socket.c 83.3 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

L
Linus Torvalds 已提交
116
static int sock_no_open(struct inode *irrelevant, struct file *dontcare);
117 118 119 120
static ssize_t sock_aio_read(struct kiocb *iocb, const struct iovec *iov,
			 unsigned long nr_segs, loff_t pos);
static ssize_t sock_aio_write(struct kiocb *iocb, const struct iovec *iov,
			  unsigned long nr_segs, loff_t pos);
121
static int sock_mmap(struct file *file, struct vm_area_struct *vma);
L
Linus Torvalds 已提交
122 123 124 125

static int sock_close(struct inode *inode, struct file *file);
static unsigned int sock_poll(struct file *file,
			      struct poll_table_struct *wait);
126
static long sock_ioctl(struct file *file, unsigned int cmd, unsigned long arg);
127 128
#ifdef CONFIG_COMPAT
static long compat_sock_ioctl(struct file *file,
129
			      unsigned int cmd, unsigned long arg);
130
#endif
L
Linus Torvalds 已提交
131 132 133
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
134
static ssize_t sock_splice_read(struct file *file, loff_t *ppos,
135
				struct pipe_inode_info *pipe, size_t len,
J
Jens Axboe 已提交
136
				unsigned int flags);
L
Linus Torvalds 已提交
137 138 139 140 141 142

/*
 *	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.
 */

143
static const struct file_operations socket_file_ops = {
L
Linus Torvalds 已提交
144 145 146 147 148 149
	.owner =	THIS_MODULE,
	.llseek =	no_llseek,
	.aio_read =	sock_aio_read,
	.aio_write =	sock_aio_write,
	.poll =		sock_poll,
	.unlocked_ioctl = sock_ioctl,
150 151 152
#ifdef CONFIG_COMPAT
	.compat_ioctl = compat_sock_ioctl,
#endif
L
Linus Torvalds 已提交
153 154 155 156
	.mmap =		sock_mmap,
	.open =		sock_no_open,	/* special open code to disallow open via /proc */
	.release =	sock_close,
	.fasync =	sock_fasync,
157 158
	.sendpage =	sock_sendpage,
	.splice_write = generic_splice_sendpage,
J
Jens Axboe 已提交
159
	.splice_read =	sock_splice_read,
L
Linus Torvalds 已提交
160 161 162 163 164 165 166
};

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

static DEFINE_SPINLOCK(net_family_lock);
E
Eric Dumazet 已提交
167
static const struct net_proto_family __rcu *net_families[NPROTO] __read_mostly;
L
Linus Torvalds 已提交
168 169 170 171 172

/*
 *	Statistics counters of the socket lists
 */

173
static DEFINE_PER_CPU(int, sockets_in_use);
L
Linus Torvalds 已提交
174 175

/*
176 177 178
 * 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
179 180 181 182 183 184 185 186 187 188 189 190 191
 */

/**
 *	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.
 */

192
int move_addr_to_kernel(void __user *uaddr, int ulen, struct sockaddr_storage *kaddr)
L
Linus Torvalds 已提交
193
{
194
	if (ulen < 0 || ulen > sizeof(struct sockaddr_storage))
L
Linus Torvalds 已提交
195
		return -EINVAL;
196
	if (ulen == 0)
L
Linus Torvalds 已提交
197
		return 0;
198
	if (copy_from_user(kaddr, uaddr, ulen))
L
Linus Torvalds 已提交
199
		return -EFAULT;
200
	return audit_sockaddr(ulen, kaddr);
L
Linus Torvalds 已提交
201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218
}

/**
 *	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.
 */
219

220
static int move_addr_to_user(struct sockaddr_storage *kaddr, int klen,
S
stephen hemminger 已提交
221
			     void __user *uaddr, int __user *ulen)
L
Linus Torvalds 已提交
222 223 224 225
{
	int err;
	int len;

226
	BUG_ON(klen > sizeof(struct sockaddr_storage));
227 228
	err = get_user(len, ulen);
	if (err)
L
Linus Torvalds 已提交
229
		return err;
230 231
	if (len > klen)
		len = klen;
232
	if (len < 0)
L
Linus Torvalds 已提交
233
		return -EINVAL;
234
	if (len) {
S
Steve Grubb 已提交
235 236
		if (audit_sockaddr(klen, kaddr))
			return -ENOMEM;
237
		if (copy_to_user(uaddr, kaddr, len))
L
Linus Torvalds 已提交
238 239 240
			return -EFAULT;
	}
	/*
241 242
	 *      "fromlen shall refer to the value before truncation.."
	 *                      1003.1g
L
Linus Torvalds 已提交
243 244 245 246
	 */
	return __put_user(klen, ulen);
}

247
static struct kmem_cache *sock_inode_cachep __read_mostly;
L
Linus Torvalds 已提交
248 249 250 251

static struct inode *sock_alloc_inode(struct super_block *sb)
{
	struct socket_alloc *ei;
252
	struct socket_wq *wq;
253

254
	ei = kmem_cache_alloc(sock_inode_cachep, GFP_KERNEL);
L
Linus Torvalds 已提交
255 256
	if (!ei)
		return NULL;
257 258
	wq = kmalloc(sizeof(*wq), GFP_KERNEL);
	if (!wq) {
259 260 261
		kmem_cache_free(sock_inode_cachep, ei);
		return NULL;
	}
262 263 264
	init_waitqueue_head(&wq->wait);
	wq->fasync_list = NULL;
	RCU_INIT_POINTER(ei->socket.wq, wq);
265

L
Linus Torvalds 已提交
266 267 268 269 270 271 272 273 274 275 276
	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)
{
277
	struct socket_alloc *ei;
278
	struct socket_wq *wq;
279 280

	ei = container_of(inode, struct socket_alloc, vfs_inode);
281
	wq = rcu_dereference_protected(ei->socket.wq, 1);
282
	kfree_rcu(wq, rcu);
283
	kmem_cache_free(sock_inode_cachep, ei);
L
Linus Torvalds 已提交
284 285
}

286
static void init_once(void *foo)
L
Linus Torvalds 已提交
287
{
288
	struct socket_alloc *ei = (struct socket_alloc *)foo;
L
Linus Torvalds 已提交
289

C
Christoph Lameter 已提交
290
	inode_init_once(&ei->vfs_inode);
L
Linus Torvalds 已提交
291
}
292

L
Linus Torvalds 已提交
293 294 295
static int init_inodecache(void)
{
	sock_inode_cachep = kmem_cache_create("sock_inode_cache",
296 297 298 299 300
					      sizeof(struct socket_alloc),
					      0,
					      (SLAB_HWCACHE_ALIGN |
					       SLAB_RECLAIM_ACCOUNT |
					       SLAB_MEM_SPREAD),
301
					      init_once);
L
Linus Torvalds 已提交
302 303 304 305 306
	if (sock_inode_cachep == NULL)
		return -ENOMEM;
	return 0;
}

307
static const struct super_operations sockfs_ops = {
308 309 310
	.alloc_inode	= sock_alloc_inode,
	.destroy_inode	= sock_destroy_inode,
	.statfs		= simple_statfs,
L
Linus Torvalds 已提交
311 312
};

313 314 315 316 317 318 319 320 321
/*
 * sockfs_dname() is called from d_path().
 */
static char *sockfs_dname(struct dentry *dentry, char *buffer, int buflen)
{
	return dynamic_dname(dentry, buffer, buflen, "socket:[%lu]",
				dentry->d_inode->i_ino);
}

A
Al Viro 已提交
322
static const struct dentry_operations sockfs_dentry_operations = {
323
	.d_dname  = sockfs_dname,
L
Linus Torvalds 已提交
324 325
};

326 327 328 329 330 331 332 333 334 335 336 337 338 339 340
static struct dentry *sockfs_mount(struct file_system_type *fs_type,
			 int flags, const char *dev_name, void *data)
{
	return mount_pseudo(fs_type, "socket:", &sockfs_ops,
		&sockfs_dentry_operations, SOCKFS_MAGIC);
}

static struct vfsmount *sock_mnt __read_mostly;

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

L
Linus Torvalds 已提交
341 342 343
/*
 *	Obtains the first available file descriptor and sets it up for use.
 *
344 345
 *	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
346 347 348 349 350 351 352 353 354 355 356 357
 *	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.
 */

358
struct file *sock_alloc_file(struct socket *sock, int flags, const char *dname)
L
Linus Torvalds 已提交
359
{
360
	struct qstr name = { .name = "" };
361
	struct path path;
362
	struct file *file;
L
Linus Torvalds 已提交
363

364 365 366 367 368 369 370
	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);
	}
371
	path.dentry = d_alloc_pseudo(sock_mnt->mnt_sb, &name);
372 373
	if (unlikely(!path.dentry))
		return ERR_PTR(-ENOMEM);
374
	path.mnt = mntget(sock_mnt);
375

376
	d_instantiate(path.dentry, SOCK_INODE(sock));
A
Al Viro 已提交
377
	SOCK_INODE(sock)->i_fop = &socket_file_ops;
378

379
	file = alloc_file(&path, FMODE_READ | FMODE_WRITE,
380
		  &socket_file_ops);
381
	if (unlikely(IS_ERR(file))) {
A
Al Viro 已提交
382
		/* drop dentry, keep inode */
A
Al Viro 已提交
383
		ihold(path.dentry->d_inode);
384
		path_put(&path);
385
		return file;
A
Al Viro 已提交
386 387 388
	}

	sock->file = file;
389
	file->f_flags = O_RDWR | (flags & O_NONBLOCK);
390
	file->private_data = sock;
391
	return file;
392
}
393
EXPORT_SYMBOL(sock_alloc_file);
394

395
static int sock_map_fd(struct socket *sock, int flags)
396 397
{
	struct file *newfile;
398 399 400
	int fd = get_unused_fd_flags(flags);
	if (unlikely(fd < 0))
		return fd;
401

402
	newfile = sock_alloc_file(sock, flags, NULL);
403
	if (likely(!IS_ERR(newfile))) {
404
		fd_install(fd, newfile);
405 406
		return fd;
	}
407

408 409
	put_unused_fd(fd);
	return PTR_ERR(newfile);
L
Linus Torvalds 已提交
410 411
}

412
struct socket *sock_from_file(struct file *file, int *err)
413 414 415 416
{
	if (file->f_op == &socket_file_ops)
		return file->private_data;	/* set in sock_map_fd */

E
Eric Dumazet 已提交
417 418
	*err = -ENOTSOCK;
	return NULL;
419
}
420
EXPORT_SYMBOL(sock_from_file);
421

L
Linus Torvalds 已提交
422
/**
423
 *	sockfd_lookup - Go from a file number to its socket slot
L
Linus Torvalds 已提交
424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 439
 *	@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;

440 441
	file = fget(fd);
	if (!file) {
L
Linus Torvalds 已提交
442 443 444
		*err = -EBADF;
		return NULL;
	}
445

446 447
	sock = sock_from_file(file, err);
	if (!sock)
L
Linus Torvalds 已提交
448
		fput(file);
449 450
	return sock;
}
451
EXPORT_SYMBOL(sockfd_lookup);
L
Linus Torvalds 已提交
452

453 454
static struct socket *sockfd_lookup_light(int fd, int *err, int *fput_needed)
{
455
	struct fd f = fdget(fd);
456 457
	struct socket *sock;

458
	*err = -EBADF;
459 460 461 462
	if (f.file) {
		sock = sock_from_file(f.file, err);
		if (likely(sock)) {
			*fput_needed = f.flags;
463
			return sock;
464 465
		}
		fdput(f);
L
Linus Torvalds 已提交
466
	}
467
	return NULL;
L
Linus Torvalds 已提交
468 469
}

470 471 472 473 474 475 476 477 478 479 480 481 482 483 484 485 486 487 488 489 490 491 492 493 494 495 496 497 498 499 500 501 502 503 504 505 506 507 508 509 510 511 512 513 514 515 516 517 518 519 520 521 522 523 524 525 526 527 528 529 530 531
#define XATTR_SOCKPROTONAME_SUFFIX "sockprotoname"
#define XATTR_NAME_SOCKPROTONAME (XATTR_SYSTEM_PREFIX XATTR_SOCKPROTONAME_SUFFIX)
#define XATTR_NAME_SOCKPROTONAME_LEN (sizeof(XATTR_NAME_SOCKPROTONAME)-1)
static ssize_t sockfs_getxattr(struct dentry *dentry,
			       const char *name, void *value, size_t size)
{
	const char *proto_name;
	size_t proto_size;
	int error;

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

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

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

out:
	return error;
}

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

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

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

	return used;
}

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

L
Linus Torvalds 已提交
532 533
/**
 *	sock_alloc	-	allocate a socket
534
 *
L
Linus Torvalds 已提交
535 536 537 538 539 540 541
 *	Allocate a new inode and socket object. The two are bound together
 *	and initialised. The socket is then returned. If we are out of inodes
 *	NULL is returned.
 */

static struct socket *sock_alloc(void)
{
542 543
	struct inode *inode;
	struct socket *sock;
L
Linus Torvalds 已提交
544

545
	inode = new_inode_pseudo(sock_mnt->mnt_sb);
L
Linus Torvalds 已提交
546 547 548 549 550
	if (!inode)
		return NULL;

	sock = SOCKET_I(inode);

551
	kmemcheck_annotate_bitfield(sock, type);
552
	inode->i_ino = get_next_ino();
553
	inode->i_mode = S_IFSOCK | S_IRWXUGO;
554 555
	inode->i_uid = current_fsuid();
	inode->i_gid = current_fsgid();
556
	inode->i_op = &sockfs_inode_ops;
L
Linus Torvalds 已提交
557

558
	this_cpu_add(sockets_in_use, 1);
L
Linus Torvalds 已提交
559 560 561 562 563 564 565 566
	return sock;
}

/*
 *	In theory you can't get an open on this inode, but /proc provides
 *	a back door. Remember to keep it shut otherwise you'll let the
 *	creepy crawlies in.
 */
567

L
Linus Torvalds 已提交
568 569 570 571 572
static int sock_no_open(struct inode *irrelevant, struct file *dontcare)
{
	return -ENXIO;
}

573
const struct file_operations bad_sock_fops = {
L
Linus Torvalds 已提交
574 575
	.owner = THIS_MODULE,
	.open = sock_no_open,
576
	.llseek = noop_llseek,
L
Linus Torvalds 已提交
577 578 579 580 581 582 583 584
};

/**
 *	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
585
 *	an inode not a file.
L
Linus Torvalds 已提交
586
 */
587

L
Linus Torvalds 已提交
588 589 590 591 592 593 594 595 596 597
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);
	}

598
	if (rcu_dereference_protected(sock->wq, 1)->fasync_list)
599
		pr_err("%s: fasync list not empty!\n", __func__);
L
Linus Torvalds 已提交
600

601 602 603
	if (test_bit(SOCK_EXTERNALLY_ALLOCATED, &sock->flags))
		return;

604
	this_cpu_sub(sockets_in_use, 1);
L
Linus Torvalds 已提交
605 606 607 608
	if (!sock->file) {
		iput(SOCK_INODE(sock));
		return;
	}
609
	sock->file = NULL;
L
Linus Torvalds 已提交
610
}
611
EXPORT_SYMBOL(sock_release);
L
Linus Torvalds 已提交
612

613
void __sock_tx_timestamp(const struct sock *sk, __u8 *tx_flags)
614
{
615 616
	u8 flags = *tx_flags;

617
	if (sk->sk_tsflags & SOF_TIMESTAMPING_TX_HARDWARE)
618 619
		flags |= SKBTX_HW_TSTAMP;

620
	if (sk->sk_tsflags & SOF_TIMESTAMPING_TX_SOFTWARE)
621 622
		flags |= SKBTX_SW_TSTAMP;

623
	if (sk->sk_tsflags & SOF_TIMESTAMPING_TX_SCHED)
624 625
		flags |= SKBTX_SCHED_TSTAMP;

626
	if (sk->sk_tsflags & SOF_TIMESTAMPING_TX_ACK)
627
		flags |= SKBTX_ACK_TSTAMP;
628

629
	*tx_flags = flags;
630
}
631
EXPORT_SYMBOL(__sock_tx_timestamp);
632

633 634
static inline int __sock_sendmsg_nosec(struct kiocb *iocb, struct socket *sock,
				       struct msghdr *msg, size_t size)
L
Linus Torvalds 已提交
635 636 637 638 639 640 641 642 643 644 645
{
	struct sock_iocb *si = kiocb_to_siocb(iocb);

	si->sock = sock;
	si->scm = NULL;
	si->msg = msg;
	si->size = size;

	return sock->ops->sendmsg(iocb, sock, msg, size);
}

646 647 648 649 650 651 652 653
static inline int __sock_sendmsg(struct kiocb *iocb, struct socket *sock,
				 struct msghdr *msg, size_t size)
{
	int err = security_socket_sendmsg(sock, msg, size);

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

L
Linus Torvalds 已提交
654 655 656 657 658 659 660 661 662 663 664 665 666
int sock_sendmsg(struct socket *sock, struct msghdr *msg, size_t size)
{
	struct kiocb iocb;
	struct sock_iocb siocb;
	int ret;

	init_sync_kiocb(&iocb, NULL);
	iocb.private = &siocb;
	ret = __sock_sendmsg(&iocb, sock, msg, size);
	if (-EIOCBQUEUED == ret)
		ret = wait_on_sync_kiocb(&iocb);
	return ret;
}
667
EXPORT_SYMBOL(sock_sendmsg);
L
Linus Torvalds 已提交
668

669
static int sock_sendmsg_nosec(struct socket *sock, struct msghdr *msg, size_t size)
670 671 672 673 674 675 676 677 678 679 680 681 682
{
	struct kiocb iocb;
	struct sock_iocb siocb;
	int ret;

	init_sync_kiocb(&iocb, NULL);
	iocb.private = &siocb;
	ret = __sock_sendmsg_nosec(&iocb, sock, msg, size);
	if (-EIOCBQUEUED == ret)
		ret = wait_on_sync_kiocb(&iocb);
	return ret;
}

L
Linus Torvalds 已提交
683 684 685 686 687 688 689 690 691 692 693
int kernel_sendmsg(struct socket *sock, struct msghdr *msg,
		   struct kvec *vec, size_t num, size_t size)
{
	mm_segment_t oldfs = get_fs();
	int result;

	set_fs(KERNEL_DS);
	/*
	 * the following is safe, since for compiler definitions of kvec and
	 * iovec are identical, yielding the same in-core layout and alignment
	 */
694
	msg->msg_iov = (struct iovec *)vec;
L
Linus Torvalds 已提交
695 696 697 698 699
	msg->msg_iovlen = num;
	result = sock_sendmsg(sock, msg, size);
	set_fs(oldfs);
	return result;
}
700
EXPORT_SYMBOL(kernel_sendmsg);
L
Linus Torvalds 已提交
701

702 703 704 705 706 707
/*
 * 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)
{
708
	int need_software_tstamp = sock_flag(sk, SOCK_RCVTSTAMP);
709
	struct scm_timestamping tss;
710 711 712 713 714 715 716 717 718 719 720 721 722 723 724 725
	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 {
726 727
			struct timespec ts;
			skb_get_timestampns(skb, &ts);
728
			put_cmsg(msg, SOL_SOCKET, SCM_TIMESTAMPNS,
729
				 sizeof(ts), &ts);
730 731 732
		}
	}

733
	memset(&tss, 0, sizeof(tss));
734
	if ((sk->sk_tsflags & SOF_TIMESTAMPING_SOFTWARE) &&
735
	    ktime_to_timespec_cond(skb->tstamp, tss.ts + 0))
736
		empty = 0;
737
	if (shhwtstamps &&
738
	    (sk->sk_tsflags & SOF_TIMESTAMPING_RAW_HARDWARE) &&
739
	    ktime_to_timespec_cond(shhwtstamps->hwtstamp, tss.ts + 2))
740
		empty = 0;
741 742
	if (!empty)
		put_cmsg(msg, SOL_SOCKET,
743
			 SCM_TIMESTAMPING, sizeof(tss), &tss);
744
}
745 746
EXPORT_SYMBOL_GPL(__sock_recv_timestamp);

747 748 749 750 751 752 753 754 755 756 757 758 759 760 761 762
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
763 764
static inline void sock_recv_drops(struct msghdr *msg, struct sock *sk,
				   struct sk_buff *skb)
765 766 767 768 769 770
{
	if (sock_flag(sk, SOCK_RXQ_OVFL) && skb && skb->dropcount)
		put_cmsg(msg, SOL_SOCKET, SO_RXQ_OVFL,
			sizeof(__u32), &skb->dropcount);
}

771
void __sock_recv_ts_and_drops(struct msghdr *msg, struct sock *sk,
772 773 774 775 776
	struct sk_buff *skb)
{
	sock_recv_timestamp(msg, sk, skb);
	sock_recv_drops(msg, sk, skb);
}
777
EXPORT_SYMBOL_GPL(__sock_recv_ts_and_drops);
778

779 780
static inline int __sock_recvmsg_nosec(struct kiocb *iocb, struct socket *sock,
				       struct msghdr *msg, size_t size, int flags)
L
Linus Torvalds 已提交
781 782 783 784 785 786 787 788 789 790 791 792
{
	struct sock_iocb *si = kiocb_to_siocb(iocb);

	si->sock = sock;
	si->scm = NULL;
	si->msg = msg;
	si->size = size;
	si->flags = flags;

	return sock->ops->recvmsg(iocb, sock, msg, size, flags);
}

793 794 795 796 797 798 799 800
static inline int __sock_recvmsg(struct kiocb *iocb, struct socket *sock,
				 struct msghdr *msg, size_t size, int flags)
{
	int err = security_socket_recvmsg(sock, msg, size, flags);

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

801
int sock_recvmsg(struct socket *sock, struct msghdr *msg,
L
Linus Torvalds 已提交
802 803 804 805 806 807
		 size_t size, int flags)
{
	struct kiocb iocb;
	struct sock_iocb siocb;
	int ret;

808
	init_sync_kiocb(&iocb, NULL);
L
Linus Torvalds 已提交
809 810 811 812 813 814
	iocb.private = &siocb;
	ret = __sock_recvmsg(&iocb, sock, msg, size, flags);
	if (-EIOCBQUEUED == ret)
		ret = wait_on_sync_kiocb(&iocb);
	return ret;
}
815
EXPORT_SYMBOL(sock_recvmsg);
L
Linus Torvalds 已提交
816

817 818 819 820 821 822 823 824 825 826 827 828 829 830 831
static int sock_recvmsg_nosec(struct socket *sock, struct msghdr *msg,
			      size_t size, int flags)
{
	struct kiocb iocb;
	struct sock_iocb siocb;
	int ret;

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

832 833 834 835 836 837 838 839 840 841 842 843 844 845 846
/**
 * 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.
 */
847 848
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
849 850 851 852 853 854 855 856 857
{
	mm_segment_t oldfs = get_fs();
	int result;

	set_fs(KERNEL_DS);
	/*
	 * the following is safe, since for compiler definitions of kvec and
	 * iovec are identical, yielding the same in-core layout and alignment
	 */
858
	msg->msg_iov = (struct iovec *)vec, msg->msg_iovlen = num;
L
Linus Torvalds 已提交
859 860 861 862
	result = sock_recvmsg(sock, msg, size, flags);
	set_fs(oldfs);
	return result;
}
863
EXPORT_SYMBOL(kernel_recvmsg);
L
Linus Torvalds 已提交
864

865 866
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
867 868 869 870
{
	struct socket *sock;
	int flags;

871 872
	sock = file->private_data;

873 874 875
	flags = (file->f_flags & O_NONBLOCK) ? MSG_DONTWAIT : 0;
	/* more is a combination of MSG_MORE and MSG_SENDPAGE_NOTLAST */
	flags |= more;
876

877
	return kernel_sendpage(sock, page, offset, size, flags);
878
}
L
Linus Torvalds 已提交
879

J
Jens Axboe 已提交
880
static ssize_t sock_splice_read(struct file *file, loff_t *ppos,
881
				struct pipe_inode_info *pipe, size_t len,
J
Jens Axboe 已提交
882 883 884 885
				unsigned int flags)
{
	struct socket *sock = file->private_data;

886 887 888
	if (unlikely(!sock->ops->splice_read))
		return -EINVAL;

J
Jens Axboe 已提交
889 890 891
	return sock->ops->splice_read(sock, ppos, pipe, len, flags);
}

892
static struct sock_iocb *alloc_sock_iocb(struct kiocb *iocb,
893
					 struct sock_iocb *siocb)
894
{
K
Kent Overstreet 已提交
895 896
	if (!is_sync_kiocb(iocb))
		BUG();
L
Linus Torvalds 已提交
897

898 899 900
	siocb->kiocb = iocb;
	iocb->private = siocb;
	return siocb;
L
Linus Torvalds 已提交
901 902
}

903
static ssize_t do_sock_read(struct msghdr *msg, struct kiocb *iocb,
904 905
		struct file *file, const struct iovec *iov,
		unsigned long nr_segs)
906 907 908 909
{
	struct socket *sock = file->private_data;
	size_t size = 0;
	int i;
L
Linus Torvalds 已提交
910

911 912
	for (i = 0; i < nr_segs; i++)
		size += iov[i].iov_len;
L
Linus Torvalds 已提交
913

914 915 916 917
	msg->msg_name = NULL;
	msg->msg_namelen = 0;
	msg->msg_control = NULL;
	msg->msg_controllen = 0;
918
	msg->msg_iov = (struct iovec *)iov;
919 920 921 922 923 924
	msg->msg_iovlen = nr_segs;
	msg->msg_flags = (file->f_flags & O_NONBLOCK) ? MSG_DONTWAIT : 0;

	return __sock_recvmsg(iocb, sock, msg, size, msg->msg_flags);
}

925 926
static ssize_t sock_aio_read(struct kiocb *iocb, const struct iovec *iov,
				unsigned long nr_segs, loff_t pos)
927 928 929
{
	struct sock_iocb siocb, *x;

L
Linus Torvalds 已提交
930 931
	if (pos != 0)
		return -ESPIPE;
932

K
Kent Overstreet 已提交
933
	if (iocb->ki_nbytes == 0)	/* Match SYS5 behaviour */
L
Linus Torvalds 已提交
934 935
		return 0;

936 937

	x = alloc_sock_iocb(iocb, &siocb);
938 939
	if (!x)
		return -ENOMEM;
940
	return do_sock_read(&x->async_msg, iocb, iocb->ki_filp, iov, nr_segs);
L
Linus Torvalds 已提交
941 942
}

943
static ssize_t do_sock_write(struct msghdr *msg, struct kiocb *iocb,
944 945
			struct file *file, const struct iovec *iov,
			unsigned long nr_segs)
L
Linus Torvalds 已提交
946
{
947 948 949
	struct socket *sock = file->private_data;
	size_t size = 0;
	int i;
L
Linus Torvalds 已提交
950

951 952
	for (i = 0; i < nr_segs; i++)
		size += iov[i].iov_len;
L
Linus Torvalds 已提交
953

954 955 956 957
	msg->msg_name = NULL;
	msg->msg_namelen = 0;
	msg->msg_control = NULL;
	msg->msg_controllen = 0;
958
	msg->msg_iov = (struct iovec *)iov;
959 960 961 962
	msg->msg_iovlen = nr_segs;
	msg->msg_flags = (file->f_flags & O_NONBLOCK) ? MSG_DONTWAIT : 0;
	if (sock->type == SOCK_SEQPACKET)
		msg->msg_flags |= MSG_EOR;
L
Linus Torvalds 已提交
963

964
	return __sock_sendmsg(iocb, sock, msg, size);
L
Linus Torvalds 已提交
965 966
}

967 968
static ssize_t sock_aio_write(struct kiocb *iocb, const struct iovec *iov,
			  unsigned long nr_segs, loff_t pos)
969 970
{
	struct sock_iocb siocb, *x;
L
Linus Torvalds 已提交
971

972 973
	if (pos != 0)
		return -ESPIPE;
974 975

	x = alloc_sock_iocb(iocb, &siocb);
976 977
	if (!x)
		return -ENOMEM;
L
Linus Torvalds 已提交
978

979
	return do_sock_write(&x->async_msg, iocb, iocb->ki_filp, iov, nr_segs);
L
Linus Torvalds 已提交
980 981 982 983 984 985 986
}

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

A
Arjan van de Ven 已提交
987
static DEFINE_MUTEX(br_ioctl_mutex);
988
static int (*br_ioctl_hook) (struct net *, unsigned int cmd, void __user *arg);
L
Linus Torvalds 已提交
989

990
void brioctl_set(int (*hook) (struct net *, unsigned int, void __user *))
L
Linus Torvalds 已提交
991
{
A
Arjan van de Ven 已提交
992
	mutex_lock(&br_ioctl_mutex);
L
Linus Torvalds 已提交
993
	br_ioctl_hook = hook;
A
Arjan van de Ven 已提交
994
	mutex_unlock(&br_ioctl_mutex);
L
Linus Torvalds 已提交
995 996 997
}
EXPORT_SYMBOL(brioctl_set);

A
Arjan van de Ven 已提交
998
static DEFINE_MUTEX(vlan_ioctl_mutex);
999
static int (*vlan_ioctl_hook) (struct net *, void __user *arg);
L
Linus Torvalds 已提交
1000

1001
void vlan_ioctl_set(int (*hook) (struct net *, void __user *))
L
Linus Torvalds 已提交
1002
{
A
Arjan van de Ven 已提交
1003
	mutex_lock(&vlan_ioctl_mutex);
L
Linus Torvalds 已提交
1004
	vlan_ioctl_hook = hook;
A
Arjan van de Ven 已提交
1005
	mutex_unlock(&vlan_ioctl_mutex);
L
Linus Torvalds 已提交
1006 1007 1008
}
EXPORT_SYMBOL(vlan_ioctl_set);

A
Arjan van de Ven 已提交
1009
static DEFINE_MUTEX(dlci_ioctl_mutex);
1010
static int (*dlci_ioctl_hook) (unsigned int, void __user *);
L
Linus Torvalds 已提交
1011

1012
void dlci_ioctl_set(int (*hook) (unsigned int, void __user *))
L
Linus Torvalds 已提交
1013
{
A
Arjan van de Ven 已提交
1014
	mutex_lock(&dlci_ioctl_mutex);
L
Linus Torvalds 已提交
1015
	dlci_ioctl_hook = hook;
A
Arjan van de Ven 已提交
1016
	mutex_unlock(&dlci_ioctl_mutex);
L
Linus Torvalds 已提交
1017 1018 1019
}
EXPORT_SYMBOL(dlci_ioctl_set);

1020 1021 1022 1023 1024 1025 1026 1027 1028 1029 1030 1031 1032 1033 1034 1035 1036 1037
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
1038 1039 1040 1041 1042 1043 1044 1045
/*
 *	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;
1046
	struct sock *sk;
L
Linus Torvalds 已提交
1047 1048
	void __user *argp = (void __user *)arg;
	int pid, err;
1049
	struct net *net;
L
Linus Torvalds 已提交
1050

1051
	sock = file->private_data;
1052
	sk = sock->sk;
1053
	net = sock_net(sk);
L
Linus Torvalds 已提交
1054
	if (cmd >= SIOCDEVPRIVATE && cmd <= (SIOCDEVPRIVATE + 15)) {
1055
		err = dev_ioctl(net, cmd, argp);
L
Linus Torvalds 已提交
1056
	} else
J
Johannes Berg 已提交
1057
#ifdef CONFIG_WEXT_CORE
L
Linus Torvalds 已提交
1058
	if (cmd >= SIOCIWFIRST && cmd <= SIOCIWLAST) {
1059
		err = dev_ioctl(net, cmd, argp);
L
Linus Torvalds 已提交
1060
	} else
J
Johannes Berg 已提交
1061
#endif
1062
		switch (cmd) {
L
Linus Torvalds 已提交
1063 1064 1065 1066 1067
		case FIOSETOWN:
		case SIOCSPGRP:
			err = -EFAULT;
			if (get_user(pid, (int __user *)argp))
				break;
1068 1069
			f_setown(sock->file, pid, 1);
			err = 0;
L
Linus Torvalds 已提交
1070 1071 1072
			break;
		case FIOGETOWN:
		case SIOCGPGRP:
1073
			err = put_user(f_getown(sock->file),
1074
				       (int __user *)argp);
L
Linus Torvalds 已提交
1075 1076 1077 1078 1079 1080 1081 1082 1083
			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
1084
			mutex_lock(&br_ioctl_mutex);
1085
			if (br_ioctl_hook)
1086
				err = br_ioctl_hook(net, cmd, argp);
A
Arjan van de Ven 已提交
1087
			mutex_unlock(&br_ioctl_mutex);
L
Linus Torvalds 已提交
1088 1089 1090 1091 1092 1093 1094
			break;
		case SIOCGIFVLAN:
		case SIOCSIFVLAN:
			err = -ENOPKG;
			if (!vlan_ioctl_hook)
				request_module("8021q");

A
Arjan van de Ven 已提交
1095
			mutex_lock(&vlan_ioctl_mutex);
L
Linus Torvalds 已提交
1096
			if (vlan_ioctl_hook)
1097
				err = vlan_ioctl_hook(net, argp);
A
Arjan van de Ven 已提交
1098
			mutex_unlock(&vlan_ioctl_mutex);
L
Linus Torvalds 已提交
1099 1100 1101 1102 1103 1104 1105
			break;
		case SIOCADDDLCI:
		case SIOCDELDLCI:
			err = -ENOPKG;
			if (!dlci_ioctl_hook)
				request_module("dlci");

1106 1107
			mutex_lock(&dlci_ioctl_mutex);
			if (dlci_ioctl_hook)
L
Linus Torvalds 已提交
1108
				err = dlci_ioctl_hook(cmd, argp);
1109
			mutex_unlock(&dlci_ioctl_mutex);
L
Linus Torvalds 已提交
1110 1111
			break;
		default:
1112
			err = sock_do_ioctl(net, sock, cmd, arg);
L
Linus Torvalds 已提交
1113
			break;
1114
		}
L
Linus Torvalds 已提交
1115 1116 1117 1118 1119 1120 1121
	return err;
}

int sock_create_lite(int family, int type, int protocol, struct socket **res)
{
	int err;
	struct socket *sock = NULL;
1122

L
Linus Torvalds 已提交
1123 1124 1125 1126 1127 1128 1129 1130 1131 1132 1133
	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
1134 1135 1136 1137
	err = security_socket_post_create(sock, family, type, protocol, 1);
	if (err)
		goto out_release;

L
Linus Torvalds 已提交
1138 1139 1140
out:
	*res = sock;
	return err;
V
Venkat Yekkirala 已提交
1141 1142 1143 1144
out_release:
	sock_release(sock);
	sock = NULL;
	goto out;
L
Linus Torvalds 已提交
1145
}
1146
EXPORT_SYMBOL(sock_create_lite);
L
Linus Torvalds 已提交
1147 1148

/* No kernel lock held - perfect */
1149
static unsigned int sock_poll(struct file *file, poll_table *wait)
L
Linus Torvalds 已提交
1150
{
1151
	unsigned int busy_flag = 0;
L
Linus Torvalds 已提交
1152 1153 1154
	struct socket *sock;

	/*
1155
	 *      We can't return errors to poll, so it's either yes or no.
L
Linus Torvalds 已提交
1156
	 */
1157
	sock = file->private_data;
1158

1159
	if (sk_can_busy_loop(sock->sk)) {
1160
		/* this socket can poll_ll so tell the system call */
1161
		busy_flag = POLL_BUSY_LOOP;
1162 1163

		/* once, only if requested by syscall */
1164 1165
		if (wait && (wait->_key & POLL_BUSY_LOOP))
			sk_busy_loop(sock->sk, 1);
1166 1167
	}

1168
	return busy_flag | sock->ops->poll(file, sock, wait);
L
Linus Torvalds 已提交
1169 1170
}

1171
static int sock_mmap(struct file *file, struct vm_area_struct *vma)
L
Linus Torvalds 已提交
1172
{
1173
	struct socket *sock = file->private_data;
L
Linus Torvalds 已提交
1174 1175 1176 1177

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

1178
static int sock_close(struct inode *inode, struct file *filp)
L
Linus Torvalds 已提交
1179 1180 1181 1182 1183 1184 1185 1186 1187 1188 1189 1190 1191
{
	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)
1192
 *	   or under socket lock
L
Linus Torvalds 已提交
1193 1194 1195 1196
 */

static int sock_fasync(int fd, struct file *filp, int on)
{
1197 1198
	struct socket *sock = filp->private_data;
	struct sock *sk = sock->sk;
1199
	struct socket_wq *wq;
L
Linus Torvalds 已提交
1200

1201
	if (sk == NULL)
L
Linus Torvalds 已提交
1202 1203 1204
		return -EINVAL;

	lock_sock(sk);
1205 1206
	wq = rcu_dereference_protected(sock->wq, sock_owned_by_user(sk));
	fasync_helper(fd, filp, on, &wq->fasync_list);
L
Linus Torvalds 已提交
1207

1208
	if (!wq->fasync_list)
1209 1210
		sock_reset_flag(sk, SOCK_FASYNC);
	else
E
Eric Dumazet 已提交
1211
		sock_set_flag(sk, SOCK_FASYNC);
L
Linus Torvalds 已提交
1212

1213
	release_sock(sk);
L
Linus Torvalds 已提交
1214 1215 1216
	return 0;
}

1217
/* This function may be called only under socket lock or callback_lock or rcu_lock */
L
Linus Torvalds 已提交
1218 1219 1220

int sock_wake_async(struct socket *sock, int how, int band)
{
1221 1222 1223 1224 1225 1226 1227 1228
	struct socket_wq *wq;

	if (!sock)
		return -1;
	rcu_read_lock();
	wq = rcu_dereference(sock->wq);
	if (!wq || !wq->fasync_list) {
		rcu_read_unlock();
L
Linus Torvalds 已提交
1229
		return -1;
1230
	}
1231
	switch (how) {
1232
	case SOCK_WAKE_WAITD:
L
Linus Torvalds 已提交
1233 1234 1235
		if (test_bit(SOCK_ASYNC_WAITDATA, &sock->flags))
			break;
		goto call_kill;
1236
	case SOCK_WAKE_SPACE:
L
Linus Torvalds 已提交
1237 1238 1239
		if (!test_and_clear_bit(SOCK_ASYNC_NOSPACE, &sock->flags))
			break;
		/* fall through */
1240
	case SOCK_WAKE_IO:
1241
call_kill:
1242
		kill_fasync(&wq->fasync_list, SIGIO, band);
L
Linus Torvalds 已提交
1243
		break;
1244
	case SOCK_WAKE_URG:
1245
		kill_fasync(&wq->fasync_list, SIGURG, band);
L
Linus Torvalds 已提交
1246
	}
1247
	rcu_read_unlock();
L
Linus Torvalds 已提交
1248 1249
	return 0;
}
1250
EXPORT_SYMBOL(sock_wake_async);
L
Linus Torvalds 已提交
1251

P
Pavel Emelyanov 已提交
1252
int __sock_create(struct net *net, int family, int type, int protocol,
1253
			 struct socket **res, int kern)
L
Linus Torvalds 已提交
1254 1255 1256
{
	int err;
	struct socket *sock;
1257
	const struct net_proto_family *pf;
L
Linus Torvalds 已提交
1258 1259

	/*
1260
	 *      Check protocol is in range
L
Linus Torvalds 已提交
1261 1262 1263 1264 1265 1266 1267 1268 1269 1270 1271 1272
	 */
	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) {
1273
		static int warned;
L
Linus Torvalds 已提交
1274 1275
		if (!warned) {
			warned = 1;
1276 1277
			pr_info("%s uses obsolete (PF_INET,SOCK_PACKET)\n",
				current->comm);
L
Linus Torvalds 已提交
1278 1279 1280 1281 1282 1283 1284
		}
		family = PF_PACKET;
	}

	err = security_socket_create(family, type, protocol, kern);
	if (err)
		return err;
1285

1286 1287 1288 1289 1290 1291 1292
	/*
	 *	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) {
1293
		net_warn_ratelimited("socket: no more sockets\n");
1294 1295 1296 1297 1298 1299
		return -ENFILE;	/* Not exactly a match, but its the
				   closest posix thing */
	}

	sock->type = type;

1300
#ifdef CONFIG_MODULES
1301 1302 1303
	/* 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
1304 1305 1306
	 * requested real, full-featured networking support upon configuration.
	 * Otherwise module support will break!
	 */
E
Eric Dumazet 已提交
1307
	if (rcu_access_pointer(net_families[family]) == NULL)
1308
		request_module("net-pf-%d", family);
L
Linus Torvalds 已提交
1309 1310
#endif

1311 1312 1313 1314 1315
	rcu_read_lock();
	pf = rcu_dereference(net_families[family]);
	err = -EAFNOSUPPORT;
	if (!pf)
		goto out_release;
L
Linus Torvalds 已提交
1316 1317 1318 1319 1320

	/*
	 * We will call the ->create function, that possibly is in a loadable
	 * module, so we have to bump that loadable module refcnt first.
	 */
1321
	if (!try_module_get(pf->owner))
L
Linus Torvalds 已提交
1322 1323
		goto out_release;

1324 1325 1326
	/* Now protected by module ref count */
	rcu_read_unlock();

1327
	err = pf->create(net, sock, protocol, kern);
1328
	if (err < 0)
L
Linus Torvalds 已提交
1329
		goto out_module_put;
1330

L
Linus Torvalds 已提交
1331 1332 1333 1334
	/*
	 * Now to bump the refcnt of the [loadable] module that owns this
	 * socket at sock_release time we decrement its refcnt.
	 */
1335 1336 1337
	if (!try_module_get(sock->ops->owner))
		goto out_module_busy;

L
Linus Torvalds 已提交
1338 1339 1340 1341
	/*
	 * Now that we're done with the ->create function, the [loadable]
	 * module can have its refcnt decremented
	 */
1342
	module_put(pf->owner);
V
Venkat Yekkirala 已提交
1343 1344
	err = security_socket_post_create(sock, family, type, protocol, kern);
	if (err)
1345
		goto out_sock_release;
1346
	*res = sock;
L
Linus Torvalds 已提交
1347

1348 1349 1350 1351
	return 0;

out_module_busy:
	err = -EAFNOSUPPORT;
L
Linus Torvalds 已提交
1352
out_module_put:
1353 1354 1355
	sock->ops = NULL;
	module_put(pf->owner);
out_sock_release:
L
Linus Torvalds 已提交
1356
	sock_release(sock);
1357 1358 1359 1360 1361
	return err;

out_release:
	rcu_read_unlock();
	goto out_sock_release;
L
Linus Torvalds 已提交
1362
}
P
Pavel Emelyanov 已提交
1363
EXPORT_SYMBOL(__sock_create);
L
Linus Torvalds 已提交
1364 1365 1366

int sock_create(int family, int type, int protocol, struct socket **res)
{
1367
	return __sock_create(current->nsproxy->net_ns, family, type, protocol, res, 0);
L
Linus Torvalds 已提交
1368
}
1369
EXPORT_SYMBOL(sock_create);
L
Linus Torvalds 已提交
1370 1371 1372

int sock_create_kern(int family, int type, int protocol, struct socket **res)
{
1373
	return __sock_create(&init_net, family, type, protocol, res, 1);
L
Linus Torvalds 已提交
1374
}
1375
EXPORT_SYMBOL(sock_create_kern);
L
Linus Torvalds 已提交
1376

1377
SYSCALL_DEFINE3(socket, int, family, int, type, int, protocol)
L
Linus Torvalds 已提交
1378 1379 1380
{
	int retval;
	struct socket *sock;
1381 1382
	int flags;

1383 1384 1385 1386 1387 1388
	/* 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);

1389
	flags = type & ~SOCK_TYPE_MASK;
1390
	if (flags & ~(SOCK_CLOEXEC | SOCK_NONBLOCK))
1391 1392
		return -EINVAL;
	type &= SOCK_TYPE_MASK;
L
Linus Torvalds 已提交
1393

U
Ulrich Drepper 已提交
1394 1395 1396
	if (SOCK_NONBLOCK != O_NONBLOCK && (flags & SOCK_NONBLOCK))
		flags = (flags & ~SOCK_NONBLOCK) | O_NONBLOCK;

L
Linus Torvalds 已提交
1397 1398 1399 1400
	retval = sock_create(family, type, protocol, &sock);
	if (retval < 0)
		goto out;

1401
	retval = sock_map_fd(sock, flags & (O_CLOEXEC | O_NONBLOCK));
L
Linus Torvalds 已提交
1402 1403 1404 1405 1406 1407 1408 1409 1410 1411 1412 1413 1414 1415 1416 1417
	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.
 */

1418 1419
SYSCALL_DEFINE4(socketpair, int, family, int, type, int, protocol,
		int __user *, usockvec)
L
Linus Torvalds 已提交
1420 1421 1422
{
	struct socket *sock1, *sock2;
	int fd1, fd2, err;
A
Al Viro 已提交
1423
	struct file *newfile1, *newfile2;
1424 1425 1426
	int flags;

	flags = type & ~SOCK_TYPE_MASK;
1427
	if (flags & ~(SOCK_CLOEXEC | SOCK_NONBLOCK))
1428 1429
		return -EINVAL;
	type &= SOCK_TYPE_MASK;
L
Linus Torvalds 已提交
1430

U
Ulrich Drepper 已提交
1431 1432 1433
	if (SOCK_NONBLOCK != O_NONBLOCK && (flags & SOCK_NONBLOCK))
		flags = (flags & ~SOCK_NONBLOCK) | O_NONBLOCK;

L
Linus Torvalds 已提交
1434 1435 1436 1437 1438 1439 1440 1441 1442 1443 1444 1445 1446 1447
	/*
	 * 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);
1448
	if (err < 0)
L
Linus Torvalds 已提交
1449 1450
		goto out_release_both;

1451
	fd1 = get_unused_fd_flags(flags);
1452 1453
	if (unlikely(fd1 < 0)) {
		err = fd1;
A
Al Viro 已提交
1454
		goto out_release_both;
1455
	}
1456

1457
	fd2 = get_unused_fd_flags(flags);
1458 1459
	if (unlikely(fd2 < 0)) {
		err = fd2;
1460
		goto out_put_unused_1;
1461 1462
	}

1463
	newfile1 = sock_alloc_file(sock1, flags, NULL);
1464 1465
	if (unlikely(IS_ERR(newfile1))) {
		err = PTR_ERR(newfile1);
1466
		goto out_put_unused_both;
1467 1468
	}

1469
	newfile2 = sock_alloc_file(sock2, flags, NULL);
1470 1471
	if (IS_ERR(newfile2)) {
		err = PTR_ERR(newfile2);
1472
		goto out_fput_1;
A
Al Viro 已提交
1473 1474
	}

1475 1476 1477 1478 1479 1480 1481 1482
	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
1483
	audit_fd_pair(fd1, fd2);
1484

A
Al Viro 已提交
1485 1486
	fd_install(fd1, newfile1);
	fd_install(fd2, newfile2);
L
Linus Torvalds 已提交
1487 1488 1489 1490
	/* fd1 and fd2 may be already another descriptors.
	 * Not kernel problem.
	 */

1491
	return 0;
L
Linus Torvalds 已提交
1492

1493 1494 1495 1496 1497 1498 1499 1500 1501 1502 1503 1504 1505
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
1506

1507 1508 1509 1510
out_put_unused_both:
	put_unused_fd(fd2);
out_put_unused_1:
	put_unused_fd(fd1);
L
Linus Torvalds 已提交
1511
out_release_both:
1512
	sock_release(sock2);
L
Linus Torvalds 已提交
1513
out_release_1:
1514
	sock_release(sock1);
L
Linus Torvalds 已提交
1515 1516 1517 1518 1519 1520 1521 1522 1523 1524 1525 1526
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).
 */

1527
SYSCALL_DEFINE3(bind, int, fd, struct sockaddr __user *, umyaddr, int, addrlen)
L
Linus Torvalds 已提交
1528 1529
{
	struct socket *sock;
1530
	struct sockaddr_storage address;
1531
	int err, fput_needed;
L
Linus Torvalds 已提交
1532

1533
	sock = sockfd_lookup_light(fd, &err, &fput_needed);
1534
	if (sock) {
1535
		err = move_addr_to_kernel(umyaddr, addrlen, &address);
1536 1537
		if (err >= 0) {
			err = security_socket_bind(sock,
1538
						   (struct sockaddr *)&address,
1539
						   addrlen);
1540 1541
			if (!err)
				err = sock->ops->bind(sock,
1542
						      (struct sockaddr *)
1543
						      &address, addrlen);
L
Linus Torvalds 已提交
1544
		}
1545
		fput_light(sock->file, fput_needed);
1546
	}
L
Linus Torvalds 已提交
1547 1548 1549 1550 1551 1552 1553 1554 1555
	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.
 */

1556
SYSCALL_DEFINE2(listen, int, fd, int, backlog)
L
Linus Torvalds 已提交
1557 1558
{
	struct socket *sock;
1559
	int err, fput_needed;
1560
	int somaxconn;
1561 1562 1563

	sock = sockfd_lookup_light(fd, &err, &fput_needed);
	if (sock) {
1564
		somaxconn = sock_net(sock->sk)->core.sysctl_somaxconn;
1565
		if ((unsigned int)backlog > somaxconn)
1566
			backlog = somaxconn;
L
Linus Torvalds 已提交
1567 1568

		err = security_socket_listen(sock, backlog);
1569 1570
		if (!err)
			err = sock->ops->listen(sock, backlog);
L
Linus Torvalds 已提交
1571

1572
		fput_light(sock->file, fput_needed);
L
Linus Torvalds 已提交
1573 1574 1575 1576 1577 1578 1579 1580 1581 1582 1583 1584 1585 1586 1587 1588
	}
	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.
 */

1589 1590
SYSCALL_DEFINE4(accept4, int, fd, struct sockaddr __user *, upeer_sockaddr,
		int __user *, upeer_addrlen, int, flags)
L
Linus Torvalds 已提交
1591 1592
{
	struct socket *sock, *newsock;
1593
	struct file *newfile;
1594
	int err, len, newfd, fput_needed;
1595
	struct sockaddr_storage address;
L
Linus Torvalds 已提交
1596

1597
	if (flags & ~(SOCK_CLOEXEC | SOCK_NONBLOCK))
U
Ulrich Drepper 已提交
1598 1599 1600 1601 1602
		return -EINVAL;

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

1603
	sock = sockfd_lookup_light(fd, &err, &fput_needed);
L
Linus Torvalds 已提交
1604 1605 1606 1607
	if (!sock)
		goto out;

	err = -ENFILE;
1608 1609
	newsock = sock_alloc();
	if (!newsock)
L
Linus Torvalds 已提交
1610 1611 1612 1613 1614 1615 1616 1617 1618 1619 1620
		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);

1621
	newfd = get_unused_fd_flags(flags);
1622 1623
	if (unlikely(newfd < 0)) {
		err = newfd;
1624 1625
		sock_release(newsock);
		goto out_put;
1626
	}
1627
	newfile = sock_alloc_file(newsock, flags, sock->sk->sk_prot_creator->name);
1628 1629 1630 1631 1632 1633
	if (unlikely(IS_ERR(newfile))) {
		err = PTR_ERR(newfile);
		put_unused_fd(newfd);
		sock_release(newsock);
		goto out_put;
	}
1634

1635 1636
	err = security_socket_accept(sock, newsock);
	if (err)
1637
		goto out_fd;
1638

L
Linus Torvalds 已提交
1639 1640
	err = sock->ops->accept(sock, newsock, sock->file->f_flags);
	if (err < 0)
1641
		goto out_fd;
L
Linus Torvalds 已提交
1642 1643

	if (upeer_sockaddr) {
1644
		if (newsock->ops->getname(newsock, (struct sockaddr *)&address,
1645
					  &len, 2) < 0) {
L
Linus Torvalds 已提交
1646
			err = -ECONNABORTED;
1647
			goto out_fd;
L
Linus Torvalds 已提交
1648
		}
1649
		err = move_addr_to_user(&address,
1650
					len, upeer_sockaddr, upeer_addrlen);
L
Linus Torvalds 已提交
1651
		if (err < 0)
1652
			goto out_fd;
L
Linus Torvalds 已提交
1653 1654 1655 1656
	}

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

1657 1658
	fd_install(newfd, newfile);
	err = newfd;
L
Linus Torvalds 已提交
1659 1660

out_put:
1661
	fput_light(sock->file, fput_needed);
L
Linus Torvalds 已提交
1662 1663
out:
	return err;
1664
out_fd:
1665
	fput(newfile);
1666
	put_unused_fd(newfd);
L
Linus Torvalds 已提交
1667 1668 1669
	goto out_put;
}

1670 1671
SYSCALL_DEFINE3(accept, int, fd, struct sockaddr __user *, upeer_sockaddr,
		int __user *, upeer_addrlen)
U
Ulrich Drepper 已提交
1672
{
U
Ulrich Drepper 已提交
1673
	return sys_accept4(fd, upeer_sockaddr, upeer_addrlen, 0);
U
Ulrich Drepper 已提交
1674 1675
}

L
Linus Torvalds 已提交
1676 1677 1678 1679 1680 1681 1682 1683 1684 1685 1686 1687
/*
 *	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.
 */

1688 1689
SYSCALL_DEFINE3(connect, int, fd, struct sockaddr __user *, uservaddr,
		int, addrlen)
L
Linus Torvalds 已提交
1690 1691
{
	struct socket *sock;
1692
	struct sockaddr_storage address;
1693
	int err, fput_needed;
L
Linus Torvalds 已提交
1694

1695
	sock = sockfd_lookup_light(fd, &err, &fput_needed);
L
Linus Torvalds 已提交
1696 1697
	if (!sock)
		goto out;
1698
	err = move_addr_to_kernel(uservaddr, addrlen, &address);
L
Linus Torvalds 已提交
1699 1700 1701
	if (err < 0)
		goto out_put;

1702
	err =
1703
	    security_socket_connect(sock, (struct sockaddr *)&address, addrlen);
L
Linus Torvalds 已提交
1704 1705 1706
	if (err)
		goto out_put;

1707
	err = sock->ops->connect(sock, (struct sockaddr *)&address, addrlen,
L
Linus Torvalds 已提交
1708 1709
				 sock->file->f_flags);
out_put:
1710
	fput_light(sock->file, fput_needed);
L
Linus Torvalds 已提交
1711 1712 1713 1714 1715 1716 1717 1718 1719
out:
	return err;
}

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

1720 1721
SYSCALL_DEFINE3(getsockname, int, fd, struct sockaddr __user *, usockaddr,
		int __user *, usockaddr_len)
L
Linus Torvalds 已提交
1722 1723
{
	struct socket *sock;
1724
	struct sockaddr_storage address;
1725
	int len, err, fput_needed;
1726

1727
	sock = sockfd_lookup_light(fd, &err, &fput_needed);
L
Linus Torvalds 已提交
1728 1729 1730 1731 1732 1733 1734
	if (!sock)
		goto out;

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

1735
	err = sock->ops->getname(sock, (struct sockaddr *)&address, &len, 0);
L
Linus Torvalds 已提交
1736 1737
	if (err)
		goto out_put;
1738
	err = move_addr_to_user(&address, len, usockaddr, usockaddr_len);
L
Linus Torvalds 已提交
1739 1740

out_put:
1741
	fput_light(sock->file, fput_needed);
L
Linus Torvalds 已提交
1742 1743 1744 1745 1746 1747 1748 1749 1750
out:
	return err;
}

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

1751 1752
SYSCALL_DEFINE3(getpeername, int, fd, struct sockaddr __user *, usockaddr,
		int __user *, usockaddr_len)
L
Linus Torvalds 已提交
1753 1754
{
	struct socket *sock;
1755
	struct sockaddr_storage address;
1756
	int len, err, fput_needed;
L
Linus Torvalds 已提交
1757

1758 1759
	sock = sockfd_lookup_light(fd, &err, &fput_needed);
	if (sock != NULL) {
L
Linus Torvalds 已提交
1760 1761
		err = security_socket_getpeername(sock);
		if (err) {
1762
			fput_light(sock->file, fput_needed);
L
Linus Torvalds 已提交
1763 1764 1765
			return err;
		}

1766
		err =
1767
		    sock->ops->getname(sock, (struct sockaddr *)&address, &len,
1768
				       1);
L
Linus Torvalds 已提交
1769
		if (!err)
1770
			err = move_addr_to_user(&address, len, usockaddr,
1771
						usockaddr_len);
1772
		fput_light(sock->file, fput_needed);
L
Linus Torvalds 已提交
1773 1774 1775 1776 1777 1778 1779 1780 1781 1782
	}
	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.
 */

1783
SYSCALL_DEFINE6(sendto, int, fd, void __user *, buff, size_t, len,
1784
		unsigned int, flags, struct sockaddr __user *, addr,
1785
		int, addr_len)
L
Linus Torvalds 已提交
1786 1787
{
	struct socket *sock;
1788
	struct sockaddr_storage address;
L
Linus Torvalds 已提交
1789 1790 1791
	int err;
	struct msghdr msg;
	struct iovec iov;
1792 1793
	int fput_needed;

1794 1795
	if (len > INT_MAX)
		len = INT_MAX;
1796 1797
	sock = sockfd_lookup_light(fd, &err, &fput_needed);
	if (!sock)
1798
		goto out;
1799

1800 1801 1802 1803 1804 1805 1806 1807
	iov.iov_base = buff;
	iov.iov_len = len;
	msg.msg_name = NULL;
	msg.msg_iov = &iov;
	msg.msg_iovlen = 1;
	msg.msg_control = NULL;
	msg.msg_controllen = 0;
	msg.msg_namelen = 0;
1808
	if (addr) {
1809
		err = move_addr_to_kernel(addr, addr_len, &address);
L
Linus Torvalds 已提交
1810 1811
		if (err < 0)
			goto out_put;
1812
		msg.msg_name = (struct sockaddr *)&address;
1813
		msg.msg_namelen = addr_len;
L
Linus Torvalds 已提交
1814 1815 1816 1817 1818 1819
	}
	if (sock->file->f_flags & O_NONBLOCK)
		flags |= MSG_DONTWAIT;
	msg.msg_flags = flags;
	err = sock_sendmsg(sock, &msg, len);

1820
out_put:
1821
	fput_light(sock->file, fput_needed);
1822
out:
L
Linus Torvalds 已提交
1823 1824 1825 1826
	return err;
}

/*
1827
 *	Send a datagram down a socket.
L
Linus Torvalds 已提交
1828 1829
 */

1830
SYSCALL_DEFINE4(send, int, fd, void __user *, buff, size_t, len,
1831
		unsigned int, flags)
L
Linus Torvalds 已提交
1832 1833 1834 1835 1836
{
	return sys_sendto(fd, buff, len, flags, NULL, 0);
}

/*
1837
 *	Receive a frame from the socket and optionally record the address of the
L
Linus Torvalds 已提交
1838 1839 1840 1841
 *	sender. We verify the buffers are writable and if needed move the
 *	sender address from kernel to user space.
 */

1842
SYSCALL_DEFINE6(recvfrom, int, fd, void __user *, ubuf, size_t, size,
1843
		unsigned int, flags, struct sockaddr __user *, addr,
1844
		int __user *, addr_len)
L
Linus Torvalds 已提交
1845 1846 1847 1848
{
	struct socket *sock;
	struct iovec iov;
	struct msghdr msg;
1849
	struct sockaddr_storage address;
1850
	int err, err2;
1851 1852
	int fput_needed;

1853 1854
	if (size > INT_MAX)
		size = INT_MAX;
1855
	sock = sockfd_lookup_light(fd, &err, &fput_needed);
L
Linus Torvalds 已提交
1856
	if (!sock)
1857
		goto out;
L
Linus Torvalds 已提交
1858

1859 1860 1861 1862 1863 1864
	msg.msg_control = NULL;
	msg.msg_controllen = 0;
	msg.msg_iovlen = 1;
	msg.msg_iov = &iov;
	iov.iov_len = size;
	iov.iov_base = ubuf;
1865 1866 1867 1868
	/* Save some cycles and don't copy the address if not needed */
	msg.msg_name = addr ? (struct sockaddr *)&address : NULL;
	/* We assume all kernel code knows the size of sockaddr_storage */
	msg.msg_namelen = 0;
L
Linus Torvalds 已提交
1869 1870
	if (sock->file->f_flags & O_NONBLOCK)
		flags |= MSG_DONTWAIT;
1871
	err = sock_recvmsg(sock, &msg, size, flags);
L
Linus Torvalds 已提交
1872

1873
	if (err >= 0 && addr != NULL) {
1874
		err2 = move_addr_to_user(&address,
1875
					 msg.msg_namelen, addr, addr_len);
1876 1877
		if (err2 < 0)
			err = err2;
L
Linus Torvalds 已提交
1878
	}
1879 1880

	fput_light(sock->file, fput_needed);
1881
out:
L
Linus Torvalds 已提交
1882 1883 1884 1885
	return err;
}

/*
1886
 *	Receive a datagram from a socket.
L
Linus Torvalds 已提交
1887 1888
 */

1889 1890
SYSCALL_DEFINE4(recv, int, fd, void __user *, ubuf, size_t, size,
		unsigned int, flags)
L
Linus Torvalds 已提交
1891 1892 1893 1894 1895 1896 1897 1898 1899
{
	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.
 */

1900 1901
SYSCALL_DEFINE5(setsockopt, int, fd, int, level, int, optname,
		char __user *, optval, int, optlen)
L
Linus Torvalds 已提交
1902
{
1903
	int err, fput_needed;
L
Linus Torvalds 已提交
1904 1905 1906 1907
	struct socket *sock;

	if (optlen < 0)
		return -EINVAL;
1908 1909 1910 1911

	sock = sockfd_lookup_light(fd, &err, &fput_needed);
	if (sock != NULL) {
		err = security_socket_setsockopt(sock, level, optname);
1912 1913
		if (err)
			goto out_put;
L
Linus Torvalds 已提交
1914 1915

		if (level == SOL_SOCKET)
1916 1917 1918
			err =
			    sock_setsockopt(sock, level, optname, optval,
					    optlen);
L
Linus Torvalds 已提交
1919
		else
1920 1921 1922
			err =
			    sock->ops->setsockopt(sock, level, optname, optval,
						  optlen);
1923 1924
out_put:
		fput_light(sock->file, fput_needed);
L
Linus Torvalds 已提交
1925 1926 1927 1928 1929 1930 1931 1932 1933
	}
	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.
 */

1934 1935
SYSCALL_DEFINE5(getsockopt, int, fd, int, level, int, optname,
		char __user *, optval, int __user *, optlen)
L
Linus Torvalds 已提交
1936
{
1937
	int err, fput_needed;
L
Linus Torvalds 已提交
1938 1939
	struct socket *sock;

1940 1941
	sock = sockfd_lookup_light(fd, &err, &fput_needed);
	if (sock != NULL) {
1942 1943 1944
		err = security_socket_getsockopt(sock, level, optname);
		if (err)
			goto out_put;
L
Linus Torvalds 已提交
1945 1946

		if (level == SOL_SOCKET)
1947 1948 1949
			err =
			    sock_getsockopt(sock, level, optname, optval,
					    optlen);
L
Linus Torvalds 已提交
1950
		else
1951 1952 1953
			err =
			    sock->ops->getsockopt(sock, level, optname, optval,
						  optlen);
1954 1955
out_put:
		fput_light(sock->file, fput_needed);
L
Linus Torvalds 已提交
1956 1957 1958 1959 1960 1961 1962 1963
	}
	return err;
}

/*
 *	Shutdown a socket.
 */

1964
SYSCALL_DEFINE2(shutdown, int, fd, int, how)
L
Linus Torvalds 已提交
1965
{
1966
	int err, fput_needed;
L
Linus Torvalds 已提交
1967 1968
	struct socket *sock;

1969 1970
	sock = sockfd_lookup_light(fd, &err, &fput_needed);
	if (sock != NULL) {
L
Linus Torvalds 已提交
1971
		err = security_socket_shutdown(sock, how);
1972 1973 1974
		if (!err)
			err = sock->ops->shutdown(sock, how);
		fput_light(sock->file, fput_needed);
L
Linus Torvalds 已提交
1975 1976 1977 1978
	}
	return err;
}

1979
/* A couple of helpful macros for getting the address of the 32/64 bit
L
Linus Torvalds 已提交
1980 1981 1982 1983 1984 1985
 * 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)

1986 1987 1988 1989 1990
struct used_address {
	struct sockaddr_storage name;
	unsigned int name_len;
};

1991
static int copy_msghdr_from_user(struct msghdr *kmsg,
1992
				 struct user_msghdr __user *umsg)
1993
{
1994 1995 1996
	/* We are relying on the (currently) identical layouts.  Once
	 * the kernel-side changes, this place will need to be updated
	 */
1997 1998
	if (copy_from_user(kmsg, umsg, sizeof(struct msghdr)))
		return -EFAULT;
1999

2000 2001 2002
	if (kmsg->msg_name == NULL)
		kmsg->msg_namelen = 0;

2003 2004 2005
	if (kmsg->msg_namelen < 0)
		return -EINVAL;

2006
	if (kmsg->msg_namelen > sizeof(struct sockaddr_storage))
2007
		kmsg->msg_namelen = sizeof(struct sockaddr_storage);
2008 2009 2010
	return 0;
}

2011
static int ___sys_sendmsg(struct socket *sock, struct user_msghdr __user *msg,
2012
			 struct msghdr *msg_sys, unsigned int flags,
2013
			 struct used_address *used_address)
L
Linus Torvalds 已提交
2014
{
2015 2016
	struct compat_msghdr __user *msg_compat =
	    (struct compat_msghdr __user *)msg;
2017
	struct sockaddr_storage address;
L
Linus Torvalds 已提交
2018
	struct iovec iovstack[UIO_FASTIOV], *iov = iovstack;
2019
	unsigned char ctl[sizeof(struct cmsghdr) + 20]
2020 2021
	    __attribute__ ((aligned(sizeof(__kernel_size_t))));
	/* 20 is size of ipv6_pktinfo */
L
Linus Torvalds 已提交
2022
	unsigned char *ctl_buf = ctl;
E
Eric Dumazet 已提交
2023
	int err, ctl_len, total_len;
2024

L
Linus Torvalds 已提交
2025 2026
	err = -EFAULT;
	if (MSG_CMSG_COMPAT & flags) {
2027
		if (get_compat_msghdr(msg_sys, msg_compat))
L
Linus Torvalds 已提交
2028
			return -EFAULT;
2029 2030 2031 2032 2033
	} else {
		err = copy_msghdr_from_user(msg_sys, msg);
		if (err)
			return err;
	}
L
Linus Torvalds 已提交
2034 2035

	/* This will also move the address data into kernel space */
2036 2037 2038 2039
	if (MSG_CMSG_COMPAT & flags)
		err = verify_compat_iovec(msg_sys, iovstack, &address, WRITE);
	else
		err = verify_iovec(msg_sys, iovstack, &address, WRITE);
2040
	if (err < 0)
L
Linus Torvalds 已提交
2041
		goto out_freeiov;
2042
	iov = msg_sys->msg_iov;
L
Linus Torvalds 已提交
2043 2044 2045 2046
	total_len = err;

	err = -ENOBUFS;

2047
	if (msg_sys->msg_controllen > INT_MAX)
L
Linus Torvalds 已提交
2048
		goto out_freeiov;
2049
	ctl_len = msg_sys->msg_controllen;
L
Linus Torvalds 已提交
2050
	if ((MSG_CMSG_COMPAT & flags) && ctl_len) {
2051
		err =
2052
		    cmsghdr_from_user_compat_to_kern(msg_sys, sock->sk, ctl,
2053
						     sizeof(ctl));
L
Linus Torvalds 已提交
2054 2055
		if (err)
			goto out_freeiov;
2056 2057
		ctl_buf = msg_sys->msg_control;
		ctl_len = msg_sys->msg_controllen;
L
Linus Torvalds 已提交
2058
	} else if (ctl_len) {
2059
		if (ctl_len > sizeof(ctl)) {
L
Linus Torvalds 已提交
2060
			ctl_buf = sock_kmalloc(sock->sk, ctl_len, GFP_KERNEL);
2061
			if (ctl_buf == NULL)
L
Linus Torvalds 已提交
2062 2063 2064 2065
				goto out_freeiov;
		}
		err = -EFAULT;
		/*
2066
		 * Careful! Before this, msg_sys->msg_control contains a user pointer.
L
Linus Torvalds 已提交
2067 2068 2069
		 * Afterwards, it will be a kernel pointer. Thus the compiler-assisted
		 * checking falls down on this.
		 */
2070
		if (copy_from_user(ctl_buf,
2071
				   (void __user __force *)msg_sys->msg_control,
2072
				   ctl_len))
L
Linus Torvalds 已提交
2073
			goto out_freectl;
2074
		msg_sys->msg_control = ctl_buf;
L
Linus Torvalds 已提交
2075
	}
2076
	msg_sys->msg_flags = flags;
L
Linus Torvalds 已提交
2077 2078

	if (sock->file->f_flags & O_NONBLOCK)
2079
		msg_sys->msg_flags |= MSG_DONTWAIT;
2080 2081 2082 2083 2084 2085
	/*
	 * 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.
	 */
2086 2087 2088
	if (used_address && msg_sys->msg_name &&
	    used_address->name_len == msg_sys->msg_namelen &&
	    !memcmp(&used_address->name, msg_sys->msg_name,
2089 2090 2091 2092 2093 2094 2095 2096 2097 2098 2099
		    used_address->name_len)) {
		err = sock_sendmsg_nosec(sock, msg_sys, total_len);
		goto out_freectl;
	}
	err = sock_sendmsg(sock, msg_sys, total_len);
	/*
	 * If this is sendmmsg() and sending to current destination address was
	 * successful, remember it.
	 */
	if (used_address && err >= 0) {
		used_address->name_len = msg_sys->msg_namelen;
2100 2101 2102
		if (msg_sys->msg_name)
			memcpy(&used_address->name, msg_sys->msg_name,
			       used_address->name_len);
2103
	}
L
Linus Torvalds 已提交
2104 2105

out_freectl:
2106
	if (ctl_buf != ctl)
L
Linus Torvalds 已提交
2107 2108 2109
		sock_kfree_s(sock->sk, ctl_buf, ctl_len);
out_freeiov:
	if (iov != iovstack)
E
Eric Dumazet 已提交
2110
		kfree(iov);
2111 2112 2113 2114 2115 2116 2117
	return err;
}

/*
 *	BSD sendmsg interface
 */

2118
long __sys_sendmsg(int fd, struct user_msghdr __user *msg, unsigned flags)
2119 2120 2121
{
	int fput_needed, err;
	struct msghdr msg_sys;
2122 2123 2124
	struct socket *sock;

	sock = sockfd_lookup_light(fd, &err, &fput_needed);
2125 2126 2127
	if (!sock)
		goto out;

2128
	err = ___sys_sendmsg(sock, msg, &msg_sys, flags, NULL);
2129

2130
	fput_light(sock->file, fput_needed);
2131
out:
L
Linus Torvalds 已提交
2132 2133 2134
	return err;
}

2135
SYSCALL_DEFINE3(sendmsg, int, fd, struct user_msghdr __user *, msg, unsigned int, flags)
2136 2137 2138 2139 2140 2141
{
	if (flags & MSG_CMSG_COMPAT)
		return -EINVAL;
	return __sys_sendmsg(fd, msg, flags);
}

2142 2143 2144 2145 2146 2147 2148 2149 2150 2151 2152 2153
/*
 *	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;
2154
	struct used_address used_address;
2155

2156 2157
	if (vlen > UIO_MAXIOV)
		vlen = UIO_MAXIOV;
2158 2159 2160 2161 2162 2163 2164

	datagrams = 0;

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

2165
	used_address.name_len = UINT_MAX;
2166 2167
	entry = mmsg;
	compat_entry = (struct compat_mmsghdr __user *)mmsg;
2168
	err = 0;
2169 2170 2171

	while (datagrams < vlen) {
		if (MSG_CMSG_COMPAT & flags) {
2172
			err = ___sys_sendmsg(sock, (struct user_msghdr __user *)compat_entry,
2173
					     &msg_sys, flags, &used_address);
2174 2175 2176 2177 2178
			if (err < 0)
				break;
			err = __put_user(err, &compat_entry->msg_len);
			++compat_entry;
		} else {
2179
			err = ___sys_sendmsg(sock,
2180
					     (struct user_msghdr __user *)entry,
2181
					     &msg_sys, flags, &used_address);
2182 2183 2184 2185 2186 2187 2188 2189 2190 2191 2192 2193 2194
			if (err < 0)
				break;
			err = put_user(err, &entry->msg_len);
			++entry;
		}

		if (err)
			break;
		++datagrams;
	}

	fput_light(sock->file, fput_needed);

2195 2196
	/* We only return an error if no datagrams were able to be sent */
	if (datagrams != 0)
2197 2198 2199 2200 2201 2202 2203 2204
		return datagrams;

	return err;
}

SYSCALL_DEFINE4(sendmmsg, int, fd, struct mmsghdr __user *, mmsg,
		unsigned int, vlen, unsigned int, flags)
{
2205 2206
	if (flags & MSG_CMSG_COMPAT)
		return -EINVAL;
2207 2208 2209
	return __sys_sendmmsg(fd, mmsg, vlen, flags);
}

2210
static int ___sys_recvmsg(struct socket *sock, struct user_msghdr __user *msg,
2211
			 struct msghdr *msg_sys, unsigned int flags, int nosec)
L
Linus Torvalds 已提交
2212
{
2213 2214
	struct compat_msghdr __user *msg_compat =
	    (struct compat_msghdr __user *)msg;
L
Linus Torvalds 已提交
2215
	struct iovec iovstack[UIO_FASTIOV];
2216
	struct iovec *iov = iovstack;
L
Linus Torvalds 已提交
2217
	unsigned long cmsg_ptr;
E
Eric Dumazet 已提交
2218
	int err, total_len, len;
L
Linus Torvalds 已提交
2219 2220

	/* kernel mode address */
2221
	struct sockaddr_storage addr;
L
Linus Torvalds 已提交
2222 2223 2224 2225

	/* user mode address pointers */
	struct sockaddr __user *uaddr;
	int __user *uaddr_len;
2226

L
Linus Torvalds 已提交
2227
	if (MSG_CMSG_COMPAT & flags) {
2228
		if (get_compat_msghdr(msg_sys, msg_compat))
L
Linus Torvalds 已提交
2229
			return -EFAULT;
2230 2231 2232 2233 2234
	} else {
		err = copy_msghdr_from_user(msg_sys, msg);
		if (err)
			return err;
	}
L
Linus Torvalds 已提交
2235

2236 2237
	/* Save the user-mode address (verify_iovec will change the
	 * kernel msghdr to use the kernel address space)
L
Linus Torvalds 已提交
2238
	 */
2239
	uaddr = (__force void __user *)msg_sys->msg_name;
L
Linus Torvalds 已提交
2240
	uaddr_len = COMPAT_NAMELEN(msg);
2241
	if (MSG_CMSG_COMPAT & flags)
2242
		err = verify_compat_iovec(msg_sys, iovstack, &addr, READ);
2243
	else
2244
		err = verify_iovec(msg_sys, iovstack, &addr, READ);
L
Linus Torvalds 已提交
2245 2246
	if (err < 0)
		goto out_freeiov;
2247
	iov = msg_sys->msg_iov;
2248
	total_len = err;
L
Linus Torvalds 已提交
2249

2250 2251
	cmsg_ptr = (unsigned long)msg_sys->msg_control;
	msg_sys->msg_flags = flags & (MSG_CMSG_CLOEXEC|MSG_CMSG_COMPAT);
2252

2253 2254 2255
	/* We assume all kernel code knows the size of sockaddr_storage */
	msg_sys->msg_namelen = 0;

L
Linus Torvalds 已提交
2256 2257
	if (sock->file->f_flags & O_NONBLOCK)
		flags |= MSG_DONTWAIT;
2258 2259
	err = (nosec ? sock_recvmsg_nosec : sock_recvmsg)(sock, msg_sys,
							  total_len, flags);
L
Linus Torvalds 已提交
2260 2261 2262 2263 2264
	if (err < 0)
		goto out_freeiov;
	len = err;

	if (uaddr != NULL) {
2265
		err = move_addr_to_user(&addr,
2266
					msg_sys->msg_namelen, uaddr,
2267
					uaddr_len);
L
Linus Torvalds 已提交
2268 2269 2270
		if (err < 0)
			goto out_freeiov;
	}
2271
	err = __put_user((msg_sys->msg_flags & ~MSG_CMSG_COMPAT),
2272
			 COMPAT_FLAGS(msg));
L
Linus Torvalds 已提交
2273 2274 2275
	if (err)
		goto out_freeiov;
	if (MSG_CMSG_COMPAT & flags)
2276
		err = __put_user((unsigned long)msg_sys->msg_control - cmsg_ptr,
L
Linus Torvalds 已提交
2277 2278
				 &msg_compat->msg_controllen);
	else
2279
		err = __put_user((unsigned long)msg_sys->msg_control - cmsg_ptr,
L
Linus Torvalds 已提交
2280 2281 2282 2283 2284 2285 2286
				 &msg->msg_controllen);
	if (err)
		goto out_freeiov;
	err = len;

out_freeiov:
	if (iov != iovstack)
E
Eric Dumazet 已提交
2287
		kfree(iov);
2288 2289 2290 2291 2292 2293 2294
	return err;
}

/*
 *	BSD recvmsg interface
 */

2295
long __sys_recvmsg(int fd, struct user_msghdr __user *msg, unsigned flags)
2296 2297 2298
{
	int fput_needed, err;
	struct msghdr msg_sys;
2299 2300 2301
	struct socket *sock;

	sock = sockfd_lookup_light(fd, &err, &fput_needed);
2302 2303 2304
	if (!sock)
		goto out;

2305
	err = ___sys_recvmsg(sock, msg, &msg_sys, flags, 0);
2306

2307
	fput_light(sock->file, fput_needed);
L
Linus Torvalds 已提交
2308 2309 2310 2311
out:
	return err;
}

2312
SYSCALL_DEFINE3(recvmsg, int, fd, struct user_msghdr __user *, msg,
2313 2314 2315 2316 2317 2318 2319
		unsigned int, flags)
{
	if (flags & MSG_CMSG_COMPAT)
		return -EINVAL;
	return __sys_recvmsg(fd, msg, flags);
}

2320 2321 2322 2323 2324 2325 2326 2327 2328 2329
/*
 *     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;
2330
	struct compat_mmsghdr __user *compat_entry;
2331 2332 2333 2334 2335 2336 2337 2338 2339 2340 2341 2342 2343 2344 2345 2346 2347 2348 2349
	struct msghdr msg_sys;
	struct timespec end_time;

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

	datagrams = 0;

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

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

	entry = mmsg;
2350
	compat_entry = (struct compat_mmsghdr __user *)mmsg;
2351 2352 2353 2354 2355

	while (datagrams < vlen) {
		/*
		 * No need to ask LSM for more than the first datagram.
		 */
2356
		if (MSG_CMSG_COMPAT & flags) {
2357
			err = ___sys_recvmsg(sock, (struct user_msghdr __user *)compat_entry,
2358 2359
					     &msg_sys, flags & ~MSG_WAITFORONE,
					     datagrams);
2360 2361 2362 2363 2364
			if (err < 0)
				break;
			err = __put_user(err, &compat_entry->msg_len);
			++compat_entry;
		} else {
2365
			err = ___sys_recvmsg(sock,
2366
					     (struct user_msghdr __user *)entry,
2367 2368
					     &msg_sys, flags & ~MSG_WAITFORONE,
					     datagrams);
2369 2370 2371 2372 2373 2374
			if (err < 0)
				break;
			err = put_user(err, &entry->msg_len);
			++entry;
		}

2375 2376 2377 2378
		if (err)
			break;
		++datagrams;

2379 2380 2381 2382
		/* MSG_WAITFORONE turns on MSG_DONTWAIT after one packet */
		if (flags & MSG_WAITFORONE)
			flags |= MSG_DONTWAIT;

2383 2384 2385 2386 2387 2388 2389 2390 2391 2392 2393 2394 2395 2396 2397 2398 2399 2400 2401 2402
		if (timeout) {
			ktime_get_ts(timeout);
			*timeout = timespec_sub(end_time, *timeout);
			if (timeout->tv_sec < 0) {
				timeout->tv_sec = timeout->tv_nsec = 0;
				break;
			}

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

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

out_put:
	fput_light(sock->file, fput_needed);
L
Linus Torvalds 已提交
2403

2404 2405 2406 2407 2408 2409 2410 2411 2412 2413 2414 2415 2416 2417 2418 2419 2420 2421 2422 2423 2424 2425 2426 2427 2428 2429 2430 2431 2432 2433 2434
	if (err == 0)
		return datagrams;

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

		return datagrams;
	}

	return err;
}

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

2435 2436 2437
	if (flags & MSG_CMSG_COMPAT)
		return -EINVAL;

2438 2439 2440 2441 2442 2443 2444 2445 2446 2447 2448 2449 2450 2451 2452 2453
	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
2454 2455
/* Argument list sizes for sys_socketcall */
#define AL(x) ((x) * sizeof(unsigned long))
2456
static const unsigned char nargs[21] = {
2457 2458 2459
	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),
2460
	AL(4), AL(5), AL(4)
2461 2462
};

L
Linus Torvalds 已提交
2463 2464 2465
#undef AL

/*
2466
 *	System call vectors.
L
Linus Torvalds 已提交
2467 2468 2469
 *
 *	Argument checking cleaned up. Saved 20% in size.
 *  This function doesn't need to set the kernel lock because
2470
 *  it is set by the callees.
L
Linus Torvalds 已提交
2471 2472
 */

2473
SYSCALL_DEFINE2(socketcall, int, call, unsigned long __user *, args)
L
Linus Torvalds 已提交
2474
{
2475
	unsigned long a[AUDITSC_ARGS];
2476
	unsigned long a0, a1;
L
Linus Torvalds 已提交
2477
	int err;
2478
	unsigned int len;
L
Linus Torvalds 已提交
2479

2480
	if (call < 1 || call > SYS_SENDMMSG)
L
Linus Torvalds 已提交
2481 2482
		return -EINVAL;

2483 2484 2485 2486
	len = nargs[call];
	if (len > sizeof(a))
		return -EINVAL;

L
Linus Torvalds 已提交
2487
	/* copy_from_user should be SMP safe. */
2488
	if (copy_from_user(a, args, len))
L
Linus Torvalds 已提交
2489
		return -EFAULT;
2490

2491 2492 2493
	err = audit_socketcall(nargs[call] / sizeof(unsigned long), a);
	if (err)
		return err;
2494

2495 2496 2497 2498 2499 2500 2501 2502 2503 2504 2505 2506 2507 2508 2509 2510 2511
	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
2512 2513
		err = sys_accept4(a0, (struct sockaddr __user *)a1,
				  (int __user *)a[2], 0);
2514 2515 2516 2517 2518 2519 2520 2521 2522 2523 2524 2525 2526 2527 2528 2529 2530 2531 2532 2533 2534 2535 2536 2537 2538 2539 2540 2541 2542 2543 2544 2545 2546 2547 2548 2549 2550 2551 2552 2553 2554
		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:
2555
		err = sys_sendmsg(a0, (struct user_msghdr __user *)a1, a[2]);
2556
		break;
2557 2558 2559
	case SYS_SENDMMSG:
		err = sys_sendmmsg(a0, (struct mmsghdr __user *)a1, a[2], a[3]);
		break;
2560
	case SYS_RECVMSG:
2561
		err = sys_recvmsg(a0, (struct user_msghdr __user *)a1, a[2]);
2562
		break;
2563 2564 2565 2566
	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
2567 2568 2569
	case SYS_ACCEPT4:
		err = sys_accept4(a0, (struct sockaddr __user *)a1,
				  (int __user *)a[2], a[3]);
U
Ulrich Drepper 已提交
2570
		break;
2571 2572 2573
	default:
		err = -EINVAL;
		break;
L
Linus Torvalds 已提交
2574 2575 2576 2577
	}
	return err;
}

2578
#endif				/* __ARCH_WANT_SYS_SOCKETCALL */
L
Linus Torvalds 已提交
2579

2580 2581 2582 2583
/**
 *	sock_register - add a socket protocol handler
 *	@ops: description of protocol
 *
L
Linus Torvalds 已提交
2584 2585
 *	This function is called by a protocol handler that wants to
 *	advertise its address family, and have it linked into the
2586
 *	socket interface. The value ops->family corresponds to the
2587
 *	socket system call protocol family.
L
Linus Torvalds 已提交
2588
 */
2589
int sock_register(const struct net_proto_family *ops)
L
Linus Torvalds 已提交
2590 2591 2592 2593
{
	int err;

	if (ops->family >= NPROTO) {
2594
		pr_crit("protocol %d >= NPROTO(%d)\n", ops->family, NPROTO);
L
Linus Torvalds 已提交
2595 2596
		return -ENOBUFS;
	}
2597 2598

	spin_lock(&net_family_lock);
E
Eric Dumazet 已提交
2599 2600
	if (rcu_dereference_protected(net_families[ops->family],
				      lockdep_is_held(&net_family_lock)))
2601 2602
		err = -EEXIST;
	else {
2603
		rcu_assign_pointer(net_families[ops->family], ops);
L
Linus Torvalds 已提交
2604 2605
		err = 0;
	}
2606 2607
	spin_unlock(&net_family_lock);

2608
	pr_info("NET: Registered protocol family %d\n", ops->family);
L
Linus Torvalds 已提交
2609 2610
	return err;
}
2611
EXPORT_SYMBOL(sock_register);
L
Linus Torvalds 已提交
2612

2613 2614 2615 2616
/**
 *	sock_unregister - remove a protocol handler
 *	@family: protocol family to remove
 *
L
Linus Torvalds 已提交
2617 2618
 *	This function is called by a protocol handler that wants to
 *	remove its address family, and have it unlinked from the
2619 2620 2621 2622 2623 2624
 *	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
2625
 */
2626
void sock_unregister(int family)
L
Linus Torvalds 已提交
2627
{
2628
	BUG_ON(family < 0 || family >= NPROTO);
L
Linus Torvalds 已提交
2629

2630
	spin_lock(&net_family_lock);
2631
	RCU_INIT_POINTER(net_families[family], NULL);
2632 2633 2634 2635
	spin_unlock(&net_family_lock);

	synchronize_rcu();

2636
	pr_info("NET: Unregistered protocol family %d\n", family);
L
Linus Torvalds 已提交
2637
}
2638
EXPORT_SYMBOL(sock_unregister);
L
Linus Torvalds 已提交
2639

2640
static int __init sock_init(void)
L
Linus Torvalds 已提交
2641
{
N
Nick Piggin 已提交
2642
	int err;
2643 2644 2645 2646 2647 2648
	/*
	 *      Initialize the network sysctl infrastructure.
	 */
	err = net_sysctl_init();
	if (err)
		goto out;
N
Nick Piggin 已提交
2649

L
Linus Torvalds 已提交
2650
	/*
2651
	 *      Initialize skbuff SLAB cache
L
Linus Torvalds 已提交
2652 2653 2654 2655
	 */
	skb_init();

	/*
2656
	 *      Initialize the protocols module.
L
Linus Torvalds 已提交
2657 2658 2659
	 */

	init_inodecache();
N
Nick Piggin 已提交
2660 2661 2662 2663

	err = register_filesystem(&sock_fs_type);
	if (err)
		goto out_fs;
L
Linus Torvalds 已提交
2664
	sock_mnt = kern_mount(&sock_fs_type);
N
Nick Piggin 已提交
2665 2666 2667 2668
	if (IS_ERR(sock_mnt)) {
		err = PTR_ERR(sock_mnt);
		goto out_mount;
	}
2669 2670

	/* The real protocol initialization is performed in later initcalls.
L
Linus Torvalds 已提交
2671 2672 2673
	 */

#ifdef CONFIG_NETFILTER
2674 2675 2676
	err = netfilter_init();
	if (err)
		goto out;
L
Linus Torvalds 已提交
2677
#endif
2678

2679
	ptp_classifier_init();
2680

N
Nick Piggin 已提交
2681 2682 2683 2684 2685 2686 2687
out:
	return err;

out_mount:
	unregister_filesystem(&sock_fs_type);
out_fs:
	goto out;
L
Linus Torvalds 已提交
2688 2689
}

2690 2691
core_initcall(sock_init);	/* early initcall */

L
Linus Torvalds 已提交
2692 2693 2694 2695 2696 2697
#ifdef CONFIG_PROC_FS
void socket_seq_show(struct seq_file *seq)
{
	int cpu;
	int counter = 0;

2698
	for_each_possible_cpu(cpu)
2699
	    counter += per_cpu(sockets_in_use, cpu);
L
Linus Torvalds 已提交
2700 2701 2702 2703 2704 2705 2706

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

	seq_printf(seq, "sockets: used %d\n", counter);
}
2707
#endif				/* CONFIG_PROC_FS */
L
Linus Torvalds 已提交
2708

2709
#ifdef CONFIG_COMPAT
2710
static int do_siocgstamp(struct net *net, struct socket *sock,
2711
			 unsigned int cmd, void __user *up)
2712 2713 2714 2715 2716 2717
{
	mm_segment_t old_fs = get_fs();
	struct timeval ktv;
	int err;

	set_fs(KERNEL_DS);
2718
	err = sock_do_ioctl(net, sock, cmd, (unsigned long)&ktv);
2719
	set_fs(old_fs);
2720
	if (!err)
2721
		err = compat_put_timeval(&ktv, up);
2722

2723 2724 2725
	return err;
}

2726
static int do_siocgstampns(struct net *net, struct socket *sock,
2727
			   unsigned int cmd, void __user *up)
2728 2729 2730 2731 2732 2733
{
	mm_segment_t old_fs = get_fs();
	struct timespec kts;
	int err;

	set_fs(KERNEL_DS);
2734
	err = sock_do_ioctl(net, sock, cmd, (unsigned long)&kts);
2735
	set_fs(old_fs);
2736
	if (!err)
2737
		err = compat_put_timespec(&kts, up);
2738

2739 2740 2741
	return err;
}

2742
static int dev_ifname32(struct net *net, struct compat_ifreq __user *uifr32)
2743 2744 2745 2746 2747
{
	struct ifreq __user *uifr;
	int err;

	uifr = compat_alloc_user_space(sizeof(struct ifreq));
2748
	if (copy_in_user(uifr, uifr32, sizeof(struct compat_ifreq)))
2749 2750
		return -EFAULT;

2751
	err = dev_ioctl(net, SIOCGIFNAME, uifr);
2752 2753 2754
	if (err)
		return err;

2755
	if (copy_in_user(uifr32, uifr, sizeof(struct compat_ifreq)))
2756 2757 2758 2759 2760
		return -EFAULT;

	return 0;
}

2761
static int dev_ifconf(struct net *net, struct compat_ifconf __user *uifc32)
2762
{
2763
	struct compat_ifconf ifc32;
2764 2765
	struct ifconf ifc;
	struct ifconf __user *uifc;
2766
	struct compat_ifreq __user *ifr32;
2767 2768 2769 2770
	struct ifreq __user *ifr;
	unsigned int i, j;
	int err;

2771
	if (copy_from_user(&ifc32, uifc32, sizeof(struct compat_ifconf)))
2772 2773
		return -EFAULT;

2774
	memset(&ifc, 0, sizeof(ifc));
2775 2776 2777 2778 2779 2780
	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 {
2781 2782
		size_t len = ((ifc32.ifc_len / sizeof(struct compat_ifreq)) + 1) *
			sizeof(struct ifreq);
2783 2784 2785 2786
		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);
2787
		for (i = 0; i < ifc32.ifc_len; i += sizeof(struct compat_ifreq)) {
2788
			if (copy_in_user(ifr, ifr32, sizeof(struct compat_ifreq)))
2789 2790 2791 2792 2793 2794 2795 2796
				return -EFAULT;
			ifr++;
			ifr32++;
		}
	}
	if (copy_to_user(uifc, &ifc, sizeof(struct ifconf)))
		return -EFAULT;

2797
	err = dev_ioctl(net, SIOCGIFCONF, uifc);
2798 2799 2800 2801 2802 2803 2804 2805 2806
	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;
2807 2808 2809
	     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)))
2810 2811 2812 2813 2814 2815 2816 2817 2818 2819
			return -EFAULT;
		ifr32++;
		ifr++;
	}

	if (ifc32.ifcbuf == 0) {
		/* Translate from 64-bit structure multiple to
		 * a 32-bit one.
		 */
		i = ifc.ifc_len;
2820
		i = ((i / sizeof(struct ifreq)) * sizeof(struct compat_ifreq));
2821 2822 2823 2824
		ifc32.ifc_len = i;
	} else {
		ifc32.ifc_len = i;
	}
2825
	if (copy_to_user(uifc32, &ifc32, sizeof(struct compat_ifconf)))
2826 2827 2828 2829 2830
		return -EFAULT;

	return 0;
}

2831
static int ethtool_ioctl(struct net *net, struct compat_ifreq __user *ifr32)
2832
{
2833 2834 2835 2836
	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;
2837
	struct ifreq __user *ifr;
2838 2839
	u32 rule_cnt = 0, actual_rule_cnt;
	u32 ethcmd;
2840
	u32 data;
2841
	int ret;
2842

2843 2844
	if (get_user(data, &ifr32->ifr_ifru.ifru_data))
		return -EFAULT;
2845

2846 2847 2848
	compat_rxnfc = compat_ptr(data);

	if (get_user(ethcmd, &compat_rxnfc->cmd))
2849 2850
		return -EFAULT;

2851 2852 2853 2854 2855 2856 2857 2858 2859 2860 2861 2862 2863 2864 2865 2866 2867
	/* 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:
2868
	case ETHTOOL_SRXCLSRLINS:
2869 2870 2871 2872 2873 2874 2875 2876 2877
		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
2878
	rxnfc = (void __user *)ifr + ALIGN(sizeof(struct ifreq), 8);
2879 2880

	if (copy_in_user(&ifr->ifr_name, &ifr32->ifr_name, IFNAMSIZ))
2881 2882
		return -EFAULT;

2883 2884
	if (put_user(convert_in ? rxnfc : compat_ptr(data),
		     &ifr->ifr_ifru.ifru_data))
2885 2886
		return -EFAULT;

2887
	if (convert_in) {
2888
		/* We expect there to be holes between fs.m_ext and
2889 2890
		 * fs.ring_cookie and at the end of fs, but nowhere else.
		 */
2891 2892 2893 2894
		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));
2895 2896 2897 2898 2899 2900 2901
		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
2902 2903
				 (void __user *)(&rxnfc->fs.m_ext + 1) -
				 (void __user *)rxnfc) ||
2904 2905
		    copy_in_user(&rxnfc->fs.ring_cookie,
				 &compat_rxnfc->fs.ring_cookie,
S
Stephen Hemminger 已提交
2906 2907
				 (void __user *)(&rxnfc->fs.location + 1) -
				 (void __user *)&rxnfc->fs.ring_cookie) ||
2908 2909 2910 2911 2912 2913 2914 2915 2916 2917 2918
		    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
2919 2920
				 (const void __user *)(&rxnfc->fs.m_ext + 1) -
				 (const void __user *)rxnfc) ||
2921 2922
		    copy_in_user(&compat_rxnfc->fs.ring_cookie,
				 &rxnfc->fs.ring_cookie,
S
Stephen Hemminger 已提交
2923 2924
				 (const void __user *)(&rxnfc->fs.location + 1) -
				 (const void __user *)&rxnfc->fs.ring_cookie) ||
2925 2926 2927 2928 2929 2930 2931 2932 2933 2934 2935 2936 2937 2938 2939 2940 2941 2942 2943 2944 2945 2946 2947 2948 2949
		    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;
2950 2951
}

2952 2953 2954 2955 2956 2957
static int compat_siocwandev(struct net *net, struct compat_ifreq __user *uifr32)
{
	void __user *uptr;
	compat_uptr_t uptr32;
	struct ifreq __user *uifr;

2958
	uifr = compat_alloc_user_space(sizeof(*uifr));
2959 2960 2961 2962 2963 2964 2965 2966 2967 2968 2969 2970 2971 2972
	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);
}

2973 2974
static int bond_ioctl(struct net *net, unsigned int cmd,
			 struct compat_ifreq __user *ifr32)
2975 2976 2977 2978 2979 2980 2981 2982 2983 2984
{
	struct ifreq kifr;
	mm_segment_t old_fs;
	int err;

	switch (cmd) {
	case SIOCBONDENSLAVE:
	case SIOCBONDRELEASE:
	case SIOCBONDSETHWADDR:
	case SIOCBONDCHANGEACTIVE:
2985
		if (copy_from_user(&kifr, ifr32, sizeof(struct compat_ifreq)))
2986 2987 2988
			return -EFAULT;

		old_fs = get_fs();
2989
		set_fs(KERNEL_DS);
2990 2991
		err = dev_ioctl(net, cmd,
				(struct ifreq __user __force *) &kifr);
2992
		set_fs(old_fs);
2993 2994 2995

		return err;
	default:
2996
		return -ENOIOCTLCMD;
2997
	}
2998 2999
}

3000 3001
/* 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,
3002
				 struct compat_ifreq __user *u_ifreq32)
3003 3004 3005 3006 3007 3008 3009 3010 3011
{
	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;
3012
	if (get_user(data32, &u_ifreq32->ifr_ifru.ifru_data))
3013 3014 3015 3016 3017 3018 3019 3020
		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;
3021
	if (put_user(data64, &u_ifreq64->ifr_ifru.ifru_data))
3022 3023
		return -EFAULT;

3024
	return dev_ioctl(net, cmd, u_ifreq64);
3025 3026
}

3027 3028
static int dev_ifsioc(struct net *net, struct socket *sock,
			 unsigned int cmd, struct compat_ifreq __user *uifr32)
3029
{
3030
	struct ifreq __user *uifr;
3031 3032
	int err;

3033 3034 3035 3036 3037 3038
	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);

3039 3040 3041 3042 3043 3044 3045 3046 3047 3048 3049 3050
	if (!err) {
		switch (cmd) {
		case SIOCGIFFLAGS:
		case SIOCGIFMETRIC:
		case SIOCGIFMTU:
		case SIOCGIFMEM:
		case SIOCGIFHWADDR:
		case SIOCGIFINDEX:
		case SIOCGIFADDR:
		case SIOCGIFBRDADDR:
		case SIOCGIFDSTADDR:
		case SIOCGIFNETMASK:
3051
		case SIOCGIFPFLAGS:
3052
		case SIOCGIFTXQLEN:
3053 3054
		case SIOCGMIIPHY:
		case SIOCGMIIREG:
3055
			if (copy_in_user(uifr32, uifr, sizeof(*uifr32)))
3056 3057 3058 3059 3060 3061 3062
				err = -EFAULT;
			break;
		}
	}
	return err;
}

3063 3064 3065 3066 3067 3068 3069 3070 3071 3072
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));
3073 3074 3075 3076 3077 3078
	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);
3079 3080 3081 3082
	if (err)
		return -EFAULT;

	old_fs = get_fs();
3083
	set_fs(KERNEL_DS);
3084
	err = dev_ioctl(net, cmd, (void  __user __force *)&ifr);
3085
	set_fs(old_fs);
3086 3087 3088

	if (cmd == SIOCGIFMAP && !err) {
		err = copy_to_user(uifr32, &ifr, sizeof(ifr.ifr_name));
3089 3090 3091 3092 3093 3094
		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);
3095 3096 3097 3098 3099 3100
		if (err)
			err = -EFAULT;
	}
	return err;
}

3101
struct rtentry32 {
3102
	u32		rt_pad1;
3103 3104 3105
	struct sockaddr rt_dst;         /* target address               */
	struct sockaddr rt_gateway;     /* gateway addr (RTF_GATEWAY)   */
	struct sockaddr rt_genmask;     /* target network mask (IP)     */
3106 3107 3108 3109 3110 3111 3112
	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! */
3113
	/* char * */ u32 rt_dev;        /* forcing the device at add    */
3114 3115
	u32		rt_mtu;         /* per route MTU/Window         */
	u32		rt_window;      /* Window clamping              */
3116 3117 3118 3119 3120 3121 3122 3123 3124 3125 3126 3127 3128 3129 3130 3131
	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;
};

3132 3133
static int routing_ioctl(struct net *net, struct socket *sock,
			 unsigned int cmd, void __user *argp)
3134 3135 3136 3137 3138 3139 3140 3141 3142
{
	int ret;
	void *r = NULL;
	struct in6_rtmsg r6;
	struct rtentry r4;
	char devname[16];
	u32 rtdev;
	mm_segment_t old_fs = get_fs();

3143 3144
	if (sock && sock->sk && sock->sk->sk_family == AF_INET6) { /* ipv6 */
		struct in6_rtmsg32 __user *ur6 = argp;
3145
		ret = copy_from_user(&r6.rtmsg_dst, &(ur6->rtmsg_dst),
3146
			3 * sizeof(struct in6_addr));
3147 3148 3149 3150 3151 3152 3153
		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));
3154 3155 3156

		r = (void *) &r6;
	} else { /* ipv4 */
3157
		struct rtentry32 __user *ur4 = argp;
3158
		ret = copy_from_user(&r4.rt_dst, &(ur4->rt_dst),
3159
					3 * sizeof(struct sockaddr));
3160 3161 3162 3163 3164 3165
		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));
3166
		if (rtdev) {
3167
			ret |= copy_from_user(devname, compat_ptr(rtdev), 15);
3168 3169
			r4.rt_dev = (char __user __force *)devname;
			devname[15] = 0;
3170 3171 3172 3173 3174 3175 3176 3177 3178 3179 3180
		} else
			r4.rt_dev = NULL;

		r = (void *) &r4;
	}

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

3181
	set_fs(KERNEL_DS);
3182
	ret = sock_do_ioctl(net, sock, cmd, (unsigned long) r);
3183
	set_fs(old_fs);
3184 3185 3186 3187 3188 3189 3190

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
3191
 * use compatible ioctls
3192
 */
3193
static int old_bridge_ioctl(compat_ulong_t __user *argp)
3194
{
3195
	compat_ulong_t tmp;
3196

3197
	if (get_user(tmp, argp))
3198 3199 3200 3201 3202 3203
		return -EFAULT;
	if (tmp == BRCTL_GET_VERSION)
		return BRCTL_VERSION + 1;
	return -EINVAL;
}

3204 3205 3206 3207 3208 3209
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);
3210

3211
	if (cmd >= SIOCDEVPRIVATE && cmd <= (SIOCDEVPRIVATE + 15))
3212
		return compat_ifr_data_ioctl(net, cmd, argp);
3213 3214 3215 3216 3217 3218 3219 3220 3221 3222 3223

	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);
3224 3225
	case SIOCWANDEV:
		return compat_siocwandev(net, argp);
3226 3227 3228
	case SIOCGIFMAP:
	case SIOCSIFMAP:
		return compat_sioc_ifmap(net, cmd, argp);
3229 3230 3231 3232 3233 3234 3235 3236 3237 3238 3239 3240
	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);
3241 3242
	case SIOCBONDSLAVEINFOQUERY:
	case SIOCBONDINFOQUERY:
3243
	case SIOCSHWTSTAMP:
3244
	case SIOCGHWTSTAMP:
3245
		return compat_ifr_data_ioctl(net, cmd, argp);
3246 3247 3248 3249 3250 3251 3252 3253 3254 3255 3256 3257 3258 3259 3260 3261 3262 3263 3264 3265 3266 3267 3268 3269 3270 3271 3272 3273 3274 3275 3276 3277 3278 3279 3280 3281 3282 3283 3284 3285 3286 3287

	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:
3288 3289 3290 3291
	case SIOCSIFNAME:
	case SIOCGMIIPHY:
	case SIOCGMIIREG:
	case SIOCSMIIREG:
3292
		return dev_ifsioc(net, sock, cmd, argp);
3293

3294 3295 3296 3297
	case SIOCSARP:
	case SIOCGARP:
	case SIOCDARP:
	case SIOCATMARK:
3298 3299 3300
		return sock_do_ioctl(net, sock, cmd, arg);
	}

3301 3302
	return -ENOIOCTLCMD;
}
3303

3304
static long compat_sock_ioctl(struct file *file, unsigned int cmd,
3305
			      unsigned long arg)
3306 3307 3308
{
	struct socket *sock = file->private_data;
	int ret = -ENOIOCTLCMD;
3309 3310 3311 3312 3313
	struct sock *sk;
	struct net *net;

	sk = sock->sk;
	net = sock_net(sk);
3314 3315 3316 3317

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

3318 3319 3320 3321
	if (ret == -ENOIOCTLCMD &&
	    (cmd >= SIOCIWFIRST && cmd <= SIOCIWLAST))
		ret = compat_wext_handle_ioctl(net, cmd, arg);

3322 3323 3324
	if (ret == -ENOIOCTLCMD)
		ret = compat_sock_ioctl_trans(file, sock, cmd, arg);

3325 3326 3327 3328
	return ret;
}
#endif

3329 3330 3331 3332
int kernel_bind(struct socket *sock, struct sockaddr *addr, int addrlen)
{
	return sock->ops->bind(sock, addr, addrlen);
}
3333
EXPORT_SYMBOL(kernel_bind);
3334 3335 3336 3337 3338

int kernel_listen(struct socket *sock, int backlog)
{
	return sock->ops->listen(sock, backlog);
}
3339
EXPORT_SYMBOL(kernel_listen);
3340 3341 3342 3343 3344 3345 3346 3347 3348 3349 3350 3351 3352 3353

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);
3354
		*newsock = NULL;
3355 3356 3357 3358
		goto done;
	}

	(*newsock)->ops = sock->ops;
3359
	__module_get((*newsock)->ops->owner);
3360 3361 3362 3363

done:
	return err;
}
3364
EXPORT_SYMBOL(kernel_accept);
3365 3366

int kernel_connect(struct socket *sock, struct sockaddr *addr, int addrlen,
3367
		   int flags)
3368 3369 3370
{
	return sock->ops->connect(sock, addr, addrlen, flags);
}
3371
EXPORT_SYMBOL(kernel_connect);
3372 3373 3374 3375 3376 3377

int kernel_getsockname(struct socket *sock, struct sockaddr *addr,
			 int *addrlen)
{
	return sock->ops->getname(sock, addr, addrlen, 0);
}
3378
EXPORT_SYMBOL(kernel_getsockname);
3379 3380 3381 3382 3383 3384

int kernel_getpeername(struct socket *sock, struct sockaddr *addr,
			 int *addrlen)
{
	return sock->ops->getname(sock, addr, addrlen, 1);
}
3385
EXPORT_SYMBOL(kernel_getpeername);
3386 3387 3388 3389 3390

int kernel_getsockopt(struct socket *sock, int level, int optname,
			char *optval, int *optlen)
{
	mm_segment_t oldfs = get_fs();
3391 3392
	char __user *uoptval;
	int __user *uoptlen;
3393 3394
	int err;

3395 3396 3397
	uoptval = (char __user __force *) optval;
	uoptlen = (int __user __force *) optlen;

3398 3399
	set_fs(KERNEL_DS);
	if (level == SOL_SOCKET)
3400
		err = sock_getsockopt(sock, level, optname, uoptval, uoptlen);
3401
	else
3402 3403
		err = sock->ops->getsockopt(sock, level, optname, uoptval,
					    uoptlen);
3404 3405 3406
	set_fs(oldfs);
	return err;
}
3407
EXPORT_SYMBOL(kernel_getsockopt);
3408 3409

int kernel_setsockopt(struct socket *sock, int level, int optname,
3410
			char *optval, unsigned int optlen)
3411 3412
{
	mm_segment_t oldfs = get_fs();
3413
	char __user *uoptval;
3414 3415
	int err;

3416 3417
	uoptval = (char __user __force *) optval;

3418 3419
	set_fs(KERNEL_DS);
	if (level == SOL_SOCKET)
3420
		err = sock_setsockopt(sock, level, optname, uoptval, optlen);
3421
	else
3422
		err = sock->ops->setsockopt(sock, level, optname, uoptval,
3423 3424 3425 3426
					    optlen);
	set_fs(oldfs);
	return err;
}
3427
EXPORT_SYMBOL(kernel_setsockopt);
3428 3429 3430 3431 3432 3433 3434 3435 3436

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);
}
3437
EXPORT_SYMBOL(kernel_sendpage);
3438 3439 3440 3441 3442 3443 3444 3445 3446 3447 3448 3449

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;
}
3450
EXPORT_SYMBOL(kernel_sock_ioctl);
3451

3452 3453 3454 3455 3456
int kernel_sock_shutdown(struct socket *sock, enum sock_shutdown_cmd how)
{
	return sock->ops->shutdown(sock, how);
}
EXPORT_SYMBOL(kernel_sock_shutdown);