socket.c 79.6 KB
Newer Older
L
Linus Torvalds 已提交
1 2 3 4 5 6
/*
 * NET		An implementation of the SOCKET network access protocol.
 *
 * Version:	@(#)socket.c	1.1.93	18/02/95
 *
 * Authors:	Orest Zborowski, <obz@Kodak.COM>
7
 *		Ross Biro
L
Linus Torvalds 已提交
8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44
 *		Fred N. van Kempen, <waltje@uWalt.NL.Mugnet.ORG>
 *
 * Fixes:
 *		Anonymous	:	NOTSOCK/BADF cleanup. Error fix in
 *					shutdown()
 *		Alan Cox	:	verify_area() fixes
 *		Alan Cox	:	Removed DDI
 *		Jonathan Kamens	:	SOCK_DGRAM reconnect bug
 *		Alan Cox	:	Moved a load of checks to the very
 *					top level.
 *		Alan Cox	:	Move address structures to/from user
 *					mode above the protocol layers.
 *		Rob Janssen	:	Allow 0 length sends.
 *		Alan Cox	:	Asynchronous I/O support (cribbed from the
 *					tty drivers).
 *		Niibe Yutaka	:	Asynchronous I/O for writes (4.4BSD style)
 *		Jeff Uphoff	:	Made max number of sockets command-line
 *					configurable.
 *		Matti Aarnio	:	Made the number of sockets dynamic,
 *					to be allocated when needed, and mr.
 *					Uphoff's max is used as max to be
 *					allowed to allocate.
 *		Linus		:	Argh. removed all the socket allocation
 *					altogether: it's in the inode now.
 *		Alan Cox	:	Made sock_alloc()/sock_release() public
 *					for NetROM and future kernel nfsd type
 *					stuff.
 *		Alan Cox	:	sendmsg/recvmsg basics.
 *		Tom Dyas	:	Export net symbols.
 *		Marcin Dalecki	:	Fixed problems with CONFIG_NET="n".
 *		Alan Cox	:	Added thread locking to sys_* calls
 *					for sockets. May have errors at the
 *					moment.
 *		Kevin Buhr	:	Fixed the dumb errors in the above.
 *		Andi Kleen	:	Some small cleanups, optimizations,
 *					and fixed a copy_from_user() bug.
 *		Tigran Aivazian	:	sys_send(args) calls sys_sendto(args, NULL, 0)
45
 *		Tigran Aivazian	:	Made listen(2) backlog sanity checks
L
Linus Torvalds 已提交
46 47 48 49 50 51 52 53 54 55
 *					protocol-independent
 *
 *
 *		This program is free software; you can redistribute it and/or
 *		modify it under the terms of the GNU General Public License
 *		as published by the Free Software Foundation; either version
 *		2 of the License, or (at your option) any later version.
 *
 *
 *	This module is effectively the top level interface to the BSD socket
56
 *	paradigm.
L
Linus Torvalds 已提交
57 58 59 60 61 62 63 64 65
 *
 *	Based upon Swansea University Computer Society NET3.039
 */

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

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

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

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

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

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

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

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

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

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

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

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

/*
 *	Statistics counters of the socket lists
 */

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

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

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

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

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

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

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

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

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

250
	ei = kmem_cache_alloc(sock_inode_cachep, GFP_KERNEL);
L
Linus Torvalds 已提交
251 252
	if (!ei)
		return NULL;
253 254
	wq = kmalloc(sizeof(*wq), GFP_KERNEL);
	if (!wq) {
255 256 257
		kmem_cache_free(sock_inode_cachep, ei);
		return NULL;
	}
258 259
	init_waitqueue_head(&wq->wait);
	wq->fasync_list = NULL;
260
	wq->flags = 0;
261
	RCU_INIT_POINTER(ei->socket.wq, wq);
262

L
Linus Torvalds 已提交
263 264 265 266 267 268 269 270 271 272 273
	ei->socket.state = SS_UNCONNECTED;
	ei->socket.flags = 0;
	ei->socket.ops = NULL;
	ei->socket.sk = NULL;
	ei->socket.file = NULL;

	return &ei->vfs_inode;
}

static void sock_destroy_inode(struct inode *inode)
{
274
	struct socket_alloc *ei;
275
	struct socket_wq *wq;
276 277

	ei = container_of(inode, struct socket_alloc, vfs_inode);
278
	wq = rcu_dereference_protected(ei->socket.wq, 1);
279
	kfree_rcu(wq, rcu);
280
	kmem_cache_free(sock_inode_cachep, ei);
L
Linus Torvalds 已提交
281 282
}

283
static void init_once(void *foo)
L
Linus Torvalds 已提交
284
{
285
	struct socket_alloc *ei = (struct socket_alloc *)foo;
L
Linus Torvalds 已提交
286

C
Christoph Lameter 已提交
287
	inode_init_once(&ei->vfs_inode);
L
Linus Torvalds 已提交
288
}
289

L
Linus Torvalds 已提交
290 291 292
static int init_inodecache(void)
{
	sock_inode_cachep = kmem_cache_create("sock_inode_cache",
293 294 295 296
					      sizeof(struct socket_alloc),
					      0,
					      (SLAB_HWCACHE_ALIGN |
					       SLAB_RECLAIM_ACCOUNT |
297
					       SLAB_MEM_SPREAD | SLAB_ACCOUNT),
298
					      init_once);
L
Linus Torvalds 已提交
299 300 301 302 303
	if (sock_inode_cachep == NULL)
		return -ENOMEM;
	return 0;
}

304
static const struct super_operations sockfs_ops = {
305 306 307
	.alloc_inode	= sock_alloc_inode,
	.destroy_inode	= sock_destroy_inode,
	.statfs		= simple_statfs,
L
Linus Torvalds 已提交
308 309
};

310 311 312 313 314 315
/*
 * sockfs_dname() is called from d_path().
 */
static char *sockfs_dname(struct dentry *dentry, char *buffer, int buflen)
{
	return dynamic_dname(dentry, buffer, buflen, "socket:[%lu]",
D
David Howells 已提交
316
				d_inode(dentry)->i_ino);
317 318
}

A
Al Viro 已提交
319
static const struct dentry_operations sockfs_dentry_operations = {
320
	.d_dname  = sockfs_dname,
L
Linus Torvalds 已提交
321 322
};

323 324 325 326 327 328 329 330 331 332 333 334 335 336 337
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
338 339 340
/*
 *	Obtains the first available file descriptor and sets it up for use.
 *
341 342
 *	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
343 344 345 346 347 348 349 350 351 352 353 354
 *	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.
 */

355
struct file *sock_alloc_file(struct socket *sock, int flags, const char *dname)
L
Linus Torvalds 已提交
356
{
357
	struct qstr name = { .name = "" };
358
	struct path path;
359
	struct file *file;
L
Linus Torvalds 已提交
360

361 362 363 364 365 366 367
	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);
	}
368
	path.dentry = d_alloc_pseudo(sock_mnt->mnt_sb, &name);
369 370
	if (unlikely(!path.dentry))
		return ERR_PTR(-ENOMEM);
371
	path.mnt = mntget(sock_mnt);
372

373
	d_instantiate(path.dentry, SOCK_INODE(sock));
374

375
	file = alloc_file(&path, FMODE_READ | FMODE_WRITE,
376
		  &socket_file_ops);
377
	if (IS_ERR(file)) {
A
Al Viro 已提交
378
		/* drop dentry, keep inode */
D
David Howells 已提交
379
		ihold(d_inode(path.dentry));
380
		path_put(&path);
381
		return file;
A
Al Viro 已提交
382 383 384
	}

	sock->file = file;
385
	file->f_flags = O_RDWR | (flags & O_NONBLOCK);
386
	file->private_data = sock;
387
	return file;
388
}
389
EXPORT_SYMBOL(sock_alloc_file);
390

391
static int sock_map_fd(struct socket *sock, int flags)
392 393
{
	struct file *newfile;
394 395 396
	int fd = get_unused_fd_flags(flags);
	if (unlikely(fd < 0))
		return fd;
397

398
	newfile = sock_alloc_file(sock, flags, NULL);
399
	if (likely(!IS_ERR(newfile))) {
400
		fd_install(fd, newfile);
401 402
		return fd;
	}
403

404 405
	put_unused_fd(fd);
	return PTR_ERR(newfile);
L
Linus Torvalds 已提交
406 407
}

408
struct socket *sock_from_file(struct file *file, int *err)
409 410 411 412
{
	if (file->f_op == &socket_file_ops)
		return file->private_data;	/* set in sock_map_fd */

E
Eric Dumazet 已提交
413 414
	*err = -ENOTSOCK;
	return NULL;
415
}
416
EXPORT_SYMBOL(sock_from_file);
417

L
Linus Torvalds 已提交
418
/**
419
 *	sockfd_lookup - Go from a file number to its socket slot
L
Linus Torvalds 已提交
420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435
 *	@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;

436 437
	file = fget(fd);
	if (!file) {
L
Linus Torvalds 已提交
438 439 440
		*err = -EBADF;
		return NULL;
	}
441

442 443
	sock = sock_from_file(file, err);
	if (!sock)
L
Linus Torvalds 已提交
444
		fput(file);
445 446
	return sock;
}
447
EXPORT_SYMBOL(sockfd_lookup);
L
Linus Torvalds 已提交
448

449 450
static struct socket *sockfd_lookup_light(int fd, int *err, int *fput_needed)
{
451
	struct fd f = fdget(fd);
452 453
	struct socket *sock;

454
	*err = -EBADF;
455 456 457 458
	if (f.file) {
		sock = sock_from_file(f.file, err);
		if (likely(sock)) {
			*fput_needed = f.flags;
459
			return sock;
460 461
		}
		fdput(f);
L
Linus Torvalds 已提交
462
	}
463
	return NULL;
L
Linus Torvalds 已提交
464 465
}

466 467 468 469 470 471 472 473 474 475 476 477 478 479 480 481 482 483 484 485 486 487 488 489 490 491 492 493 494 495 496 497 498 499 500
#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;

D
David Howells 已提交
501
	len = security_inode_listsecurity(d_inode(dentry), buffer, size);
502 503 504 505 506 507 508 509 510 511 512 513 514 515 516 517 518 519 520 521 522 523 524 525 526 527
	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
528 529
/**
 *	sock_alloc	-	allocate a socket
530
 *
L
Linus Torvalds 已提交
531 532 533 534 535
 *	Allocate a new inode and socket object. The two are bound together
 *	and initialised. The socket is then returned. If we are out of inodes
 *	NULL is returned.
 */

T
Tom Herbert 已提交
536
struct socket *sock_alloc(void)
L
Linus Torvalds 已提交
537
{
538 539
	struct inode *inode;
	struct socket *sock;
L
Linus Torvalds 已提交
540

541
	inode = new_inode_pseudo(sock_mnt->mnt_sb);
L
Linus Torvalds 已提交
542 543 544 545 546
	if (!inode)
		return NULL;

	sock = SOCKET_I(inode);

547
	kmemcheck_annotate_bitfield(sock, type);
548
	inode->i_ino = get_next_ino();
549
	inode->i_mode = S_IFSOCK | S_IRWXUGO;
550 551
	inode->i_uid = current_fsuid();
	inode->i_gid = current_fsgid();
552
	inode->i_op = &sockfs_inode_ops;
L
Linus Torvalds 已提交
553

554
	this_cpu_add(sockets_in_use, 1);
L
Linus Torvalds 已提交
555 556
	return sock;
}
T
Tom Herbert 已提交
557
EXPORT_SYMBOL(sock_alloc);
L
Linus Torvalds 已提交
558 559 560 561 562 563 564

/**
 *	sock_release	-	close a socket
 *	@sock: socket to close
 *
 *	The socket is released from the protocol stack if it has a release
 *	callback, and the inode is then released if the socket is bound to
565
 *	an inode not a file.
L
Linus Torvalds 已提交
566
 */
567

L
Linus Torvalds 已提交
568 569 570 571 572 573 574 575 576 577
void sock_release(struct socket *sock)
{
	if (sock->ops) {
		struct module *owner = sock->ops->owner;

		sock->ops->release(sock);
		sock->ops = NULL;
		module_put(owner);
	}

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

581
	this_cpu_sub(sockets_in_use, 1);
L
Linus Torvalds 已提交
582 583 584 585
	if (!sock->file) {
		iput(SOCK_INODE(sock));
		return;
	}
586
	sock->file = NULL;
L
Linus Torvalds 已提交
587
}
588
EXPORT_SYMBOL(sock_release);
L
Linus Torvalds 已提交
589

590
void __sock_tx_timestamp(const struct sock *sk, __u8 *tx_flags)
591
{
592 593
	u8 flags = *tx_flags;

594
	if (sk->sk_tsflags & SOF_TIMESTAMPING_TX_HARDWARE)
595 596
		flags |= SKBTX_HW_TSTAMP;

597
	if (sk->sk_tsflags & SOF_TIMESTAMPING_TX_SOFTWARE)
598 599
		flags |= SKBTX_SW_TSTAMP;

600
	if (sk->sk_tsflags & SOF_TIMESTAMPING_TX_SCHED)
601 602
		flags |= SKBTX_SCHED_TSTAMP;

603
	if (sk->sk_tsflags & SOF_TIMESTAMPING_TX_ACK)
604
		flags |= SKBTX_ACK_TSTAMP;
605

606
	*tx_flags = flags;
607
}
608
EXPORT_SYMBOL(__sock_tx_timestamp);
609

610
static inline int sock_sendmsg_nosec(struct socket *sock, struct msghdr *msg)
L
Linus Torvalds 已提交
611
{
A
Al Viro 已提交
612
	int ret = sock->ops->sendmsg(sock, msg, msg_data_left(msg));
613 614
	BUG_ON(ret == -EIOCBQUEUED);
	return ret;
L
Linus Torvalds 已提交
615 616
}

617
int sock_sendmsg(struct socket *sock, struct msghdr *msg)
618
{
619
	int err = security_socket_sendmsg(sock, msg,
A
Al Viro 已提交
620
					  msg_data_left(msg));
621

622
	return err ?: sock_sendmsg_nosec(sock, msg);
623
}
624
EXPORT_SYMBOL(sock_sendmsg);
L
Linus Torvalds 已提交
625 626 627 628

int kernel_sendmsg(struct socket *sock, struct msghdr *msg,
		   struct kvec *vec, size_t num, size_t size)
{
629
	iov_iter_kvec(&msg->msg_iter, WRITE | ITER_KVEC, vec, num, size);
630
	return sock_sendmsg(sock, msg);
L
Linus Torvalds 已提交
631
}
632
EXPORT_SYMBOL(kernel_sendmsg);
L
Linus Torvalds 已提交
633

634 635 636 637 638 639
/*
 * 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)
{
640
	int need_software_tstamp = sock_flag(sk, SOCK_RCVTSTAMP);
641
	struct scm_timestamping tss;
642 643 644 645 646 647 648 649 650 651 652 653 654 655 656 657
	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 {
658 659
			struct timespec ts;
			skb_get_timestampns(skb, &ts);
660
			put_cmsg(msg, SOL_SOCKET, SCM_TIMESTAMPNS,
661
				 sizeof(ts), &ts);
662 663 664
		}
	}

665
	memset(&tss, 0, sizeof(tss));
666
	if ((sk->sk_tsflags & SOF_TIMESTAMPING_SOFTWARE) &&
667
	    ktime_to_timespec_cond(skb->tstamp, tss.ts + 0))
668
		empty = 0;
669
	if (shhwtstamps &&
670
	    (sk->sk_tsflags & SOF_TIMESTAMPING_RAW_HARDWARE) &&
671
	    ktime_to_timespec_cond(shhwtstamps->hwtstamp, tss.ts + 2))
672
		empty = 0;
673 674
	if (!empty)
		put_cmsg(msg, SOL_SOCKET,
675
			 SCM_TIMESTAMPING, sizeof(tss), &tss);
676
}
677 678
EXPORT_SYMBOL_GPL(__sock_recv_timestamp);

679 680 681 682 683 684 685 686 687 688 689 690 691 692 693 694
void __sock_recv_wifi_status(struct msghdr *msg, struct sock *sk,
	struct sk_buff *skb)
{
	int ack;

	if (!sock_flag(sk, SOCK_WIFI_STATUS))
		return;
	if (!skb->wifi_acked_valid)
		return;

	ack = skb->wifi_acked;

	put_cmsg(msg, SOL_SOCKET, SCM_WIFI_STATUS, sizeof(ack), &ack);
}
EXPORT_SYMBOL_GPL(__sock_recv_wifi_status);

S
stephen hemminger 已提交
695 696
static inline void sock_recv_drops(struct msghdr *msg, struct sock *sk,
				   struct sk_buff *skb)
697
{
698
	if (sock_flag(sk, SOCK_RXQ_OVFL) && skb && SOCK_SKB_CB(skb)->dropcount)
699
		put_cmsg(msg, SOL_SOCKET, SO_RXQ_OVFL,
700
			sizeof(__u32), &SOCK_SKB_CB(skb)->dropcount);
701 702
}

703
void __sock_recv_ts_and_drops(struct msghdr *msg, struct sock *sk,
704 705 706 707 708
	struct sk_buff *skb)
{
	sock_recv_timestamp(msg, sk, skb);
	sock_recv_drops(msg, sk, skb);
}
709
EXPORT_SYMBOL_GPL(__sock_recv_ts_and_drops);
710

711 712
static inline int sock_recvmsg_nosec(struct socket *sock, struct msghdr *msg,
				     size_t size, int flags)
L
Linus Torvalds 已提交
713
{
714
	return sock->ops->recvmsg(sock, msg, size, flags);
L
Linus Torvalds 已提交
715 716
}

717 718
int sock_recvmsg(struct socket *sock, struct msghdr *msg, size_t size,
		 int flags)
719 720 721
{
	int err = security_socket_recvmsg(sock, msg, size, flags);

722
	return err ?: sock_recvmsg_nosec(sock, msg, size, flags);
L
Linus Torvalds 已提交
723
}
724
EXPORT_SYMBOL(sock_recvmsg);
L
Linus Torvalds 已提交
725

726 727 728 729 730 731 732 733 734 735 736 737 738 739 740
/**
 * 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.
 */
741 742
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
743 744 745 746
{
	mm_segment_t oldfs = get_fs();
	int result;

747
	iov_iter_kvec(&msg->msg_iter, READ | ITER_KVEC, vec, num, size);
L
Linus Torvalds 已提交
748 749 750 751 752
	set_fs(KERNEL_DS);
	result = sock_recvmsg(sock, msg, size, flags);
	set_fs(oldfs);
	return result;
}
753
EXPORT_SYMBOL(kernel_recvmsg);
L
Linus Torvalds 已提交
754

755 756
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
757 758 759 760
{
	struct socket *sock;
	int flags;

761 762
	sock = file->private_data;

763 764 765
	flags = (file->f_flags & O_NONBLOCK) ? MSG_DONTWAIT : 0;
	/* more is a combination of MSG_MORE and MSG_SENDPAGE_NOTLAST */
	flags |= more;
766

767
	return kernel_sendpage(sock, page, offset, size, flags);
768
}
L
Linus Torvalds 已提交
769

J
Jens Axboe 已提交
770
static ssize_t sock_splice_read(struct file *file, loff_t *ppos,
771
				struct pipe_inode_info *pipe, size_t len,
J
Jens Axboe 已提交
772 773 774 775
				unsigned int flags)
{
	struct socket *sock = file->private_data;

776 777 778
	if (unlikely(!sock->ops->splice_read))
		return -EINVAL;

J
Jens Axboe 已提交
779 780 781
	return sock->ops->splice_read(sock, ppos, pipe, len, flags);
}

782
static ssize_t sock_read_iter(struct kiocb *iocb, struct iov_iter *to)
783
{
784 785
	struct file *file = iocb->ki_filp;
	struct socket *sock = file->private_data;
786 787
	struct msghdr msg = {.msg_iter = *to,
			     .msg_iocb = iocb};
788
	ssize_t res;
789

790 791 792 793
	if (file->f_flags & O_NONBLOCK)
		msg.msg_flags = MSG_DONTWAIT;

	if (iocb->ki_pos != 0)
L
Linus Torvalds 已提交
794
		return -ESPIPE;
795

C
Christoph Hellwig 已提交
796
	if (!iov_iter_count(to))	/* Match SYS5 behaviour */
L
Linus Torvalds 已提交
797 798
		return 0;

A
Al Viro 已提交
799
	res = sock_recvmsg(sock, &msg, iov_iter_count(to), msg.msg_flags);
800 801
	*to = msg.msg_iter;
	return res;
L
Linus Torvalds 已提交
802 803
}

804
static ssize_t sock_write_iter(struct kiocb *iocb, struct iov_iter *from)
805
{
806 807
	struct file *file = iocb->ki_filp;
	struct socket *sock = file->private_data;
808 809
	struct msghdr msg = {.msg_iter = *from,
			     .msg_iocb = iocb};
810
	ssize_t res;
L
Linus Torvalds 已提交
811

812
	if (iocb->ki_pos != 0)
813
		return -ESPIPE;
814

815 816 817
	if (file->f_flags & O_NONBLOCK)
		msg.msg_flags = MSG_DONTWAIT;

818 819 820
	if (sock->type == SOCK_SEQPACKET)
		msg.msg_flags |= MSG_EOR;

821
	res = sock_sendmsg(sock, &msg);
822 823
	*from = msg.msg_iter;
	return res;
L
Linus Torvalds 已提交
824 825 826 827 828 829 830
}

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

A
Arjan van de Ven 已提交
831
static DEFINE_MUTEX(br_ioctl_mutex);
832
static int (*br_ioctl_hook) (struct net *, unsigned int cmd, void __user *arg);
L
Linus Torvalds 已提交
833

834
void brioctl_set(int (*hook) (struct net *, unsigned int, void __user *))
L
Linus Torvalds 已提交
835
{
A
Arjan van de Ven 已提交
836
	mutex_lock(&br_ioctl_mutex);
L
Linus Torvalds 已提交
837
	br_ioctl_hook = hook;
A
Arjan van de Ven 已提交
838
	mutex_unlock(&br_ioctl_mutex);
L
Linus Torvalds 已提交
839 840 841
}
EXPORT_SYMBOL(brioctl_set);

A
Arjan van de Ven 已提交
842
static DEFINE_MUTEX(vlan_ioctl_mutex);
843
static int (*vlan_ioctl_hook) (struct net *, void __user *arg);
L
Linus Torvalds 已提交
844

845
void vlan_ioctl_set(int (*hook) (struct net *, void __user *))
L
Linus Torvalds 已提交
846
{
A
Arjan van de Ven 已提交
847
	mutex_lock(&vlan_ioctl_mutex);
L
Linus Torvalds 已提交
848
	vlan_ioctl_hook = hook;
A
Arjan van de Ven 已提交
849
	mutex_unlock(&vlan_ioctl_mutex);
L
Linus Torvalds 已提交
850 851 852
}
EXPORT_SYMBOL(vlan_ioctl_set);

A
Arjan van de Ven 已提交
853
static DEFINE_MUTEX(dlci_ioctl_mutex);
854
static int (*dlci_ioctl_hook) (unsigned int, void __user *);
L
Linus Torvalds 已提交
855

856
void dlci_ioctl_set(int (*hook) (unsigned int, void __user *))
L
Linus Torvalds 已提交
857
{
A
Arjan van de Ven 已提交
858
	mutex_lock(&dlci_ioctl_mutex);
L
Linus Torvalds 已提交
859
	dlci_ioctl_hook = hook;
A
Arjan van de Ven 已提交
860
	mutex_unlock(&dlci_ioctl_mutex);
L
Linus Torvalds 已提交
861 862 863
}
EXPORT_SYMBOL(dlci_ioctl_set);

864 865 866 867 868 869 870 871 872 873 874 875 876 877 878 879 880 881
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
882 883 884 885 886 887 888 889
/*
 *	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;
890
	struct sock *sk;
L
Linus Torvalds 已提交
891 892
	void __user *argp = (void __user *)arg;
	int pid, err;
893
	struct net *net;
L
Linus Torvalds 已提交
894

895
	sock = file->private_data;
896
	sk = sock->sk;
897
	net = sock_net(sk);
L
Linus Torvalds 已提交
898
	if (cmd >= SIOCDEVPRIVATE && cmd <= (SIOCDEVPRIVATE + 15)) {
899
		err = dev_ioctl(net, cmd, argp);
L
Linus Torvalds 已提交
900
	} else
J
Johannes Berg 已提交
901
#ifdef CONFIG_WEXT_CORE
L
Linus Torvalds 已提交
902
	if (cmd >= SIOCIWFIRST && cmd <= SIOCIWLAST) {
903
		err = dev_ioctl(net, cmd, argp);
L
Linus Torvalds 已提交
904
	} else
J
Johannes Berg 已提交
905
#endif
906
		switch (cmd) {
L
Linus Torvalds 已提交
907 908 909 910 911
		case FIOSETOWN:
		case SIOCSPGRP:
			err = -EFAULT;
			if (get_user(pid, (int __user *)argp))
				break;
912 913
			f_setown(sock->file, pid, 1);
			err = 0;
L
Linus Torvalds 已提交
914 915 916
			break;
		case FIOGETOWN:
		case SIOCGPGRP:
917
			err = put_user(f_getown(sock->file),
918
				       (int __user *)argp);
L
Linus Torvalds 已提交
919 920 921 922 923 924 925 926 927
			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
928
			mutex_lock(&br_ioctl_mutex);
929
			if (br_ioctl_hook)
930
				err = br_ioctl_hook(net, cmd, argp);
A
Arjan van de Ven 已提交
931
			mutex_unlock(&br_ioctl_mutex);
L
Linus Torvalds 已提交
932 933 934 935 936 937 938
			break;
		case SIOCGIFVLAN:
		case SIOCSIFVLAN:
			err = -ENOPKG;
			if (!vlan_ioctl_hook)
				request_module("8021q");

A
Arjan van de Ven 已提交
939
			mutex_lock(&vlan_ioctl_mutex);
L
Linus Torvalds 已提交
940
			if (vlan_ioctl_hook)
941
				err = vlan_ioctl_hook(net, argp);
A
Arjan van de Ven 已提交
942
			mutex_unlock(&vlan_ioctl_mutex);
L
Linus Torvalds 已提交
943 944 945 946 947 948 949
			break;
		case SIOCADDDLCI:
		case SIOCDELDLCI:
			err = -ENOPKG;
			if (!dlci_ioctl_hook)
				request_module("dlci");

950 951
			mutex_lock(&dlci_ioctl_mutex);
			if (dlci_ioctl_hook)
L
Linus Torvalds 已提交
952
				err = dlci_ioctl_hook(cmd, argp);
953
			mutex_unlock(&dlci_ioctl_mutex);
L
Linus Torvalds 已提交
954 955
			break;
		default:
956
			err = sock_do_ioctl(net, sock, cmd, arg);
L
Linus Torvalds 已提交
957
			break;
958
		}
L
Linus Torvalds 已提交
959 960 961 962 963 964 965
	return err;
}

int sock_create_lite(int family, int type, int protocol, struct socket **res)
{
	int err;
	struct socket *sock = NULL;
966

L
Linus Torvalds 已提交
967 968 969 970 971 972 973 974 975 976 977
	err = security_socket_create(family, type, protocol, 1);
	if (err)
		goto out;

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

	sock->type = type;
V
Venkat Yekkirala 已提交
978 979 980 981
	err = security_socket_post_create(sock, family, type, protocol, 1);
	if (err)
		goto out_release;

L
Linus Torvalds 已提交
982 983 984
out:
	*res = sock;
	return err;
V
Venkat Yekkirala 已提交
985 986 987 988
out_release:
	sock_release(sock);
	sock = NULL;
	goto out;
L
Linus Torvalds 已提交
989
}
990
EXPORT_SYMBOL(sock_create_lite);
L
Linus Torvalds 已提交
991 992

/* No kernel lock held - perfect */
993
static unsigned int sock_poll(struct file *file, poll_table *wait)
L
Linus Torvalds 已提交
994
{
995
	unsigned int busy_flag = 0;
L
Linus Torvalds 已提交
996 997 998
	struct socket *sock;

	/*
999
	 *      We can't return errors to poll, so it's either yes or no.
L
Linus Torvalds 已提交
1000
	 */
1001
	sock = file->private_data;
1002

1003
	if (sk_can_busy_loop(sock->sk)) {
1004
		/* this socket can poll_ll so tell the system call */
1005
		busy_flag = POLL_BUSY_LOOP;
1006 1007

		/* once, only if requested by syscall */
1008 1009
		if (wait && (wait->_key & POLL_BUSY_LOOP))
			sk_busy_loop(sock->sk, 1);
1010 1011
	}

1012
	return busy_flag | sock->ops->poll(file, sock, wait);
L
Linus Torvalds 已提交
1013 1014
}

1015
static int sock_mmap(struct file *file, struct vm_area_struct *vma)
L
Linus Torvalds 已提交
1016
{
1017
	struct socket *sock = file->private_data;
L
Linus Torvalds 已提交
1018 1019 1020 1021

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

1022
static int sock_close(struct inode *inode, struct file *filp)
L
Linus Torvalds 已提交
1023 1024 1025 1026 1027 1028 1029 1030 1031 1032 1033 1034 1035
{
	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)
1036
 *	   or under socket lock
L
Linus Torvalds 已提交
1037 1038 1039 1040
 */

static int sock_fasync(int fd, struct file *filp, int on)
{
1041 1042
	struct socket *sock = filp->private_data;
	struct sock *sk = sock->sk;
1043
	struct socket_wq *wq;
L
Linus Torvalds 已提交
1044

1045
	if (sk == NULL)
L
Linus Torvalds 已提交
1046 1047 1048
		return -EINVAL;

	lock_sock(sk);
1049 1050
	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
1051

1052
	if (!wq->fasync_list)
1053 1054
		sock_reset_flag(sk, SOCK_FASYNC);
	else
E
Eric Dumazet 已提交
1055
		sock_set_flag(sk, SOCK_FASYNC);
L
Linus Torvalds 已提交
1056

1057
	release_sock(sk);
L
Linus Torvalds 已提交
1058 1059 1060
	return 0;
}

1061
/* This function may be called only under rcu_lock */
L
Linus Torvalds 已提交
1062

1063
int sock_wake_async(struct socket_wq *wq, int how, int band)
L
Linus Torvalds 已提交
1064
{
1065
	if (!wq || !wq->fasync_list)
L
Linus Torvalds 已提交
1066
		return -1;
1067

1068
	switch (how) {
1069
	case SOCK_WAKE_WAITD:
1070
		if (test_bit(SOCKWQ_ASYNC_WAITDATA, &wq->flags))
L
Linus Torvalds 已提交
1071 1072
			break;
		goto call_kill;
1073
	case SOCK_WAKE_SPACE:
1074
		if (!test_and_clear_bit(SOCKWQ_ASYNC_NOSPACE, &wq->flags))
L
Linus Torvalds 已提交
1075 1076
			break;
		/* fall through */
1077
	case SOCK_WAKE_IO:
1078
call_kill:
1079
		kill_fasync(&wq->fasync_list, SIGIO, band);
L
Linus Torvalds 已提交
1080
		break;
1081
	case SOCK_WAKE_URG:
1082
		kill_fasync(&wq->fasync_list, SIGURG, band);
L
Linus Torvalds 已提交
1083
	}
1084

L
Linus Torvalds 已提交
1085 1086
	return 0;
}
1087
EXPORT_SYMBOL(sock_wake_async);
L
Linus Torvalds 已提交
1088

P
Pavel Emelyanov 已提交
1089
int __sock_create(struct net *net, int family, int type, int protocol,
1090
			 struct socket **res, int kern)
L
Linus Torvalds 已提交
1091 1092 1093
{
	int err;
	struct socket *sock;
1094
	const struct net_proto_family *pf;
L
Linus Torvalds 已提交
1095 1096

	/*
1097
	 *      Check protocol is in range
L
Linus Torvalds 已提交
1098 1099 1100 1101 1102 1103 1104 1105 1106 1107 1108 1109
	 */
	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) {
1110
		static int warned;
L
Linus Torvalds 已提交
1111 1112
		if (!warned) {
			warned = 1;
1113 1114
			pr_info("%s uses obsolete (PF_INET,SOCK_PACKET)\n",
				current->comm);
L
Linus Torvalds 已提交
1115 1116 1117 1118 1119 1120 1121
		}
		family = PF_PACKET;
	}

	err = security_socket_create(family, type, protocol, kern);
	if (err)
		return err;
1122

1123 1124 1125 1126 1127 1128 1129
	/*
	 *	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) {
1130
		net_warn_ratelimited("socket: no more sockets\n");
1131 1132 1133 1134 1135 1136
		return -ENFILE;	/* Not exactly a match, but its the
				   closest posix thing */
	}

	sock->type = type;

1137
#ifdef CONFIG_MODULES
1138 1139 1140
	/* 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
1141 1142 1143
	 * requested real, full-featured networking support upon configuration.
	 * Otherwise module support will break!
	 */
E
Eric Dumazet 已提交
1144
	if (rcu_access_pointer(net_families[family]) == NULL)
1145
		request_module("net-pf-%d", family);
L
Linus Torvalds 已提交
1146 1147
#endif

1148 1149 1150 1151 1152
	rcu_read_lock();
	pf = rcu_dereference(net_families[family]);
	err = -EAFNOSUPPORT;
	if (!pf)
		goto out_release;
L
Linus Torvalds 已提交
1153 1154 1155 1156 1157

	/*
	 * We will call the ->create function, that possibly is in a loadable
	 * module, so we have to bump that loadable module refcnt first.
	 */
1158
	if (!try_module_get(pf->owner))
L
Linus Torvalds 已提交
1159 1160
		goto out_release;

1161 1162 1163
	/* Now protected by module ref count */
	rcu_read_unlock();

1164
	err = pf->create(net, sock, protocol, kern);
1165
	if (err < 0)
L
Linus Torvalds 已提交
1166
		goto out_module_put;
1167

L
Linus Torvalds 已提交
1168 1169 1170 1171
	/*
	 * Now to bump the refcnt of the [loadable] module that owns this
	 * socket at sock_release time we decrement its refcnt.
	 */
1172 1173 1174
	if (!try_module_get(sock->ops->owner))
		goto out_module_busy;

L
Linus Torvalds 已提交
1175 1176 1177 1178
	/*
	 * Now that we're done with the ->create function, the [loadable]
	 * module can have its refcnt decremented
	 */
1179
	module_put(pf->owner);
V
Venkat Yekkirala 已提交
1180 1181
	err = security_socket_post_create(sock, family, type, protocol, kern);
	if (err)
1182
		goto out_sock_release;
1183
	*res = sock;
L
Linus Torvalds 已提交
1184

1185 1186 1187 1188
	return 0;

out_module_busy:
	err = -EAFNOSUPPORT;
L
Linus Torvalds 已提交
1189
out_module_put:
1190 1191 1192
	sock->ops = NULL;
	module_put(pf->owner);
out_sock_release:
L
Linus Torvalds 已提交
1193
	sock_release(sock);
1194 1195 1196 1197 1198
	return err;

out_release:
	rcu_read_unlock();
	goto out_sock_release;
L
Linus Torvalds 已提交
1199
}
P
Pavel Emelyanov 已提交
1200
EXPORT_SYMBOL(__sock_create);
L
Linus Torvalds 已提交
1201 1202 1203

int sock_create(int family, int type, int protocol, struct socket **res)
{
1204
	return __sock_create(current->nsproxy->net_ns, family, type, protocol, res, 0);
L
Linus Torvalds 已提交
1205
}
1206
EXPORT_SYMBOL(sock_create);
L
Linus Torvalds 已提交
1207

1208
int sock_create_kern(struct net *net, int family, int type, int protocol, struct socket **res)
L
Linus Torvalds 已提交
1209
{
1210
	return __sock_create(net, family, type, protocol, res, 1);
L
Linus Torvalds 已提交
1211
}
1212
EXPORT_SYMBOL(sock_create_kern);
L
Linus Torvalds 已提交
1213

1214
SYSCALL_DEFINE3(socket, int, family, int, type, int, protocol)
L
Linus Torvalds 已提交
1215 1216 1217
{
	int retval;
	struct socket *sock;
1218 1219
	int flags;

1220 1221 1222 1223 1224 1225
	/* 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);

1226
	flags = type & ~SOCK_TYPE_MASK;
1227
	if (flags & ~(SOCK_CLOEXEC | SOCK_NONBLOCK))
1228 1229
		return -EINVAL;
	type &= SOCK_TYPE_MASK;
L
Linus Torvalds 已提交
1230

U
Ulrich Drepper 已提交
1231 1232 1233
	if (SOCK_NONBLOCK != O_NONBLOCK && (flags & SOCK_NONBLOCK))
		flags = (flags & ~SOCK_NONBLOCK) | O_NONBLOCK;

L
Linus Torvalds 已提交
1234 1235 1236 1237
	retval = sock_create(family, type, protocol, &sock);
	if (retval < 0)
		goto out;

1238
	retval = sock_map_fd(sock, flags & (O_CLOEXEC | O_NONBLOCK));
L
Linus Torvalds 已提交
1239 1240 1241 1242 1243 1244 1245 1246 1247 1248 1249 1250 1251 1252 1253 1254
	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.
 */

1255 1256
SYSCALL_DEFINE4(socketpair, int, family, int, type, int, protocol,
		int __user *, usockvec)
L
Linus Torvalds 已提交
1257 1258 1259
{
	struct socket *sock1, *sock2;
	int fd1, fd2, err;
A
Al Viro 已提交
1260
	struct file *newfile1, *newfile2;
1261 1262 1263
	int flags;

	flags = type & ~SOCK_TYPE_MASK;
1264
	if (flags & ~(SOCK_CLOEXEC | SOCK_NONBLOCK))
1265 1266
		return -EINVAL;
	type &= SOCK_TYPE_MASK;
L
Linus Torvalds 已提交
1267

U
Ulrich Drepper 已提交
1268 1269 1270
	if (SOCK_NONBLOCK != O_NONBLOCK && (flags & SOCK_NONBLOCK))
		flags = (flags & ~SOCK_NONBLOCK) | O_NONBLOCK;

L
Linus Torvalds 已提交
1271 1272 1273 1274 1275 1276 1277 1278 1279 1280 1281 1282 1283 1284
	/*
	 * 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);
1285
	if (err < 0)
L
Linus Torvalds 已提交
1286 1287
		goto out_release_both;

1288
	fd1 = get_unused_fd_flags(flags);
1289 1290
	if (unlikely(fd1 < 0)) {
		err = fd1;
A
Al Viro 已提交
1291
		goto out_release_both;
1292
	}
1293

1294
	fd2 = get_unused_fd_flags(flags);
1295 1296
	if (unlikely(fd2 < 0)) {
		err = fd2;
1297
		goto out_put_unused_1;
1298 1299
	}

1300
	newfile1 = sock_alloc_file(sock1, flags, NULL);
1301
	if (IS_ERR(newfile1)) {
1302
		err = PTR_ERR(newfile1);
1303
		goto out_put_unused_both;
1304 1305
	}

1306
	newfile2 = sock_alloc_file(sock2, flags, NULL);
1307 1308
	if (IS_ERR(newfile2)) {
		err = PTR_ERR(newfile2);
1309
		goto out_fput_1;
A
Al Viro 已提交
1310 1311
	}

1312 1313 1314 1315 1316 1317 1318 1319
	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
1320
	audit_fd_pair(fd1, fd2);
1321

A
Al Viro 已提交
1322 1323
	fd_install(fd1, newfile1);
	fd_install(fd2, newfile2);
L
Linus Torvalds 已提交
1324 1325 1326 1327
	/* fd1 and fd2 may be already another descriptors.
	 * Not kernel problem.
	 */

1328
	return 0;
L
Linus Torvalds 已提交
1329

1330 1331 1332 1333 1334 1335 1336 1337 1338 1339 1340 1341 1342
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
1343

1344 1345 1346 1347
out_put_unused_both:
	put_unused_fd(fd2);
out_put_unused_1:
	put_unused_fd(fd1);
L
Linus Torvalds 已提交
1348
out_release_both:
1349
	sock_release(sock2);
L
Linus Torvalds 已提交
1350
out_release_1:
1351
	sock_release(sock1);
L
Linus Torvalds 已提交
1352 1353 1354 1355 1356 1357 1358 1359 1360 1361 1362 1363
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).
 */

1364
SYSCALL_DEFINE3(bind, int, fd, struct sockaddr __user *, umyaddr, int, addrlen)
L
Linus Torvalds 已提交
1365 1366
{
	struct socket *sock;
1367
	struct sockaddr_storage address;
1368
	int err, fput_needed;
L
Linus Torvalds 已提交
1369

1370
	sock = sockfd_lookup_light(fd, &err, &fput_needed);
1371
	if (sock) {
1372
		err = move_addr_to_kernel(umyaddr, addrlen, &address);
1373 1374
		if (err >= 0) {
			err = security_socket_bind(sock,
1375
						   (struct sockaddr *)&address,
1376
						   addrlen);
1377 1378
			if (!err)
				err = sock->ops->bind(sock,
1379
						      (struct sockaddr *)
1380
						      &address, addrlen);
L
Linus Torvalds 已提交
1381
		}
1382
		fput_light(sock->file, fput_needed);
1383
	}
L
Linus Torvalds 已提交
1384 1385 1386 1387 1388 1389 1390 1391 1392
	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.
 */

1393
SYSCALL_DEFINE2(listen, int, fd, int, backlog)
L
Linus Torvalds 已提交
1394 1395
{
	struct socket *sock;
1396
	int err, fput_needed;
1397
	int somaxconn;
1398 1399 1400

	sock = sockfd_lookup_light(fd, &err, &fput_needed);
	if (sock) {
1401
		somaxconn = sock_net(sock->sk)->core.sysctl_somaxconn;
1402
		if ((unsigned int)backlog > somaxconn)
1403
			backlog = somaxconn;
L
Linus Torvalds 已提交
1404 1405

		err = security_socket_listen(sock, backlog);
1406 1407
		if (!err)
			err = sock->ops->listen(sock, backlog);
L
Linus Torvalds 已提交
1408

1409
		fput_light(sock->file, fput_needed);
L
Linus Torvalds 已提交
1410 1411 1412 1413 1414 1415 1416 1417 1418 1419 1420 1421 1422 1423 1424 1425
	}
	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.
 */

1426 1427
SYSCALL_DEFINE4(accept4, int, fd, struct sockaddr __user *, upeer_sockaddr,
		int __user *, upeer_addrlen, int, flags)
L
Linus Torvalds 已提交
1428 1429
{
	struct socket *sock, *newsock;
1430
	struct file *newfile;
1431
	int err, len, newfd, fput_needed;
1432
	struct sockaddr_storage address;
L
Linus Torvalds 已提交
1433

1434
	if (flags & ~(SOCK_CLOEXEC | SOCK_NONBLOCK))
U
Ulrich Drepper 已提交
1435 1436 1437 1438 1439
		return -EINVAL;

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

1440
	sock = sockfd_lookup_light(fd, &err, &fput_needed);
L
Linus Torvalds 已提交
1441 1442 1443 1444
	if (!sock)
		goto out;

	err = -ENFILE;
1445 1446
	newsock = sock_alloc();
	if (!newsock)
L
Linus Torvalds 已提交
1447 1448 1449 1450 1451 1452 1453 1454 1455 1456 1457
		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);

1458
	newfd = get_unused_fd_flags(flags);
1459 1460
	if (unlikely(newfd < 0)) {
		err = newfd;
1461 1462
		sock_release(newsock);
		goto out_put;
1463
	}
1464
	newfile = sock_alloc_file(newsock, flags, sock->sk->sk_prot_creator->name);
1465
	if (IS_ERR(newfile)) {
1466 1467 1468 1469 1470
		err = PTR_ERR(newfile);
		put_unused_fd(newfd);
		sock_release(newsock);
		goto out_put;
	}
1471

1472 1473
	err = security_socket_accept(sock, newsock);
	if (err)
1474
		goto out_fd;
1475

L
Linus Torvalds 已提交
1476 1477
	err = sock->ops->accept(sock, newsock, sock->file->f_flags);
	if (err < 0)
1478
		goto out_fd;
L
Linus Torvalds 已提交
1479 1480

	if (upeer_sockaddr) {
1481
		if (newsock->ops->getname(newsock, (struct sockaddr *)&address,
1482
					  &len, 2) < 0) {
L
Linus Torvalds 已提交
1483
			err = -ECONNABORTED;
1484
			goto out_fd;
L
Linus Torvalds 已提交
1485
		}
1486
		err = move_addr_to_user(&address,
1487
					len, upeer_sockaddr, upeer_addrlen);
L
Linus Torvalds 已提交
1488
		if (err < 0)
1489
			goto out_fd;
L
Linus Torvalds 已提交
1490 1491 1492 1493
	}

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

1494 1495
	fd_install(newfd, newfile);
	err = newfd;
L
Linus Torvalds 已提交
1496 1497

out_put:
1498
	fput_light(sock->file, fput_needed);
L
Linus Torvalds 已提交
1499 1500
out:
	return err;
1501
out_fd:
1502
	fput(newfile);
1503
	put_unused_fd(newfd);
L
Linus Torvalds 已提交
1504 1505 1506
	goto out_put;
}

1507 1508
SYSCALL_DEFINE3(accept, int, fd, struct sockaddr __user *, upeer_sockaddr,
		int __user *, upeer_addrlen)
U
Ulrich Drepper 已提交
1509
{
U
Ulrich Drepper 已提交
1510
	return sys_accept4(fd, upeer_sockaddr, upeer_addrlen, 0);
U
Ulrich Drepper 已提交
1511 1512
}

L
Linus Torvalds 已提交
1513 1514 1515 1516 1517 1518 1519 1520 1521 1522 1523 1524
/*
 *	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.
 */

1525 1526
SYSCALL_DEFINE3(connect, int, fd, struct sockaddr __user *, uservaddr,
		int, addrlen)
L
Linus Torvalds 已提交
1527 1528
{
	struct socket *sock;
1529
	struct sockaddr_storage address;
1530
	int err, fput_needed;
L
Linus Torvalds 已提交
1531

1532
	sock = sockfd_lookup_light(fd, &err, &fput_needed);
L
Linus Torvalds 已提交
1533 1534
	if (!sock)
		goto out;
1535
	err = move_addr_to_kernel(uservaddr, addrlen, &address);
L
Linus Torvalds 已提交
1536 1537 1538
	if (err < 0)
		goto out_put;

1539
	err =
1540
	    security_socket_connect(sock, (struct sockaddr *)&address, addrlen);
L
Linus Torvalds 已提交
1541 1542 1543
	if (err)
		goto out_put;

1544
	err = sock->ops->connect(sock, (struct sockaddr *)&address, addrlen,
L
Linus Torvalds 已提交
1545 1546
				 sock->file->f_flags);
out_put:
1547
	fput_light(sock->file, fput_needed);
L
Linus Torvalds 已提交
1548 1549 1550 1551 1552 1553 1554 1555 1556
out:
	return err;
}

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

1557 1558
SYSCALL_DEFINE3(getsockname, int, fd, struct sockaddr __user *, usockaddr,
		int __user *, usockaddr_len)
L
Linus Torvalds 已提交
1559 1560
{
	struct socket *sock;
1561
	struct sockaddr_storage address;
1562
	int len, err, fput_needed;
1563

1564
	sock = sockfd_lookup_light(fd, &err, &fput_needed);
L
Linus Torvalds 已提交
1565 1566 1567 1568 1569 1570 1571
	if (!sock)
		goto out;

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

1572
	err = sock->ops->getname(sock, (struct sockaddr *)&address, &len, 0);
L
Linus Torvalds 已提交
1573 1574
	if (err)
		goto out_put;
1575
	err = move_addr_to_user(&address, len, usockaddr, usockaddr_len);
L
Linus Torvalds 已提交
1576 1577

out_put:
1578
	fput_light(sock->file, fput_needed);
L
Linus Torvalds 已提交
1579 1580 1581 1582 1583 1584 1585 1586 1587
out:
	return err;
}

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

1588 1589
SYSCALL_DEFINE3(getpeername, int, fd, struct sockaddr __user *, usockaddr,
		int __user *, usockaddr_len)
L
Linus Torvalds 已提交
1590 1591
{
	struct socket *sock;
1592
	struct sockaddr_storage address;
1593
	int len, err, fput_needed;
L
Linus Torvalds 已提交
1594

1595 1596
	sock = sockfd_lookup_light(fd, &err, &fput_needed);
	if (sock != NULL) {
L
Linus Torvalds 已提交
1597 1598
		err = security_socket_getpeername(sock);
		if (err) {
1599
			fput_light(sock->file, fput_needed);
L
Linus Torvalds 已提交
1600 1601 1602
			return err;
		}

1603
		err =
1604
		    sock->ops->getname(sock, (struct sockaddr *)&address, &len,
1605
				       1);
L
Linus Torvalds 已提交
1606
		if (!err)
1607
			err = move_addr_to_user(&address, len, usockaddr,
1608
						usockaddr_len);
1609
		fput_light(sock->file, fput_needed);
L
Linus Torvalds 已提交
1610 1611 1612 1613 1614 1615 1616 1617 1618 1619
	}
	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.
 */

1620
SYSCALL_DEFINE6(sendto, int, fd, void __user *, buff, size_t, len,
1621
		unsigned int, flags, struct sockaddr __user *, addr,
1622
		int, addr_len)
L
Linus Torvalds 已提交
1623 1624
{
	struct socket *sock;
1625
	struct sockaddr_storage address;
L
Linus Torvalds 已提交
1626 1627 1628
	int err;
	struct msghdr msg;
	struct iovec iov;
1629 1630
	int fput_needed;

1631 1632 1633
	err = import_single_range(WRITE, buff, len, &iov, &msg.msg_iter);
	if (unlikely(err))
		return err;
1634 1635
	sock = sockfd_lookup_light(fd, &err, &fput_needed);
	if (!sock)
1636
		goto out;
1637

1638 1639 1640 1641
	msg.msg_name = NULL;
	msg.msg_control = NULL;
	msg.msg_controllen = 0;
	msg.msg_namelen = 0;
1642
	if (addr) {
1643
		err = move_addr_to_kernel(addr, addr_len, &address);
L
Linus Torvalds 已提交
1644 1645
		if (err < 0)
			goto out_put;
1646
		msg.msg_name = (struct sockaddr *)&address;
1647
		msg.msg_namelen = addr_len;
L
Linus Torvalds 已提交
1648 1649 1650 1651
	}
	if (sock->file->f_flags & O_NONBLOCK)
		flags |= MSG_DONTWAIT;
	msg.msg_flags = flags;
1652
	err = sock_sendmsg(sock, &msg);
L
Linus Torvalds 已提交
1653

1654
out_put:
1655
	fput_light(sock->file, fput_needed);
1656
out:
L
Linus Torvalds 已提交
1657 1658 1659 1660
	return err;
}

/*
1661
 *	Send a datagram down a socket.
L
Linus Torvalds 已提交
1662 1663
 */

1664
SYSCALL_DEFINE4(send, int, fd, void __user *, buff, size_t, len,
1665
		unsigned int, flags)
L
Linus Torvalds 已提交
1666 1667 1668 1669 1670
{
	return sys_sendto(fd, buff, len, flags, NULL, 0);
}

/*
1671
 *	Receive a frame from the socket and optionally record the address of the
L
Linus Torvalds 已提交
1672 1673 1674 1675
 *	sender. We verify the buffers are writable and if needed move the
 *	sender address from kernel to user space.
 */

1676
SYSCALL_DEFINE6(recvfrom, int, fd, void __user *, ubuf, size_t, size,
1677
		unsigned int, flags, struct sockaddr __user *, addr,
1678
		int __user *, addr_len)
L
Linus Torvalds 已提交
1679 1680 1681 1682
{
	struct socket *sock;
	struct iovec iov;
	struct msghdr msg;
1683
	struct sockaddr_storage address;
1684
	int err, err2;
1685 1686
	int fput_needed;

1687 1688 1689
	err = import_single_range(READ, ubuf, size, &iov, &msg.msg_iter);
	if (unlikely(err))
		return err;
1690
	sock = sockfd_lookup_light(fd, &err, &fput_needed);
L
Linus Torvalds 已提交
1691
	if (!sock)
1692
		goto out;
L
Linus Torvalds 已提交
1693

1694 1695
	msg.msg_control = NULL;
	msg.msg_controllen = 0;
1696 1697 1698 1699
	/* 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;
1700
	msg.msg_iocb = NULL;
L
Linus Torvalds 已提交
1701 1702
	if (sock->file->f_flags & O_NONBLOCK)
		flags |= MSG_DONTWAIT;
1703
	err = sock_recvmsg(sock, &msg, iov_iter_count(&msg.msg_iter), flags);
L
Linus Torvalds 已提交
1704

1705
	if (err >= 0 && addr != NULL) {
1706
		err2 = move_addr_to_user(&address,
1707
					 msg.msg_namelen, addr, addr_len);
1708 1709
		if (err2 < 0)
			err = err2;
L
Linus Torvalds 已提交
1710
	}
1711 1712

	fput_light(sock->file, fput_needed);
1713
out:
L
Linus Torvalds 已提交
1714 1715 1716 1717
	return err;
}

/*
1718
 *	Receive a datagram from a socket.
L
Linus Torvalds 已提交
1719 1720
 */

1721 1722
SYSCALL_DEFINE4(recv, int, fd, void __user *, ubuf, size_t, size,
		unsigned int, flags)
L
Linus Torvalds 已提交
1723 1724 1725 1726 1727 1728 1729 1730 1731
{
	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.
 */

1732 1733
SYSCALL_DEFINE5(setsockopt, int, fd, int, level, int, optname,
		char __user *, optval, int, optlen)
L
Linus Torvalds 已提交
1734
{
1735
	int err, fput_needed;
L
Linus Torvalds 已提交
1736 1737 1738 1739
	struct socket *sock;

	if (optlen < 0)
		return -EINVAL;
1740 1741 1742 1743

	sock = sockfd_lookup_light(fd, &err, &fput_needed);
	if (sock != NULL) {
		err = security_socket_setsockopt(sock, level, optname);
1744 1745
		if (err)
			goto out_put;
L
Linus Torvalds 已提交
1746 1747

		if (level == SOL_SOCKET)
1748 1749 1750
			err =
			    sock_setsockopt(sock, level, optname, optval,
					    optlen);
L
Linus Torvalds 已提交
1751
		else
1752 1753 1754
			err =
			    sock->ops->setsockopt(sock, level, optname, optval,
						  optlen);
1755 1756
out_put:
		fput_light(sock->file, fput_needed);
L
Linus Torvalds 已提交
1757 1758 1759 1760 1761 1762 1763 1764 1765
	}
	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.
 */

1766 1767
SYSCALL_DEFINE5(getsockopt, int, fd, int, level, int, optname,
		char __user *, optval, int __user *, optlen)
L
Linus Torvalds 已提交
1768
{
1769
	int err, fput_needed;
L
Linus Torvalds 已提交
1770 1771
	struct socket *sock;

1772 1773
	sock = sockfd_lookup_light(fd, &err, &fput_needed);
	if (sock != NULL) {
1774 1775 1776
		err = security_socket_getsockopt(sock, level, optname);
		if (err)
			goto out_put;
L
Linus Torvalds 已提交
1777 1778

		if (level == SOL_SOCKET)
1779 1780 1781
			err =
			    sock_getsockopt(sock, level, optname, optval,
					    optlen);
L
Linus Torvalds 已提交
1782
		else
1783 1784 1785
			err =
			    sock->ops->getsockopt(sock, level, optname, optval,
						  optlen);
1786 1787
out_put:
		fput_light(sock->file, fput_needed);
L
Linus Torvalds 已提交
1788 1789 1790 1791 1792 1793 1794 1795
	}
	return err;
}

/*
 *	Shutdown a socket.
 */

1796
SYSCALL_DEFINE2(shutdown, int, fd, int, how)
L
Linus Torvalds 已提交
1797
{
1798
	int err, fput_needed;
L
Linus Torvalds 已提交
1799 1800
	struct socket *sock;

1801 1802
	sock = sockfd_lookup_light(fd, &err, &fput_needed);
	if (sock != NULL) {
L
Linus Torvalds 已提交
1803
		err = security_socket_shutdown(sock, how);
1804 1805 1806
		if (!err)
			err = sock->ops->shutdown(sock, how);
		fput_light(sock->file, fput_needed);
L
Linus Torvalds 已提交
1807 1808 1809 1810
	}
	return err;
}

1811
/* A couple of helpful macros for getting the address of the 32/64 bit
L
Linus Torvalds 已提交
1812 1813 1814 1815 1816 1817
 * 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)

1818 1819 1820 1821 1822
struct used_address {
	struct sockaddr_storage name;
	unsigned int name_len;
};

1823 1824 1825 1826
static int copy_msghdr_from_user(struct msghdr *kmsg,
				 struct user_msghdr __user *umsg,
				 struct sockaddr __user **save_addr,
				 struct iovec **iov)
1827
{
1828 1829
	struct sockaddr __user *uaddr;
	struct iovec __user *uiov;
A
Al Viro 已提交
1830
	size_t nr_segs;
1831 1832 1833 1834 1835 1836
	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
1837
	    __get_user(nr_segs, &umsg->msg_iovlen) ||
1838 1839 1840
	    __get_user(kmsg->msg_control, &umsg->msg_control) ||
	    __get_user(kmsg->msg_controllen, &umsg->msg_controllen) ||
	    __get_user(kmsg->msg_flags, &umsg->msg_flags))
1841
		return -EFAULT;
1842

1843
	if (!uaddr)
1844 1845
		kmsg->msg_namelen = 0;

1846 1847 1848
	if (kmsg->msg_namelen < 0)
		return -EINVAL;

1849
	if (kmsg->msg_namelen > sizeof(struct sockaddr_storage))
1850
		kmsg->msg_namelen = sizeof(struct sockaddr_storage);
1851 1852 1853 1854 1855 1856 1857 1858 1859 1860 1861 1862 1863 1864 1865 1866

	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
1867
	if (nr_segs > UIO_MAXIOV)
1868 1869
		return -EMSGSIZE;

1870 1871
	kmsg->msg_iocb = NULL;

1872 1873
	return import_iovec(save_addr ? READ : WRITE, uiov, nr_segs,
			    UIO_FASTIOV, iov, &kmsg->msg_iter);
1874 1875
}

1876
static int ___sys_sendmsg(struct socket *sock, struct user_msghdr __user *msg,
1877
			 struct msghdr *msg_sys, unsigned int flags,
1878
			 struct used_address *used_address)
L
Linus Torvalds 已提交
1879
{
1880 1881
	struct compat_msghdr __user *msg_compat =
	    (struct compat_msghdr __user *)msg;
1882
	struct sockaddr_storage address;
L
Linus Torvalds 已提交
1883
	struct iovec iovstack[UIO_FASTIOV], *iov = iovstack;
1884
	unsigned char ctl[sizeof(struct cmsghdr) + 20]
1885 1886
	    __attribute__ ((aligned(sizeof(__kernel_size_t))));
	/* 20 is size of ipv6_pktinfo */
L
Linus Torvalds 已提交
1887
	unsigned char *ctl_buf = ctl;
1888
	int ctl_len;
1889
	ssize_t err;
1890

1891
	msg_sys->msg_name = &address;
L
Linus Torvalds 已提交
1892

1893
	if (MSG_CMSG_COMPAT & flags)
1894
		err = get_compat_msghdr(msg_sys, msg_compat, NULL, &iov);
1895
	else
1896
		err = copy_msghdr_from_user(msg_sys, msg, NULL, &iov);
1897
	if (err < 0)
1898
		return err;
L
Linus Torvalds 已提交
1899 1900 1901

	err = -ENOBUFS;

1902
	if (msg_sys->msg_controllen > INT_MAX)
L
Linus Torvalds 已提交
1903
		goto out_freeiov;
1904
	ctl_len = msg_sys->msg_controllen;
L
Linus Torvalds 已提交
1905
	if ((MSG_CMSG_COMPAT & flags) && ctl_len) {
1906
		err =
1907
		    cmsghdr_from_user_compat_to_kern(msg_sys, sock->sk, ctl,
1908
						     sizeof(ctl));
L
Linus Torvalds 已提交
1909 1910
		if (err)
			goto out_freeiov;
1911 1912
		ctl_buf = msg_sys->msg_control;
		ctl_len = msg_sys->msg_controllen;
L
Linus Torvalds 已提交
1913
	} else if (ctl_len) {
1914
		if (ctl_len > sizeof(ctl)) {
L
Linus Torvalds 已提交
1915
			ctl_buf = sock_kmalloc(sock->sk, ctl_len, GFP_KERNEL);
1916
			if (ctl_buf == NULL)
L
Linus Torvalds 已提交
1917 1918 1919 1920
				goto out_freeiov;
		}
		err = -EFAULT;
		/*
1921
		 * Careful! Before this, msg_sys->msg_control contains a user pointer.
L
Linus Torvalds 已提交
1922 1923 1924
		 * Afterwards, it will be a kernel pointer. Thus the compiler-assisted
		 * checking falls down on this.
		 */
1925
		if (copy_from_user(ctl_buf,
1926
				   (void __user __force *)msg_sys->msg_control,
1927
				   ctl_len))
L
Linus Torvalds 已提交
1928
			goto out_freectl;
1929
		msg_sys->msg_control = ctl_buf;
L
Linus Torvalds 已提交
1930
	}
1931
	msg_sys->msg_flags = flags;
L
Linus Torvalds 已提交
1932 1933

	if (sock->file->f_flags & O_NONBLOCK)
1934
		msg_sys->msg_flags |= MSG_DONTWAIT;
1935 1936 1937 1938 1939 1940
	/*
	 * 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.
	 */
1941 1942 1943
	if (used_address && msg_sys->msg_name &&
	    used_address->name_len == msg_sys->msg_namelen &&
	    !memcmp(&used_address->name, msg_sys->msg_name,
1944
		    used_address->name_len)) {
1945
		err = sock_sendmsg_nosec(sock, msg_sys);
1946 1947
		goto out_freectl;
	}
1948
	err = sock_sendmsg(sock, msg_sys);
1949 1950 1951 1952 1953 1954
	/*
	 * 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;
1955 1956 1957
		if (msg_sys->msg_name)
			memcpy(&used_address->name, msg_sys->msg_name,
			       used_address->name_len);
1958
	}
L
Linus Torvalds 已提交
1959 1960

out_freectl:
1961
	if (ctl_buf != ctl)
L
Linus Torvalds 已提交
1962 1963
		sock_kfree_s(sock->sk, ctl_buf, ctl_len);
out_freeiov:
1964
	kfree(iov);
1965 1966 1967 1968 1969 1970 1971
	return err;
}

/*
 *	BSD sendmsg interface
 */

1972
long __sys_sendmsg(int fd, struct user_msghdr __user *msg, unsigned flags)
1973 1974 1975
{
	int fput_needed, err;
	struct msghdr msg_sys;
1976 1977 1978
	struct socket *sock;

	sock = sockfd_lookup_light(fd, &err, &fput_needed);
1979 1980 1981
	if (!sock)
		goto out;

1982
	err = ___sys_sendmsg(sock, msg, &msg_sys, flags, NULL);
1983

1984
	fput_light(sock->file, fput_needed);
1985
out:
L
Linus Torvalds 已提交
1986 1987 1988
	return err;
}

1989
SYSCALL_DEFINE3(sendmsg, int, fd, struct user_msghdr __user *, msg, unsigned int, flags)
1990 1991 1992 1993 1994 1995
{
	if (flags & MSG_CMSG_COMPAT)
		return -EINVAL;
	return __sys_sendmsg(fd, msg, flags);
}

1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007
/*
 *	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;
2008
	struct used_address used_address;
2009

2010 2011
	if (vlen > UIO_MAXIOV)
		vlen = UIO_MAXIOV;
2012 2013 2014 2015 2016 2017 2018

	datagrams = 0;

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

2019
	used_address.name_len = UINT_MAX;
2020 2021
	entry = mmsg;
	compat_entry = (struct compat_mmsghdr __user *)mmsg;
2022
	err = 0;
2023 2024 2025

	while (datagrams < vlen) {
		if (MSG_CMSG_COMPAT & flags) {
2026
			err = ___sys_sendmsg(sock, (struct user_msghdr __user *)compat_entry,
2027
					     &msg_sys, flags, &used_address);
2028 2029 2030 2031 2032
			if (err < 0)
				break;
			err = __put_user(err, &compat_entry->msg_len);
			++compat_entry;
		} else {
2033
			err = ___sys_sendmsg(sock,
2034
					     (struct user_msghdr __user *)entry,
2035
					     &msg_sys, flags, &used_address);
2036 2037 2038 2039 2040 2041 2042 2043 2044
			if (err < 0)
				break;
			err = put_user(err, &entry->msg_len);
			++entry;
		}

		if (err)
			break;
		++datagrams;
2045
		cond_resched();
2046 2047 2048 2049
	}

	fput_light(sock->file, fput_needed);

2050 2051
	/* We only return an error if no datagrams were able to be sent */
	if (datagrams != 0)
2052 2053 2054 2055 2056 2057 2058 2059
		return datagrams;

	return err;
}

SYSCALL_DEFINE4(sendmmsg, int, fd, struct mmsghdr __user *, mmsg,
		unsigned int, vlen, unsigned int, flags)
{
2060 2061
	if (flags & MSG_CMSG_COMPAT)
		return -EINVAL;
2062 2063 2064
	return __sys_sendmmsg(fd, mmsg, vlen, flags);
}

2065
static int ___sys_recvmsg(struct socket *sock, struct user_msghdr __user *msg,
2066
			 struct msghdr *msg_sys, unsigned int flags, int nosec)
L
Linus Torvalds 已提交
2067
{
2068 2069
	struct compat_msghdr __user *msg_compat =
	    (struct compat_msghdr __user *)msg;
L
Linus Torvalds 已提交
2070
	struct iovec iovstack[UIO_FASTIOV];
2071
	struct iovec *iov = iovstack;
L
Linus Torvalds 已提交
2072
	unsigned long cmsg_ptr;
2073 2074
	int total_len, len;
	ssize_t err;
L
Linus Torvalds 已提交
2075 2076

	/* kernel mode address */
2077
	struct sockaddr_storage addr;
L
Linus Torvalds 已提交
2078 2079 2080

	/* user mode address pointers */
	struct sockaddr __user *uaddr;
2081
	int __user *uaddr_len = COMPAT_NAMELEN(msg);
2082

2083
	msg_sys->msg_name = &addr;
L
Linus Torvalds 已提交
2084

2085
	if (MSG_CMSG_COMPAT & flags)
2086
		err = get_compat_msghdr(msg_sys, msg_compat, &uaddr, &iov);
2087
	else
2088
		err = copy_msghdr_from_user(msg_sys, msg, &uaddr, &iov);
L
Linus Torvalds 已提交
2089
	if (err < 0)
2090 2091
		return err;
	total_len = iov_iter_count(&msg_sys->msg_iter);
L
Linus Torvalds 已提交
2092

2093 2094
	cmsg_ptr = (unsigned long)msg_sys->msg_control;
	msg_sys->msg_flags = flags & (MSG_CMSG_CLOEXEC|MSG_CMSG_COMPAT);
2095

2096 2097 2098
	/* We assume all kernel code knows the size of sockaddr_storage */
	msg_sys->msg_namelen = 0;

L
Linus Torvalds 已提交
2099 2100
	if (sock->file->f_flags & O_NONBLOCK)
		flags |= MSG_DONTWAIT;
2101 2102
	err = (nosec ? sock_recvmsg_nosec : sock_recvmsg)(sock, msg_sys,
							  total_len, flags);
L
Linus Torvalds 已提交
2103 2104 2105 2106 2107
	if (err < 0)
		goto out_freeiov;
	len = err;

	if (uaddr != NULL) {
2108
		err = move_addr_to_user(&addr,
2109
					msg_sys->msg_namelen, uaddr,
2110
					uaddr_len);
L
Linus Torvalds 已提交
2111 2112 2113
		if (err < 0)
			goto out_freeiov;
	}
2114
	err = __put_user((msg_sys->msg_flags & ~MSG_CMSG_COMPAT),
2115
			 COMPAT_FLAGS(msg));
L
Linus Torvalds 已提交
2116 2117 2118
	if (err)
		goto out_freeiov;
	if (MSG_CMSG_COMPAT & flags)
2119
		err = __put_user((unsigned long)msg_sys->msg_control - cmsg_ptr,
L
Linus Torvalds 已提交
2120 2121
				 &msg_compat->msg_controllen);
	else
2122
		err = __put_user((unsigned long)msg_sys->msg_control - cmsg_ptr,
L
Linus Torvalds 已提交
2123 2124 2125 2126 2127 2128
				 &msg->msg_controllen);
	if (err)
		goto out_freeiov;
	err = len;

out_freeiov:
2129
	kfree(iov);
2130 2131 2132 2133 2134 2135 2136
	return err;
}

/*
 *	BSD recvmsg interface
 */

2137
long __sys_recvmsg(int fd, struct user_msghdr __user *msg, unsigned flags)
2138 2139 2140
{
	int fput_needed, err;
	struct msghdr msg_sys;
2141 2142 2143
	struct socket *sock;

	sock = sockfd_lookup_light(fd, &err, &fput_needed);
2144 2145 2146
	if (!sock)
		goto out;

2147
	err = ___sys_recvmsg(sock, msg, &msg_sys, flags, 0);
2148

2149
	fput_light(sock->file, fput_needed);
L
Linus Torvalds 已提交
2150 2151 2152 2153
out:
	return err;
}

2154
SYSCALL_DEFINE3(recvmsg, int, fd, struct user_msghdr __user *, msg,
2155 2156 2157 2158 2159 2160 2161
		unsigned int, flags)
{
	if (flags & MSG_CMSG_COMPAT)
		return -EINVAL;
	return __sys_recvmsg(fd, msg, flags);
}

2162 2163 2164 2165 2166 2167 2168 2169 2170 2171
/*
 *     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;
2172
	struct compat_mmsghdr __user *compat_entry;
2173 2174 2175 2176 2177 2178 2179 2180 2181 2182 2183 2184 2185 2186 2187 2188 2189 2190 2191
	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;
2192
	compat_entry = (struct compat_mmsghdr __user *)mmsg;
2193 2194 2195 2196 2197

	while (datagrams < vlen) {
		/*
		 * No need to ask LSM for more than the first datagram.
		 */
2198
		if (MSG_CMSG_COMPAT & flags) {
2199
			err = ___sys_recvmsg(sock, (struct user_msghdr __user *)compat_entry,
2200 2201
					     &msg_sys, flags & ~MSG_WAITFORONE,
					     datagrams);
2202 2203 2204 2205 2206
			if (err < 0)
				break;
			err = __put_user(err, &compat_entry->msg_len);
			++compat_entry;
		} else {
2207
			err = ___sys_recvmsg(sock,
2208
					     (struct user_msghdr __user *)entry,
2209 2210
					     &msg_sys, flags & ~MSG_WAITFORONE,
					     datagrams);
2211 2212 2213 2214 2215 2216
			if (err < 0)
				break;
			err = put_user(err, &entry->msg_len);
			++entry;
		}

2217 2218 2219 2220
		if (err)
			break;
		++datagrams;

2221 2222 2223 2224
		/* MSG_WAITFORONE turns on MSG_DONTWAIT after one packet */
		if (flags & MSG_WAITFORONE)
			flags |= MSG_DONTWAIT;

2225 2226 2227 2228 2229 2230 2231 2232 2233 2234 2235 2236 2237 2238 2239 2240
		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;
2241
		cond_resched();
2242 2243 2244 2245
	}

out_put:
	fput_light(sock->file, fput_needed);
L
Linus Torvalds 已提交
2246

2247 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
	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;

2278 2279 2280
	if (flags & MSG_CMSG_COMPAT)
		return -EINVAL;

2281 2282 2283 2284 2285 2286 2287 2288 2289 2290 2291 2292 2293 2294 2295 2296
	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
2297 2298
/* Argument list sizes for sys_socketcall */
#define AL(x) ((x) * sizeof(unsigned long))
2299
static const unsigned char nargs[21] = {
2300 2301 2302
	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),
2303
	AL(4), AL(5), AL(4)
2304 2305
};

L
Linus Torvalds 已提交
2306 2307 2308
#undef AL

/*
2309
 *	System call vectors.
L
Linus Torvalds 已提交
2310 2311 2312
 *
 *	Argument checking cleaned up. Saved 20% in size.
 *  This function doesn't need to set the kernel lock because
2313
 *  it is set by the callees.
L
Linus Torvalds 已提交
2314 2315
 */

2316
SYSCALL_DEFINE2(socketcall, int, call, unsigned long __user *, args)
L
Linus Torvalds 已提交
2317
{
2318
	unsigned long a[AUDITSC_ARGS];
2319
	unsigned long a0, a1;
L
Linus Torvalds 已提交
2320
	int err;
2321
	unsigned int len;
L
Linus Torvalds 已提交
2322

2323
	if (call < 1 || call > SYS_SENDMMSG)
L
Linus Torvalds 已提交
2324 2325
		return -EINVAL;

2326 2327 2328 2329
	len = nargs[call];
	if (len > sizeof(a))
		return -EINVAL;

L
Linus Torvalds 已提交
2330
	/* copy_from_user should be SMP safe. */
2331
	if (copy_from_user(a, args, len))
L
Linus Torvalds 已提交
2332
		return -EFAULT;
2333

2334 2335 2336
	err = audit_socketcall(nargs[call] / sizeof(unsigned long), a);
	if (err)
		return err;
2337

2338 2339 2340 2341 2342 2343 2344 2345 2346 2347 2348 2349 2350 2351 2352 2353 2354
	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
2355 2356
		err = sys_accept4(a0, (struct sockaddr __user *)a1,
				  (int __user *)a[2], 0);
2357 2358 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
		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:
2398
		err = sys_sendmsg(a0, (struct user_msghdr __user *)a1, a[2]);
2399
		break;
2400 2401 2402
	case SYS_SENDMMSG:
		err = sys_sendmmsg(a0, (struct mmsghdr __user *)a1, a[2], a[3]);
		break;
2403
	case SYS_RECVMSG:
2404
		err = sys_recvmsg(a0, (struct user_msghdr __user *)a1, a[2]);
2405
		break;
2406 2407 2408 2409
	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
2410 2411 2412
	case SYS_ACCEPT4:
		err = sys_accept4(a0, (struct sockaddr __user *)a1,
				  (int __user *)a[2], a[3]);
U
Ulrich Drepper 已提交
2413
		break;
2414 2415 2416
	default:
		err = -EINVAL;
		break;
L
Linus Torvalds 已提交
2417 2418 2419 2420
	}
	return err;
}

2421
#endif				/* __ARCH_WANT_SYS_SOCKETCALL */
L
Linus Torvalds 已提交
2422

2423 2424 2425 2426
/**
 *	sock_register - add a socket protocol handler
 *	@ops: description of protocol
 *
L
Linus Torvalds 已提交
2427 2428
 *	This function is called by a protocol handler that wants to
 *	advertise its address family, and have it linked into the
2429
 *	socket interface. The value ops->family corresponds to the
2430
 *	socket system call protocol family.
L
Linus Torvalds 已提交
2431
 */
2432
int sock_register(const struct net_proto_family *ops)
L
Linus Torvalds 已提交
2433 2434 2435 2436
{
	int err;

	if (ops->family >= NPROTO) {
2437
		pr_crit("protocol %d >= NPROTO(%d)\n", ops->family, NPROTO);
L
Linus Torvalds 已提交
2438 2439
		return -ENOBUFS;
	}
2440 2441

	spin_lock(&net_family_lock);
E
Eric Dumazet 已提交
2442 2443
	if (rcu_dereference_protected(net_families[ops->family],
				      lockdep_is_held(&net_family_lock)))
2444 2445
		err = -EEXIST;
	else {
2446
		rcu_assign_pointer(net_families[ops->family], ops);
L
Linus Torvalds 已提交
2447 2448
		err = 0;
	}
2449 2450
	spin_unlock(&net_family_lock);

2451
	pr_info("NET: Registered protocol family %d\n", ops->family);
L
Linus Torvalds 已提交
2452 2453
	return err;
}
2454
EXPORT_SYMBOL(sock_register);
L
Linus Torvalds 已提交
2455

2456 2457 2458 2459
/**
 *	sock_unregister - remove a protocol handler
 *	@family: protocol family to remove
 *
L
Linus Torvalds 已提交
2460 2461
 *	This function is called by a protocol handler that wants to
 *	remove its address family, and have it unlinked from the
2462 2463 2464 2465 2466 2467
 *	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
2468
 */
2469
void sock_unregister(int family)
L
Linus Torvalds 已提交
2470
{
2471
	BUG_ON(family < 0 || family >= NPROTO);
L
Linus Torvalds 已提交
2472

2473
	spin_lock(&net_family_lock);
2474
	RCU_INIT_POINTER(net_families[family], NULL);
2475 2476 2477 2478
	spin_unlock(&net_family_lock);

	synchronize_rcu();

2479
	pr_info("NET: Unregistered protocol family %d\n", family);
L
Linus Torvalds 已提交
2480
}
2481
EXPORT_SYMBOL(sock_unregister);
L
Linus Torvalds 已提交
2482

2483
static int __init sock_init(void)
L
Linus Torvalds 已提交
2484
{
N
Nick Piggin 已提交
2485
	int err;
2486 2487 2488 2489 2490 2491
	/*
	 *      Initialize the network sysctl infrastructure.
	 */
	err = net_sysctl_init();
	if (err)
		goto out;
N
Nick Piggin 已提交
2492

L
Linus Torvalds 已提交
2493
	/*
2494
	 *      Initialize skbuff SLAB cache
L
Linus Torvalds 已提交
2495 2496 2497 2498
	 */
	skb_init();

	/*
2499
	 *      Initialize the protocols module.
L
Linus Torvalds 已提交
2500 2501 2502
	 */

	init_inodecache();
N
Nick Piggin 已提交
2503 2504 2505 2506

	err = register_filesystem(&sock_fs_type);
	if (err)
		goto out_fs;
L
Linus Torvalds 已提交
2507
	sock_mnt = kern_mount(&sock_fs_type);
N
Nick Piggin 已提交
2508 2509 2510 2511
	if (IS_ERR(sock_mnt)) {
		err = PTR_ERR(sock_mnt);
		goto out_mount;
	}
2512 2513

	/* The real protocol initialization is performed in later initcalls.
L
Linus Torvalds 已提交
2514 2515 2516
	 */

#ifdef CONFIG_NETFILTER
2517 2518 2519
	err = netfilter_init();
	if (err)
		goto out;
L
Linus Torvalds 已提交
2520
#endif
2521

2522
	ptp_classifier_init();
2523

N
Nick Piggin 已提交
2524 2525 2526 2527 2528 2529 2530
out:
	return err;

out_mount:
	unregister_filesystem(&sock_fs_type);
out_fs:
	goto out;
L
Linus Torvalds 已提交
2531 2532
}

2533 2534
core_initcall(sock_init);	/* early initcall */

L
Linus Torvalds 已提交
2535 2536 2537 2538 2539 2540
#ifdef CONFIG_PROC_FS
void socket_seq_show(struct seq_file *seq)
{
	int cpu;
	int counter = 0;

2541
	for_each_possible_cpu(cpu)
2542
	    counter += per_cpu(sockets_in_use, cpu);
L
Linus Torvalds 已提交
2543 2544 2545 2546 2547 2548 2549

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

	seq_printf(seq, "sockets: used %d\n", counter);
}
2550
#endif				/* CONFIG_PROC_FS */
L
Linus Torvalds 已提交
2551

2552
#ifdef CONFIG_COMPAT
2553
static int do_siocgstamp(struct net *net, struct socket *sock,
2554
			 unsigned int cmd, void __user *up)
2555 2556 2557 2558 2559 2560
{
	mm_segment_t old_fs = get_fs();
	struct timeval ktv;
	int err;

	set_fs(KERNEL_DS);
2561
	err = sock_do_ioctl(net, sock, cmd, (unsigned long)&ktv);
2562
	set_fs(old_fs);
2563
	if (!err)
2564
		err = compat_put_timeval(&ktv, up);
2565

2566 2567 2568
	return err;
}

2569
static int do_siocgstampns(struct net *net, struct socket *sock,
2570
			   unsigned int cmd, void __user *up)
2571 2572 2573 2574 2575 2576
{
	mm_segment_t old_fs = get_fs();
	struct timespec kts;
	int err;

	set_fs(KERNEL_DS);
2577
	err = sock_do_ioctl(net, sock, cmd, (unsigned long)&kts);
2578
	set_fs(old_fs);
2579
	if (!err)
2580
		err = compat_put_timespec(&kts, up);
2581

2582 2583 2584
	return err;
}

2585
static int dev_ifname32(struct net *net, struct compat_ifreq __user *uifr32)
2586 2587 2588 2589 2590
{
	struct ifreq __user *uifr;
	int err;

	uifr = compat_alloc_user_space(sizeof(struct ifreq));
2591
	if (copy_in_user(uifr, uifr32, sizeof(struct compat_ifreq)))
2592 2593
		return -EFAULT;

2594
	err = dev_ioctl(net, SIOCGIFNAME, uifr);
2595 2596 2597
	if (err)
		return err;

2598
	if (copy_in_user(uifr32, uifr, sizeof(struct compat_ifreq)))
2599 2600 2601 2602 2603
		return -EFAULT;

	return 0;
}

2604
static int dev_ifconf(struct net *net, struct compat_ifconf __user *uifc32)
2605
{
2606
	struct compat_ifconf ifc32;
2607 2608
	struct ifconf ifc;
	struct ifconf __user *uifc;
2609
	struct compat_ifreq __user *ifr32;
2610 2611 2612 2613
	struct ifreq __user *ifr;
	unsigned int i, j;
	int err;

2614
	if (copy_from_user(&ifc32, uifc32, sizeof(struct compat_ifconf)))
2615 2616
		return -EFAULT;

2617
	memset(&ifc, 0, sizeof(ifc));
2618 2619 2620 2621 2622 2623
	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 {
2624 2625
		size_t len = ((ifc32.ifc_len / sizeof(struct compat_ifreq)) + 1) *
			sizeof(struct ifreq);
2626 2627 2628 2629
		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);
2630
		for (i = 0; i < ifc32.ifc_len; i += sizeof(struct compat_ifreq)) {
2631
			if (copy_in_user(ifr, ifr32, sizeof(struct compat_ifreq)))
2632 2633 2634 2635 2636 2637 2638 2639
				return -EFAULT;
			ifr++;
			ifr32++;
		}
	}
	if (copy_to_user(uifc, &ifc, sizeof(struct ifconf)))
		return -EFAULT;

2640
	err = dev_ioctl(net, SIOCGIFCONF, uifc);
2641 2642 2643 2644 2645 2646 2647 2648 2649
	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;
2650 2651 2652
	     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)))
2653 2654 2655 2656 2657 2658 2659 2660 2661 2662
			return -EFAULT;
		ifr32++;
		ifr++;
	}

	if (ifc32.ifcbuf == 0) {
		/* Translate from 64-bit structure multiple to
		 * a 32-bit one.
		 */
		i = ifc.ifc_len;
2663
		i = ((i / sizeof(struct ifreq)) * sizeof(struct compat_ifreq));
2664 2665 2666 2667
		ifc32.ifc_len = i;
	} else {
		ifc32.ifc_len = i;
	}
2668
	if (copy_to_user(uifc32, &ifc32, sizeof(struct compat_ifconf)))
2669 2670 2671 2672 2673
		return -EFAULT;

	return 0;
}

2674
static int ethtool_ioctl(struct net *net, struct compat_ifreq __user *ifr32)
2675
{
2676 2677 2678 2679
	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;
2680
	struct ifreq __user *ifr;
2681 2682
	u32 rule_cnt = 0, actual_rule_cnt;
	u32 ethcmd;
2683
	u32 data;
2684
	int ret;
2685

2686 2687
	if (get_user(data, &ifr32->ifr_ifru.ifru_data))
		return -EFAULT;
2688

2689 2690 2691
	compat_rxnfc = compat_ptr(data);

	if (get_user(ethcmd, &compat_rxnfc->cmd))
2692 2693
		return -EFAULT;

2694 2695 2696 2697 2698 2699 2700 2701 2702 2703 2704 2705 2706 2707 2708 2709 2710
	/* 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:
2711
	case ETHTOOL_SRXCLSRLINS:
2712 2713 2714 2715 2716 2717 2718 2719 2720
		convert_out = true;
		/* fall through */
	case ETHTOOL_SRXCLSRLDEL:
		buf_size += sizeof(struct ethtool_rxnfc);
		convert_in = true;
		break;
	}

	ifr = compat_alloc_user_space(buf_size);
S
Stephen Hemminger 已提交
2721
	rxnfc = (void __user *)ifr + ALIGN(sizeof(struct ifreq), 8);
2722 2723

	if (copy_in_user(&ifr->ifr_name, &ifr32->ifr_name, IFNAMSIZ))
2724 2725
		return -EFAULT;

2726 2727
	if (put_user(convert_in ? rxnfc : compat_ptr(data),
		     &ifr->ifr_ifru.ifru_data))
2728 2729
		return -EFAULT;

2730
	if (convert_in) {
2731
		/* We expect there to be holes between fs.m_ext and
2732 2733
		 * fs.ring_cookie and at the end of fs, but nowhere else.
		 */
2734 2735 2736 2737
		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));
2738 2739 2740 2741 2742 2743 2744
		BUILD_BUG_ON(
			offsetof(struct compat_ethtool_rxnfc, fs.location) -
			offsetof(struct compat_ethtool_rxnfc, fs.ring_cookie) !=
			offsetof(struct ethtool_rxnfc, fs.location) -
			offsetof(struct ethtool_rxnfc, fs.ring_cookie));

		if (copy_in_user(rxnfc, compat_rxnfc,
S
Stephen Hemminger 已提交
2745 2746
				 (void __user *)(&rxnfc->fs.m_ext + 1) -
				 (void __user *)rxnfc) ||
2747 2748
		    copy_in_user(&rxnfc->fs.ring_cookie,
				 &compat_rxnfc->fs.ring_cookie,
S
Stephen Hemminger 已提交
2749 2750
				 (void __user *)(&rxnfc->fs.location + 1) -
				 (void __user *)&rxnfc->fs.ring_cookie) ||
2751 2752 2753 2754 2755 2756 2757 2758 2759 2760 2761
		    copy_in_user(&rxnfc->rule_cnt, &compat_rxnfc->rule_cnt,
				 sizeof(rxnfc->rule_cnt)))
			return -EFAULT;
	}

	ret = dev_ioctl(net, SIOCETHTOOL, ifr);
	if (ret)
		return ret;

	if (convert_out) {
		if (copy_in_user(compat_rxnfc, rxnfc,
S
Stephen Hemminger 已提交
2762 2763
				 (const void __user *)(&rxnfc->fs.m_ext + 1) -
				 (const void __user *)rxnfc) ||
2764 2765
		    copy_in_user(&compat_rxnfc->fs.ring_cookie,
				 &rxnfc->fs.ring_cookie,
S
Stephen Hemminger 已提交
2766 2767
				 (const void __user *)(&rxnfc->fs.location + 1) -
				 (const void __user *)&rxnfc->fs.ring_cookie) ||
2768 2769 2770 2771 2772 2773 2774 2775 2776 2777 2778 2779 2780 2781 2782 2783 2784 2785 2786 2787 2788 2789 2790 2791 2792
		    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;
2793 2794
}

2795 2796 2797 2798 2799 2800
static int compat_siocwandev(struct net *net, struct compat_ifreq __user *uifr32)
{
	void __user *uptr;
	compat_uptr_t uptr32;
	struct ifreq __user *uifr;

2801
	uifr = compat_alloc_user_space(sizeof(*uifr));
2802 2803 2804 2805 2806 2807 2808 2809 2810 2811 2812 2813 2814 2815
	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);
}

2816 2817
static int bond_ioctl(struct net *net, unsigned int cmd,
			 struct compat_ifreq __user *ifr32)
2818 2819 2820 2821 2822 2823 2824 2825 2826 2827
{
	struct ifreq kifr;
	mm_segment_t old_fs;
	int err;

	switch (cmd) {
	case SIOCBONDENSLAVE:
	case SIOCBONDRELEASE:
	case SIOCBONDSETHWADDR:
	case SIOCBONDCHANGEACTIVE:
2828
		if (copy_from_user(&kifr, ifr32, sizeof(struct compat_ifreq)))
2829 2830 2831
			return -EFAULT;

		old_fs = get_fs();
2832
		set_fs(KERNEL_DS);
2833 2834
		err = dev_ioctl(net, cmd,
				(struct ifreq __user __force *) &kifr);
2835
		set_fs(old_fs);
2836 2837 2838

		return err;
	default:
2839
		return -ENOIOCTLCMD;
2840
	}
2841 2842
}

2843 2844
/* 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,
2845
				 struct compat_ifreq __user *u_ifreq32)
2846 2847 2848 2849 2850 2851 2852 2853 2854
{
	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;
2855
	if (get_user(data32, &u_ifreq32->ifr_ifru.ifru_data))
2856 2857 2858 2859 2860 2861 2862 2863
		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;
2864
	if (put_user(data64, &u_ifreq64->ifr_ifru.ifru_data))
2865 2866
		return -EFAULT;

2867
	return dev_ioctl(net, cmd, u_ifreq64);
2868 2869
}

2870 2871
static int dev_ifsioc(struct net *net, struct socket *sock,
			 unsigned int cmd, struct compat_ifreq __user *uifr32)
2872
{
2873
	struct ifreq __user *uifr;
2874 2875
	int err;

2876 2877 2878 2879 2880 2881
	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);

2882 2883 2884 2885 2886 2887 2888 2889 2890 2891 2892 2893
	if (!err) {
		switch (cmd) {
		case SIOCGIFFLAGS:
		case SIOCGIFMETRIC:
		case SIOCGIFMTU:
		case SIOCGIFMEM:
		case SIOCGIFHWADDR:
		case SIOCGIFINDEX:
		case SIOCGIFADDR:
		case SIOCGIFBRDADDR:
		case SIOCGIFDSTADDR:
		case SIOCGIFNETMASK:
2894
		case SIOCGIFPFLAGS:
2895
		case SIOCGIFTXQLEN:
2896 2897
		case SIOCGMIIPHY:
		case SIOCGMIIREG:
2898
			if (copy_in_user(uifr32, uifr, sizeof(*uifr32)))
2899 2900 2901 2902 2903 2904 2905
				err = -EFAULT;
			break;
		}
	}
	return err;
}

2906 2907 2908 2909 2910 2911 2912 2913 2914 2915
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));
2916 2917 2918 2919 2920 2921
	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);
2922 2923 2924 2925
	if (err)
		return -EFAULT;

	old_fs = get_fs();
2926
	set_fs(KERNEL_DS);
2927
	err = dev_ioctl(net, cmd, (void  __user __force *)&ifr);
2928
	set_fs(old_fs);
2929 2930 2931

	if (cmd == SIOCGIFMAP && !err) {
		err = copy_to_user(uifr32, &ifr, sizeof(ifr.ifr_name));
2932 2933 2934 2935 2936 2937
		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);
2938 2939 2940 2941 2942 2943
		if (err)
			err = -EFAULT;
	}
	return err;
}

2944
struct rtentry32 {
2945
	u32		rt_pad1;
2946 2947 2948
	struct sockaddr rt_dst;         /* target address               */
	struct sockaddr rt_gateway;     /* gateway addr (RTF_GATEWAY)   */
	struct sockaddr rt_genmask;     /* target network mask (IP)     */
2949 2950 2951 2952 2953 2954 2955
	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! */
2956
	/* char * */ u32 rt_dev;        /* forcing the device at add    */
2957 2958
	u32		rt_mtu;         /* per route MTU/Window         */
	u32		rt_window;      /* Window clamping              */
2959 2960 2961 2962 2963 2964 2965 2966 2967 2968 2969 2970 2971 2972 2973 2974
	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;
};

2975 2976
static int routing_ioctl(struct net *net, struct socket *sock,
			 unsigned int cmd, void __user *argp)
2977 2978 2979 2980 2981 2982 2983 2984 2985
{
	int ret;
	void *r = NULL;
	struct in6_rtmsg r6;
	struct rtentry r4;
	char devname[16];
	u32 rtdev;
	mm_segment_t old_fs = get_fs();

2986 2987
	if (sock && sock->sk && sock->sk->sk_family == AF_INET6) { /* ipv6 */
		struct in6_rtmsg32 __user *ur6 = argp;
2988
		ret = copy_from_user(&r6.rtmsg_dst, &(ur6->rtmsg_dst),
2989
			3 * sizeof(struct in6_addr));
2990 2991 2992 2993 2994 2995 2996
		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));
2997 2998 2999

		r = (void *) &r6;
	} else { /* ipv4 */
3000
		struct rtentry32 __user *ur4 = argp;
3001
		ret = copy_from_user(&r4.rt_dst, &(ur4->rt_dst),
3002
					3 * sizeof(struct sockaddr));
3003 3004 3005 3006 3007 3008
		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));
3009
		if (rtdev) {
3010
			ret |= copy_from_user(devname, compat_ptr(rtdev), 15);
3011 3012
			r4.rt_dev = (char __user __force *)devname;
			devname[15] = 0;
3013 3014 3015 3016 3017 3018 3019 3020 3021 3022 3023
		} else
			r4.rt_dev = NULL;

		r = (void *) &r4;
	}

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

3024
	set_fs(KERNEL_DS);
3025
	ret = sock_do_ioctl(net, sock, cmd, (unsigned long) r);
3026
	set_fs(old_fs);
3027 3028 3029 3030 3031 3032 3033

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
3034
 * use compatible ioctls
3035
 */
3036
static int old_bridge_ioctl(compat_ulong_t __user *argp)
3037
{
3038
	compat_ulong_t tmp;
3039

3040
	if (get_user(tmp, argp))
3041 3042 3043 3044 3045 3046
		return -EFAULT;
	if (tmp == BRCTL_GET_VERSION)
		return BRCTL_VERSION + 1;
	return -EINVAL;
}

3047 3048 3049 3050 3051 3052
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);
3053

3054
	if (cmd >= SIOCDEVPRIVATE && cmd <= (SIOCDEVPRIVATE + 15))
3055
		return compat_ifr_data_ioctl(net, cmd, argp);
3056 3057 3058 3059 3060 3061 3062 3063 3064 3065 3066

	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);
3067 3068
	case SIOCWANDEV:
		return compat_siocwandev(net, argp);
3069 3070 3071
	case SIOCGIFMAP:
	case SIOCSIFMAP:
		return compat_sioc_ifmap(net, cmd, argp);
3072 3073 3074 3075 3076 3077 3078 3079 3080 3081 3082 3083
	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);
3084 3085
	case SIOCBONDSLAVEINFOQUERY:
	case SIOCBONDINFOQUERY:
3086
	case SIOCSHWTSTAMP:
3087
	case SIOCGHWTSTAMP:
3088
		return compat_ifr_data_ioctl(net, cmd, argp);
3089 3090 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

	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:
3131 3132 3133 3134
	case SIOCSIFNAME:
	case SIOCGMIIPHY:
	case SIOCGMIIREG:
	case SIOCSMIIREG:
3135
		return dev_ifsioc(net, sock, cmd, argp);
3136

3137 3138 3139 3140
	case SIOCSARP:
	case SIOCGARP:
	case SIOCDARP:
	case SIOCATMARK:
3141 3142 3143
		return sock_do_ioctl(net, sock, cmd, arg);
	}

3144 3145
	return -ENOIOCTLCMD;
}
3146

3147
static long compat_sock_ioctl(struct file *file, unsigned int cmd,
3148
			      unsigned long arg)
3149 3150 3151
{
	struct socket *sock = file->private_data;
	int ret = -ENOIOCTLCMD;
3152 3153 3154 3155 3156
	struct sock *sk;
	struct net *net;

	sk = sock->sk;
	net = sock_net(sk);
3157 3158 3159 3160

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

3161 3162 3163 3164
	if (ret == -ENOIOCTLCMD &&
	    (cmd >= SIOCIWFIRST && cmd <= SIOCIWLAST))
		ret = compat_wext_handle_ioctl(net, cmd, arg);

3165 3166 3167
	if (ret == -ENOIOCTLCMD)
		ret = compat_sock_ioctl_trans(file, sock, cmd, arg);

3168 3169 3170 3171
	return ret;
}
#endif

3172 3173 3174 3175
int kernel_bind(struct socket *sock, struct sockaddr *addr, int addrlen)
{
	return sock->ops->bind(sock, addr, addrlen);
}
3176
EXPORT_SYMBOL(kernel_bind);
3177 3178 3179 3180 3181

int kernel_listen(struct socket *sock, int backlog)
{
	return sock->ops->listen(sock, backlog);
}
3182
EXPORT_SYMBOL(kernel_listen);
3183 3184 3185 3186 3187 3188 3189 3190 3191 3192 3193 3194 3195 3196

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);
3197
		*newsock = NULL;
3198 3199 3200 3201
		goto done;
	}

	(*newsock)->ops = sock->ops;
3202
	__module_get((*newsock)->ops->owner);
3203 3204 3205 3206

done:
	return err;
}
3207
EXPORT_SYMBOL(kernel_accept);
3208 3209

int kernel_connect(struct socket *sock, struct sockaddr *addr, int addrlen,
3210
		   int flags)
3211 3212 3213
{
	return sock->ops->connect(sock, addr, addrlen, flags);
}
3214
EXPORT_SYMBOL(kernel_connect);
3215 3216 3217 3218 3219 3220

int kernel_getsockname(struct socket *sock, struct sockaddr *addr,
			 int *addrlen)
{
	return sock->ops->getname(sock, addr, addrlen, 0);
}
3221
EXPORT_SYMBOL(kernel_getsockname);
3222 3223 3224 3225 3226 3227

int kernel_getpeername(struct socket *sock, struct sockaddr *addr,
			 int *addrlen)
{
	return sock->ops->getname(sock, addr, addrlen, 1);
}
3228
EXPORT_SYMBOL(kernel_getpeername);
3229 3230 3231 3232 3233

int kernel_getsockopt(struct socket *sock, int level, int optname,
			char *optval, int *optlen)
{
	mm_segment_t oldfs = get_fs();
3234 3235
	char __user *uoptval;
	int __user *uoptlen;
3236 3237
	int err;

3238 3239 3240
	uoptval = (char __user __force *) optval;
	uoptlen = (int __user __force *) optlen;

3241 3242
	set_fs(KERNEL_DS);
	if (level == SOL_SOCKET)
3243
		err = sock_getsockopt(sock, level, optname, uoptval, uoptlen);
3244
	else
3245 3246
		err = sock->ops->getsockopt(sock, level, optname, uoptval,
					    uoptlen);
3247 3248 3249
	set_fs(oldfs);
	return err;
}
3250
EXPORT_SYMBOL(kernel_getsockopt);
3251 3252

int kernel_setsockopt(struct socket *sock, int level, int optname,
3253
			char *optval, unsigned int optlen)
3254 3255
{
	mm_segment_t oldfs = get_fs();
3256
	char __user *uoptval;
3257 3258
	int err;

3259 3260
	uoptval = (char __user __force *) optval;

3261 3262
	set_fs(KERNEL_DS);
	if (level == SOL_SOCKET)
3263
		err = sock_setsockopt(sock, level, optname, uoptval, optlen);
3264
	else
3265
		err = sock->ops->setsockopt(sock, level, optname, uoptval,
3266 3267 3268 3269
					    optlen);
	set_fs(oldfs);
	return err;
}
3270
EXPORT_SYMBOL(kernel_setsockopt);
3271 3272 3273 3274 3275 3276 3277 3278 3279

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);
}
3280
EXPORT_SYMBOL(kernel_sendpage);
3281 3282 3283 3284 3285 3286 3287 3288 3289 3290 3291 3292

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;
}
3293
EXPORT_SYMBOL(kernel_sock_ioctl);
3294

3295 3296 3297 3298 3299
int kernel_sock_shutdown(struct socket *sock, enum sock_shutdown_cmd how)
{
	return sock->ops->shutdown(sock, how);
}
EXPORT_SYMBOL(kernel_sock_shutdown);