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
};

323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348
static int sockfs_xattr_get(const struct xattr_handler *handler,
			    struct dentry *dentry, struct inode *inode,
			    const char *suffix, void *value, size_t size)
{
	if (value) {
		if (dentry->d_name.len + 1 > size)
			return -ERANGE;
		memcpy(value, dentry->d_name.name, dentry->d_name.len + 1);
	}
	return dentry->d_name.len + 1;
}

#define XATTR_SOCKPROTONAME_SUFFIX "sockprotoname"
#define XATTR_NAME_SOCKPROTONAME (XATTR_SYSTEM_PREFIX XATTR_SOCKPROTONAME_SUFFIX)
#define XATTR_NAME_SOCKPROTONAME_LEN (sizeof(XATTR_NAME_SOCKPROTONAME)-1)

static const struct xattr_handler sockfs_xattr_handler = {
	.name = XATTR_NAME_SOCKPROTONAME,
	.get = sockfs_xattr_get,
};

static const struct xattr_handler *sockfs_xattr_handlers[] = {
	&sockfs_xattr_handler,
	NULL
};

349 350 351
static struct dentry *sockfs_mount(struct file_system_type *fs_type,
			 int flags, const char *dev_name, void *data)
{
352 353 354
	return mount_pseudo_xattr(fs_type, "socket:", &sockfs_ops,
				  sockfs_xattr_handlers,
				  &sockfs_dentry_operations, SOCKFS_MAGIC);
355 356 357 358 359 360 361 362 363 364
}

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
365 366 367
/*
 *	Obtains the first available file descriptor and sets it up for use.
 *
368 369
 *	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
370 371 372 373 374 375 376 377 378 379 380 381
 *	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.
 */

382
struct file *sock_alloc_file(struct socket *sock, int flags, const char *dname)
L
Linus Torvalds 已提交
383
{
384
	struct qstr name = { .name = "" };
385
	struct path path;
386
	struct file *file;
L
Linus Torvalds 已提交
387

388 389 390 391 392 393 394
	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);
	}
395
	path.dentry = d_alloc_pseudo(sock_mnt->mnt_sb, &name);
396 397
	if (unlikely(!path.dentry))
		return ERR_PTR(-ENOMEM);
398
	path.mnt = mntget(sock_mnt);
399

400
	d_instantiate(path.dentry, SOCK_INODE(sock));
401

402
	file = alloc_file(&path, FMODE_READ | FMODE_WRITE,
403
		  &socket_file_ops);
404
	if (IS_ERR(file)) {
A
Al Viro 已提交
405
		/* drop dentry, keep inode */
D
David Howells 已提交
406
		ihold(d_inode(path.dentry));
407
		path_put(&path);
408
		return file;
A
Al Viro 已提交
409 410 411
	}

	sock->file = file;
412
	file->f_flags = O_RDWR | (flags & O_NONBLOCK);
413
	file->private_data = sock;
414
	return file;
415
}
416
EXPORT_SYMBOL(sock_alloc_file);
417

418
static int sock_map_fd(struct socket *sock, int flags)
419 420
{
	struct file *newfile;
421 422 423
	int fd = get_unused_fd_flags(flags);
	if (unlikely(fd < 0))
		return fd;
424

425
	newfile = sock_alloc_file(sock, flags, NULL);
426
	if (likely(!IS_ERR(newfile))) {
427
		fd_install(fd, newfile);
428 429
		return fd;
	}
430

431 432
	put_unused_fd(fd);
	return PTR_ERR(newfile);
L
Linus Torvalds 已提交
433 434
}

435
struct socket *sock_from_file(struct file *file, int *err)
436 437 438 439
{
	if (file->f_op == &socket_file_ops)
		return file->private_data;	/* set in sock_map_fd */

E
Eric Dumazet 已提交
440 441
	*err = -ENOTSOCK;
	return NULL;
442
}
443
EXPORT_SYMBOL(sock_from_file);
444

L
Linus Torvalds 已提交
445
/**
446
 *	sockfd_lookup - Go from a file number to its socket slot
L
Linus Torvalds 已提交
447 448 449 450 451 452 453 454 455 456 457 458 459 460 461 462
 *	@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;

463 464
	file = fget(fd);
	if (!file) {
L
Linus Torvalds 已提交
465 466 467
		*err = -EBADF;
		return NULL;
	}
468

469 470
	sock = sock_from_file(file, err);
	if (!sock)
L
Linus Torvalds 已提交
471
		fput(file);
472 473
	return sock;
}
474
EXPORT_SYMBOL(sockfd_lookup);
L
Linus Torvalds 已提交
475

476 477
static struct socket *sockfd_lookup_light(int fd, int *err, int *fput_needed)
{
478
	struct fd f = fdget(fd);
479 480
	struct socket *sock;

481
	*err = -EBADF;
482 483 484 485
	if (f.file) {
		sock = sock_from_file(f.file, err);
		if (likely(sock)) {
			*fput_needed = f.flags;
486
			return sock;
487 488
		}
		fdput(f);
L
Linus Torvalds 已提交
489
	}
490
	return NULL;
L
Linus Torvalds 已提交
491 492
}

493 494 495 496 497 498
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
499
	len = security_inode_listsecurity(d_inode(dentry), buffer, size);
500 501 502 503 504 505 506 507 508 509 510 511 512 513 514 515 516 517 518 519 520 521 522 523 524
	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 = {
	.listxattr = sockfs_listxattr,
};

L
Linus Torvalds 已提交
525 526
/**
 *	sock_alloc	-	allocate a socket
527
 *
L
Linus Torvalds 已提交
528 529 530 531 532
 *	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
533
struct socket *sock_alloc(void)
L
Linus Torvalds 已提交
534
{
535 536
	struct inode *inode;
	struct socket *sock;
L
Linus Torvalds 已提交
537

538
	inode = new_inode_pseudo(sock_mnt->mnt_sb);
L
Linus Torvalds 已提交
539 540 541 542 543
	if (!inode)
		return NULL;

	sock = SOCKET_I(inode);

544
	kmemcheck_annotate_bitfield(sock, type);
545
	inode->i_ino = get_next_ino();
546
	inode->i_mode = S_IFSOCK | S_IRWXUGO;
547 548
	inode->i_uid = current_fsuid();
	inode->i_gid = current_fsgid();
549
	inode->i_op = &sockfs_inode_ops;
L
Linus Torvalds 已提交
550

551
	this_cpu_add(sockets_in_use, 1);
L
Linus Torvalds 已提交
552 553
	return sock;
}
T
Tom Herbert 已提交
554
EXPORT_SYMBOL(sock_alloc);
L
Linus Torvalds 已提交
555 556 557 558 559 560 561

/**
 *	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
562
 *	an inode not a file.
L
Linus Torvalds 已提交
563
 */
564

L
Linus Torvalds 已提交
565 566 567 568 569 570 571 572 573 574
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);
	}

575
	if (rcu_dereference_protected(sock->wq, 1)->fasync_list)
576
		pr_err("%s: fasync list not empty!\n", __func__);
L
Linus Torvalds 已提交
577

578
	this_cpu_sub(sockets_in_use, 1);
L
Linus Torvalds 已提交
579 580 581 582
	if (!sock->file) {
		iput(SOCK_INODE(sock));
		return;
	}
583
	sock->file = NULL;
L
Linus Torvalds 已提交
584
}
585
EXPORT_SYMBOL(sock_release);
L
Linus Torvalds 已提交
586

587
void __sock_tx_timestamp(__u16 tsflags, __u8 *tx_flags)
588
{
589 590
	u8 flags = *tx_flags;

591
	if (tsflags & SOF_TIMESTAMPING_TX_HARDWARE)
592 593
		flags |= SKBTX_HW_TSTAMP;

594
	if (tsflags & SOF_TIMESTAMPING_TX_SOFTWARE)
595 596
		flags |= SKBTX_SW_TSTAMP;

597
	if (tsflags & SOF_TIMESTAMPING_TX_SCHED)
598 599 600
		flags |= SKBTX_SCHED_TSTAMP;

	*tx_flags = flags;
601
}
602
EXPORT_SYMBOL(__sock_tx_timestamp);
603

604
static inline int sock_sendmsg_nosec(struct socket *sock, struct msghdr *msg)
L
Linus Torvalds 已提交
605
{
A
Al Viro 已提交
606
	int ret = sock->ops->sendmsg(sock, msg, msg_data_left(msg));
607 608
	BUG_ON(ret == -EIOCBQUEUED);
	return ret;
L
Linus Torvalds 已提交
609 610
}

611
int sock_sendmsg(struct socket *sock, struct msghdr *msg)
612
{
613
	int err = security_socket_sendmsg(sock, msg,
A
Al Viro 已提交
614
					  msg_data_left(msg));
615

616
	return err ?: sock_sendmsg_nosec(sock, msg);
617
}
618
EXPORT_SYMBOL(sock_sendmsg);
L
Linus Torvalds 已提交
619 620 621 622

int kernel_sendmsg(struct socket *sock, struct msghdr *msg,
		   struct kvec *vec, size_t num, size_t size)
{
623
	iov_iter_kvec(&msg->msg_iter, WRITE | ITER_KVEC, vec, num, size);
624
	return sock_sendmsg(sock, msg);
L
Linus Torvalds 已提交
625
}
626
EXPORT_SYMBOL(kernel_sendmsg);
L
Linus Torvalds 已提交
627

628 629 630 631 632 633
/*
 * 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)
{
634
	int need_software_tstamp = sock_flag(sk, SOCK_RCVTSTAMP);
635
	struct scm_timestamping tss;
636 637 638 639 640 641 642 643 644 645 646 647 648 649 650 651
	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 {
652 653
			struct timespec ts;
			skb_get_timestampns(skb, &ts);
654
			put_cmsg(msg, SOL_SOCKET, SCM_TIMESTAMPNS,
655
				 sizeof(ts), &ts);
656 657 658
		}
	}

659
	memset(&tss, 0, sizeof(tss));
660
	if ((sk->sk_tsflags & SOF_TIMESTAMPING_SOFTWARE) &&
661
	    ktime_to_timespec_cond(skb->tstamp, tss.ts + 0))
662
		empty = 0;
663
	if (shhwtstamps &&
664
	    (sk->sk_tsflags & SOF_TIMESTAMPING_RAW_HARDWARE) &&
665
	    ktime_to_timespec_cond(shhwtstamps->hwtstamp, tss.ts + 2))
666
		empty = 0;
667 668
	if (!empty)
		put_cmsg(msg, SOL_SOCKET,
669
			 SCM_TIMESTAMPING, sizeof(tss), &tss);
670
}
671 672
EXPORT_SYMBOL_GPL(__sock_recv_timestamp);

673 674 675 676 677 678 679 680 681 682 683 684 685 686 687 688
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
689 690
static inline void sock_recv_drops(struct msghdr *msg, struct sock *sk,
				   struct sk_buff *skb)
691
{
692
	if (sock_flag(sk, SOCK_RXQ_OVFL) && skb && SOCK_SKB_CB(skb)->dropcount)
693
		put_cmsg(msg, SOL_SOCKET, SO_RXQ_OVFL,
694
			sizeof(__u32), &SOCK_SKB_CB(skb)->dropcount);
695 696
}

697
void __sock_recv_ts_and_drops(struct msghdr *msg, struct sock *sk,
698 699 700 701 702
	struct sk_buff *skb)
{
	sock_recv_timestamp(msg, sk, skb);
	sock_recv_drops(msg, sk, skb);
}
703
EXPORT_SYMBOL_GPL(__sock_recv_ts_and_drops);
704

705
static inline int sock_recvmsg_nosec(struct socket *sock, struct msghdr *msg,
706
				     int flags)
L
Linus Torvalds 已提交
707
{
708
	return sock->ops->recvmsg(sock, msg, msg_data_left(msg), flags);
L
Linus Torvalds 已提交
709 710
}

711
int sock_recvmsg(struct socket *sock, struct msghdr *msg, int flags)
712
{
713
	int err = security_socket_recvmsg(sock, msg, msg_data_left(msg), flags);
714

715
	return err ?: sock_recvmsg_nosec(sock, msg, flags);
L
Linus Torvalds 已提交
716
}
717
EXPORT_SYMBOL(sock_recvmsg);
L
Linus Torvalds 已提交
718

719 720 721 722 723 724 725 726 727 728 729 730 731 732 733
/**
 * 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.
 */
734 735
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
736 737 738 739
{
	mm_segment_t oldfs = get_fs();
	int result;

740
	iov_iter_kvec(&msg->msg_iter, READ | ITER_KVEC, vec, num, size);
L
Linus Torvalds 已提交
741
	set_fs(KERNEL_DS);
742
	result = sock_recvmsg(sock, msg, flags);
L
Linus Torvalds 已提交
743 744 745
	set_fs(oldfs);
	return result;
}
746
EXPORT_SYMBOL(kernel_recvmsg);
L
Linus Torvalds 已提交
747

748 749
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
750 751 752 753
{
	struct socket *sock;
	int flags;

754 755
	sock = file->private_data;

756 757 758
	flags = (file->f_flags & O_NONBLOCK) ? MSG_DONTWAIT : 0;
	/* more is a combination of MSG_MORE and MSG_SENDPAGE_NOTLAST */
	flags |= more;
759

760
	return kernel_sendpage(sock, page, offset, size, flags);
761
}
L
Linus Torvalds 已提交
762

J
Jens Axboe 已提交
763
static ssize_t sock_splice_read(struct file *file, loff_t *ppos,
764
				struct pipe_inode_info *pipe, size_t len,
J
Jens Axboe 已提交
765 766 767 768
				unsigned int flags)
{
	struct socket *sock = file->private_data;

769 770 771
	if (unlikely(!sock->ops->splice_read))
		return -EINVAL;

J
Jens Axboe 已提交
772 773 774
	return sock->ops->splice_read(sock, ppos, pipe, len, flags);
}

775
static ssize_t sock_read_iter(struct kiocb *iocb, struct iov_iter *to)
776
{
777 778
	struct file *file = iocb->ki_filp;
	struct socket *sock = file->private_data;
779 780
	struct msghdr msg = {.msg_iter = *to,
			     .msg_iocb = iocb};
781
	ssize_t res;
782

783 784 785 786
	if (file->f_flags & O_NONBLOCK)
		msg.msg_flags = MSG_DONTWAIT;

	if (iocb->ki_pos != 0)
L
Linus Torvalds 已提交
787
		return -ESPIPE;
788

C
Christoph Hellwig 已提交
789
	if (!iov_iter_count(to))	/* Match SYS5 behaviour */
L
Linus Torvalds 已提交
790 791
		return 0;

792
	res = sock_recvmsg(sock, &msg, msg.msg_flags);
793 794
	*to = msg.msg_iter;
	return res;
L
Linus Torvalds 已提交
795 796
}

797
static ssize_t sock_write_iter(struct kiocb *iocb, struct iov_iter *from)
798
{
799 800
	struct file *file = iocb->ki_filp;
	struct socket *sock = file->private_data;
801 802
	struct msghdr msg = {.msg_iter = *from,
			     .msg_iocb = iocb};
803
	ssize_t res;
L
Linus Torvalds 已提交
804

805
	if (iocb->ki_pos != 0)
806
		return -ESPIPE;
807

808 809 810
	if (file->f_flags & O_NONBLOCK)
		msg.msg_flags = MSG_DONTWAIT;

811 812 813
	if (sock->type == SOCK_SEQPACKET)
		msg.msg_flags |= MSG_EOR;

814
	res = sock_sendmsg(sock, &msg);
815 816
	*from = msg.msg_iter;
	return res;
L
Linus Torvalds 已提交
817 818 819 820 821 822 823
}

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

A
Arjan van de Ven 已提交
824
static DEFINE_MUTEX(br_ioctl_mutex);
825
static int (*br_ioctl_hook) (struct net *, unsigned int cmd, void __user *arg);
L
Linus Torvalds 已提交
826

827
void brioctl_set(int (*hook) (struct net *, unsigned int, void __user *))
L
Linus Torvalds 已提交
828
{
A
Arjan van de Ven 已提交
829
	mutex_lock(&br_ioctl_mutex);
L
Linus Torvalds 已提交
830
	br_ioctl_hook = hook;
A
Arjan van de Ven 已提交
831
	mutex_unlock(&br_ioctl_mutex);
L
Linus Torvalds 已提交
832 833 834
}
EXPORT_SYMBOL(brioctl_set);

A
Arjan van de Ven 已提交
835
static DEFINE_MUTEX(vlan_ioctl_mutex);
836
static int (*vlan_ioctl_hook) (struct net *, void __user *arg);
L
Linus Torvalds 已提交
837

838
void vlan_ioctl_set(int (*hook) (struct net *, void __user *))
L
Linus Torvalds 已提交
839
{
A
Arjan van de Ven 已提交
840
	mutex_lock(&vlan_ioctl_mutex);
L
Linus Torvalds 已提交
841
	vlan_ioctl_hook = hook;
A
Arjan van de Ven 已提交
842
	mutex_unlock(&vlan_ioctl_mutex);
L
Linus Torvalds 已提交
843 844 845
}
EXPORT_SYMBOL(vlan_ioctl_set);

A
Arjan van de Ven 已提交
846
static DEFINE_MUTEX(dlci_ioctl_mutex);
847
static int (*dlci_ioctl_hook) (unsigned int, void __user *);
L
Linus Torvalds 已提交
848

849
void dlci_ioctl_set(int (*hook) (unsigned int, void __user *))
L
Linus Torvalds 已提交
850
{
A
Arjan van de Ven 已提交
851
	mutex_lock(&dlci_ioctl_mutex);
L
Linus Torvalds 已提交
852
	dlci_ioctl_hook = hook;
A
Arjan van de Ven 已提交
853
	mutex_unlock(&dlci_ioctl_mutex);
L
Linus Torvalds 已提交
854 855 856
}
EXPORT_SYMBOL(dlci_ioctl_set);

857 858 859 860 861 862 863 864 865 866 867 868 869 870 871 872 873 874
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
875 876 877 878 879 880 881 882
/*
 *	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;
883
	struct sock *sk;
L
Linus Torvalds 已提交
884 885
	void __user *argp = (void __user *)arg;
	int pid, err;
886
	struct net *net;
L
Linus Torvalds 已提交
887

888
	sock = file->private_data;
889
	sk = sock->sk;
890
	net = sock_net(sk);
L
Linus Torvalds 已提交
891
	if (cmd >= SIOCDEVPRIVATE && cmd <= (SIOCDEVPRIVATE + 15)) {
892
		err = dev_ioctl(net, cmd, argp);
L
Linus Torvalds 已提交
893
	} else
J
Johannes Berg 已提交
894
#ifdef CONFIG_WEXT_CORE
L
Linus Torvalds 已提交
895
	if (cmd >= SIOCIWFIRST && cmd <= SIOCIWLAST) {
896
		err = dev_ioctl(net, cmd, argp);
L
Linus Torvalds 已提交
897
	} else
J
Johannes Berg 已提交
898
#endif
899
		switch (cmd) {
L
Linus Torvalds 已提交
900 901 902 903 904
		case FIOSETOWN:
		case SIOCSPGRP:
			err = -EFAULT;
			if (get_user(pid, (int __user *)argp))
				break;
905 906
			f_setown(sock->file, pid, 1);
			err = 0;
L
Linus Torvalds 已提交
907 908 909
			break;
		case FIOGETOWN:
		case SIOCGPGRP:
910
			err = put_user(f_getown(sock->file),
911
				       (int __user *)argp);
L
Linus Torvalds 已提交
912 913 914 915 916 917 918 919 920
			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
921
			mutex_lock(&br_ioctl_mutex);
922
			if (br_ioctl_hook)
923
				err = br_ioctl_hook(net, cmd, argp);
A
Arjan van de Ven 已提交
924
			mutex_unlock(&br_ioctl_mutex);
L
Linus Torvalds 已提交
925 926 927 928 929 930 931
			break;
		case SIOCGIFVLAN:
		case SIOCSIFVLAN:
			err = -ENOPKG;
			if (!vlan_ioctl_hook)
				request_module("8021q");

A
Arjan van de Ven 已提交
932
			mutex_lock(&vlan_ioctl_mutex);
L
Linus Torvalds 已提交
933
			if (vlan_ioctl_hook)
934
				err = vlan_ioctl_hook(net, argp);
A
Arjan van de Ven 已提交
935
			mutex_unlock(&vlan_ioctl_mutex);
L
Linus Torvalds 已提交
936 937 938 939 940 941 942
			break;
		case SIOCADDDLCI:
		case SIOCDELDLCI:
			err = -ENOPKG;
			if (!dlci_ioctl_hook)
				request_module("dlci");

943 944
			mutex_lock(&dlci_ioctl_mutex);
			if (dlci_ioctl_hook)
L
Linus Torvalds 已提交
945
				err = dlci_ioctl_hook(cmd, argp);
946
			mutex_unlock(&dlci_ioctl_mutex);
L
Linus Torvalds 已提交
947 948
			break;
		default:
949
			err = sock_do_ioctl(net, sock, cmd, arg);
L
Linus Torvalds 已提交
950
			break;
951
		}
L
Linus Torvalds 已提交
952 953 954 955 956 957 958
	return err;
}

int sock_create_lite(int family, int type, int protocol, struct socket **res)
{
	int err;
	struct socket *sock = NULL;
959

L
Linus Torvalds 已提交
960 961 962 963 964 965 966 967 968 969 970
	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
971 972 973 974
	err = security_socket_post_create(sock, family, type, protocol, 1);
	if (err)
		goto out_release;

L
Linus Torvalds 已提交
975 976 977
out:
	*res = sock;
	return err;
V
Venkat Yekkirala 已提交
978 979 980 981
out_release:
	sock_release(sock);
	sock = NULL;
	goto out;
L
Linus Torvalds 已提交
982
}
983
EXPORT_SYMBOL(sock_create_lite);
L
Linus Torvalds 已提交
984 985

/* No kernel lock held - perfect */
986
static unsigned int sock_poll(struct file *file, poll_table *wait)
L
Linus Torvalds 已提交
987
{
988
	unsigned int busy_flag = 0;
L
Linus Torvalds 已提交
989 990 991
	struct socket *sock;

	/*
992
	 *      We can't return errors to poll, so it's either yes or no.
L
Linus Torvalds 已提交
993
	 */
994
	sock = file->private_data;
995

996
	if (sk_can_busy_loop(sock->sk)) {
997
		/* this socket can poll_ll so tell the system call */
998
		busy_flag = POLL_BUSY_LOOP;
999 1000

		/* once, only if requested by syscall */
1001 1002
		if (wait && (wait->_key & POLL_BUSY_LOOP))
			sk_busy_loop(sock->sk, 1);
1003 1004
	}

1005
	return busy_flag | sock->ops->poll(file, sock, wait);
L
Linus Torvalds 已提交
1006 1007
}

1008
static int sock_mmap(struct file *file, struct vm_area_struct *vma)
L
Linus Torvalds 已提交
1009
{
1010
	struct socket *sock = file->private_data;
L
Linus Torvalds 已提交
1011 1012 1013 1014

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

1015
static int sock_close(struct inode *inode, struct file *filp)
L
Linus Torvalds 已提交
1016 1017 1018 1019 1020 1021 1022 1023 1024 1025 1026 1027 1028
{
	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)
1029
 *	   or under socket lock
L
Linus Torvalds 已提交
1030 1031 1032 1033
 */

static int sock_fasync(int fd, struct file *filp, int on)
{
1034 1035
	struct socket *sock = filp->private_data;
	struct sock *sk = sock->sk;
1036
	struct socket_wq *wq;
L
Linus Torvalds 已提交
1037

1038
	if (sk == NULL)
L
Linus Torvalds 已提交
1039 1040 1041
		return -EINVAL;

	lock_sock(sk);
1042
	wq = rcu_dereference_protected(sock->wq, lockdep_sock_is_held(sk));
1043
	fasync_helper(fd, filp, on, &wq->fasync_list);
L
Linus Torvalds 已提交
1044

1045
	if (!wq->fasync_list)
1046 1047
		sock_reset_flag(sk, SOCK_FASYNC);
	else
E
Eric Dumazet 已提交
1048
		sock_set_flag(sk, SOCK_FASYNC);
L
Linus Torvalds 已提交
1049

1050
	release_sock(sk);
L
Linus Torvalds 已提交
1051 1052 1053
	return 0;
}

1054
/* This function may be called only under rcu_lock */
L
Linus Torvalds 已提交
1055

1056
int sock_wake_async(struct socket_wq *wq, int how, int band)
L
Linus Torvalds 已提交
1057
{
1058
	if (!wq || !wq->fasync_list)
L
Linus Torvalds 已提交
1059
		return -1;
1060

1061
	switch (how) {
1062
	case SOCK_WAKE_WAITD:
1063
		if (test_bit(SOCKWQ_ASYNC_WAITDATA, &wq->flags))
L
Linus Torvalds 已提交
1064 1065
			break;
		goto call_kill;
1066
	case SOCK_WAKE_SPACE:
1067
		if (!test_and_clear_bit(SOCKWQ_ASYNC_NOSPACE, &wq->flags))
L
Linus Torvalds 已提交
1068 1069
			break;
		/* fall through */
1070
	case SOCK_WAKE_IO:
1071
call_kill:
1072
		kill_fasync(&wq->fasync_list, SIGIO, band);
L
Linus Torvalds 已提交
1073
		break;
1074
	case SOCK_WAKE_URG:
1075
		kill_fasync(&wq->fasync_list, SIGURG, band);
L
Linus Torvalds 已提交
1076
	}
1077

L
Linus Torvalds 已提交
1078 1079
	return 0;
}
1080
EXPORT_SYMBOL(sock_wake_async);
L
Linus Torvalds 已提交
1081

P
Pavel Emelyanov 已提交
1082
int __sock_create(struct net *net, int family, int type, int protocol,
1083
			 struct socket **res, int kern)
L
Linus Torvalds 已提交
1084 1085 1086
{
	int err;
	struct socket *sock;
1087
	const struct net_proto_family *pf;
L
Linus Torvalds 已提交
1088 1089

	/*
1090
	 *      Check protocol is in range
L
Linus Torvalds 已提交
1091 1092 1093 1094 1095 1096 1097 1098 1099 1100 1101 1102
	 */
	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) {
1103 1104
		pr_info_once("%s uses obsolete (PF_INET,SOCK_PACKET)\n",
			     current->comm);
L
Linus Torvalds 已提交
1105 1106 1107 1108 1109 1110
		family = PF_PACKET;
	}

	err = security_socket_create(family, type, protocol, kern);
	if (err)
		return err;
1111

1112 1113 1114 1115 1116 1117 1118
	/*
	 *	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) {
1119
		net_warn_ratelimited("socket: no more sockets\n");
1120 1121 1122 1123 1124 1125
		return -ENFILE;	/* Not exactly a match, but its the
				   closest posix thing */
	}

	sock->type = type;

1126
#ifdef CONFIG_MODULES
1127 1128 1129
	/* 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
1130 1131 1132
	 * requested real, full-featured networking support upon configuration.
	 * Otherwise module support will break!
	 */
E
Eric Dumazet 已提交
1133
	if (rcu_access_pointer(net_families[family]) == NULL)
1134
		request_module("net-pf-%d", family);
L
Linus Torvalds 已提交
1135 1136
#endif

1137 1138 1139 1140 1141
	rcu_read_lock();
	pf = rcu_dereference(net_families[family]);
	err = -EAFNOSUPPORT;
	if (!pf)
		goto out_release;
L
Linus Torvalds 已提交
1142 1143 1144 1145 1146

	/*
	 * We will call the ->create function, that possibly is in a loadable
	 * module, so we have to bump that loadable module refcnt first.
	 */
1147
	if (!try_module_get(pf->owner))
L
Linus Torvalds 已提交
1148 1149
		goto out_release;

1150 1151 1152
	/* Now protected by module ref count */
	rcu_read_unlock();

1153
	err = pf->create(net, sock, protocol, kern);
1154
	if (err < 0)
L
Linus Torvalds 已提交
1155
		goto out_module_put;
1156

L
Linus Torvalds 已提交
1157 1158 1159 1160
	/*
	 * Now to bump the refcnt of the [loadable] module that owns this
	 * socket at sock_release time we decrement its refcnt.
	 */
1161 1162 1163
	if (!try_module_get(sock->ops->owner))
		goto out_module_busy;

L
Linus Torvalds 已提交
1164 1165 1166 1167
	/*
	 * Now that we're done with the ->create function, the [loadable]
	 * module can have its refcnt decremented
	 */
1168
	module_put(pf->owner);
V
Venkat Yekkirala 已提交
1169 1170
	err = security_socket_post_create(sock, family, type, protocol, kern);
	if (err)
1171
		goto out_sock_release;
1172
	*res = sock;
L
Linus Torvalds 已提交
1173

1174 1175 1176 1177
	return 0;

out_module_busy:
	err = -EAFNOSUPPORT;
L
Linus Torvalds 已提交
1178
out_module_put:
1179 1180 1181
	sock->ops = NULL;
	module_put(pf->owner);
out_sock_release:
L
Linus Torvalds 已提交
1182
	sock_release(sock);
1183 1184 1185 1186 1187
	return err;

out_release:
	rcu_read_unlock();
	goto out_sock_release;
L
Linus Torvalds 已提交
1188
}
P
Pavel Emelyanov 已提交
1189
EXPORT_SYMBOL(__sock_create);
L
Linus Torvalds 已提交
1190 1191 1192

int sock_create(int family, int type, int protocol, struct socket **res)
{
1193
	return __sock_create(current->nsproxy->net_ns, family, type, protocol, res, 0);
L
Linus Torvalds 已提交
1194
}
1195
EXPORT_SYMBOL(sock_create);
L
Linus Torvalds 已提交
1196

1197
int sock_create_kern(struct net *net, int family, int type, int protocol, struct socket **res)
L
Linus Torvalds 已提交
1198
{
1199
	return __sock_create(net, family, type, protocol, res, 1);
L
Linus Torvalds 已提交
1200
}
1201
EXPORT_SYMBOL(sock_create_kern);
L
Linus Torvalds 已提交
1202

1203
SYSCALL_DEFINE3(socket, int, family, int, type, int, protocol)
L
Linus Torvalds 已提交
1204 1205 1206
{
	int retval;
	struct socket *sock;
1207 1208
	int flags;

1209 1210 1211 1212 1213 1214
	/* 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);

1215
	flags = type & ~SOCK_TYPE_MASK;
1216
	if (flags & ~(SOCK_CLOEXEC | SOCK_NONBLOCK))
1217 1218
		return -EINVAL;
	type &= SOCK_TYPE_MASK;
L
Linus Torvalds 已提交
1219

U
Ulrich Drepper 已提交
1220 1221 1222
	if (SOCK_NONBLOCK != O_NONBLOCK && (flags & SOCK_NONBLOCK))
		flags = (flags & ~SOCK_NONBLOCK) | O_NONBLOCK;

L
Linus Torvalds 已提交
1223 1224 1225 1226
	retval = sock_create(family, type, protocol, &sock);
	if (retval < 0)
		goto out;

1227
	retval = sock_map_fd(sock, flags & (O_CLOEXEC | O_NONBLOCK));
L
Linus Torvalds 已提交
1228 1229 1230 1231 1232 1233 1234 1235 1236 1237 1238 1239 1240 1241 1242 1243
	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.
 */

1244 1245
SYSCALL_DEFINE4(socketpair, int, family, int, type, int, protocol,
		int __user *, usockvec)
L
Linus Torvalds 已提交
1246 1247 1248
{
	struct socket *sock1, *sock2;
	int fd1, fd2, err;
A
Al Viro 已提交
1249
	struct file *newfile1, *newfile2;
1250 1251 1252
	int flags;

	flags = type & ~SOCK_TYPE_MASK;
1253
	if (flags & ~(SOCK_CLOEXEC | SOCK_NONBLOCK))
1254 1255
		return -EINVAL;
	type &= SOCK_TYPE_MASK;
L
Linus Torvalds 已提交
1256

U
Ulrich Drepper 已提交
1257 1258 1259
	if (SOCK_NONBLOCK != O_NONBLOCK && (flags & SOCK_NONBLOCK))
		flags = (flags & ~SOCK_NONBLOCK) | O_NONBLOCK;

L
Linus Torvalds 已提交
1260 1261 1262 1263 1264 1265 1266 1267 1268 1269 1270 1271 1272 1273
	/*
	 * 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);
1274
	if (err < 0)
L
Linus Torvalds 已提交
1275 1276
		goto out_release_both;

1277
	fd1 = get_unused_fd_flags(flags);
1278 1279
	if (unlikely(fd1 < 0)) {
		err = fd1;
A
Al Viro 已提交
1280
		goto out_release_both;
1281
	}
1282

1283
	fd2 = get_unused_fd_flags(flags);
1284 1285
	if (unlikely(fd2 < 0)) {
		err = fd2;
1286
		goto out_put_unused_1;
1287 1288
	}

1289
	newfile1 = sock_alloc_file(sock1, flags, NULL);
1290
	if (IS_ERR(newfile1)) {
1291
		err = PTR_ERR(newfile1);
1292
		goto out_put_unused_both;
1293 1294
	}

1295
	newfile2 = sock_alloc_file(sock2, flags, NULL);
1296 1297
	if (IS_ERR(newfile2)) {
		err = PTR_ERR(newfile2);
1298
		goto out_fput_1;
A
Al Viro 已提交
1299 1300
	}

1301 1302 1303 1304 1305 1306 1307 1308
	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
1309
	audit_fd_pair(fd1, fd2);
1310

A
Al Viro 已提交
1311 1312
	fd_install(fd1, newfile1);
	fd_install(fd2, newfile2);
L
Linus Torvalds 已提交
1313 1314 1315 1316
	/* fd1 and fd2 may be already another descriptors.
	 * Not kernel problem.
	 */

1317
	return 0;
L
Linus Torvalds 已提交
1318

1319 1320 1321 1322 1323 1324 1325 1326 1327 1328 1329 1330 1331
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
1332

1333 1334 1335 1336
out_put_unused_both:
	put_unused_fd(fd2);
out_put_unused_1:
	put_unused_fd(fd1);
L
Linus Torvalds 已提交
1337
out_release_both:
1338
	sock_release(sock2);
L
Linus Torvalds 已提交
1339
out_release_1:
1340
	sock_release(sock1);
L
Linus Torvalds 已提交
1341 1342 1343 1344 1345 1346 1347 1348 1349 1350 1351 1352
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).
 */

1353
SYSCALL_DEFINE3(bind, int, fd, struct sockaddr __user *, umyaddr, int, addrlen)
L
Linus Torvalds 已提交
1354 1355
{
	struct socket *sock;
1356
	struct sockaddr_storage address;
1357
	int err, fput_needed;
L
Linus Torvalds 已提交
1358

1359
	sock = sockfd_lookup_light(fd, &err, &fput_needed);
1360
	if (sock) {
1361
		err = move_addr_to_kernel(umyaddr, addrlen, &address);
1362 1363
		if (err >= 0) {
			err = security_socket_bind(sock,
1364
						   (struct sockaddr *)&address,
1365
						   addrlen);
1366 1367
			if (!err)
				err = sock->ops->bind(sock,
1368
						      (struct sockaddr *)
1369
						      &address, addrlen);
L
Linus Torvalds 已提交
1370
		}
1371
		fput_light(sock->file, fput_needed);
1372
	}
L
Linus Torvalds 已提交
1373 1374 1375 1376 1377 1378 1379 1380 1381
	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.
 */

1382
SYSCALL_DEFINE2(listen, int, fd, int, backlog)
L
Linus Torvalds 已提交
1383 1384
{
	struct socket *sock;
1385
	int err, fput_needed;
1386
	int somaxconn;
1387 1388 1389

	sock = sockfd_lookup_light(fd, &err, &fput_needed);
	if (sock) {
1390
		somaxconn = sock_net(sock->sk)->core.sysctl_somaxconn;
1391
		if ((unsigned int)backlog > somaxconn)
1392
			backlog = somaxconn;
L
Linus Torvalds 已提交
1393 1394

		err = security_socket_listen(sock, backlog);
1395 1396
		if (!err)
			err = sock->ops->listen(sock, backlog);
L
Linus Torvalds 已提交
1397

1398
		fput_light(sock->file, fput_needed);
L
Linus Torvalds 已提交
1399 1400 1401 1402 1403 1404 1405 1406 1407 1408 1409 1410 1411 1412 1413 1414
	}
	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.
 */

1415 1416
SYSCALL_DEFINE4(accept4, int, fd, struct sockaddr __user *, upeer_sockaddr,
		int __user *, upeer_addrlen, int, flags)
L
Linus Torvalds 已提交
1417 1418
{
	struct socket *sock, *newsock;
1419
	struct file *newfile;
1420
	int err, len, newfd, fput_needed;
1421
	struct sockaddr_storage address;
L
Linus Torvalds 已提交
1422

1423
	if (flags & ~(SOCK_CLOEXEC | SOCK_NONBLOCK))
U
Ulrich Drepper 已提交
1424 1425 1426 1427 1428
		return -EINVAL;

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

1429
	sock = sockfd_lookup_light(fd, &err, &fput_needed);
L
Linus Torvalds 已提交
1430 1431 1432 1433
	if (!sock)
		goto out;

	err = -ENFILE;
1434 1435
	newsock = sock_alloc();
	if (!newsock)
L
Linus Torvalds 已提交
1436 1437 1438 1439 1440 1441 1442 1443 1444 1445 1446
		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);

1447
	newfd = get_unused_fd_flags(flags);
1448 1449
	if (unlikely(newfd < 0)) {
		err = newfd;
1450 1451
		sock_release(newsock);
		goto out_put;
1452
	}
1453
	newfile = sock_alloc_file(newsock, flags, sock->sk->sk_prot_creator->name);
1454
	if (IS_ERR(newfile)) {
1455 1456 1457 1458 1459
		err = PTR_ERR(newfile);
		put_unused_fd(newfd);
		sock_release(newsock);
		goto out_put;
	}
1460

1461 1462
	err = security_socket_accept(sock, newsock);
	if (err)
1463
		goto out_fd;
1464

L
Linus Torvalds 已提交
1465 1466
	err = sock->ops->accept(sock, newsock, sock->file->f_flags);
	if (err < 0)
1467
		goto out_fd;
L
Linus Torvalds 已提交
1468 1469

	if (upeer_sockaddr) {
1470
		if (newsock->ops->getname(newsock, (struct sockaddr *)&address,
1471
					  &len, 2) < 0) {
L
Linus Torvalds 已提交
1472
			err = -ECONNABORTED;
1473
			goto out_fd;
L
Linus Torvalds 已提交
1474
		}
1475
		err = move_addr_to_user(&address,
1476
					len, upeer_sockaddr, upeer_addrlen);
L
Linus Torvalds 已提交
1477
		if (err < 0)
1478
			goto out_fd;
L
Linus Torvalds 已提交
1479 1480 1481 1482
	}

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

1483 1484
	fd_install(newfd, newfile);
	err = newfd;
L
Linus Torvalds 已提交
1485 1486

out_put:
1487
	fput_light(sock->file, fput_needed);
L
Linus Torvalds 已提交
1488 1489
out:
	return err;
1490
out_fd:
1491
	fput(newfile);
1492
	put_unused_fd(newfd);
L
Linus Torvalds 已提交
1493 1494 1495
	goto out_put;
}

1496 1497
SYSCALL_DEFINE3(accept, int, fd, struct sockaddr __user *, upeer_sockaddr,
		int __user *, upeer_addrlen)
U
Ulrich Drepper 已提交
1498
{
U
Ulrich Drepper 已提交
1499
	return sys_accept4(fd, upeer_sockaddr, upeer_addrlen, 0);
U
Ulrich Drepper 已提交
1500 1501
}

L
Linus Torvalds 已提交
1502 1503 1504 1505 1506 1507 1508 1509 1510 1511 1512 1513
/*
 *	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.
 */

1514 1515
SYSCALL_DEFINE3(connect, int, fd, struct sockaddr __user *, uservaddr,
		int, addrlen)
L
Linus Torvalds 已提交
1516 1517
{
	struct socket *sock;
1518
	struct sockaddr_storage address;
1519
	int err, fput_needed;
L
Linus Torvalds 已提交
1520

1521
	sock = sockfd_lookup_light(fd, &err, &fput_needed);
L
Linus Torvalds 已提交
1522 1523
	if (!sock)
		goto out;
1524
	err = move_addr_to_kernel(uservaddr, addrlen, &address);
L
Linus Torvalds 已提交
1525 1526 1527
	if (err < 0)
		goto out_put;

1528
	err =
1529
	    security_socket_connect(sock, (struct sockaddr *)&address, addrlen);
L
Linus Torvalds 已提交
1530 1531 1532
	if (err)
		goto out_put;

1533
	err = sock->ops->connect(sock, (struct sockaddr *)&address, addrlen,
L
Linus Torvalds 已提交
1534 1535
				 sock->file->f_flags);
out_put:
1536
	fput_light(sock->file, fput_needed);
L
Linus Torvalds 已提交
1537 1538 1539 1540 1541 1542 1543 1544 1545
out:
	return err;
}

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

1546 1547
SYSCALL_DEFINE3(getsockname, int, fd, struct sockaddr __user *, usockaddr,
		int __user *, usockaddr_len)
L
Linus Torvalds 已提交
1548 1549
{
	struct socket *sock;
1550
	struct sockaddr_storage address;
1551
	int len, err, fput_needed;
1552

1553
	sock = sockfd_lookup_light(fd, &err, &fput_needed);
L
Linus Torvalds 已提交
1554 1555 1556 1557 1558 1559 1560
	if (!sock)
		goto out;

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

1561
	err = sock->ops->getname(sock, (struct sockaddr *)&address, &len, 0);
L
Linus Torvalds 已提交
1562 1563
	if (err)
		goto out_put;
1564
	err = move_addr_to_user(&address, len, usockaddr, usockaddr_len);
L
Linus Torvalds 已提交
1565 1566

out_put:
1567
	fput_light(sock->file, fput_needed);
L
Linus Torvalds 已提交
1568 1569 1570 1571 1572 1573 1574 1575 1576
out:
	return err;
}

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

1577 1578
SYSCALL_DEFINE3(getpeername, int, fd, struct sockaddr __user *, usockaddr,
		int __user *, usockaddr_len)
L
Linus Torvalds 已提交
1579 1580
{
	struct socket *sock;
1581
	struct sockaddr_storage address;
1582
	int len, err, fput_needed;
L
Linus Torvalds 已提交
1583

1584 1585
	sock = sockfd_lookup_light(fd, &err, &fput_needed);
	if (sock != NULL) {
L
Linus Torvalds 已提交
1586 1587
		err = security_socket_getpeername(sock);
		if (err) {
1588
			fput_light(sock->file, fput_needed);
L
Linus Torvalds 已提交
1589 1590 1591
			return err;
		}

1592
		err =
1593
		    sock->ops->getname(sock, (struct sockaddr *)&address, &len,
1594
				       1);
L
Linus Torvalds 已提交
1595
		if (!err)
1596
			err = move_addr_to_user(&address, len, usockaddr,
1597
						usockaddr_len);
1598
		fput_light(sock->file, fput_needed);
L
Linus Torvalds 已提交
1599 1600 1601 1602 1603 1604 1605 1606 1607 1608
	}
	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.
 */

1609
SYSCALL_DEFINE6(sendto, int, fd, void __user *, buff, size_t, len,
1610
		unsigned int, flags, struct sockaddr __user *, addr,
1611
		int, addr_len)
L
Linus Torvalds 已提交
1612 1613
{
	struct socket *sock;
1614
	struct sockaddr_storage address;
L
Linus Torvalds 已提交
1615 1616 1617
	int err;
	struct msghdr msg;
	struct iovec iov;
1618 1619
	int fput_needed;

1620 1621 1622
	err = import_single_range(WRITE, buff, len, &iov, &msg.msg_iter);
	if (unlikely(err))
		return err;
1623 1624
	sock = sockfd_lookup_light(fd, &err, &fput_needed);
	if (!sock)
1625
		goto out;
1626

1627 1628 1629 1630
	msg.msg_name = NULL;
	msg.msg_control = NULL;
	msg.msg_controllen = 0;
	msg.msg_namelen = 0;
1631
	if (addr) {
1632
		err = move_addr_to_kernel(addr, addr_len, &address);
L
Linus Torvalds 已提交
1633 1634
		if (err < 0)
			goto out_put;
1635
		msg.msg_name = (struct sockaddr *)&address;
1636
		msg.msg_namelen = addr_len;
L
Linus Torvalds 已提交
1637 1638 1639 1640
	}
	if (sock->file->f_flags & O_NONBLOCK)
		flags |= MSG_DONTWAIT;
	msg.msg_flags = flags;
1641
	err = sock_sendmsg(sock, &msg);
L
Linus Torvalds 已提交
1642

1643
out_put:
1644
	fput_light(sock->file, fput_needed);
1645
out:
L
Linus Torvalds 已提交
1646 1647 1648 1649
	return err;
}

/*
1650
 *	Send a datagram down a socket.
L
Linus Torvalds 已提交
1651 1652
 */

1653
SYSCALL_DEFINE4(send, int, fd, void __user *, buff, size_t, len,
1654
		unsigned int, flags)
L
Linus Torvalds 已提交
1655 1656 1657 1658 1659
{
	return sys_sendto(fd, buff, len, flags, NULL, 0);
}

/*
1660
 *	Receive a frame from the socket and optionally record the address of the
L
Linus Torvalds 已提交
1661 1662 1663 1664
 *	sender. We verify the buffers are writable and if needed move the
 *	sender address from kernel to user space.
 */

1665
SYSCALL_DEFINE6(recvfrom, int, fd, void __user *, ubuf, size_t, size,
1666
		unsigned int, flags, struct sockaddr __user *, addr,
1667
		int __user *, addr_len)
L
Linus Torvalds 已提交
1668 1669 1670 1671
{
	struct socket *sock;
	struct iovec iov;
	struct msghdr msg;
1672
	struct sockaddr_storage address;
1673
	int err, err2;
1674 1675
	int fput_needed;

1676 1677 1678
	err = import_single_range(READ, ubuf, size, &iov, &msg.msg_iter);
	if (unlikely(err))
		return err;
1679
	sock = sockfd_lookup_light(fd, &err, &fput_needed);
L
Linus Torvalds 已提交
1680
	if (!sock)
1681
		goto out;
L
Linus Torvalds 已提交
1682

1683 1684
	msg.msg_control = NULL;
	msg.msg_controllen = 0;
1685 1686 1687 1688
	/* 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;
1689
	msg.msg_iocb = NULL;
L
Linus Torvalds 已提交
1690 1691
	if (sock->file->f_flags & O_NONBLOCK)
		flags |= MSG_DONTWAIT;
1692
	err = sock_recvmsg(sock, &msg, flags);
L
Linus Torvalds 已提交
1693

1694
	if (err >= 0 && addr != NULL) {
1695
		err2 = move_addr_to_user(&address,
1696
					 msg.msg_namelen, addr, addr_len);
1697 1698
		if (err2 < 0)
			err = err2;
L
Linus Torvalds 已提交
1699
	}
1700 1701

	fput_light(sock->file, fput_needed);
1702
out:
L
Linus Torvalds 已提交
1703 1704 1705 1706
	return err;
}

/*
1707
 *	Receive a datagram from a socket.
L
Linus Torvalds 已提交
1708 1709
 */

1710 1711
SYSCALL_DEFINE4(recv, int, fd, void __user *, ubuf, size_t, size,
		unsigned int, flags)
L
Linus Torvalds 已提交
1712 1713 1714 1715 1716 1717 1718 1719 1720
{
	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.
 */

1721 1722
SYSCALL_DEFINE5(setsockopt, int, fd, int, level, int, optname,
		char __user *, optval, int, optlen)
L
Linus Torvalds 已提交
1723
{
1724
	int err, fput_needed;
L
Linus Torvalds 已提交
1725 1726 1727 1728
	struct socket *sock;

	if (optlen < 0)
		return -EINVAL;
1729 1730 1731 1732

	sock = sockfd_lookup_light(fd, &err, &fput_needed);
	if (sock != NULL) {
		err = security_socket_setsockopt(sock, level, optname);
1733 1734
		if (err)
			goto out_put;
L
Linus Torvalds 已提交
1735 1736

		if (level == SOL_SOCKET)
1737 1738 1739
			err =
			    sock_setsockopt(sock, level, optname, optval,
					    optlen);
L
Linus Torvalds 已提交
1740
		else
1741 1742 1743
			err =
			    sock->ops->setsockopt(sock, level, optname, optval,
						  optlen);
1744 1745
out_put:
		fput_light(sock->file, fput_needed);
L
Linus Torvalds 已提交
1746 1747 1748 1749 1750 1751 1752 1753 1754
	}
	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.
 */

1755 1756
SYSCALL_DEFINE5(getsockopt, int, fd, int, level, int, optname,
		char __user *, optval, int __user *, optlen)
L
Linus Torvalds 已提交
1757
{
1758
	int err, fput_needed;
L
Linus Torvalds 已提交
1759 1760
	struct socket *sock;

1761 1762
	sock = sockfd_lookup_light(fd, &err, &fput_needed);
	if (sock != NULL) {
1763 1764 1765
		err = security_socket_getsockopt(sock, level, optname);
		if (err)
			goto out_put;
L
Linus Torvalds 已提交
1766 1767

		if (level == SOL_SOCKET)
1768 1769 1770
			err =
			    sock_getsockopt(sock, level, optname, optval,
					    optlen);
L
Linus Torvalds 已提交
1771
		else
1772 1773 1774
			err =
			    sock->ops->getsockopt(sock, level, optname, optval,
						  optlen);
1775 1776
out_put:
		fput_light(sock->file, fput_needed);
L
Linus Torvalds 已提交
1777 1778 1779 1780 1781 1782 1783 1784
	}
	return err;
}

/*
 *	Shutdown a socket.
 */

1785
SYSCALL_DEFINE2(shutdown, int, fd, int, how)
L
Linus Torvalds 已提交
1786
{
1787
	int err, fput_needed;
L
Linus Torvalds 已提交
1788 1789
	struct socket *sock;

1790 1791
	sock = sockfd_lookup_light(fd, &err, &fput_needed);
	if (sock != NULL) {
L
Linus Torvalds 已提交
1792
		err = security_socket_shutdown(sock, how);
1793 1794 1795
		if (!err)
			err = sock->ops->shutdown(sock, how);
		fput_light(sock->file, fput_needed);
L
Linus Torvalds 已提交
1796 1797 1798 1799
	}
	return err;
}

1800
/* A couple of helpful macros for getting the address of the 32/64 bit
L
Linus Torvalds 已提交
1801 1802 1803 1804 1805 1806
 * 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)

1807 1808 1809 1810 1811
struct used_address {
	struct sockaddr_storage name;
	unsigned int name_len;
};

1812 1813 1814 1815
static int copy_msghdr_from_user(struct msghdr *kmsg,
				 struct user_msghdr __user *umsg,
				 struct sockaddr __user **save_addr,
				 struct iovec **iov)
1816
{
1817 1818
	struct sockaddr __user *uaddr;
	struct iovec __user *uiov;
A
Al Viro 已提交
1819
	size_t nr_segs;
1820 1821 1822 1823 1824 1825
	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
1826
	    __get_user(nr_segs, &umsg->msg_iovlen) ||
1827 1828 1829
	    __get_user(kmsg->msg_control, &umsg->msg_control) ||
	    __get_user(kmsg->msg_controllen, &umsg->msg_controllen) ||
	    __get_user(kmsg->msg_flags, &umsg->msg_flags))
1830
		return -EFAULT;
1831

1832
	if (!uaddr)
1833 1834
		kmsg->msg_namelen = 0;

1835 1836 1837
	if (kmsg->msg_namelen < 0)
		return -EINVAL;

1838
	if (kmsg->msg_namelen > sizeof(struct sockaddr_storage))
1839
		kmsg->msg_namelen = sizeof(struct sockaddr_storage);
1840 1841 1842 1843 1844 1845 1846 1847 1848 1849 1850 1851 1852 1853 1854 1855

	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
1856
	if (nr_segs > UIO_MAXIOV)
1857 1858
		return -EMSGSIZE;

1859 1860
	kmsg->msg_iocb = NULL;

1861 1862
	return import_iovec(save_addr ? READ : WRITE, uiov, nr_segs,
			    UIO_FASTIOV, iov, &kmsg->msg_iter);
1863 1864
}

1865
static int ___sys_sendmsg(struct socket *sock, struct user_msghdr __user *msg,
1866
			 struct msghdr *msg_sys, unsigned int flags,
1867 1868
			 struct used_address *used_address,
			 unsigned int allowed_msghdr_flags)
L
Linus Torvalds 已提交
1869
{
1870 1871
	struct compat_msghdr __user *msg_compat =
	    (struct compat_msghdr __user *)msg;
1872
	struct sockaddr_storage address;
L
Linus Torvalds 已提交
1873
	struct iovec iovstack[UIO_FASTIOV], *iov = iovstack;
1874
	unsigned char ctl[sizeof(struct cmsghdr) + 20]
1875 1876
	    __attribute__ ((aligned(sizeof(__kernel_size_t))));
	/* 20 is size of ipv6_pktinfo */
L
Linus Torvalds 已提交
1877
	unsigned char *ctl_buf = ctl;
1878
	int ctl_len;
1879
	ssize_t err;
1880

1881
	msg_sys->msg_name = &address;
L
Linus Torvalds 已提交
1882

1883
	if (MSG_CMSG_COMPAT & flags)
1884
		err = get_compat_msghdr(msg_sys, msg_compat, NULL, &iov);
1885
	else
1886
		err = copy_msghdr_from_user(msg_sys, msg, NULL, &iov);
1887
	if (err < 0)
1888
		return err;
L
Linus Torvalds 已提交
1889 1890 1891

	err = -ENOBUFS;

1892
	if (msg_sys->msg_controllen > INT_MAX)
L
Linus Torvalds 已提交
1893
		goto out_freeiov;
1894
	flags |= (msg_sys->msg_flags & allowed_msghdr_flags);
1895
	ctl_len = msg_sys->msg_controllen;
L
Linus Torvalds 已提交
1896
	if ((MSG_CMSG_COMPAT & flags) && ctl_len) {
1897
		err =
1898
		    cmsghdr_from_user_compat_to_kern(msg_sys, sock->sk, ctl,
1899
						     sizeof(ctl));
L
Linus Torvalds 已提交
1900 1901
		if (err)
			goto out_freeiov;
1902 1903
		ctl_buf = msg_sys->msg_control;
		ctl_len = msg_sys->msg_controllen;
L
Linus Torvalds 已提交
1904
	} else if (ctl_len) {
1905
		if (ctl_len > sizeof(ctl)) {
L
Linus Torvalds 已提交
1906
			ctl_buf = sock_kmalloc(sock->sk, ctl_len, GFP_KERNEL);
1907
			if (ctl_buf == NULL)
L
Linus Torvalds 已提交
1908 1909 1910 1911
				goto out_freeiov;
		}
		err = -EFAULT;
		/*
1912
		 * Careful! Before this, msg_sys->msg_control contains a user pointer.
L
Linus Torvalds 已提交
1913 1914 1915
		 * Afterwards, it will be a kernel pointer. Thus the compiler-assisted
		 * checking falls down on this.
		 */
1916
		if (copy_from_user(ctl_buf,
1917
				   (void __user __force *)msg_sys->msg_control,
1918
				   ctl_len))
L
Linus Torvalds 已提交
1919
			goto out_freectl;
1920
		msg_sys->msg_control = ctl_buf;
L
Linus Torvalds 已提交
1921
	}
1922
	msg_sys->msg_flags = flags;
L
Linus Torvalds 已提交
1923 1924

	if (sock->file->f_flags & O_NONBLOCK)
1925
		msg_sys->msg_flags |= MSG_DONTWAIT;
1926 1927 1928 1929 1930 1931
	/*
	 * 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.
	 */
1932 1933 1934
	if (used_address && msg_sys->msg_name &&
	    used_address->name_len == msg_sys->msg_namelen &&
	    !memcmp(&used_address->name, msg_sys->msg_name,
1935
		    used_address->name_len)) {
1936
		err = sock_sendmsg_nosec(sock, msg_sys);
1937 1938
		goto out_freectl;
	}
1939
	err = sock_sendmsg(sock, msg_sys);
1940 1941 1942 1943 1944 1945
	/*
	 * 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;
1946 1947 1948
		if (msg_sys->msg_name)
			memcpy(&used_address->name, msg_sys->msg_name,
			       used_address->name_len);
1949
	}
L
Linus Torvalds 已提交
1950 1951

out_freectl:
1952
	if (ctl_buf != ctl)
L
Linus Torvalds 已提交
1953 1954
		sock_kfree_s(sock->sk, ctl_buf, ctl_len);
out_freeiov:
1955
	kfree(iov);
1956 1957 1958 1959 1960 1961 1962
	return err;
}

/*
 *	BSD sendmsg interface
 */

1963
long __sys_sendmsg(int fd, struct user_msghdr __user *msg, unsigned flags)
1964 1965 1966
{
	int fput_needed, err;
	struct msghdr msg_sys;
1967 1968 1969
	struct socket *sock;

	sock = sockfd_lookup_light(fd, &err, &fput_needed);
1970 1971 1972
	if (!sock)
		goto out;

1973
	err = ___sys_sendmsg(sock, msg, &msg_sys, flags, NULL, 0);
1974

1975
	fput_light(sock->file, fput_needed);
1976
out:
L
Linus Torvalds 已提交
1977 1978 1979
	return err;
}

1980
SYSCALL_DEFINE3(sendmsg, int, fd, struct user_msghdr __user *, msg, unsigned int, flags)
1981 1982 1983 1984 1985 1986
{
	if (flags & MSG_CMSG_COMPAT)
		return -EINVAL;
	return __sys_sendmsg(fd, msg, flags);
}

1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998
/*
 *	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;
1999
	struct used_address used_address;
T
Tom Herbert 已提交
2000
	unsigned int oflags = flags;
2001

2002 2003
	if (vlen > UIO_MAXIOV)
		vlen = UIO_MAXIOV;
2004 2005 2006 2007 2008 2009 2010

	datagrams = 0;

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

2011
	used_address.name_len = UINT_MAX;
2012 2013
	entry = mmsg;
	compat_entry = (struct compat_mmsghdr __user *)mmsg;
2014
	err = 0;
T
Tom Herbert 已提交
2015
	flags |= MSG_BATCH;
2016 2017

	while (datagrams < vlen) {
T
Tom Herbert 已提交
2018 2019 2020
		if (datagrams == vlen - 1)
			flags = oflags;

2021
		if (MSG_CMSG_COMPAT & flags) {
2022
			err = ___sys_sendmsg(sock, (struct user_msghdr __user *)compat_entry,
2023
					     &msg_sys, flags, &used_address, MSG_EOR);
2024 2025 2026 2027 2028
			if (err < 0)
				break;
			err = __put_user(err, &compat_entry->msg_len);
			++compat_entry;
		} else {
2029
			err = ___sys_sendmsg(sock,
2030
					     (struct user_msghdr __user *)entry,
2031
					     &msg_sys, flags, &used_address, MSG_EOR);
2032 2033 2034 2035 2036 2037 2038 2039 2040
			if (err < 0)
				break;
			err = put_user(err, &entry->msg_len);
			++entry;
		}

		if (err)
			break;
		++datagrams;
2041
		cond_resched();
2042 2043 2044 2045
	}

	fput_light(sock->file, fput_needed);

2046 2047
	/* We only return an error if no datagrams were able to be sent */
	if (datagrams != 0)
2048 2049 2050 2051 2052 2053 2054 2055
		return datagrams;

	return err;
}

SYSCALL_DEFINE4(sendmmsg, int, fd, struct mmsghdr __user *, mmsg,
		unsigned int, vlen, unsigned int, flags)
{
2056 2057
	if (flags & MSG_CMSG_COMPAT)
		return -EINVAL;
2058 2059 2060
	return __sys_sendmmsg(fd, mmsg, vlen, flags);
}

2061
static int ___sys_recvmsg(struct socket *sock, struct user_msghdr __user *msg,
2062
			 struct msghdr *msg_sys, unsigned int flags, int nosec)
L
Linus Torvalds 已提交
2063
{
2064 2065
	struct compat_msghdr __user *msg_compat =
	    (struct compat_msghdr __user *)msg;
L
Linus Torvalds 已提交
2066
	struct iovec iovstack[UIO_FASTIOV];
2067
	struct iovec *iov = iovstack;
L
Linus Torvalds 已提交
2068
	unsigned long cmsg_ptr;
2069
	int len;
2070
	ssize_t err;
L
Linus Torvalds 已提交
2071 2072

	/* kernel mode address */
2073
	struct sockaddr_storage addr;
L
Linus Torvalds 已提交
2074 2075 2076

	/* user mode address pointers */
	struct sockaddr __user *uaddr;
2077
	int __user *uaddr_len = COMPAT_NAMELEN(msg);
2078

2079
	msg_sys->msg_name = &addr;
L
Linus Torvalds 已提交
2080

2081
	if (MSG_CMSG_COMPAT & flags)
2082
		err = get_compat_msghdr(msg_sys, msg_compat, &uaddr, &iov);
2083
	else
2084
		err = copy_msghdr_from_user(msg_sys, msg, &uaddr, &iov);
L
Linus Torvalds 已提交
2085
	if (err < 0)
2086
		return err;
L
Linus Torvalds 已提交
2087

2088 2089
	cmsg_ptr = (unsigned long)msg_sys->msg_control;
	msg_sys->msg_flags = flags & (MSG_CMSG_CLOEXEC|MSG_CMSG_COMPAT);
2090

2091 2092 2093
	/* We assume all kernel code knows the size of sockaddr_storage */
	msg_sys->msg_namelen = 0;

L
Linus Torvalds 已提交
2094 2095
	if (sock->file->f_flags & O_NONBLOCK)
		flags |= MSG_DONTWAIT;
2096
	err = (nosec ? sock_recvmsg_nosec : sock_recvmsg)(sock, msg_sys, flags);
L
Linus Torvalds 已提交
2097 2098 2099 2100 2101
	if (err < 0)
		goto out_freeiov;
	len = err;

	if (uaddr != NULL) {
2102
		err = move_addr_to_user(&addr,
2103
					msg_sys->msg_namelen, uaddr,
2104
					uaddr_len);
L
Linus Torvalds 已提交
2105 2106 2107
		if (err < 0)
			goto out_freeiov;
	}
2108
	err = __put_user((msg_sys->msg_flags & ~MSG_CMSG_COMPAT),
2109
			 COMPAT_FLAGS(msg));
L
Linus Torvalds 已提交
2110 2111 2112
	if (err)
		goto out_freeiov;
	if (MSG_CMSG_COMPAT & flags)
2113
		err = __put_user((unsigned long)msg_sys->msg_control - cmsg_ptr,
L
Linus Torvalds 已提交
2114 2115
				 &msg_compat->msg_controllen);
	else
2116
		err = __put_user((unsigned long)msg_sys->msg_control - cmsg_ptr,
L
Linus Torvalds 已提交
2117 2118 2119 2120 2121 2122
				 &msg->msg_controllen);
	if (err)
		goto out_freeiov;
	err = len;

out_freeiov:
2123
	kfree(iov);
2124 2125 2126 2127 2128 2129 2130
	return err;
}

/*
 *	BSD recvmsg interface
 */

2131
long __sys_recvmsg(int fd, struct user_msghdr __user *msg, unsigned flags)
2132 2133 2134
{
	int fput_needed, err;
	struct msghdr msg_sys;
2135 2136 2137
	struct socket *sock;

	sock = sockfd_lookup_light(fd, &err, &fput_needed);
2138 2139 2140
	if (!sock)
		goto out;

2141
	err = ___sys_recvmsg(sock, msg, &msg_sys, flags, 0);
2142

2143
	fput_light(sock->file, fput_needed);
L
Linus Torvalds 已提交
2144 2145 2146 2147
out:
	return err;
}

2148
SYSCALL_DEFINE3(recvmsg, int, fd, struct user_msghdr __user *, msg,
2149 2150 2151 2152 2153 2154 2155
		unsigned int, flags)
{
	if (flags & MSG_CMSG_COMPAT)
		return -EINVAL;
	return __sys_recvmsg(fd, msg, flags);
}

2156 2157 2158 2159 2160 2161 2162 2163 2164 2165
/*
 *     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;
2166
	struct compat_mmsghdr __user *compat_entry;
2167
	struct msghdr msg_sys;
2168 2169
	struct timespec64 end_time;
	struct timespec64 timeout64;
2170 2171 2172 2173 2174 2175 2176 2177 2178 2179 2180 2181 2182 2183 2184 2185 2186

	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;
2187
	compat_entry = (struct compat_mmsghdr __user *)mmsg;
2188 2189 2190 2191 2192

	while (datagrams < vlen) {
		/*
		 * No need to ask LSM for more than the first datagram.
		 */
2193
		if (MSG_CMSG_COMPAT & flags) {
2194
			err = ___sys_recvmsg(sock, (struct user_msghdr __user *)compat_entry,
2195 2196
					     &msg_sys, flags & ~MSG_WAITFORONE,
					     datagrams);
2197 2198 2199 2200 2201
			if (err < 0)
				break;
			err = __put_user(err, &compat_entry->msg_len);
			++compat_entry;
		} else {
2202
			err = ___sys_recvmsg(sock,
2203
					     (struct user_msghdr __user *)entry,
2204 2205
					     &msg_sys, flags & ~MSG_WAITFORONE,
					     datagrams);
2206 2207 2208 2209 2210 2211
			if (err < 0)
				break;
			err = put_user(err, &entry->msg_len);
			++entry;
		}

2212 2213 2214 2215
		if (err)
			break;
		++datagrams;

2216 2217 2218 2219
		/* MSG_WAITFORONE turns on MSG_DONTWAIT after one packet */
		if (flags & MSG_WAITFORONE)
			flags |= MSG_DONTWAIT;

2220
		if (timeout) {
2221 2222 2223
			ktime_get_ts64(&timeout64);
			*timeout = timespec64_to_timespec(
					timespec64_sub(end_time, timeout64));
2224 2225 2226 2227 2228 2229 2230 2231 2232 2233 2234 2235 2236
			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;
2237
		cond_resched();
2238 2239 2240
	}

	if (err == 0)
2241 2242 2243 2244 2245 2246
		goto out_put;

	if (datagrams == 0) {
		datagrams = err;
		goto out_put;
	}
2247

2248 2249 2250 2251 2252
	/*
	 * We may return less entries than requested (vlen) if the
	 * sock is non block and there aren't enough datagrams...
	 */
	if (err != -EAGAIN) {
2253
		/*
2254 2255 2256 2257
		 * ... 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).
2258
		 */
2259
		sock->sk->sk_err = -err;
2260
	}
2261 2262
out_put:
	fput_light(sock->file, fput_needed);
2263

2264
	return datagrams;
2265 2266 2267 2268 2269 2270 2271 2272 2273
}

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;

2274 2275 2276
	if (flags & MSG_CMSG_COMPAT)
		return -EINVAL;

2277 2278 2279 2280 2281 2282 2283 2284 2285 2286 2287 2288 2289 2290 2291 2292
	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
2293 2294
/* Argument list sizes for sys_socketcall */
#define AL(x) ((x) * sizeof(unsigned long))
2295
static const unsigned char nargs[21] = {
2296 2297 2298
	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),
2299
	AL(4), AL(5), AL(4)
2300 2301
};

L
Linus Torvalds 已提交
2302 2303 2304
#undef AL

/*
2305
 *	System call vectors.
L
Linus Torvalds 已提交
2306 2307 2308
 *
 *	Argument checking cleaned up. Saved 20% in size.
 *  This function doesn't need to set the kernel lock because
2309
 *  it is set by the callees.
L
Linus Torvalds 已提交
2310 2311
 */

2312
SYSCALL_DEFINE2(socketcall, int, call, unsigned long __user *, args)
L
Linus Torvalds 已提交
2313
{
2314
	unsigned long a[AUDITSC_ARGS];
2315
	unsigned long a0, a1;
L
Linus Torvalds 已提交
2316
	int err;
2317
	unsigned int len;
L
Linus Torvalds 已提交
2318

2319
	if (call < 1 || call > SYS_SENDMMSG)
L
Linus Torvalds 已提交
2320 2321
		return -EINVAL;

2322 2323 2324 2325
	len = nargs[call];
	if (len > sizeof(a))
		return -EINVAL;

L
Linus Torvalds 已提交
2326
	/* copy_from_user should be SMP safe. */
2327
	if (copy_from_user(a, args, len))
L
Linus Torvalds 已提交
2328
		return -EFAULT;
2329

2330 2331 2332
	err = audit_socketcall(nargs[call] / sizeof(unsigned long), a);
	if (err)
		return err;
2333

2334 2335 2336 2337 2338 2339 2340 2341 2342 2343 2344 2345 2346 2347 2348 2349 2350
	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
2351 2352
		err = sys_accept4(a0, (struct sockaddr __user *)a1,
				  (int __user *)a[2], 0);
2353 2354 2355 2356 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
		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:
2394
		err = sys_sendmsg(a0, (struct user_msghdr __user *)a1, a[2]);
2395
		break;
2396 2397 2398
	case SYS_SENDMMSG:
		err = sys_sendmmsg(a0, (struct mmsghdr __user *)a1, a[2], a[3]);
		break;
2399
	case SYS_RECVMSG:
2400
		err = sys_recvmsg(a0, (struct user_msghdr __user *)a1, a[2]);
2401
		break;
2402 2403 2404 2405
	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
2406 2407 2408
	case SYS_ACCEPT4:
		err = sys_accept4(a0, (struct sockaddr __user *)a1,
				  (int __user *)a[2], a[3]);
U
Ulrich Drepper 已提交
2409
		break;
2410 2411 2412
	default:
		err = -EINVAL;
		break;
L
Linus Torvalds 已提交
2413 2414 2415 2416
	}
	return err;
}

2417
#endif				/* __ARCH_WANT_SYS_SOCKETCALL */
L
Linus Torvalds 已提交
2418

2419 2420 2421 2422
/**
 *	sock_register - add a socket protocol handler
 *	@ops: description of protocol
 *
L
Linus Torvalds 已提交
2423 2424
 *	This function is called by a protocol handler that wants to
 *	advertise its address family, and have it linked into the
2425
 *	socket interface. The value ops->family corresponds to the
2426
 *	socket system call protocol family.
L
Linus Torvalds 已提交
2427
 */
2428
int sock_register(const struct net_proto_family *ops)
L
Linus Torvalds 已提交
2429 2430 2431 2432
{
	int err;

	if (ops->family >= NPROTO) {
2433
		pr_crit("protocol %d >= NPROTO(%d)\n", ops->family, NPROTO);
L
Linus Torvalds 已提交
2434 2435
		return -ENOBUFS;
	}
2436 2437

	spin_lock(&net_family_lock);
E
Eric Dumazet 已提交
2438 2439
	if (rcu_dereference_protected(net_families[ops->family],
				      lockdep_is_held(&net_family_lock)))
2440 2441
		err = -EEXIST;
	else {
2442
		rcu_assign_pointer(net_families[ops->family], ops);
L
Linus Torvalds 已提交
2443 2444
		err = 0;
	}
2445 2446
	spin_unlock(&net_family_lock);

2447
	pr_info("NET: Registered protocol family %d\n", ops->family);
L
Linus Torvalds 已提交
2448 2449
	return err;
}
2450
EXPORT_SYMBOL(sock_register);
L
Linus Torvalds 已提交
2451

2452 2453 2454 2455
/**
 *	sock_unregister - remove a protocol handler
 *	@family: protocol family to remove
 *
L
Linus Torvalds 已提交
2456 2457
 *	This function is called by a protocol handler that wants to
 *	remove its address family, and have it unlinked from the
2458 2459 2460 2461 2462 2463
 *	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
2464
 */
2465
void sock_unregister(int family)
L
Linus Torvalds 已提交
2466
{
2467
	BUG_ON(family < 0 || family >= NPROTO);
L
Linus Torvalds 已提交
2468

2469
	spin_lock(&net_family_lock);
2470
	RCU_INIT_POINTER(net_families[family], NULL);
2471 2472 2473 2474
	spin_unlock(&net_family_lock);

	synchronize_rcu();

2475
	pr_info("NET: Unregistered protocol family %d\n", family);
L
Linus Torvalds 已提交
2476
}
2477
EXPORT_SYMBOL(sock_unregister);
L
Linus Torvalds 已提交
2478

2479
static int __init sock_init(void)
L
Linus Torvalds 已提交
2480
{
N
Nick Piggin 已提交
2481
	int err;
2482 2483 2484 2485 2486 2487
	/*
	 *      Initialize the network sysctl infrastructure.
	 */
	err = net_sysctl_init();
	if (err)
		goto out;
N
Nick Piggin 已提交
2488

L
Linus Torvalds 已提交
2489
	/*
2490
	 *      Initialize skbuff SLAB cache
L
Linus Torvalds 已提交
2491 2492 2493 2494
	 */
	skb_init();

	/*
2495
	 *      Initialize the protocols module.
L
Linus Torvalds 已提交
2496 2497 2498
	 */

	init_inodecache();
N
Nick Piggin 已提交
2499 2500 2501 2502

	err = register_filesystem(&sock_fs_type);
	if (err)
		goto out_fs;
L
Linus Torvalds 已提交
2503
	sock_mnt = kern_mount(&sock_fs_type);
N
Nick Piggin 已提交
2504 2505 2506 2507
	if (IS_ERR(sock_mnt)) {
		err = PTR_ERR(sock_mnt);
		goto out_mount;
	}
2508 2509

	/* The real protocol initialization is performed in later initcalls.
L
Linus Torvalds 已提交
2510 2511 2512
	 */

#ifdef CONFIG_NETFILTER
2513 2514 2515
	err = netfilter_init();
	if (err)
		goto out;
L
Linus Torvalds 已提交
2516
#endif
2517

2518
	ptp_classifier_init();
2519

N
Nick Piggin 已提交
2520 2521 2522 2523 2524 2525 2526
out:
	return err;

out_mount:
	unregister_filesystem(&sock_fs_type);
out_fs:
	goto out;
L
Linus Torvalds 已提交
2527 2528
}

2529 2530
core_initcall(sock_init);	/* early initcall */

L
Linus Torvalds 已提交
2531 2532 2533 2534 2535 2536
#ifdef CONFIG_PROC_FS
void socket_seq_show(struct seq_file *seq)
{
	int cpu;
	int counter = 0;

2537
	for_each_possible_cpu(cpu)
2538
	    counter += per_cpu(sockets_in_use, cpu);
L
Linus Torvalds 已提交
2539 2540 2541 2542 2543 2544 2545

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

	seq_printf(seq, "sockets: used %d\n", counter);
}
2546
#endif				/* CONFIG_PROC_FS */
L
Linus Torvalds 已提交
2547

2548
#ifdef CONFIG_COMPAT
2549
static int do_siocgstamp(struct net *net, struct socket *sock,
2550
			 unsigned int cmd, void __user *up)
2551 2552 2553 2554 2555 2556
{
	mm_segment_t old_fs = get_fs();
	struct timeval ktv;
	int err;

	set_fs(KERNEL_DS);
2557
	err = sock_do_ioctl(net, sock, cmd, (unsigned long)&ktv);
2558
	set_fs(old_fs);
2559
	if (!err)
2560
		err = compat_put_timeval(&ktv, up);
2561

2562 2563 2564
	return err;
}

2565
static int do_siocgstampns(struct net *net, struct socket *sock,
2566
			   unsigned int cmd, void __user *up)
2567 2568 2569 2570 2571 2572
{
	mm_segment_t old_fs = get_fs();
	struct timespec kts;
	int err;

	set_fs(KERNEL_DS);
2573
	err = sock_do_ioctl(net, sock, cmd, (unsigned long)&kts);
2574
	set_fs(old_fs);
2575
	if (!err)
2576
		err = compat_put_timespec(&kts, up);
2577

2578 2579 2580
	return err;
}

2581
static int dev_ifname32(struct net *net, struct compat_ifreq __user *uifr32)
2582 2583 2584 2585 2586
{
	struct ifreq __user *uifr;
	int err;

	uifr = compat_alloc_user_space(sizeof(struct ifreq));
2587
	if (copy_in_user(uifr, uifr32, sizeof(struct compat_ifreq)))
2588 2589
		return -EFAULT;

2590
	err = dev_ioctl(net, SIOCGIFNAME, uifr);
2591 2592 2593
	if (err)
		return err;

2594
	if (copy_in_user(uifr32, uifr, sizeof(struct compat_ifreq)))
2595 2596 2597 2598 2599
		return -EFAULT;

	return 0;
}

2600
static int dev_ifconf(struct net *net, struct compat_ifconf __user *uifc32)
2601
{
2602
	struct compat_ifconf ifc32;
2603 2604
	struct ifconf ifc;
	struct ifconf __user *uifc;
2605
	struct compat_ifreq __user *ifr32;
2606 2607 2608 2609
	struct ifreq __user *ifr;
	unsigned int i, j;
	int err;

2610
	if (copy_from_user(&ifc32, uifc32, sizeof(struct compat_ifconf)))
2611 2612
		return -EFAULT;

2613
	memset(&ifc, 0, sizeof(ifc));
2614 2615 2616 2617 2618 2619
	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 {
2620 2621
		size_t len = ((ifc32.ifc_len / sizeof(struct compat_ifreq)) + 1) *
			sizeof(struct ifreq);
2622 2623 2624 2625
		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);
2626
		for (i = 0; i < ifc32.ifc_len; i += sizeof(struct compat_ifreq)) {
2627
			if (copy_in_user(ifr, ifr32, sizeof(struct compat_ifreq)))
2628 2629 2630 2631 2632 2633 2634 2635
				return -EFAULT;
			ifr++;
			ifr32++;
		}
	}
	if (copy_to_user(uifc, &ifc, sizeof(struct ifconf)))
		return -EFAULT;

2636
	err = dev_ioctl(net, SIOCGIFCONF, uifc);
2637 2638 2639 2640 2641 2642 2643 2644 2645
	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;
2646 2647 2648
	     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)))
2649 2650 2651 2652 2653 2654 2655 2656 2657 2658
			return -EFAULT;
		ifr32++;
		ifr++;
	}

	if (ifc32.ifcbuf == 0) {
		/* Translate from 64-bit structure multiple to
		 * a 32-bit one.
		 */
		i = ifc.ifc_len;
2659
		i = ((i / sizeof(struct ifreq)) * sizeof(struct compat_ifreq));
2660 2661 2662 2663
		ifc32.ifc_len = i;
	} else {
		ifc32.ifc_len = i;
	}
2664
	if (copy_to_user(uifc32, &ifc32, sizeof(struct compat_ifconf)))
2665 2666 2667 2668 2669
		return -EFAULT;

	return 0;
}

2670
static int ethtool_ioctl(struct net *net, struct compat_ifreq __user *ifr32)
2671
{
2672 2673 2674 2675
	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;
2676
	struct ifreq __user *ifr;
2677 2678
	u32 rule_cnt = 0, actual_rule_cnt;
	u32 ethcmd;
2679
	u32 data;
2680
	int ret;
2681

2682 2683
	if (get_user(data, &ifr32->ifr_ifru.ifru_data))
		return -EFAULT;
2684

2685 2686 2687
	compat_rxnfc = compat_ptr(data);

	if (get_user(ethcmd, &compat_rxnfc->cmd))
2688 2689
		return -EFAULT;

2690 2691 2692 2693 2694 2695 2696 2697 2698 2699 2700 2701 2702 2703 2704 2705 2706
	/* 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:
2707
	case ETHTOOL_SRXCLSRLINS:
2708 2709 2710 2711 2712 2713 2714 2715 2716
		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
2717
	rxnfc = (void __user *)ifr + ALIGN(sizeof(struct ifreq), 8);
2718 2719

	if (copy_in_user(&ifr->ifr_name, &ifr32->ifr_name, IFNAMSIZ))
2720 2721
		return -EFAULT;

2722 2723
	if (put_user(convert_in ? rxnfc : compat_ptr(data),
		     &ifr->ifr_ifru.ifru_data))
2724 2725
		return -EFAULT;

2726
	if (convert_in) {
2727
		/* We expect there to be holes between fs.m_ext and
2728 2729
		 * fs.ring_cookie and at the end of fs, but nowhere else.
		 */
2730 2731 2732 2733
		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));
2734 2735 2736 2737 2738 2739 2740
		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
2741 2742
				 (void __user *)(&rxnfc->fs.m_ext + 1) -
				 (void __user *)rxnfc) ||
2743 2744
		    copy_in_user(&rxnfc->fs.ring_cookie,
				 &compat_rxnfc->fs.ring_cookie,
S
Stephen Hemminger 已提交
2745 2746
				 (void __user *)(&rxnfc->fs.location + 1) -
				 (void __user *)&rxnfc->fs.ring_cookie) ||
2747 2748 2749 2750 2751 2752 2753 2754 2755 2756 2757
		    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
2758 2759
				 (const void __user *)(&rxnfc->fs.m_ext + 1) -
				 (const void __user *)rxnfc) ||
2760 2761
		    copy_in_user(&compat_rxnfc->fs.ring_cookie,
				 &rxnfc->fs.ring_cookie,
S
Stephen Hemminger 已提交
2762 2763
				 (const void __user *)(&rxnfc->fs.location + 1) -
				 (const void __user *)&rxnfc->fs.ring_cookie) ||
2764 2765 2766 2767 2768 2769 2770 2771 2772 2773 2774 2775 2776 2777 2778 2779 2780 2781 2782 2783 2784 2785 2786 2787 2788
		    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;
2789 2790
}

2791 2792 2793 2794 2795 2796
static int compat_siocwandev(struct net *net, struct compat_ifreq __user *uifr32)
{
	void __user *uptr;
	compat_uptr_t uptr32;
	struct ifreq __user *uifr;

2797
	uifr = compat_alloc_user_space(sizeof(*uifr));
2798 2799 2800 2801 2802 2803 2804 2805 2806 2807 2808 2809 2810 2811
	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);
}

2812 2813
static int bond_ioctl(struct net *net, unsigned int cmd,
			 struct compat_ifreq __user *ifr32)
2814 2815 2816 2817 2818 2819 2820 2821 2822 2823
{
	struct ifreq kifr;
	mm_segment_t old_fs;
	int err;

	switch (cmd) {
	case SIOCBONDENSLAVE:
	case SIOCBONDRELEASE:
	case SIOCBONDSETHWADDR:
	case SIOCBONDCHANGEACTIVE:
2824
		if (copy_from_user(&kifr, ifr32, sizeof(struct compat_ifreq)))
2825 2826 2827
			return -EFAULT;

		old_fs = get_fs();
2828
		set_fs(KERNEL_DS);
2829 2830
		err = dev_ioctl(net, cmd,
				(struct ifreq __user __force *) &kifr);
2831
		set_fs(old_fs);
2832 2833 2834

		return err;
	default:
2835
		return -ENOIOCTLCMD;
2836
	}
2837 2838
}

2839 2840
/* 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,
2841
				 struct compat_ifreq __user *u_ifreq32)
2842 2843 2844 2845 2846 2847 2848 2849 2850
{
	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;
2851
	if (get_user(data32, &u_ifreq32->ifr_ifru.ifru_data))
2852 2853 2854 2855 2856 2857 2858 2859
		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;
2860
	if (put_user(data64, &u_ifreq64->ifr_ifru.ifru_data))
2861 2862
		return -EFAULT;

2863
	return dev_ioctl(net, cmd, u_ifreq64);
2864 2865
}

2866 2867
static int dev_ifsioc(struct net *net, struct socket *sock,
			 unsigned int cmd, struct compat_ifreq __user *uifr32)
2868
{
2869
	struct ifreq __user *uifr;
2870 2871
	int err;

2872 2873 2874 2875 2876 2877
	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);

2878 2879 2880 2881 2882 2883 2884 2885 2886 2887 2888 2889
	if (!err) {
		switch (cmd) {
		case SIOCGIFFLAGS:
		case SIOCGIFMETRIC:
		case SIOCGIFMTU:
		case SIOCGIFMEM:
		case SIOCGIFHWADDR:
		case SIOCGIFINDEX:
		case SIOCGIFADDR:
		case SIOCGIFBRDADDR:
		case SIOCGIFDSTADDR:
		case SIOCGIFNETMASK:
2890
		case SIOCGIFPFLAGS:
2891
		case SIOCGIFTXQLEN:
2892 2893
		case SIOCGMIIPHY:
		case SIOCGMIIREG:
2894
			if (copy_in_user(uifr32, uifr, sizeof(*uifr32)))
2895 2896 2897 2898 2899 2900 2901
				err = -EFAULT;
			break;
		}
	}
	return err;
}

2902 2903 2904 2905 2906 2907 2908 2909 2910 2911
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));
2912 2913 2914 2915 2916 2917
	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);
2918 2919 2920 2921
	if (err)
		return -EFAULT;

	old_fs = get_fs();
2922
	set_fs(KERNEL_DS);
2923
	err = dev_ioctl(net, cmd, (void  __user __force *)&ifr);
2924
	set_fs(old_fs);
2925 2926 2927

	if (cmd == SIOCGIFMAP && !err) {
		err = copy_to_user(uifr32, &ifr, sizeof(ifr.ifr_name));
2928 2929 2930 2931 2932 2933
		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);
2934 2935 2936 2937 2938 2939
		if (err)
			err = -EFAULT;
	}
	return err;
}

2940
struct rtentry32 {
2941
	u32		rt_pad1;
2942 2943 2944
	struct sockaddr rt_dst;         /* target address               */
	struct sockaddr rt_gateway;     /* gateway addr (RTF_GATEWAY)   */
	struct sockaddr rt_genmask;     /* target network mask (IP)     */
2945 2946 2947 2948 2949 2950 2951
	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! */
2952
	/* char * */ u32 rt_dev;        /* forcing the device at add    */
2953 2954
	u32		rt_mtu;         /* per route MTU/Window         */
	u32		rt_window;      /* Window clamping              */
2955 2956 2957 2958 2959 2960 2961 2962 2963 2964 2965 2966 2967 2968 2969 2970
	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;
};

2971 2972
static int routing_ioctl(struct net *net, struct socket *sock,
			 unsigned int cmd, void __user *argp)
2973 2974 2975 2976 2977 2978 2979 2980 2981
{
	int ret;
	void *r = NULL;
	struct in6_rtmsg r6;
	struct rtentry r4;
	char devname[16];
	u32 rtdev;
	mm_segment_t old_fs = get_fs();

2982 2983
	if (sock && sock->sk && sock->sk->sk_family == AF_INET6) { /* ipv6 */
		struct in6_rtmsg32 __user *ur6 = argp;
2984
		ret = copy_from_user(&r6.rtmsg_dst, &(ur6->rtmsg_dst),
2985
			3 * sizeof(struct in6_addr));
2986 2987 2988 2989 2990 2991 2992
		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));
2993 2994 2995

		r = (void *) &r6;
	} else { /* ipv4 */
2996
		struct rtentry32 __user *ur4 = argp;
2997
		ret = copy_from_user(&r4.rt_dst, &(ur4->rt_dst),
2998
					3 * sizeof(struct sockaddr));
2999 3000 3001 3002 3003 3004
		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));
3005
		if (rtdev) {
3006
			ret |= copy_from_user(devname, compat_ptr(rtdev), 15);
3007 3008
			r4.rt_dev = (char __user __force *)devname;
			devname[15] = 0;
3009 3010 3011 3012 3013 3014 3015 3016 3017 3018 3019
		} else
			r4.rt_dev = NULL;

		r = (void *) &r4;
	}

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

3020
	set_fs(KERNEL_DS);
3021
	ret = sock_do_ioctl(net, sock, cmd, (unsigned long) r);
3022
	set_fs(old_fs);
3023 3024 3025 3026 3027 3028 3029

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
3030
 * use compatible ioctls
3031
 */
3032
static int old_bridge_ioctl(compat_ulong_t __user *argp)
3033
{
3034
	compat_ulong_t tmp;
3035

3036
	if (get_user(tmp, argp))
3037 3038 3039 3040 3041 3042
		return -EFAULT;
	if (tmp == BRCTL_GET_VERSION)
		return BRCTL_VERSION + 1;
	return -EINVAL;
}

3043 3044 3045 3046 3047 3048
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);
3049

3050
	if (cmd >= SIOCDEVPRIVATE && cmd <= (SIOCDEVPRIVATE + 15))
3051
		return compat_ifr_data_ioctl(net, cmd, argp);
3052 3053 3054 3055 3056 3057 3058 3059 3060 3061 3062

	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);
3063 3064
	case SIOCWANDEV:
		return compat_siocwandev(net, argp);
3065 3066 3067
	case SIOCGIFMAP:
	case SIOCSIFMAP:
		return compat_sioc_ifmap(net, cmd, argp);
3068 3069 3070 3071 3072 3073 3074 3075 3076 3077 3078 3079
	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);
3080 3081
	case SIOCBONDSLAVEINFOQUERY:
	case SIOCBONDINFOQUERY:
3082
	case SIOCSHWTSTAMP:
3083
	case SIOCGHWTSTAMP:
3084
		return compat_ifr_data_ioctl(net, cmd, argp);
3085 3086 3087 3088 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

	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:
3127 3128 3129 3130
	case SIOCSIFNAME:
	case SIOCGMIIPHY:
	case SIOCGMIIREG:
	case SIOCSMIIREG:
3131
		return dev_ifsioc(net, sock, cmd, argp);
3132

3133 3134 3135 3136
	case SIOCSARP:
	case SIOCGARP:
	case SIOCDARP:
	case SIOCATMARK:
3137 3138 3139
		return sock_do_ioctl(net, sock, cmd, arg);
	}

3140 3141
	return -ENOIOCTLCMD;
}
3142

3143
static long compat_sock_ioctl(struct file *file, unsigned int cmd,
3144
			      unsigned long arg)
3145 3146 3147
{
	struct socket *sock = file->private_data;
	int ret = -ENOIOCTLCMD;
3148 3149 3150 3151 3152
	struct sock *sk;
	struct net *net;

	sk = sock->sk;
	net = sock_net(sk);
3153 3154 3155 3156

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

3157 3158 3159 3160
	if (ret == -ENOIOCTLCMD &&
	    (cmd >= SIOCIWFIRST && cmd <= SIOCIWLAST))
		ret = compat_wext_handle_ioctl(net, cmd, arg);

3161 3162 3163
	if (ret == -ENOIOCTLCMD)
		ret = compat_sock_ioctl_trans(file, sock, cmd, arg);

3164 3165 3166 3167
	return ret;
}
#endif

3168 3169 3170 3171
int kernel_bind(struct socket *sock, struct sockaddr *addr, int addrlen)
{
	return sock->ops->bind(sock, addr, addrlen);
}
3172
EXPORT_SYMBOL(kernel_bind);
3173 3174 3175 3176 3177

int kernel_listen(struct socket *sock, int backlog)
{
	return sock->ops->listen(sock, backlog);
}
3178
EXPORT_SYMBOL(kernel_listen);
3179 3180 3181 3182 3183 3184 3185 3186 3187 3188 3189 3190 3191 3192

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);
3193
		*newsock = NULL;
3194 3195 3196 3197
		goto done;
	}

	(*newsock)->ops = sock->ops;
3198
	__module_get((*newsock)->ops->owner);
3199 3200 3201 3202

done:
	return err;
}
3203
EXPORT_SYMBOL(kernel_accept);
3204 3205

int kernel_connect(struct socket *sock, struct sockaddr *addr, int addrlen,
3206
		   int flags)
3207 3208 3209
{
	return sock->ops->connect(sock, addr, addrlen, flags);
}
3210
EXPORT_SYMBOL(kernel_connect);
3211 3212 3213 3214 3215 3216

int kernel_getsockname(struct socket *sock, struct sockaddr *addr,
			 int *addrlen)
{
	return sock->ops->getname(sock, addr, addrlen, 0);
}
3217
EXPORT_SYMBOL(kernel_getsockname);
3218 3219 3220 3221 3222 3223

int kernel_getpeername(struct socket *sock, struct sockaddr *addr,
			 int *addrlen)
{
	return sock->ops->getname(sock, addr, addrlen, 1);
}
3224
EXPORT_SYMBOL(kernel_getpeername);
3225 3226 3227 3228 3229

int kernel_getsockopt(struct socket *sock, int level, int optname,
			char *optval, int *optlen)
{
	mm_segment_t oldfs = get_fs();
3230 3231
	char __user *uoptval;
	int __user *uoptlen;
3232 3233
	int err;

3234 3235 3236
	uoptval = (char __user __force *) optval;
	uoptlen = (int __user __force *) optlen;

3237 3238
	set_fs(KERNEL_DS);
	if (level == SOL_SOCKET)
3239
		err = sock_getsockopt(sock, level, optname, uoptval, uoptlen);
3240
	else
3241 3242
		err = sock->ops->getsockopt(sock, level, optname, uoptval,
					    uoptlen);
3243 3244 3245
	set_fs(oldfs);
	return err;
}
3246
EXPORT_SYMBOL(kernel_getsockopt);
3247 3248

int kernel_setsockopt(struct socket *sock, int level, int optname,
3249
			char *optval, unsigned int optlen)
3250 3251
{
	mm_segment_t oldfs = get_fs();
3252
	char __user *uoptval;
3253 3254
	int err;

3255 3256
	uoptval = (char __user __force *) optval;

3257 3258
	set_fs(KERNEL_DS);
	if (level == SOL_SOCKET)
3259
		err = sock_setsockopt(sock, level, optname, uoptval, optlen);
3260
	else
3261
		err = sock->ops->setsockopt(sock, level, optname, uoptval,
3262 3263 3264 3265
					    optlen);
	set_fs(oldfs);
	return err;
}
3266
EXPORT_SYMBOL(kernel_setsockopt);
3267 3268 3269 3270 3271 3272 3273 3274 3275

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);
}
3276
EXPORT_SYMBOL(kernel_sendpage);
3277 3278 3279 3280 3281 3282 3283 3284 3285 3286 3287 3288

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;
}
3289
EXPORT_SYMBOL(kernel_sock_ioctl);
3290

3291 3292 3293 3294 3295
int kernel_sock_shutdown(struct socket *sock, enum sock_shutdown_cmd how)
{
	return sock->ops->shutdown(sock, how);
}
EXPORT_SYMBOL(kernel_sock_shutdown);