socket.c 81.7 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

613 614
static inline int __sock_sendmsg_nosec(struct kiocb *iocb, struct socket *sock,
				       struct msghdr *msg, size_t size)
L
Linus Torvalds 已提交
615 616 617 618
{
	return sock->ops->sendmsg(iocb, sock, msg, size);
}

619 620 621 622 623 624 625 626
static inline int __sock_sendmsg(struct kiocb *iocb, struct socket *sock,
				 struct msghdr *msg, size_t size)
{
	int err = security_socket_sendmsg(sock, msg, size);

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

627 628
static int do_sock_sendmsg(struct socket *sock, struct msghdr *msg,
			   size_t size, bool nosec)
L
Linus Torvalds 已提交
629 630 631 632 633
{
	struct kiocb iocb;
	int ret;

	init_sync_kiocb(&iocb, NULL);
634 635
	ret = nosec ? __sock_sendmsg_nosec(&iocb, sock, msg, size) :
		      __sock_sendmsg(&iocb, sock, msg, size);
L
Linus Torvalds 已提交
636 637 638 639
	if (-EIOCBQUEUED == ret)
		ret = wait_on_sync_kiocb(&iocb);
	return ret;
}
640 641 642 643 644

int sock_sendmsg(struct socket *sock, struct msghdr *msg, size_t size)
{
	return do_sock_sendmsg(sock, msg, size, false);
}
645
EXPORT_SYMBOL(sock_sendmsg);
L
Linus Torvalds 已提交
646

647
static int sock_sendmsg_nosec(struct socket *sock, struct msghdr *msg, size_t size)
648
{
649
	return do_sock_sendmsg(sock, msg, size, true);
650 651
}

L
Linus Torvalds 已提交
652 653 654 655 656 657 658 659 660 661 662
int kernel_sendmsg(struct socket *sock, struct msghdr *msg,
		   struct kvec *vec, size_t num, size_t size)
{
	mm_segment_t oldfs = get_fs();
	int result;

	set_fs(KERNEL_DS);
	/*
	 * the following is safe, since for compiler definitions of kvec and
	 * iovec are identical, yielding the same in-core layout and alignment
	 */
A
Al Viro 已提交
663
	iov_iter_init(&msg->msg_iter, WRITE, (struct iovec *)vec, num, size);
L
Linus Torvalds 已提交
664 665 666 667
	result = sock_sendmsg(sock, msg, size);
	set_fs(oldfs);
	return result;
}
668
EXPORT_SYMBOL(kernel_sendmsg);
L
Linus Torvalds 已提交
669

670 671 672 673 674 675
/*
 * 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)
{
676
	int need_software_tstamp = sock_flag(sk, SOCK_RCVTSTAMP);
677
	struct scm_timestamping tss;
678 679 680 681 682 683 684 685 686 687 688 689 690 691 692 693
	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 {
694 695
			struct timespec ts;
			skb_get_timestampns(skb, &ts);
696
			put_cmsg(msg, SOL_SOCKET, SCM_TIMESTAMPNS,
697
				 sizeof(ts), &ts);
698 699 700
		}
	}

701
	memset(&tss, 0, sizeof(tss));
702
	if ((sk->sk_tsflags & SOF_TIMESTAMPING_SOFTWARE) &&
703
	    ktime_to_timespec_cond(skb->tstamp, tss.ts + 0))
704
		empty = 0;
705
	if (shhwtstamps &&
706
	    (sk->sk_tsflags & SOF_TIMESTAMPING_RAW_HARDWARE) &&
707
	    ktime_to_timespec_cond(shhwtstamps->hwtstamp, tss.ts + 2))
708
		empty = 0;
709 710
	if (!empty)
		put_cmsg(msg, SOL_SOCKET,
711
			 SCM_TIMESTAMPING, sizeof(tss), &tss);
712
}
713 714
EXPORT_SYMBOL_GPL(__sock_recv_timestamp);

715 716 717 718 719 720 721 722 723 724 725 726 727 728 729 730
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
731 732
static inline void sock_recv_drops(struct msghdr *msg, struct sock *sk,
				   struct sk_buff *skb)
733
{
734
	if (sock_flag(sk, SOCK_RXQ_OVFL) && skb && SOCK_SKB_CB(skb)->dropcount)
735
		put_cmsg(msg, SOL_SOCKET, SO_RXQ_OVFL,
736
			sizeof(__u32), &SOCK_SKB_CB(skb)->dropcount);
737 738
}

739
void __sock_recv_ts_and_drops(struct msghdr *msg, struct sock *sk,
740 741 742 743 744
	struct sk_buff *skb)
{
	sock_recv_timestamp(msg, sk, skb);
	sock_recv_drops(msg, sk, skb);
}
745
EXPORT_SYMBOL_GPL(__sock_recv_ts_and_drops);
746

747 748
static inline int __sock_recvmsg_nosec(struct kiocb *iocb, struct socket *sock,
				       struct msghdr *msg, size_t size, int flags)
L
Linus Torvalds 已提交
749 750 751 752
{
	return sock->ops->recvmsg(iocb, sock, msg, size, flags);
}

753 754 755 756 757 758 759 760
static inline int __sock_recvmsg(struct kiocb *iocb, struct socket *sock,
				 struct msghdr *msg, size_t size, int flags)
{
	int err = security_socket_recvmsg(sock, msg, size, flags);

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

761
int sock_recvmsg(struct socket *sock, struct msghdr *msg,
L
Linus Torvalds 已提交
762 763 764 765 766
		 size_t size, int flags)
{
	struct kiocb iocb;
	int ret;

767
	init_sync_kiocb(&iocb, NULL);
L
Linus Torvalds 已提交
768 769 770 771 772
	ret = __sock_recvmsg(&iocb, sock, msg, size, flags);
	if (-EIOCBQUEUED == ret)
		ret = wait_on_sync_kiocb(&iocb);
	return ret;
}
773
EXPORT_SYMBOL(sock_recvmsg);
L
Linus Torvalds 已提交
774

775 776 777 778 779 780 781 782 783 784 785 786 787
static int sock_recvmsg_nosec(struct socket *sock, struct msghdr *msg,
			      size_t size, int flags)
{
	struct kiocb iocb;
	int ret;

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

788 789 790 791 792 793 794 795 796 797 798 799 800 801 802
/**
 * 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.
 */
803 804
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
805 806 807 808 809 810 811 812 813
{
	mm_segment_t oldfs = get_fs();
	int result;

	set_fs(KERNEL_DS);
	/*
	 * the following is safe, since for compiler definitions of kvec and
	 * iovec are identical, yielding the same in-core layout and alignment
	 */
A
Al Viro 已提交
814
	iov_iter_init(&msg->msg_iter, READ, (struct iovec *)vec, num, size);
L
Linus Torvalds 已提交
815 816 817 818
	result = sock_recvmsg(sock, msg, size, flags);
	set_fs(oldfs);
	return result;
}
819
EXPORT_SYMBOL(kernel_recvmsg);
L
Linus Torvalds 已提交
820

821 822
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
823 824 825 826
{
	struct socket *sock;
	int flags;

827 828
	sock = file->private_data;

829 830 831
	flags = (file->f_flags & O_NONBLOCK) ? MSG_DONTWAIT : 0;
	/* more is a combination of MSG_MORE and MSG_SENDPAGE_NOTLAST */
	flags |= more;
832

833
	return kernel_sendpage(sock, page, offset, size, flags);
834
}
L
Linus Torvalds 已提交
835

J
Jens Axboe 已提交
836
static ssize_t sock_splice_read(struct file *file, loff_t *ppos,
837
				struct pipe_inode_info *pipe, size_t len,
J
Jens Axboe 已提交
838 839 840 841
				unsigned int flags)
{
	struct socket *sock = file->private_data;

842 843 844
	if (unlikely(!sock->ops->splice_read))
		return -EINVAL;

J
Jens Axboe 已提交
845 846 847
	return sock->ops->splice_read(sock, ppos, pipe, len, flags);
}

848
static ssize_t sock_read_iter(struct kiocb *iocb, struct iov_iter *to)
849
{
850 851
	struct file *file = iocb->ki_filp;
	struct socket *sock = file->private_data;
852 853
	struct msghdr msg = {.msg_iter = *to};
	ssize_t res;
854

855 856 857 858
	if (file->f_flags & O_NONBLOCK)
		msg.msg_flags = MSG_DONTWAIT;

	if (iocb->ki_pos != 0)
L
Linus Torvalds 已提交
859
		return -ESPIPE;
860

K
Kent Overstreet 已提交
861
	if (iocb->ki_nbytes == 0)	/* Match SYS5 behaviour */
L
Linus Torvalds 已提交
862 863
		return 0;

864 865 866 867
	res = __sock_recvmsg(iocb, sock, &msg,
			     iocb->ki_nbytes, msg.msg_flags);
	*to = msg.msg_iter;
	return res;
L
Linus Torvalds 已提交
868 869
}

870
static ssize_t sock_write_iter(struct kiocb *iocb, struct iov_iter *from)
871
{
872 873
	struct file *file = iocb->ki_filp;
	struct socket *sock = file->private_data;
874 875
	struct msghdr msg = {.msg_iter = *from};
	ssize_t res;
L
Linus Torvalds 已提交
876

877
	if (iocb->ki_pos != 0)
878
		return -ESPIPE;
879

880 881 882
	if (file->f_flags & O_NONBLOCK)
		msg.msg_flags = MSG_DONTWAIT;

883 884 885
	if (sock->type == SOCK_SEQPACKET)
		msg.msg_flags |= MSG_EOR;

886 887 888
	res = __sock_sendmsg(iocb, sock, &msg, iocb->ki_nbytes);
	*from = msg.msg_iter;
	return res;
L
Linus Torvalds 已提交
889 890 891 892 893 894 895
}

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

A
Arjan van de Ven 已提交
896
static DEFINE_MUTEX(br_ioctl_mutex);
897
static int (*br_ioctl_hook) (struct net *, unsigned int cmd, void __user *arg);
L
Linus Torvalds 已提交
898

899
void brioctl_set(int (*hook) (struct net *, unsigned int, void __user *))
L
Linus Torvalds 已提交
900
{
A
Arjan van de Ven 已提交
901
	mutex_lock(&br_ioctl_mutex);
L
Linus Torvalds 已提交
902
	br_ioctl_hook = hook;
A
Arjan van de Ven 已提交
903
	mutex_unlock(&br_ioctl_mutex);
L
Linus Torvalds 已提交
904 905 906
}
EXPORT_SYMBOL(brioctl_set);

A
Arjan van de Ven 已提交
907
static DEFINE_MUTEX(vlan_ioctl_mutex);
908
static int (*vlan_ioctl_hook) (struct net *, void __user *arg);
L
Linus Torvalds 已提交
909

910
void vlan_ioctl_set(int (*hook) (struct net *, void __user *))
L
Linus Torvalds 已提交
911
{
A
Arjan van de Ven 已提交
912
	mutex_lock(&vlan_ioctl_mutex);
L
Linus Torvalds 已提交
913
	vlan_ioctl_hook = hook;
A
Arjan van de Ven 已提交
914
	mutex_unlock(&vlan_ioctl_mutex);
L
Linus Torvalds 已提交
915 916 917
}
EXPORT_SYMBOL(vlan_ioctl_set);

A
Arjan van de Ven 已提交
918
static DEFINE_MUTEX(dlci_ioctl_mutex);
919
static int (*dlci_ioctl_hook) (unsigned int, void __user *);
L
Linus Torvalds 已提交
920

921
void dlci_ioctl_set(int (*hook) (unsigned int, void __user *))
L
Linus Torvalds 已提交
922
{
A
Arjan van de Ven 已提交
923
	mutex_lock(&dlci_ioctl_mutex);
L
Linus Torvalds 已提交
924
	dlci_ioctl_hook = hook;
A
Arjan van de Ven 已提交
925
	mutex_unlock(&dlci_ioctl_mutex);
L
Linus Torvalds 已提交
926 927 928
}
EXPORT_SYMBOL(dlci_ioctl_set);

929 930 931 932 933 934 935 936 937 938 939 940 941 942 943 944 945 946
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
947 948 949 950 951 952 953 954
/*
 *	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;
955
	struct sock *sk;
L
Linus Torvalds 已提交
956 957
	void __user *argp = (void __user *)arg;
	int pid, err;
958
	struct net *net;
L
Linus Torvalds 已提交
959

960
	sock = file->private_data;
961
	sk = sock->sk;
962
	net = sock_net(sk);
L
Linus Torvalds 已提交
963
	if (cmd >= SIOCDEVPRIVATE && cmd <= (SIOCDEVPRIVATE + 15)) {
964
		err = dev_ioctl(net, cmd, argp);
L
Linus Torvalds 已提交
965
	} else
J
Johannes Berg 已提交
966
#ifdef CONFIG_WEXT_CORE
L
Linus Torvalds 已提交
967
	if (cmd >= SIOCIWFIRST && cmd <= SIOCIWLAST) {
968
		err = dev_ioctl(net, cmd, argp);
L
Linus Torvalds 已提交
969
	} else
J
Johannes Berg 已提交
970
#endif
971
		switch (cmd) {
L
Linus Torvalds 已提交
972 973 974 975 976
		case FIOSETOWN:
		case SIOCSPGRP:
			err = -EFAULT;
			if (get_user(pid, (int __user *)argp))
				break;
977 978
			f_setown(sock->file, pid, 1);
			err = 0;
L
Linus Torvalds 已提交
979 980 981
			break;
		case FIOGETOWN:
		case SIOCGPGRP:
982
			err = put_user(f_getown(sock->file),
983
				       (int __user *)argp);
L
Linus Torvalds 已提交
984 985 986 987 988 989 990 991 992
			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
993
			mutex_lock(&br_ioctl_mutex);
994
			if (br_ioctl_hook)
995
				err = br_ioctl_hook(net, cmd, argp);
A
Arjan van de Ven 已提交
996
			mutex_unlock(&br_ioctl_mutex);
L
Linus Torvalds 已提交
997 998 999 1000 1001 1002 1003
			break;
		case SIOCGIFVLAN:
		case SIOCSIFVLAN:
			err = -ENOPKG;
			if (!vlan_ioctl_hook)
				request_module("8021q");

A
Arjan van de Ven 已提交
1004
			mutex_lock(&vlan_ioctl_mutex);
L
Linus Torvalds 已提交
1005
			if (vlan_ioctl_hook)
1006
				err = vlan_ioctl_hook(net, argp);
A
Arjan van de Ven 已提交
1007
			mutex_unlock(&vlan_ioctl_mutex);
L
Linus Torvalds 已提交
1008 1009 1010 1011 1012 1013 1014
			break;
		case SIOCADDDLCI:
		case SIOCDELDLCI:
			err = -ENOPKG;
			if (!dlci_ioctl_hook)
				request_module("dlci");

1015 1016
			mutex_lock(&dlci_ioctl_mutex);
			if (dlci_ioctl_hook)
L
Linus Torvalds 已提交
1017
				err = dlci_ioctl_hook(cmd, argp);
1018
			mutex_unlock(&dlci_ioctl_mutex);
L
Linus Torvalds 已提交
1019 1020
			break;
		default:
1021
			err = sock_do_ioctl(net, sock, cmd, arg);
L
Linus Torvalds 已提交
1022
			break;
1023
		}
L
Linus Torvalds 已提交
1024 1025 1026 1027 1028 1029 1030
	return err;
}

int sock_create_lite(int family, int type, int protocol, struct socket **res)
{
	int err;
	struct socket *sock = NULL;
1031

L
Linus Torvalds 已提交
1032 1033 1034 1035 1036 1037 1038 1039 1040 1041 1042
	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
1043 1044 1045 1046
	err = security_socket_post_create(sock, family, type, protocol, 1);
	if (err)
		goto out_release;

L
Linus Torvalds 已提交
1047 1048 1049
out:
	*res = sock;
	return err;
V
Venkat Yekkirala 已提交
1050 1051 1052 1053
out_release:
	sock_release(sock);
	sock = NULL;
	goto out;
L
Linus Torvalds 已提交
1054
}
1055
EXPORT_SYMBOL(sock_create_lite);
L
Linus Torvalds 已提交
1056 1057

/* No kernel lock held - perfect */
1058
static unsigned int sock_poll(struct file *file, poll_table *wait)
L
Linus Torvalds 已提交
1059
{
1060
	unsigned int busy_flag = 0;
L
Linus Torvalds 已提交
1061 1062 1063
	struct socket *sock;

	/*
1064
	 *      We can't return errors to poll, so it's either yes or no.
L
Linus Torvalds 已提交
1065
	 */
1066
	sock = file->private_data;
1067

1068
	if (sk_can_busy_loop(sock->sk)) {
1069
		/* this socket can poll_ll so tell the system call */
1070
		busy_flag = POLL_BUSY_LOOP;
1071 1072

		/* once, only if requested by syscall */
1073 1074
		if (wait && (wait->_key & POLL_BUSY_LOOP))
			sk_busy_loop(sock->sk, 1);
1075 1076
	}

1077
	return busy_flag | sock->ops->poll(file, sock, wait);
L
Linus Torvalds 已提交
1078 1079
}

1080
static int sock_mmap(struct file *file, struct vm_area_struct *vma)
L
Linus Torvalds 已提交
1081
{
1082
	struct socket *sock = file->private_data;
L
Linus Torvalds 已提交
1083 1084 1085 1086

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

1087
static int sock_close(struct inode *inode, struct file *filp)
L
Linus Torvalds 已提交
1088 1089 1090 1091 1092 1093 1094 1095 1096 1097 1098 1099 1100
{
	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)
1101
 *	   or under socket lock
L
Linus Torvalds 已提交
1102 1103 1104 1105
 */

static int sock_fasync(int fd, struct file *filp, int on)
{
1106 1107
	struct socket *sock = filp->private_data;
	struct sock *sk = sock->sk;
1108
	struct socket_wq *wq;
L
Linus Torvalds 已提交
1109

1110
	if (sk == NULL)
L
Linus Torvalds 已提交
1111 1112 1113
		return -EINVAL;

	lock_sock(sk);
1114 1115
	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
1116

1117
	if (!wq->fasync_list)
1118 1119
		sock_reset_flag(sk, SOCK_FASYNC);
	else
E
Eric Dumazet 已提交
1120
		sock_set_flag(sk, SOCK_FASYNC);
L
Linus Torvalds 已提交
1121

1122
	release_sock(sk);
L
Linus Torvalds 已提交
1123 1124 1125
	return 0;
}

1126
/* This function may be called only under socket lock or callback_lock or rcu_lock */
L
Linus Torvalds 已提交
1127 1128 1129

int sock_wake_async(struct socket *sock, int how, int band)
{
1130 1131 1132 1133 1134 1135 1136 1137
	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
1138
		return -1;
1139
	}
1140
	switch (how) {
1141
	case SOCK_WAKE_WAITD:
L
Linus Torvalds 已提交
1142 1143 1144
		if (test_bit(SOCK_ASYNC_WAITDATA, &sock->flags))
			break;
		goto call_kill;
1145
	case SOCK_WAKE_SPACE:
L
Linus Torvalds 已提交
1146 1147 1148
		if (!test_and_clear_bit(SOCK_ASYNC_NOSPACE, &sock->flags))
			break;
		/* fall through */
1149
	case SOCK_WAKE_IO:
1150
call_kill:
1151
		kill_fasync(&wq->fasync_list, SIGIO, band);
L
Linus Torvalds 已提交
1152
		break;
1153
	case SOCK_WAKE_URG:
1154
		kill_fasync(&wq->fasync_list, SIGURG, band);
L
Linus Torvalds 已提交
1155
	}
1156
	rcu_read_unlock();
L
Linus Torvalds 已提交
1157 1158
	return 0;
}
1159
EXPORT_SYMBOL(sock_wake_async);
L
Linus Torvalds 已提交
1160

P
Pavel Emelyanov 已提交
1161
int __sock_create(struct net *net, int family, int type, int protocol,
1162
			 struct socket **res, int kern)
L
Linus Torvalds 已提交
1163 1164 1165
{
	int err;
	struct socket *sock;
1166
	const struct net_proto_family *pf;
L
Linus Torvalds 已提交
1167 1168

	/*
1169
	 *      Check protocol is in range
L
Linus Torvalds 已提交
1170 1171 1172 1173 1174 1175 1176 1177 1178 1179 1180 1181
	 */
	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) {
1182
		static int warned;
L
Linus Torvalds 已提交
1183 1184
		if (!warned) {
			warned = 1;
1185 1186
			pr_info("%s uses obsolete (PF_INET,SOCK_PACKET)\n",
				current->comm);
L
Linus Torvalds 已提交
1187 1188 1189 1190 1191 1192 1193
		}
		family = PF_PACKET;
	}

	err = security_socket_create(family, type, protocol, kern);
	if (err)
		return err;
1194

1195 1196 1197 1198 1199 1200 1201
	/*
	 *	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) {
1202
		net_warn_ratelimited("socket: no more sockets\n");
1203 1204 1205 1206 1207 1208
		return -ENFILE;	/* Not exactly a match, but its the
				   closest posix thing */
	}

	sock->type = type;

1209
#ifdef CONFIG_MODULES
1210 1211 1212
	/* 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
1213 1214 1215
	 * requested real, full-featured networking support upon configuration.
	 * Otherwise module support will break!
	 */
E
Eric Dumazet 已提交
1216
	if (rcu_access_pointer(net_families[family]) == NULL)
1217
		request_module("net-pf-%d", family);
L
Linus Torvalds 已提交
1218 1219
#endif

1220 1221 1222 1223 1224
	rcu_read_lock();
	pf = rcu_dereference(net_families[family]);
	err = -EAFNOSUPPORT;
	if (!pf)
		goto out_release;
L
Linus Torvalds 已提交
1225 1226 1227 1228 1229

	/*
	 * We will call the ->create function, that possibly is in a loadable
	 * module, so we have to bump that loadable module refcnt first.
	 */
1230
	if (!try_module_get(pf->owner))
L
Linus Torvalds 已提交
1231 1232
		goto out_release;

1233 1234 1235
	/* Now protected by module ref count */
	rcu_read_unlock();

1236
	err = pf->create(net, sock, protocol, kern);
1237
	if (err < 0)
L
Linus Torvalds 已提交
1238
		goto out_module_put;
1239

L
Linus Torvalds 已提交
1240 1241 1242 1243
	/*
	 * Now to bump the refcnt of the [loadable] module that owns this
	 * socket at sock_release time we decrement its refcnt.
	 */
1244 1245 1246
	if (!try_module_get(sock->ops->owner))
		goto out_module_busy;

L
Linus Torvalds 已提交
1247 1248 1249 1250
	/*
	 * Now that we're done with the ->create function, the [loadable]
	 * module can have its refcnt decremented
	 */
1251
	module_put(pf->owner);
V
Venkat Yekkirala 已提交
1252 1253
	err = security_socket_post_create(sock, family, type, protocol, kern);
	if (err)
1254
		goto out_sock_release;
1255
	*res = sock;
L
Linus Torvalds 已提交
1256

1257 1258 1259 1260
	return 0;

out_module_busy:
	err = -EAFNOSUPPORT;
L
Linus Torvalds 已提交
1261
out_module_put:
1262 1263 1264
	sock->ops = NULL;
	module_put(pf->owner);
out_sock_release:
L
Linus Torvalds 已提交
1265
	sock_release(sock);
1266 1267 1268 1269 1270
	return err;

out_release:
	rcu_read_unlock();
	goto out_sock_release;
L
Linus Torvalds 已提交
1271
}
P
Pavel Emelyanov 已提交
1272
EXPORT_SYMBOL(__sock_create);
L
Linus Torvalds 已提交
1273 1274 1275

int sock_create(int family, int type, int protocol, struct socket **res)
{
1276
	return __sock_create(current->nsproxy->net_ns, family, type, protocol, res, 0);
L
Linus Torvalds 已提交
1277
}
1278
EXPORT_SYMBOL(sock_create);
L
Linus Torvalds 已提交
1279 1280 1281

int sock_create_kern(int family, int type, int protocol, struct socket **res)
{
1282
	return __sock_create(&init_net, family, type, protocol, res, 1);
L
Linus Torvalds 已提交
1283
}
1284
EXPORT_SYMBOL(sock_create_kern);
L
Linus Torvalds 已提交
1285

1286
SYSCALL_DEFINE3(socket, int, family, int, type, int, protocol)
L
Linus Torvalds 已提交
1287 1288 1289
{
	int retval;
	struct socket *sock;
1290 1291
	int flags;

1292 1293 1294 1295 1296 1297
	/* 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);

1298
	flags = type & ~SOCK_TYPE_MASK;
1299
	if (flags & ~(SOCK_CLOEXEC | SOCK_NONBLOCK))
1300 1301
		return -EINVAL;
	type &= SOCK_TYPE_MASK;
L
Linus Torvalds 已提交
1302

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

L
Linus Torvalds 已提交
1306 1307 1308 1309
	retval = sock_create(family, type, protocol, &sock);
	if (retval < 0)
		goto out;

1310
	retval = sock_map_fd(sock, flags & (O_CLOEXEC | O_NONBLOCK));
L
Linus Torvalds 已提交
1311 1312 1313 1314 1315 1316 1317 1318 1319 1320 1321 1322 1323 1324 1325 1326
	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.
 */

1327 1328
SYSCALL_DEFINE4(socketpair, int, family, int, type, int, protocol,
		int __user *, usockvec)
L
Linus Torvalds 已提交
1329 1330 1331
{
	struct socket *sock1, *sock2;
	int fd1, fd2, err;
A
Al Viro 已提交
1332
	struct file *newfile1, *newfile2;
1333 1334 1335
	int flags;

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

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

L
Linus Torvalds 已提交
1343 1344 1345 1346 1347 1348 1349 1350 1351 1352 1353 1354 1355 1356
	/*
	 * 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);
1357
	if (err < 0)
L
Linus Torvalds 已提交
1358 1359
		goto out_release_both;

1360
	fd1 = get_unused_fd_flags(flags);
1361 1362
	if (unlikely(fd1 < 0)) {
		err = fd1;
A
Al Viro 已提交
1363
		goto out_release_both;
1364
	}
1365

1366
	fd2 = get_unused_fd_flags(flags);
1367 1368
	if (unlikely(fd2 < 0)) {
		err = fd2;
1369
		goto out_put_unused_1;
1370 1371
	}

1372
	newfile1 = sock_alloc_file(sock1, flags, NULL);
1373 1374
	if (unlikely(IS_ERR(newfile1))) {
		err = PTR_ERR(newfile1);
1375
		goto out_put_unused_both;
1376 1377
	}

1378
	newfile2 = sock_alloc_file(sock2, flags, NULL);
1379 1380
	if (IS_ERR(newfile2)) {
		err = PTR_ERR(newfile2);
1381
		goto out_fput_1;
A
Al Viro 已提交
1382 1383
	}

1384 1385 1386 1387 1388 1389 1390 1391
	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
1392
	audit_fd_pair(fd1, fd2);
1393

A
Al Viro 已提交
1394 1395
	fd_install(fd1, newfile1);
	fd_install(fd2, newfile2);
L
Linus Torvalds 已提交
1396 1397 1398 1399
	/* fd1 and fd2 may be already another descriptors.
	 * Not kernel problem.
	 */

1400
	return 0;
L
Linus Torvalds 已提交
1401

1402 1403 1404 1405 1406 1407 1408 1409 1410 1411 1412 1413 1414
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
1415

1416 1417 1418 1419
out_put_unused_both:
	put_unused_fd(fd2);
out_put_unused_1:
	put_unused_fd(fd1);
L
Linus Torvalds 已提交
1420
out_release_both:
1421
	sock_release(sock2);
L
Linus Torvalds 已提交
1422
out_release_1:
1423
	sock_release(sock1);
L
Linus Torvalds 已提交
1424 1425 1426 1427 1428 1429 1430 1431 1432 1433 1434 1435
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).
 */

1436
SYSCALL_DEFINE3(bind, int, fd, struct sockaddr __user *, umyaddr, int, addrlen)
L
Linus Torvalds 已提交
1437 1438
{
	struct socket *sock;
1439
	struct sockaddr_storage address;
1440
	int err, fput_needed;
L
Linus Torvalds 已提交
1441

1442
	sock = sockfd_lookup_light(fd, &err, &fput_needed);
1443
	if (sock) {
1444
		err = move_addr_to_kernel(umyaddr, addrlen, &address);
1445 1446
		if (err >= 0) {
			err = security_socket_bind(sock,
1447
						   (struct sockaddr *)&address,
1448
						   addrlen);
1449 1450
			if (!err)
				err = sock->ops->bind(sock,
1451
						      (struct sockaddr *)
1452
						      &address, addrlen);
L
Linus Torvalds 已提交
1453
		}
1454
		fput_light(sock->file, fput_needed);
1455
	}
L
Linus Torvalds 已提交
1456 1457 1458 1459 1460 1461 1462 1463 1464
	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.
 */

1465
SYSCALL_DEFINE2(listen, int, fd, int, backlog)
L
Linus Torvalds 已提交
1466 1467
{
	struct socket *sock;
1468
	int err, fput_needed;
1469
	int somaxconn;
1470 1471 1472

	sock = sockfd_lookup_light(fd, &err, &fput_needed);
	if (sock) {
1473
		somaxconn = sock_net(sock->sk)->core.sysctl_somaxconn;
1474
		if ((unsigned int)backlog > somaxconn)
1475
			backlog = somaxconn;
L
Linus Torvalds 已提交
1476 1477

		err = security_socket_listen(sock, backlog);
1478 1479
		if (!err)
			err = sock->ops->listen(sock, backlog);
L
Linus Torvalds 已提交
1480

1481
		fput_light(sock->file, fput_needed);
L
Linus Torvalds 已提交
1482 1483 1484 1485 1486 1487 1488 1489 1490 1491 1492 1493 1494 1495 1496 1497
	}
	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.
 */

1498 1499
SYSCALL_DEFINE4(accept4, int, fd, struct sockaddr __user *, upeer_sockaddr,
		int __user *, upeer_addrlen, int, flags)
L
Linus Torvalds 已提交
1500 1501
{
	struct socket *sock, *newsock;
1502
	struct file *newfile;
1503
	int err, len, newfd, fput_needed;
1504
	struct sockaddr_storage address;
L
Linus Torvalds 已提交
1505

1506
	if (flags & ~(SOCK_CLOEXEC | SOCK_NONBLOCK))
U
Ulrich Drepper 已提交
1507 1508 1509 1510 1511
		return -EINVAL;

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

1512
	sock = sockfd_lookup_light(fd, &err, &fput_needed);
L
Linus Torvalds 已提交
1513 1514 1515 1516
	if (!sock)
		goto out;

	err = -ENFILE;
1517 1518
	newsock = sock_alloc();
	if (!newsock)
L
Linus Torvalds 已提交
1519 1520 1521 1522 1523 1524 1525 1526 1527 1528 1529
		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);

1530
	newfd = get_unused_fd_flags(flags);
1531 1532
	if (unlikely(newfd < 0)) {
		err = newfd;
1533 1534
		sock_release(newsock);
		goto out_put;
1535
	}
1536
	newfile = sock_alloc_file(newsock, flags, sock->sk->sk_prot_creator->name);
1537 1538 1539 1540 1541 1542
	if (unlikely(IS_ERR(newfile))) {
		err = PTR_ERR(newfile);
		put_unused_fd(newfd);
		sock_release(newsock);
		goto out_put;
	}
1543

1544 1545
	err = security_socket_accept(sock, newsock);
	if (err)
1546
		goto out_fd;
1547

L
Linus Torvalds 已提交
1548 1549
	err = sock->ops->accept(sock, newsock, sock->file->f_flags);
	if (err < 0)
1550
		goto out_fd;
L
Linus Torvalds 已提交
1551 1552

	if (upeer_sockaddr) {
1553
		if (newsock->ops->getname(newsock, (struct sockaddr *)&address,
1554
					  &len, 2) < 0) {
L
Linus Torvalds 已提交
1555
			err = -ECONNABORTED;
1556
			goto out_fd;
L
Linus Torvalds 已提交
1557
		}
1558
		err = move_addr_to_user(&address,
1559
					len, upeer_sockaddr, upeer_addrlen);
L
Linus Torvalds 已提交
1560
		if (err < 0)
1561
			goto out_fd;
L
Linus Torvalds 已提交
1562 1563 1564 1565
	}

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

1566 1567
	fd_install(newfd, newfile);
	err = newfd;
L
Linus Torvalds 已提交
1568 1569

out_put:
1570
	fput_light(sock->file, fput_needed);
L
Linus Torvalds 已提交
1571 1572
out:
	return err;
1573
out_fd:
1574
	fput(newfile);
1575
	put_unused_fd(newfd);
L
Linus Torvalds 已提交
1576 1577 1578
	goto out_put;
}

1579 1580
SYSCALL_DEFINE3(accept, int, fd, struct sockaddr __user *, upeer_sockaddr,
		int __user *, upeer_addrlen)
U
Ulrich Drepper 已提交
1581
{
U
Ulrich Drepper 已提交
1582
	return sys_accept4(fd, upeer_sockaddr, upeer_addrlen, 0);
U
Ulrich Drepper 已提交
1583 1584
}

L
Linus Torvalds 已提交
1585 1586 1587 1588 1589 1590 1591 1592 1593 1594 1595 1596
/*
 *	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.
 */

1597 1598
SYSCALL_DEFINE3(connect, int, fd, struct sockaddr __user *, uservaddr,
		int, addrlen)
L
Linus Torvalds 已提交
1599 1600
{
	struct socket *sock;
1601
	struct sockaddr_storage address;
1602
	int err, fput_needed;
L
Linus Torvalds 已提交
1603

1604
	sock = sockfd_lookup_light(fd, &err, &fput_needed);
L
Linus Torvalds 已提交
1605 1606
	if (!sock)
		goto out;
1607
	err = move_addr_to_kernel(uservaddr, addrlen, &address);
L
Linus Torvalds 已提交
1608 1609 1610
	if (err < 0)
		goto out_put;

1611
	err =
1612
	    security_socket_connect(sock, (struct sockaddr *)&address, addrlen);
L
Linus Torvalds 已提交
1613 1614 1615
	if (err)
		goto out_put;

1616
	err = sock->ops->connect(sock, (struct sockaddr *)&address, addrlen,
L
Linus Torvalds 已提交
1617 1618
				 sock->file->f_flags);
out_put:
1619
	fput_light(sock->file, fput_needed);
L
Linus Torvalds 已提交
1620 1621 1622 1623 1624 1625 1626 1627 1628
out:
	return err;
}

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

1629 1630
SYSCALL_DEFINE3(getsockname, int, fd, struct sockaddr __user *, usockaddr,
		int __user *, usockaddr_len)
L
Linus Torvalds 已提交
1631 1632
{
	struct socket *sock;
1633
	struct sockaddr_storage address;
1634
	int len, err, fput_needed;
1635

1636
	sock = sockfd_lookup_light(fd, &err, &fput_needed);
L
Linus Torvalds 已提交
1637 1638 1639 1640 1641 1642 1643
	if (!sock)
		goto out;

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

1644
	err = sock->ops->getname(sock, (struct sockaddr *)&address, &len, 0);
L
Linus Torvalds 已提交
1645 1646
	if (err)
		goto out_put;
1647
	err = move_addr_to_user(&address, len, usockaddr, usockaddr_len);
L
Linus Torvalds 已提交
1648 1649

out_put:
1650
	fput_light(sock->file, fput_needed);
L
Linus Torvalds 已提交
1651 1652 1653 1654 1655 1656 1657 1658 1659
out:
	return err;
}

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

1660 1661
SYSCALL_DEFINE3(getpeername, int, fd, struct sockaddr __user *, usockaddr,
		int __user *, usockaddr_len)
L
Linus Torvalds 已提交
1662 1663
{
	struct socket *sock;
1664
	struct sockaddr_storage address;
1665
	int len, err, fput_needed;
L
Linus Torvalds 已提交
1666

1667 1668
	sock = sockfd_lookup_light(fd, &err, &fput_needed);
	if (sock != NULL) {
L
Linus Torvalds 已提交
1669 1670
		err = security_socket_getpeername(sock);
		if (err) {
1671
			fput_light(sock->file, fput_needed);
L
Linus Torvalds 已提交
1672 1673 1674
			return err;
		}

1675
		err =
1676
		    sock->ops->getname(sock, (struct sockaddr *)&address, &len,
1677
				       1);
L
Linus Torvalds 已提交
1678
		if (!err)
1679
			err = move_addr_to_user(&address, len, usockaddr,
1680
						usockaddr_len);
1681
		fput_light(sock->file, fput_needed);
L
Linus Torvalds 已提交
1682 1683 1684 1685 1686 1687 1688 1689 1690 1691
	}
	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.
 */

1692
SYSCALL_DEFINE6(sendto, int, fd, void __user *, buff, size_t, len,
1693
		unsigned int, flags, struct sockaddr __user *, addr,
1694
		int, addr_len)
L
Linus Torvalds 已提交
1695 1696
{
	struct socket *sock;
1697
	struct sockaddr_storage address;
L
Linus Torvalds 已提交
1698 1699 1700
	int err;
	struct msghdr msg;
	struct iovec iov;
1701 1702
	int fput_needed;

1703 1704
	if (len > INT_MAX)
		len = INT_MAX;
1705 1706
	sock = sockfd_lookup_light(fd, &err, &fput_needed);
	if (!sock)
1707
		goto out;
1708

1709 1710 1711
	iov.iov_base = buff;
	iov.iov_len = len;
	msg.msg_name = NULL;
A
Al Viro 已提交
1712
	iov_iter_init(&msg.msg_iter, WRITE, &iov, 1, len);
1713 1714 1715
	msg.msg_control = NULL;
	msg.msg_controllen = 0;
	msg.msg_namelen = 0;
1716
	if (addr) {
1717
		err = move_addr_to_kernel(addr, addr_len, &address);
L
Linus Torvalds 已提交
1718 1719
		if (err < 0)
			goto out_put;
1720
		msg.msg_name = (struct sockaddr *)&address;
1721
		msg.msg_namelen = addr_len;
L
Linus Torvalds 已提交
1722 1723 1724 1725 1726 1727
	}
	if (sock->file->f_flags & O_NONBLOCK)
		flags |= MSG_DONTWAIT;
	msg.msg_flags = flags;
	err = sock_sendmsg(sock, &msg, len);

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

/*
1735
 *	Send a datagram down a socket.
L
Linus Torvalds 已提交
1736 1737
 */

1738
SYSCALL_DEFINE4(send, int, fd, void __user *, buff, size_t, len,
1739
		unsigned int, flags)
L
Linus Torvalds 已提交
1740 1741 1742 1743 1744
{
	return sys_sendto(fd, buff, len, flags, NULL, 0);
}

/*
1745
 *	Receive a frame from the socket and optionally record the address of the
L
Linus Torvalds 已提交
1746 1747 1748 1749
 *	sender. We verify the buffers are writable and if needed move the
 *	sender address from kernel to user space.
 */

1750
SYSCALL_DEFINE6(recvfrom, int, fd, void __user *, ubuf, size_t, size,
1751
		unsigned int, flags, struct sockaddr __user *, addr,
1752
		int __user *, addr_len)
L
Linus Torvalds 已提交
1753 1754 1755 1756
{
	struct socket *sock;
	struct iovec iov;
	struct msghdr msg;
1757
	struct sockaddr_storage address;
1758
	int err, err2;
1759 1760
	int fput_needed;

1761 1762
	if (size > INT_MAX)
		size = INT_MAX;
1763
	sock = sockfd_lookup_light(fd, &err, &fput_needed);
L
Linus Torvalds 已提交
1764
	if (!sock)
1765
		goto out;
L
Linus Torvalds 已提交
1766

1767 1768 1769 1770
	msg.msg_control = NULL;
	msg.msg_controllen = 0;
	iov.iov_len = size;
	iov.iov_base = ubuf;
A
Al Viro 已提交
1771
	iov_iter_init(&msg.msg_iter, READ, &iov, 1, size);
1772 1773 1774 1775
	/* 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
1776 1777
	if (sock->file->f_flags & O_NONBLOCK)
		flags |= MSG_DONTWAIT;
1778
	err = sock_recvmsg(sock, &msg, size, flags);
L
Linus Torvalds 已提交
1779

1780
	if (err >= 0 && addr != NULL) {
1781
		err2 = move_addr_to_user(&address,
1782
					 msg.msg_namelen, addr, addr_len);
1783 1784
		if (err2 < 0)
			err = err2;
L
Linus Torvalds 已提交
1785
	}
1786 1787

	fput_light(sock->file, fput_needed);
1788
out:
L
Linus Torvalds 已提交
1789 1790 1791 1792
	return err;
}

/*
1793
 *	Receive a datagram from a socket.
L
Linus Torvalds 已提交
1794 1795
 */

1796 1797
SYSCALL_DEFINE4(recv, int, fd, void __user *, ubuf, size_t, size,
		unsigned int, flags)
L
Linus Torvalds 已提交
1798 1799 1800 1801 1802 1803 1804 1805 1806
{
	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.
 */

1807 1808
SYSCALL_DEFINE5(setsockopt, int, fd, int, level, int, optname,
		char __user *, optval, int, optlen)
L
Linus Torvalds 已提交
1809
{
1810
	int err, fput_needed;
L
Linus Torvalds 已提交
1811 1812 1813 1814
	struct socket *sock;

	if (optlen < 0)
		return -EINVAL;
1815 1816 1817 1818

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

		if (level == SOL_SOCKET)
1823 1824 1825
			err =
			    sock_setsockopt(sock, level, optname, optval,
					    optlen);
L
Linus Torvalds 已提交
1826
		else
1827 1828 1829
			err =
			    sock->ops->setsockopt(sock, level, optname, optval,
						  optlen);
1830 1831
out_put:
		fput_light(sock->file, fput_needed);
L
Linus Torvalds 已提交
1832 1833 1834 1835 1836 1837 1838 1839 1840
	}
	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.
 */

1841 1842
SYSCALL_DEFINE5(getsockopt, int, fd, int, level, int, optname,
		char __user *, optval, int __user *, optlen)
L
Linus Torvalds 已提交
1843
{
1844
	int err, fput_needed;
L
Linus Torvalds 已提交
1845 1846
	struct socket *sock;

1847 1848
	sock = sockfd_lookup_light(fd, &err, &fput_needed);
	if (sock != NULL) {
1849 1850 1851
		err = security_socket_getsockopt(sock, level, optname);
		if (err)
			goto out_put;
L
Linus Torvalds 已提交
1852 1853

		if (level == SOL_SOCKET)
1854 1855 1856
			err =
			    sock_getsockopt(sock, level, optname, optval,
					    optlen);
L
Linus Torvalds 已提交
1857
		else
1858 1859 1860
			err =
			    sock->ops->getsockopt(sock, level, optname, optval,
						  optlen);
1861 1862
out_put:
		fput_light(sock->file, fput_needed);
L
Linus Torvalds 已提交
1863 1864 1865 1866 1867 1868 1869 1870
	}
	return err;
}

/*
 *	Shutdown a socket.
 */

1871
SYSCALL_DEFINE2(shutdown, int, fd, int, how)
L
Linus Torvalds 已提交
1872
{
1873
	int err, fput_needed;
L
Linus Torvalds 已提交
1874 1875
	struct socket *sock;

1876 1877
	sock = sockfd_lookup_light(fd, &err, &fput_needed);
	if (sock != NULL) {
L
Linus Torvalds 已提交
1878
		err = security_socket_shutdown(sock, how);
1879 1880 1881
		if (!err)
			err = sock->ops->shutdown(sock, how);
		fput_light(sock->file, fput_needed);
L
Linus Torvalds 已提交
1882 1883 1884 1885
	}
	return err;
}

1886
/* A couple of helpful macros for getting the address of the 32/64 bit
L
Linus Torvalds 已提交
1887 1888 1889 1890 1891 1892
 * 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)

1893 1894 1895 1896 1897
struct used_address {
	struct sockaddr_storage name;
	unsigned int name_len;
};

1898 1899 1900 1901
static ssize_t copy_msghdr_from_user(struct msghdr *kmsg,
				     struct user_msghdr __user *umsg,
				     struct sockaddr __user **save_addr,
				     struct iovec **iov)
1902
{
1903 1904
	struct sockaddr __user *uaddr;
	struct iovec __user *uiov;
A
Al Viro 已提交
1905
	size_t nr_segs;
1906 1907 1908 1909 1910 1911
	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
1912
	    __get_user(nr_segs, &umsg->msg_iovlen) ||
1913 1914 1915
	    __get_user(kmsg->msg_control, &umsg->msg_control) ||
	    __get_user(kmsg->msg_controllen, &umsg->msg_controllen) ||
	    __get_user(kmsg->msg_flags, &umsg->msg_flags))
1916
		return -EFAULT;
1917

1918
	if (!uaddr)
1919 1920
		kmsg->msg_namelen = 0;

1921 1922 1923
	if (kmsg->msg_namelen < 0)
		return -EINVAL;

1924
	if (kmsg->msg_namelen > sizeof(struct sockaddr_storage))
1925
		kmsg->msg_namelen = sizeof(struct sockaddr_storage);
1926 1927 1928 1929 1930 1931 1932 1933 1934 1935 1936 1937 1938 1939 1940 1941

	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
1942
	if (nr_segs > UIO_MAXIOV)
1943 1944 1945
		return -EMSGSIZE;

	err = rw_copy_check_uvector(save_addr ? READ : WRITE,
A
Al Viro 已提交
1946
				    uiov, nr_segs,
1947 1948
				    UIO_FASTIOV, *iov, iov);
	if (err >= 0)
A
Al Viro 已提交
1949 1950
		iov_iter_init(&kmsg->msg_iter, save_addr ? READ : WRITE,
			      *iov, nr_segs, err);
1951
	return err;
1952 1953
}

1954
static int ___sys_sendmsg(struct socket *sock, struct user_msghdr __user *msg,
1955
			 struct msghdr *msg_sys, unsigned int flags,
1956
			 struct used_address *used_address)
L
Linus Torvalds 已提交
1957
{
1958 1959
	struct compat_msghdr __user *msg_compat =
	    (struct compat_msghdr __user *)msg;
1960
	struct sockaddr_storage address;
L
Linus Torvalds 已提交
1961
	struct iovec iovstack[UIO_FASTIOV], *iov = iovstack;
1962
	unsigned char ctl[sizeof(struct cmsghdr) + 20]
1963 1964
	    __attribute__ ((aligned(sizeof(__kernel_size_t))));
	/* 20 is size of ipv6_pktinfo */
L
Linus Torvalds 已提交
1965
	unsigned char *ctl_buf = ctl;
1966 1967
	int ctl_len, total_len;
	ssize_t err;
1968

1969
	msg_sys->msg_name = &address;
L
Linus Torvalds 已提交
1970

1971
	if (MSG_CMSG_COMPAT & flags)
1972
		err = get_compat_msghdr(msg_sys, msg_compat, NULL, &iov);
1973
	else
1974
		err = copy_msghdr_from_user(msg_sys, msg, NULL, &iov);
1975
	if (err < 0)
L
Linus Torvalds 已提交
1976 1977 1978 1979 1980
		goto out_freeiov;
	total_len = err;

	err = -ENOBUFS;

1981
	if (msg_sys->msg_controllen > INT_MAX)
L
Linus Torvalds 已提交
1982
		goto out_freeiov;
1983
	ctl_len = msg_sys->msg_controllen;
L
Linus Torvalds 已提交
1984
	if ((MSG_CMSG_COMPAT & flags) && ctl_len) {
1985
		err =
1986
		    cmsghdr_from_user_compat_to_kern(msg_sys, sock->sk, ctl,
1987
						     sizeof(ctl));
L
Linus Torvalds 已提交
1988 1989
		if (err)
			goto out_freeiov;
1990 1991
		ctl_buf = msg_sys->msg_control;
		ctl_len = msg_sys->msg_controllen;
L
Linus Torvalds 已提交
1992
	} else if (ctl_len) {
1993
		if (ctl_len > sizeof(ctl)) {
L
Linus Torvalds 已提交
1994
			ctl_buf = sock_kmalloc(sock->sk, ctl_len, GFP_KERNEL);
1995
			if (ctl_buf == NULL)
L
Linus Torvalds 已提交
1996 1997 1998 1999
				goto out_freeiov;
		}
		err = -EFAULT;
		/*
2000
		 * Careful! Before this, msg_sys->msg_control contains a user pointer.
L
Linus Torvalds 已提交
2001 2002 2003
		 * Afterwards, it will be a kernel pointer. Thus the compiler-assisted
		 * checking falls down on this.
		 */
2004
		if (copy_from_user(ctl_buf,
2005
				   (void __user __force *)msg_sys->msg_control,
2006
				   ctl_len))
L
Linus Torvalds 已提交
2007
			goto out_freectl;
2008
		msg_sys->msg_control = ctl_buf;
L
Linus Torvalds 已提交
2009
	}
2010
	msg_sys->msg_flags = flags;
L
Linus Torvalds 已提交
2011 2012

	if (sock->file->f_flags & O_NONBLOCK)
2013
		msg_sys->msg_flags |= MSG_DONTWAIT;
2014 2015 2016 2017 2018 2019
	/*
	 * 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.
	 */
2020 2021 2022
	if (used_address && msg_sys->msg_name &&
	    used_address->name_len == msg_sys->msg_namelen &&
	    !memcmp(&used_address->name, msg_sys->msg_name,
2023 2024 2025 2026 2027 2028 2029 2030 2031 2032 2033
		    used_address->name_len)) {
		err = sock_sendmsg_nosec(sock, msg_sys, total_len);
		goto out_freectl;
	}
	err = sock_sendmsg(sock, msg_sys, total_len);
	/*
	 * If this is sendmmsg() and sending to current destination address was
	 * successful, remember it.
	 */
	if (used_address && err >= 0) {
		used_address->name_len = msg_sys->msg_namelen;
2034 2035 2036
		if (msg_sys->msg_name)
			memcpy(&used_address->name, msg_sys->msg_name,
			       used_address->name_len);
2037
	}
L
Linus Torvalds 已提交
2038 2039

out_freectl:
2040
	if (ctl_buf != ctl)
L
Linus Torvalds 已提交
2041 2042 2043
		sock_kfree_s(sock->sk, ctl_buf, ctl_len);
out_freeiov:
	if (iov != iovstack)
E
Eric Dumazet 已提交
2044
		kfree(iov);
2045 2046 2047 2048 2049 2050 2051
	return err;
}

/*
 *	BSD sendmsg interface
 */

2052
long __sys_sendmsg(int fd, struct user_msghdr __user *msg, unsigned flags)
2053 2054 2055
{
	int fput_needed, err;
	struct msghdr msg_sys;
2056 2057 2058
	struct socket *sock;

	sock = sockfd_lookup_light(fd, &err, &fput_needed);
2059 2060 2061
	if (!sock)
		goto out;

2062
	err = ___sys_sendmsg(sock, msg, &msg_sys, flags, NULL);
2063

2064
	fput_light(sock->file, fput_needed);
2065
out:
L
Linus Torvalds 已提交
2066 2067 2068
	return err;
}

2069
SYSCALL_DEFINE3(sendmsg, int, fd, struct user_msghdr __user *, msg, unsigned int, flags)
2070 2071 2072 2073 2074 2075
{
	if (flags & MSG_CMSG_COMPAT)
		return -EINVAL;
	return __sys_sendmsg(fd, msg, flags);
}

2076 2077 2078 2079 2080 2081 2082 2083 2084 2085 2086 2087
/*
 *	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;
2088
	struct used_address used_address;
2089

2090 2091
	if (vlen > UIO_MAXIOV)
		vlen = UIO_MAXIOV;
2092 2093 2094 2095 2096 2097 2098

	datagrams = 0;

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

2099
	used_address.name_len = UINT_MAX;
2100 2101
	entry = mmsg;
	compat_entry = (struct compat_mmsghdr __user *)mmsg;
2102
	err = 0;
2103 2104 2105

	while (datagrams < vlen) {
		if (MSG_CMSG_COMPAT & flags) {
2106
			err = ___sys_sendmsg(sock, (struct user_msghdr __user *)compat_entry,
2107
					     &msg_sys, flags, &used_address);
2108 2109 2110 2111 2112
			if (err < 0)
				break;
			err = __put_user(err, &compat_entry->msg_len);
			++compat_entry;
		} else {
2113
			err = ___sys_sendmsg(sock,
2114
					     (struct user_msghdr __user *)entry,
2115
					     &msg_sys, flags, &used_address);
2116 2117 2118 2119 2120 2121 2122 2123 2124 2125 2126 2127 2128
			if (err < 0)
				break;
			err = put_user(err, &entry->msg_len);
			++entry;
		}

		if (err)
			break;
		++datagrams;
	}

	fput_light(sock->file, fput_needed);

2129 2130
	/* We only return an error if no datagrams were able to be sent */
	if (datagrams != 0)
2131 2132 2133 2134 2135 2136 2137 2138
		return datagrams;

	return err;
}

SYSCALL_DEFINE4(sendmmsg, int, fd, struct mmsghdr __user *, mmsg,
		unsigned int, vlen, unsigned int, flags)
{
2139 2140
	if (flags & MSG_CMSG_COMPAT)
		return -EINVAL;
2141 2142 2143
	return __sys_sendmmsg(fd, mmsg, vlen, flags);
}

2144
static int ___sys_recvmsg(struct socket *sock, struct user_msghdr __user *msg,
2145
			 struct msghdr *msg_sys, unsigned int flags, int nosec)
L
Linus Torvalds 已提交
2146
{
2147 2148
	struct compat_msghdr __user *msg_compat =
	    (struct compat_msghdr __user *)msg;
L
Linus Torvalds 已提交
2149
	struct iovec iovstack[UIO_FASTIOV];
2150
	struct iovec *iov = iovstack;
L
Linus Torvalds 已提交
2151
	unsigned long cmsg_ptr;
2152 2153
	int total_len, len;
	ssize_t err;
L
Linus Torvalds 已提交
2154 2155

	/* kernel mode address */
2156
	struct sockaddr_storage addr;
L
Linus Torvalds 已提交
2157 2158 2159

	/* user mode address pointers */
	struct sockaddr __user *uaddr;
2160
	int __user *uaddr_len = COMPAT_NAMELEN(msg);
2161

2162
	msg_sys->msg_name = &addr;
L
Linus Torvalds 已提交
2163

2164
	if (MSG_CMSG_COMPAT & flags)
2165
		err = get_compat_msghdr(msg_sys, msg_compat, &uaddr, &iov);
2166
	else
2167
		err = copy_msghdr_from_user(msg_sys, msg, &uaddr, &iov);
L
Linus Torvalds 已提交
2168 2169
	if (err < 0)
		goto out_freeiov;
2170
	total_len = err;
L
Linus Torvalds 已提交
2171

2172 2173
	cmsg_ptr = (unsigned long)msg_sys->msg_control;
	msg_sys->msg_flags = flags & (MSG_CMSG_CLOEXEC|MSG_CMSG_COMPAT);
2174

2175 2176 2177
	/* We assume all kernel code knows the size of sockaddr_storage */
	msg_sys->msg_namelen = 0;

L
Linus Torvalds 已提交
2178 2179
	if (sock->file->f_flags & O_NONBLOCK)
		flags |= MSG_DONTWAIT;
2180 2181
	err = (nosec ? sock_recvmsg_nosec : sock_recvmsg)(sock, msg_sys,
							  total_len, flags);
L
Linus Torvalds 已提交
2182 2183 2184 2185 2186
	if (err < 0)
		goto out_freeiov;
	len = err;

	if (uaddr != NULL) {
2187
		err = move_addr_to_user(&addr,
2188
					msg_sys->msg_namelen, uaddr,
2189
					uaddr_len);
L
Linus Torvalds 已提交
2190 2191 2192
		if (err < 0)
			goto out_freeiov;
	}
2193
	err = __put_user((msg_sys->msg_flags & ~MSG_CMSG_COMPAT),
2194
			 COMPAT_FLAGS(msg));
L
Linus Torvalds 已提交
2195 2196 2197
	if (err)
		goto out_freeiov;
	if (MSG_CMSG_COMPAT & flags)
2198
		err = __put_user((unsigned long)msg_sys->msg_control - cmsg_ptr,
L
Linus Torvalds 已提交
2199 2200
				 &msg_compat->msg_controllen);
	else
2201
		err = __put_user((unsigned long)msg_sys->msg_control - cmsg_ptr,
L
Linus Torvalds 已提交
2202 2203 2204 2205 2206 2207 2208
				 &msg->msg_controllen);
	if (err)
		goto out_freeiov;
	err = len;

out_freeiov:
	if (iov != iovstack)
E
Eric Dumazet 已提交
2209
		kfree(iov);
2210 2211 2212 2213 2214 2215 2216
	return err;
}

/*
 *	BSD recvmsg interface
 */

2217
long __sys_recvmsg(int fd, struct user_msghdr __user *msg, unsigned flags)
2218 2219 2220
{
	int fput_needed, err;
	struct msghdr msg_sys;
2221 2222 2223
	struct socket *sock;

	sock = sockfd_lookup_light(fd, &err, &fput_needed);
2224 2225 2226
	if (!sock)
		goto out;

2227
	err = ___sys_recvmsg(sock, msg, &msg_sys, flags, 0);
2228

2229
	fput_light(sock->file, fput_needed);
L
Linus Torvalds 已提交
2230 2231 2232 2233
out:
	return err;
}

2234
SYSCALL_DEFINE3(recvmsg, int, fd, struct user_msghdr __user *, msg,
2235 2236 2237 2238 2239 2240 2241
		unsigned int, flags)
{
	if (flags & MSG_CMSG_COMPAT)
		return -EINVAL;
	return __sys_recvmsg(fd, msg, flags);
}

2242 2243 2244 2245 2246 2247 2248 2249 2250 2251
/*
 *     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;
2252
	struct compat_mmsghdr __user *compat_entry;
2253 2254 2255 2256 2257 2258 2259 2260 2261 2262 2263 2264 2265 2266 2267 2268 2269 2270 2271
	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;
2272
	compat_entry = (struct compat_mmsghdr __user *)mmsg;
2273 2274 2275 2276 2277

	while (datagrams < vlen) {
		/*
		 * No need to ask LSM for more than the first datagram.
		 */
2278
		if (MSG_CMSG_COMPAT & flags) {
2279
			err = ___sys_recvmsg(sock, (struct user_msghdr __user *)compat_entry,
2280 2281
					     &msg_sys, flags & ~MSG_WAITFORONE,
					     datagrams);
2282 2283 2284 2285 2286
			if (err < 0)
				break;
			err = __put_user(err, &compat_entry->msg_len);
			++compat_entry;
		} else {
2287
			err = ___sys_recvmsg(sock,
2288
					     (struct user_msghdr __user *)entry,
2289 2290
					     &msg_sys, flags & ~MSG_WAITFORONE,
					     datagrams);
2291 2292 2293 2294 2295 2296
			if (err < 0)
				break;
			err = put_user(err, &entry->msg_len);
			++entry;
		}

2297 2298 2299 2300
		if (err)
			break;
		++datagrams;

2301 2302 2303 2304
		/* MSG_WAITFORONE turns on MSG_DONTWAIT after one packet */
		if (flags & MSG_WAITFORONE)
			flags |= MSG_DONTWAIT;

2305 2306 2307 2308 2309 2310 2311 2312 2313 2314 2315 2316 2317 2318 2319 2320 2321 2322 2323 2324
		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
2325

2326 2327 2328 2329 2330 2331 2332 2333 2334 2335 2336 2337 2338 2339 2340 2341 2342 2343 2344 2345 2346 2347 2348 2349 2350 2351 2352 2353 2354 2355 2356
	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;

2357 2358 2359
	if (flags & MSG_CMSG_COMPAT)
		return -EINVAL;

2360 2361 2362 2363 2364 2365 2366 2367 2368 2369 2370 2371 2372 2373 2374 2375
	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
2376 2377
/* Argument list sizes for sys_socketcall */
#define AL(x) ((x) * sizeof(unsigned long))
2378
static const unsigned char nargs[21] = {
2379 2380 2381
	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),
2382
	AL(4), AL(5), AL(4)
2383 2384
};

L
Linus Torvalds 已提交
2385 2386 2387
#undef AL

/*
2388
 *	System call vectors.
L
Linus Torvalds 已提交
2389 2390 2391
 *
 *	Argument checking cleaned up. Saved 20% in size.
 *  This function doesn't need to set the kernel lock because
2392
 *  it is set by the callees.
L
Linus Torvalds 已提交
2393 2394
 */

2395
SYSCALL_DEFINE2(socketcall, int, call, unsigned long __user *, args)
L
Linus Torvalds 已提交
2396
{
2397
	unsigned long a[AUDITSC_ARGS];
2398
	unsigned long a0, a1;
L
Linus Torvalds 已提交
2399
	int err;
2400
	unsigned int len;
L
Linus Torvalds 已提交
2401

2402
	if (call < 1 || call > SYS_SENDMMSG)
L
Linus Torvalds 已提交
2403 2404
		return -EINVAL;

2405 2406 2407 2408
	len = nargs[call];
	if (len > sizeof(a))
		return -EINVAL;

L
Linus Torvalds 已提交
2409
	/* copy_from_user should be SMP safe. */
2410
	if (copy_from_user(a, args, len))
L
Linus Torvalds 已提交
2411
		return -EFAULT;
2412

2413 2414 2415
	err = audit_socketcall(nargs[call] / sizeof(unsigned long), a);
	if (err)
		return err;
2416

2417 2418 2419 2420 2421 2422 2423 2424 2425 2426 2427 2428 2429 2430 2431 2432 2433
	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
2434 2435
		err = sys_accept4(a0, (struct sockaddr __user *)a1,
				  (int __user *)a[2], 0);
2436 2437 2438 2439 2440 2441 2442 2443 2444 2445 2446 2447 2448 2449 2450 2451 2452 2453 2454 2455 2456 2457 2458 2459 2460 2461 2462 2463 2464 2465 2466 2467 2468 2469 2470 2471 2472 2473 2474 2475 2476
		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:
2477
		err = sys_sendmsg(a0, (struct user_msghdr __user *)a1, a[2]);
2478
		break;
2479 2480 2481
	case SYS_SENDMMSG:
		err = sys_sendmmsg(a0, (struct mmsghdr __user *)a1, a[2], a[3]);
		break;
2482
	case SYS_RECVMSG:
2483
		err = sys_recvmsg(a0, (struct user_msghdr __user *)a1, a[2]);
2484
		break;
2485 2486 2487 2488
	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
2489 2490 2491
	case SYS_ACCEPT4:
		err = sys_accept4(a0, (struct sockaddr __user *)a1,
				  (int __user *)a[2], a[3]);
U
Ulrich Drepper 已提交
2492
		break;
2493 2494 2495
	default:
		err = -EINVAL;
		break;
L
Linus Torvalds 已提交
2496 2497 2498 2499
	}
	return err;
}

2500
#endif				/* __ARCH_WANT_SYS_SOCKETCALL */
L
Linus Torvalds 已提交
2501

2502 2503 2504 2505
/**
 *	sock_register - add a socket protocol handler
 *	@ops: description of protocol
 *
L
Linus Torvalds 已提交
2506 2507
 *	This function is called by a protocol handler that wants to
 *	advertise its address family, and have it linked into the
2508
 *	socket interface. The value ops->family corresponds to the
2509
 *	socket system call protocol family.
L
Linus Torvalds 已提交
2510
 */
2511
int sock_register(const struct net_proto_family *ops)
L
Linus Torvalds 已提交
2512 2513 2514 2515
{
	int err;

	if (ops->family >= NPROTO) {
2516
		pr_crit("protocol %d >= NPROTO(%d)\n", ops->family, NPROTO);
L
Linus Torvalds 已提交
2517 2518
		return -ENOBUFS;
	}
2519 2520

	spin_lock(&net_family_lock);
E
Eric Dumazet 已提交
2521 2522
	if (rcu_dereference_protected(net_families[ops->family],
				      lockdep_is_held(&net_family_lock)))
2523 2524
		err = -EEXIST;
	else {
2525
		rcu_assign_pointer(net_families[ops->family], ops);
L
Linus Torvalds 已提交
2526 2527
		err = 0;
	}
2528 2529
	spin_unlock(&net_family_lock);

2530
	pr_info("NET: Registered protocol family %d\n", ops->family);
L
Linus Torvalds 已提交
2531 2532
	return err;
}
2533
EXPORT_SYMBOL(sock_register);
L
Linus Torvalds 已提交
2534

2535 2536 2537 2538
/**
 *	sock_unregister - remove a protocol handler
 *	@family: protocol family to remove
 *
L
Linus Torvalds 已提交
2539 2540
 *	This function is called by a protocol handler that wants to
 *	remove its address family, and have it unlinked from the
2541 2542 2543 2544 2545 2546
 *	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
2547
 */
2548
void sock_unregister(int family)
L
Linus Torvalds 已提交
2549
{
2550
	BUG_ON(family < 0 || family >= NPROTO);
L
Linus Torvalds 已提交
2551

2552
	spin_lock(&net_family_lock);
2553
	RCU_INIT_POINTER(net_families[family], NULL);
2554 2555 2556 2557
	spin_unlock(&net_family_lock);

	synchronize_rcu();

2558
	pr_info("NET: Unregistered protocol family %d\n", family);
L
Linus Torvalds 已提交
2559
}
2560
EXPORT_SYMBOL(sock_unregister);
L
Linus Torvalds 已提交
2561

2562
static int __init sock_init(void)
L
Linus Torvalds 已提交
2563
{
N
Nick Piggin 已提交
2564
	int err;
2565 2566 2567 2568 2569 2570
	/*
	 *      Initialize the network sysctl infrastructure.
	 */
	err = net_sysctl_init();
	if (err)
		goto out;
N
Nick Piggin 已提交
2571

L
Linus Torvalds 已提交
2572
	/*
2573
	 *      Initialize skbuff SLAB cache
L
Linus Torvalds 已提交
2574 2575 2576 2577
	 */
	skb_init();

	/*
2578
	 *      Initialize the protocols module.
L
Linus Torvalds 已提交
2579 2580 2581
	 */

	init_inodecache();
N
Nick Piggin 已提交
2582 2583 2584 2585

	err = register_filesystem(&sock_fs_type);
	if (err)
		goto out_fs;
L
Linus Torvalds 已提交
2586
	sock_mnt = kern_mount(&sock_fs_type);
N
Nick Piggin 已提交
2587 2588 2589 2590
	if (IS_ERR(sock_mnt)) {
		err = PTR_ERR(sock_mnt);
		goto out_mount;
	}
2591 2592

	/* The real protocol initialization is performed in later initcalls.
L
Linus Torvalds 已提交
2593 2594 2595
	 */

#ifdef CONFIG_NETFILTER
2596 2597 2598
	err = netfilter_init();
	if (err)
		goto out;
L
Linus Torvalds 已提交
2599
#endif
2600

2601
	ptp_classifier_init();
2602

N
Nick Piggin 已提交
2603 2604 2605 2606 2607 2608 2609
out:
	return err;

out_mount:
	unregister_filesystem(&sock_fs_type);
out_fs:
	goto out;
L
Linus Torvalds 已提交
2610 2611
}

2612 2613
core_initcall(sock_init);	/* early initcall */

L
Linus Torvalds 已提交
2614 2615 2616 2617 2618 2619
#ifdef CONFIG_PROC_FS
void socket_seq_show(struct seq_file *seq)
{
	int cpu;
	int counter = 0;

2620
	for_each_possible_cpu(cpu)
2621
	    counter += per_cpu(sockets_in_use, cpu);
L
Linus Torvalds 已提交
2622 2623 2624 2625 2626 2627 2628

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

	seq_printf(seq, "sockets: used %d\n", counter);
}
2629
#endif				/* CONFIG_PROC_FS */
L
Linus Torvalds 已提交
2630

2631
#ifdef CONFIG_COMPAT
2632
static int do_siocgstamp(struct net *net, struct socket *sock,
2633
			 unsigned int cmd, void __user *up)
2634 2635 2636 2637 2638 2639
{
	mm_segment_t old_fs = get_fs();
	struct timeval ktv;
	int err;

	set_fs(KERNEL_DS);
2640
	err = sock_do_ioctl(net, sock, cmd, (unsigned long)&ktv);
2641
	set_fs(old_fs);
2642
	if (!err)
2643
		err = compat_put_timeval(&ktv, up);
2644

2645 2646 2647
	return err;
}

2648
static int do_siocgstampns(struct net *net, struct socket *sock,
2649
			   unsigned int cmd, void __user *up)
2650 2651 2652 2653 2654 2655
{
	mm_segment_t old_fs = get_fs();
	struct timespec kts;
	int err;

	set_fs(KERNEL_DS);
2656
	err = sock_do_ioctl(net, sock, cmd, (unsigned long)&kts);
2657
	set_fs(old_fs);
2658
	if (!err)
2659
		err = compat_put_timespec(&kts, up);
2660

2661 2662 2663
	return err;
}

2664
static int dev_ifname32(struct net *net, struct compat_ifreq __user *uifr32)
2665 2666 2667 2668 2669
{
	struct ifreq __user *uifr;
	int err;

	uifr = compat_alloc_user_space(sizeof(struct ifreq));
2670
	if (copy_in_user(uifr, uifr32, sizeof(struct compat_ifreq)))
2671 2672
		return -EFAULT;

2673
	err = dev_ioctl(net, SIOCGIFNAME, uifr);
2674 2675 2676
	if (err)
		return err;

2677
	if (copy_in_user(uifr32, uifr, sizeof(struct compat_ifreq)))
2678 2679 2680 2681 2682
		return -EFAULT;

	return 0;
}

2683
static int dev_ifconf(struct net *net, struct compat_ifconf __user *uifc32)
2684
{
2685
	struct compat_ifconf ifc32;
2686 2687
	struct ifconf ifc;
	struct ifconf __user *uifc;
2688
	struct compat_ifreq __user *ifr32;
2689 2690 2691 2692
	struct ifreq __user *ifr;
	unsigned int i, j;
	int err;

2693
	if (copy_from_user(&ifc32, uifc32, sizeof(struct compat_ifconf)))
2694 2695
		return -EFAULT;

2696
	memset(&ifc, 0, sizeof(ifc));
2697 2698 2699 2700 2701 2702
	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 {
2703 2704
		size_t len = ((ifc32.ifc_len / sizeof(struct compat_ifreq)) + 1) *
			sizeof(struct ifreq);
2705 2706 2707 2708
		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);
2709
		for (i = 0; i < ifc32.ifc_len; i += sizeof(struct compat_ifreq)) {
2710
			if (copy_in_user(ifr, ifr32, sizeof(struct compat_ifreq)))
2711 2712 2713 2714 2715 2716 2717 2718
				return -EFAULT;
			ifr++;
			ifr32++;
		}
	}
	if (copy_to_user(uifc, &ifc, sizeof(struct ifconf)))
		return -EFAULT;

2719
	err = dev_ioctl(net, SIOCGIFCONF, uifc);
2720 2721 2722 2723 2724 2725 2726 2727 2728
	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;
2729 2730 2731
	     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)))
2732 2733 2734 2735 2736 2737 2738 2739 2740 2741
			return -EFAULT;
		ifr32++;
		ifr++;
	}

	if (ifc32.ifcbuf == 0) {
		/* Translate from 64-bit structure multiple to
		 * a 32-bit one.
		 */
		i = ifc.ifc_len;
2742
		i = ((i / sizeof(struct ifreq)) * sizeof(struct compat_ifreq));
2743 2744 2745 2746
		ifc32.ifc_len = i;
	} else {
		ifc32.ifc_len = i;
	}
2747
	if (copy_to_user(uifc32, &ifc32, sizeof(struct compat_ifconf)))
2748 2749 2750 2751 2752
		return -EFAULT;

	return 0;
}

2753
static int ethtool_ioctl(struct net *net, struct compat_ifreq __user *ifr32)
2754
{
2755 2756 2757 2758
	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;
2759
	struct ifreq __user *ifr;
2760 2761
	u32 rule_cnt = 0, actual_rule_cnt;
	u32 ethcmd;
2762
	u32 data;
2763
	int ret;
2764

2765 2766
	if (get_user(data, &ifr32->ifr_ifru.ifru_data))
		return -EFAULT;
2767

2768 2769 2770
	compat_rxnfc = compat_ptr(data);

	if (get_user(ethcmd, &compat_rxnfc->cmd))
2771 2772
		return -EFAULT;

2773 2774 2775 2776 2777 2778 2779 2780 2781 2782 2783 2784 2785 2786 2787 2788 2789
	/* 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:
2790
	case ETHTOOL_SRXCLSRLINS:
2791 2792 2793 2794 2795 2796 2797 2798 2799
		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
2800
	rxnfc = (void __user *)ifr + ALIGN(sizeof(struct ifreq), 8);
2801 2802

	if (copy_in_user(&ifr->ifr_name, &ifr32->ifr_name, IFNAMSIZ))
2803 2804
		return -EFAULT;

2805 2806
	if (put_user(convert_in ? rxnfc : compat_ptr(data),
		     &ifr->ifr_ifru.ifru_data))
2807 2808
		return -EFAULT;

2809
	if (convert_in) {
2810
		/* We expect there to be holes between fs.m_ext and
2811 2812
		 * fs.ring_cookie and at the end of fs, but nowhere else.
		 */
2813 2814 2815 2816
		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));
2817 2818 2819 2820 2821 2822 2823
		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
2824 2825
				 (void __user *)(&rxnfc->fs.m_ext + 1) -
				 (void __user *)rxnfc) ||
2826 2827
		    copy_in_user(&rxnfc->fs.ring_cookie,
				 &compat_rxnfc->fs.ring_cookie,
S
Stephen Hemminger 已提交
2828 2829
				 (void __user *)(&rxnfc->fs.location + 1) -
				 (void __user *)&rxnfc->fs.ring_cookie) ||
2830 2831 2832 2833 2834 2835 2836 2837 2838 2839 2840
		    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
2841 2842
				 (const void __user *)(&rxnfc->fs.m_ext + 1) -
				 (const void __user *)rxnfc) ||
2843 2844
		    copy_in_user(&compat_rxnfc->fs.ring_cookie,
				 &rxnfc->fs.ring_cookie,
S
Stephen Hemminger 已提交
2845 2846
				 (const void __user *)(&rxnfc->fs.location + 1) -
				 (const void __user *)&rxnfc->fs.ring_cookie) ||
2847 2848 2849 2850 2851 2852 2853 2854 2855 2856 2857 2858 2859 2860 2861 2862 2863 2864 2865 2866 2867 2868 2869 2870 2871
		    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;
2872 2873
}

2874 2875 2876 2877 2878 2879
static int compat_siocwandev(struct net *net, struct compat_ifreq __user *uifr32)
{
	void __user *uptr;
	compat_uptr_t uptr32;
	struct ifreq __user *uifr;

2880
	uifr = compat_alloc_user_space(sizeof(*uifr));
2881 2882 2883 2884 2885 2886 2887 2888 2889 2890 2891 2892 2893 2894
	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);
}

2895 2896
static int bond_ioctl(struct net *net, unsigned int cmd,
			 struct compat_ifreq __user *ifr32)
2897 2898 2899 2900 2901 2902 2903 2904 2905 2906
{
	struct ifreq kifr;
	mm_segment_t old_fs;
	int err;

	switch (cmd) {
	case SIOCBONDENSLAVE:
	case SIOCBONDRELEASE:
	case SIOCBONDSETHWADDR:
	case SIOCBONDCHANGEACTIVE:
2907
		if (copy_from_user(&kifr, ifr32, sizeof(struct compat_ifreq)))
2908 2909 2910
			return -EFAULT;

		old_fs = get_fs();
2911
		set_fs(KERNEL_DS);
2912 2913
		err = dev_ioctl(net, cmd,
				(struct ifreq __user __force *) &kifr);
2914
		set_fs(old_fs);
2915 2916 2917

		return err;
	default:
2918
		return -ENOIOCTLCMD;
2919
	}
2920 2921
}

2922 2923
/* 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,
2924
				 struct compat_ifreq __user *u_ifreq32)
2925 2926 2927 2928 2929 2930 2931 2932 2933
{
	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;
2934
	if (get_user(data32, &u_ifreq32->ifr_ifru.ifru_data))
2935 2936 2937 2938 2939 2940 2941 2942
		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;
2943
	if (put_user(data64, &u_ifreq64->ifr_ifru.ifru_data))
2944 2945
		return -EFAULT;

2946
	return dev_ioctl(net, cmd, u_ifreq64);
2947 2948
}

2949 2950
static int dev_ifsioc(struct net *net, struct socket *sock,
			 unsigned int cmd, struct compat_ifreq __user *uifr32)
2951
{
2952
	struct ifreq __user *uifr;
2953 2954
	int err;

2955 2956 2957 2958 2959 2960
	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);

2961 2962 2963 2964 2965 2966 2967 2968 2969 2970 2971 2972
	if (!err) {
		switch (cmd) {
		case SIOCGIFFLAGS:
		case SIOCGIFMETRIC:
		case SIOCGIFMTU:
		case SIOCGIFMEM:
		case SIOCGIFHWADDR:
		case SIOCGIFINDEX:
		case SIOCGIFADDR:
		case SIOCGIFBRDADDR:
		case SIOCGIFDSTADDR:
		case SIOCGIFNETMASK:
2973
		case SIOCGIFPFLAGS:
2974
		case SIOCGIFTXQLEN:
2975 2976
		case SIOCGMIIPHY:
		case SIOCGMIIREG:
2977
			if (copy_in_user(uifr32, uifr, sizeof(*uifr32)))
2978 2979 2980 2981 2982 2983 2984
				err = -EFAULT;
			break;
		}
	}
	return err;
}

2985 2986 2987 2988 2989 2990 2991 2992 2993 2994
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));
2995 2996 2997 2998 2999 3000
	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);
3001 3002 3003 3004
	if (err)
		return -EFAULT;

	old_fs = get_fs();
3005
	set_fs(KERNEL_DS);
3006
	err = dev_ioctl(net, cmd, (void  __user __force *)&ifr);
3007
	set_fs(old_fs);
3008 3009 3010

	if (cmd == SIOCGIFMAP && !err) {
		err = copy_to_user(uifr32, &ifr, sizeof(ifr.ifr_name));
3011 3012 3013 3014 3015 3016
		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);
3017 3018 3019 3020 3021 3022
		if (err)
			err = -EFAULT;
	}
	return err;
}

3023
struct rtentry32 {
3024
	u32		rt_pad1;
3025 3026 3027
	struct sockaddr rt_dst;         /* target address               */
	struct sockaddr rt_gateway;     /* gateway addr (RTF_GATEWAY)   */
	struct sockaddr rt_genmask;     /* target network mask (IP)     */
3028 3029 3030 3031 3032 3033 3034
	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! */
3035
	/* char * */ u32 rt_dev;        /* forcing the device at add    */
3036 3037
	u32		rt_mtu;         /* per route MTU/Window         */
	u32		rt_window;      /* Window clamping              */
3038 3039 3040 3041 3042 3043 3044 3045 3046 3047 3048 3049 3050 3051 3052 3053
	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;
};

3054 3055
static int routing_ioctl(struct net *net, struct socket *sock,
			 unsigned int cmd, void __user *argp)
3056 3057 3058 3059 3060 3061 3062 3063 3064
{
	int ret;
	void *r = NULL;
	struct in6_rtmsg r6;
	struct rtentry r4;
	char devname[16];
	u32 rtdev;
	mm_segment_t old_fs = get_fs();

3065 3066
	if (sock && sock->sk && sock->sk->sk_family == AF_INET6) { /* ipv6 */
		struct in6_rtmsg32 __user *ur6 = argp;
3067
		ret = copy_from_user(&r6.rtmsg_dst, &(ur6->rtmsg_dst),
3068
			3 * sizeof(struct in6_addr));
3069 3070 3071 3072 3073 3074 3075
		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));
3076 3077 3078

		r = (void *) &r6;
	} else { /* ipv4 */
3079
		struct rtentry32 __user *ur4 = argp;
3080
		ret = copy_from_user(&r4.rt_dst, &(ur4->rt_dst),
3081
					3 * sizeof(struct sockaddr));
3082 3083 3084 3085 3086 3087
		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));
3088
		if (rtdev) {
3089
			ret |= copy_from_user(devname, compat_ptr(rtdev), 15);
3090 3091
			r4.rt_dev = (char __user __force *)devname;
			devname[15] = 0;
3092 3093 3094 3095 3096 3097 3098 3099 3100 3101 3102
		} else
			r4.rt_dev = NULL;

		r = (void *) &r4;
	}

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

3103
	set_fs(KERNEL_DS);
3104
	ret = sock_do_ioctl(net, sock, cmd, (unsigned long) r);
3105
	set_fs(old_fs);
3106 3107 3108 3109 3110 3111 3112

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
3113
 * use compatible ioctls
3114
 */
3115
static int old_bridge_ioctl(compat_ulong_t __user *argp)
3116
{
3117
	compat_ulong_t tmp;
3118

3119
	if (get_user(tmp, argp))
3120 3121 3122 3123 3124 3125
		return -EFAULT;
	if (tmp == BRCTL_GET_VERSION)
		return BRCTL_VERSION + 1;
	return -EINVAL;
}

3126 3127 3128 3129 3130 3131
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);
3132

3133
	if (cmd >= SIOCDEVPRIVATE && cmd <= (SIOCDEVPRIVATE + 15))
3134
		return compat_ifr_data_ioctl(net, cmd, argp);
3135 3136 3137 3138 3139 3140 3141 3142 3143 3144 3145

	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);
3146 3147
	case SIOCWANDEV:
		return compat_siocwandev(net, argp);
3148 3149 3150
	case SIOCGIFMAP:
	case SIOCSIFMAP:
		return compat_sioc_ifmap(net, cmd, argp);
3151 3152 3153 3154 3155 3156 3157 3158 3159 3160 3161 3162
	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);
3163 3164
	case SIOCBONDSLAVEINFOQUERY:
	case SIOCBONDINFOQUERY:
3165
	case SIOCSHWTSTAMP:
3166
	case SIOCGHWTSTAMP:
3167
		return compat_ifr_data_ioctl(net, cmd, argp);
3168 3169 3170 3171 3172 3173 3174 3175 3176 3177 3178 3179 3180 3181 3182 3183 3184 3185 3186 3187 3188 3189 3190 3191 3192 3193 3194 3195 3196 3197 3198 3199 3200 3201 3202 3203 3204 3205 3206 3207 3208 3209

	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:
3210 3211 3212 3213
	case SIOCSIFNAME:
	case SIOCGMIIPHY:
	case SIOCGMIIREG:
	case SIOCSMIIREG:
3214
		return dev_ifsioc(net, sock, cmd, argp);
3215

3216 3217 3218 3219
	case SIOCSARP:
	case SIOCGARP:
	case SIOCDARP:
	case SIOCATMARK:
3220 3221 3222
		return sock_do_ioctl(net, sock, cmd, arg);
	}

3223 3224
	return -ENOIOCTLCMD;
}
3225

3226
static long compat_sock_ioctl(struct file *file, unsigned int cmd,
3227
			      unsigned long arg)
3228 3229 3230
{
	struct socket *sock = file->private_data;
	int ret = -ENOIOCTLCMD;
3231 3232 3233 3234 3235
	struct sock *sk;
	struct net *net;

	sk = sock->sk;
	net = sock_net(sk);
3236 3237 3238 3239

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

3240 3241 3242 3243
	if (ret == -ENOIOCTLCMD &&
	    (cmd >= SIOCIWFIRST && cmd <= SIOCIWLAST))
		ret = compat_wext_handle_ioctl(net, cmd, arg);

3244 3245 3246
	if (ret == -ENOIOCTLCMD)
		ret = compat_sock_ioctl_trans(file, sock, cmd, arg);

3247 3248 3249 3250
	return ret;
}
#endif

3251 3252 3253 3254
int kernel_bind(struct socket *sock, struct sockaddr *addr, int addrlen)
{
	return sock->ops->bind(sock, addr, addrlen);
}
3255
EXPORT_SYMBOL(kernel_bind);
3256 3257 3258 3259 3260

int kernel_listen(struct socket *sock, int backlog)
{
	return sock->ops->listen(sock, backlog);
}
3261
EXPORT_SYMBOL(kernel_listen);
3262 3263 3264 3265 3266 3267 3268 3269 3270 3271 3272 3273 3274 3275

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);
3276
		*newsock = NULL;
3277 3278 3279 3280
		goto done;
	}

	(*newsock)->ops = sock->ops;
3281
	__module_get((*newsock)->ops->owner);
3282 3283 3284 3285

done:
	return err;
}
3286
EXPORT_SYMBOL(kernel_accept);
3287 3288

int kernel_connect(struct socket *sock, struct sockaddr *addr, int addrlen,
3289
		   int flags)
3290 3291 3292
{
	return sock->ops->connect(sock, addr, addrlen, flags);
}
3293
EXPORT_SYMBOL(kernel_connect);
3294 3295 3296 3297 3298 3299

int kernel_getsockname(struct socket *sock, struct sockaddr *addr,
			 int *addrlen)
{
	return sock->ops->getname(sock, addr, addrlen, 0);
}
3300
EXPORT_SYMBOL(kernel_getsockname);
3301 3302 3303 3304 3305 3306

int kernel_getpeername(struct socket *sock, struct sockaddr *addr,
			 int *addrlen)
{
	return sock->ops->getname(sock, addr, addrlen, 1);
}
3307
EXPORT_SYMBOL(kernel_getpeername);
3308 3309 3310 3311 3312

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

3317 3318 3319
	uoptval = (char __user __force *) optval;
	uoptlen = (int __user __force *) optlen;

3320 3321
	set_fs(KERNEL_DS);
	if (level == SOL_SOCKET)
3322
		err = sock_getsockopt(sock, level, optname, uoptval, uoptlen);
3323
	else
3324 3325
		err = sock->ops->getsockopt(sock, level, optname, uoptval,
					    uoptlen);
3326 3327 3328
	set_fs(oldfs);
	return err;
}
3329
EXPORT_SYMBOL(kernel_getsockopt);
3330 3331

int kernel_setsockopt(struct socket *sock, int level, int optname,
3332
			char *optval, unsigned int optlen)
3333 3334
{
	mm_segment_t oldfs = get_fs();
3335
	char __user *uoptval;
3336 3337
	int err;

3338 3339
	uoptval = (char __user __force *) optval;

3340 3341
	set_fs(KERNEL_DS);
	if (level == SOL_SOCKET)
3342
		err = sock_setsockopt(sock, level, optname, uoptval, optlen);
3343
	else
3344
		err = sock->ops->setsockopt(sock, level, optname, uoptval,
3345 3346 3347 3348
					    optlen);
	set_fs(oldfs);
	return err;
}
3349
EXPORT_SYMBOL(kernel_setsockopt);
3350 3351 3352 3353 3354 3355 3356 3357 3358

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);
}
3359
EXPORT_SYMBOL(kernel_sendpage);
3360 3361 3362 3363 3364 3365 3366 3367 3368 3369 3370 3371

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;
}
3372
EXPORT_SYMBOL(kernel_sock_ioctl);
3373

3374 3375 3376 3377 3378
int kernel_sock_shutdown(struct socket *sock, enum sock_shutdown_cmd how)
{
	return sock->ops->shutdown(sock, how);
}
EXPORT_SYMBOL(kernel_sock_shutdown);