socket.c 79.8 KB
Newer Older
L
Linus Torvalds 已提交
1 2 3 4 5 6
/*
 * NET		An implementation of the SOCKET network access protocol.
 *
 * Version:	@(#)socket.c	1.1.93	18/02/95
 *
 * Authors:	Orest Zborowski, <obz@Kodak.COM>
7
 *		Ross Biro
L
Linus Torvalds 已提交
8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44
 *		Fred N. van Kempen, <waltje@uWalt.NL.Mugnet.ORG>
 *
 * Fixes:
 *		Anonymous	:	NOTSOCK/BADF cleanup. Error fix in
 *					shutdown()
 *		Alan Cox	:	verify_area() fixes
 *		Alan Cox	:	Removed DDI
 *		Jonathan Kamens	:	SOCK_DGRAM reconnect bug
 *		Alan Cox	:	Moved a load of checks to the very
 *					top level.
 *		Alan Cox	:	Move address structures to/from user
 *					mode above the protocol layers.
 *		Rob Janssen	:	Allow 0 length sends.
 *		Alan Cox	:	Asynchronous I/O support (cribbed from the
 *					tty drivers).
 *		Niibe Yutaka	:	Asynchronous I/O for writes (4.4BSD style)
 *		Jeff Uphoff	:	Made max number of sockets command-line
 *					configurable.
 *		Matti Aarnio	:	Made the number of sockets dynamic,
 *					to be allocated when needed, and mr.
 *					Uphoff's max is used as max to be
 *					allowed to allocate.
 *		Linus		:	Argh. removed all the socket allocation
 *					altogether: it's in the inode now.
 *		Alan Cox	:	Made sock_alloc()/sock_release() public
 *					for NetROM and future kernel nfsd type
 *					stuff.
 *		Alan Cox	:	sendmsg/recvmsg basics.
 *		Tom Dyas	:	Export net symbols.
 *		Marcin Dalecki	:	Fixed problems with CONFIG_NET="n".
 *		Alan Cox	:	Added thread locking to sys_* calls
 *					for sockets. May have errors at the
 *					moment.
 *		Kevin Buhr	:	Fixed the dumb errors in the above.
 *		Andi Kleen	:	Some small cleanups, optimizations,
 *					and fixed a copy_from_user() bug.
 *		Tigran Aivazian	:	sys_send(args) calls sys_sendto(args, NULL, 0)
45
 *		Tigran Aivazian	:	Made listen(2) backlog sanity checks
L
Linus Torvalds 已提交
46 47 48 49 50 51 52 53 54 55
 *					protocol-independent
 *
 *
 *		This program is free software; you can redistribute it and/or
 *		modify it under the terms of the GNU General Public License
 *		as published by the Free Software Foundation; either version
 *		2 of the License, or (at your option) any later version.
 *
 *
 *	This module is effectively the top level interface to the BSD socket
56
 *	paradigm.
L
Linus Torvalds 已提交
57 58 59 60 61 62 63 64 65
 *
 *	Based upon Swansea University Computer Society NET3.039
 */

#include <linux/mm.h>
#include <linux/socket.h>
#include <linux/file.h>
#include <linux/net.h>
#include <linux/interrupt.h>
U
Ulrich Drepper 已提交
66
#include <linux/thread_info.h>
67
#include <linux/rcupdate.h>
L
Linus Torvalds 已提交
68 69 70
#include <linux/netdevice.h>
#include <linux/proc_fs.h>
#include <linux/seq_file.h>
A
Arjan van de Ven 已提交
71
#include <linux/mutex.h>
L
Linus Torvalds 已提交
72
#include <linux/if_bridge.h>
73 74
#include <linux/if_frad.h>
#include <linux/if_vlan.h>
75
#include <linux/ptp_classify.h>
L
Linus Torvalds 已提交
76 77 78 79 80 81 82 83 84 85
#include <linux/init.h>
#include <linux/poll.h>
#include <linux/cache.h>
#include <linux/module.h>
#include <linux/highmem.h>
#include <linux/mount.h>
#include <linux/security.h>
#include <linux/syscalls.h>
#include <linux/compat.h>
#include <linux/kmod.h>
86
#include <linux/audit.h>
87
#include <linux/wireless.h>
88
#include <linux/nsproxy.h>
N
Nick Black 已提交
89
#include <linux/magic.h>
90
#include <linux/slab.h>
91
#include <linux/xattr.h>
L
Linus Torvalds 已提交
92 93 94 95 96

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

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

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

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

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

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

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

/*
 *	Socket files have a set of 'special' operations as well as the generic file ones. These don't appear
 *	in the operation structures but are done directly via the socketcall() multiplexor.
 */

140
static const struct file_operations socket_file_ops = {
L
Linus Torvalds 已提交
141 142
	.owner =	THIS_MODULE,
	.llseek =	no_llseek,
143 144 145 146
	.read =		new_sync_read,
	.write =	new_sync_write,
	.read_iter =	sock_read_iter,
	.write_iter =	sock_write_iter,
L
Linus Torvalds 已提交
147 148
	.poll =		sock_poll,
	.unlocked_ioctl = sock_ioctl,
149 150 151
#ifdef CONFIG_COMPAT
	.compat_ioctl = compat_sock_ioctl,
#endif
L
Linus Torvalds 已提交
152 153 154
	.mmap =		sock_mmap,
	.release =	sock_close,
	.fasync =	sock_fasync,
155 156
	.sendpage =	sock_sendpage,
	.splice_write = generic_splice_sendpage,
J
Jens Axboe 已提交
157
	.splice_read =	sock_splice_read,
L
Linus Torvalds 已提交
158 159 160 161 162 163 164
};

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

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

/*
 *	Statistics counters of the socket lists
 */

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

/*
174 175 176
 * Support routines.
 * Move socket addresses back and forth across the kernel/user
 * divide and look after the messy bits.
L
Linus Torvalds 已提交
177 178 179 180 181 182 183 184 185 186 187 188 189
 */

/**
 *	move_addr_to_kernel	-	copy a socket address into kernel space
 *	@uaddr: Address in user space
 *	@kaddr: Address in kernel space
 *	@ulen: Length in user space
 *
 *	The address is copied into kernel space. If the provided address is
 *	too long an error code of -EINVAL is returned. If the copy gives
 *	invalid addresses -EFAULT is returned. On a success 0 is returned.
 */

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

/**
 *	move_addr_to_user	-	copy an address to user space
 *	@kaddr: kernel space address
 *	@klen: length of address in kernel
 *	@uaddr: user space address
 *	@ulen: pointer to user length field
 *
 *	The value pointed to by ulen on entry is the buffer length available.
 *	This is overwritten with the buffer space used. -EINVAL is returned
 *	if an overlong buffer is specified or a negative buffer size. -EFAULT
 *	is returned if either the buffer or the length field are not
 *	accessible.
 *	After copying the data up to the limit the user specifies, the true
 *	length of the data is written over the length limit the user
 *	specified. Zero is returned for a success.
 */
217

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

224
	BUG_ON(klen > sizeof(struct sockaddr_storage));
225 226
	err = get_user(len, ulen);
	if (err)
L
Linus Torvalds 已提交
227
		return err;
228 229
	if (len > klen)
		len = klen;
230
	if (len < 0)
L
Linus Torvalds 已提交
231
		return -EINVAL;
232
	if (len) {
S
Steve Grubb 已提交
233 234
		if (audit_sockaddr(klen, kaddr))
			return -ENOMEM;
235
		if (copy_to_user(uaddr, kaddr, len))
L
Linus Torvalds 已提交
236 237 238
			return -EFAULT;
	}
	/*
239 240
	 *      "fromlen shall refer to the value before truncation.."
	 *                      1003.1g
L
Linus Torvalds 已提交
241 242 243 244
	 */
	return __put_user(klen, ulen);
}

245
static struct kmem_cache *sock_inode_cachep __read_mostly;
L
Linus Torvalds 已提交
246 247 248 249

static struct inode *sock_alloc_inode(struct super_block *sb)
{
	struct socket_alloc *ei;
250
	struct socket_wq *wq;
251

252
	ei = kmem_cache_alloc(sock_inode_cachep, GFP_KERNEL);
L
Linus Torvalds 已提交
253 254
	if (!ei)
		return NULL;
255 256
	wq = kmalloc(sizeof(*wq), GFP_KERNEL);
	if (!wq) {
257 258 259
		kmem_cache_free(sock_inode_cachep, ei);
		return NULL;
	}
260 261 262
	init_waitqueue_head(&wq->wait);
	wq->fasync_list = NULL;
	RCU_INIT_POINTER(ei->socket.wq, wq);
263

L
Linus Torvalds 已提交
264 265 266 267 268 269 270 271 272 273 274
	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)
{
275
	struct socket_alloc *ei;
276
	struct socket_wq *wq;
277 278

	ei = container_of(inode, struct socket_alloc, vfs_inode);
279
	wq = rcu_dereference_protected(ei->socket.wq, 1);
280
	kfree_rcu(wq, rcu);
281
	kmem_cache_free(sock_inode_cachep, ei);
L
Linus Torvalds 已提交
282 283
}

284
static void init_once(void *foo)
L
Linus Torvalds 已提交
285
{
286
	struct socket_alloc *ei = (struct socket_alloc *)foo;
L
Linus Torvalds 已提交
287

C
Christoph Lameter 已提交
288
	inode_init_once(&ei->vfs_inode);
L
Linus Torvalds 已提交
289
}
290

L
Linus Torvalds 已提交
291 292 293
static int init_inodecache(void)
{
	sock_inode_cachep = kmem_cache_create("sock_inode_cache",
294 295 296 297 298
					      sizeof(struct socket_alloc),
					      0,
					      (SLAB_HWCACHE_ALIGN |
					       SLAB_RECLAIM_ACCOUNT |
					       SLAB_MEM_SPREAD),
299
					      init_once);
L
Linus Torvalds 已提交
300 301 302 303 304
	if (sock_inode_cachep == NULL)
		return -ENOMEM;
	return 0;
}

305
static const struct super_operations sockfs_ops = {
306 307 308
	.alloc_inode	= sock_alloc_inode,
	.destroy_inode	= sock_destroy_inode,
	.statfs		= simple_statfs,
L
Linus Torvalds 已提交
309 310
};

311 312 313 314 315 316 317 318 319
/*
 * sockfs_dname() is called from d_path().
 */
static char *sockfs_dname(struct dentry *dentry, char *buffer, int buflen)
{
	return dynamic_dname(dentry, buffer, buflen, "socket:[%lu]",
				dentry->d_inode->i_ino);
}

A
Al Viro 已提交
320
static const struct dentry_operations sockfs_dentry_operations = {
321
	.d_dname  = sockfs_dname,
L
Linus Torvalds 已提交
322 323
};

324 325 326 327 328 329 330 331 332 333 334 335 336 337 338
static struct dentry *sockfs_mount(struct file_system_type *fs_type,
			 int flags, const char *dev_name, void *data)
{
	return mount_pseudo(fs_type, "socket:", &sockfs_ops,
		&sockfs_dentry_operations, SOCKFS_MAGIC);
}

static struct vfsmount *sock_mnt __read_mostly;

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

L
Linus Torvalds 已提交
339 340 341
/*
 *	Obtains the first available file descriptor and sets it up for use.
 *
342 343
 *	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
344 345 346 347 348 349 350 351 352 353 354 355
 *	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.
 */

356
struct file *sock_alloc_file(struct socket *sock, int flags, const char *dname)
L
Linus Torvalds 已提交
357
{
358
	struct qstr name = { .name = "" };
359
	struct path path;
360
	struct file *file;
L
Linus Torvalds 已提交
361

362 363 364 365 366 367 368
	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);
	}
369
	path.dentry = d_alloc_pseudo(sock_mnt->mnt_sb, &name);
370 371
	if (unlikely(!path.dentry))
		return ERR_PTR(-ENOMEM);
372
	path.mnt = mntget(sock_mnt);
373

374
	d_instantiate(path.dentry, SOCK_INODE(sock));
375

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

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

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

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

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

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

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

L
Linus Torvalds 已提交
419
/**
420
 *	sockfd_lookup - Go from a file number to its socket slot
L
Linus Torvalds 已提交
421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436
 *	@fd: file handle
 *	@err: pointer to an error code return
 *
 *	The file handle passed in is locked and the socket it is bound
 *	too is returned. If an error occurs the err pointer is overwritten
 *	with a negative errno code and NULL is returned. The function checks
 *	for both invalid handles and passing a handle which is not a socket.
 *
 *	On a success the socket object pointer is returned.
 */

struct socket *sockfd_lookup(int fd, int *err)
{
	struct file *file;
	struct socket *sock;

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

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

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

455
	*err = -EBADF;
456 457 458 459
	if (f.file) {
		sock = sock_from_file(f.file, err);
		if (likely(sock)) {
			*fput_needed = f.flags;
460
			return sock;
461 462
		}
		fdput(f);
L
Linus Torvalds 已提交
463
	}
464
	return NULL;
L
Linus Torvalds 已提交
465 466
}

467 468 469 470 471 472 473 474 475 476 477 478 479 480 481 482 483 484 485 486 487 488 489 490 491 492 493 494 495 496 497 498 499 500 501 502 503 504 505 506 507 508 509 510 511 512 513 514 515 516 517 518 519 520 521 522 523 524 525 526 527 528
#define XATTR_SOCKPROTONAME_SUFFIX "sockprotoname"
#define XATTR_NAME_SOCKPROTONAME (XATTR_SYSTEM_PREFIX XATTR_SOCKPROTONAME_SUFFIX)
#define XATTR_NAME_SOCKPROTONAME_LEN (sizeof(XATTR_NAME_SOCKPROTONAME)-1)
static ssize_t sockfs_getxattr(struct dentry *dentry,
			       const char *name, void *value, size_t size)
{
	const char *proto_name;
	size_t proto_size;
	int error;

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

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

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

out:
	return error;
}

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

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

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

	return used;
}

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

L
Linus Torvalds 已提交
529 530
/**
 *	sock_alloc	-	allocate a socket
531
 *
L
Linus Torvalds 已提交
532 533 534 535 536 537 538
 *	Allocate a new inode and socket object. The two are bound together
 *	and initialised. The socket is then returned. If we are out of inodes
 *	NULL is returned.
 */

static struct socket *sock_alloc(void)
{
539 540
	struct inode *inode;
	struct socket *sock;
L
Linus Torvalds 已提交
541

542
	inode = new_inode_pseudo(sock_mnt->mnt_sb);
L
Linus Torvalds 已提交
543 544 545 546 547
	if (!inode)
		return NULL;

	sock = SOCKET_I(inode);

548
	kmemcheck_annotate_bitfield(sock, type);
549
	inode->i_ino = get_next_ino();
550
	inode->i_mode = S_IFSOCK | S_IRWXUGO;
551 552
	inode->i_uid = current_fsuid();
	inode->i_gid = current_fsgid();
553
	inode->i_op = &sockfs_inode_ops;
L
Linus Torvalds 已提交
554

555
	this_cpu_add(sockets_in_use, 1);
L
Linus Torvalds 已提交
556 557 558 559 560 561 562 563 564
	return sock;
}

/**
 *	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
565
 *	an inode not a file.
L
Linus Torvalds 已提交
566
 */
567

L
Linus Torvalds 已提交
568 569 570 571 572 573 574 575 576 577
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);
	}

578
	if (rcu_dereference_protected(sock->wq, 1)->fasync_list)
579
		pr_err("%s: fasync list not empty!\n", __func__);
L
Linus Torvalds 已提交
580

581 582 583
	if (test_bit(SOCK_EXTERNALLY_ALLOCATED, &sock->flags))
		return;

584
	this_cpu_sub(sockets_in_use, 1);
L
Linus Torvalds 已提交
585 586 587 588
	if (!sock->file) {
		iput(SOCK_INODE(sock));
		return;
	}
589
	sock->file = NULL;
L
Linus Torvalds 已提交
590
}
591
EXPORT_SYMBOL(sock_release);
L
Linus Torvalds 已提交
592

593
void __sock_tx_timestamp(const struct sock *sk, __u8 *tx_flags)
594
{
595 596
	u8 flags = *tx_flags;

597
	if (sk->sk_tsflags & SOF_TIMESTAMPING_TX_HARDWARE)
598 599
		flags |= SKBTX_HW_TSTAMP;

600
	if (sk->sk_tsflags & SOF_TIMESTAMPING_TX_SOFTWARE)
601 602
		flags |= SKBTX_SW_TSTAMP;

603
	if (sk->sk_tsflags & SOF_TIMESTAMPING_TX_SCHED)
604 605
		flags |= SKBTX_SCHED_TSTAMP;

606
	if (sk->sk_tsflags & SOF_TIMESTAMPING_TX_ACK)
607
		flags |= SKBTX_ACK_TSTAMP;
608

609
	*tx_flags = flags;
610
}
611
EXPORT_SYMBOL(__sock_tx_timestamp);
612

613
static inline int sock_sendmsg_nosec(struct socket *sock, struct msghdr *msg)
L
Linus Torvalds 已提交
614
{
615 616 617
	int ret = sock->ops->sendmsg(sock, msg, iov_iter_count(&msg->msg_iter));
	BUG_ON(ret == -EIOCBQUEUED);
	return ret;
L
Linus Torvalds 已提交
618 619
}

620
int sock_sendmsg(struct socket *sock, struct msghdr *msg)
621
{
622 623
	int err = security_socket_sendmsg(sock, msg,
					  iov_iter_count(&msg->msg_iter));
624

625
	return err ?: sock_sendmsg_nosec(sock, msg);
626
}
627
EXPORT_SYMBOL(sock_sendmsg);
L
Linus Torvalds 已提交
628 629 630 631

int kernel_sendmsg(struct socket *sock, struct msghdr *msg,
		   struct kvec *vec, size_t num, size_t size)
{
632
	iov_iter_kvec(&msg->msg_iter, WRITE | ITER_KVEC, vec, num, size);
633
	return sock_sendmsg(sock, msg);
L
Linus Torvalds 已提交
634
}
635
EXPORT_SYMBOL(kernel_sendmsg);
L
Linus Torvalds 已提交
636

637 638 639 640 641 642
/*
 * 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)
{
643
	int need_software_tstamp = sock_flag(sk, SOCK_RCVTSTAMP);
644
	struct scm_timestamping tss;
645 646 647 648 649 650 651 652 653 654 655 656 657 658 659 660
	int empty = 1;
	struct skb_shared_hwtstamps *shhwtstamps =
		skb_hwtstamps(skb);

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

	if (need_software_tstamp) {
		if (!sock_flag(sk, SOCK_RCVTSTAMPNS)) {
			struct timeval tv;
			skb_get_timestamp(skb, &tv);
			put_cmsg(msg, SOL_SOCKET, SCM_TIMESTAMP,
				 sizeof(tv), &tv);
		} else {
661 662
			struct timespec ts;
			skb_get_timestampns(skb, &ts);
663
			put_cmsg(msg, SOL_SOCKET, SCM_TIMESTAMPNS,
664
				 sizeof(ts), &ts);
665 666 667
		}
	}

668
	memset(&tss, 0, sizeof(tss));
669
	if ((sk->sk_tsflags & SOF_TIMESTAMPING_SOFTWARE) &&
670
	    ktime_to_timespec_cond(skb->tstamp, tss.ts + 0))
671
		empty = 0;
672
	if (shhwtstamps &&
673
	    (sk->sk_tsflags & SOF_TIMESTAMPING_RAW_HARDWARE) &&
674
	    ktime_to_timespec_cond(shhwtstamps->hwtstamp, tss.ts + 2))
675
		empty = 0;
676 677
	if (!empty)
		put_cmsg(msg, SOL_SOCKET,
678
			 SCM_TIMESTAMPING, sizeof(tss), &tss);
679
}
680 681
EXPORT_SYMBOL_GPL(__sock_recv_timestamp);

682 683 684 685 686 687 688 689 690 691 692 693 694 695 696 697
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
698 699
static inline void sock_recv_drops(struct msghdr *msg, struct sock *sk,
				   struct sk_buff *skb)
700
{
701
	if (sock_flag(sk, SOCK_RXQ_OVFL) && skb && SOCK_SKB_CB(skb)->dropcount)
702
		put_cmsg(msg, SOL_SOCKET, SO_RXQ_OVFL,
703
			sizeof(__u32), &SOCK_SKB_CB(skb)->dropcount);
704 705
}

706
void __sock_recv_ts_and_drops(struct msghdr *msg, struct sock *sk,
707 708 709 710 711
	struct sk_buff *skb)
{
	sock_recv_timestamp(msg, sk, skb);
	sock_recv_drops(msg, sk, skb);
}
712
EXPORT_SYMBOL_GPL(__sock_recv_ts_and_drops);
713

714 715
static inline int sock_recvmsg_nosec(struct socket *sock, struct msghdr *msg,
				     size_t size, int flags)
L
Linus Torvalds 已提交
716
{
717
	return sock->ops->recvmsg(sock, msg, size, flags);
L
Linus Torvalds 已提交
718 719
}

720 721
int sock_recvmsg(struct socket *sock, struct msghdr *msg, size_t size,
		 int flags)
722 723 724
{
	int err = security_socket_recvmsg(sock, msg, size, flags);

725
	return err ?: sock_recvmsg_nosec(sock, msg, size, flags);
L
Linus Torvalds 已提交
726
}
727
EXPORT_SYMBOL(sock_recvmsg);
L
Linus Torvalds 已提交
728

729 730 731 732 733 734 735 736 737 738 739 740 741 742 743
/**
 * 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.
 */
744 745
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
746 747 748 749
{
	mm_segment_t oldfs = get_fs();
	int result;

750
	iov_iter_kvec(&msg->msg_iter, READ | ITER_KVEC, vec, num, size);
L
Linus Torvalds 已提交
751 752 753 754 755
	set_fs(KERNEL_DS);
	result = sock_recvmsg(sock, msg, size, flags);
	set_fs(oldfs);
	return result;
}
756
EXPORT_SYMBOL(kernel_recvmsg);
L
Linus Torvalds 已提交
757

758 759
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
760 761 762 763
{
	struct socket *sock;
	int flags;

764 765
	sock = file->private_data;

766 767 768
	flags = (file->f_flags & O_NONBLOCK) ? MSG_DONTWAIT : 0;
	/* more is a combination of MSG_MORE and MSG_SENDPAGE_NOTLAST */
	flags |= more;
769

770
	return kernel_sendpage(sock, page, offset, size, flags);
771
}
L
Linus Torvalds 已提交
772

J
Jens Axboe 已提交
773
static ssize_t sock_splice_read(struct file *file, loff_t *ppos,
774
				struct pipe_inode_info *pipe, size_t len,
J
Jens Axboe 已提交
775 776 777 778
				unsigned int flags)
{
	struct socket *sock = file->private_data;

779 780 781
	if (unlikely(!sock->ops->splice_read))
		return -EINVAL;

J
Jens Axboe 已提交
782 783 784
	return sock->ops->splice_read(sock, ppos, pipe, len, flags);
}

785
static ssize_t sock_read_iter(struct kiocb *iocb, struct iov_iter *to)
786
{
787 788
	struct file *file = iocb->ki_filp;
	struct socket *sock = file->private_data;
789 790
	struct msghdr msg = {.msg_iter = *to,
			     .msg_iocb = iocb};
791
	ssize_t res;
792

793 794 795 796
	if (file->f_flags & O_NONBLOCK)
		msg.msg_flags = MSG_DONTWAIT;

	if (iocb->ki_pos != 0)
L
Linus Torvalds 已提交
797
		return -ESPIPE;
798

C
Christoph Hellwig 已提交
799
	if (!iov_iter_count(to))	/* Match SYS5 behaviour */
L
Linus Torvalds 已提交
800 801
		return 0;

A
Al Viro 已提交
802
	res = sock_recvmsg(sock, &msg, iov_iter_count(to), msg.msg_flags);
803 804
	*to = msg.msg_iter;
	return res;
L
Linus Torvalds 已提交
805 806
}

807
static ssize_t sock_write_iter(struct kiocb *iocb, struct iov_iter *from)
808
{
809 810
	struct file *file = iocb->ki_filp;
	struct socket *sock = file->private_data;
811 812
	struct msghdr msg = {.msg_iter = *from,
			     .msg_iocb = iocb};
813
	ssize_t res;
L
Linus Torvalds 已提交
814

815
	if (iocb->ki_pos != 0)
816
		return -ESPIPE;
817

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

821 822 823
	if (sock->type == SOCK_SEQPACKET)
		msg.msg_flags |= MSG_EOR;

824
	res = sock_sendmsg(sock, &msg);
825 826
	*from = msg.msg_iter;
	return res;
L
Linus Torvalds 已提交
827 828 829 830 831 832 833
}

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

A
Arjan van de Ven 已提交
834
static DEFINE_MUTEX(br_ioctl_mutex);
835
static int (*br_ioctl_hook) (struct net *, unsigned int cmd, void __user *arg);
L
Linus Torvalds 已提交
836

837
void brioctl_set(int (*hook) (struct net *, unsigned int, void __user *))
L
Linus Torvalds 已提交
838
{
A
Arjan van de Ven 已提交
839
	mutex_lock(&br_ioctl_mutex);
L
Linus Torvalds 已提交
840
	br_ioctl_hook = hook;
A
Arjan van de Ven 已提交
841
	mutex_unlock(&br_ioctl_mutex);
L
Linus Torvalds 已提交
842 843 844
}
EXPORT_SYMBOL(brioctl_set);

A
Arjan van de Ven 已提交
845
static DEFINE_MUTEX(vlan_ioctl_mutex);
846
static int (*vlan_ioctl_hook) (struct net *, void __user *arg);
L
Linus Torvalds 已提交
847

848
void vlan_ioctl_set(int (*hook) (struct net *, void __user *))
L
Linus Torvalds 已提交
849
{
A
Arjan van de Ven 已提交
850
	mutex_lock(&vlan_ioctl_mutex);
L
Linus Torvalds 已提交
851
	vlan_ioctl_hook = hook;
A
Arjan van de Ven 已提交
852
	mutex_unlock(&vlan_ioctl_mutex);
L
Linus Torvalds 已提交
853 854 855
}
EXPORT_SYMBOL(vlan_ioctl_set);

A
Arjan van de Ven 已提交
856
static DEFINE_MUTEX(dlci_ioctl_mutex);
857
static int (*dlci_ioctl_hook) (unsigned int, void __user *);
L
Linus Torvalds 已提交
858

859
void dlci_ioctl_set(int (*hook) (unsigned int, void __user *))
L
Linus Torvalds 已提交
860
{
A
Arjan van de Ven 已提交
861
	mutex_lock(&dlci_ioctl_mutex);
L
Linus Torvalds 已提交
862
	dlci_ioctl_hook = hook;
A
Arjan van de Ven 已提交
863
	mutex_unlock(&dlci_ioctl_mutex);
L
Linus Torvalds 已提交
864 865 866
}
EXPORT_SYMBOL(dlci_ioctl_set);

867 868 869 870 871 872 873 874 875 876 877 878 879 880 881 882 883 884
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
885 886 887 888 889 890 891 892
/*
 *	With an ioctl, arg may well be a user mode pointer, but we don't know
 *	what to do with it - that's up to the protocol still.
 */

static long sock_ioctl(struct file *file, unsigned cmd, unsigned long arg)
{
	struct socket *sock;
893
	struct sock *sk;
L
Linus Torvalds 已提交
894 895
	void __user *argp = (void __user *)arg;
	int pid, err;
896
	struct net *net;
L
Linus Torvalds 已提交
897

898
	sock = file->private_data;
899
	sk = sock->sk;
900
	net = sock_net(sk);
L
Linus Torvalds 已提交
901
	if (cmd >= SIOCDEVPRIVATE && cmd <= (SIOCDEVPRIVATE + 15)) {
902
		err = dev_ioctl(net, cmd, argp);
L
Linus Torvalds 已提交
903
	} else
J
Johannes Berg 已提交
904
#ifdef CONFIG_WEXT_CORE
L
Linus Torvalds 已提交
905
	if (cmd >= SIOCIWFIRST && cmd <= SIOCIWLAST) {
906
		err = dev_ioctl(net, cmd, argp);
L
Linus Torvalds 已提交
907
	} else
J
Johannes Berg 已提交
908
#endif
909
		switch (cmd) {
L
Linus Torvalds 已提交
910 911 912 913 914
		case FIOSETOWN:
		case SIOCSPGRP:
			err = -EFAULT;
			if (get_user(pid, (int __user *)argp))
				break;
915 916
			f_setown(sock->file, pid, 1);
			err = 0;
L
Linus Torvalds 已提交
917 918 919
			break;
		case FIOGETOWN:
		case SIOCGPGRP:
920
			err = put_user(f_getown(sock->file),
921
				       (int __user *)argp);
L
Linus Torvalds 已提交
922 923 924 925 926 927 928 929 930
			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
931
			mutex_lock(&br_ioctl_mutex);
932
			if (br_ioctl_hook)
933
				err = br_ioctl_hook(net, cmd, argp);
A
Arjan van de Ven 已提交
934
			mutex_unlock(&br_ioctl_mutex);
L
Linus Torvalds 已提交
935 936 937 938 939 940 941
			break;
		case SIOCGIFVLAN:
		case SIOCSIFVLAN:
			err = -ENOPKG;
			if (!vlan_ioctl_hook)
				request_module("8021q");

A
Arjan van de Ven 已提交
942
			mutex_lock(&vlan_ioctl_mutex);
L
Linus Torvalds 已提交
943
			if (vlan_ioctl_hook)
944
				err = vlan_ioctl_hook(net, argp);
A
Arjan van de Ven 已提交
945
			mutex_unlock(&vlan_ioctl_mutex);
L
Linus Torvalds 已提交
946 947 948 949 950 951 952
			break;
		case SIOCADDDLCI:
		case SIOCDELDLCI:
			err = -ENOPKG;
			if (!dlci_ioctl_hook)
				request_module("dlci");

953 954
			mutex_lock(&dlci_ioctl_mutex);
			if (dlci_ioctl_hook)
L
Linus Torvalds 已提交
955
				err = dlci_ioctl_hook(cmd, argp);
956
			mutex_unlock(&dlci_ioctl_mutex);
L
Linus Torvalds 已提交
957 958
			break;
		default:
959
			err = sock_do_ioctl(net, sock, cmd, arg);
L
Linus Torvalds 已提交
960
			break;
961
		}
L
Linus Torvalds 已提交
962 963 964 965 966 967 968
	return err;
}

int sock_create_lite(int family, int type, int protocol, struct socket **res)
{
	int err;
	struct socket *sock = NULL;
969

L
Linus Torvalds 已提交
970 971 972 973 974 975 976 977 978 979 980
	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
981 982 983 984
	err = security_socket_post_create(sock, family, type, protocol, 1);
	if (err)
		goto out_release;

L
Linus Torvalds 已提交
985 986 987
out:
	*res = sock;
	return err;
V
Venkat Yekkirala 已提交
988 989 990 991
out_release:
	sock_release(sock);
	sock = NULL;
	goto out;
L
Linus Torvalds 已提交
992
}
993
EXPORT_SYMBOL(sock_create_lite);
L
Linus Torvalds 已提交
994 995

/* No kernel lock held - perfect */
996
static unsigned int sock_poll(struct file *file, poll_table *wait)
L
Linus Torvalds 已提交
997
{
998
	unsigned int busy_flag = 0;
L
Linus Torvalds 已提交
999 1000 1001
	struct socket *sock;

	/*
1002
	 *      We can't return errors to poll, so it's either yes or no.
L
Linus Torvalds 已提交
1003
	 */
1004
	sock = file->private_data;
1005

1006
	if (sk_can_busy_loop(sock->sk)) {
1007
		/* this socket can poll_ll so tell the system call */
1008
		busy_flag = POLL_BUSY_LOOP;
1009 1010

		/* once, only if requested by syscall */
1011 1012
		if (wait && (wait->_key & POLL_BUSY_LOOP))
			sk_busy_loop(sock->sk, 1);
1013 1014
	}

1015
	return busy_flag | sock->ops->poll(file, sock, wait);
L
Linus Torvalds 已提交
1016 1017
}

1018
static int sock_mmap(struct file *file, struct vm_area_struct *vma)
L
Linus Torvalds 已提交
1019
{
1020
	struct socket *sock = file->private_data;
L
Linus Torvalds 已提交
1021 1022 1023 1024

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

1025
static int sock_close(struct inode *inode, struct file *filp)
L
Linus Torvalds 已提交
1026 1027 1028 1029 1030 1031 1032 1033 1034 1035 1036 1037 1038
{
	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)
1039
 *	   or under socket lock
L
Linus Torvalds 已提交
1040 1041 1042 1043
 */

static int sock_fasync(int fd, struct file *filp, int on)
{
1044 1045
	struct socket *sock = filp->private_data;
	struct sock *sk = sock->sk;
1046
	struct socket_wq *wq;
L
Linus Torvalds 已提交
1047

1048
	if (sk == NULL)
L
Linus Torvalds 已提交
1049 1050 1051
		return -EINVAL;

	lock_sock(sk);
1052 1053
	wq = rcu_dereference_protected(sock->wq, sock_owned_by_user(sk));
	fasync_helper(fd, filp, on, &wq->fasync_list);
L
Linus Torvalds 已提交
1054

1055
	if (!wq->fasync_list)
1056 1057
		sock_reset_flag(sk, SOCK_FASYNC);
	else
E
Eric Dumazet 已提交
1058
		sock_set_flag(sk, SOCK_FASYNC);
L
Linus Torvalds 已提交
1059

1060
	release_sock(sk);
L
Linus Torvalds 已提交
1061 1062 1063
	return 0;
}

1064
/* This function may be called only under socket lock or callback_lock or rcu_lock */
L
Linus Torvalds 已提交
1065 1066 1067

int sock_wake_async(struct socket *sock, int how, int band)
{
1068 1069 1070 1071 1072 1073 1074 1075
	struct socket_wq *wq;

	if (!sock)
		return -1;
	rcu_read_lock();
	wq = rcu_dereference(sock->wq);
	if (!wq || !wq->fasync_list) {
		rcu_read_unlock();
L
Linus Torvalds 已提交
1076
		return -1;
1077
	}
1078
	switch (how) {
1079
	case SOCK_WAKE_WAITD:
L
Linus Torvalds 已提交
1080 1081 1082
		if (test_bit(SOCK_ASYNC_WAITDATA, &sock->flags))
			break;
		goto call_kill;
1083
	case SOCK_WAKE_SPACE:
L
Linus Torvalds 已提交
1084 1085 1086
		if (!test_and_clear_bit(SOCK_ASYNC_NOSPACE, &sock->flags))
			break;
		/* fall through */
1087
	case SOCK_WAKE_IO:
1088
call_kill:
1089
		kill_fasync(&wq->fasync_list, SIGIO, band);
L
Linus Torvalds 已提交
1090
		break;
1091
	case SOCK_WAKE_URG:
1092
		kill_fasync(&wq->fasync_list, SIGURG, band);
L
Linus Torvalds 已提交
1093
	}
1094
	rcu_read_unlock();
L
Linus Torvalds 已提交
1095 1096
	return 0;
}
1097
EXPORT_SYMBOL(sock_wake_async);
L
Linus Torvalds 已提交
1098

P
Pavel Emelyanov 已提交
1099
int __sock_create(struct net *net, int family, int type, int protocol,
1100
			 struct socket **res, int kern)
L
Linus Torvalds 已提交
1101 1102 1103
{
	int err;
	struct socket *sock;
1104
	const struct net_proto_family *pf;
L
Linus Torvalds 已提交
1105 1106

	/*
1107
	 *      Check protocol is in range
L
Linus Torvalds 已提交
1108 1109 1110 1111 1112 1113 1114 1115 1116 1117 1118 1119
	 */
	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) {
1120
		static int warned;
L
Linus Torvalds 已提交
1121 1122
		if (!warned) {
			warned = 1;
1123 1124
			pr_info("%s uses obsolete (PF_INET,SOCK_PACKET)\n",
				current->comm);
L
Linus Torvalds 已提交
1125 1126 1127 1128 1129 1130 1131
		}
		family = PF_PACKET;
	}

	err = security_socket_create(family, type, protocol, kern);
	if (err)
		return err;
1132

1133 1134 1135 1136 1137 1138 1139
	/*
	 *	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) {
1140
		net_warn_ratelimited("socket: no more sockets\n");
1141 1142 1143 1144 1145 1146
		return -ENFILE;	/* Not exactly a match, but its the
				   closest posix thing */
	}

	sock->type = type;

1147
#ifdef CONFIG_MODULES
1148 1149 1150
	/* 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
1151 1152 1153
	 * requested real, full-featured networking support upon configuration.
	 * Otherwise module support will break!
	 */
E
Eric Dumazet 已提交
1154
	if (rcu_access_pointer(net_families[family]) == NULL)
1155
		request_module("net-pf-%d", family);
L
Linus Torvalds 已提交
1156 1157
#endif

1158 1159 1160 1161 1162
	rcu_read_lock();
	pf = rcu_dereference(net_families[family]);
	err = -EAFNOSUPPORT;
	if (!pf)
		goto out_release;
L
Linus Torvalds 已提交
1163 1164 1165 1166 1167

	/*
	 * We will call the ->create function, that possibly is in a loadable
	 * module, so we have to bump that loadable module refcnt first.
	 */
1168
	if (!try_module_get(pf->owner))
L
Linus Torvalds 已提交
1169 1170
		goto out_release;

1171 1172 1173
	/* Now protected by module ref count */
	rcu_read_unlock();

1174
	err = pf->create(net, sock, protocol, kern);
1175
	if (err < 0)
L
Linus Torvalds 已提交
1176
		goto out_module_put;
1177

L
Linus Torvalds 已提交
1178 1179 1180 1181
	/*
	 * Now to bump the refcnt of the [loadable] module that owns this
	 * socket at sock_release time we decrement its refcnt.
	 */
1182 1183 1184
	if (!try_module_get(sock->ops->owner))
		goto out_module_busy;

L
Linus Torvalds 已提交
1185 1186 1187 1188
	/*
	 * Now that we're done with the ->create function, the [loadable]
	 * module can have its refcnt decremented
	 */
1189
	module_put(pf->owner);
V
Venkat Yekkirala 已提交
1190 1191
	err = security_socket_post_create(sock, family, type, protocol, kern);
	if (err)
1192
		goto out_sock_release;
1193
	*res = sock;
L
Linus Torvalds 已提交
1194

1195 1196 1197 1198
	return 0;

out_module_busy:
	err = -EAFNOSUPPORT;
L
Linus Torvalds 已提交
1199
out_module_put:
1200 1201 1202
	sock->ops = NULL;
	module_put(pf->owner);
out_sock_release:
L
Linus Torvalds 已提交
1203
	sock_release(sock);
1204 1205 1206 1207 1208
	return err;

out_release:
	rcu_read_unlock();
	goto out_sock_release;
L
Linus Torvalds 已提交
1209
}
P
Pavel Emelyanov 已提交
1210
EXPORT_SYMBOL(__sock_create);
L
Linus Torvalds 已提交
1211 1212 1213

int sock_create(int family, int type, int protocol, struct socket **res)
{
1214
	return __sock_create(current->nsproxy->net_ns, family, type, protocol, res, 0);
L
Linus Torvalds 已提交
1215
}
1216
EXPORT_SYMBOL(sock_create);
L
Linus Torvalds 已提交
1217 1218 1219

int sock_create_kern(int family, int type, int protocol, struct socket **res)
{
1220
	return __sock_create(&init_net, family, type, protocol, res, 1);
L
Linus Torvalds 已提交
1221
}
1222
EXPORT_SYMBOL(sock_create_kern);
L
Linus Torvalds 已提交
1223

1224
SYSCALL_DEFINE3(socket, int, family, int, type, int, protocol)
L
Linus Torvalds 已提交
1225 1226 1227
{
	int retval;
	struct socket *sock;
1228 1229
	int flags;

1230 1231 1232 1233 1234 1235
	/* 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);

1236
	flags = type & ~SOCK_TYPE_MASK;
1237
	if (flags & ~(SOCK_CLOEXEC | SOCK_NONBLOCK))
1238 1239
		return -EINVAL;
	type &= SOCK_TYPE_MASK;
L
Linus Torvalds 已提交
1240

U
Ulrich Drepper 已提交
1241 1242 1243
	if (SOCK_NONBLOCK != O_NONBLOCK && (flags & SOCK_NONBLOCK))
		flags = (flags & ~SOCK_NONBLOCK) | O_NONBLOCK;

L
Linus Torvalds 已提交
1244 1245 1246 1247
	retval = sock_create(family, type, protocol, &sock);
	if (retval < 0)
		goto out;

1248
	retval = sock_map_fd(sock, flags & (O_CLOEXEC | O_NONBLOCK));
L
Linus Torvalds 已提交
1249 1250 1251 1252 1253 1254 1255 1256 1257 1258 1259 1260 1261 1262 1263 1264
	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.
 */

1265 1266
SYSCALL_DEFINE4(socketpair, int, family, int, type, int, protocol,
		int __user *, usockvec)
L
Linus Torvalds 已提交
1267 1268 1269
{
	struct socket *sock1, *sock2;
	int fd1, fd2, err;
A
Al Viro 已提交
1270
	struct file *newfile1, *newfile2;
1271 1272 1273
	int flags;

	flags = type & ~SOCK_TYPE_MASK;
1274
	if (flags & ~(SOCK_CLOEXEC | SOCK_NONBLOCK))
1275 1276
		return -EINVAL;
	type &= SOCK_TYPE_MASK;
L
Linus Torvalds 已提交
1277

U
Ulrich Drepper 已提交
1278 1279 1280
	if (SOCK_NONBLOCK != O_NONBLOCK && (flags & SOCK_NONBLOCK))
		flags = (flags & ~SOCK_NONBLOCK) | O_NONBLOCK;

L
Linus Torvalds 已提交
1281 1282 1283 1284 1285 1286 1287 1288 1289 1290 1291 1292 1293 1294
	/*
	 * 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);
1295
	if (err < 0)
L
Linus Torvalds 已提交
1296 1297
		goto out_release_both;

1298
	fd1 = get_unused_fd_flags(flags);
1299 1300
	if (unlikely(fd1 < 0)) {
		err = fd1;
A
Al Viro 已提交
1301
		goto out_release_both;
1302
	}
1303

1304
	fd2 = get_unused_fd_flags(flags);
1305 1306
	if (unlikely(fd2 < 0)) {
		err = fd2;
1307
		goto out_put_unused_1;
1308 1309
	}

1310
	newfile1 = sock_alloc_file(sock1, flags, NULL);
1311 1312
	if (unlikely(IS_ERR(newfile1))) {
		err = PTR_ERR(newfile1);
1313
		goto out_put_unused_both;
1314 1315
	}

1316
	newfile2 = sock_alloc_file(sock2, flags, NULL);
1317 1318
	if (IS_ERR(newfile2)) {
		err = PTR_ERR(newfile2);
1319
		goto out_fput_1;
A
Al Viro 已提交
1320 1321
	}

1322 1323 1324 1325 1326 1327 1328 1329
	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
1330
	audit_fd_pair(fd1, fd2);
1331

A
Al Viro 已提交
1332 1333
	fd_install(fd1, newfile1);
	fd_install(fd2, newfile2);
L
Linus Torvalds 已提交
1334 1335 1336 1337
	/* fd1 and fd2 may be already another descriptors.
	 * Not kernel problem.
	 */

1338
	return 0;
L
Linus Torvalds 已提交
1339

1340 1341 1342 1343 1344 1345 1346 1347 1348 1349 1350 1351 1352
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
1353

1354 1355 1356 1357
out_put_unused_both:
	put_unused_fd(fd2);
out_put_unused_1:
	put_unused_fd(fd1);
L
Linus Torvalds 已提交
1358
out_release_both:
1359
	sock_release(sock2);
L
Linus Torvalds 已提交
1360
out_release_1:
1361
	sock_release(sock1);
L
Linus Torvalds 已提交
1362 1363 1364 1365 1366 1367 1368 1369 1370 1371 1372 1373
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).
 */

1374
SYSCALL_DEFINE3(bind, int, fd, struct sockaddr __user *, umyaddr, int, addrlen)
L
Linus Torvalds 已提交
1375 1376
{
	struct socket *sock;
1377
	struct sockaddr_storage address;
1378
	int err, fput_needed;
L
Linus Torvalds 已提交
1379

1380
	sock = sockfd_lookup_light(fd, &err, &fput_needed);
1381
	if (sock) {
1382
		err = move_addr_to_kernel(umyaddr, addrlen, &address);
1383 1384
		if (err >= 0) {
			err = security_socket_bind(sock,
1385
						   (struct sockaddr *)&address,
1386
						   addrlen);
1387 1388
			if (!err)
				err = sock->ops->bind(sock,
1389
						      (struct sockaddr *)
1390
						      &address, addrlen);
L
Linus Torvalds 已提交
1391
		}
1392
		fput_light(sock->file, fput_needed);
1393
	}
L
Linus Torvalds 已提交
1394 1395 1396 1397 1398 1399 1400 1401 1402
	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.
 */

1403
SYSCALL_DEFINE2(listen, int, fd, int, backlog)
L
Linus Torvalds 已提交
1404 1405
{
	struct socket *sock;
1406
	int err, fput_needed;
1407
	int somaxconn;
1408 1409 1410

	sock = sockfd_lookup_light(fd, &err, &fput_needed);
	if (sock) {
1411
		somaxconn = sock_net(sock->sk)->core.sysctl_somaxconn;
1412
		if ((unsigned int)backlog > somaxconn)
1413
			backlog = somaxconn;
L
Linus Torvalds 已提交
1414 1415

		err = security_socket_listen(sock, backlog);
1416 1417
		if (!err)
			err = sock->ops->listen(sock, backlog);
L
Linus Torvalds 已提交
1418

1419
		fput_light(sock->file, fput_needed);
L
Linus Torvalds 已提交
1420 1421 1422 1423 1424 1425 1426 1427 1428 1429 1430 1431 1432 1433 1434 1435
	}
	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.
 */

1436 1437
SYSCALL_DEFINE4(accept4, int, fd, struct sockaddr __user *, upeer_sockaddr,
		int __user *, upeer_addrlen, int, flags)
L
Linus Torvalds 已提交
1438 1439
{
	struct socket *sock, *newsock;
1440
	struct file *newfile;
1441
	int err, len, newfd, fput_needed;
1442
	struct sockaddr_storage address;
L
Linus Torvalds 已提交
1443

1444
	if (flags & ~(SOCK_CLOEXEC | SOCK_NONBLOCK))
U
Ulrich Drepper 已提交
1445 1446 1447 1448 1449
		return -EINVAL;

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

1450
	sock = sockfd_lookup_light(fd, &err, &fput_needed);
L
Linus Torvalds 已提交
1451 1452 1453 1454
	if (!sock)
		goto out;

	err = -ENFILE;
1455 1456
	newsock = sock_alloc();
	if (!newsock)
L
Linus Torvalds 已提交
1457 1458 1459 1460 1461 1462 1463 1464 1465 1466 1467
		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);

1468
	newfd = get_unused_fd_flags(flags);
1469 1470
	if (unlikely(newfd < 0)) {
		err = newfd;
1471 1472
		sock_release(newsock);
		goto out_put;
1473
	}
1474
	newfile = sock_alloc_file(newsock, flags, sock->sk->sk_prot_creator->name);
1475 1476 1477 1478 1479 1480
	if (unlikely(IS_ERR(newfile))) {
		err = PTR_ERR(newfile);
		put_unused_fd(newfd);
		sock_release(newsock);
		goto out_put;
	}
1481

1482 1483
	err = security_socket_accept(sock, newsock);
	if (err)
1484
		goto out_fd;
1485

L
Linus Torvalds 已提交
1486 1487
	err = sock->ops->accept(sock, newsock, sock->file->f_flags);
	if (err < 0)
1488
		goto out_fd;
L
Linus Torvalds 已提交
1489 1490

	if (upeer_sockaddr) {
1491
		if (newsock->ops->getname(newsock, (struct sockaddr *)&address,
1492
					  &len, 2) < 0) {
L
Linus Torvalds 已提交
1493
			err = -ECONNABORTED;
1494
			goto out_fd;
L
Linus Torvalds 已提交
1495
		}
1496
		err = move_addr_to_user(&address,
1497
					len, upeer_sockaddr, upeer_addrlen);
L
Linus Torvalds 已提交
1498
		if (err < 0)
1499
			goto out_fd;
L
Linus Torvalds 已提交
1500 1501 1502 1503
	}

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

1504 1505
	fd_install(newfd, newfile);
	err = newfd;
L
Linus Torvalds 已提交
1506 1507

out_put:
1508
	fput_light(sock->file, fput_needed);
L
Linus Torvalds 已提交
1509 1510
out:
	return err;
1511
out_fd:
1512
	fput(newfile);
1513
	put_unused_fd(newfd);
L
Linus Torvalds 已提交
1514 1515 1516
	goto out_put;
}

1517 1518
SYSCALL_DEFINE3(accept, int, fd, struct sockaddr __user *, upeer_sockaddr,
		int __user *, upeer_addrlen)
U
Ulrich Drepper 已提交
1519
{
U
Ulrich Drepper 已提交
1520
	return sys_accept4(fd, upeer_sockaddr, upeer_addrlen, 0);
U
Ulrich Drepper 已提交
1521 1522
}

L
Linus Torvalds 已提交
1523 1524 1525 1526 1527 1528 1529 1530 1531 1532 1533 1534
/*
 *	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.
 */

1535 1536
SYSCALL_DEFINE3(connect, int, fd, struct sockaddr __user *, uservaddr,
		int, addrlen)
L
Linus Torvalds 已提交
1537 1538
{
	struct socket *sock;
1539
	struct sockaddr_storage address;
1540
	int err, fput_needed;
L
Linus Torvalds 已提交
1541

1542
	sock = sockfd_lookup_light(fd, &err, &fput_needed);
L
Linus Torvalds 已提交
1543 1544
	if (!sock)
		goto out;
1545
	err = move_addr_to_kernel(uservaddr, addrlen, &address);
L
Linus Torvalds 已提交
1546 1547 1548
	if (err < 0)
		goto out_put;

1549
	err =
1550
	    security_socket_connect(sock, (struct sockaddr *)&address, addrlen);
L
Linus Torvalds 已提交
1551 1552 1553
	if (err)
		goto out_put;

1554
	err = sock->ops->connect(sock, (struct sockaddr *)&address, addrlen,
L
Linus Torvalds 已提交
1555 1556
				 sock->file->f_flags);
out_put:
1557
	fput_light(sock->file, fput_needed);
L
Linus Torvalds 已提交
1558 1559 1560 1561 1562 1563 1564 1565 1566
out:
	return err;
}

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

1567 1568
SYSCALL_DEFINE3(getsockname, int, fd, struct sockaddr __user *, usockaddr,
		int __user *, usockaddr_len)
L
Linus Torvalds 已提交
1569 1570
{
	struct socket *sock;
1571
	struct sockaddr_storage address;
1572
	int len, err, fput_needed;
1573

1574
	sock = sockfd_lookup_light(fd, &err, &fput_needed);
L
Linus Torvalds 已提交
1575 1576 1577 1578 1579 1580 1581
	if (!sock)
		goto out;

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

1582
	err = sock->ops->getname(sock, (struct sockaddr *)&address, &len, 0);
L
Linus Torvalds 已提交
1583 1584
	if (err)
		goto out_put;
1585
	err = move_addr_to_user(&address, len, usockaddr, usockaddr_len);
L
Linus Torvalds 已提交
1586 1587

out_put:
1588
	fput_light(sock->file, fput_needed);
L
Linus Torvalds 已提交
1589 1590 1591 1592 1593 1594 1595 1596 1597
out:
	return err;
}

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

1598 1599
SYSCALL_DEFINE3(getpeername, int, fd, struct sockaddr __user *, usockaddr,
		int __user *, usockaddr_len)
L
Linus Torvalds 已提交
1600 1601
{
	struct socket *sock;
1602
	struct sockaddr_storage address;
1603
	int len, err, fput_needed;
L
Linus Torvalds 已提交
1604

1605 1606
	sock = sockfd_lookup_light(fd, &err, &fput_needed);
	if (sock != NULL) {
L
Linus Torvalds 已提交
1607 1608
		err = security_socket_getpeername(sock);
		if (err) {
1609
			fput_light(sock->file, fput_needed);
L
Linus Torvalds 已提交
1610 1611 1612
			return err;
		}

1613
		err =
1614
		    sock->ops->getname(sock, (struct sockaddr *)&address, &len,
1615
				       1);
L
Linus Torvalds 已提交
1616
		if (!err)
1617
			err = move_addr_to_user(&address, len, usockaddr,
1618
						usockaddr_len);
1619
		fput_light(sock->file, fput_needed);
L
Linus Torvalds 已提交
1620 1621 1622 1623 1624 1625 1626 1627 1628 1629
	}
	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.
 */

1630
SYSCALL_DEFINE6(sendto, int, fd, void __user *, buff, size_t, len,
1631
		unsigned int, flags, struct sockaddr __user *, addr,
1632
		int, addr_len)
L
Linus Torvalds 已提交
1633 1634
{
	struct socket *sock;
1635
	struct sockaddr_storage address;
L
Linus Torvalds 已提交
1636 1637 1638
	int err;
	struct msghdr msg;
	struct iovec iov;
1639 1640
	int fput_needed;

1641 1642 1643
	err = import_single_range(WRITE, buff, len, &iov, &msg.msg_iter);
	if (unlikely(err))
		return err;
1644 1645
	sock = sockfd_lookup_light(fd, &err, &fput_needed);
	if (!sock)
1646
		goto out;
1647

1648 1649 1650 1651
	msg.msg_name = NULL;
	msg.msg_control = NULL;
	msg.msg_controllen = 0;
	msg.msg_namelen = 0;
1652
	if (addr) {
1653
		err = move_addr_to_kernel(addr, addr_len, &address);
L
Linus Torvalds 已提交
1654 1655
		if (err < 0)
			goto out_put;
1656
		msg.msg_name = (struct sockaddr *)&address;
1657
		msg.msg_namelen = addr_len;
L
Linus Torvalds 已提交
1658 1659 1660 1661
	}
	if (sock->file->f_flags & O_NONBLOCK)
		flags |= MSG_DONTWAIT;
	msg.msg_flags = flags;
1662
	err = sock_sendmsg(sock, &msg);
L
Linus Torvalds 已提交
1663

1664
out_put:
1665
	fput_light(sock->file, fput_needed);
1666
out:
L
Linus Torvalds 已提交
1667 1668 1669 1670
	return err;
}

/*
1671
 *	Send a datagram down a socket.
L
Linus Torvalds 已提交
1672 1673
 */

1674
SYSCALL_DEFINE4(send, int, fd, void __user *, buff, size_t, len,
1675
		unsigned int, flags)
L
Linus Torvalds 已提交
1676 1677 1678 1679 1680
{
	return sys_sendto(fd, buff, len, flags, NULL, 0);
}

/*
1681
 *	Receive a frame from the socket and optionally record the address of the
L
Linus Torvalds 已提交
1682 1683 1684 1685
 *	sender. We verify the buffers are writable and if needed move the
 *	sender address from kernel to user space.
 */

1686
SYSCALL_DEFINE6(recvfrom, int, fd, void __user *, ubuf, size_t, size,
1687
		unsigned int, flags, struct sockaddr __user *, addr,
1688
		int __user *, addr_len)
L
Linus Torvalds 已提交
1689 1690 1691 1692
{
	struct socket *sock;
	struct iovec iov;
	struct msghdr msg;
1693
	struct sockaddr_storage address;
1694
	int err, err2;
1695 1696
	int fput_needed;

1697 1698 1699
	err = import_single_range(READ, ubuf, size, &iov, &msg.msg_iter);
	if (unlikely(err))
		return err;
1700
	sock = sockfd_lookup_light(fd, &err, &fput_needed);
L
Linus Torvalds 已提交
1701
	if (!sock)
1702
		goto out;
L
Linus Torvalds 已提交
1703

1704 1705
	msg.msg_control = NULL;
	msg.msg_controllen = 0;
1706 1707 1708 1709
	/* 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;
L
Linus Torvalds 已提交
1710 1711
	if (sock->file->f_flags & O_NONBLOCK)
		flags |= MSG_DONTWAIT;
1712
	err = sock_recvmsg(sock, &msg, iov_iter_count(&msg.msg_iter), flags);
L
Linus Torvalds 已提交
1713

1714
	if (err >= 0 && addr != NULL) {
1715
		err2 = move_addr_to_user(&address,
1716
					 msg.msg_namelen, addr, addr_len);
1717 1718
		if (err2 < 0)
			err = err2;
L
Linus Torvalds 已提交
1719
	}
1720 1721

	fput_light(sock->file, fput_needed);
1722
out:
L
Linus Torvalds 已提交
1723 1724 1725 1726
	return err;
}

/*
1727
 *	Receive a datagram from a socket.
L
Linus Torvalds 已提交
1728 1729
 */

1730 1731
SYSCALL_DEFINE4(recv, int, fd, void __user *, ubuf, size_t, size,
		unsigned int, flags)
L
Linus Torvalds 已提交
1732 1733 1734 1735 1736 1737 1738 1739 1740
{
	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.
 */

1741 1742
SYSCALL_DEFINE5(setsockopt, int, fd, int, level, int, optname,
		char __user *, optval, int, optlen)
L
Linus Torvalds 已提交
1743
{
1744
	int err, fput_needed;
L
Linus Torvalds 已提交
1745 1746 1747 1748
	struct socket *sock;

	if (optlen < 0)
		return -EINVAL;
1749 1750 1751 1752

	sock = sockfd_lookup_light(fd, &err, &fput_needed);
	if (sock != NULL) {
		err = security_socket_setsockopt(sock, level, optname);
1753 1754
		if (err)
			goto out_put;
L
Linus Torvalds 已提交
1755 1756

		if (level == SOL_SOCKET)
1757 1758 1759
			err =
			    sock_setsockopt(sock, level, optname, optval,
					    optlen);
L
Linus Torvalds 已提交
1760
		else
1761 1762 1763
			err =
			    sock->ops->setsockopt(sock, level, optname, optval,
						  optlen);
1764 1765
out_put:
		fput_light(sock->file, fput_needed);
L
Linus Torvalds 已提交
1766 1767 1768 1769 1770 1771 1772 1773 1774
	}
	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.
 */

1775 1776
SYSCALL_DEFINE5(getsockopt, int, fd, int, level, int, optname,
		char __user *, optval, int __user *, optlen)
L
Linus Torvalds 已提交
1777
{
1778
	int err, fput_needed;
L
Linus Torvalds 已提交
1779 1780
	struct socket *sock;

1781 1782
	sock = sockfd_lookup_light(fd, &err, &fput_needed);
	if (sock != NULL) {
1783 1784 1785
		err = security_socket_getsockopt(sock, level, optname);
		if (err)
			goto out_put;
L
Linus Torvalds 已提交
1786 1787

		if (level == SOL_SOCKET)
1788 1789 1790
			err =
			    sock_getsockopt(sock, level, optname, optval,
					    optlen);
L
Linus Torvalds 已提交
1791
		else
1792 1793 1794
			err =
			    sock->ops->getsockopt(sock, level, optname, optval,
						  optlen);
1795 1796
out_put:
		fput_light(sock->file, fput_needed);
L
Linus Torvalds 已提交
1797 1798 1799 1800 1801 1802 1803 1804
	}
	return err;
}

/*
 *	Shutdown a socket.
 */

1805
SYSCALL_DEFINE2(shutdown, int, fd, int, how)
L
Linus Torvalds 已提交
1806
{
1807
	int err, fput_needed;
L
Linus Torvalds 已提交
1808 1809
	struct socket *sock;

1810 1811
	sock = sockfd_lookup_light(fd, &err, &fput_needed);
	if (sock != NULL) {
L
Linus Torvalds 已提交
1812
		err = security_socket_shutdown(sock, how);
1813 1814 1815
		if (!err)
			err = sock->ops->shutdown(sock, how);
		fput_light(sock->file, fput_needed);
L
Linus Torvalds 已提交
1816 1817 1818 1819
	}
	return err;
}

1820
/* A couple of helpful macros for getting the address of the 32/64 bit
L
Linus Torvalds 已提交
1821 1822 1823 1824 1825 1826
 * 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)

1827 1828 1829 1830 1831
struct used_address {
	struct sockaddr_storage name;
	unsigned int name_len;
};

1832 1833 1834 1835
static int copy_msghdr_from_user(struct msghdr *kmsg,
				 struct user_msghdr __user *umsg,
				 struct sockaddr __user **save_addr,
				 struct iovec **iov)
1836
{
1837 1838
	struct sockaddr __user *uaddr;
	struct iovec __user *uiov;
A
Al Viro 已提交
1839
	size_t nr_segs;
1840 1841 1842 1843 1844 1845
	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
1846
	    __get_user(nr_segs, &umsg->msg_iovlen) ||
1847 1848 1849
	    __get_user(kmsg->msg_control, &umsg->msg_control) ||
	    __get_user(kmsg->msg_controllen, &umsg->msg_controllen) ||
	    __get_user(kmsg->msg_flags, &umsg->msg_flags))
1850
		return -EFAULT;
1851

1852
	if (!uaddr)
1853 1854
		kmsg->msg_namelen = 0;

1855 1856 1857
	if (kmsg->msg_namelen < 0)
		return -EINVAL;

1858
	if (kmsg->msg_namelen > sizeof(struct sockaddr_storage))
1859
		kmsg->msg_namelen = sizeof(struct sockaddr_storage);
1860 1861 1862 1863 1864 1865 1866 1867 1868 1869 1870 1871 1872 1873 1874 1875

	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
1876
	if (nr_segs > UIO_MAXIOV)
1877 1878
		return -EMSGSIZE;

1879 1880
	kmsg->msg_iocb = NULL;

1881 1882
	return import_iovec(save_addr ? READ : WRITE, uiov, nr_segs,
			    UIO_FASTIOV, iov, &kmsg->msg_iter);
1883 1884
}

1885
static int ___sys_sendmsg(struct socket *sock, struct user_msghdr __user *msg,
1886
			 struct msghdr *msg_sys, unsigned int flags,
1887
			 struct used_address *used_address)
L
Linus Torvalds 已提交
1888
{
1889 1890
	struct compat_msghdr __user *msg_compat =
	    (struct compat_msghdr __user *)msg;
1891
	struct sockaddr_storage address;
L
Linus Torvalds 已提交
1892
	struct iovec iovstack[UIO_FASTIOV], *iov = iovstack;
1893
	unsigned char ctl[sizeof(struct cmsghdr) + 20]
1894 1895
	    __attribute__ ((aligned(sizeof(__kernel_size_t))));
	/* 20 is size of ipv6_pktinfo */
L
Linus Torvalds 已提交
1896
	unsigned char *ctl_buf = ctl;
1897
	int ctl_len;
1898
	ssize_t err;
1899

1900
	msg_sys->msg_name = &address;
L
Linus Torvalds 已提交
1901

1902
	if (MSG_CMSG_COMPAT & flags)
1903
		err = get_compat_msghdr(msg_sys, msg_compat, NULL, &iov);
1904
	else
1905
		err = copy_msghdr_from_user(msg_sys, msg, NULL, &iov);
1906
	if (err < 0)
1907
		return err;
L
Linus Torvalds 已提交
1908 1909 1910

	err = -ENOBUFS;

1911
	if (msg_sys->msg_controllen > INT_MAX)
L
Linus Torvalds 已提交
1912
		goto out_freeiov;
1913
	ctl_len = msg_sys->msg_controllen;
L
Linus Torvalds 已提交
1914
	if ((MSG_CMSG_COMPAT & flags) && ctl_len) {
1915
		err =
1916
		    cmsghdr_from_user_compat_to_kern(msg_sys, sock->sk, ctl,
1917
						     sizeof(ctl));
L
Linus Torvalds 已提交
1918 1919
		if (err)
			goto out_freeiov;
1920 1921
		ctl_buf = msg_sys->msg_control;
		ctl_len = msg_sys->msg_controllen;
L
Linus Torvalds 已提交
1922
	} else if (ctl_len) {
1923
		if (ctl_len > sizeof(ctl)) {
L
Linus Torvalds 已提交
1924
			ctl_buf = sock_kmalloc(sock->sk, ctl_len, GFP_KERNEL);
1925
			if (ctl_buf == NULL)
L
Linus Torvalds 已提交
1926 1927 1928 1929
				goto out_freeiov;
		}
		err = -EFAULT;
		/*
1930
		 * Careful! Before this, msg_sys->msg_control contains a user pointer.
L
Linus Torvalds 已提交
1931 1932 1933
		 * Afterwards, it will be a kernel pointer. Thus the compiler-assisted
		 * checking falls down on this.
		 */
1934
		if (copy_from_user(ctl_buf,
1935
				   (void __user __force *)msg_sys->msg_control,
1936
				   ctl_len))
L
Linus Torvalds 已提交
1937
			goto out_freectl;
1938
		msg_sys->msg_control = ctl_buf;
L
Linus Torvalds 已提交
1939
	}
1940
	msg_sys->msg_flags = flags;
L
Linus Torvalds 已提交
1941 1942

	if (sock->file->f_flags & O_NONBLOCK)
1943
		msg_sys->msg_flags |= MSG_DONTWAIT;
1944 1945 1946 1947 1948 1949
	/*
	 * 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.
	 */
1950 1951 1952
	if (used_address && msg_sys->msg_name &&
	    used_address->name_len == msg_sys->msg_namelen &&
	    !memcmp(&used_address->name, msg_sys->msg_name,
1953
		    used_address->name_len)) {
1954
		err = sock_sendmsg_nosec(sock, msg_sys);
1955 1956
		goto out_freectl;
	}
1957
	err = sock_sendmsg(sock, msg_sys);
1958 1959 1960 1961 1962 1963
	/*
	 * 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;
1964 1965 1966
		if (msg_sys->msg_name)
			memcpy(&used_address->name, msg_sys->msg_name,
			       used_address->name_len);
1967
	}
L
Linus Torvalds 已提交
1968 1969

out_freectl:
1970
	if (ctl_buf != ctl)
L
Linus Torvalds 已提交
1971 1972
		sock_kfree_s(sock->sk, ctl_buf, ctl_len);
out_freeiov:
1973
	kfree(iov);
1974 1975 1976 1977 1978 1979 1980
	return err;
}

/*
 *	BSD sendmsg interface
 */

1981
long __sys_sendmsg(int fd, struct user_msghdr __user *msg, unsigned flags)
1982 1983 1984
{
	int fput_needed, err;
	struct msghdr msg_sys;
1985 1986 1987
	struct socket *sock;

	sock = sockfd_lookup_light(fd, &err, &fput_needed);
1988 1989 1990
	if (!sock)
		goto out;

1991
	err = ___sys_sendmsg(sock, msg, &msg_sys, flags, NULL);
1992

1993
	fput_light(sock->file, fput_needed);
1994
out:
L
Linus Torvalds 已提交
1995 1996 1997
	return err;
}

1998
SYSCALL_DEFINE3(sendmsg, int, fd, struct user_msghdr __user *, msg, unsigned int, flags)
1999 2000 2001 2002 2003 2004
{
	if (flags & MSG_CMSG_COMPAT)
		return -EINVAL;
	return __sys_sendmsg(fd, msg, flags);
}

2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016
/*
 *	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;
2017
	struct used_address used_address;
2018

2019 2020
	if (vlen > UIO_MAXIOV)
		vlen = UIO_MAXIOV;
2021 2022 2023 2024 2025 2026 2027

	datagrams = 0;

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

2028
	used_address.name_len = UINT_MAX;
2029 2030
	entry = mmsg;
	compat_entry = (struct compat_mmsghdr __user *)mmsg;
2031
	err = 0;
2032 2033 2034

	while (datagrams < vlen) {
		if (MSG_CMSG_COMPAT & flags) {
2035
			err = ___sys_sendmsg(sock, (struct user_msghdr __user *)compat_entry,
2036
					     &msg_sys, flags, &used_address);
2037 2038 2039 2040 2041
			if (err < 0)
				break;
			err = __put_user(err, &compat_entry->msg_len);
			++compat_entry;
		} else {
2042
			err = ___sys_sendmsg(sock,
2043
					     (struct user_msghdr __user *)entry,
2044
					     &msg_sys, flags, &used_address);
2045 2046 2047 2048 2049 2050 2051 2052 2053 2054 2055 2056 2057
			if (err < 0)
				break;
			err = put_user(err, &entry->msg_len);
			++entry;
		}

		if (err)
			break;
		++datagrams;
	}

	fput_light(sock->file, fput_needed);

2058 2059
	/* We only return an error if no datagrams were able to be sent */
	if (datagrams != 0)
2060 2061 2062 2063 2064 2065 2066 2067
		return datagrams;

	return err;
}

SYSCALL_DEFINE4(sendmmsg, int, fd, struct mmsghdr __user *, mmsg,
		unsigned int, vlen, unsigned int, flags)
{
2068 2069
	if (flags & MSG_CMSG_COMPAT)
		return -EINVAL;
2070 2071 2072
	return __sys_sendmmsg(fd, mmsg, vlen, flags);
}

2073
static int ___sys_recvmsg(struct socket *sock, struct user_msghdr __user *msg,
2074
			 struct msghdr *msg_sys, unsigned int flags, int nosec)
L
Linus Torvalds 已提交
2075
{
2076 2077
	struct compat_msghdr __user *msg_compat =
	    (struct compat_msghdr __user *)msg;
L
Linus Torvalds 已提交
2078
	struct iovec iovstack[UIO_FASTIOV];
2079
	struct iovec *iov = iovstack;
L
Linus Torvalds 已提交
2080
	unsigned long cmsg_ptr;
2081 2082
	int total_len, len;
	ssize_t err;
L
Linus Torvalds 已提交
2083 2084

	/* kernel mode address */
2085
	struct sockaddr_storage addr;
L
Linus Torvalds 已提交
2086 2087 2088

	/* user mode address pointers */
	struct sockaddr __user *uaddr;
2089
	int __user *uaddr_len = COMPAT_NAMELEN(msg);
2090

2091
	msg_sys->msg_name = &addr;
L
Linus Torvalds 已提交
2092

2093
	if (MSG_CMSG_COMPAT & flags)
2094
		err = get_compat_msghdr(msg_sys, msg_compat, &uaddr, &iov);
2095
	else
2096
		err = copy_msghdr_from_user(msg_sys, msg, &uaddr, &iov);
L
Linus Torvalds 已提交
2097
	if (err < 0)
2098 2099
		return err;
	total_len = iov_iter_count(&msg_sys->msg_iter);
L
Linus Torvalds 已提交
2100

2101 2102
	cmsg_ptr = (unsigned long)msg_sys->msg_control;
	msg_sys->msg_flags = flags & (MSG_CMSG_CLOEXEC|MSG_CMSG_COMPAT);
2103

2104 2105 2106
	/* We assume all kernel code knows the size of sockaddr_storage */
	msg_sys->msg_namelen = 0;

L
Linus Torvalds 已提交
2107 2108
	if (sock->file->f_flags & O_NONBLOCK)
		flags |= MSG_DONTWAIT;
2109 2110
	err = (nosec ? sock_recvmsg_nosec : sock_recvmsg)(sock, msg_sys,
							  total_len, flags);
L
Linus Torvalds 已提交
2111 2112 2113 2114 2115
	if (err < 0)
		goto out_freeiov;
	len = err;

	if (uaddr != NULL) {
2116
		err = move_addr_to_user(&addr,
2117
					msg_sys->msg_namelen, uaddr,
2118
					uaddr_len);
L
Linus Torvalds 已提交
2119 2120 2121
		if (err < 0)
			goto out_freeiov;
	}
2122
	err = __put_user((msg_sys->msg_flags & ~MSG_CMSG_COMPAT),
2123
			 COMPAT_FLAGS(msg));
L
Linus Torvalds 已提交
2124 2125 2126
	if (err)
		goto out_freeiov;
	if (MSG_CMSG_COMPAT & flags)
2127
		err = __put_user((unsigned long)msg_sys->msg_control - cmsg_ptr,
L
Linus Torvalds 已提交
2128 2129
				 &msg_compat->msg_controllen);
	else
2130
		err = __put_user((unsigned long)msg_sys->msg_control - cmsg_ptr,
L
Linus Torvalds 已提交
2131 2132 2133 2134 2135 2136
				 &msg->msg_controllen);
	if (err)
		goto out_freeiov;
	err = len;

out_freeiov:
2137
	kfree(iov);
2138 2139 2140 2141 2142 2143 2144
	return err;
}

/*
 *	BSD recvmsg interface
 */

2145
long __sys_recvmsg(int fd, struct user_msghdr __user *msg, unsigned flags)
2146 2147 2148
{
	int fput_needed, err;
	struct msghdr msg_sys;
2149 2150 2151
	struct socket *sock;

	sock = sockfd_lookup_light(fd, &err, &fput_needed);
2152 2153 2154
	if (!sock)
		goto out;

2155
	err = ___sys_recvmsg(sock, msg, &msg_sys, flags, 0);
2156

2157
	fput_light(sock->file, fput_needed);
L
Linus Torvalds 已提交
2158 2159 2160 2161
out:
	return err;
}

2162
SYSCALL_DEFINE3(recvmsg, int, fd, struct user_msghdr __user *, msg,
2163 2164 2165 2166 2167 2168 2169
		unsigned int, flags)
{
	if (flags & MSG_CMSG_COMPAT)
		return -EINVAL;
	return __sys_recvmsg(fd, msg, flags);
}

2170 2171 2172 2173 2174 2175 2176 2177 2178 2179
/*
 *     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;
2180
	struct compat_mmsghdr __user *compat_entry;
2181 2182 2183 2184 2185 2186 2187 2188 2189 2190 2191 2192 2193 2194 2195 2196 2197 2198 2199
	struct msghdr msg_sys;
	struct timespec end_time;

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

	datagrams = 0;

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

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

	entry = mmsg;
2200
	compat_entry = (struct compat_mmsghdr __user *)mmsg;
2201 2202 2203 2204 2205

	while (datagrams < vlen) {
		/*
		 * No need to ask LSM for more than the first datagram.
		 */
2206
		if (MSG_CMSG_COMPAT & flags) {
2207
			err = ___sys_recvmsg(sock, (struct user_msghdr __user *)compat_entry,
2208 2209
					     &msg_sys, flags & ~MSG_WAITFORONE,
					     datagrams);
2210 2211 2212 2213 2214
			if (err < 0)
				break;
			err = __put_user(err, &compat_entry->msg_len);
			++compat_entry;
		} else {
2215
			err = ___sys_recvmsg(sock,
2216
					     (struct user_msghdr __user *)entry,
2217 2218
					     &msg_sys, flags & ~MSG_WAITFORONE,
					     datagrams);
2219 2220 2221 2222 2223 2224
			if (err < 0)
				break;
			err = put_user(err, &entry->msg_len);
			++entry;
		}

2225 2226 2227 2228
		if (err)
			break;
		++datagrams;

2229 2230 2231 2232
		/* MSG_WAITFORONE turns on MSG_DONTWAIT after one packet */
		if (flags & MSG_WAITFORONE)
			flags |= MSG_DONTWAIT;

2233 2234 2235 2236 2237 2238 2239 2240 2241 2242 2243 2244 2245 2246 2247 2248 2249 2250 2251 2252
		if (timeout) {
			ktime_get_ts(timeout);
			*timeout = timespec_sub(end_time, *timeout);
			if (timeout->tv_sec < 0) {
				timeout->tv_sec = timeout->tv_nsec = 0;
				break;
			}

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

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

out_put:
	fput_light(sock->file, fput_needed);
L
Linus Torvalds 已提交
2253

2254 2255 2256 2257 2258 2259 2260 2261 2262 2263 2264 2265 2266 2267 2268 2269 2270 2271 2272 2273 2274 2275 2276 2277 2278 2279 2280 2281 2282 2283 2284
	if (err == 0)
		return datagrams;

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

		return datagrams;
	}

	return err;
}

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

2285 2286 2287
	if (flags & MSG_CMSG_COMPAT)
		return -EINVAL;

2288 2289 2290 2291 2292 2293 2294 2295 2296 2297 2298 2299 2300 2301 2302 2303
	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
2304 2305
/* Argument list sizes for sys_socketcall */
#define AL(x) ((x) * sizeof(unsigned long))
2306
static const unsigned char nargs[21] = {
2307 2308 2309
	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),
2310
	AL(4), AL(5), AL(4)
2311 2312
};

L
Linus Torvalds 已提交
2313 2314 2315
#undef AL

/*
2316
 *	System call vectors.
L
Linus Torvalds 已提交
2317 2318 2319
 *
 *	Argument checking cleaned up. Saved 20% in size.
 *  This function doesn't need to set the kernel lock because
2320
 *  it is set by the callees.
L
Linus Torvalds 已提交
2321 2322
 */

2323
SYSCALL_DEFINE2(socketcall, int, call, unsigned long __user *, args)
L
Linus Torvalds 已提交
2324
{
2325
	unsigned long a[AUDITSC_ARGS];
2326
	unsigned long a0, a1;
L
Linus Torvalds 已提交
2327
	int err;
2328
	unsigned int len;
L
Linus Torvalds 已提交
2329

2330
	if (call < 1 || call > SYS_SENDMMSG)
L
Linus Torvalds 已提交
2331 2332
		return -EINVAL;

2333 2334 2335 2336
	len = nargs[call];
	if (len > sizeof(a))
		return -EINVAL;

L
Linus Torvalds 已提交
2337
	/* copy_from_user should be SMP safe. */
2338
	if (copy_from_user(a, args, len))
L
Linus Torvalds 已提交
2339
		return -EFAULT;
2340

2341 2342 2343
	err = audit_socketcall(nargs[call] / sizeof(unsigned long), a);
	if (err)
		return err;
2344

2345 2346 2347 2348 2349 2350 2351 2352 2353 2354 2355 2356 2357 2358 2359 2360 2361
	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
2362 2363
		err = sys_accept4(a0, (struct sockaddr __user *)a1,
				  (int __user *)a[2], 0);
2364 2365 2366 2367 2368 2369 2370 2371 2372 2373 2374 2375 2376 2377 2378 2379 2380 2381 2382 2383 2384 2385 2386 2387 2388 2389 2390 2391 2392 2393 2394 2395 2396 2397 2398 2399 2400 2401 2402 2403 2404
		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:
2405
		err = sys_sendmsg(a0, (struct user_msghdr __user *)a1, a[2]);
2406
		break;
2407 2408 2409
	case SYS_SENDMMSG:
		err = sys_sendmmsg(a0, (struct mmsghdr __user *)a1, a[2], a[3]);
		break;
2410
	case SYS_RECVMSG:
2411
		err = sys_recvmsg(a0, (struct user_msghdr __user *)a1, a[2]);
2412
		break;
2413 2414 2415 2416
	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
2417 2418 2419
	case SYS_ACCEPT4:
		err = sys_accept4(a0, (struct sockaddr __user *)a1,
				  (int __user *)a[2], a[3]);
U
Ulrich Drepper 已提交
2420
		break;
2421 2422 2423
	default:
		err = -EINVAL;
		break;
L
Linus Torvalds 已提交
2424 2425 2426 2427
	}
	return err;
}

2428
#endif				/* __ARCH_WANT_SYS_SOCKETCALL */
L
Linus Torvalds 已提交
2429

2430 2431 2432 2433
/**
 *	sock_register - add a socket protocol handler
 *	@ops: description of protocol
 *
L
Linus Torvalds 已提交
2434 2435
 *	This function is called by a protocol handler that wants to
 *	advertise its address family, and have it linked into the
2436
 *	socket interface. The value ops->family corresponds to the
2437
 *	socket system call protocol family.
L
Linus Torvalds 已提交
2438
 */
2439
int sock_register(const struct net_proto_family *ops)
L
Linus Torvalds 已提交
2440 2441 2442 2443
{
	int err;

	if (ops->family >= NPROTO) {
2444
		pr_crit("protocol %d >= NPROTO(%d)\n", ops->family, NPROTO);
L
Linus Torvalds 已提交
2445 2446
		return -ENOBUFS;
	}
2447 2448

	spin_lock(&net_family_lock);
E
Eric Dumazet 已提交
2449 2450
	if (rcu_dereference_protected(net_families[ops->family],
				      lockdep_is_held(&net_family_lock)))
2451 2452
		err = -EEXIST;
	else {
2453
		rcu_assign_pointer(net_families[ops->family], ops);
L
Linus Torvalds 已提交
2454 2455
		err = 0;
	}
2456 2457
	spin_unlock(&net_family_lock);

2458
	pr_info("NET: Registered protocol family %d\n", ops->family);
L
Linus Torvalds 已提交
2459 2460
	return err;
}
2461
EXPORT_SYMBOL(sock_register);
L
Linus Torvalds 已提交
2462

2463 2464 2465 2466
/**
 *	sock_unregister - remove a protocol handler
 *	@family: protocol family to remove
 *
L
Linus Torvalds 已提交
2467 2468
 *	This function is called by a protocol handler that wants to
 *	remove its address family, and have it unlinked from the
2469 2470 2471 2472 2473 2474
 *	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
2475
 */
2476
void sock_unregister(int family)
L
Linus Torvalds 已提交
2477
{
2478
	BUG_ON(family < 0 || family >= NPROTO);
L
Linus Torvalds 已提交
2479

2480
	spin_lock(&net_family_lock);
2481
	RCU_INIT_POINTER(net_families[family], NULL);
2482 2483 2484 2485
	spin_unlock(&net_family_lock);

	synchronize_rcu();

2486
	pr_info("NET: Unregistered protocol family %d\n", family);
L
Linus Torvalds 已提交
2487
}
2488
EXPORT_SYMBOL(sock_unregister);
L
Linus Torvalds 已提交
2489

2490
static int __init sock_init(void)
L
Linus Torvalds 已提交
2491
{
N
Nick Piggin 已提交
2492
	int err;
2493 2494 2495 2496 2497 2498
	/*
	 *      Initialize the network sysctl infrastructure.
	 */
	err = net_sysctl_init();
	if (err)
		goto out;
N
Nick Piggin 已提交
2499

L
Linus Torvalds 已提交
2500
	/*
2501
	 *      Initialize skbuff SLAB cache
L
Linus Torvalds 已提交
2502 2503 2504 2505
	 */
	skb_init();

	/*
2506
	 *      Initialize the protocols module.
L
Linus Torvalds 已提交
2507 2508 2509
	 */

	init_inodecache();
N
Nick Piggin 已提交
2510 2511 2512 2513

	err = register_filesystem(&sock_fs_type);
	if (err)
		goto out_fs;
L
Linus Torvalds 已提交
2514
	sock_mnt = kern_mount(&sock_fs_type);
N
Nick Piggin 已提交
2515 2516 2517 2518
	if (IS_ERR(sock_mnt)) {
		err = PTR_ERR(sock_mnt);
		goto out_mount;
	}
2519 2520

	/* The real protocol initialization is performed in later initcalls.
L
Linus Torvalds 已提交
2521 2522 2523
	 */

#ifdef CONFIG_NETFILTER
2524 2525 2526
	err = netfilter_init();
	if (err)
		goto out;
L
Linus Torvalds 已提交
2527
#endif
2528

2529
	ptp_classifier_init();
2530

N
Nick Piggin 已提交
2531 2532 2533 2534 2535 2536 2537
out:
	return err;

out_mount:
	unregister_filesystem(&sock_fs_type);
out_fs:
	goto out;
L
Linus Torvalds 已提交
2538 2539
}

2540 2541
core_initcall(sock_init);	/* early initcall */

L
Linus Torvalds 已提交
2542 2543 2544 2545 2546 2547
#ifdef CONFIG_PROC_FS
void socket_seq_show(struct seq_file *seq)
{
	int cpu;
	int counter = 0;

2548
	for_each_possible_cpu(cpu)
2549
	    counter += per_cpu(sockets_in_use, cpu);
L
Linus Torvalds 已提交
2550 2551 2552 2553 2554 2555 2556

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

	seq_printf(seq, "sockets: used %d\n", counter);
}
2557
#endif				/* CONFIG_PROC_FS */
L
Linus Torvalds 已提交
2558

2559
#ifdef CONFIG_COMPAT
2560
static int do_siocgstamp(struct net *net, struct socket *sock,
2561
			 unsigned int cmd, void __user *up)
2562 2563 2564 2565 2566 2567
{
	mm_segment_t old_fs = get_fs();
	struct timeval ktv;
	int err;

	set_fs(KERNEL_DS);
2568
	err = sock_do_ioctl(net, sock, cmd, (unsigned long)&ktv);
2569
	set_fs(old_fs);
2570
	if (!err)
2571
		err = compat_put_timeval(&ktv, up);
2572

2573 2574 2575
	return err;
}

2576
static int do_siocgstampns(struct net *net, struct socket *sock,
2577
			   unsigned int cmd, void __user *up)
2578 2579 2580 2581 2582 2583
{
	mm_segment_t old_fs = get_fs();
	struct timespec kts;
	int err;

	set_fs(KERNEL_DS);
2584
	err = sock_do_ioctl(net, sock, cmd, (unsigned long)&kts);
2585
	set_fs(old_fs);
2586
	if (!err)
2587
		err = compat_put_timespec(&kts, up);
2588

2589 2590 2591
	return err;
}

2592
static int dev_ifname32(struct net *net, struct compat_ifreq __user *uifr32)
2593 2594 2595 2596 2597
{
	struct ifreq __user *uifr;
	int err;

	uifr = compat_alloc_user_space(sizeof(struct ifreq));
2598
	if (copy_in_user(uifr, uifr32, sizeof(struct compat_ifreq)))
2599 2600
		return -EFAULT;

2601
	err = dev_ioctl(net, SIOCGIFNAME, uifr);
2602 2603 2604
	if (err)
		return err;

2605
	if (copy_in_user(uifr32, uifr, sizeof(struct compat_ifreq)))
2606 2607 2608 2609 2610
		return -EFAULT;

	return 0;
}

2611
static int dev_ifconf(struct net *net, struct compat_ifconf __user *uifc32)
2612
{
2613
	struct compat_ifconf ifc32;
2614 2615
	struct ifconf ifc;
	struct ifconf __user *uifc;
2616
	struct compat_ifreq __user *ifr32;
2617 2618 2619 2620
	struct ifreq __user *ifr;
	unsigned int i, j;
	int err;

2621
	if (copy_from_user(&ifc32, uifc32, sizeof(struct compat_ifconf)))
2622 2623
		return -EFAULT;

2624
	memset(&ifc, 0, sizeof(ifc));
2625 2626 2627 2628 2629 2630
	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 {
2631 2632
		size_t len = ((ifc32.ifc_len / sizeof(struct compat_ifreq)) + 1) *
			sizeof(struct ifreq);
2633 2634 2635 2636
		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);
2637
		for (i = 0; i < ifc32.ifc_len; i += sizeof(struct compat_ifreq)) {
2638
			if (copy_in_user(ifr, ifr32, sizeof(struct compat_ifreq)))
2639 2640 2641 2642 2643 2644 2645 2646
				return -EFAULT;
			ifr++;
			ifr32++;
		}
	}
	if (copy_to_user(uifc, &ifc, sizeof(struct ifconf)))
		return -EFAULT;

2647
	err = dev_ioctl(net, SIOCGIFCONF, uifc);
2648 2649 2650 2651 2652 2653 2654 2655 2656
	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;
2657 2658 2659
	     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)))
2660 2661 2662 2663 2664 2665 2666 2667 2668 2669
			return -EFAULT;
		ifr32++;
		ifr++;
	}

	if (ifc32.ifcbuf == 0) {
		/* Translate from 64-bit structure multiple to
		 * a 32-bit one.
		 */
		i = ifc.ifc_len;
2670
		i = ((i / sizeof(struct ifreq)) * sizeof(struct compat_ifreq));
2671 2672 2673 2674
		ifc32.ifc_len = i;
	} else {
		ifc32.ifc_len = i;
	}
2675
	if (copy_to_user(uifc32, &ifc32, sizeof(struct compat_ifconf)))
2676 2677 2678 2679 2680
		return -EFAULT;

	return 0;
}

2681
static int ethtool_ioctl(struct net *net, struct compat_ifreq __user *ifr32)
2682
{
2683 2684 2685 2686
	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;
2687
	struct ifreq __user *ifr;
2688 2689
	u32 rule_cnt = 0, actual_rule_cnt;
	u32 ethcmd;
2690
	u32 data;
2691
	int ret;
2692

2693 2694
	if (get_user(data, &ifr32->ifr_ifru.ifru_data))
		return -EFAULT;
2695

2696 2697 2698
	compat_rxnfc = compat_ptr(data);

	if (get_user(ethcmd, &compat_rxnfc->cmd))
2699 2700
		return -EFAULT;

2701 2702 2703 2704 2705 2706 2707 2708 2709 2710 2711 2712 2713 2714 2715 2716 2717
	/* 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:
2718
	case ETHTOOL_SRXCLSRLINS:
2719 2720 2721 2722 2723 2724 2725 2726 2727
		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
2728
	rxnfc = (void __user *)ifr + ALIGN(sizeof(struct ifreq), 8);
2729 2730

	if (copy_in_user(&ifr->ifr_name, &ifr32->ifr_name, IFNAMSIZ))
2731 2732
		return -EFAULT;

2733 2734
	if (put_user(convert_in ? rxnfc : compat_ptr(data),
		     &ifr->ifr_ifru.ifru_data))
2735 2736
		return -EFAULT;

2737
	if (convert_in) {
2738
		/* We expect there to be holes between fs.m_ext and
2739 2740
		 * fs.ring_cookie and at the end of fs, but nowhere else.
		 */
2741 2742 2743 2744
		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));
2745 2746 2747 2748 2749 2750 2751
		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
2752 2753
				 (void __user *)(&rxnfc->fs.m_ext + 1) -
				 (void __user *)rxnfc) ||
2754 2755
		    copy_in_user(&rxnfc->fs.ring_cookie,
				 &compat_rxnfc->fs.ring_cookie,
S
Stephen Hemminger 已提交
2756 2757
				 (void __user *)(&rxnfc->fs.location + 1) -
				 (void __user *)&rxnfc->fs.ring_cookie) ||
2758 2759 2760 2761 2762 2763 2764 2765 2766 2767 2768
		    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
2769 2770
				 (const void __user *)(&rxnfc->fs.m_ext + 1) -
				 (const void __user *)rxnfc) ||
2771 2772
		    copy_in_user(&compat_rxnfc->fs.ring_cookie,
				 &rxnfc->fs.ring_cookie,
S
Stephen Hemminger 已提交
2773 2774
				 (const void __user *)(&rxnfc->fs.location + 1) -
				 (const void __user *)&rxnfc->fs.ring_cookie) ||
2775 2776 2777 2778 2779 2780 2781 2782 2783 2784 2785 2786 2787 2788 2789 2790 2791 2792 2793 2794 2795 2796 2797 2798 2799
		    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;
2800 2801
}

2802 2803 2804 2805 2806 2807
static int compat_siocwandev(struct net *net, struct compat_ifreq __user *uifr32)
{
	void __user *uptr;
	compat_uptr_t uptr32;
	struct ifreq __user *uifr;

2808
	uifr = compat_alloc_user_space(sizeof(*uifr));
2809 2810 2811 2812 2813 2814 2815 2816 2817 2818 2819 2820 2821 2822
	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);
}

2823 2824
static int bond_ioctl(struct net *net, unsigned int cmd,
			 struct compat_ifreq __user *ifr32)
2825 2826 2827 2828 2829 2830 2831 2832 2833 2834
{
	struct ifreq kifr;
	mm_segment_t old_fs;
	int err;

	switch (cmd) {
	case SIOCBONDENSLAVE:
	case SIOCBONDRELEASE:
	case SIOCBONDSETHWADDR:
	case SIOCBONDCHANGEACTIVE:
2835
		if (copy_from_user(&kifr, ifr32, sizeof(struct compat_ifreq)))
2836 2837 2838
			return -EFAULT;

		old_fs = get_fs();
2839
		set_fs(KERNEL_DS);
2840 2841
		err = dev_ioctl(net, cmd,
				(struct ifreq __user __force *) &kifr);
2842
		set_fs(old_fs);
2843 2844 2845

		return err;
	default:
2846
		return -ENOIOCTLCMD;
2847
	}
2848 2849
}

2850 2851
/* 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,
2852
				 struct compat_ifreq __user *u_ifreq32)
2853 2854 2855 2856 2857 2858 2859 2860 2861
{
	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;
2862
	if (get_user(data32, &u_ifreq32->ifr_ifru.ifru_data))
2863 2864 2865 2866 2867 2868 2869 2870
		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;
2871
	if (put_user(data64, &u_ifreq64->ifr_ifru.ifru_data))
2872 2873
		return -EFAULT;

2874
	return dev_ioctl(net, cmd, u_ifreq64);
2875 2876
}

2877 2878
static int dev_ifsioc(struct net *net, struct socket *sock,
			 unsigned int cmd, struct compat_ifreq __user *uifr32)
2879
{
2880
	struct ifreq __user *uifr;
2881 2882
	int err;

2883 2884 2885 2886 2887 2888
	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);

2889 2890 2891 2892 2893 2894 2895 2896 2897 2898 2899 2900
	if (!err) {
		switch (cmd) {
		case SIOCGIFFLAGS:
		case SIOCGIFMETRIC:
		case SIOCGIFMTU:
		case SIOCGIFMEM:
		case SIOCGIFHWADDR:
		case SIOCGIFINDEX:
		case SIOCGIFADDR:
		case SIOCGIFBRDADDR:
		case SIOCGIFDSTADDR:
		case SIOCGIFNETMASK:
2901
		case SIOCGIFPFLAGS:
2902
		case SIOCGIFTXQLEN:
2903 2904
		case SIOCGMIIPHY:
		case SIOCGMIIREG:
2905
			if (copy_in_user(uifr32, uifr, sizeof(*uifr32)))
2906 2907 2908 2909 2910 2911 2912
				err = -EFAULT;
			break;
		}
	}
	return err;
}

2913 2914 2915 2916 2917 2918 2919 2920 2921 2922
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));
2923 2924 2925 2926 2927 2928
	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);
2929 2930 2931 2932
	if (err)
		return -EFAULT;

	old_fs = get_fs();
2933
	set_fs(KERNEL_DS);
2934
	err = dev_ioctl(net, cmd, (void  __user __force *)&ifr);
2935
	set_fs(old_fs);
2936 2937 2938

	if (cmd == SIOCGIFMAP && !err) {
		err = copy_to_user(uifr32, &ifr, sizeof(ifr.ifr_name));
2939 2940 2941 2942 2943 2944
		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);
2945 2946 2947 2948 2949 2950
		if (err)
			err = -EFAULT;
	}
	return err;
}

2951
struct rtentry32 {
2952
	u32		rt_pad1;
2953 2954 2955
	struct sockaddr rt_dst;         /* target address               */
	struct sockaddr rt_gateway;     /* gateway addr (RTF_GATEWAY)   */
	struct sockaddr rt_genmask;     /* target network mask (IP)     */
2956 2957 2958 2959 2960 2961 2962
	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! */
2963
	/* char * */ u32 rt_dev;        /* forcing the device at add    */
2964 2965
	u32		rt_mtu;         /* per route MTU/Window         */
	u32		rt_window;      /* Window clamping              */
2966 2967 2968 2969 2970 2971 2972 2973 2974 2975 2976 2977 2978 2979 2980 2981
	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;
};

2982 2983
static int routing_ioctl(struct net *net, struct socket *sock,
			 unsigned int cmd, void __user *argp)
2984 2985 2986 2987 2988 2989 2990 2991 2992
{
	int ret;
	void *r = NULL;
	struct in6_rtmsg r6;
	struct rtentry r4;
	char devname[16];
	u32 rtdev;
	mm_segment_t old_fs = get_fs();

2993 2994
	if (sock && sock->sk && sock->sk->sk_family == AF_INET6) { /* ipv6 */
		struct in6_rtmsg32 __user *ur6 = argp;
2995
		ret = copy_from_user(&r6.rtmsg_dst, &(ur6->rtmsg_dst),
2996
			3 * sizeof(struct in6_addr));
2997 2998 2999 3000 3001 3002 3003
		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));
3004 3005 3006

		r = (void *) &r6;
	} else { /* ipv4 */
3007
		struct rtentry32 __user *ur4 = argp;
3008
		ret = copy_from_user(&r4.rt_dst, &(ur4->rt_dst),
3009
					3 * sizeof(struct sockaddr));
3010 3011 3012 3013 3014 3015
		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));
3016
		if (rtdev) {
3017
			ret |= copy_from_user(devname, compat_ptr(rtdev), 15);
3018 3019
			r4.rt_dev = (char __user __force *)devname;
			devname[15] = 0;
3020 3021 3022 3023 3024 3025 3026 3027 3028 3029 3030
		} else
			r4.rt_dev = NULL;

		r = (void *) &r4;
	}

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

3031
	set_fs(KERNEL_DS);
3032
	ret = sock_do_ioctl(net, sock, cmd, (unsigned long) r);
3033
	set_fs(old_fs);
3034 3035 3036 3037 3038 3039 3040

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
3041
 * use compatible ioctls
3042
 */
3043
static int old_bridge_ioctl(compat_ulong_t __user *argp)
3044
{
3045
	compat_ulong_t tmp;
3046

3047
	if (get_user(tmp, argp))
3048 3049 3050 3051 3052 3053
		return -EFAULT;
	if (tmp == BRCTL_GET_VERSION)
		return BRCTL_VERSION + 1;
	return -EINVAL;
}

3054 3055 3056 3057 3058 3059
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);
3060

3061
	if (cmd >= SIOCDEVPRIVATE && cmd <= (SIOCDEVPRIVATE + 15))
3062
		return compat_ifr_data_ioctl(net, cmd, argp);
3063 3064 3065 3066 3067 3068 3069 3070 3071 3072 3073

	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);
3074 3075
	case SIOCWANDEV:
		return compat_siocwandev(net, argp);
3076 3077 3078
	case SIOCGIFMAP:
	case SIOCSIFMAP:
		return compat_sioc_ifmap(net, cmd, argp);
3079 3080 3081 3082 3083 3084 3085 3086 3087 3088 3089 3090
	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);
3091 3092
	case SIOCBONDSLAVEINFOQUERY:
	case SIOCBONDINFOQUERY:
3093
	case SIOCSHWTSTAMP:
3094
	case SIOCGHWTSTAMP:
3095
		return compat_ifr_data_ioctl(net, cmd, argp);
3096 3097 3098 3099 3100 3101 3102 3103 3104 3105 3106 3107 3108 3109 3110 3111 3112 3113 3114 3115 3116 3117 3118 3119 3120 3121 3122 3123 3124 3125 3126 3127 3128 3129 3130 3131 3132 3133 3134 3135 3136 3137

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

	case SIOCGIFFLAGS:
	case SIOCSIFFLAGS:
	case SIOCGIFMETRIC:
	case SIOCSIFMETRIC:
	case SIOCGIFMTU:
	case SIOCSIFMTU:
	case SIOCGIFMEM:
	case SIOCSIFMEM:
	case SIOCGIFHWADDR:
	case SIOCSIFHWADDR:
	case SIOCADDMULTI:
	case SIOCDELMULTI:
	case SIOCGIFINDEX:
	case SIOCGIFADDR:
	case SIOCSIFADDR:
	case SIOCSIFHWBROADCAST:
	case SIOCDIFADDR:
	case SIOCGIFBRDADDR:
	case SIOCSIFBRDADDR:
	case SIOCGIFDSTADDR:
	case SIOCSIFDSTADDR:
	case SIOCGIFNETMASK:
	case SIOCSIFNETMASK:
	case SIOCSIFPFLAGS:
	case SIOCGIFPFLAGS:
	case SIOCGIFTXQLEN:
	case SIOCSIFTXQLEN:
	case SIOCBRADDIF:
	case SIOCBRDELIF:
3138 3139 3140 3141
	case SIOCSIFNAME:
	case SIOCGMIIPHY:
	case SIOCGMIIREG:
	case SIOCSMIIREG:
3142
		return dev_ifsioc(net, sock, cmd, argp);
3143

3144 3145 3146 3147
	case SIOCSARP:
	case SIOCGARP:
	case SIOCDARP:
	case SIOCATMARK:
3148 3149 3150
		return sock_do_ioctl(net, sock, cmd, arg);
	}

3151 3152
	return -ENOIOCTLCMD;
}
3153

3154
static long compat_sock_ioctl(struct file *file, unsigned int cmd,
3155
			      unsigned long arg)
3156 3157 3158
{
	struct socket *sock = file->private_data;
	int ret = -ENOIOCTLCMD;
3159 3160 3161 3162 3163
	struct sock *sk;
	struct net *net;

	sk = sock->sk;
	net = sock_net(sk);
3164 3165 3166 3167

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

3168 3169 3170 3171
	if (ret == -ENOIOCTLCMD &&
	    (cmd >= SIOCIWFIRST && cmd <= SIOCIWLAST))
		ret = compat_wext_handle_ioctl(net, cmd, arg);

3172 3173 3174
	if (ret == -ENOIOCTLCMD)
		ret = compat_sock_ioctl_trans(file, sock, cmd, arg);

3175 3176 3177 3178
	return ret;
}
#endif

3179 3180 3181 3182
int kernel_bind(struct socket *sock, struct sockaddr *addr, int addrlen)
{
	return sock->ops->bind(sock, addr, addrlen);
}
3183
EXPORT_SYMBOL(kernel_bind);
3184 3185 3186 3187 3188

int kernel_listen(struct socket *sock, int backlog)
{
	return sock->ops->listen(sock, backlog);
}
3189
EXPORT_SYMBOL(kernel_listen);
3190 3191 3192 3193 3194 3195 3196 3197 3198 3199 3200 3201 3202 3203

int kernel_accept(struct socket *sock, struct socket **newsock, int flags)
{
	struct sock *sk = sock->sk;
	int err;

	err = sock_create_lite(sk->sk_family, sk->sk_type, sk->sk_protocol,
			       newsock);
	if (err < 0)
		goto done;

	err = sock->ops->accept(sock, *newsock, flags);
	if (err < 0) {
		sock_release(*newsock);
3204
		*newsock = NULL;
3205 3206 3207 3208
		goto done;
	}

	(*newsock)->ops = sock->ops;
3209
	__module_get((*newsock)->ops->owner);
3210 3211 3212 3213

done:
	return err;
}
3214
EXPORT_SYMBOL(kernel_accept);
3215 3216

int kernel_connect(struct socket *sock, struct sockaddr *addr, int addrlen,
3217
		   int flags)
3218 3219 3220
{
	return sock->ops->connect(sock, addr, addrlen, flags);
}
3221
EXPORT_SYMBOL(kernel_connect);
3222 3223 3224 3225 3226 3227

int kernel_getsockname(struct socket *sock, struct sockaddr *addr,
			 int *addrlen)
{
	return sock->ops->getname(sock, addr, addrlen, 0);
}
3228
EXPORT_SYMBOL(kernel_getsockname);
3229 3230 3231 3232 3233 3234

int kernel_getpeername(struct socket *sock, struct sockaddr *addr,
			 int *addrlen)
{
	return sock->ops->getname(sock, addr, addrlen, 1);
}
3235
EXPORT_SYMBOL(kernel_getpeername);
3236 3237 3238 3239 3240

int kernel_getsockopt(struct socket *sock, int level, int optname,
			char *optval, int *optlen)
{
	mm_segment_t oldfs = get_fs();
3241 3242
	char __user *uoptval;
	int __user *uoptlen;
3243 3244
	int err;

3245 3246 3247
	uoptval = (char __user __force *) optval;
	uoptlen = (int __user __force *) optlen;

3248 3249
	set_fs(KERNEL_DS);
	if (level == SOL_SOCKET)
3250
		err = sock_getsockopt(sock, level, optname, uoptval, uoptlen);
3251
	else
3252 3253
		err = sock->ops->getsockopt(sock, level, optname, uoptval,
					    uoptlen);
3254 3255 3256
	set_fs(oldfs);
	return err;
}
3257
EXPORT_SYMBOL(kernel_getsockopt);
3258 3259

int kernel_setsockopt(struct socket *sock, int level, int optname,
3260
			char *optval, unsigned int optlen)
3261 3262
{
	mm_segment_t oldfs = get_fs();
3263
	char __user *uoptval;
3264 3265
	int err;

3266 3267
	uoptval = (char __user __force *) optval;

3268 3269
	set_fs(KERNEL_DS);
	if (level == SOL_SOCKET)
3270
		err = sock_setsockopt(sock, level, optname, uoptval, optlen);
3271
	else
3272
		err = sock->ops->setsockopt(sock, level, optname, uoptval,
3273 3274 3275 3276
					    optlen);
	set_fs(oldfs);
	return err;
}
3277
EXPORT_SYMBOL(kernel_setsockopt);
3278 3279 3280 3281 3282 3283 3284 3285 3286

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);
}
3287
EXPORT_SYMBOL(kernel_sendpage);
3288 3289 3290 3291 3292 3293 3294 3295 3296 3297 3298 3299

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;
}
3300
EXPORT_SYMBOL(kernel_sock_ioctl);
3301

3302 3303 3304 3305 3306
int kernel_sock_shutdown(struct socket *sock, enum sock_shutdown_cmd how)
{
	return sock->ops->shutdown(sock, how);
}
EXPORT_SYMBOL(kernel_sock_shutdown);