“d6bfde8765668c8687de72f7a40f52acdf4f2f19”上不存在“README”
socket.c 79.7 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>
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>
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;
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);
131
static ssize_t sock_splice_read(struct file *file, loff_t *ppos,
132
				struct pipe_inode_info *pipe, size_t len,
133
				unsigned int flags);
L
Linus Torvalds 已提交
134 135 136 137 138 139

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

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

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

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

/*
 *	Statistics counters of the socket lists
 */

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

/*
172 173 174
 * Support routines.
 * Move socket addresses back and forth across the kernel/user
 * divide and look after the messy bits.
L
Linus Torvalds 已提交
175 176 177 178 179 180 181 182 183 184 185 186 187
 */

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

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

/**
 *	move_addr_to_user	-	copy an address to user space
 *	@kaddr: kernel space address
 *	@klen: length of address in kernel
 *	@uaddr: user space address
 *	@ulen: pointer to user length field
 *
 *	The value pointed to by ulen on entry is the buffer length available.
 *	This is overwritten with the buffer space used. -EINVAL is returned
 *	if an overlong buffer is specified or a negative buffer size. -EFAULT
 *	is returned if either the buffer or the length field are not
 *	accessible.
 *	After copying the data up to the limit the user specifies, the true
 *	length of the data is written over the length limit the user
 *	specified. Zero is returned for a success.
 */
215

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

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

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

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

250
	ei = kmem_cache_alloc(sock_inode_cachep, GFP_KERNEL);
L
Linus Torvalds 已提交
251 252
	if (!ei)
		return NULL;
253 254
	wq = kmalloc(sizeof(*wq), GFP_KERNEL);
	if (!wq) {
255 256 257
		kmem_cache_free(sock_inode_cachep, ei);
		return NULL;
	}
258 259 260
	init_waitqueue_head(&wq->wait);
	wq->fasync_list = NULL;
	RCU_INIT_POINTER(ei->socket.wq, wq);
261

L
Linus Torvalds 已提交
262 263 264 265 266 267 268 269 270 271 272
	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)
{
273
	struct socket_alloc *ei;
274
	struct socket_wq *wq;
275 276

	ei = container_of(inode, struct socket_alloc, vfs_inode);
277
	wq = rcu_dereference_protected(ei->socket.wq, 1);
278
	kfree_rcu(wq, rcu);
279
	kmem_cache_free(sock_inode_cachep, ei);
L
Linus Torvalds 已提交
280 281
}

282
static void init_once(void *foo)
L
Linus Torvalds 已提交
283
{
284
	struct socket_alloc *ei = (struct socket_alloc *)foo;
L
Linus Torvalds 已提交
285

286
	inode_init_once(&ei->vfs_inode);
L
Linus Torvalds 已提交
287
}
288

L
Linus Torvalds 已提交
289 290 291
static int init_inodecache(void)
{
	sock_inode_cachep = kmem_cache_create("sock_inode_cache",
292 293 294 295 296
					      sizeof(struct socket_alloc),
					      0,
					      (SLAB_HWCACHE_ALIGN |
					       SLAB_RECLAIM_ACCOUNT |
					       SLAB_MEM_SPREAD),
297
					      init_once);
L
Linus Torvalds 已提交
298 299 300 301 302
	if (sock_inode_cachep == NULL)
		return -ENOMEM;
	return 0;
}

303
static const struct super_operations sockfs_ops = {
304 305 306
	.alloc_inode	= sock_alloc_inode,
	.destroy_inode	= sock_destroy_inode,
	.statfs		= simple_statfs,
L
Linus Torvalds 已提交
307 308
};

309 310 311 312 313 314
/*
 * 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]",
315
				d_inode(dentry)->i_ino);
316 317
}

318
static const struct dentry_operations sockfs_dentry_operations = {
319
	.d_dname  = sockfs_dname,
L
Linus Torvalds 已提交
320 321
};

322 323 324 325 326 327 328 329 330 331 332 333 334 335 336
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
337 338 339
/*
 *	Obtains the first available file descriptor and sets it up for use.
 *
340 341
 *	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
342 343 344 345 346 347 348 349 350 351 352 353
 *	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.
 */

354
struct file *sock_alloc_file(struct socket *sock, int flags, const char *dname)
L
Linus Torvalds 已提交
355
{
356
	struct qstr name = { .name = "" };
357
	struct path path;
358
	struct file *file;
L
Linus Torvalds 已提交
359

360 361 362 363 364 365 366
	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);
	}
367
	path.dentry = d_alloc_pseudo(sock_mnt->mnt_sb, &name);
368 369
	if (unlikely(!path.dentry))
		return ERR_PTR(-ENOMEM);
370
	path.mnt = mntget(sock_mnt);
371

372
	d_instantiate(path.dentry, SOCK_INODE(sock));
373

374
	file = alloc_file(&path, FMODE_READ | FMODE_WRITE,
375
		  &socket_file_ops);
376
	if (unlikely(IS_ERR(file))) {
377
		/* drop dentry, keep inode */
378
		ihold(d_inode(path.dentry));
379
		path_put(&path);
380
		return file;
381 382 383
	}

	sock->file = file;
384
	file->f_flags = O_RDWR | (flags & O_NONBLOCK);
385
	file->private_data = sock;
386
	return file;
387
}
388
EXPORT_SYMBOL(sock_alloc_file);
389

390
static int sock_map_fd(struct socket *sock, int flags)
391 392
{
	struct file *newfile;
393 394 395
	int fd = get_unused_fd_flags(flags);
	if (unlikely(fd < 0))
		return fd;
396

397
	newfile = sock_alloc_file(sock, flags, NULL);
398
	if (likely(!IS_ERR(newfile))) {
399
		fd_install(fd, newfile);
400 401
		return fd;
	}
402

403 404
	put_unused_fd(fd);
	return PTR_ERR(newfile);
L
Linus Torvalds 已提交
405 406
}

407
struct socket *sock_from_file(struct file *file, int *err)
408 409 410 411
{
	if (file->f_op == &socket_file_ops)
		return file->private_data;	/* set in sock_map_fd */

412 413
	*err = -ENOTSOCK;
	return NULL;
414
}
415
EXPORT_SYMBOL(sock_from_file);
416

L
Linus Torvalds 已提交
417
/**
418
 *	sockfd_lookup - Go from a file number to its socket slot
L
Linus Torvalds 已提交
419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434
 *	@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;

435 436
	file = fget(fd);
	if (!file) {
L
Linus Torvalds 已提交
437 438 439
		*err = -EBADF;
		return NULL;
	}
440

441 442
	sock = sock_from_file(file, err);
	if (!sock)
L
Linus Torvalds 已提交
443
		fput(file);
444 445
	return sock;
}
446
EXPORT_SYMBOL(sockfd_lookup);
L
Linus Torvalds 已提交
447

448 449
static struct socket *sockfd_lookup_light(int fd, int *err, int *fput_needed)
{
450
	struct fd f = fdget(fd);
451 452
	struct socket *sock;

453
	*err = -EBADF;
454 455 456 457
	if (f.file) {
		sock = sock_from_file(f.file, err);
		if (likely(sock)) {
			*fput_needed = f.flags;
458
			return sock;
459 460
		}
		fdput(f);
L
Linus Torvalds 已提交
461
	}
462
	return NULL;
L
Linus Torvalds 已提交
463 464
}

465 466 467 468 469 470 471 472 473 474 475 476 477 478 479 480 481 482 483 484 485 486 487 488 489 490 491 492 493 494 495 496 497 498 499
#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;

500
	len = security_inode_listsecurity(d_inode(dentry), buffer, size);
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
	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
527 528
/**
 *	sock_alloc	-	allocate a socket
529
 *
L
Linus Torvalds 已提交
530 531 532 533 534 535 536
 *	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)
{
537 538
	struct inode *inode;
	struct socket *sock;
L
Linus Torvalds 已提交
539

540
	inode = new_inode_pseudo(sock_mnt->mnt_sb);
L
Linus Torvalds 已提交
541 542 543 544 545
	if (!inode)
		return NULL;

	sock = SOCKET_I(inode);

546
	kmemcheck_annotate_bitfield(sock, type);
547
	inode->i_ino = get_next_ino();
548
	inode->i_mode = S_IFSOCK | S_IRWXUGO;
549 550
	inode->i_uid = current_fsuid();
	inode->i_gid = current_fsgid();
551
	inode->i_op = &sockfs_inode_ops;
L
Linus Torvalds 已提交
552

553
	this_cpu_add(sockets_in_use, 1);
L
Linus Torvalds 已提交
554 555 556 557 558 559 560 561 562
	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
563
 *	an inode not a file.
L
Linus Torvalds 已提交
564
 */
565

L
Linus Torvalds 已提交
566 567 568 569 570 571 572 573 574 575
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);
	}

576
	if (rcu_dereference_protected(sock->wq, 1)->fasync_list)
577
		pr_err("%s: fasync list not empty!\n", __func__);
L
Linus Torvalds 已提交
578

579
	this_cpu_sub(sockets_in_use, 1);
L
Linus Torvalds 已提交
580 581 582 583
	if (!sock->file) {
		iput(SOCK_INODE(sock));
		return;
	}
584
	sock->file = NULL;
L
Linus Torvalds 已提交
585
}
586
EXPORT_SYMBOL(sock_release);
L
Linus Torvalds 已提交
587

588
void __sock_tx_timestamp(const struct sock *sk, __u8 *tx_flags)
589
{
590 591
	u8 flags = *tx_flags;

592
	if (sk->sk_tsflags & SOF_TIMESTAMPING_TX_HARDWARE)
593 594
		flags |= SKBTX_HW_TSTAMP;

595
	if (sk->sk_tsflags & SOF_TIMESTAMPING_TX_SOFTWARE)
596 597
		flags |= SKBTX_SW_TSTAMP;

598
	if (sk->sk_tsflags & SOF_TIMESTAMPING_TX_SCHED)
599 600
		flags |= SKBTX_SCHED_TSTAMP;

601
	if (sk->sk_tsflags & SOF_TIMESTAMPING_TX_ACK)
602
		flags |= SKBTX_ACK_TSTAMP;
603

604
	*tx_flags = flags;
605
}
606
EXPORT_SYMBOL(__sock_tx_timestamp);
607

608
static inline int sock_sendmsg_nosec(struct socket *sock, struct msghdr *msg)
L
Linus Torvalds 已提交
609
{
A
Al Viro 已提交
610
	int ret = sock->ops->sendmsg(sock, msg, msg_data_left(msg));
611 612
	BUG_ON(ret == -EIOCBQUEUED);
	return ret;
L
Linus Torvalds 已提交
613 614
}

615
int sock_sendmsg(struct socket *sock, struct msghdr *msg)
616
{
617
	int err = security_socket_sendmsg(sock, msg,
A
Al Viro 已提交
618
					  msg_data_left(msg));
619

620
	return err ?: sock_sendmsg_nosec(sock, msg);
621
}
622
EXPORT_SYMBOL(sock_sendmsg);
L
Linus Torvalds 已提交
623 624 625 626

int kernel_sendmsg(struct socket *sock, struct msghdr *msg,
		   struct kvec *vec, size_t num, size_t size)
{
627
	iov_iter_kvec(&msg->msg_iter, WRITE | ITER_KVEC, vec, num, size);
628
	return sock_sendmsg(sock, msg);
L
Linus Torvalds 已提交
629
}
630
EXPORT_SYMBOL(kernel_sendmsg);
L
Linus Torvalds 已提交
631

632 633 634 635 636 637
/*
 * 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)
{
638
	int need_software_tstamp = sock_flag(sk, SOCK_RCVTSTAMP);
639
	struct scm_timestamping tss;
640 641 642 643 644 645 646 647 648 649 650 651 652 653 654 655
	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 {
656 657
			struct timespec ts;
			skb_get_timestampns(skb, &ts);
658
			put_cmsg(msg, SOL_SOCKET, SCM_TIMESTAMPNS,
659
				 sizeof(ts), &ts);
660 661 662
		}
	}

663
	memset(&tss, 0, sizeof(tss));
664
	if ((sk->sk_tsflags & SOF_TIMESTAMPING_SOFTWARE) &&
665
	    ktime_to_timespec_cond(skb->tstamp, tss.ts + 0))
666
		empty = 0;
667
	if (shhwtstamps &&
668
	    (sk->sk_tsflags & SOF_TIMESTAMPING_RAW_HARDWARE) &&
669
	    ktime_to_timespec_cond(shhwtstamps->hwtstamp, tss.ts + 2))
670
		empty = 0;
671 672
	if (!empty)
		put_cmsg(msg, SOL_SOCKET,
673
			 SCM_TIMESTAMPING, sizeof(tss), &tss);
674
}
675 676
EXPORT_SYMBOL_GPL(__sock_recv_timestamp);

677 678 679 680 681 682 683 684 685 686 687 688 689 690 691 692
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);

693 694
static inline void sock_recv_drops(struct msghdr *msg, struct sock *sk,
				   struct sk_buff *skb)
695
{
696
	if (sock_flag(sk, SOCK_RXQ_OVFL) && skb && SOCK_SKB_CB(skb)->dropcount)
697
		put_cmsg(msg, SOL_SOCKET, SO_RXQ_OVFL,
698
			sizeof(__u32), &SOCK_SKB_CB(skb)->dropcount);
699 700
}

701
void __sock_recv_ts_and_drops(struct msghdr *msg, struct sock *sk,
702 703 704 705 706
	struct sk_buff *skb)
{
	sock_recv_timestamp(msg, sk, skb);
	sock_recv_drops(msg, sk, skb);
}
707
EXPORT_SYMBOL_GPL(__sock_recv_ts_and_drops);
708

709 710
static inline int sock_recvmsg_nosec(struct socket *sock, struct msghdr *msg,
				     size_t size, int flags)
L
Linus Torvalds 已提交
711
{
712
	return sock->ops->recvmsg(sock, msg, size, flags);
L
Linus Torvalds 已提交
713 714
}

715 716
int sock_recvmsg(struct socket *sock, struct msghdr *msg, size_t size,
		 int flags)
717 718 719
{
	int err = security_socket_recvmsg(sock, msg, size, flags);

720
	return err ?: sock_recvmsg_nosec(sock, msg, size, flags);
L
Linus Torvalds 已提交
721
}
722
EXPORT_SYMBOL(sock_recvmsg);
L
Linus Torvalds 已提交
723

724 725 726 727 728 729 730 731 732 733 734 735 736 737 738
/**
 * 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.
 */
739 740
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
741 742 743 744
{
	mm_segment_t oldfs = get_fs();
	int result;

745
	iov_iter_kvec(&msg->msg_iter, READ | ITER_KVEC, vec, num, size);
L
Linus Torvalds 已提交
746 747 748 749 750
	set_fs(KERNEL_DS);
	result = sock_recvmsg(sock, msg, size, flags);
	set_fs(oldfs);
	return result;
}
751
EXPORT_SYMBOL(kernel_recvmsg);
L
Linus Torvalds 已提交
752

753 754
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
755 756 757 758
{
	struct socket *sock;
	int flags;

759 760
	sock = file->private_data;

761 762 763
	flags = (file->f_flags & O_NONBLOCK) ? MSG_DONTWAIT : 0;
	/* more is a combination of MSG_MORE and MSG_SENDPAGE_NOTLAST */
	flags |= more;
764

765
	return kernel_sendpage(sock, page, offset, size, flags);
766
}
L
Linus Torvalds 已提交
767

768
static ssize_t sock_splice_read(struct file *file, loff_t *ppos,
769
				struct pipe_inode_info *pipe, size_t len,
770 771 772 773
				unsigned int flags)
{
	struct socket *sock = file->private_data;

774 775 776
	if (unlikely(!sock->ops->splice_read))
		return -EINVAL;

777 778 779
	return sock->ops->splice_read(sock, ppos, pipe, len, flags);
}

780
static ssize_t sock_read_iter(struct kiocb *iocb, struct iov_iter *to)
781
{
782 783
	struct file *file = iocb->ki_filp;
	struct socket *sock = file->private_data;
784 785
	struct msghdr msg = {.msg_iter = *to,
			     .msg_iocb = iocb};
786
	ssize_t res;
787

788 789 790 791
	if (file->f_flags & O_NONBLOCK)
		msg.msg_flags = MSG_DONTWAIT;

	if (iocb->ki_pos != 0)
L
Linus Torvalds 已提交
792
		return -ESPIPE;
793

794
	if (!iov_iter_count(to))	/* Match SYS5 behaviour */
L
Linus Torvalds 已提交
795 796
		return 0;

797
	res = sock_recvmsg(sock, &msg, iov_iter_count(to), msg.msg_flags);
798 799
	*to = msg.msg_iter;
	return res;
L
Linus Torvalds 已提交
800 801
}

802
static ssize_t sock_write_iter(struct kiocb *iocb, struct iov_iter *from)
803
{
804 805
	struct file *file = iocb->ki_filp;
	struct socket *sock = file->private_data;
806 807
	struct msghdr msg = {.msg_iter = *from,
			     .msg_iocb = iocb};
808
	ssize_t res;
L
Linus Torvalds 已提交
809

810
	if (iocb->ki_pos != 0)
811
		return -ESPIPE;
812

813 814 815
	if (file->f_flags & O_NONBLOCK)
		msg.msg_flags = MSG_DONTWAIT;

816 817 818
	if (sock->type == SOCK_SEQPACKET)
		msg.msg_flags |= MSG_EOR;

819
	res = sock_sendmsg(sock, &msg);
820 821
	*from = msg.msg_iter;
	return res;
L
Linus Torvalds 已提交
822 823 824 825 826 827 828
}

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

A
Arjan van de Ven 已提交
829
static DEFINE_MUTEX(br_ioctl_mutex);
830
static int (*br_ioctl_hook) (struct net *, unsigned int cmd, void __user *arg);
L
Linus Torvalds 已提交
831

832
void brioctl_set(int (*hook) (struct net *, unsigned int, void __user *))
L
Linus Torvalds 已提交
833
{
A
Arjan van de Ven 已提交
834
	mutex_lock(&br_ioctl_mutex);
L
Linus Torvalds 已提交
835
	br_ioctl_hook = hook;
A
Arjan van de Ven 已提交
836
	mutex_unlock(&br_ioctl_mutex);
L
Linus Torvalds 已提交
837 838 839
}
EXPORT_SYMBOL(brioctl_set);

A
Arjan van de Ven 已提交
840
static DEFINE_MUTEX(vlan_ioctl_mutex);
841
static int (*vlan_ioctl_hook) (struct net *, void __user *arg);
L
Linus Torvalds 已提交
842

843
void vlan_ioctl_set(int (*hook) (struct net *, void __user *))
L
Linus Torvalds 已提交
844
{
A
Arjan van de Ven 已提交
845
	mutex_lock(&vlan_ioctl_mutex);
L
Linus Torvalds 已提交
846
	vlan_ioctl_hook = hook;
A
Arjan van de Ven 已提交
847
	mutex_unlock(&vlan_ioctl_mutex);
L
Linus Torvalds 已提交
848 849 850
}
EXPORT_SYMBOL(vlan_ioctl_set);

A
Arjan van de Ven 已提交
851
static DEFINE_MUTEX(dlci_ioctl_mutex);
852
static int (*dlci_ioctl_hook) (unsigned int, void __user *);
L
Linus Torvalds 已提交
853

854
void dlci_ioctl_set(int (*hook) (unsigned int, void __user *))
L
Linus Torvalds 已提交
855
{
A
Arjan van de Ven 已提交
856
	mutex_lock(&dlci_ioctl_mutex);
L
Linus Torvalds 已提交
857
	dlci_ioctl_hook = hook;
A
Arjan van de Ven 已提交
858
	mutex_unlock(&dlci_ioctl_mutex);
L
Linus Torvalds 已提交
859 860 861
}
EXPORT_SYMBOL(dlci_ioctl_set);

862 863 864 865 866 867 868 869 870 871 872 873 874 875 876 877 878 879
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
880 881 882 883 884 885 886 887
/*
 *	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;
888
	struct sock *sk;
L
Linus Torvalds 已提交
889 890
	void __user *argp = (void __user *)arg;
	int pid, err;
891
	struct net *net;
L
Linus Torvalds 已提交
892

893
	sock = file->private_data;
894
	sk = sock->sk;
895
	net = sock_net(sk);
L
Linus Torvalds 已提交
896
	if (cmd >= SIOCDEVPRIVATE && cmd <= (SIOCDEVPRIVATE + 15)) {
897
		err = dev_ioctl(net, cmd, argp);
L
Linus Torvalds 已提交
898
	} else
J
Johannes Berg 已提交
899
#ifdef CONFIG_WEXT_CORE
L
Linus Torvalds 已提交
900
	if (cmd >= SIOCIWFIRST && cmd <= SIOCIWLAST) {
901
		err = dev_ioctl(net, cmd, argp);
L
Linus Torvalds 已提交
902
	} else
J
Johannes Berg 已提交
903
#endif
904
		switch (cmd) {
L
Linus Torvalds 已提交
905 906 907 908 909
		case FIOSETOWN:
		case SIOCSPGRP:
			err = -EFAULT;
			if (get_user(pid, (int __user *)argp))
				break;
910 911
			f_setown(sock->file, pid, 1);
			err = 0;
L
Linus Torvalds 已提交
912 913 914
			break;
		case FIOGETOWN:
		case SIOCGPGRP:
915
			err = put_user(f_getown(sock->file),
916
				       (int __user *)argp);
L
Linus Torvalds 已提交
917 918 919 920 921 922 923 924 925
			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
926
			mutex_lock(&br_ioctl_mutex);
927
			if (br_ioctl_hook)
928
				err = br_ioctl_hook(net, cmd, argp);
A
Arjan van de Ven 已提交
929
			mutex_unlock(&br_ioctl_mutex);
L
Linus Torvalds 已提交
930 931 932 933 934 935 936
			break;
		case SIOCGIFVLAN:
		case SIOCSIFVLAN:
			err = -ENOPKG;
			if (!vlan_ioctl_hook)
				request_module("8021q");

A
Arjan van de Ven 已提交
937
			mutex_lock(&vlan_ioctl_mutex);
L
Linus Torvalds 已提交
938
			if (vlan_ioctl_hook)
939
				err = vlan_ioctl_hook(net, argp);
A
Arjan van de Ven 已提交
940
			mutex_unlock(&vlan_ioctl_mutex);
L
Linus Torvalds 已提交
941 942 943 944 945 946 947
			break;
		case SIOCADDDLCI:
		case SIOCDELDLCI:
			err = -ENOPKG;
			if (!dlci_ioctl_hook)
				request_module("dlci");

948 949
			mutex_lock(&dlci_ioctl_mutex);
			if (dlci_ioctl_hook)
L
Linus Torvalds 已提交
950
				err = dlci_ioctl_hook(cmd, argp);
951
			mutex_unlock(&dlci_ioctl_mutex);
L
Linus Torvalds 已提交
952 953
			break;
		default:
954
			err = sock_do_ioctl(net, sock, cmd, arg);
L
Linus Torvalds 已提交
955
			break;
956
		}
L
Linus Torvalds 已提交
957 958 959 960 961 962 963
	return err;
}

int sock_create_lite(int family, int type, int protocol, struct socket **res)
{
	int err;
	struct socket *sock = NULL;
964

L
Linus Torvalds 已提交
965 966 967 968 969 970 971 972 973 974 975
	err = security_socket_create(family, type, protocol, 1);
	if (err)
		goto out;

	sock = sock_alloc();
	if (!sock) {
		err = -ENOMEM;
		goto out;
	}

	sock->type = type;
976 977 978 979
	err = security_socket_post_create(sock, family, type, protocol, 1);
	if (err)
		goto out_release;

L
Linus Torvalds 已提交
980 981 982
out:
	*res = sock;
	return err;
983 984 985 986
out_release:
	sock_release(sock);
	sock = NULL;
	goto out;
L
Linus Torvalds 已提交
987
}
988
EXPORT_SYMBOL(sock_create_lite);
L
Linus Torvalds 已提交
989 990

/* No kernel lock held - perfect */
991
static unsigned int sock_poll(struct file *file, poll_table *wait)
L
Linus Torvalds 已提交
992
{
993
	unsigned int busy_flag = 0;
L
Linus Torvalds 已提交
994 995 996
	struct socket *sock;

	/*
997
	 *      We can't return errors to poll, so it's either yes or no.
L
Linus Torvalds 已提交
998
	 */
999
	sock = file->private_data;
1000

1001
	if (sk_can_busy_loop(sock->sk)) {
1002
		/* this socket can poll_ll so tell the system call */
1003
		busy_flag = POLL_BUSY_LOOP;
1004 1005

		/* once, only if requested by syscall */
1006 1007
		if (wait && (wait->_key & POLL_BUSY_LOOP))
			sk_busy_loop(sock->sk, 1);
1008 1009
	}

1010
	return busy_flag | sock->ops->poll(file, sock, wait);
L
Linus Torvalds 已提交
1011 1012
}

1013
static int sock_mmap(struct file *file, struct vm_area_struct *vma)
L
Linus Torvalds 已提交
1014
{
1015
	struct socket *sock = file->private_data;
L
Linus Torvalds 已提交
1016 1017 1018 1019

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

1020
static int sock_close(struct inode *inode, struct file *filp)
L
Linus Torvalds 已提交
1021 1022 1023 1024 1025 1026 1027 1028 1029 1030 1031 1032 1033
{
	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)
1034
 *	   or under socket lock
L
Linus Torvalds 已提交
1035 1036 1037 1038
 */

static int sock_fasync(int fd, struct file *filp, int on)
{
1039 1040
	struct socket *sock = filp->private_data;
	struct sock *sk = sock->sk;
1041
	struct socket_wq *wq;
L
Linus Torvalds 已提交
1042

1043
	if (sk == NULL)
L
Linus Torvalds 已提交
1044 1045 1046
		return -EINVAL;

	lock_sock(sk);
1047 1048
	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
1049

1050
	if (!wq->fasync_list)
1051 1052
		sock_reset_flag(sk, SOCK_FASYNC);
	else
1053
		sock_set_flag(sk, SOCK_FASYNC);
L
Linus Torvalds 已提交
1054

1055
	release_sock(sk);
L
Linus Torvalds 已提交
1056 1057 1058
	return 0;
}

1059
/* This function may be called only under socket lock or callback_lock or rcu_lock */
L
Linus Torvalds 已提交
1060 1061 1062

int sock_wake_async(struct socket *sock, int how, int band)
{
1063 1064 1065 1066 1067 1068 1069 1070
	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
1071
		return -1;
1072
	}
1073
	switch (how) {
1074
	case SOCK_WAKE_WAITD:
L
Linus Torvalds 已提交
1075 1076 1077
		if (test_bit(SOCK_ASYNC_WAITDATA, &sock->flags))
			break;
		goto call_kill;
1078
	case SOCK_WAKE_SPACE:
L
Linus Torvalds 已提交
1079 1080 1081
		if (!test_and_clear_bit(SOCK_ASYNC_NOSPACE, &sock->flags))
			break;
		/* fall through */
1082
	case SOCK_WAKE_IO:
1083
call_kill:
1084
		kill_fasync(&wq->fasync_list, SIGIO, band);
L
Linus Torvalds 已提交
1085
		break;
1086
	case SOCK_WAKE_URG:
1087
		kill_fasync(&wq->fasync_list, SIGURG, band);
L
Linus Torvalds 已提交
1088
	}
1089
	rcu_read_unlock();
L
Linus Torvalds 已提交
1090 1091
	return 0;
}
1092
EXPORT_SYMBOL(sock_wake_async);
L
Linus Torvalds 已提交
1093

1094
int __sock_create(struct net *net, int family, int type, int protocol,
1095
			 struct socket **res, int kern)
L
Linus Torvalds 已提交
1096 1097 1098
{
	int err;
	struct socket *sock;
1099
	const struct net_proto_family *pf;
L
Linus Torvalds 已提交
1100 1101

	/*
1102
	 *      Check protocol is in range
L
Linus Torvalds 已提交
1103 1104 1105 1106 1107 1108 1109 1110 1111 1112 1113 1114
	 */
	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) {
1115
		static int warned;
L
Linus Torvalds 已提交
1116 1117
		if (!warned) {
			warned = 1;
1118 1119
			pr_info("%s uses obsolete (PF_INET,SOCK_PACKET)\n",
				current->comm);
L
Linus Torvalds 已提交
1120 1121 1122 1123 1124 1125 1126
		}
		family = PF_PACKET;
	}

	err = security_socket_create(family, type, protocol, kern);
	if (err)
		return err;
1127

1128 1129 1130 1131 1132 1133 1134
	/*
	 *	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) {
1135
		net_warn_ratelimited("socket: no more sockets\n");
1136 1137 1138 1139 1140 1141
		return -ENFILE;	/* Not exactly a match, but its the
				   closest posix thing */
	}

	sock->type = type;

1142
#ifdef CONFIG_MODULES
1143 1144 1145
	/* 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
1146 1147 1148
	 * requested real, full-featured networking support upon configuration.
	 * Otherwise module support will break!
	 */
1149
	if (rcu_access_pointer(net_families[family]) == NULL)
1150
		request_module("net-pf-%d", family);
L
Linus Torvalds 已提交
1151 1152
#endif

1153 1154 1155 1156 1157
	rcu_read_lock();
	pf = rcu_dereference(net_families[family]);
	err = -EAFNOSUPPORT;
	if (!pf)
		goto out_release;
L
Linus Torvalds 已提交
1158 1159 1160 1161 1162

	/*
	 * We will call the ->create function, that possibly is in a loadable
	 * module, so we have to bump that loadable module refcnt first.
	 */
1163
	if (!try_module_get(pf->owner))
L
Linus Torvalds 已提交
1164 1165
		goto out_release;

1166 1167 1168
	/* Now protected by module ref count */
	rcu_read_unlock();

1169
	err = pf->create(net, sock, protocol, kern);
1170
	if (err < 0)
L
Linus Torvalds 已提交
1171
		goto out_module_put;
1172

L
Linus Torvalds 已提交
1173 1174 1175 1176
	/*
	 * Now to bump the refcnt of the [loadable] module that owns this
	 * socket at sock_release time we decrement its refcnt.
	 */
1177 1178 1179
	if (!try_module_get(sock->ops->owner))
		goto out_module_busy;

L
Linus Torvalds 已提交
1180 1181 1182 1183
	/*
	 * Now that we're done with the ->create function, the [loadable]
	 * module can have its refcnt decremented
	 */
1184
	module_put(pf->owner);
1185 1186
	err = security_socket_post_create(sock, family, type, protocol, kern);
	if (err)
1187
		goto out_sock_release;
1188
	*res = sock;
L
Linus Torvalds 已提交
1189

1190 1191 1192 1193
	return 0;

out_module_busy:
	err = -EAFNOSUPPORT;
L
Linus Torvalds 已提交
1194
out_module_put:
1195 1196 1197
	sock->ops = NULL;
	module_put(pf->owner);
out_sock_release:
L
Linus Torvalds 已提交
1198
	sock_release(sock);
1199 1200 1201 1202 1203
	return err;

out_release:
	rcu_read_unlock();
	goto out_sock_release;
L
Linus Torvalds 已提交
1204
}
1205
EXPORT_SYMBOL(__sock_create);
L
Linus Torvalds 已提交
1206 1207 1208

int sock_create(int family, int type, int protocol, struct socket **res)
{
1209
	return __sock_create(current->nsproxy->net_ns, family, type, protocol, res, 0);
L
Linus Torvalds 已提交
1210
}
1211
EXPORT_SYMBOL(sock_create);
L
Linus Torvalds 已提交
1212

1213
int sock_create_kern(struct net *net, int family, int type, int protocol, struct socket **res)
L
Linus Torvalds 已提交
1214
{
1215
	return __sock_create(net, family, type, protocol, res, 1);
L
Linus Torvalds 已提交
1216
}
1217
EXPORT_SYMBOL(sock_create_kern);
L
Linus Torvalds 已提交
1218

1219
SYSCALL_DEFINE3(socket, int, family, int, type, int, protocol)
L
Linus Torvalds 已提交
1220 1221 1222
{
	int retval;
	struct socket *sock;
1223 1224
	int flags;

1225 1226 1227 1228 1229 1230
	/* 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);

1231
	flags = type & ~SOCK_TYPE_MASK;
1232
	if (flags & ~(SOCK_CLOEXEC | SOCK_NONBLOCK))
1233 1234
		return -EINVAL;
	type &= SOCK_TYPE_MASK;
L
Linus Torvalds 已提交
1235

U
Ulrich Drepper 已提交
1236 1237 1238
	if (SOCK_NONBLOCK != O_NONBLOCK && (flags & SOCK_NONBLOCK))
		flags = (flags & ~SOCK_NONBLOCK) | O_NONBLOCK;

L
Linus Torvalds 已提交
1239 1240 1241 1242
	retval = sock_create(family, type, protocol, &sock);
	if (retval < 0)
		goto out;

1243
	retval = sock_map_fd(sock, flags & (O_CLOEXEC | O_NONBLOCK));
L
Linus Torvalds 已提交
1244 1245 1246 1247 1248 1249 1250 1251 1252 1253 1254 1255 1256 1257 1258 1259
	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.
 */

1260 1261
SYSCALL_DEFINE4(socketpair, int, family, int, type, int, protocol,
		int __user *, usockvec)
L
Linus Torvalds 已提交
1262 1263 1264
{
	struct socket *sock1, *sock2;
	int fd1, fd2, err;
A
Al Viro 已提交
1265
	struct file *newfile1, *newfile2;
1266 1267 1268
	int flags;

	flags = type & ~SOCK_TYPE_MASK;
1269
	if (flags & ~(SOCK_CLOEXEC | SOCK_NONBLOCK))
1270 1271
		return -EINVAL;
	type &= SOCK_TYPE_MASK;
L
Linus Torvalds 已提交
1272

U
Ulrich Drepper 已提交
1273 1274 1275
	if (SOCK_NONBLOCK != O_NONBLOCK && (flags & SOCK_NONBLOCK))
		flags = (flags & ~SOCK_NONBLOCK) | O_NONBLOCK;

L
Linus Torvalds 已提交
1276 1277 1278 1279 1280 1281 1282 1283 1284 1285 1286 1287 1288 1289
	/*
	 * 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);
1290
	if (err < 0)
L
Linus Torvalds 已提交
1291 1292
		goto out_release_both;

1293
	fd1 = get_unused_fd_flags(flags);
1294 1295
	if (unlikely(fd1 < 0)) {
		err = fd1;
A
Al Viro 已提交
1296
		goto out_release_both;
1297
	}
1298

1299
	fd2 = get_unused_fd_flags(flags);
1300 1301
	if (unlikely(fd2 < 0)) {
		err = fd2;
1302
		goto out_put_unused_1;
1303 1304
	}

1305
	newfile1 = sock_alloc_file(sock1, flags, NULL);
1306 1307
	if (unlikely(IS_ERR(newfile1))) {
		err = PTR_ERR(newfile1);
1308
		goto out_put_unused_both;
1309 1310
	}

1311
	newfile2 = sock_alloc_file(sock2, flags, NULL);
1312 1313
	if (IS_ERR(newfile2)) {
		err = PTR_ERR(newfile2);
1314
		goto out_fput_1;
A
Al Viro 已提交
1315 1316
	}

1317 1318 1319 1320 1321 1322 1323 1324
	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
1325
	audit_fd_pair(fd1, fd2);
1326

A
Al Viro 已提交
1327 1328
	fd_install(fd1, newfile1);
	fd_install(fd2, newfile2);
L
Linus Torvalds 已提交
1329 1330 1331 1332
	/* fd1 and fd2 may be already another descriptors.
	 * Not kernel problem.
	 */

1333
	return 0;
L
Linus Torvalds 已提交
1334

1335 1336 1337 1338 1339 1340 1341 1342 1343 1344 1345 1346 1347
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
1348

1349 1350 1351 1352
out_put_unused_both:
	put_unused_fd(fd2);
out_put_unused_1:
	put_unused_fd(fd1);
L
Linus Torvalds 已提交
1353
out_release_both:
1354
	sock_release(sock2);
L
Linus Torvalds 已提交
1355
out_release_1:
1356
	sock_release(sock1);
L
Linus Torvalds 已提交
1357 1358 1359 1360 1361 1362 1363 1364 1365 1366 1367 1368
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).
 */

1369
SYSCALL_DEFINE3(bind, int, fd, struct sockaddr __user *, umyaddr, int, addrlen)
L
Linus Torvalds 已提交
1370 1371
{
	struct socket *sock;
1372
	struct sockaddr_storage address;
1373
	int err, fput_needed;
L
Linus Torvalds 已提交
1374

1375
	sock = sockfd_lookup_light(fd, &err, &fput_needed);
1376
	if (sock) {
1377
		err = move_addr_to_kernel(umyaddr, addrlen, &address);
1378 1379
		if (err >= 0) {
			err = security_socket_bind(sock,
1380
						   (struct sockaddr *)&address,
1381
						   addrlen);
1382 1383
			if (!err)
				err = sock->ops->bind(sock,
1384
						      (struct sockaddr *)
1385
						      &address, addrlen);
L
Linus Torvalds 已提交
1386
		}
1387
		fput_light(sock->file, fput_needed);
1388
	}
L
Linus Torvalds 已提交
1389 1390 1391 1392 1393 1394 1395 1396 1397
	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.
 */

1398
SYSCALL_DEFINE2(listen, int, fd, int, backlog)
L
Linus Torvalds 已提交
1399 1400
{
	struct socket *sock;
1401
	int err, fput_needed;
1402
	int somaxconn;
1403 1404 1405

	sock = sockfd_lookup_light(fd, &err, &fput_needed);
	if (sock) {
1406
		somaxconn = sock_net(sock->sk)->core.sysctl_somaxconn;
1407
		if ((unsigned int)backlog > somaxconn)
1408
			backlog = somaxconn;
L
Linus Torvalds 已提交
1409 1410

		err = security_socket_listen(sock, backlog);
1411 1412
		if (!err)
			err = sock->ops->listen(sock, backlog);
L
Linus Torvalds 已提交
1413

1414
		fput_light(sock->file, fput_needed);
L
Linus Torvalds 已提交
1415 1416 1417 1418 1419 1420 1421 1422 1423 1424 1425 1426 1427 1428 1429 1430
	}
	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.
 */

1431 1432
SYSCALL_DEFINE4(accept4, int, fd, struct sockaddr __user *, upeer_sockaddr,
		int __user *, upeer_addrlen, int, flags)
L
Linus Torvalds 已提交
1433 1434
{
	struct socket *sock, *newsock;
1435
	struct file *newfile;
1436
	int err, len, newfd, fput_needed;
1437
	struct sockaddr_storage address;
L
Linus Torvalds 已提交
1438

1439
	if (flags & ~(SOCK_CLOEXEC | SOCK_NONBLOCK))
U
Ulrich Drepper 已提交
1440 1441 1442 1443 1444
		return -EINVAL;

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

1445
	sock = sockfd_lookup_light(fd, &err, &fput_needed);
L
Linus Torvalds 已提交
1446 1447 1448 1449
	if (!sock)
		goto out;

	err = -ENFILE;
1450 1451
	newsock = sock_alloc();
	if (!newsock)
L
Linus Torvalds 已提交
1452 1453 1454 1455 1456 1457 1458 1459 1460 1461 1462
		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);

1463
	newfd = get_unused_fd_flags(flags);
1464 1465
	if (unlikely(newfd < 0)) {
		err = newfd;
1466 1467
		sock_release(newsock);
		goto out_put;
1468
	}
1469
	newfile = sock_alloc_file(newsock, flags, sock->sk->sk_prot_creator->name);
1470 1471 1472 1473 1474 1475
	if (unlikely(IS_ERR(newfile))) {
		err = PTR_ERR(newfile);
		put_unused_fd(newfd);
		sock_release(newsock);
		goto out_put;
	}
1476

1477 1478
	err = security_socket_accept(sock, newsock);
	if (err)
1479
		goto out_fd;
1480

L
Linus Torvalds 已提交
1481 1482
	err = sock->ops->accept(sock, newsock, sock->file->f_flags);
	if (err < 0)
1483
		goto out_fd;
L
Linus Torvalds 已提交
1484 1485

	if (upeer_sockaddr) {
1486
		if (newsock->ops->getname(newsock, (struct sockaddr *)&address,
1487
					  &len, 2) < 0) {
L
Linus Torvalds 已提交
1488
			err = -ECONNABORTED;
1489
			goto out_fd;
L
Linus Torvalds 已提交
1490
		}
1491
		err = move_addr_to_user(&address,
1492
					len, upeer_sockaddr, upeer_addrlen);
L
Linus Torvalds 已提交
1493
		if (err < 0)
1494
			goto out_fd;
L
Linus Torvalds 已提交
1495 1496 1497 1498
	}

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

1499 1500
	fd_install(newfd, newfile);
	err = newfd;
L
Linus Torvalds 已提交
1501 1502

out_put:
1503
	fput_light(sock->file, fput_needed);
L
Linus Torvalds 已提交
1504 1505
out:
	return err;
1506
out_fd:
1507
	fput(newfile);
1508
	put_unused_fd(newfd);
L
Linus Torvalds 已提交
1509 1510 1511
	goto out_put;
}

1512 1513
SYSCALL_DEFINE3(accept, int, fd, struct sockaddr __user *, upeer_sockaddr,
		int __user *, upeer_addrlen)
U
Ulrich Drepper 已提交
1514
{
U
Ulrich Drepper 已提交
1515
	return sys_accept4(fd, upeer_sockaddr, upeer_addrlen, 0);
U
Ulrich Drepper 已提交
1516 1517
}

L
Linus Torvalds 已提交
1518 1519 1520 1521 1522 1523 1524 1525 1526 1527 1528 1529
/*
 *	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.
 */

1530 1531
SYSCALL_DEFINE3(connect, int, fd, struct sockaddr __user *, uservaddr,
		int, addrlen)
L
Linus Torvalds 已提交
1532 1533
{
	struct socket *sock;
1534
	struct sockaddr_storage address;
1535
	int err, fput_needed;
L
Linus Torvalds 已提交
1536

1537
	sock = sockfd_lookup_light(fd, &err, &fput_needed);
L
Linus Torvalds 已提交
1538 1539
	if (!sock)
		goto out;
1540
	err = move_addr_to_kernel(uservaddr, addrlen, &address);
L
Linus Torvalds 已提交
1541 1542 1543
	if (err < 0)
		goto out_put;

1544
	err =
1545
	    security_socket_connect(sock, (struct sockaddr *)&address, addrlen);
L
Linus Torvalds 已提交
1546 1547 1548
	if (err)
		goto out_put;

1549
	err = sock->ops->connect(sock, (struct sockaddr *)&address, addrlen,
L
Linus Torvalds 已提交
1550 1551
				 sock->file->f_flags);
out_put:
1552
	fput_light(sock->file, fput_needed);
L
Linus Torvalds 已提交
1553 1554 1555 1556 1557 1558 1559 1560 1561
out:
	return err;
}

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

1562 1563
SYSCALL_DEFINE3(getsockname, int, fd, struct sockaddr __user *, usockaddr,
		int __user *, usockaddr_len)
L
Linus Torvalds 已提交
1564 1565
{
	struct socket *sock;
1566
	struct sockaddr_storage address;
1567
	int len, err, fput_needed;
1568

1569
	sock = sockfd_lookup_light(fd, &err, &fput_needed);
L
Linus Torvalds 已提交
1570 1571 1572 1573 1574 1575 1576
	if (!sock)
		goto out;

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

1577
	err = sock->ops->getname(sock, (struct sockaddr *)&address, &len, 0);
L
Linus Torvalds 已提交
1578 1579
	if (err)
		goto out_put;
1580
	err = move_addr_to_user(&address, len, usockaddr, usockaddr_len);
L
Linus Torvalds 已提交
1581 1582

out_put:
1583
	fput_light(sock->file, fput_needed);
L
Linus Torvalds 已提交
1584 1585 1586 1587 1588 1589 1590 1591 1592
out:
	return err;
}

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

1593 1594
SYSCALL_DEFINE3(getpeername, int, fd, struct sockaddr __user *, usockaddr,
		int __user *, usockaddr_len)
L
Linus Torvalds 已提交
1595 1596
{
	struct socket *sock;
1597
	struct sockaddr_storage address;
1598
	int len, err, fput_needed;
L
Linus Torvalds 已提交
1599

1600 1601
	sock = sockfd_lookup_light(fd, &err, &fput_needed);
	if (sock != NULL) {
L
Linus Torvalds 已提交
1602 1603
		err = security_socket_getpeername(sock);
		if (err) {
1604
			fput_light(sock->file, fput_needed);
L
Linus Torvalds 已提交
1605 1606 1607
			return err;
		}

1608
		err =
1609
		    sock->ops->getname(sock, (struct sockaddr *)&address, &len,
1610
				       1);
L
Linus Torvalds 已提交
1611
		if (!err)
1612
			err = move_addr_to_user(&address, len, usockaddr,
1613
						usockaddr_len);
1614
		fput_light(sock->file, fput_needed);
L
Linus Torvalds 已提交
1615 1616 1617 1618 1619 1620 1621 1622 1623 1624
	}
	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.
 */

1625
SYSCALL_DEFINE6(sendto, int, fd, void __user *, buff, size_t, len,
1626
		unsigned int, flags, struct sockaddr __user *, addr,
1627
		int, addr_len)
L
Linus Torvalds 已提交
1628 1629
{
	struct socket *sock;
1630
	struct sockaddr_storage address;
L
Linus Torvalds 已提交
1631 1632 1633
	int err;
	struct msghdr msg;
	struct iovec iov;
1634 1635
	int fput_needed;

1636 1637 1638
	err = import_single_range(WRITE, buff, len, &iov, &msg.msg_iter);
	if (unlikely(err))
		return err;
1639 1640
	sock = sockfd_lookup_light(fd, &err, &fput_needed);
	if (!sock)
1641
		goto out;
1642

1643 1644 1645 1646
	msg.msg_name = NULL;
	msg.msg_control = NULL;
	msg.msg_controllen = 0;
	msg.msg_namelen = 0;
1647
	if (addr) {
1648
		err = move_addr_to_kernel(addr, addr_len, &address);
L
Linus Torvalds 已提交
1649 1650
		if (err < 0)
			goto out_put;
1651
		msg.msg_name = (struct sockaddr *)&address;
1652
		msg.msg_namelen = addr_len;
L
Linus Torvalds 已提交
1653 1654 1655 1656
	}
	if (sock->file->f_flags & O_NONBLOCK)
		flags |= MSG_DONTWAIT;
	msg.msg_flags = flags;
1657
	err = sock_sendmsg(sock, &msg);
L
Linus Torvalds 已提交
1658

1659
out_put:
1660
	fput_light(sock->file, fput_needed);
1661
out:
L
Linus Torvalds 已提交
1662 1663 1664 1665
	return err;
}

/*
1666
 *	Send a datagram down a socket.
L
Linus Torvalds 已提交
1667 1668
 */

1669
SYSCALL_DEFINE4(send, int, fd, void __user *, buff, size_t, len,
1670
		unsigned int, flags)
L
Linus Torvalds 已提交
1671 1672 1673 1674 1675
{
	return sys_sendto(fd, buff, len, flags, NULL, 0);
}

/*
1676
 *	Receive a frame from the socket and optionally record the address of the
L
Linus Torvalds 已提交
1677 1678 1679 1680
 *	sender. We verify the buffers are writable and if needed move the
 *	sender address from kernel to user space.
 */

1681
SYSCALL_DEFINE6(recvfrom, int, fd, void __user *, ubuf, size_t, size,
1682
		unsigned int, flags, struct sockaddr __user *, addr,
1683
		int __user *, addr_len)
L
Linus Torvalds 已提交
1684 1685 1686 1687
{
	struct socket *sock;
	struct iovec iov;
	struct msghdr msg;
1688
	struct sockaddr_storage address;
1689
	int err, err2;
1690 1691
	int fput_needed;

1692 1693 1694
	err = import_single_range(READ, ubuf, size, &iov, &msg.msg_iter);
	if (unlikely(err))
		return err;
1695
	sock = sockfd_lookup_light(fd, &err, &fput_needed);
L
Linus Torvalds 已提交
1696
	if (!sock)
1697
		goto out;
L
Linus Torvalds 已提交
1698

1699 1700
	msg.msg_control = NULL;
	msg.msg_controllen = 0;
1701 1702 1703 1704
	/* 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
1705 1706
	if (sock->file->f_flags & O_NONBLOCK)
		flags |= MSG_DONTWAIT;
1707
	err = sock_recvmsg(sock, &msg, iov_iter_count(&msg.msg_iter), flags);
L
Linus Torvalds 已提交
1708

1709
	if (err >= 0 && addr != NULL) {
1710
		err2 = move_addr_to_user(&address,
1711
					 msg.msg_namelen, addr, addr_len);
1712 1713
		if (err2 < 0)
			err = err2;
L
Linus Torvalds 已提交
1714
	}
1715 1716

	fput_light(sock->file, fput_needed);
1717
out:
L
Linus Torvalds 已提交
1718 1719 1720 1721
	return err;
}

/*
1722
 *	Receive a datagram from a socket.
L
Linus Torvalds 已提交
1723 1724
 */

1725 1726
SYSCALL_DEFINE4(recv, int, fd, void __user *, ubuf, size_t, size,
		unsigned int, flags)
L
Linus Torvalds 已提交
1727 1728 1729 1730 1731 1732 1733 1734 1735
{
	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.
 */

1736 1737
SYSCALL_DEFINE5(setsockopt, int, fd, int, level, int, optname,
		char __user *, optval, int, optlen)
L
Linus Torvalds 已提交
1738
{
1739
	int err, fput_needed;
L
Linus Torvalds 已提交
1740 1741 1742 1743
	struct socket *sock;

	if (optlen < 0)
		return -EINVAL;
1744 1745 1746 1747

	sock = sockfd_lookup_light(fd, &err, &fput_needed);
	if (sock != NULL) {
		err = security_socket_setsockopt(sock, level, optname);
1748 1749
		if (err)
			goto out_put;
L
Linus Torvalds 已提交
1750 1751

		if (level == SOL_SOCKET)
1752 1753 1754
			err =
			    sock_setsockopt(sock, level, optname, optval,
					    optlen);
L
Linus Torvalds 已提交
1755
		else
1756 1757 1758
			err =
			    sock->ops->setsockopt(sock, level, optname, optval,
						  optlen);
1759 1760
out_put:
		fput_light(sock->file, fput_needed);
L
Linus Torvalds 已提交
1761 1762 1763 1764 1765 1766 1767 1768 1769
	}
	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.
 */

1770 1771
SYSCALL_DEFINE5(getsockopt, int, fd, int, level, int, optname,
		char __user *, optval, int __user *, optlen)
L
Linus Torvalds 已提交
1772
{
1773
	int err, fput_needed;
L
Linus Torvalds 已提交
1774 1775
	struct socket *sock;

1776 1777
	sock = sockfd_lookup_light(fd, &err, &fput_needed);
	if (sock != NULL) {
1778 1779 1780
		err = security_socket_getsockopt(sock, level, optname);
		if (err)
			goto out_put;
L
Linus Torvalds 已提交
1781 1782

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

/*
 *	Shutdown a socket.
 */

1800
SYSCALL_DEFINE2(shutdown, int, fd, int, how)
L
Linus Torvalds 已提交
1801
{
1802
	int err, fput_needed;
L
Linus Torvalds 已提交
1803 1804
	struct socket *sock;

1805 1806
	sock = sockfd_lookup_light(fd, &err, &fput_needed);
	if (sock != NULL) {
L
Linus Torvalds 已提交
1807
		err = security_socket_shutdown(sock, how);
1808 1809 1810
		if (!err)
			err = sock->ops->shutdown(sock, how);
		fput_light(sock->file, fput_needed);
L
Linus Torvalds 已提交
1811 1812 1813 1814
	}
	return err;
}

1815
/* A couple of helpful macros for getting the address of the 32/64 bit
L
Linus Torvalds 已提交
1816 1817 1818 1819 1820 1821
 * 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)

1822 1823 1824 1825 1826
struct used_address {
	struct sockaddr_storage name;
	unsigned int name_len;
};

1827 1828 1829 1830
static int copy_msghdr_from_user(struct msghdr *kmsg,
				 struct user_msghdr __user *umsg,
				 struct sockaddr __user **save_addr,
				 struct iovec **iov)
1831
{
1832 1833
	struct sockaddr __user *uaddr;
	struct iovec __user *uiov;
A
Al Viro 已提交
1834
	size_t nr_segs;
1835 1836 1837 1838 1839 1840
	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
1841
	    __get_user(nr_segs, &umsg->msg_iovlen) ||
1842 1843 1844
	    __get_user(kmsg->msg_control, &umsg->msg_control) ||
	    __get_user(kmsg->msg_controllen, &umsg->msg_controllen) ||
	    __get_user(kmsg->msg_flags, &umsg->msg_flags))
1845
		return -EFAULT;
1846

1847
	if (!uaddr)
1848 1849
		kmsg->msg_namelen = 0;

1850 1851 1852
	if (kmsg->msg_namelen < 0)
		return -EINVAL;

1853
	if (kmsg->msg_namelen > sizeof(struct sockaddr_storage))
1854
		kmsg->msg_namelen = sizeof(struct sockaddr_storage);
1855 1856 1857 1858 1859 1860 1861 1862 1863 1864 1865 1866 1867 1868 1869 1870

	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
1871
	if (nr_segs > UIO_MAXIOV)
1872 1873
		return -EMSGSIZE;

1874 1875
	kmsg->msg_iocb = NULL;

1876 1877
	return import_iovec(save_addr ? READ : WRITE, uiov, nr_segs,
			    UIO_FASTIOV, iov, &kmsg->msg_iter);
1878 1879
}

1880
static int ___sys_sendmsg(struct socket *sock, struct user_msghdr __user *msg,
1881
			 struct msghdr *msg_sys, unsigned int flags,
1882
			 struct used_address *used_address)
L
Linus Torvalds 已提交
1883
{
1884 1885
	struct compat_msghdr __user *msg_compat =
	    (struct compat_msghdr __user *)msg;
1886
	struct sockaddr_storage address;
L
Linus Torvalds 已提交
1887
	struct iovec iovstack[UIO_FASTIOV], *iov = iovstack;
1888
	unsigned char ctl[sizeof(struct cmsghdr) + 20]
1889 1890
	    __attribute__ ((aligned(sizeof(__kernel_size_t))));
	/* 20 is size of ipv6_pktinfo */
L
Linus Torvalds 已提交
1891
	unsigned char *ctl_buf = ctl;
1892
	int ctl_len;
1893
	ssize_t err;
1894

1895
	msg_sys->msg_name = &address;
L
Linus Torvalds 已提交
1896

1897
	if (MSG_CMSG_COMPAT & flags)
1898
		err = get_compat_msghdr(msg_sys, msg_compat, NULL, &iov);
1899
	else
1900
		err = copy_msghdr_from_user(msg_sys, msg, NULL, &iov);
1901
	if (err < 0)
1902
		return err;
L
Linus Torvalds 已提交
1903 1904 1905

	err = -ENOBUFS;

1906
	if (msg_sys->msg_controllen > INT_MAX)
L
Linus Torvalds 已提交
1907
		goto out_freeiov;
1908
	ctl_len = msg_sys->msg_controllen;
L
Linus Torvalds 已提交
1909
	if ((MSG_CMSG_COMPAT & flags) && ctl_len) {
1910
		err =
1911
		    cmsghdr_from_user_compat_to_kern(msg_sys, sock->sk, ctl,
1912
						     sizeof(ctl));
L
Linus Torvalds 已提交
1913 1914
		if (err)
			goto out_freeiov;
1915 1916
		ctl_buf = msg_sys->msg_control;
		ctl_len = msg_sys->msg_controllen;
L
Linus Torvalds 已提交
1917
	} else if (ctl_len) {
1918
		if (ctl_len > sizeof(ctl)) {
L
Linus Torvalds 已提交
1919
			ctl_buf = sock_kmalloc(sock->sk, ctl_len, GFP_KERNEL);
1920
			if (ctl_buf == NULL)
L
Linus Torvalds 已提交
1921 1922 1923 1924
				goto out_freeiov;
		}
		err = -EFAULT;
		/*
1925
		 * Careful! Before this, msg_sys->msg_control contains a user pointer.
L
Linus Torvalds 已提交
1926 1927 1928
		 * Afterwards, it will be a kernel pointer. Thus the compiler-assisted
		 * checking falls down on this.
		 */
1929
		if (copy_from_user(ctl_buf,
1930
				   (void __user __force *)msg_sys->msg_control,
1931
				   ctl_len))
L
Linus Torvalds 已提交
1932
			goto out_freectl;
1933
		msg_sys->msg_control = ctl_buf;
L
Linus Torvalds 已提交
1934
	}
1935
	msg_sys->msg_flags = flags;
L
Linus Torvalds 已提交
1936 1937

	if (sock->file->f_flags & O_NONBLOCK)
1938
		msg_sys->msg_flags |= MSG_DONTWAIT;
1939 1940 1941 1942 1943 1944
	/*
	 * 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.
	 */
1945 1946 1947
	if (used_address && msg_sys->msg_name &&
	    used_address->name_len == msg_sys->msg_namelen &&
	    !memcmp(&used_address->name, msg_sys->msg_name,
1948
		    used_address->name_len)) {
1949
		err = sock_sendmsg_nosec(sock, msg_sys);
1950 1951
		goto out_freectl;
	}
1952
	err = sock_sendmsg(sock, msg_sys);
1953 1954 1955 1956 1957 1958
	/*
	 * 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;
1959 1960 1961
		if (msg_sys->msg_name)
			memcpy(&used_address->name, msg_sys->msg_name,
			       used_address->name_len);
1962
	}
L
Linus Torvalds 已提交
1963 1964

out_freectl:
1965
	if (ctl_buf != ctl)
L
Linus Torvalds 已提交
1966 1967
		sock_kfree_s(sock->sk, ctl_buf, ctl_len);
out_freeiov:
1968
	kfree(iov);
1969 1970 1971 1972 1973 1974 1975
	return err;
}

/*
 *	BSD sendmsg interface
 */

1976
long __sys_sendmsg(int fd, struct user_msghdr __user *msg, unsigned flags)
1977 1978 1979
{
	int fput_needed, err;
	struct msghdr msg_sys;
1980 1981 1982
	struct socket *sock;

	sock = sockfd_lookup_light(fd, &err, &fput_needed);
1983 1984 1985
	if (!sock)
		goto out;

1986
	err = ___sys_sendmsg(sock, msg, &msg_sys, flags, NULL);
1987

1988
	fput_light(sock->file, fput_needed);
1989
out:
L
Linus Torvalds 已提交
1990 1991 1992
	return err;
}

1993
SYSCALL_DEFINE3(sendmsg, int, fd, struct user_msghdr __user *, msg, unsigned int, flags)
1994 1995 1996 1997 1998 1999
{
	if (flags & MSG_CMSG_COMPAT)
		return -EINVAL;
	return __sys_sendmsg(fd, msg, flags);
}

2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011
/*
 *	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;
2012
	struct used_address used_address;
2013

2014 2015
	if (vlen > UIO_MAXIOV)
		vlen = UIO_MAXIOV;
2016 2017 2018 2019 2020 2021 2022

	datagrams = 0;

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

2023
	used_address.name_len = UINT_MAX;
2024 2025
	entry = mmsg;
	compat_entry = (struct compat_mmsghdr __user *)mmsg;
2026
	err = 0;
2027 2028 2029

	while (datagrams < vlen) {
		if (MSG_CMSG_COMPAT & flags) {
2030
			err = ___sys_sendmsg(sock, (struct user_msghdr __user *)compat_entry,
2031
					     &msg_sys, flags, &used_address);
2032 2033 2034 2035 2036
			if (err < 0)
				break;
			err = __put_user(err, &compat_entry->msg_len);
			++compat_entry;
		} else {
2037
			err = ___sys_sendmsg(sock,
2038
					     (struct user_msghdr __user *)entry,
2039
					     &msg_sys, flags, &used_address);
2040 2041 2042 2043 2044 2045 2046 2047 2048 2049 2050 2051 2052
			if (err < 0)
				break;
			err = put_user(err, &entry->msg_len);
			++entry;
		}

		if (err)
			break;
		++datagrams;
	}

	fput_light(sock->file, fput_needed);

2053 2054
	/* We only return an error if no datagrams were able to be sent */
	if (datagrams != 0)
2055 2056 2057 2058 2059 2060 2061 2062
		return datagrams;

	return err;
}

SYSCALL_DEFINE4(sendmmsg, int, fd, struct mmsghdr __user *, mmsg,
		unsigned int, vlen, unsigned int, flags)
{
2063 2064
	if (flags & MSG_CMSG_COMPAT)
		return -EINVAL;
2065 2066 2067
	return __sys_sendmmsg(fd, mmsg, vlen, flags);
}

2068
static int ___sys_recvmsg(struct socket *sock, struct user_msghdr __user *msg,
2069
			 struct msghdr *msg_sys, unsigned int flags, int nosec)
L
Linus Torvalds 已提交
2070
{
2071 2072
	struct compat_msghdr __user *msg_compat =
	    (struct compat_msghdr __user *)msg;
L
Linus Torvalds 已提交
2073
	struct iovec iovstack[UIO_FASTIOV];
2074
	struct iovec *iov = iovstack;
L
Linus Torvalds 已提交
2075
	unsigned long cmsg_ptr;
2076 2077
	int total_len, len;
	ssize_t err;
L
Linus Torvalds 已提交
2078 2079

	/* kernel mode address */
2080
	struct sockaddr_storage addr;
L
Linus Torvalds 已提交
2081 2082 2083

	/* user mode address pointers */
	struct sockaddr __user *uaddr;
2084
	int __user *uaddr_len = COMPAT_NAMELEN(msg);
2085

2086
	msg_sys->msg_name = &addr;
L
Linus Torvalds 已提交
2087

2088
	if (MSG_CMSG_COMPAT & flags)
2089
		err = get_compat_msghdr(msg_sys, msg_compat, &uaddr, &iov);
2090
	else
2091
		err = copy_msghdr_from_user(msg_sys, msg, &uaddr, &iov);
L
Linus Torvalds 已提交
2092
	if (err < 0)
2093 2094
		return err;
	total_len = iov_iter_count(&msg_sys->msg_iter);
L
Linus Torvalds 已提交
2095

2096 2097
	cmsg_ptr = (unsigned long)msg_sys->msg_control;
	msg_sys->msg_flags = flags & (MSG_CMSG_CLOEXEC|MSG_CMSG_COMPAT);
2098

2099 2100 2101
	/* We assume all kernel code knows the size of sockaddr_storage */
	msg_sys->msg_namelen = 0;

L
Linus Torvalds 已提交
2102 2103
	if (sock->file->f_flags & O_NONBLOCK)
		flags |= MSG_DONTWAIT;
2104 2105
	err = (nosec ? sock_recvmsg_nosec : sock_recvmsg)(sock, msg_sys,
							  total_len, flags);
L
Linus Torvalds 已提交
2106 2107 2108 2109 2110
	if (err < 0)
		goto out_freeiov;
	len = err;

	if (uaddr != NULL) {
2111
		err = move_addr_to_user(&addr,
2112
					msg_sys->msg_namelen, uaddr,
2113
					uaddr_len);
L
Linus Torvalds 已提交
2114 2115 2116
		if (err < 0)
			goto out_freeiov;
	}
2117
	err = __put_user((msg_sys->msg_flags & ~MSG_CMSG_COMPAT),
2118
			 COMPAT_FLAGS(msg));
L
Linus Torvalds 已提交
2119 2120 2121
	if (err)
		goto out_freeiov;
	if (MSG_CMSG_COMPAT & flags)
2122
		err = __put_user((unsigned long)msg_sys->msg_control - cmsg_ptr,
L
Linus Torvalds 已提交
2123 2124
				 &msg_compat->msg_controllen);
	else
2125
		err = __put_user((unsigned long)msg_sys->msg_control - cmsg_ptr,
L
Linus Torvalds 已提交
2126 2127 2128 2129 2130 2131
				 &msg->msg_controllen);
	if (err)
		goto out_freeiov;
	err = len;

out_freeiov:
2132
	kfree(iov);
2133 2134 2135 2136 2137 2138 2139
	return err;
}

/*
 *	BSD recvmsg interface
 */

2140
long __sys_recvmsg(int fd, struct user_msghdr __user *msg, unsigned flags)
2141 2142 2143
{
	int fput_needed, err;
	struct msghdr msg_sys;
2144 2145 2146
	struct socket *sock;

	sock = sockfd_lookup_light(fd, &err, &fput_needed);
2147 2148 2149
	if (!sock)
		goto out;

2150
	err = ___sys_recvmsg(sock, msg, &msg_sys, flags, 0);
2151

2152
	fput_light(sock->file, fput_needed);
L
Linus Torvalds 已提交
2153 2154 2155 2156
out:
	return err;
}

2157
SYSCALL_DEFINE3(recvmsg, int, fd, struct user_msghdr __user *, msg,
2158 2159 2160 2161 2162 2163 2164
		unsigned int, flags)
{
	if (flags & MSG_CMSG_COMPAT)
		return -EINVAL;
	return __sys_recvmsg(fd, msg, flags);
}

2165 2166 2167 2168 2169 2170 2171 2172 2173 2174
/*
 *     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;
2175
	struct compat_mmsghdr __user *compat_entry;
2176 2177 2178 2179 2180 2181 2182 2183 2184 2185 2186 2187 2188 2189 2190 2191 2192 2193 2194
	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;
2195
	compat_entry = (struct compat_mmsghdr __user *)mmsg;
2196 2197 2198 2199 2200

	while (datagrams < vlen) {
		/*
		 * No need to ask LSM for more than the first datagram.
		 */
2201
		if (MSG_CMSG_COMPAT & flags) {
2202
			err = ___sys_recvmsg(sock, (struct user_msghdr __user *)compat_entry,
2203 2204
					     &msg_sys, flags & ~MSG_WAITFORONE,
					     datagrams);
2205 2206 2207 2208 2209
			if (err < 0)
				break;
			err = __put_user(err, &compat_entry->msg_len);
			++compat_entry;
		} else {
2210
			err = ___sys_recvmsg(sock,
2211
					     (struct user_msghdr __user *)entry,
2212 2213
					     &msg_sys, flags & ~MSG_WAITFORONE,
					     datagrams);
2214 2215 2216 2217 2218 2219
			if (err < 0)
				break;
			err = put_user(err, &entry->msg_len);
			++entry;
		}

2220 2221 2222 2223
		if (err)
			break;
		++datagrams;

2224 2225 2226 2227
		/* MSG_WAITFORONE turns on MSG_DONTWAIT after one packet */
		if (flags & MSG_WAITFORONE)
			flags |= MSG_DONTWAIT;

2228 2229 2230 2231 2232 2233 2234 2235 2236 2237 2238 2239 2240 2241 2242 2243 2244 2245 2246 2247
		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
2248

2249 2250 2251 2252 2253 2254 2255 2256 2257 2258 2259 2260 2261 2262 2263 2264 2265 2266 2267 2268 2269 2270 2271 2272 2273 2274 2275 2276 2277 2278 2279
	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;

2280 2281 2282
	if (flags & MSG_CMSG_COMPAT)
		return -EINVAL;

2283 2284 2285 2286 2287 2288 2289 2290 2291 2292 2293 2294 2295 2296 2297 2298
	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
2299 2300
/* Argument list sizes for sys_socketcall */
#define AL(x) ((x) * sizeof(unsigned long))
2301
static const unsigned char nargs[21] = {
2302 2303 2304
	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),
2305
	AL(4), AL(5), AL(4)
2306 2307
};

L
Linus Torvalds 已提交
2308 2309 2310
#undef AL

/*
2311
 *	System call vectors.
L
Linus Torvalds 已提交
2312 2313 2314
 *
 *	Argument checking cleaned up. Saved 20% in size.
 *  This function doesn't need to set the kernel lock because
2315
 *  it is set by the callees.
L
Linus Torvalds 已提交
2316 2317
 */

2318
SYSCALL_DEFINE2(socketcall, int, call, unsigned long __user *, args)
L
Linus Torvalds 已提交
2319
{
2320
	unsigned long a[AUDITSC_ARGS];
2321
	unsigned long a0, a1;
L
Linus Torvalds 已提交
2322
	int err;
2323
	unsigned int len;
L
Linus Torvalds 已提交
2324

2325
	if (call < 1 || call > SYS_SENDMMSG)
L
Linus Torvalds 已提交
2326 2327
		return -EINVAL;

2328 2329 2330 2331
	len = nargs[call];
	if (len > sizeof(a))
		return -EINVAL;

L
Linus Torvalds 已提交
2332
	/* copy_from_user should be SMP safe. */
2333
	if (copy_from_user(a, args, len))
L
Linus Torvalds 已提交
2334
		return -EFAULT;
2335

2336 2337 2338
	err = audit_socketcall(nargs[call] / sizeof(unsigned long), a);
	if (err)
		return err;
2339

2340 2341 2342 2343 2344 2345 2346 2347 2348 2349 2350 2351 2352 2353 2354 2355 2356
	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
2357 2358
		err = sys_accept4(a0, (struct sockaddr __user *)a1,
				  (int __user *)a[2], 0);
2359 2360 2361 2362 2363 2364 2365 2366 2367 2368 2369 2370 2371 2372 2373 2374 2375 2376 2377 2378 2379 2380 2381 2382 2383 2384 2385 2386 2387 2388 2389 2390 2391 2392 2393 2394 2395 2396 2397 2398 2399
		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:
2400
		err = sys_sendmsg(a0, (struct user_msghdr __user *)a1, a[2]);
2401
		break;
2402 2403 2404
	case SYS_SENDMMSG:
		err = sys_sendmmsg(a0, (struct mmsghdr __user *)a1, a[2], a[3]);
		break;
2405
	case SYS_RECVMSG:
2406
		err = sys_recvmsg(a0, (struct user_msghdr __user *)a1, a[2]);
2407
		break;
2408 2409 2410 2411
	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
2412 2413 2414
	case SYS_ACCEPT4:
		err = sys_accept4(a0, (struct sockaddr __user *)a1,
				  (int __user *)a[2], a[3]);
U
Ulrich Drepper 已提交
2415
		break;
2416 2417 2418
	default:
		err = -EINVAL;
		break;
L
Linus Torvalds 已提交
2419 2420 2421 2422
	}
	return err;
}

2423
#endif				/* __ARCH_WANT_SYS_SOCKETCALL */
L
Linus Torvalds 已提交
2424

2425 2426 2427 2428
/**
 *	sock_register - add a socket protocol handler
 *	@ops: description of protocol
 *
L
Linus Torvalds 已提交
2429 2430
 *	This function is called by a protocol handler that wants to
 *	advertise its address family, and have it linked into the
2431
 *	socket interface. The value ops->family corresponds to the
2432
 *	socket system call protocol family.
L
Linus Torvalds 已提交
2433
 */
2434
int sock_register(const struct net_proto_family *ops)
L
Linus Torvalds 已提交
2435 2436 2437 2438
{
	int err;

	if (ops->family >= NPROTO) {
2439
		pr_crit("protocol %d >= NPROTO(%d)\n", ops->family, NPROTO);
L
Linus Torvalds 已提交
2440 2441
		return -ENOBUFS;
	}
2442 2443

	spin_lock(&net_family_lock);
2444 2445
	if (rcu_dereference_protected(net_families[ops->family],
				      lockdep_is_held(&net_family_lock)))
2446 2447
		err = -EEXIST;
	else {
2448
		rcu_assign_pointer(net_families[ops->family], ops);
L
Linus Torvalds 已提交
2449 2450
		err = 0;
	}
2451 2452
	spin_unlock(&net_family_lock);

2453
	pr_info("NET: Registered protocol family %d\n", ops->family);
L
Linus Torvalds 已提交
2454 2455
	return err;
}
2456
EXPORT_SYMBOL(sock_register);
L
Linus Torvalds 已提交
2457

2458 2459 2460 2461
/**
 *	sock_unregister - remove a protocol handler
 *	@family: protocol family to remove
 *
L
Linus Torvalds 已提交
2462 2463
 *	This function is called by a protocol handler that wants to
 *	remove its address family, and have it unlinked from the
2464 2465 2466 2467 2468 2469
 *	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
2470
 */
2471
void sock_unregister(int family)
L
Linus Torvalds 已提交
2472
{
2473
	BUG_ON(family < 0 || family >= NPROTO);
L
Linus Torvalds 已提交
2474

2475
	spin_lock(&net_family_lock);
2476
	RCU_INIT_POINTER(net_families[family], NULL);
2477 2478 2479 2480
	spin_unlock(&net_family_lock);

	synchronize_rcu();

2481
	pr_info("NET: Unregistered protocol family %d\n", family);
L
Linus Torvalds 已提交
2482
}
2483
EXPORT_SYMBOL(sock_unregister);
L
Linus Torvalds 已提交
2484

2485
static int __init sock_init(void)
L
Linus Torvalds 已提交
2486
{
N
Nick Piggin 已提交
2487
	int err;
2488 2489 2490 2491 2492 2493
	/*
	 *      Initialize the network sysctl infrastructure.
	 */
	err = net_sysctl_init();
	if (err)
		goto out;
N
Nick Piggin 已提交
2494

L
Linus Torvalds 已提交
2495
	/*
2496
	 *      Initialize skbuff SLAB cache
L
Linus Torvalds 已提交
2497 2498 2499 2500
	 */
	skb_init();

	/*
2501
	 *      Initialize the protocols module.
L
Linus Torvalds 已提交
2502 2503 2504
	 */

	init_inodecache();
N
Nick Piggin 已提交
2505 2506 2507 2508

	err = register_filesystem(&sock_fs_type);
	if (err)
		goto out_fs;
L
Linus Torvalds 已提交
2509
	sock_mnt = kern_mount(&sock_fs_type);
N
Nick Piggin 已提交
2510 2511 2512 2513
	if (IS_ERR(sock_mnt)) {
		err = PTR_ERR(sock_mnt);
		goto out_mount;
	}
2514 2515

	/* The real protocol initialization is performed in later initcalls.
L
Linus Torvalds 已提交
2516 2517 2518
	 */

#ifdef CONFIG_NETFILTER
2519 2520 2521
	err = netfilter_init();
	if (err)
		goto out;
L
Linus Torvalds 已提交
2522
#endif
2523

2524
	ptp_classifier_init();
2525

N
Nick Piggin 已提交
2526 2527 2528 2529 2530 2531 2532
out:
	return err;

out_mount:
	unregister_filesystem(&sock_fs_type);
out_fs:
	goto out;
L
Linus Torvalds 已提交
2533 2534
}

2535 2536
core_initcall(sock_init);	/* early initcall */

L
Linus Torvalds 已提交
2537 2538 2539 2540 2541 2542
#ifdef CONFIG_PROC_FS
void socket_seq_show(struct seq_file *seq)
{
	int cpu;
	int counter = 0;

2543
	for_each_possible_cpu(cpu)
2544
	    counter += per_cpu(sockets_in_use, cpu);
L
Linus Torvalds 已提交
2545 2546 2547 2548 2549 2550 2551

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

	seq_printf(seq, "sockets: used %d\n", counter);
}
2552
#endif				/* CONFIG_PROC_FS */
L
Linus Torvalds 已提交
2553

2554
#ifdef CONFIG_COMPAT
2555
static int do_siocgstamp(struct net *net, struct socket *sock,
2556
			 unsigned int cmd, void __user *up)
2557 2558 2559 2560 2561 2562
{
	mm_segment_t old_fs = get_fs();
	struct timeval ktv;
	int err;

	set_fs(KERNEL_DS);
2563
	err = sock_do_ioctl(net, sock, cmd, (unsigned long)&ktv);
2564
	set_fs(old_fs);
2565
	if (!err)
2566
		err = compat_put_timeval(&ktv, up);
2567

2568 2569 2570
	return err;
}

2571
static int do_siocgstampns(struct net *net, struct socket *sock,
2572
			   unsigned int cmd, void __user *up)
2573 2574 2575 2576 2577 2578
{
	mm_segment_t old_fs = get_fs();
	struct timespec kts;
	int err;

	set_fs(KERNEL_DS);
2579
	err = sock_do_ioctl(net, sock, cmd, (unsigned long)&kts);
2580
	set_fs(old_fs);
2581
	if (!err)
2582
		err = compat_put_timespec(&kts, up);
2583

2584 2585 2586
	return err;
}

2587
static int dev_ifname32(struct net *net, struct compat_ifreq __user *uifr32)
2588 2589 2590 2591 2592
{
	struct ifreq __user *uifr;
	int err;

	uifr = compat_alloc_user_space(sizeof(struct ifreq));
2593
	if (copy_in_user(uifr, uifr32, sizeof(struct compat_ifreq)))
2594 2595
		return -EFAULT;

2596
	err = dev_ioctl(net, SIOCGIFNAME, uifr);
2597 2598 2599
	if (err)
		return err;

2600
	if (copy_in_user(uifr32, uifr, sizeof(struct compat_ifreq)))
2601 2602 2603 2604 2605
		return -EFAULT;

	return 0;
}

2606
static int dev_ifconf(struct net *net, struct compat_ifconf __user *uifc32)
2607
{
2608
	struct compat_ifconf ifc32;
2609 2610
	struct ifconf ifc;
	struct ifconf __user *uifc;
2611
	struct compat_ifreq __user *ifr32;
2612 2613 2614 2615
	struct ifreq __user *ifr;
	unsigned int i, j;
	int err;

2616
	if (copy_from_user(&ifc32, uifc32, sizeof(struct compat_ifconf)))
2617 2618
		return -EFAULT;

2619
	memset(&ifc, 0, sizeof(ifc));
2620 2621 2622 2623 2624 2625
	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 {
2626 2627
		size_t len = ((ifc32.ifc_len / sizeof(struct compat_ifreq)) + 1) *
			sizeof(struct ifreq);
2628 2629 2630 2631
		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);
2632
		for (i = 0; i < ifc32.ifc_len; i += sizeof(struct compat_ifreq)) {
2633
			if (copy_in_user(ifr, ifr32, sizeof(struct compat_ifreq)))
2634 2635 2636 2637 2638 2639 2640 2641
				return -EFAULT;
			ifr++;
			ifr32++;
		}
	}
	if (copy_to_user(uifc, &ifc, sizeof(struct ifconf)))
		return -EFAULT;

2642
	err = dev_ioctl(net, SIOCGIFCONF, uifc);
2643 2644 2645 2646 2647 2648 2649 2650 2651
	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;
2652 2653 2654
	     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)))
2655 2656 2657 2658 2659 2660 2661 2662 2663 2664
			return -EFAULT;
		ifr32++;
		ifr++;
	}

	if (ifc32.ifcbuf == 0) {
		/* Translate from 64-bit structure multiple to
		 * a 32-bit one.
		 */
		i = ifc.ifc_len;
2665
		i = ((i / sizeof(struct ifreq)) * sizeof(struct compat_ifreq));
2666 2667 2668 2669
		ifc32.ifc_len = i;
	} else {
		ifc32.ifc_len = i;
	}
2670
	if (copy_to_user(uifc32, &ifc32, sizeof(struct compat_ifconf)))
2671 2672 2673 2674 2675
		return -EFAULT;

	return 0;
}

2676
static int ethtool_ioctl(struct net *net, struct compat_ifreq __user *ifr32)
2677
{
2678 2679 2680 2681
	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;
2682
	struct ifreq __user *ifr;
2683 2684
	u32 rule_cnt = 0, actual_rule_cnt;
	u32 ethcmd;
2685
	u32 data;
2686
	int ret;
2687

2688 2689
	if (get_user(data, &ifr32->ifr_ifru.ifru_data))
		return -EFAULT;
2690

2691 2692 2693
	compat_rxnfc = compat_ptr(data);

	if (get_user(ethcmd, &compat_rxnfc->cmd))
2694 2695
		return -EFAULT;

2696 2697 2698 2699 2700 2701 2702 2703 2704 2705 2706 2707 2708 2709 2710 2711 2712
	/* 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:
2713
	case ETHTOOL_SRXCLSRLINS:
2714 2715 2716 2717 2718 2719 2720 2721 2722
		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);
2723
	rxnfc = (void __user *)ifr + ALIGN(sizeof(struct ifreq), 8);
2724 2725

	if (copy_in_user(&ifr->ifr_name, &ifr32->ifr_name, IFNAMSIZ))
2726 2727
		return -EFAULT;

2728 2729
	if (put_user(convert_in ? rxnfc : compat_ptr(data),
		     &ifr->ifr_ifru.ifru_data))
2730 2731
		return -EFAULT;

2732
	if (convert_in) {
2733
		/* We expect there to be holes between fs.m_ext and
2734 2735
		 * fs.ring_cookie and at the end of fs, but nowhere else.
		 */
2736 2737 2738 2739
		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));
2740 2741 2742 2743 2744 2745 2746
		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,
2747 2748
				 (void __user *)(&rxnfc->fs.m_ext + 1) -
				 (void __user *)rxnfc) ||
2749 2750
		    copy_in_user(&rxnfc->fs.ring_cookie,
				 &compat_rxnfc->fs.ring_cookie,
2751 2752
				 (void __user *)(&rxnfc->fs.location + 1) -
				 (void __user *)&rxnfc->fs.ring_cookie) ||
2753 2754 2755 2756 2757 2758 2759 2760 2761 2762 2763
		    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,
2764 2765
				 (const void __user *)(&rxnfc->fs.m_ext + 1) -
				 (const void __user *)rxnfc) ||
2766 2767
		    copy_in_user(&compat_rxnfc->fs.ring_cookie,
				 &rxnfc->fs.ring_cookie,
2768 2769
				 (const void __user *)(&rxnfc->fs.location + 1) -
				 (const void __user *)&rxnfc->fs.ring_cookie) ||
2770 2771 2772 2773 2774 2775 2776 2777 2778 2779 2780 2781 2782 2783 2784 2785 2786 2787 2788 2789 2790 2791 2792 2793 2794
		    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;
2795 2796
}

2797 2798 2799 2800 2801 2802
static int compat_siocwandev(struct net *net, struct compat_ifreq __user *uifr32)
{
	void __user *uptr;
	compat_uptr_t uptr32;
	struct ifreq __user *uifr;

2803
	uifr = compat_alloc_user_space(sizeof(*uifr));
2804 2805 2806 2807 2808 2809 2810 2811 2812 2813 2814 2815 2816 2817
	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);
}

2818 2819
static int bond_ioctl(struct net *net, unsigned int cmd,
			 struct compat_ifreq __user *ifr32)
2820 2821 2822 2823 2824 2825 2826 2827 2828 2829
{
	struct ifreq kifr;
	mm_segment_t old_fs;
	int err;

	switch (cmd) {
	case SIOCBONDENSLAVE:
	case SIOCBONDRELEASE:
	case SIOCBONDSETHWADDR:
	case SIOCBONDCHANGEACTIVE:
2830
		if (copy_from_user(&kifr, ifr32, sizeof(struct compat_ifreq)))
2831 2832 2833
			return -EFAULT;

		old_fs = get_fs();
2834
		set_fs(KERNEL_DS);
2835 2836
		err = dev_ioctl(net, cmd,
				(struct ifreq __user __force *) &kifr);
2837
		set_fs(old_fs);
2838 2839 2840

		return err;
	default:
2841
		return -ENOIOCTLCMD;
2842
	}
2843 2844
}

2845 2846
/* 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,
2847
				 struct compat_ifreq __user *u_ifreq32)
2848 2849 2850 2851 2852 2853 2854 2855 2856
{
	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;
2857
	if (get_user(data32, &u_ifreq32->ifr_ifru.ifru_data))
2858 2859 2860 2861 2862 2863 2864 2865
		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;
2866
	if (put_user(data64, &u_ifreq64->ifr_ifru.ifru_data))
2867 2868
		return -EFAULT;

2869
	return dev_ioctl(net, cmd, u_ifreq64);
2870 2871
}

2872 2873
static int dev_ifsioc(struct net *net, struct socket *sock,
			 unsigned int cmd, struct compat_ifreq __user *uifr32)
2874
{
2875
	struct ifreq __user *uifr;
2876 2877
	int err;

2878 2879 2880 2881 2882 2883
	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);

2884 2885 2886 2887 2888 2889 2890 2891 2892 2893 2894 2895
	if (!err) {
		switch (cmd) {
		case SIOCGIFFLAGS:
		case SIOCGIFMETRIC:
		case SIOCGIFMTU:
		case SIOCGIFMEM:
		case SIOCGIFHWADDR:
		case SIOCGIFINDEX:
		case SIOCGIFADDR:
		case SIOCGIFBRDADDR:
		case SIOCGIFDSTADDR:
		case SIOCGIFNETMASK:
2896
		case SIOCGIFPFLAGS:
2897
		case SIOCGIFTXQLEN:
2898 2899
		case SIOCGMIIPHY:
		case SIOCGMIIREG:
2900
			if (copy_in_user(uifr32, uifr, sizeof(*uifr32)))
2901 2902 2903 2904 2905 2906 2907
				err = -EFAULT;
			break;
		}
	}
	return err;
}

2908 2909 2910 2911 2912 2913 2914 2915 2916 2917
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));
2918 2919 2920 2921 2922 2923
	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);
2924 2925 2926 2927
	if (err)
		return -EFAULT;

	old_fs = get_fs();
2928
	set_fs(KERNEL_DS);
2929
	err = dev_ioctl(net, cmd, (void  __user __force *)&ifr);
2930
	set_fs(old_fs);
2931 2932 2933

	if (cmd == SIOCGIFMAP && !err) {
		err = copy_to_user(uifr32, &ifr, sizeof(ifr.ifr_name));
2934 2935 2936 2937 2938 2939
		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);
2940 2941 2942 2943 2944 2945
		if (err)
			err = -EFAULT;
	}
	return err;
}

2946
struct rtentry32 {
2947
	u32		rt_pad1;
2948 2949 2950
	struct sockaddr rt_dst;         /* target address               */
	struct sockaddr rt_gateway;     /* gateway addr (RTF_GATEWAY)   */
	struct sockaddr rt_genmask;     /* target network mask (IP)     */
2951 2952 2953 2954 2955 2956 2957
	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! */
2958
	/* char * */ u32 rt_dev;        /* forcing the device at add    */
2959 2960
	u32		rt_mtu;         /* per route MTU/Window         */
	u32		rt_window;      /* Window clamping              */
2961 2962 2963 2964 2965 2966 2967 2968 2969 2970 2971 2972 2973 2974 2975 2976
	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;
};

2977 2978
static int routing_ioctl(struct net *net, struct socket *sock,
			 unsigned int cmd, void __user *argp)
2979 2980 2981 2982 2983 2984 2985 2986 2987
{
	int ret;
	void *r = NULL;
	struct in6_rtmsg r6;
	struct rtentry r4;
	char devname[16];
	u32 rtdev;
	mm_segment_t old_fs = get_fs();

2988 2989
	if (sock && sock->sk && sock->sk->sk_family == AF_INET6) { /* ipv6 */
		struct in6_rtmsg32 __user *ur6 = argp;
2990
		ret = copy_from_user(&r6.rtmsg_dst, &(ur6->rtmsg_dst),
2991
			3 * sizeof(struct in6_addr));
2992 2993 2994 2995 2996 2997 2998
		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));
2999 3000 3001

		r = (void *) &r6;
	} else { /* ipv4 */
3002
		struct rtentry32 __user *ur4 = argp;
3003
		ret = copy_from_user(&r4.rt_dst, &(ur4->rt_dst),
3004
					3 * sizeof(struct sockaddr));
3005 3006 3007 3008 3009 3010
		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));
3011
		if (rtdev) {
3012
			ret |= copy_from_user(devname, compat_ptr(rtdev), 15);
3013 3014
			r4.rt_dev = (char __user __force *)devname;
			devname[15] = 0;
3015 3016 3017 3018 3019 3020 3021 3022 3023 3024 3025
		} else
			r4.rt_dev = NULL;

		r = (void *) &r4;
	}

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

3026
	set_fs(KERNEL_DS);
3027
	ret = sock_do_ioctl(net, sock, cmd, (unsigned long) r);
3028
	set_fs(old_fs);
3029 3030 3031 3032 3033 3034 3035

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
3036
 * use compatible ioctls
3037
 */
3038
static int old_bridge_ioctl(compat_ulong_t __user *argp)
3039
{
3040
	compat_ulong_t tmp;
3041

3042
	if (get_user(tmp, argp))
3043 3044 3045 3046 3047 3048
		return -EFAULT;
	if (tmp == BRCTL_GET_VERSION)
		return BRCTL_VERSION + 1;
	return -EINVAL;
}

3049 3050 3051 3052 3053 3054
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);
3055

3056
	if (cmd >= SIOCDEVPRIVATE && cmd <= (SIOCDEVPRIVATE + 15))
3057
		return compat_ifr_data_ioctl(net, cmd, argp);
3058 3059 3060 3061 3062 3063 3064 3065 3066 3067 3068

	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);
3069 3070
	case SIOCWANDEV:
		return compat_siocwandev(net, argp);
3071 3072 3073
	case SIOCGIFMAP:
	case SIOCSIFMAP:
		return compat_sioc_ifmap(net, cmd, argp);
3074 3075 3076 3077 3078 3079 3080 3081 3082 3083 3084 3085
	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);
3086 3087
	case SIOCBONDSLAVEINFOQUERY:
	case SIOCBONDINFOQUERY:
3088
	case SIOCSHWTSTAMP:
3089
	case SIOCGHWTSTAMP:
3090
		return compat_ifr_data_ioctl(net, cmd, argp);
3091 3092 3093 3094 3095 3096 3097 3098 3099 3100 3101 3102 3103 3104 3105 3106 3107 3108 3109 3110 3111 3112 3113 3114 3115 3116 3117 3118 3119 3120 3121 3122 3123 3124 3125 3126 3127 3128 3129 3130 3131 3132

	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:
3133 3134 3135 3136
	case SIOCSIFNAME:
	case SIOCGMIIPHY:
	case SIOCGMIIREG:
	case SIOCSMIIREG:
3137
		return dev_ifsioc(net, sock, cmd, argp);
3138

3139 3140 3141 3142
	case SIOCSARP:
	case SIOCGARP:
	case SIOCDARP:
	case SIOCATMARK:
3143 3144 3145
		return sock_do_ioctl(net, sock, cmd, arg);
	}

3146 3147
	return -ENOIOCTLCMD;
}
3148

3149
static long compat_sock_ioctl(struct file *file, unsigned int cmd,
3150
			      unsigned long arg)
3151 3152 3153
{
	struct socket *sock = file->private_data;
	int ret = -ENOIOCTLCMD;
3154 3155 3156 3157 3158
	struct sock *sk;
	struct net *net;

	sk = sock->sk;
	net = sock_net(sk);
3159 3160 3161 3162

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

3163 3164 3165 3166
	if (ret == -ENOIOCTLCMD &&
	    (cmd >= SIOCIWFIRST && cmd <= SIOCIWLAST))
		ret = compat_wext_handle_ioctl(net, cmd, arg);

3167 3168 3169
	if (ret == -ENOIOCTLCMD)
		ret = compat_sock_ioctl_trans(file, sock, cmd, arg);

3170 3171 3172 3173
	return ret;
}
#endif

3174 3175 3176 3177
int kernel_bind(struct socket *sock, struct sockaddr *addr, int addrlen)
{
	return sock->ops->bind(sock, addr, addrlen);
}
3178
EXPORT_SYMBOL(kernel_bind);
3179 3180 3181 3182 3183

int kernel_listen(struct socket *sock, int backlog)
{
	return sock->ops->listen(sock, backlog);
}
3184
EXPORT_SYMBOL(kernel_listen);
3185 3186 3187 3188 3189 3190 3191 3192 3193 3194 3195 3196 3197 3198

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);
3199
		*newsock = NULL;
3200 3201 3202 3203
		goto done;
	}

	(*newsock)->ops = sock->ops;
3204
	__module_get((*newsock)->ops->owner);
3205 3206 3207 3208

done:
	return err;
}
3209
EXPORT_SYMBOL(kernel_accept);
3210 3211

int kernel_connect(struct socket *sock, struct sockaddr *addr, int addrlen,
3212
		   int flags)
3213 3214 3215
{
	return sock->ops->connect(sock, addr, addrlen, flags);
}
3216
EXPORT_SYMBOL(kernel_connect);
3217 3218 3219 3220 3221 3222

int kernel_getsockname(struct socket *sock, struct sockaddr *addr,
			 int *addrlen)
{
	return sock->ops->getname(sock, addr, addrlen, 0);
}
3223
EXPORT_SYMBOL(kernel_getsockname);
3224 3225 3226 3227 3228 3229

int kernel_getpeername(struct socket *sock, struct sockaddr *addr,
			 int *addrlen)
{
	return sock->ops->getname(sock, addr, addrlen, 1);
}
3230
EXPORT_SYMBOL(kernel_getpeername);
3231 3232 3233 3234 3235

int kernel_getsockopt(struct socket *sock, int level, int optname,
			char *optval, int *optlen)
{
	mm_segment_t oldfs = get_fs();
3236 3237
	char __user *uoptval;
	int __user *uoptlen;
3238 3239
	int err;

3240 3241 3242
	uoptval = (char __user __force *) optval;
	uoptlen = (int __user __force *) optlen;

3243 3244
	set_fs(KERNEL_DS);
	if (level == SOL_SOCKET)
3245
		err = sock_getsockopt(sock, level, optname, uoptval, uoptlen);
3246
	else
3247 3248
		err = sock->ops->getsockopt(sock, level, optname, uoptval,
					    uoptlen);
3249 3250 3251
	set_fs(oldfs);
	return err;
}
3252
EXPORT_SYMBOL(kernel_getsockopt);
3253 3254

int kernel_setsockopt(struct socket *sock, int level, int optname,
3255
			char *optval, unsigned int optlen)
3256 3257
{
	mm_segment_t oldfs = get_fs();
3258
	char __user *uoptval;
3259 3260
	int err;

3261 3262
	uoptval = (char __user __force *) optval;

3263 3264
	set_fs(KERNEL_DS);
	if (level == SOL_SOCKET)
3265
		err = sock_setsockopt(sock, level, optname, uoptval, optlen);
3266
	else
3267
		err = sock->ops->setsockopt(sock, level, optname, uoptval,
3268 3269 3270 3271
					    optlen);
	set_fs(oldfs);
	return err;
}
3272
EXPORT_SYMBOL(kernel_setsockopt);
3273 3274 3275 3276 3277 3278 3279 3280 3281

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);
}
3282
EXPORT_SYMBOL(kernel_sendpage);
3283 3284 3285 3286 3287 3288 3289 3290 3291 3292 3293 3294

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;
}
3295
EXPORT_SYMBOL(kernel_sock_ioctl);
3296

3297 3298 3299 3300 3301
int kernel_sock_shutdown(struct socket *sock, enum sock_shutdown_cmd how)
{
	return sock->ops->shutdown(sock, how);
}
EXPORT_SYMBOL(kernel_sock_shutdown);
反馈
建议
客服 返回
顶部