if (optlen < 0)
		return -EINVAL;
1785 1786 1787 1788

	sock = sockfd_lookup_light(fd, &err, &fput_needed);
	if (sock != NULL) {
		err = security_socket_setsockopt(sock, level, optname);
1789 1790
		if (err)
			goto out_put;
L
Linus Torvalds 已提交
1791 1792

		if (level == SOL_SOCKET)
1793 1794 1795
			err =
			    sock_setsockopt(sock, level, optname, optval,
					    optlen);
L
Linus Torvalds 已提交
1796
		else
1797 1798 1799
			err =
			    sock->ops->setsockopt(sock, level, optname, optval,
						  optlen);
1800 1801
out_put:
		fput_light(sock->file, fput_needed);
L
Linus Torvalds 已提交
1802 1803 1804 1805 1806 1807 1808 1809 1810
	}
	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.
 */

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

1817 1818
	sock = sockfd_lookup_light(fd, &err, &fput_needed);
	if (sock != NULL) {
1819 1820 1821
		err = security_socket_getsockopt(sock, level, optname);
		if (err)
			goto out_put;
L
Linus Torvalds 已提交
1822 1823

		if (level == SOL_SOCKET)
1824 1825 1826
			err =
			    sock_getsockopt(sock, level, optname, optval,
					    optlen);
L
Linus Torvalds 已提交
1827
		else
1828 1829 1830
			err =
			    sock->ops->getsockopt(sock, level, optname, optval,
						  optlen);
1831 1832
out_put:
		fput_light(sock->file, fput_needed);
L
Linus Torvalds 已提交
1833 1834 1835 1836 1837 1838 1839 1840
	}
	return err;
}

/*
 *	Shutdown a socket.
 */

1841
SYSCALL_DEFINE2(shutdown, int, fd, int, how)
L
Linus Torvalds 已提交
1842
{
1843
	int err, fput_needed;
L
Linus Torvalds 已提交
1844 1845
	struct socket *sock;

1846 1847
	sock = sockfd_lookup_light(fd, &err, &fput_needed);
	if (sock != NULL) {
L
Linus Torvalds 已提交
1848
		err = security_socket_shutdown(sock, how);
1849 1850 1851
		if (!err)
			err = sock->ops->shutdown(sock, how);
		fput_light(sock->file, fput_needed);
L
Linus Torvalds 已提交
1852 1853 1854 1855
	}
	return err;
}

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

1863 1864 1865 1866 1867
struct used_address {
	struct sockaddr_storage name;
	unsigned int name_len;
};

1868 1869 1870 1871
static int copy_msghdr_from_user(struct msghdr *kmsg,
				 struct user_msghdr __user *umsg,
				 struct sockaddr __user **save_addr,
				 struct iovec **iov)
1872
{
1873 1874
	struct sockaddr __user *uaddr;
	struct iovec __user *uiov;
A
Al Viro 已提交
1875
	size_t nr_segs;
1876 1877 1878 1879 1880 1881
	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
1882
	    __get_user(nr_segs, &umsg->msg_iovlen) ||
1883 1884 1885
	    __get_user(kmsg->msg_control, &umsg->msg_control) ||
	    __get_user(kmsg->msg_controllen, &umsg->msg_controllen) ||
	    __get_user(kmsg->msg_flags, &umsg->msg_flags))
1886
		return -EFAULT;
1887

1888
	if (!uaddr)
1889 1890
		kmsg->msg_namelen = 0;

1891 1892 1893
	if (kmsg->msg_namelen < 0)
		return -EINVAL;

1894
	if (kmsg->msg_namelen > sizeof(struct sockaddr_storage))
1895
		kmsg->msg_namelen = sizeof(struct sockaddr_storage);
1896 1897 1898 1899 1900 1901 1902 1903 1904 1905 1906 1907 1908 1909 1910 1911

	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
1912
	if (nr_segs > UIO_MAXIOV)
1913 1914
		return -EMSGSIZE;

1915 1916
	kmsg->msg_iocb = NULL;

1917 1918
	return import_iovec(save_addr ? READ : WRITE, uiov, nr_segs,
			    UIO_FASTIOV, iov, &kmsg->msg_iter);
1919 1920
}

1921
static int ___sys_sendmsg(struct socket *sock, struct user_msghdr __user *msg,
1922
			 struct msghdr *msg_sys, unsigned int flags,
1923 1924
			 struct used_address *used_address,
			 unsigned int allowed_msghdr_flags)
L
Linus Torvalds 已提交
1925
{
1926 1927
	struct compat_msghdr __user *msg_compat =
	    (struct compat_msghdr __user *)msg;
1928
	struct sockaddr_storage address;
L
Linus Torvalds 已提交
1929
	struct iovec iovstack[UIO_FASTIOV], *iov = iovstack;
1930
	unsigned char ctl[sizeof(struct cmsghdr) + 20]
1931
				__aligned(sizeof(__kernel_size_t));
1932
	/* 20 is size of ipv6_pktinfo */
L
Linus Torvalds 已提交
1933
	unsigned char *ctl_buf = ctl;
1934
	int ctl_len;
1935
	ssize_t err;
1936

1937
	msg_sys->msg_name = &address;
L
Linus Torvalds 已提交
1938

1939
	if (MSG_CMSG_COMPAT & flags)
1940
		err = get_compat_msghdr(msg_sys, msg_compat, NULL, &iov);
1941
	else
1942
		err = copy_msghdr_from_user(msg_sys, msg, NULL, &iov);
1943
	if (err < 0)
1944
		return err;
L
Linus Torvalds 已提交
1945 1946 1947

	err = -ENOBUFS;

1948
	if (msg_sys->msg_controllen > INT_MAX)
L
Linus Torvalds 已提交
1949
		goto out_freeiov;
1950
	flags |= (msg_sys->msg_flags & allowed_msghdr_flags);
1951
	ctl_len = msg_sys->msg_controllen;
L
Linus Torvalds 已提交
1952
	if ((MSG_CMSG_COMPAT & flags) && ctl_len) {
1953
		err =
1954
		    cmsghdr_from_user_compat_to_kern(msg_sys, sock->sk, ctl,
1955
						     sizeof(ctl));
L
Linus Torvalds 已提交
1956 1957
		if (err)
			goto out_freeiov;
1958 1959
		ctl_buf = msg_sys->msg_control;
		ctl_len = msg_sys->msg_controllen;
L
Linus Torvalds 已提交
1960
	} else if (ctl_len) {
1961 1962
		BUILD_BUG_ON(sizeof(struct cmsghdr) !=
			     CMSG_ALIGN(sizeof(struct cmsghdr)));
1963
		if (ctl_len > sizeof(ctl)) {
L
Linus Torvalds 已提交
1964
			ctl_buf = sock_kmalloc(sock->sk, ctl_len, GFP_KERNEL);
1965
			if (ctl_buf == NULL)
L
Linus Torvalds 已提交
1966 1967 1968 1969
				goto out_freeiov;
		}
		err = -EFAULT;
		/*
1970
		 * Careful! Before this, msg_sys->msg_control contains a user pointer.
L
Linus Torvalds 已提交
1971 1972 1973
		 * Afterwards, it will be a kernel pointer. Thus the compiler-assisted
		 * checking falls down on this.
		 */
1974
		if (copy_from_user(ctl_buf,
1975
				   (void __user __force *)msg_sys->msg_control,
1976
				   ctl_len))
L
Linus Torvalds 已提交
1977
			goto out_freectl;
1978
		msg_sys->msg_control = ctl_buf;
L
Linus Torvalds 已提交
1979
	}
1980
	msg_sys->msg_flags = flags;
L
Linus Torvalds 已提交
1981 1982

	if (sock->file->f_flags & O_NONBLOCK)
1983
		msg_sys->msg_flags |= MSG_DONTWAIT;
1984 1985 1986 1987 1988 1989
	/*
	 * 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.
	 */
1990 1991 1992
	if (used_address && msg_sys->msg_name &&
	    used_address->name_len == msg_sys->msg_namelen &&
	    !memcmp(&used_address->name, msg_sys->msg_name,
1993
		    used_address->name_len)) {
1994
		err = sock_sendmsg_nosec(sock, msg_sys);
1995 1996
		goto out_freectl;
	}
1997
	err = sock_sendmsg(sock, msg_sys);
1998 1999 2000 2001 2002 2003
	/*
	 * 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;
2004 2005 2006
		if (msg_sys->msg_name)
			memcpy(&used_address->name, msg_sys->msg_name,
			       used_address->name_len);
2007
	}
L
Linus Torvalds 已提交
2008 2009

out_freectl:
2010
	if (ctl_buf != ctl)
L
Linus Torvalds 已提交
2011 2012
		sock_kfree_s(sock->sk, ctl_buf, ctl_len);
out_freeiov:
2013
	kfree(iov);
2014 2015 2016 2017 2018 2019 2020
	return err;
}

/*
 *	BSD sendmsg interface
 */

2021
long __sys_sendmsg(int fd, struct user_msghdr __user *msg, unsigned flags)
2022 2023 2024
{
	int fput_needed, err;
	struct msghdr msg_sys;
2025 2026 2027
	struct socket *sock;

	sock = sockfd_lookup_light(fd, &err, &fput_needed);
2028 2029 2030
	if (!sock)
		goto out;

2031
	err = ___sys_sendmsg(sock, msg, &msg_sys, flags, NULL, 0);
2032

2033
	fput_light(sock->file, fput_needed);
2034
out:
L
Linus Torvalds 已提交
2035 2036 2037
	return err;
}

2038
SYSCALL_DEFINE3(sendmsg, int, fd, struct user_msghdr __user *, msg, unsigned int, flags)
2039 2040 2041 2042 2043 2044
{
	if (flags & MSG_CMSG_COMPAT)
		return -EINVAL;
	return __sys_sendmsg(fd, msg, flags);
}

2045 2046 2047 2048 2049 2050 2051 2052 2053 2054 2055 2056
/*
 *	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;
2057
	struct used_address used_address;
T
Tom Herbert 已提交
2058
	unsigned int oflags = flags;
2059

2060 2061
	if (vlen > UIO_MAXIOV)
		vlen = UIO_MAXIOV;
2062 2063 2064 2065 2066 2067 2068

	datagrams = 0;

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

2069
	used_address.name_len = UINT_MAX;
2070 2071
	entry = mmsg;
	compat_entry = (struct compat_mmsghdr __user *)mmsg;
2072
	err = 0;
T
Tom Herbert 已提交
2073
	flags |= MSG_BATCH;
2074 2075

	while (datagrams < vlen) {
T
Tom Herbert 已提交
2076 2077 2078
		if (datagrams == vlen - 1)
			flags = oflags;

2079
		if (MSG_CMSG_COMPAT & flags) {
2080
			err = ___sys_sendmsg(sock, (struct user_msghdr __user *)compat_entry,
2081
					     &msg_sys, flags, &used_address, MSG_EOR);
2082 2083 2084 2085 2086
			if (err < 0)
				break;
			err = __put_user(err, &compat_entry->msg_len);
			++compat_entry;
		} else {
2087
			err = ___sys_sendmsg(sock,
2088
					     (struct user_msghdr __user *)entry,
2089
					     &msg_sys, flags, &used_address, MSG_EOR);
2090 2091 2092 2093 2094 2095 2096 2097 2098
			if (err < 0)
				break;
			err = put_user(err, &entry->msg_len);
			++entry;
		}

		if (err)
			break;
		++datagrams;
2099 2100
		if (msg_data_left(&msg_sys))
			break;
2101
		cond_resched();
2102 2103 2104 2105
	}

	fput_light(sock->file, fput_needed);

2106 2107
	/* We only return an error if no datagrams were able to be sent */
	if (datagrams != 0)
2108 2109 2110 2111 2112 2113 2114 2115
		return datagrams;

	return err;
}

SYSCALL_DEFINE4(sendmmsg, int, fd, struct mmsghdr __user *, mmsg,
		unsigned int, vlen, unsigned int, flags)
{
2116 2117
	if (flags & MSG_CMSG_COMPAT)
		return -EINVAL;
2118 2119 2120
	return __sys_sendmmsg(fd, mmsg, vlen, flags);
}

2121
static int ___sys_recvmsg(struct socket *sock, struct user_msghdr __user *msg,
2122
			 struct msghdr *msg_sys, unsigned int flags, int nosec)
L
Linus Torvalds 已提交
2123
{
2124 2125
	struct compat_msghdr __user *msg_compat =
	    (struct compat_msghdr __user *)msg;
L
Linus Torvalds 已提交
2126
	struct iovec iovstack[UIO_FASTIOV];
2127
	struct iovec *iov = iovstack;
L
Linus Torvalds 已提交
2128
	unsigned long cmsg_ptr;
2129
	int len;
2130
	ssize_t err;
L
Linus Torvalds 已提交
2131 2132

	/* kernel mode address */
2133
	struct sockaddr_storage addr;
L
Linus Torvalds 已提交
2134 2135 2136

	/* user mode address pointers */
	struct sockaddr __user *uaddr;
2137
	int __user *uaddr_len = COMPAT_NAMELEN(msg);
2138

2139
	msg_sys->msg_name = &addr;
L
Linus Torvalds 已提交
2140

2141
	if (MSG_CMSG_COMPAT & flags)
2142
		err = get_compat_msghdr(msg_sys, msg_compat, &uaddr, &iov);
2143
	else
2144
		err = copy_msghdr_from_user(msg_sys, msg, &uaddr, &iov);
L
Linus Torvalds 已提交
2145
	if (err < 0)
2146
		return err;
L
Linus Torvalds 已提交
2147

2148 2149
	cmsg_ptr = (unsigned long)msg_sys->msg_control;
	msg_sys->msg_flags = flags & (MSG_CMSG_CLOEXEC|MSG_CMSG_COMPAT);
2150

2151 2152 2153
	/* We assume all kernel code knows the size of sockaddr_storage */
	msg_sys->msg_namelen = 0;

L
Linus Torvalds 已提交
2154 2155
	if (sock->file->f_flags & O_NONBLOCK)
		flags |= MSG_DONTWAIT;
2156
	err = (nosec ? sock_recvmsg_nosec : sock_recvmsg)(sock, msg_sys, flags);
L
Linus Torvalds 已提交
2157 2158 2159 2160 2161
	if (err < 0)
		goto out_freeiov;
	len = err;

	if (uaddr != NULL) {
2162
		err = move_addr_to_user(&addr,
2163
					msg_sys->msg_namelen, uaddr,
2164
					uaddr_len);
L
Linus Torvalds 已提交
2165 2166 2167
		if (err < 0)
			goto out_freeiov;
	}
2168
	err = __put_user((msg_sys->msg_flags & ~MSG_CMSG_COMPAT),
2169
			 COMPAT_FLAGS(msg));
L
Linus Torvalds 已提交
2170 2171 2172
	if (err)
		goto out_freeiov;
	if (MSG_CMSG_COMPAT & flags)
2173
		err = __put_user((unsigned long)msg_sys->msg_control - cmsg_ptr,
L
Linus Torvalds 已提交
2174 2175
				 &msg_compat->msg_controllen);
	else
2176
		err = __put_user((unsigned long)msg_sys->msg_control - cmsg_ptr,
L
Linus Torvalds 已提交
2177 2178 2179 2180 2181 2182
				 &msg->msg_controllen);
	if (err)
		goto out_freeiov;
	err = len;

out_freeiov:
2183
	kfree(iov);
2184 2185 2186 2187 2188 2189 2190
	return err;
}

/*
 *	BSD recvmsg interface
 */

2191
long __sys_recvmsg(int fd, struct user_msghdr __user *msg, unsigned flags)
2192 2193 2194
{
	int fput_needed, err;
	struct msghdr msg_sys;
2195 2196 2197
	struct socket *sock;

	sock = sockfd_lookup_light(fd, &err, &fput_needed);
2198 2199 2200
	if (!sock)
		goto out;

2201
	err = ___sys_recvmsg(sock, msg, &msg_sys, flags, 0);
2202

2203
	fput_light(sock->file, fput_needed);
L
Linus Torvalds 已提交
2204 2205 2206 2207
out:
	return err;
}

2208
SYSCALL_DEFINE3(recvmsg, int, fd, struct user_msghdr __user *, msg,
2209 2210 2211 2212 2213 2214 2215
		unsigned int, flags)
{
	if (flags & MSG_CMSG_COMPAT)
		return -EINVAL;
	return __sys_recvmsg(fd, msg, flags);
}

2216 2217 2218 2219 2220 2221 2222 2223 2224 2225
/*
 *     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;
2226
	struct compat_mmsghdr __user *compat_entry;
2227
	struct msghdr msg_sys;
2228 2229
	struct timespec64 end_time;
	struct timespec64 timeout64;
2230 2231 2232 2233 2234 2235 2236 2237 2238 2239 2240 2241 2242

	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);
2243 2244
	if (err) {
		datagrams = err;
2245
		goto out_put;
2246
	}
2247 2248

	entry = mmsg;
2249
	compat_entry = (struct compat_mmsghdr __user *)mmsg;
2250 2251 2252 2253 2254

	while (datagrams < vlen) {
		/*
		 * No need to ask LSM for more than the first datagram.
		 */
2255
		if (MSG_CMSG_COMPAT & flags) {
2256
			err = ___sys_recvmsg(sock, (struct user_msghdr __user *)compat_entry,
2257 2258
					     &msg_sys, flags & ~MSG_WAITFORONE,
					     datagrams);
2259 2260 2261 2262 2263
			if (err < 0)
				break;
			err = __put_user(err, &compat_entry->msg_len);
			++compat_entry;
		} else {
2264
			err = ___sys_recvmsg(sock,
2265
					     (struct user_msghdr __user *)entry,
2266 2267
					     &msg_sys, flags & ~MSG_WAITFORONE,
					     datagrams);
2268 2269 2270 2271 2272 2273
			if (err < 0)
				break;
			err = put_user(err, &entry->msg_len);
			++entry;
		}

2274 2275 2276 2277
		if (err)
			break;
		++datagrams;

2278 2279 2280 2281
		/* MSG_WAITFORONE turns on MSG_DONTWAIT after one packet */
		if (flags & MSG_WAITFORONE)
			flags |= MSG_DONTWAIT;

2282
		if (timeout) {
2283 2284 2285
			ktime_get_ts64(&timeout64);
			*timeout = timespec64_to_timespec(
					timespec64_sub(end_time, timeout64));
2286 2287 2288 2289 2290 2291 2292 2293 2294 2295 2296 2297 2298
			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;
2299
		cond_resched();
2300 2301 2302
	}

	if (err == 0)
2303 2304 2305 2306 2307 2308
		goto out_put;

	if (datagrams == 0) {
		datagrams = err;
		goto out_put;
	}
2309

2310 2311 2312 2313 2314
	/*
	 * We may return less entries than requested (vlen) if the
	 * sock is non block and there aren't enough datagrams...
	 */
	if (err != -EAGAIN) {
2315
		/*
2316 2317 2318 2319
		 * ... 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).
2320
		 */
2321
		sock->sk->sk_err = -err;
2322
	}
2323 2324
out_put:
	fput_light(sock->file, fput_needed);
2325

2326
	return datagrams;
2327 2328 2329 2330 2331 2332 2333 2334 2335
}

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;

2336 2337 2338
	if (flags & MSG_CMSG_COMPAT)
		return -EINVAL;

2339 2340 2341 2342 2343 2344 2345 2346 2347 2348 2349 2350 2351 2352 2353 2354
	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
2355 2356
/* Argument list sizes for sys_socketcall */
#define AL(x) ((x) * sizeof(unsigned long))
2357
static const unsigned char nargs[21] = {
2358 2359 2360
	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),
2361
	AL(4), AL(5), AL(4)
2362 2363
};

L
Linus Torvalds 已提交
2364 2365 2366
#undef AL

/*
2367
 *	System call vectors.
L
Linus Torvalds 已提交
2368 2369 2370
 *
 *	Argument checking cleaned up. Saved 20% in size.
 *  This function doesn't need to set the kernel lock because
2371
 *  it is set by the callees.
L
Linus Torvalds 已提交
2372 2373
 */

2374
SYSCALL_DEFINE2(socketcall, int, call, unsigned long __user *, args)
L
Linus Torvalds 已提交
2375
{
2376
	unsigned long a[AUDITSC_ARGS];
2377
	unsigned long a0, a1;
L
Linus Torvalds 已提交
2378
	int err;
2379
	unsigned int len;
L
Linus Torvalds 已提交
2380

2381
	if (call < 1 || call > SYS_SENDMMSG)
L
Linus Torvalds 已提交
2382 2383
		return -EINVAL;

2384 2385 2386 2387
	len = nargs[call];
	if (len > sizeof(a))
		return -EINVAL;

L
Linus Torvalds 已提交
2388
	/* copy_from_user should be SMP safe. */
2389
	if (copy_from_user(a, args, len))
L
Linus Torvalds 已提交
2390
		return -EFAULT;
2391

2392 2393 2394
	err = audit_socketcall(nargs[call] / sizeof(unsigned long), a);
	if (err)
		return err;
2395

2396 2397 2398 2399 2400 2401 2402 2403 2404 2405 2406 2407 2408 2409 2410 2411 2412
	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
2413 2414
		err = sys_accept4(a0, (struct sockaddr __user *)a1,
				  (int __user *)a[2], 0);
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 2442 2443 2444 2445 2446 2447 2448 2449 2450 2451 2452 2453 2454 2455
		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:
2456
		err = sys_sendmsg(a0, (struct user_msghdr __user *)a1, a[2]);
2457
		break;
2458 2459 2460
	case SYS_SENDMMSG:
		err = sys_sendmmsg(a0, (struct mmsghdr __user *)a1, a[2], a[3]);
		break;
2461
	case SYS_RECVMSG:
2462
		err = sys_recvmsg(a0, (struct user_msghdr __user *)a1, a[2]);
2463
		break;
2464 2465 2466 2467
	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
2468 2469 2470
	case SYS_ACCEPT4:
		err = sys_accept4(a0, (struct sockaddr __user *)a1,
				  (int __user *)a[2], a[3]);
U
Ulrich Drepper 已提交
2471
		break;
2472 2473 2474
	default:
		err = -EINVAL;
		break;
L
Linus Torvalds 已提交
2475 2476 2477 2478
	}
	return err;
}

2479
#endif				/* __ARCH_WANT_SYS_SOCKETCALL */
L
Linus Torvalds 已提交
2480

2481 2482 2483 2484
/**
 *	sock_register - add a socket protocol handler
 *	@ops: description of protocol
 *
L
Linus Torvalds 已提交
2485 2486
 *	This function is called by a protocol handler that wants to
 *	advertise its address family, and have it linked into the
2487
 *	socket interface. The value ops->family corresponds to the
2488
 *	socket system call protocol family.
L
Linus Torvalds 已提交
2489
 */
2490
int sock_register(const struct net_proto_family *ops)
L
Linus Torvalds 已提交
2491 2492 2493 2494
{
	int err;

	if (ops->family >= NPROTO) {
2495
		pr_crit("protocol %d >= NPROTO(%d)\n", ops->family, NPROTO);
L
Linus Torvalds 已提交
2496 2497
		return -ENOBUFS;
	}
2498 2499

	spin_lock(&net_family_lock);
E
Eric Dumazet 已提交
2500 2501
	if (rcu_dereference_protected(net_families[ops->family],
				      lockdep_is_held(&net_family_lock)))
2502 2503
		err = -EEXIST;
	else {
2504
		rcu_assign_pointer(net_families[ops->family], ops);
L
Linus Torvalds 已提交
2505 2506
		err = 0;
	}
2507 2508
	spin_unlock(&net_family_lock);

2509
	pr_info("NET: Registered protocol family %d\n", ops->family);
L
Linus Torvalds 已提交
2510 2511
	return err;
}
2512
EXPORT_SYMBOL(sock_register);
L
Linus Torvalds 已提交
2513

2514 2515 2516 2517
/**
 *	sock_unregister - remove a protocol handler
 *	@family: protocol family to remove
 *
L
Linus Torvalds 已提交
2518 2519
 *	This function is called by a protocol handler that wants to
 *	remove its address family, and have it unlinked from the
2520 2521 2522 2523 2524 2525
 *	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
2526
 */
2527
void sock_unregister(int family)
L
Linus Torvalds 已提交
2528
{
2529
	BUG_ON(family < 0 || family >= NPROTO);
L
Linus Torvalds 已提交
2530

2531
	spin_lock(&net_family_lock);
2532
	RCU_INIT_POINTER(net_families[family], NULL);
2533 2534 2535 2536
	spin_unlock(&net_family_lock);

	synchronize_rcu();

2537
	pr_info("NET: Unregistered protocol family %d\n", family);
L
Linus Torvalds 已提交
2538
}
2539
EXPORT_SYMBOL(sock_unregister);
L
Linus Torvalds 已提交
2540

2541
static int __init sock_init(void)
L
Linus Torvalds 已提交
2542
{
N
Nick Piggin 已提交
2543
	int err;
2544 2545 2546 2547 2548 2549
	/*
	 *      Initialize the network sysctl infrastructure.
	 */
	err = net_sysctl_init();
	if (err)
		goto out;
N
Nick Piggin 已提交
2550

L
Linus Torvalds 已提交
2551
	/*
2552
	 *      Initialize skbuff SLAB cache
L
Linus Torvalds 已提交
2553 2554 2555 2556
	 */
	skb_init();

	/*
2557
	 *      Initialize the protocols module.
L
Linus Torvalds 已提交
2558 2559 2560
	 */

	init_inodecache();
N
Nick Piggin 已提交
2561 2562 2563 2564

	err = register_filesystem(&sock_fs_type);
	if (err)
		goto out_fs;
L
Linus Torvalds 已提交
2565
	sock_mnt = kern_mount(&sock_fs_type);
N
Nick Piggin 已提交
2566 2567 2568 2569
	if (IS_ERR(sock_mnt)) {
		err = PTR_ERR(sock_mnt);
		goto out_mount;
	}
2570 2571

	/* The real protocol initialization is performed in later initcalls.
L
Linus Torvalds 已提交
2572 2573 2574
	 */

#ifdef CONFIG_NETFILTER
2575 2576 2577
	err = netfilter_init();
	if (err)
		goto out;
L
Linus Torvalds 已提交
2578
#endif
2579

2580
	ptp_classifier_init();
2581

N
Nick Piggin 已提交
2582 2583 2584 2585 2586 2587 2588
out:
	return err;

out_mount:
	unregister_filesystem(&sock_fs_type);
out_fs:
	goto out;
L
Linus Torvalds 已提交
2589 2590
}

2591 2592
core_initcall(sock_init);	/* early initcall */

L
Linus Torvalds 已提交
2593 2594 2595 2596 2597 2598
#ifdef CONFIG_PROC_FS
void socket_seq_show(struct seq_file *seq)
{
	int cpu;
	int counter = 0;

2599
	for_each_possible_cpu(cpu)
2600
	    counter += per_cpu(sockets_in_use, cpu);
L
Linus Torvalds 已提交
2601 2602 2603 2604 2605 2606 2607

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

	seq_printf(seq, "sockets: used %d\n", counter);
}
2608
#endif				/* CONFIG_PROC_FS */
L
Linus Torvalds 已提交
2609

2610
#ifdef CONFIG_COMPAT
2611
static int do_siocgstamp(struct net *net, struct socket *sock,
2612
			 unsigned int cmd, void __user *up)
2613 2614 2615 2616 2617 2618
{
	mm_segment_t old_fs = get_fs();
	struct timeval ktv;
	int err;

	set_fs(KERNEL_DS);
2619
	err = sock_do_ioctl(net, sock, cmd, (unsigned long)&ktv);
2620
	set_fs(old_fs);
2621
	if (!err)
2622
		err = compat_put_timeval(&ktv, up);
2623

2624 2625 2626
	return err;
}

2627
static int do_siocgstampns(struct net *net, struct socket *sock,
2628
			   unsigned int cmd, void __user *up)
2629 2630 2631 2632 2633 2634
{
	mm_segment_t old_fs = get_fs();
	struct timespec kts;
	int err;

	set_fs(KERNEL_DS);
2635
	err = sock_do_ioctl(net, sock, cmd, (unsigned long)&kts);
2636
	set_fs(old_fs);
2637
	if (!err)
2638
		err = compat_put_timespec(&kts, up);
2639

2640 2641 2642
	return err;
}

2643
static int dev_ifname32(struct net *net, struct compat_ifreq __user *uifr32)
2644 2645 2646 2647 2648
{
	struct ifreq __user *uifr;
	int err;

	uifr = compat_alloc_user_space(sizeof(struct ifreq));
2649
	if (copy_in_user(uifr, uifr32, sizeof(struct compat_ifreq)))
2650 2651
		return -EFAULT;

2652
	err = dev_ioctl(net, SIOCGIFNAME, uifr);
2653 2654 2655
	if (err)
		return err;

2656
	if (copy_in_user(uifr32, uifr, sizeof(struct compat_ifreq)))
2657 2658 2659 2660 2661
		return -EFAULT;

	return 0;
}

2662
static int dev_ifconf(struct net *net, struct compat_ifconf __user *uifc32)
2663
{
2664
	struct compat_ifconf ifc32;
2665 2666
	struct ifconf ifc;
	struct ifconf __user *uifc;
2667
	struct compat_ifreq __user *ifr32;
2668 2669 2670 2671
	struct ifreq __user *ifr;
	unsigned int i, j;
	int err;

2672
	if (copy_from_user(&ifc32, uifc32, sizeof(struct compat_ifconf)))
2673 2674
		return -EFAULT;

2675
	memset(&ifc, 0, sizeof(ifc));
2676 2677 2678 2679 2680 2681
	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 {
2682 2683
		size_t len = ((ifc32.ifc_len / sizeof(struct compat_ifreq)) + 1) *
			sizeof(struct ifreq);
2684 2685 2686 2687
		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);
2688
		for (i = 0; i < ifc32.ifc_len; i += sizeof(struct compat_ifreq)) {
2689
			if (copy_in_user(ifr, ifr32, sizeof(struct compat_ifreq)))
2690 2691 2692 2693 2694 2695 2696 2697
				return -EFAULT;
			ifr++;
			ifr32++;
		}
	}
	if (copy_to_user(uifc, &ifc, sizeof(struct ifconf)))
		return -EFAULT;

2698
	err = dev_ioctl(net, SIOCGIFCONF, uifc);
2699 2700 2701 2702 2703 2704 2705 2706 2707
	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;
2708 2709 2710
	     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)))
2711 2712 2713 2714 2715 2716 2717 2718 2719 2720
			return -EFAULT;
		ifr32++;
		ifr++;
	}

	if (ifc32.ifcbuf == 0) {
		/* Translate from 64-bit structure multiple to
		 * a 32-bit one.
		 */
		i = ifc.ifc_len;
2721
		i = ((i / sizeof(struct ifreq)) * sizeof(struct compat_ifreq));
2722 2723 2724 2725
		ifc32.ifc_len = i;
	} else {
		ifc32.ifc_len = i;
	}
2726
	if (copy_to_user(uifc32, &ifc32, sizeof(struct compat_ifconf)))
2727 2728 2729 2730 2731
		return -EFAULT;

	return 0;
}

2732
static int ethtool_ioctl(struct net *net, struct compat_ifreq __user *ifr32)
2733
{
2734 2735 2736 2737
	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;
2738
	struct ifreq __user *ifr;
2739 2740
	u32 rule_cnt = 0, actual_rule_cnt;
	u32 ethcmd;
2741
	u32 data;
2742
	int ret;
2743

2744 2745
	if (get_user(data, &ifr32->ifr_ifru.ifru_data))
		return -EFAULT;
2746

2747 2748 2749
	compat_rxnfc = compat_ptr(data);

	if (get_user(ethcmd, &compat_rxnfc->cmd))
2750 2751
		return -EFAULT;

2752 2753 2754 2755 2756 2757 2758 2759 2760 2761 2762 2763 2764 2765 2766 2767 2768
	/* 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:
2769
	case ETHTOOL_SRXCLSRLINS:
2770 2771 2772 2773 2774 2775 2776 2777 2778
		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
2779
	rxnfc = (void __user *)ifr + ALIGN(sizeof(struct ifreq), 8);
2780 2781

	if (copy_in_user(&ifr->ifr_name, &ifr32->ifr_name, IFNAMSIZ))
2782 2783
		return -EFAULT;

2784 2785
	if (put_user(convert_in ? rxnfc : compat_ptr(data),
		     &ifr->ifr_ifru.ifru_data))
2786 2787
		return -EFAULT;

2788
	if (convert_in) {
2789
		/* We expect there to be holes between fs.m_ext and
2790 2791
		 * fs.ring_cookie and at the end of fs, but nowhere else.
		 */
2792 2793 2794 2795
		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));
2796 2797 2798 2799 2800 2801 2802
		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
2803 2804
				 (void __user *)(&rxnfc->fs.m_ext + 1) -
				 (void __user *)rxnfc) ||
2805 2806
		    copy_in_user(&rxnfc->fs.ring_cookie,
				 &compat_rxnfc->fs.ring_cookie,
S
Stephen Hemminger 已提交
2807 2808
				 (void __user *)(&rxnfc->fs.location + 1) -
				 (void __user *)&rxnfc->fs.ring_cookie) ||
2809 2810 2811 2812 2813 2814 2815 2816 2817 2818 2819
		    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
2820 2821
				 (const void __user *)(&rxnfc->fs.m_ext + 1) -
				 (const void __user *)rxnfc) ||
2822 2823
		    copy_in_user(&compat_rxnfc->fs.ring_cookie,
				 &rxnfc->fs.ring_cookie,
S
Stephen Hemminger 已提交
2824 2825
				 (const void __user *)(&rxnfc->fs.location + 1) -
				 (const void __user *)&rxnfc->fs.ring_cookie) ||
2826 2827 2828 2829 2830 2831 2832 2833 2834 2835 2836 2837 2838 2839 2840 2841 2842 2843 2844 2845 2846 2847 2848 2849 2850
		    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;
2851 2852
}

2853 2854 2855 2856 2857 2858
static int compat_siocwandev(struct net *net, struct compat_ifreq __user *uifr32)
{
	void __user *uptr;
	compat_uptr_t uptr32;
	struct ifreq __user *uifr;

2859
	uifr = compat_alloc_user_space(sizeof(*uifr));
2860 2861 2862 2863 2864 2865 2866 2867 2868 2869 2870 2871 2872 2873
	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);
}

2874 2875
static int bond_ioctl(struct net *net, unsigned int cmd,
			 struct compat_ifreq __user *ifr32)
2876 2877 2878 2879 2880 2881 2882 2883 2884 2885
{
	struct ifreq kifr;
	mm_segment_t old_fs;
	int err;

	switch (cmd) {
	case SIOCBONDENSLAVE:
	case SIOCBONDRELEASE:
	case SIOCBONDSETHWADDR:
	case SIOCBONDCHANGEACTIVE:
2886
		if (copy_from_user(&kifr, ifr32, sizeof(struct compat_ifreq)))
2887 2888 2889
			return -EFAULT;

		old_fs = get_fs();
2890
		set_fs(KERNEL_DS);
2891 2892
		err = dev_ioctl(net, cmd,
				(struct ifreq __user __force *) &kifr);
2893
		set_fs(old_fs);
2894 2895 2896

		return err;
	default:
2897
		return -ENOIOCTLCMD;
2898
	}
2899 2900
}

2901 2902
/* 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,
2903
				 struct compat_ifreq __user *u_ifreq32)
2904 2905 2906 2907 2908 2909 2910 2911 2912
{
	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;
2913
	if (get_user(data32, &u_ifreq32->ifr_ifru.ifru_data))
2914 2915 2916 2917 2918 2919 2920 2921
		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;
2922
	if (put_user(data64, &u_ifreq64->ifr_ifru.ifru_data))
2923 2924
		return -EFAULT;

2925
	return dev_ioctl(net, cmd, u_ifreq64);
2926 2927
}

2928 2929
static int dev_ifsioc(struct net *net, struct socket *sock,
			 unsigned int cmd, struct compat_ifreq __user *uifr32)
2930
{
2931
	struct ifreq __user *uifr;
2932 2933
	int err;

2934 2935 2936 2937 2938 2939
	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);

2940 2941 2942 2943 2944 2945 2946 2947 2948 2949 2950 2951
	if (!err) {
		switch (cmd) {
		case SIOCGIFFLAGS:
		case SIOCGIFMETRIC:
		case SIOCGIFMTU:
		case SIOCGIFMEM:
		case SIOCGIFHWADDR:
		case SIOCGIFINDEX:
		case SIOCGIFADDR:
		case SIOCGIFBRDADDR:
		case SIOCGIFDSTADDR:
		case SIOCGIFNETMASK:
2952
		case SIOCGIFPFLAGS:
2953
		case SIOCGIFTXQLEN:
2954 2955
		case SIOCGMIIPHY:
		case SIOCGMIIREG:
2956
			if (copy_in_user(uifr32, uifr, sizeof(*uifr32)))
2957 2958 2959 2960 2961 2962 2963
				err = -EFAULT;
			break;
		}
	}
	return err;
}

2964 2965 2966 2967 2968 2969 2970 2971 2972 2973
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));
2974 2975 2976 2977 2978 2979
	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);
2980 2981 2982 2983
	if (err)
		return -EFAULT;

	old_fs = get_fs();
2984
	set_fs(KERNEL_DS);
2985
	err = dev_ioctl(net, cmd, (void  __user __force *)&ifr);
2986
	set_fs(old_fs);
2987 2988 2989

	if (cmd == SIOCGIFMAP && !err) {
		err = copy_to_user(uifr32, &ifr, sizeof(ifr.ifr_name));
2990 2991 2992 2993 2994 2995
		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);
2996 2997 2998 2999 3000 3001
		if (err)
			err = -EFAULT;
	}
	return err;
}

3002
struct rtentry32 {
3003
	u32		rt_pad1;
3004 3005 3006
	struct sockaddr rt_dst;         /* target address               */
	struct sockaddr rt_gateway;     /* gateway addr (RTF_GATEWAY)   */
	struct sockaddr rt_genmask;     /* target network mask (IP)     */
3007 3008 3009 3010 3011 3012 3013
	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! */
3014
	/* char * */ u32 rt_dev;        /* forcing the device at add    */
3015 3016
	u32		rt_mtu;         /* per route MTU/Window         */
	u32		rt_window;      /* Window clamping              */
3017 3018 3019 3020 3021 3022 3023 3024 3025 3026 3027 3028 3029 3030 3031 3032
	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;
};

3033 3034
static int routing_ioctl(struct net *net, struct socket *sock,
			 unsigned int cmd, void __user *argp)
3035 3036 3037 3038 3039 3040 3041 3042 3043
{
	int ret;
	void *r = NULL;
	struct in6_rtmsg r6;
	struct rtentry r4;
	char devname[16];
	u32 rtdev;
	mm_segment_t old_fs = get_fs();

3044 3045
	if (sock && sock->sk && sock->sk->sk_family == AF_INET6) { /* ipv6 */
		struct in6_rtmsg32 __user *ur6 = argp;
3046
		ret = copy_from_user(&r6.rtmsg_dst, &(ur6->rtmsg_dst),
3047
			3 * sizeof(struct in6_addr));
3048 3049 3050 3051 3052 3053 3054
		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));
3055 3056 3057

		r = (void *) &r6;
	} else { /* ipv4 */
3058
		struct rtentry32 __user *ur4 = argp;
3059
		ret = copy_from_user(&r4.rt_dst, &(ur4->rt_dst),
3060
					3 * sizeof(struct sockaddr));
3061 3062 3063 3064 3065 3066
		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));
3067
		if (rtdev) {
3068
			ret |= copy_from_user(devname, compat_ptr(rtdev), 15);
3069 3070
			r4.rt_dev = (char __user __force *)devname;
			devname[15] = 0;
3071 3072 3073 3074 3075 3076 3077 3078 3079 3080 3081
		} else
			r4.rt_dev = NULL;

		r = (void *) &r4;
	}

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

3082
	set_fs(KERNEL_DS);
3083
	ret = sock_do_ioctl(net, sock, cmd, (unsigned long) r);
3084
	set_fs(old_fs);
3085 3086 3087 3088 3089 3090 3091

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
3092
 * use compatible ioctls
3093
 */
3094
static int old_bridge_ioctl(compat_ulong_t __user *argp)
3095
{
3096
	compat_ulong_t tmp;
3097

3098
	if (get_user(tmp, argp))
3099 3100 3101 3102 3103 3104
		return -EFAULT;
	if (tmp == BRCTL_GET_VERSION)
		return BRCTL_VERSION + 1;
	return -EINVAL;
}

3105 3106 3107 3108 3109 3110
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);
3111

3112
	if (cmd >= SIOCDEVPRIVATE && cmd <= (SIOCDEVPRIVATE + 15))
3113
		return compat_ifr_data_ioctl(net, cmd, argp);
3114 3115 3116 3117 3118 3119 3120 3121 3122 3123 3124

	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);
3125 3126
	case SIOCWANDEV:
		return compat_siocwandev(net, argp);
3127 3128 3129
	case SIOCGIFMAP:
	case SIOCSIFMAP:
		return compat_sioc_ifmap(net, cmd, argp);
3130 3131 3132 3133 3134 3135 3136 3137 3138 3139 3140 3141
	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);
3142 3143
	case SIOCBONDSLAVEINFOQUERY:
	case SIOCBONDINFOQUERY:
3144
	case SIOCSHWTSTAMP:
3145
	case SIOCGHWTSTAMP:
3146
		return compat_ifr_data_ioctl(net, cmd, argp);
3147 3148 3149 3150 3151 3152 3153 3154 3155 3156 3157

	case FIOSETOWN:
	case SIOCSPGRP:
	case FIOGETOWN:
	case SIOCGPGRP:
	case SIOCBRADDBR:
	case SIOCBRDELBR:
	case SIOCGIFVLAN:
	case SIOCSIFVLAN:
	case SIOCADDDLCI:
	case SIOCDELDLCI:
3158
	case SIOCGSKNS:
3159 3160 3161 3162 3163 3164 3165 3166 3167 3168 3169 3170 3171 3172 3173 3174 3175 3176 3177 3178 3179 3180 3181 3182 3183 3184 3185 3186 3187 3188 3189
		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:
3190 3191 3192 3193
	case SIOCSIFNAME:
	case SIOCGMIIPHY:
	case SIOCGMIIREG:
	case SIOCSMIIREG:
3194
		return dev_ifsioc(net, sock, cmd, argp);
3195

3196 3197 3198 3199
	case SIOCSARP:
	case SIOCGARP:
	case SIOCDARP:
	case SIOCATMARK:
3200 3201 3202
		return sock_do_ioctl(net, sock, cmd, arg);
	}

3203 3204
	return -ENOIOCTLCMD;
}
3205

3206
static long compat_sock_ioctl(struct file *file, unsigned int cmd,
3207
			      unsigned long arg)
3208 3209 3210
{
	struct socket *sock = file->private_data;
	int ret = -ENOIOCTLCMD;
3211 3212 3213 3214 3215
	struct sock *sk;
	struct net *net;

	sk = sock->sk;
	net = sock_net(sk);
3216 3217 3218 3219

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

3220 3221 3222 3223
	if (ret == -ENOIOCTLCMD &&
	    (cmd >= SIOCIWFIRST && cmd <= SIOCIWLAST))
		ret = compat_wext_handle_ioctl(net, cmd, arg);

3224 3225 3226
	if (ret == -ENOIOCTLCMD)
		ret = compat_sock_ioctl_trans(file, sock, cmd, arg);

3227 3228 3229 3230
	return ret;
}
#endif

3231 3232 3233 3234
int kernel_bind(struct socket *sock, struct sockaddr *addr, int addrlen)
{
	return sock->ops->bind(sock, addr, addrlen);
}
3235
EXPORT_SYMBOL(kernel_bind);
3236 3237 3238 3239 3240

int kernel_listen(struct socket *sock, int backlog)
{
	return sock->ops->listen(sock, backlog);
}
3241
EXPORT_SYMBOL(kernel_listen);
3242 3243 3244 3245 3246 3247 3248 3249 3250 3251 3252

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;

3253
	err = sock->ops->accept(sock, *newsock, flags, true);
3254 3255
	if (err < 0) {
		sock_release(*newsock);
3256
		*newsock = NULL;
3257 3258 3259 3260
		goto done;
	}

	(*newsock)->ops = sock->ops;
3261
	__module_get((*newsock)->ops->owner);
3262 3263 3264 3265

done:
	return err;
}
3266
EXPORT_SYMBOL(kernel_accept);
3267 3268

int kernel_connect(struct socket *sock, struct sockaddr *addr, int addrlen,
3269
		   int flags)
3270 3271 3272
{
	return sock->ops->connect(sock, addr, addrlen, flags);
}
3273
EXPORT_SYMBOL(kernel_connect);
3274 3275 3276 3277 3278 3279

int kernel_getsockname(struct socket *sock, struct sockaddr *addr,
			 int *addrlen)
{
	return sock->ops->getname(sock, addr, addrlen, 0);
}
3280
EXPORT_SYMBOL(kernel_getsockname);
3281 3282 3283 3284 3285 3286

int kernel_getpeername(struct socket *sock, struct sockaddr *addr,
			 int *addrlen)
{
	return sock->ops->getname(sock, addr, addrlen, 1);
}
3287
EXPORT_SYMBOL(kernel_getpeername);
3288 3289 3290 3291 3292

int kernel_getsockopt(struct socket *sock, int level, int optname,
			char *optval, int *optlen)
{
	mm_segment_t oldfs = get_fs();
3293 3294
	char __user *uoptval;
	int __user *uoptlen;
3295 3296
	int err;

3297 3298 3299
	uoptval = (char __user __force *) optval;
	uoptlen = (int __user __force *) optlen;

3300 3301
	set_fs(KERNEL_DS);
	if (level == SOL_SOCKET)
3302
		err = sock_getsockopt(sock, level, optname, uoptval, uoptlen);
3303
	else
3304 3305
		err = sock->ops->getsockopt(sock, level, optname, uoptval,
					    uoptlen);
3306 3307 3308
	set_fs(oldfs);
	return err;
}
3309
EXPORT_SYMBOL(kernel_getsockopt);
3310 3311

int kernel_setsockopt(struct socket *sock, int level, int optname,
3312
			char *optval, unsigned int optlen)
3313 3314
{
	mm_segment_t oldfs = get_fs();
3315
	char __user *uoptval;
3316 3317
	int err;

3318 3319
	uoptval = (char __user __force *) optval;

3320 3321
	set_fs(KERNEL_DS);
	if (level == SOL_SOCKET)
3322
		err = sock_setsockopt(sock, level, optname, uoptval, optlen);
3323
	else
3324
		err = sock->ops->setsockopt(sock, level, optname, uoptval,
3325 3326 3327 3328
					    optlen);
	set_fs(oldfs);
	return err;
}
3329
EXPORT_SYMBOL(kernel_setsockopt);
3330 3331 3332 3333 3334 3335 3336 3337 3338

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);
}
3339
EXPORT_SYMBOL(kernel_sendpage);
3340 3341 3342 3343 3344 3345 3346 3347 3348 3349 3350 3351

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;
}
3352
EXPORT_SYMBOL(kernel_sock_ioctl);
3353

3354 3355 3356 3357 3358
int kernel_sock_shutdown(struct socket *sock, enum sock_shutdown_cmd how)
{
	return sock->ops->shutdown(sock, how);
}
EXPORT_SYMBOL(kernel_sock_shutdown);
3359 3360 3361 3362 3363 3364 3365 3366 3367 3368 3369 3370 3371 3372 3373 3374 3375 3376 3377 3378 3379 3380 3381 3382 3383 3384 3385 3386 3387 3388 3389 3390 3391 3392 3393 3394 3395 3396 3397 3398 3399 3400 3401 3402 3403 3404

/* This routine returns the IP overhead imposed by a socket i.e.
 * the length of the underlying IP header, depending on whether
 * this is an IPv4 or IPv6 socket and the length from IP options turned
 * on at the socket.
 */
u32 kernel_sock_ip_overhead(struct sock *sk)
{
	struct inet_sock *inet;
	struct ip_options_rcu *opt;
	u32 overhead = 0;
	bool owned_by_user;
#if IS_ENABLED(CONFIG_IPV6)
	struct ipv6_pinfo *np;
	struct ipv6_txoptions *optv6 = NULL;
#endif /* IS_ENABLED(CONFIG_IPV6) */

	if (!sk)
		return overhead;

	owned_by_user = sock_owned_by_user(sk);
	switch (sk->sk_family) {
	case AF_INET:
		inet = inet_sk(sk);
		overhead += sizeof(struct iphdr);
		opt = rcu_dereference_protected(inet->inet_opt,
						owned_by_user);
		if (opt)
			overhead += opt->opt.optlen;
		return overhead;
#if IS_ENABLED(CONFIG_IPV6)
	case AF_INET6:
		np = inet6_sk(sk);
		overhead += sizeof(struct ipv6hdr);
		if (np)
			optv6 = rcu_dereference_protected(np->opt,
							  owned_by_user);
		if (optv6)
			overhead += (optv6->opt_flen + optv6->opt_nflen);
		return overhead;
#endif /* IS_ENABLED(CONFIG_IPV6) */
	default: /* Returns 0 overhead if the socket is not ipv4 or ipv6 */
		return overhead;
	}
}
EXPORT_SYMBOL(kernel_sock_ip_overhead);