socket.c 83.6 KB
Newer Older
L
Linus Torvalds 已提交
1 2 3 4 5 6
/*
 * NET		An implementation of the SOCKET network access protocol.
 *
 * Version:	@(#)socket.c	1.1.93	18/02/95
 *
 * Authors:	Orest Zborowski, <obz@Kodak.COM>
7
 *		Ross Biro
L
Linus Torvalds 已提交
8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44
 *		Fred N. van Kempen, <waltje@uWalt.NL.Mugnet.ORG>
 *
 * Fixes:
 *		Anonymous	:	NOTSOCK/BADF cleanup. Error fix in
 *					shutdown()
 *		Alan Cox	:	verify_area() fixes
 *		Alan Cox	:	Removed DDI
 *		Jonathan Kamens	:	SOCK_DGRAM reconnect bug
 *		Alan Cox	:	Moved a load of checks to the very
 *					top level.
 *		Alan Cox	:	Move address structures to/from user
 *					mode above the protocol layers.
 *		Rob Janssen	:	Allow 0 length sends.
 *		Alan Cox	:	Asynchronous I/O support (cribbed from the
 *					tty drivers).
 *		Niibe Yutaka	:	Asynchronous I/O for writes (4.4BSD style)
 *		Jeff Uphoff	:	Made max number of sockets command-line
 *					configurable.
 *		Matti Aarnio	:	Made the number of sockets dynamic,
 *					to be allocated when needed, and mr.
 *					Uphoff's max is used as max to be
 *					allowed to allocate.
 *		Linus		:	Argh. removed all the socket allocation
 *					altogether: it's in the inode now.
 *		Alan Cox	:	Made sock_alloc()/sock_release() public
 *					for NetROM and future kernel nfsd type
 *					stuff.
 *		Alan Cox	:	sendmsg/recvmsg basics.
 *		Tom Dyas	:	Export net symbols.
 *		Marcin Dalecki	:	Fixed problems with CONFIG_NET="n".
 *		Alan Cox	:	Added thread locking to sys_* calls
 *					for sockets. May have errors at the
 *					moment.
 *		Kevin Buhr	:	Fixed the dumb errors in the above.
 *		Andi Kleen	:	Some small cleanups, optimizations,
 *					and fixed a copy_from_user() bug.
 *		Tigran Aivazian	:	sys_send(args) calls sys_sendto(args, NULL, 0)
45
 *		Tigran Aivazian	:	Made listen(2) backlog sanity checks
L
Linus Torvalds 已提交
46 47 48 49 50 51 52 53 54 55
 *					protocol-independent
 *
 *
 *		This program is free software; you can redistribute it and/or
 *		modify it under the terms of the GNU General Public License
 *		as published by the Free Software Foundation; either version
 *		2 of the License, or (at your option) any later version.
 *
 *
 *	This module is effectively the top level interface to the BSD socket
56
 *	paradigm.
L
Linus Torvalds 已提交
57 58 59 60 61 62 63 64 65
 *
 *	Based upon Swansea University Computer Society NET3.039
 */

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

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

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

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

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

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

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
	if (sock_flag(sk, SOCK_WIFI_STATUS))
630 631 632
		flags |= SKBTX_WIFI_STATUS;

	*tx_flags = flags;
633 634 635
}
EXPORT_SYMBOL(sock_tx_timestamp);

636 637
static inline int __sock_sendmsg_nosec(struct kiocb *iocb, struct socket *sock,
				       struct msghdr *msg, size_t size)
L
Linus Torvalds 已提交
638 639 640 641 642 643 644 645 646 647 648
{
	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);
}

649 650 651 652 653 654 655 656
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
657 658 659 660 661 662 663 664 665 666 667 668 669
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;
}
670
EXPORT_SYMBOL(sock_sendmsg);
L
Linus Torvalds 已提交
671

672
static int sock_sendmsg_nosec(struct socket *sock, struct msghdr *msg, size_t size)
673 674 675 676 677 678 679 680 681 682 683 684 685
{
	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
686 687 688 689 690 691 692 693 694 695 696
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
	 */
697
	msg->msg_iov = (struct iovec *)vec;
L
Linus Torvalds 已提交
698 699 700 701 702
	msg->msg_iovlen = num;
	result = sock_sendmsg(sock, msg, size);
	set_fs(oldfs);
	return result;
}
703
EXPORT_SYMBOL(kernel_sendmsg);
L
Linus Torvalds 已提交
704

705 706 707 708 709 710
/*
 * 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)
{
711
	int need_software_tstamp = sock_flag(sk, SOCK_RCVTSTAMP);
712
	struct scm_timestamping tss;
713 714 715 716 717 718 719 720 721 722 723 724 725 726 727 728
	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 {
729 730
			struct timespec ts;
			skb_get_timestampns(skb, &ts);
731
			put_cmsg(msg, SOL_SOCKET, SCM_TIMESTAMPNS,
732
				 sizeof(ts), &ts);
733 734 735
		}
	}

736
	memset(&tss, 0, sizeof(tss));
737
	if ((sk->sk_tsflags & SOF_TIMESTAMPING_SOFTWARE ||
738 739
	     skb_shinfo(skb)->tx_flags & SKBTX_ANY_SW_TSTAMP) &&
	    ktime_to_timespec_cond(skb->tstamp, tss.ts + 0))
740
		empty = 0;
741
	if (shhwtstamps &&
742
	    (sk->sk_tsflags & SOF_TIMESTAMPING_RAW_HARDWARE) &&
743
	    ktime_to_timespec_cond(shhwtstamps->hwtstamp, tss.ts + 2))
744
		empty = 0;
745 746
	if (!empty)
		put_cmsg(msg, SOL_SOCKET,
747
			 SCM_TIMESTAMPING, sizeof(tss), &tss);
748
}
749 750
EXPORT_SYMBOL_GPL(__sock_recv_timestamp);

751 752 753 754 755 756 757 758 759 760 761 762 763 764 765 766
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
767 768
static inline void sock_recv_drops(struct msghdr *msg, struct sock *sk,
				   struct sk_buff *skb)
769 770 771 772 773 774
{
	if (sock_flag(sk, SOCK_RXQ_OVFL) && skb && skb->dropcount)
		put_cmsg(msg, SOL_SOCKET, SO_RXQ_OVFL,
			sizeof(__u32), &skb->dropcount);
}

775
void __sock_recv_ts_and_drops(struct msghdr *msg, struct sock *sk,
776 777 778 779 780
	struct sk_buff *skb)
{
	sock_recv_timestamp(msg, sk, skb);
	sock_recv_drops(msg, sk, skb);
}
781
EXPORT_SYMBOL_GPL(__sock_recv_ts_and_drops);
782

783 784
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
785 786 787 788 789 790 791 792 793 794 795 796
{
	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);
}

797 798 799 800 801 802 803 804
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);
}

805
int sock_recvmsg(struct socket *sock, struct msghdr *msg,
L
Linus Torvalds 已提交
806 807 808 809 810 811
		 size_t size, int flags)
{
	struct kiocb iocb;
	struct sock_iocb siocb;
	int ret;

812
	init_sync_kiocb(&iocb, NULL);
L
Linus Torvalds 已提交
813 814 815 816 817 818
	iocb.private = &siocb;
	ret = __sock_recvmsg(&iocb, sock, msg, size, flags);
	if (-EIOCBQUEUED == ret)
		ret = wait_on_sync_kiocb(&iocb);
	return ret;
}
819
EXPORT_SYMBOL(sock_recvmsg);
L
Linus Torvalds 已提交
820

821 822 823 824 825 826 827 828 829 830 831 832 833 834 835
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;
}

836 837 838 839 840 841 842 843 844 845 846 847 848 849 850
/**
 * 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.
 */
851 852
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
853 854 855 856 857 858 859 860 861
{
	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
	 */
862
	msg->msg_iov = (struct iovec *)vec, msg->msg_iovlen = num;
L
Linus Torvalds 已提交
863 864 865 866
	result = sock_recvmsg(sock, msg, size, flags);
	set_fs(oldfs);
	return result;
}
867
EXPORT_SYMBOL(kernel_recvmsg);
L
Linus Torvalds 已提交
868

869 870
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
871 872 873 874
{
	struct socket *sock;
	int flags;

875 876
	sock = file->private_data;

877 878 879
	flags = (file->f_flags & O_NONBLOCK) ? MSG_DONTWAIT : 0;
	/* more is a combination of MSG_MORE and MSG_SENDPAGE_NOTLAST */
	flags |= more;
880

881
	return kernel_sendpage(sock, page, offset, size, flags);
882
}
L
Linus Torvalds 已提交
883

J
Jens Axboe 已提交
884
static ssize_t sock_splice_read(struct file *file, loff_t *ppos,
885
				struct pipe_inode_info *pipe, size_t len,
J
Jens Axboe 已提交
886 887 888 889
				unsigned int flags)
{
	struct socket *sock = file->private_data;

890 891 892
	if (unlikely(!sock->ops->splice_read))
		return -EINVAL;

J
Jens Axboe 已提交
893 894 895
	return sock->ops->splice_read(sock, ppos, pipe, len, flags);
}

896
static struct sock_iocb *alloc_sock_iocb(struct kiocb *iocb,
897
					 struct sock_iocb *siocb)
898
{
K
Kent Overstreet 已提交
899 900
	if (!is_sync_kiocb(iocb))
		BUG();
L
Linus Torvalds 已提交
901

902 903 904
	siocb->kiocb = iocb;
	iocb->private = siocb;
	return siocb;
L
Linus Torvalds 已提交
905 906
}

907
static ssize_t do_sock_read(struct msghdr *msg, struct kiocb *iocb,
908 909
		struct file *file, const struct iovec *iov,
		unsigned long nr_segs)
910 911 912 913
{
	struct socket *sock = file->private_data;
	size_t size = 0;
	int i;
L
Linus Torvalds 已提交
914

915 916
	for (i = 0; i < nr_segs; i++)
		size += iov[i].iov_len;
L
Linus Torvalds 已提交
917

918 919 920 921
	msg->msg_name = NULL;
	msg->msg_namelen = 0;
	msg->msg_control = NULL;
	msg->msg_controllen = 0;
922
	msg->msg_iov = (struct iovec *)iov;
923 924 925 926 927 928
	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);
}

929 930
static ssize_t sock_aio_read(struct kiocb *iocb, const struct iovec *iov,
				unsigned long nr_segs, loff_t pos)
931 932 933
{
	struct sock_iocb siocb, *x;

L
Linus Torvalds 已提交
934 935
	if (pos != 0)
		return -ESPIPE;
936

K
Kent Overstreet 已提交
937
	if (iocb->ki_nbytes == 0)	/* Match SYS5 behaviour */
L
Linus Torvalds 已提交
938 939
		return 0;

940 941

	x = alloc_sock_iocb(iocb, &siocb);
942 943
	if (!x)
		return -ENOMEM;
944
	return do_sock_read(&x->async_msg, iocb, iocb->ki_filp, iov, nr_segs);
L
Linus Torvalds 已提交
945 946
}

947
static ssize_t do_sock_write(struct msghdr *msg, struct kiocb *iocb,
948 949
			struct file *file, const struct iovec *iov,
			unsigned long nr_segs)
L
Linus Torvalds 已提交
950
{
951 952 953
	struct socket *sock = file->private_data;
	size_t size = 0;
	int i;
L
Linus Torvalds 已提交
954

955 956
	for (i = 0; i < nr_segs; i++)
		size += iov[i].iov_len;
L
Linus Torvalds 已提交
957

958 959 960 961
	msg->msg_name = NULL;
	msg->msg_namelen = 0;
	msg->msg_control = NULL;
	msg->msg_controllen = 0;
962
	msg->msg_iov = (struct iovec *)iov;
963 964 965 966
	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
967

968
	return __sock_sendmsg(iocb, sock, msg, size);
L
Linus Torvalds 已提交
969 970
}

971 972
static ssize_t sock_aio_write(struct kiocb *iocb, const struct iovec *iov,
			  unsigned long nr_segs, loff_t pos)
973 974
{
	struct sock_iocb siocb, *x;
L
Linus Torvalds 已提交
975

976 977
	if (pos != 0)
		return -ESPIPE;
978 979

	x = alloc_sock_iocb(iocb, &siocb);
980 981
	if (!x)
		return -ENOMEM;
L
Linus Torvalds 已提交
982

983
	return do_sock_write(&x->async_msg, iocb, iocb->ki_filp, iov, nr_segs);
L
Linus Torvalds 已提交
984 985 986 987 988 989 990
}

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

A
Arjan van de Ven 已提交
991
static DEFINE_MUTEX(br_ioctl_mutex);
992
static int (*br_ioctl_hook) (struct net *, unsigned int cmd, void __user *arg);
L
Linus Torvalds 已提交
993

994
void brioctl_set(int (*hook) (struct net *, unsigned int, void __user *))
L
Linus Torvalds 已提交
995
{
A
Arjan van de Ven 已提交
996
	mutex_lock(&br_ioctl_mutex);
L
Linus Torvalds 已提交
997
	br_ioctl_hook = hook;
A
Arjan van de Ven 已提交
998
	mutex_unlock(&br_ioctl_mutex);
L
Linus Torvalds 已提交
999 1000 1001
}
EXPORT_SYMBOL(brioctl_set);

A
Arjan van de Ven 已提交
1002
static DEFINE_MUTEX(vlan_ioctl_mutex);
1003
static int (*vlan_ioctl_hook) (struct net *, void __user *arg);
L
Linus Torvalds 已提交
1004

1005
void vlan_ioctl_set(int (*hook) (struct net *, void __user *))
L
Linus Torvalds 已提交
1006
{
A
Arjan van de Ven 已提交
1007
	mutex_lock(&vlan_ioctl_mutex);
L
Linus Torvalds 已提交
1008
	vlan_ioctl_hook = hook;
A
Arjan van de Ven 已提交
1009
	mutex_unlock(&vlan_ioctl_mutex);
L
Linus Torvalds 已提交
1010 1011 1012
}
EXPORT_SYMBOL(vlan_ioctl_set);

A
Arjan van de Ven 已提交
1013
static DEFINE_MUTEX(dlci_ioctl_mutex);
1014
static int (*dlci_ioctl_hook) (unsigned int, void __user *);
L
Linus Torvalds 已提交
1015

1016
void dlci_ioctl_set(int (*hook) (unsigned int, void __user *))
L
Linus Torvalds 已提交
1017
{
A
Arjan van de Ven 已提交
1018
	mutex_lock(&dlci_ioctl_mutex);
L
Linus Torvalds 已提交
1019
	dlci_ioctl_hook = hook;
A
Arjan van de Ven 已提交
1020
	mutex_unlock(&dlci_ioctl_mutex);
L
Linus Torvalds 已提交
1021 1022 1023
}
EXPORT_SYMBOL(dlci_ioctl_set);

1024 1025 1026 1027 1028 1029 1030 1031 1032 1033 1034 1035 1036 1037 1038 1039 1040 1041
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
1042 1043 1044 1045 1046 1047 1048 1049
/*
 *	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;
1050
	struct sock *sk;
L
Linus Torvalds 已提交
1051 1052
	void __user *argp = (void __user *)arg;
	int pid, err;
1053
	struct net *net;
L
Linus Torvalds 已提交
1054

1055
	sock = file->private_data;
1056
	sk = sock->sk;
1057
	net = sock_net(sk);
L
Linus Torvalds 已提交
1058
	if (cmd >= SIOCDEVPRIVATE && cmd <= (SIOCDEVPRIVATE + 15)) {
1059
		err = dev_ioctl(net, cmd, argp);
L
Linus Torvalds 已提交
1060
	} else
J
Johannes Berg 已提交
1061
#ifdef CONFIG_WEXT_CORE
L
Linus Torvalds 已提交
1062
	if (cmd >= SIOCIWFIRST && cmd <= SIOCIWLAST) {
1063
		err = dev_ioctl(net, cmd, argp);
L
Linus Torvalds 已提交
1064
	} else
J
Johannes Berg 已提交
1065
#endif
1066
		switch (cmd) {
L
Linus Torvalds 已提交
1067 1068 1069 1070 1071 1072 1073 1074 1075
		case FIOSETOWN:
		case SIOCSPGRP:
			err = -EFAULT;
			if (get_user(pid, (int __user *)argp))
				break;
			err = f_setown(sock->file, pid, 1);
			break;
		case FIOGETOWN:
		case SIOCGPGRP:
1076
			err = put_user(f_getown(sock->file),
1077
				       (int __user *)argp);
L
Linus Torvalds 已提交
1078 1079 1080 1081 1082 1083 1084 1085 1086
			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
1087
			mutex_lock(&br_ioctl_mutex);
1088
			if (br_ioctl_hook)
1089
				err = br_ioctl_hook(net, cmd, argp);
A
Arjan van de Ven 已提交
1090
			mutex_unlock(&br_ioctl_mutex);
L
Linus Torvalds 已提交
1091 1092 1093 1094 1095 1096 1097
			break;
		case SIOCGIFVLAN:
		case SIOCSIFVLAN:
			err = -ENOPKG;
			if (!vlan_ioctl_hook)
				request_module("8021q");

A
Arjan van de Ven 已提交
1098
			mutex_lock(&vlan_ioctl_mutex);
L
Linus Torvalds 已提交
1099
			if (vlan_ioctl_hook)
1100
				err = vlan_ioctl_hook(net, argp);
A
Arjan van de Ven 已提交
1101
			mutex_unlock(&vlan_ioctl_mutex);
L
Linus Torvalds 已提交
1102 1103 1104 1105 1106 1107 1108
			break;
		case SIOCADDDLCI:
		case SIOCDELDLCI:
			err = -ENOPKG;
			if (!dlci_ioctl_hook)
				request_module("dlci");

1109 1110
			mutex_lock(&dlci_ioctl_mutex);
			if (dlci_ioctl_hook)
L
Linus Torvalds 已提交
1111
				err = dlci_ioctl_hook(cmd, argp);
1112
			mutex_unlock(&dlci_ioctl_mutex);
L
Linus Torvalds 已提交
1113 1114
			break;
		default:
1115
			err = sock_do_ioctl(net, sock, cmd, arg);
L
Linus Torvalds 已提交
1116
			break;
1117
		}
L
Linus Torvalds 已提交
1118 1119 1120 1121 1122 1123 1124
	return err;
}

int sock_create_lite(int family, int type, int protocol, struct socket **res)
{
	int err;
	struct socket *sock = NULL;
1125

L
Linus Torvalds 已提交
1126 1127 1128 1129 1130 1131 1132 1133 1134 1135 1136
	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
1137 1138 1139 1140
	err = security_socket_post_create(sock, family, type, protocol, 1);
	if (err)
		goto out_release;

L
Linus Torvalds 已提交
1141 1142 1143
out:
	*res = sock;
	return err;
V
Venkat Yekkirala 已提交
1144 1145 1146 1147
out_release:
	sock_release(sock);
	sock = NULL;
	goto out;
L
Linus Torvalds 已提交
1148
}
1149
EXPORT_SYMBOL(sock_create_lite);
L
Linus Torvalds 已提交
1150 1151

/* No kernel lock held - perfect */
1152
static unsigned int sock_poll(struct file *file, poll_table *wait)
L
Linus Torvalds 已提交
1153
{
1154
	unsigned int busy_flag = 0;
L
Linus Torvalds 已提交
1155 1156 1157
	struct socket *sock;

	/*
1158
	 *      We can't return errors to poll, so it's either yes or no.
L
Linus Torvalds 已提交
1159
	 */
1160
	sock = file->private_data;
1161

1162
	if (sk_can_busy_loop(sock->sk)) {
1163
		/* this socket can poll_ll so tell the system call */
1164
		busy_flag = POLL_BUSY_LOOP;
1165 1166

		/* once, only if requested by syscall */
1167 1168
		if (wait && (wait->_key & POLL_BUSY_LOOP))
			sk_busy_loop(sock->sk, 1);
1169 1170
	}

1171
	return busy_flag | sock->ops->poll(file, sock, wait);
L
Linus Torvalds 已提交
1172 1173
}

1174
static int sock_mmap(struct file *file, struct vm_area_struct *vma)
L
Linus Torvalds 已提交
1175
{
1176
	struct socket *sock = file->private_data;
L
Linus Torvalds 已提交
1177 1178 1179 1180

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

1181
static int sock_close(struct inode *inode, struct file *filp)
L
Linus Torvalds 已提交
1182 1183 1184 1185 1186 1187 1188 1189 1190 1191 1192 1193 1194
{
	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)
1195
 *	   or under socket lock
L
Linus Torvalds 已提交
1196 1197 1198 1199
 */

static int sock_fasync(int fd, struct file *filp, int on)
{
1200 1201
	struct socket *sock = filp->private_data;
	struct sock *sk = sock->sk;
1202
	struct socket_wq *wq;
L
Linus Torvalds 已提交
1203

1204
	if (sk == NULL)
L
Linus Torvalds 已提交
1205 1206 1207
		return -EINVAL;

	lock_sock(sk);
1208 1209
	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
1210

1211
	if (!wq->fasync_list)
1212 1213
		sock_reset_flag(sk, SOCK_FASYNC);
	else
E
Eric Dumazet 已提交
1214
		sock_set_flag(sk, SOCK_FASYNC);
L
Linus Torvalds 已提交
1215

1216
	release_sock(sk);
L
Linus Torvalds 已提交
1217 1218 1219
	return 0;
}

1220
/* This function may be called only under socket lock or callback_lock or rcu_lock */
L
Linus Torvalds 已提交
1221 1222 1223

int sock_wake_async(struct socket *sock, int how, int band)
{
1224 1225 1226 1227 1228 1229 1230 1231
	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
1232
		return -1;
1233
	}
1234
	switch (how) {
1235
	case SOCK_WAKE_WAITD:
L
Linus Torvalds 已提交
1236 1237 1238
		if (test_bit(SOCK_ASYNC_WAITDATA, &sock->flags))
			break;
		goto call_kill;
1239
	case SOCK_WAKE_SPACE:
L
Linus Torvalds 已提交
1240 1241 1242
		if (!test_and_clear_bit(SOCK_ASYNC_NOSPACE, &sock->flags))
			break;
		/* fall through */
1243
	case SOCK_WAKE_IO:
1244
call_kill:
1245
		kill_fasync(&wq->fasync_list, SIGIO, band);
L
Linus Torvalds 已提交
1246
		break;
1247
	case SOCK_WAKE_URG:
1248
		kill_fasync(&wq->fasync_list, SIGURG, band);
L
Linus Torvalds 已提交
1249
	}
1250
	rcu_read_unlock();
L
Linus Torvalds 已提交
1251 1252
	return 0;
}
1253
EXPORT_SYMBOL(sock_wake_async);
L
Linus Torvalds 已提交
1254

P
Pavel Emelyanov 已提交
1255
int __sock_create(struct net *net, int family, int type, int protocol,
1256
			 struct socket **res, int kern)
L
Linus Torvalds 已提交
1257 1258 1259
{
	int err;
	struct socket *sock;
1260
	const struct net_proto_family *pf;
L
Linus Torvalds 已提交
1261 1262

	/*
1263
	 *      Check protocol is in range
L
Linus Torvalds 已提交
1264 1265 1266 1267 1268 1269 1270 1271 1272 1273 1274 1275
	 */
	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) {
1276
		static int warned;
L
Linus Torvalds 已提交
1277 1278
		if (!warned) {
			warned = 1;
1279 1280
			pr_info("%s uses obsolete (PF_INET,SOCK_PACKET)\n",
				current->comm);
L
Linus Torvalds 已提交
1281 1282 1283 1284 1285 1286 1287
		}
		family = PF_PACKET;
	}

	err = security_socket_create(family, type, protocol, kern);
	if (err)
		return err;
1288

1289 1290 1291 1292 1293 1294 1295
	/*
	 *	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) {
1296
		net_warn_ratelimited("socket: no more sockets\n");
1297 1298 1299 1300 1301 1302
		return -ENFILE;	/* Not exactly a match, but its the
				   closest posix thing */
	}

	sock->type = type;

1303
#ifdef CONFIG_MODULES
1304 1305 1306
	/* 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
1307 1308 1309
	 * requested real, full-featured networking support upon configuration.
	 * Otherwise module support will break!
	 */
E
Eric Dumazet 已提交
1310
	if (rcu_access_pointer(net_families[family]) == NULL)
1311
		request_module("net-pf-%d", family);
L
Linus Torvalds 已提交
1312 1313
#endif

1314 1315 1316 1317 1318
	rcu_read_lock();
	pf = rcu_dereference(net_families[family]);
	err = -EAFNOSUPPORT;
	if (!pf)
		goto out_release;
L
Linus Torvalds 已提交
1319 1320 1321 1322 1323

	/*
	 * We will call the ->create function, that possibly is in a loadable
	 * module, so we have to bump that loadable module refcnt first.
	 */
1324
	if (!try_module_get(pf->owner))
L
Linus Torvalds 已提交
1325 1326
		goto out_release;

1327 1328 1329
	/* Now protected by module ref count */
	rcu_read_unlock();

1330
	err = pf->create(net, sock, protocol, kern);
1331
	if (err < 0)
L
Linus Torvalds 已提交
1332
		goto out_module_put;
1333

L
Linus Torvalds 已提交
1334 1335 1336 1337
	/*
	 * Now to bump the refcnt of the [loadable] module that owns this
	 * socket at sock_release time we decrement its refcnt.
	 */
1338 1339 1340
	if (!try_module_get(sock->ops->owner))
		goto out_module_busy;

L
Linus Torvalds 已提交
1341 1342 1343 1344
	/*
	 * Now that we're done with the ->create function, the [loadable]
	 * module can have its refcnt decremented
	 */
1345
	module_put(pf->owner);
V
Venkat Yekkirala 已提交
1346 1347
	err = security_socket_post_create(sock, family, type, protocol, kern);
	if (err)
1348
		goto out_sock_release;
1349
	*res = sock;
L
Linus Torvalds 已提交
1350

1351 1352 1353 1354
	return 0;

out_module_busy:
	err = -EAFNOSUPPORT;
L
Linus Torvalds 已提交
1355
out_module_put:
1356 1357 1358
	sock->ops = NULL;
	module_put(pf->owner);
out_sock_release:
L
Linus Torvalds 已提交
1359
	sock_release(sock);
1360 1361 1362 1363 1364
	return err;

out_release:
	rcu_read_unlock();
	goto out_sock_release;
L
Linus Torvalds 已提交
1365
}
P
Pavel Emelyanov 已提交
1366
EXPORT_SYMBOL(__sock_create);
L
Linus Torvalds 已提交
1367 1368 1369

int sock_create(int family, int type, int protocol, struct socket **res)
{
1370
	return __sock_create(current->nsproxy->net_ns, family, type, protocol, res, 0);
L
Linus Torvalds 已提交
1371
}
1372
EXPORT_SYMBOL(sock_create);
L
Linus Torvalds 已提交
1373 1374 1375

int sock_create_kern(int family, int type, int protocol, struct socket **res)
{
1376
	return __sock_create(&init_net, family, type, protocol, res, 1);
L
Linus Torvalds 已提交
1377
}
1378
EXPORT_SYMBOL(sock_create_kern);
L
Linus Torvalds 已提交
1379

1380
SYSCALL_DEFINE3(socket, int, family, int, type, int, protocol)
L
Linus Torvalds 已提交
1381 1382 1383
{
	int retval;
	struct socket *sock;
1384 1385
	int flags;

1386 1387 1388 1389 1390 1391
	/* 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);

1392
	flags = type & ~SOCK_TYPE_MASK;
1393
	if (flags & ~(SOCK_CLOEXEC | SOCK_NONBLOCK))
1394 1395
		return -EINVAL;
	type &= SOCK_TYPE_MASK;
L
Linus Torvalds 已提交
1396

U
Ulrich Drepper 已提交
1397 1398 1399
	if (SOCK_NONBLOCK != O_NONBLOCK && (flags & SOCK_NONBLOCK))
		flags = (flags & ~SOCK_NONBLOCK) | O_NONBLOCK;

L
Linus Torvalds 已提交
1400 1401 1402 1403
	retval = sock_create(family, type, protocol, &sock);
	if (retval < 0)
		goto out;

1404
	retval = sock_map_fd(sock, flags & (O_CLOEXEC | O_NONBLOCK));
L
Linus Torvalds 已提交
1405 1406 1407 1408 1409 1410 1411 1412 1413 1414 1415 1416 1417 1418 1419 1420
	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.
 */

1421 1422
SYSCALL_DEFINE4(socketpair, int, family, int, type, int, protocol,
		int __user *, usockvec)
L
Linus Torvalds 已提交
1423 1424 1425
{
	struct socket *sock1, *sock2;
	int fd1, fd2, err;
A
Al Viro 已提交
1426
	struct file *newfile1, *newfile2;
1427 1428 1429
	int flags;

	flags = type & ~SOCK_TYPE_MASK;
1430
	if (flags & ~(SOCK_CLOEXEC | SOCK_NONBLOCK))
1431 1432
		return -EINVAL;
	type &= SOCK_TYPE_MASK;
L
Linus Torvalds 已提交
1433

U
Ulrich Drepper 已提交
1434 1435 1436
	if (SOCK_NONBLOCK != O_NONBLOCK && (flags & SOCK_NONBLOCK))
		flags = (flags & ~SOCK_NONBLOCK) | O_NONBLOCK;

L
Linus Torvalds 已提交
1437 1438 1439 1440 1441 1442 1443 1444 1445 1446 1447 1448 1449 1450
	/*
	 * 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);
1451
	if (err < 0)
L
Linus Torvalds 已提交
1452 1453
		goto out_release_both;

1454
	fd1 = get_unused_fd_flags(flags);
1455 1456
	if (unlikely(fd1 < 0)) {
		err = fd1;
A
Al Viro 已提交
1457
		goto out_release_both;
1458
	}
1459

1460
	fd2 = get_unused_fd_flags(flags);
1461 1462
	if (unlikely(fd2 < 0)) {
		err = fd2;
1463
		goto out_put_unused_1;
1464 1465
	}

1466
	newfile1 = sock_alloc_file(sock1, flags, NULL);
1467 1468
	if (unlikely(IS_ERR(newfile1))) {
		err = PTR_ERR(newfile1);
1469
		goto out_put_unused_both;
1470 1471
	}

1472
	newfile2 = sock_alloc_file(sock2, flags, NULL);
1473 1474
	if (IS_ERR(newfile2)) {
		err = PTR_ERR(newfile2);
1475
		goto out_fput_1;
A
Al Viro 已提交
1476 1477
	}

1478 1479 1480 1481 1482 1483 1484 1485
	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
1486
	audit_fd_pair(fd1, fd2);
1487

A
Al Viro 已提交
1488 1489
	fd_install(fd1, newfile1);
	fd_install(fd2, newfile2);
L
Linus Torvalds 已提交
1490 1491 1492 1493
	/* fd1 and fd2 may be already another descriptors.
	 * Not kernel problem.
	 */

1494
	return 0;
L
Linus Torvalds 已提交
1495

1496 1497 1498 1499 1500 1501 1502 1503 1504 1505 1506 1507 1508
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
1509

1510 1511 1512 1513
out_put_unused_both:
	put_unused_fd(fd2);
out_put_unused_1:
	put_unused_fd(fd1);
L
Linus Torvalds 已提交
1514
out_release_both:
1515
	sock_release(sock2);
L
Linus Torvalds 已提交
1516
out_release_1:
1517
	sock_release(sock1);
L
Linus Torvalds 已提交
1518 1519 1520 1521 1522 1523 1524 1525 1526 1527 1528 1529
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).
 */

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

1536
	sock = sockfd_lookup_light(fd, &err, &fput_needed);
1537
	if (sock) {
1538
		err = move_addr_to_kernel(umyaddr, addrlen, &address);
1539 1540
		if (err >= 0) {
			err = security_socket_bind(sock,
1541
						   (struct sockaddr *)&address,
1542
						   addrlen);
1543 1544
			if (!err)
				err = sock->ops->bind(sock,
1545
						      (struct sockaddr *)
1546
						      &address, addrlen);
L
Linus Torvalds 已提交
1547
		}
1548
		fput_light(sock->file, fput_needed);
1549
	}
L
Linus Torvalds 已提交
1550 1551 1552 1553 1554 1555 1556 1557 1558
	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.
 */

1559
SYSCALL_DEFINE2(listen, int, fd, int, backlog)
L
Linus Torvalds 已提交
1560 1561
{
	struct socket *sock;
1562
	int err, fput_needed;
1563
	int somaxconn;
1564 1565 1566

	sock = sockfd_lookup_light(fd, &err, &fput_needed);
	if (sock) {
1567
		somaxconn = sock_net(sock->sk)->core.sysctl_somaxconn;
1568
		if ((unsigned int)backlog > somaxconn)
1569
			backlog = somaxconn;
L
Linus Torvalds 已提交
1570 1571

		err = security_socket_listen(sock, backlog);
1572 1573
		if (!err)
			err = sock->ops->listen(sock, backlog);
L
Linus Torvalds 已提交
1574

1575
		fput_light(sock->file, fput_needed);
L
Linus Torvalds 已提交
1576 1577 1578 1579 1580 1581 1582 1583 1584 1585 1586 1587 1588 1589 1590 1591
	}
	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.
 */

1592 1593
SYSCALL_DEFINE4(accept4, int, fd, struct sockaddr __user *, upeer_sockaddr,
		int __user *, upeer_addrlen, int, flags)
L
Linus Torvalds 已提交
1594 1595
{
	struct socket *sock, *newsock;
1596
	struct file *newfile;
1597
	int err, len, newfd, fput_needed;
1598
	struct sockaddr_storage address;
L
Linus Torvalds 已提交
1599

1600
	if (flags & ~(SOCK_CLOEXEC | SOCK_NONBLOCK))
U
Ulrich Drepper 已提交
1601 1602 1603 1604 1605
		return -EINVAL;

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

1606
	sock = sockfd_lookup_light(fd, &err, &fput_needed);
L
Linus Torvalds 已提交
1607 1608 1609 1610
	if (!sock)
		goto out;

	err = -ENFILE;
1611 1612
	newsock = sock_alloc();
	if (!newsock)
L
Linus Torvalds 已提交
1613 1614 1615 1616 1617 1618 1619 1620 1621 1622 1623
		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);

1624
	newfd = get_unused_fd_flags(flags);
1625 1626
	if (unlikely(newfd < 0)) {
		err = newfd;
1627 1628
		sock_release(newsock);
		goto out_put;
1629
	}
1630
	newfile = sock_alloc_file(newsock, flags, sock->sk->sk_prot_creator->name);
1631 1632 1633 1634 1635 1636
	if (unlikely(IS_ERR(newfile))) {
		err = PTR_ERR(newfile);
		put_unused_fd(newfd);
		sock_release(newsock);
		goto out_put;
	}
1637

1638 1639
	err = security_socket_accept(sock, newsock);
	if (err)
1640
		goto out_fd;
1641

L
Linus Torvalds 已提交
1642 1643
	err = sock->ops->accept(sock, newsock, sock->file->f_flags);
	if (err < 0)
1644
		goto out_fd;
L
Linus Torvalds 已提交
1645 1646

	if (upeer_sockaddr) {
1647
		if (newsock->ops->getname(newsock, (struct sockaddr *)&address,
1648
					  &len, 2) < 0) {
L
Linus Torvalds 已提交
1649
			err = -ECONNABORTED;
1650
			goto out_fd;
L
Linus Torvalds 已提交
1651
		}
1652
		err = move_addr_to_user(&address,
1653
					len, upeer_sockaddr, upeer_addrlen);
L
Linus Torvalds 已提交
1654
		if (err < 0)
1655
			goto out_fd;
L
Linus Torvalds 已提交
1656 1657 1658 1659
	}

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

1660 1661
	fd_install(newfd, newfile);
	err = newfd;
L
Linus Torvalds 已提交
1662 1663

out_put:
1664
	fput_light(sock->file, fput_needed);
L
Linus Torvalds 已提交
1665 1666
out:
	return err;
1667
out_fd:
1668
	fput(newfile);
1669
	put_unused_fd(newfd);
L
Linus Torvalds 已提交
1670 1671 1672
	goto out_put;
}

1673 1674
SYSCALL_DEFINE3(accept, int, fd, struct sockaddr __user *, upeer_sockaddr,
		int __user *, upeer_addrlen)
U
Ulrich Drepper 已提交
1675
{
U
Ulrich Drepper 已提交
1676
	return sys_accept4(fd, upeer_sockaddr, upeer_addrlen, 0);
U
Ulrich Drepper 已提交
1677 1678
}

L
Linus Torvalds 已提交
1679 1680 1681 1682 1683 1684 1685 1686 1687 1688 1689 1690
/*
 *	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.
 */

1691 1692
SYSCALL_DEFINE3(connect, int, fd, struct sockaddr __user *, uservaddr,
		int, addrlen)
L
Linus Torvalds 已提交
1693 1694
{
	struct socket *sock;
1695
	struct sockaddr_storage address;
1696
	int err, fput_needed;
L
Linus Torvalds 已提交
1697

1698
	sock = sockfd_lookup_light(fd, &err, &fput_needed);
L
Linus Torvalds 已提交
1699 1700
	if (!sock)
		goto out;
1701
	err = move_addr_to_kernel(uservaddr, addrlen, &address);
L
Linus Torvalds 已提交
1702 1703 1704
	if (err < 0)
		goto out_put;

1705
	err =
1706
	    security_socket_connect(sock, (struct sockaddr *)&address, addrlen);
L
Linus Torvalds 已提交
1707 1708 1709
	if (err)
		goto out_put;

1710
	err = sock->ops->connect(sock, (struct sockaddr *)&address, addrlen,
L
Linus Torvalds 已提交
1711 1712
				 sock->file->f_flags);
out_put:
1713
	fput_light(sock->file, fput_needed);
L
Linus Torvalds 已提交
1714 1715 1716 1717 1718 1719 1720 1721 1722
out:
	return err;
}

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

1723 1724
SYSCALL_DEFINE3(getsockname, int, fd, struct sockaddr __user *, usockaddr,
		int __user *, usockaddr_len)
L
Linus Torvalds 已提交
1725 1726
{
	struct socket *sock;
1727
	struct sockaddr_storage address;
1728
	int len, err, fput_needed;
1729

1730
	sock = sockfd_lookup_light(fd, &err, &fput_needed);
L
Linus Torvalds 已提交
1731 1732 1733 1734 1735 1736 1737
	if (!sock)
		goto out;

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

1738
	err = sock->ops->getname(sock, (struct sockaddr *)&address, &len, 0);
L
Linus Torvalds 已提交
1739 1740
	if (err)
		goto out_put;
1741
	err = move_addr_to_user(&address, len, usockaddr, usockaddr_len);
L
Linus Torvalds 已提交
1742 1743

out_put:
1744
	fput_light(sock->file, fput_needed);
L
Linus Torvalds 已提交
1745 1746 1747 1748 1749 1750 1751 1752 1753
out:
	return err;
}

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

1754 1755
SYSCALL_DEFINE3(getpeername, int, fd, struct sockaddr __user *, usockaddr,
		int __user *, usockaddr_len)
L
Linus Torvalds 已提交
1756 1757
{
	struct socket *sock;
1758
	struct sockaddr_storage address;
1759
	int len, err, fput_needed;
L
Linus Torvalds 已提交
1760

1761 1762
	sock = sockfd_lookup_light(fd, &err, &fput_needed);
	if (sock != NULL) {
L
Linus Torvalds 已提交
1763 1764
		err = security_socket_getpeername(sock);
		if (err) {
1765
			fput_light(sock->file, fput_needed);
L
Linus Torvalds 已提交
1766 1767 1768
			return err;
		}

1769
		err =
1770
		    sock->ops->getname(sock, (struct sockaddr *)&address, &len,
1771
				       1);
L
Linus Torvalds 已提交
1772
		if (!err)
1773
			err = move_addr_to_user(&address, len, usockaddr,
1774
						usockaddr_len);
1775
		fput_light(sock->file, fput_needed);
L
Linus Torvalds 已提交
1776 1777 1778 1779 1780 1781 1782 1783 1784 1785
	}
	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.
 */

1786
SYSCALL_DEFINE6(sendto, int, fd, void __user *, buff, size_t, len,
1787
		unsigned int, flags, struct sockaddr __user *, addr,
1788
		int, addr_len)
L
Linus Torvalds 已提交
1789 1790
{
	struct socket *sock;
1791
	struct sockaddr_storage address;
L
Linus Torvalds 已提交
1792 1793 1794
	int err;
	struct msghdr msg;
	struct iovec iov;
1795 1796
	int fput_needed;

1797 1798
	if (len > INT_MAX)
		len = INT_MAX;
1799 1800
	sock = sockfd_lookup_light(fd, &err, &fput_needed);
	if (!sock)
1801
		goto out;
1802

1803 1804 1805 1806 1807 1808 1809 1810
	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;
1811
	if (addr) {
1812
		err = move_addr_to_kernel(addr, addr_len, &address);
L
Linus Torvalds 已提交
1813 1814
		if (err < 0)
			goto out_put;
1815
		msg.msg_name = (struct sockaddr *)&address;
1816
		msg.msg_namelen = addr_len;
L
Linus Torvalds 已提交
1817 1818 1819 1820 1821 1822
	}
	if (sock->file->f_flags & O_NONBLOCK)
		flags |= MSG_DONTWAIT;
	msg.msg_flags = flags;
	err = sock_sendmsg(sock, &msg, len);

1823
out_put:
1824
	fput_light(sock->file, fput_needed);
1825
out:
L
Linus Torvalds 已提交
1826 1827 1828 1829
	return err;
}

/*
1830
 *	Send a datagram down a socket.
L
Linus Torvalds 已提交
1831 1832
 */

1833
SYSCALL_DEFINE4(send, int, fd, void __user *, buff, size_t, len,
1834
		unsigned int, flags)
L
Linus Torvalds 已提交
1835 1836 1837 1838 1839
{
	return sys_sendto(fd, buff, len, flags, NULL, 0);
}

/*
1840
 *	Receive a frame from the socket and optionally record the address of the
L
Linus Torvalds 已提交
1841 1842 1843 1844
 *	sender. We verify the buffers are writable and if needed move the
 *	sender address from kernel to user space.
 */

1845
SYSCALL_DEFINE6(recvfrom, int, fd, void __user *, ubuf, size_t, size,
1846
		unsigned int, flags, struct sockaddr __user *, addr,
1847
		int __user *, addr_len)
L
Linus Torvalds 已提交
1848 1849 1850 1851
{
	struct socket *sock;
	struct iovec iov;
	struct msghdr msg;
1852
	struct sockaddr_storage address;
1853
	int err, err2;
1854 1855
	int fput_needed;

1856 1857
	if (size > INT_MAX)
		size = INT_MAX;
1858
	sock = sockfd_lookup_light(fd, &err, &fput_needed);
L
Linus Torvalds 已提交
1859
	if (!sock)
1860
		goto out;
L
Linus Torvalds 已提交
1861

1862 1863 1864 1865 1866 1867
	msg.msg_control = NULL;
	msg.msg_controllen = 0;
	msg.msg_iovlen = 1;
	msg.msg_iov = &iov;
	iov.iov_len = size;
	iov.iov_base = ubuf;
1868 1869 1870 1871
	/* 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
1872 1873
	if (sock->file->f_flags & O_NONBLOCK)
		flags |= MSG_DONTWAIT;
1874
	err = sock_recvmsg(sock, &msg, size, flags);
L
Linus Torvalds 已提交
1875

1876
	if (err >= 0 && addr != NULL) {
1877
		err2 = move_addr_to_user(&address,
1878
					 msg.msg_namelen, addr, addr_len);
1879 1880
		if (err2 < 0)
			err = err2;
L
Linus Torvalds 已提交
1881
	}
1882 1883

	fput_light(sock->file, fput_needed);
1884
out:
L
Linus Torvalds 已提交
1885 1886 1887 1888
	return err;
}

/*
1889
 *	Receive a datagram from a socket.
L
Linus Torvalds 已提交
1890 1891
 */

1892 1893
SYSCALL_DEFINE4(recv, int, fd, void __user *, ubuf, size_t, size,
		unsigned int, flags)
L
Linus Torvalds 已提交
1894 1895 1896 1897 1898 1899 1900 1901 1902
{
	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.
 */

1903 1904
SYSCALL_DEFINE5(setsockopt, int, fd, int, level, int, optname,
		char __user *, optval, int, optlen)
L
Linus Torvalds 已提交
1905
{
1906
	int err, fput_needed;
L
Linus Torvalds 已提交
1907 1908 1909 1910
	struct socket *sock;

	if (optlen < 0)
		return -EINVAL;
1911 1912 1913 1914

	sock = sockfd_lookup_light(fd, &err, &fput_needed);
	if (sock != NULL) {
		err = security_socket_setsockopt(sock, level, optname);
1915 1916
		if (err)
			goto out_put;
L
Linus Torvalds 已提交
1917 1918

		if (level == SOL_SOCKET)
1919 1920 1921
			err =
			    sock_setsockopt(sock, level, optname, optval,
					    optlen);
L
Linus Torvalds 已提交
1922
		else
1923 1924 1925
			err =
			    sock->ops->setsockopt(sock, level, optname, optval,
						  optlen);
1926 1927
out_put:
		fput_light(sock->file, fput_needed);
L
Linus Torvalds 已提交
1928 1929 1930 1931 1932 1933 1934 1935 1936
	}
	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.
 */

1937 1938
SYSCALL_DEFINE5(getsockopt, int, fd, int, level, int, optname,
		char __user *, optval, int __user *, optlen)
L
Linus Torvalds 已提交
1939
{
1940
	int err, fput_needed;
L
Linus Torvalds 已提交
1941 1942
	struct socket *sock;

1943 1944
	sock = sockfd_lookup_light(fd, &err, &fput_needed);
	if (sock != NULL) {
1945 1946 1947
		err = security_socket_getsockopt(sock, level, optname);
		if (err)
			goto out_put;
L
Linus Torvalds 已提交
1948 1949

		if (level == SOL_SOCKET)
1950 1951 1952
			err =
			    sock_getsockopt(sock, level, optname, optval,
					    optlen);
L
Linus Torvalds 已提交
1953
		else
1954 1955 1956
			err =
			    sock->ops->getsockopt(sock, level, optname, optval,
						  optlen);
1957 1958
out_put:
		fput_light(sock->file, fput_needed);
L
Linus Torvalds 已提交
1959 1960 1961 1962 1963 1964 1965 1966
	}
	return err;
}

/*
 *	Shutdown a socket.
 */

1967
SYSCALL_DEFINE2(shutdown, int, fd, int, how)
L
Linus Torvalds 已提交
1968
{
1969
	int err, fput_needed;
L
Linus Torvalds 已提交
1970 1971
	struct socket *sock;

1972 1973
	sock = sockfd_lookup_light(fd, &err, &fput_needed);
	if (sock != NULL) {
L
Linus Torvalds 已提交
1974
		err = security_socket_shutdown(sock, how);
1975 1976 1977
		if (!err)
			err = sock->ops->shutdown(sock, how);
		fput_light(sock->file, fput_needed);
L
Linus Torvalds 已提交
1978 1979 1980 1981
	}
	return err;
}

1982
/* A couple of helpful macros for getting the address of the 32/64 bit
L
Linus Torvalds 已提交
1983 1984 1985 1986 1987 1988
 * 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)

1989 1990 1991 1992 1993
struct used_address {
	struct sockaddr_storage name;
	unsigned int name_len;
};

1994 1995 1996 1997 1998
static int copy_msghdr_from_user(struct msghdr *kmsg,
				 struct msghdr __user *umsg)
{
	if (copy_from_user(kmsg, umsg, sizeof(struct msghdr)))
		return -EFAULT;
1999 2000 2001 2002

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

2003
	if (kmsg->msg_namelen > sizeof(struct sockaddr_storage))
2004
		kmsg->msg_namelen = sizeof(struct sockaddr_storage);
2005 2006 2007
	return 0;
}

2008
static int ___sys_sendmsg(struct socket *sock, struct msghdr __user *msg,
2009
			 struct msghdr *msg_sys, unsigned int flags,
2010
			 struct used_address *used_address)
L
Linus Torvalds 已提交
2011
{
2012 2013
	struct compat_msghdr __user *msg_compat =
	    (struct compat_msghdr __user *)msg;
2014
	struct sockaddr_storage address;
L
Linus Torvalds 已提交
2015
	struct iovec iovstack[UIO_FASTIOV], *iov = iovstack;
2016
	unsigned char ctl[sizeof(struct cmsghdr) + 20]
2017 2018
	    __attribute__ ((aligned(sizeof(__kernel_size_t))));
	/* 20 is size of ipv6_pktinfo */
L
Linus Torvalds 已提交
2019
	unsigned char *ctl_buf = ctl;
E
Eric Dumazet 已提交
2020
	int err, ctl_len, total_len;
2021

L
Linus Torvalds 已提交
2022 2023
	err = -EFAULT;
	if (MSG_CMSG_COMPAT & flags) {
2024
		if (get_compat_msghdr(msg_sys, msg_compat))
L
Linus Torvalds 已提交
2025
			return -EFAULT;
2026 2027 2028 2029 2030
	} else {
		err = copy_msghdr_from_user(msg_sys, msg);
		if (err)
			return err;
	}
L
Linus Torvalds 已提交
2031

2032
	if (msg_sys->msg_iovlen > UIO_FASTIOV) {
E
Eric Dumazet 已提交
2033 2034 2035 2036 2037 2038
		err = -EMSGSIZE;
		if (msg_sys->msg_iovlen > UIO_MAXIOV)
			goto out;
		err = -ENOMEM;
		iov = kmalloc(msg_sys->msg_iovlen * sizeof(struct iovec),
			      GFP_KERNEL);
L
Linus Torvalds 已提交
2039
		if (!iov)
2040
			goto out;
L
Linus Torvalds 已提交
2041 2042 2043 2044
	}

	/* This will also move the address data into kernel space */
	if (MSG_CMSG_COMPAT & flags) {
2045
		err = verify_compat_iovec(msg_sys, iov, &address, VERIFY_READ);
L
Linus Torvalds 已提交
2046
	} else
2047
		err = verify_iovec(msg_sys, iov, &address, VERIFY_READ);
2048
	if (err < 0)
L
Linus Torvalds 已提交
2049 2050 2051 2052 2053
		goto out_freeiov;
	total_len = err;

	err = -ENOBUFS;

2054
	if (msg_sys->msg_controllen > INT_MAX)
L
Linus Torvalds 已提交
2055
		goto out_freeiov;
2056
	ctl_len = msg_sys->msg_controllen;
L
Linus Torvalds 已提交
2057
	if ((MSG_CMSG_COMPAT & flags) && ctl_len) {
2058
		err =
2059
		    cmsghdr_from_user_compat_to_kern(msg_sys, sock->sk, ctl,
2060
						     sizeof(ctl));
L
Linus Torvalds 已提交
2061 2062
		if (err)
			goto out_freeiov;
2063 2064
		ctl_buf = msg_sys->msg_control;
		ctl_len = msg_sys->msg_controllen;
L
Linus Torvalds 已提交
2065
	} else if (ctl_len) {
2066
		if (ctl_len > sizeof(ctl)) {
L
Linus Torvalds 已提交
2067
			ctl_buf = sock_kmalloc(sock->sk, ctl_len, GFP_KERNEL);
2068
			if (ctl_buf == NULL)
L
Linus Torvalds 已提交
2069 2070 2071 2072
				goto out_freeiov;
		}
		err = -EFAULT;
		/*
2073
		 * Careful! Before this, msg_sys->msg_control contains a user pointer.
L
Linus Torvalds 已提交
2074 2075 2076
		 * Afterwards, it will be a kernel pointer. Thus the compiler-assisted
		 * checking falls down on this.
		 */
2077
		if (copy_from_user(ctl_buf,
2078
				   (void __user __force *)msg_sys->msg_control,
2079
				   ctl_len))
L
Linus Torvalds 已提交
2080
			goto out_freectl;
2081
		msg_sys->msg_control = ctl_buf;
L
Linus Torvalds 已提交
2082
	}
2083
	msg_sys->msg_flags = flags;
L
Linus Torvalds 已提交
2084 2085

	if (sock->file->f_flags & O_NONBLOCK)
2086
		msg_sys->msg_flags |= MSG_DONTWAIT;
2087 2088 2089 2090 2091 2092
	/*
	 * 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.
	 */
2093 2094 2095
	if (used_address && msg_sys->msg_name &&
	    used_address->name_len == msg_sys->msg_namelen &&
	    !memcmp(&used_address->name, msg_sys->msg_name,
2096 2097 2098 2099 2100 2101 2102 2103 2104 2105 2106
		    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;
2107 2108 2109
		if (msg_sys->msg_name)
			memcpy(&used_address->name, msg_sys->msg_name,
			       used_address->name_len);
2110
	}
L
Linus Torvalds 已提交
2111 2112

out_freectl:
2113
	if (ctl_buf != ctl)
L
Linus Torvalds 已提交
2114 2115 2116
		sock_kfree_s(sock->sk, ctl_buf, ctl_len);
out_freeiov:
	if (iov != iovstack)
E
Eric Dumazet 已提交
2117
		kfree(iov);
2118 2119 2120 2121 2122 2123 2124 2125
out:
	return err;
}

/*
 *	BSD sendmsg interface
 */

2126
long __sys_sendmsg(int fd, struct msghdr __user *msg, unsigned flags)
2127 2128 2129
{
	int fput_needed, err;
	struct msghdr msg_sys;
2130 2131 2132
	struct socket *sock;

	sock = sockfd_lookup_light(fd, &err, &fput_needed);
2133 2134 2135
	if (!sock)
		goto out;

2136
	err = ___sys_sendmsg(sock, msg, &msg_sys, flags, NULL);
2137

2138
	fput_light(sock->file, fput_needed);
2139
out:
L
Linus Torvalds 已提交
2140 2141 2142
	return err;
}

2143 2144 2145 2146 2147 2148 2149
SYSCALL_DEFINE3(sendmsg, int, fd, struct msghdr __user *, msg, unsigned int, flags)
{
	if (flags & MSG_CMSG_COMPAT)
		return -EINVAL;
	return __sys_sendmsg(fd, msg, flags);
}

2150 2151 2152 2153 2154 2155 2156 2157 2158 2159 2160 2161
/*
 *	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;
2162
	struct used_address used_address;
2163

2164 2165
	if (vlen > UIO_MAXIOV)
		vlen = UIO_MAXIOV;
2166 2167 2168 2169 2170 2171 2172

	datagrams = 0;

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

2173
	used_address.name_len = UINT_MAX;
2174 2175
	entry = mmsg;
	compat_entry = (struct compat_mmsghdr __user *)mmsg;
2176
	err = 0;
2177 2178 2179

	while (datagrams < vlen) {
		if (MSG_CMSG_COMPAT & flags) {
2180 2181
			err = ___sys_sendmsg(sock, (struct msghdr __user *)compat_entry,
					     &msg_sys, flags, &used_address);
2182 2183 2184 2185 2186
			if (err < 0)
				break;
			err = __put_user(err, &compat_entry->msg_len);
			++compat_entry;
		} else {
2187 2188 2189
			err = ___sys_sendmsg(sock,
					     (struct msghdr __user *)entry,
					     &msg_sys, flags, &used_address);
2190 2191 2192 2193 2194 2195 2196 2197 2198 2199 2200 2201 2202
			if (err < 0)
				break;
			err = put_user(err, &entry->msg_len);
			++entry;
		}

		if (err)
			break;
		++datagrams;
	}

	fput_light(sock->file, fput_needed);

2203 2204
	/* We only return an error if no datagrams were able to be sent */
	if (datagrams != 0)
2205 2206 2207 2208 2209 2210 2211 2212
		return datagrams;

	return err;
}

SYSCALL_DEFINE4(sendmmsg, int, fd, struct mmsghdr __user *, mmsg,
		unsigned int, vlen, unsigned int, flags)
{
2213 2214
	if (flags & MSG_CMSG_COMPAT)
		return -EINVAL;
2215 2216 2217
	return __sys_sendmmsg(fd, mmsg, vlen, flags);
}

2218
static int ___sys_recvmsg(struct socket *sock, struct msghdr __user *msg,
2219
			 struct msghdr *msg_sys, unsigned int flags, int nosec)
L
Linus Torvalds 已提交
2220
{
2221 2222
	struct compat_msghdr __user *msg_compat =
	    (struct compat_msghdr __user *)msg;
L
Linus Torvalds 已提交
2223
	struct iovec iovstack[UIO_FASTIOV];
2224
	struct iovec *iov = iovstack;
L
Linus Torvalds 已提交
2225
	unsigned long cmsg_ptr;
E
Eric Dumazet 已提交
2226
	int err, total_len, len;
L
Linus Torvalds 已提交
2227 2228

	/* kernel mode address */
2229
	struct sockaddr_storage addr;
L
Linus Torvalds 已提交
2230 2231 2232 2233

	/* user mode address pointers */
	struct sockaddr __user *uaddr;
	int __user *uaddr_len;
2234

L
Linus Torvalds 已提交
2235
	if (MSG_CMSG_COMPAT & flags) {
2236
		if (get_compat_msghdr(msg_sys, msg_compat))
L
Linus Torvalds 已提交
2237
			return -EFAULT;
2238 2239 2240 2241 2242
	} else {
		err = copy_msghdr_from_user(msg_sys, msg);
		if (err)
			return err;
	}
L
Linus Torvalds 已提交
2243

2244
	if (msg_sys->msg_iovlen > UIO_FASTIOV) {
E
Eric Dumazet 已提交
2245 2246 2247 2248 2249 2250
		err = -EMSGSIZE;
		if (msg_sys->msg_iovlen > UIO_MAXIOV)
			goto out;
		err = -ENOMEM;
		iov = kmalloc(msg_sys->msg_iovlen * sizeof(struct iovec),
			      GFP_KERNEL);
L
Linus Torvalds 已提交
2251
		if (!iov)
2252
			goto out;
L
Linus Torvalds 已提交
2253 2254
	}

2255 2256
	/* Save the user-mode address (verify_iovec will change the
	 * kernel msghdr to use the kernel address space)
L
Linus Torvalds 已提交
2257
	 */
2258
	uaddr = (__force void __user *)msg_sys->msg_name;
L
Linus Torvalds 已提交
2259
	uaddr_len = COMPAT_NAMELEN(msg);
2260
	if (MSG_CMSG_COMPAT & flags)
2261
		err = verify_compat_iovec(msg_sys, iov, &addr, VERIFY_WRITE);
2262
	else
2263
		err = verify_iovec(msg_sys, iov, &addr, VERIFY_WRITE);
L
Linus Torvalds 已提交
2264 2265
	if (err < 0)
		goto out_freeiov;
2266
	total_len = err;
L
Linus Torvalds 已提交
2267

2268 2269
	cmsg_ptr = (unsigned long)msg_sys->msg_control;
	msg_sys->msg_flags = flags & (MSG_CMSG_CLOEXEC|MSG_CMSG_COMPAT);
2270

2271 2272 2273
	/* We assume all kernel code knows the size of sockaddr_storage */
	msg_sys->msg_namelen = 0;

L
Linus Torvalds 已提交
2274 2275
	if (sock->file->f_flags & O_NONBLOCK)
		flags |= MSG_DONTWAIT;
2276 2277
	err = (nosec ? sock_recvmsg_nosec : sock_recvmsg)(sock, msg_sys,
							  total_len, flags);
L
Linus Torvalds 已提交
2278 2279 2280 2281 2282
	if (err < 0)
		goto out_freeiov;
	len = err;

	if (uaddr != NULL) {
2283
		err = move_addr_to_user(&addr,
2284
					msg_sys->msg_namelen, uaddr,
2285
					uaddr_len);
L
Linus Torvalds 已提交
2286 2287 2288
		if (err < 0)
			goto out_freeiov;
	}
2289
	err = __put_user((msg_sys->msg_flags & ~MSG_CMSG_COMPAT),
2290
			 COMPAT_FLAGS(msg));
L
Linus Torvalds 已提交
2291 2292 2293
	if (err)
		goto out_freeiov;
	if (MSG_CMSG_COMPAT & flags)
2294
		err = __put_user((unsigned long)msg_sys->msg_control - cmsg_ptr,
L
Linus Torvalds 已提交
2295 2296
				 &msg_compat->msg_controllen);
	else
2297
		err = __put_user((unsigned long)msg_sys->msg_control - cmsg_ptr,
L
Linus Torvalds 已提交
2298 2299 2300 2301 2302 2303 2304
				 &msg->msg_controllen);
	if (err)
		goto out_freeiov;
	err = len;

out_freeiov:
	if (iov != iovstack)
E
Eric Dumazet 已提交
2305
		kfree(iov);
2306 2307 2308 2309 2310 2311 2312 2313
out:
	return err;
}

/*
 *	BSD recvmsg interface
 */

2314
long __sys_recvmsg(int fd, struct msghdr __user *msg, unsigned flags)
2315 2316 2317
{
	int fput_needed, err;
	struct msghdr msg_sys;
2318 2319 2320
	struct socket *sock;

	sock = sockfd_lookup_light(fd, &err, &fput_needed);
2321 2322 2323
	if (!sock)
		goto out;

2324
	err = ___sys_recvmsg(sock, msg, &msg_sys, flags, 0);
2325

2326
	fput_light(sock->file, fput_needed);
L
Linus Torvalds 已提交
2327 2328 2329 2330
out:
	return err;
}

2331 2332 2333 2334 2335 2336 2337 2338
SYSCALL_DEFINE3(recvmsg, int, fd, struct msghdr __user *, msg,
		unsigned int, flags)
{
	if (flags & MSG_CMSG_COMPAT)
		return -EINVAL;
	return __sys_recvmsg(fd, msg, flags);
}

2339 2340 2341 2342 2343 2344 2345 2346 2347 2348
/*
 *     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;
2349
	struct compat_mmsghdr __user *compat_entry;
2350 2351 2352 2353 2354 2355 2356 2357 2358 2359 2360 2361 2362 2363 2364 2365 2366 2367 2368
	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;
2369
	compat_entry = (struct compat_mmsghdr __user *)mmsg;
2370 2371 2372 2373 2374

	while (datagrams < vlen) {
		/*
		 * No need to ask LSM for more than the first datagram.
		 */
2375
		if (MSG_CMSG_COMPAT & flags) {
2376 2377 2378
			err = ___sys_recvmsg(sock, (struct msghdr __user *)compat_entry,
					     &msg_sys, flags & ~MSG_WAITFORONE,
					     datagrams);
2379 2380 2381 2382 2383
			if (err < 0)
				break;
			err = __put_user(err, &compat_entry->msg_len);
			++compat_entry;
		} else {
2384 2385 2386 2387
			err = ___sys_recvmsg(sock,
					     (struct msghdr __user *)entry,
					     &msg_sys, flags & ~MSG_WAITFORONE,
					     datagrams);
2388 2389 2390 2391 2392 2393
			if (err < 0)
				break;
			err = put_user(err, &entry->msg_len);
			++entry;
		}

2394 2395 2396 2397
		if (err)
			break;
		++datagrams;

2398 2399 2400 2401
		/* MSG_WAITFORONE turns on MSG_DONTWAIT after one packet */
		if (flags & MSG_WAITFORONE)
			flags |= MSG_DONTWAIT;

2402 2403 2404 2405 2406 2407 2408 2409 2410 2411 2412 2413 2414 2415 2416 2417 2418 2419 2420 2421
		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
2422

2423 2424 2425 2426 2427 2428 2429 2430 2431 2432 2433 2434 2435 2436 2437 2438 2439 2440 2441 2442 2443 2444 2445 2446 2447 2448 2449 2450 2451 2452 2453
	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;

2454 2455 2456
	if (flags & MSG_CMSG_COMPAT)
		return -EINVAL;

2457 2458 2459 2460 2461 2462 2463 2464 2465 2466 2467 2468 2469 2470 2471 2472
	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
2473 2474
/* Argument list sizes for sys_socketcall */
#define AL(x) ((x) * sizeof(unsigned long))
2475
static const unsigned char nargs[21] = {
2476 2477 2478
	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),
2479
	AL(4), AL(5), AL(4)
2480 2481
};

L
Linus Torvalds 已提交
2482 2483 2484
#undef AL

/*
2485
 *	System call vectors.
L
Linus Torvalds 已提交
2486 2487 2488
 *
 *	Argument checking cleaned up. Saved 20% in size.
 *  This function doesn't need to set the kernel lock because
2489
 *  it is set by the callees.
L
Linus Torvalds 已提交
2490 2491
 */

2492
SYSCALL_DEFINE2(socketcall, int, call, unsigned long __user *, args)
L
Linus Torvalds 已提交
2493
{
2494
	unsigned long a[AUDITSC_ARGS];
2495
	unsigned long a0, a1;
L
Linus Torvalds 已提交
2496
	int err;
2497
	unsigned int len;
L
Linus Torvalds 已提交
2498

2499
	if (call < 1 || call > SYS_SENDMMSG)
L
Linus Torvalds 已提交
2500 2501
		return -EINVAL;

2502 2503 2504 2505
	len = nargs[call];
	if (len > sizeof(a))
		return -EINVAL;

L
Linus Torvalds 已提交
2506
	/* copy_from_user should be SMP safe. */
2507
	if (copy_from_user(a, args, len))
L
Linus Torvalds 已提交
2508
		return -EFAULT;
2509

2510 2511 2512
	err = audit_socketcall(nargs[call] / sizeof(unsigned long), a);
	if (err)
		return err;
2513

2514 2515 2516 2517 2518 2519 2520 2521 2522 2523 2524 2525 2526 2527 2528 2529 2530
	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
2531 2532
		err = sys_accept4(a0, (struct sockaddr __user *)a1,
				  (int __user *)a[2], 0);
2533 2534 2535 2536 2537 2538 2539 2540 2541 2542 2543 2544 2545 2546 2547 2548 2549 2550 2551 2552 2553 2554 2555 2556 2557 2558 2559 2560 2561 2562 2563 2564 2565 2566 2567 2568 2569 2570 2571 2572 2573 2574 2575
		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:
		err = sys_sendmsg(a0, (struct msghdr __user *)a1, a[2]);
		break;
2576 2577 2578
	case SYS_SENDMMSG:
		err = sys_sendmmsg(a0, (struct mmsghdr __user *)a1, a[2], a[3]);
		break;
2579 2580 2581
	case SYS_RECVMSG:
		err = sys_recvmsg(a0, (struct msghdr __user *)a1, a[2]);
		break;
2582 2583 2584 2585
	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
2586 2587 2588
	case SYS_ACCEPT4:
		err = sys_accept4(a0, (struct sockaddr __user *)a1,
				  (int __user *)a[2], a[3]);
U
Ulrich Drepper 已提交
2589
		break;
2590 2591 2592
	default:
		err = -EINVAL;
		break;
L
Linus Torvalds 已提交
2593 2594 2595 2596
	}
	return err;
}

2597
#endif				/* __ARCH_WANT_SYS_SOCKETCALL */
L
Linus Torvalds 已提交
2598

2599 2600 2601 2602
/**
 *	sock_register - add a socket protocol handler
 *	@ops: description of protocol
 *
L
Linus Torvalds 已提交
2603 2604
 *	This function is called by a protocol handler that wants to
 *	advertise its address family, and have it linked into the
2605 2606
 *	socket interface. The value ops->family coresponds to the
 *	socket system call protocol family.
L
Linus Torvalds 已提交
2607
 */
2608
int sock_register(const struct net_proto_family *ops)
L
Linus Torvalds 已提交
2609 2610 2611 2612
{
	int err;

	if (ops->family >= NPROTO) {
2613
		pr_crit("protocol %d >= NPROTO(%d)\n", ops->family, NPROTO);
L
Linus Torvalds 已提交
2614 2615
		return -ENOBUFS;
	}
2616 2617

	spin_lock(&net_family_lock);
E
Eric Dumazet 已提交
2618 2619
	if (rcu_dereference_protected(net_families[ops->family],
				      lockdep_is_held(&net_family_lock)))
2620 2621
		err = -EEXIST;
	else {
2622
		rcu_assign_pointer(net_families[ops->family], ops);
L
Linus Torvalds 已提交
2623 2624
		err = 0;
	}
2625 2626
	spin_unlock(&net_family_lock);

2627
	pr_info("NET: Registered protocol family %d\n", ops->family);
L
Linus Torvalds 已提交
2628 2629
	return err;
}
2630
EXPORT_SYMBOL(sock_register);
L
Linus Torvalds 已提交
2631

2632 2633 2634 2635
/**
 *	sock_unregister - remove a protocol handler
 *	@family: protocol family to remove
 *
L
Linus Torvalds 已提交
2636 2637
 *	This function is called by a protocol handler that wants to
 *	remove its address family, and have it unlinked from the
2638 2639 2640 2641 2642 2643
 *	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
2644
 */
2645
void sock_unregister(int family)
L
Linus Torvalds 已提交
2646
{
2647
	BUG_ON(family < 0 || family >= NPROTO);
L
Linus Torvalds 已提交
2648

2649
	spin_lock(&net_family_lock);
2650
	RCU_INIT_POINTER(net_families[family], NULL);
2651 2652 2653 2654
	spin_unlock(&net_family_lock);

	synchronize_rcu();

2655
	pr_info("NET: Unregistered protocol family %d\n", family);
L
Linus Torvalds 已提交
2656
}
2657
EXPORT_SYMBOL(sock_unregister);
L
Linus Torvalds 已提交
2658

2659
static int __init sock_init(void)
L
Linus Torvalds 已提交
2660
{
N
Nick Piggin 已提交
2661
	int err;
2662 2663 2664 2665 2666 2667
	/*
	 *      Initialize the network sysctl infrastructure.
	 */
	err = net_sysctl_init();
	if (err)
		goto out;
N
Nick Piggin 已提交
2668

L
Linus Torvalds 已提交
2669
	/*
2670
	 *      Initialize skbuff SLAB cache
L
Linus Torvalds 已提交
2671 2672 2673 2674
	 */
	skb_init();

	/*
2675
	 *      Initialize the protocols module.
L
Linus Torvalds 已提交
2676 2677 2678
	 */

	init_inodecache();
N
Nick Piggin 已提交
2679 2680 2681 2682

	err = register_filesystem(&sock_fs_type);
	if (err)
		goto out_fs;
L
Linus Torvalds 已提交
2683
	sock_mnt = kern_mount(&sock_fs_type);
N
Nick Piggin 已提交
2684 2685 2686 2687
	if (IS_ERR(sock_mnt)) {
		err = PTR_ERR(sock_mnt);
		goto out_mount;
	}
2688 2689

	/* The real protocol initialization is performed in later initcalls.
L
Linus Torvalds 已提交
2690 2691 2692
	 */

#ifdef CONFIG_NETFILTER
2693 2694 2695
	err = netfilter_init();
	if (err)
		goto out;
L
Linus Torvalds 已提交
2696
#endif
2697

2698
	ptp_classifier_init();
2699

N
Nick Piggin 已提交
2700 2701 2702 2703 2704 2705 2706
out:
	return err;

out_mount:
	unregister_filesystem(&sock_fs_type);
out_fs:
	goto out;
L
Linus Torvalds 已提交
2707 2708
}

2709 2710
core_initcall(sock_init);	/* early initcall */

L
Linus Torvalds 已提交
2711 2712 2713 2714 2715 2716
#ifdef CONFIG_PROC_FS
void socket_seq_show(struct seq_file *seq)
{
	int cpu;
	int counter = 0;

2717
	for_each_possible_cpu(cpu)
2718
	    counter += per_cpu(sockets_in_use, cpu);
L
Linus Torvalds 已提交
2719 2720 2721 2722 2723 2724 2725

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

	seq_printf(seq, "sockets: used %d\n", counter);
}
2726
#endif				/* CONFIG_PROC_FS */
L
Linus Torvalds 已提交
2727

2728
#ifdef CONFIG_COMPAT
2729
static int do_siocgstamp(struct net *net, struct socket *sock,
2730
			 unsigned int cmd, void __user *up)
2731 2732 2733 2734 2735 2736
{
	mm_segment_t old_fs = get_fs();
	struct timeval ktv;
	int err;

	set_fs(KERNEL_DS);
2737
	err = sock_do_ioctl(net, sock, cmd, (unsigned long)&ktv);
2738
	set_fs(old_fs);
2739
	if (!err)
2740
		err = compat_put_timeval(&ktv, up);
2741

2742 2743 2744
	return err;
}

2745
static int do_siocgstampns(struct net *net, struct socket *sock,
2746
			   unsigned int cmd, void __user *up)
2747 2748 2749 2750 2751 2752
{
	mm_segment_t old_fs = get_fs();
	struct timespec kts;
	int err;

	set_fs(KERNEL_DS);
2753
	err = sock_do_ioctl(net, sock, cmd, (unsigned long)&kts);
2754
	set_fs(old_fs);
2755
	if (!err)
2756
		err = compat_put_timespec(&kts, up);
2757

2758 2759 2760
	return err;
}

2761
static int dev_ifname32(struct net *net, struct compat_ifreq __user *uifr32)
2762 2763 2764 2765 2766
{
	struct ifreq __user *uifr;
	int err;

	uifr = compat_alloc_user_space(sizeof(struct ifreq));
2767
	if (copy_in_user(uifr, uifr32, sizeof(struct compat_ifreq)))
2768 2769
		return -EFAULT;

2770
	err = dev_ioctl(net, SIOCGIFNAME, uifr);
2771 2772 2773
	if (err)
		return err;

2774
	if (copy_in_user(uifr32, uifr, sizeof(struct compat_ifreq)))
2775 2776 2777 2778 2779
		return -EFAULT;

	return 0;
}

2780
static int dev_ifconf(struct net *net, struct compat_ifconf __user *uifc32)
2781
{
2782
	struct compat_ifconf ifc32;
2783 2784
	struct ifconf ifc;
	struct ifconf __user *uifc;
2785
	struct compat_ifreq __user *ifr32;
2786 2787 2788 2789
	struct ifreq __user *ifr;
	unsigned int i, j;
	int err;

2790
	if (copy_from_user(&ifc32, uifc32, sizeof(struct compat_ifconf)))
2791 2792
		return -EFAULT;

2793
	memset(&ifc, 0, sizeof(ifc));
2794 2795 2796 2797 2798 2799
	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 {
2800 2801
		size_t len = ((ifc32.ifc_len / sizeof(struct compat_ifreq)) + 1) *
			sizeof(struct ifreq);
2802 2803 2804 2805
		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);
2806
		for (i = 0; i < ifc32.ifc_len; i += sizeof(struct compat_ifreq)) {
2807
			if (copy_in_user(ifr, ifr32, sizeof(struct compat_ifreq)))
2808 2809 2810 2811 2812 2813 2814 2815
				return -EFAULT;
			ifr++;
			ifr32++;
		}
	}
	if (copy_to_user(uifc, &ifc, sizeof(struct ifconf)))
		return -EFAULT;

2816
	err = dev_ioctl(net, SIOCGIFCONF, uifc);
2817 2818 2819 2820 2821 2822 2823 2824 2825
	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;
2826 2827 2828
	     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)))
2829 2830 2831 2832 2833 2834 2835 2836 2837 2838
			return -EFAULT;
		ifr32++;
		ifr++;
	}

	if (ifc32.ifcbuf == 0) {
		/* Translate from 64-bit structure multiple to
		 * a 32-bit one.
		 */
		i = ifc.ifc_len;
2839
		i = ((i / sizeof(struct ifreq)) * sizeof(struct compat_ifreq));
2840 2841 2842 2843
		ifc32.ifc_len = i;
	} else {
		ifc32.ifc_len = i;
	}
2844
	if (copy_to_user(uifc32, &ifc32, sizeof(struct compat_ifconf)))
2845 2846 2847 2848 2849
		return -EFAULT;

	return 0;
}

2850
static int ethtool_ioctl(struct net *net, struct compat_ifreq __user *ifr32)
2851
{
2852 2853 2854 2855
	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;
2856
	struct ifreq __user *ifr;
2857 2858
	u32 rule_cnt = 0, actual_rule_cnt;
	u32 ethcmd;
2859
	u32 data;
2860
	int ret;
2861

2862 2863
	if (get_user(data, &ifr32->ifr_ifru.ifru_data))
		return -EFAULT;
2864

2865 2866 2867
	compat_rxnfc = compat_ptr(data);

	if (get_user(ethcmd, &compat_rxnfc->cmd))
2868 2869
		return -EFAULT;

2870 2871 2872 2873 2874 2875 2876 2877 2878 2879 2880 2881 2882 2883 2884 2885 2886
	/* 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:
2887
	case ETHTOOL_SRXCLSRLINS:
2888 2889 2890 2891 2892 2893 2894 2895 2896
		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
2897
	rxnfc = (void __user *)ifr + ALIGN(sizeof(struct ifreq), 8);
2898 2899

	if (copy_in_user(&ifr->ifr_name, &ifr32->ifr_name, IFNAMSIZ))
2900 2901
		return -EFAULT;

2902 2903
	if (put_user(convert_in ? rxnfc : compat_ptr(data),
		     &ifr->ifr_ifru.ifru_data))
2904 2905
		return -EFAULT;

2906
	if (convert_in) {
2907
		/* We expect there to be holes between fs.m_ext and
2908 2909
		 * fs.ring_cookie and at the end of fs, but nowhere else.
		 */
2910 2911 2912 2913
		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));
2914 2915 2916 2917 2918 2919 2920
		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
2921 2922
				 (void __user *)(&rxnfc->fs.m_ext + 1) -
				 (void __user *)rxnfc) ||
2923 2924
		    copy_in_user(&rxnfc->fs.ring_cookie,
				 &compat_rxnfc->fs.ring_cookie,
S
Stephen Hemminger 已提交
2925 2926
				 (void __user *)(&rxnfc->fs.location + 1) -
				 (void __user *)&rxnfc->fs.ring_cookie) ||
2927 2928 2929 2930 2931 2932 2933 2934 2935 2936 2937
		    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
2938 2939
				 (const void __user *)(&rxnfc->fs.m_ext + 1) -
				 (const void __user *)rxnfc) ||
2940 2941
		    copy_in_user(&compat_rxnfc->fs.ring_cookie,
				 &rxnfc->fs.ring_cookie,
S
Stephen Hemminger 已提交
2942 2943
				 (const void __user *)(&rxnfc->fs.location + 1) -
				 (const void __user *)&rxnfc->fs.ring_cookie) ||
2944 2945 2946 2947 2948 2949 2950 2951 2952 2953 2954 2955 2956 2957 2958 2959 2960 2961 2962 2963 2964 2965 2966 2967 2968
		    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;
2969 2970
}

2971 2972 2973 2974 2975 2976
static int compat_siocwandev(struct net *net, struct compat_ifreq __user *uifr32)
{
	void __user *uptr;
	compat_uptr_t uptr32;
	struct ifreq __user *uifr;

2977
	uifr = compat_alloc_user_space(sizeof(*uifr));
2978 2979 2980 2981 2982 2983 2984 2985 2986 2987 2988 2989 2990 2991
	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);
}

2992 2993
static int bond_ioctl(struct net *net, unsigned int cmd,
			 struct compat_ifreq __user *ifr32)
2994 2995 2996 2997 2998 2999 3000 3001 3002 3003
{
	struct ifreq kifr;
	mm_segment_t old_fs;
	int err;

	switch (cmd) {
	case SIOCBONDENSLAVE:
	case SIOCBONDRELEASE:
	case SIOCBONDSETHWADDR:
	case SIOCBONDCHANGEACTIVE:
3004
		if (copy_from_user(&kifr, ifr32, sizeof(struct compat_ifreq)))
3005 3006 3007
			return -EFAULT;

		old_fs = get_fs();
3008
		set_fs(KERNEL_DS);
3009 3010
		err = dev_ioctl(net, cmd,
				(struct ifreq __user __force *) &kifr);
3011
		set_fs(old_fs);
3012 3013 3014

		return err;
	default:
3015
		return -ENOIOCTLCMD;
3016
	}
3017 3018
}

3019 3020
/* 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,
3021
				 struct compat_ifreq __user *u_ifreq32)
3022 3023 3024 3025 3026 3027 3028 3029 3030
{
	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;
3031
	if (get_user(data32, &u_ifreq32->ifr_ifru.ifru_data))
3032 3033 3034 3035 3036 3037 3038 3039
		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;
3040
	if (put_user(data64, &u_ifreq64->ifr_ifru.ifru_data))
3041 3042
		return -EFAULT;

3043
	return dev_ioctl(net, cmd, u_ifreq64);
3044 3045
}

3046 3047
static int dev_ifsioc(struct net *net, struct socket *sock,
			 unsigned int cmd, struct compat_ifreq __user *uifr32)
3048
{
3049
	struct ifreq __user *uifr;
3050 3051
	int err;

3052 3053 3054 3055 3056 3057
	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);

3058 3059 3060 3061 3062 3063 3064 3065 3066 3067 3068 3069
	if (!err) {
		switch (cmd) {
		case SIOCGIFFLAGS:
		case SIOCGIFMETRIC:
		case SIOCGIFMTU:
		case SIOCGIFMEM:
		case SIOCGIFHWADDR:
		case SIOCGIFINDEX:
		case SIOCGIFADDR:
		case SIOCGIFBRDADDR:
		case SIOCGIFDSTADDR:
		case SIOCGIFNETMASK:
3070
		case SIOCGIFPFLAGS:
3071
		case SIOCGIFTXQLEN:
3072 3073
		case SIOCGMIIPHY:
		case SIOCGMIIREG:
3074
			if (copy_in_user(uifr32, uifr, sizeof(*uifr32)))
3075 3076 3077 3078 3079 3080 3081
				err = -EFAULT;
			break;
		}
	}
	return err;
}

3082 3083 3084 3085 3086 3087 3088 3089 3090 3091
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));
3092 3093 3094 3095 3096 3097
	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);
3098 3099 3100 3101
	if (err)
		return -EFAULT;

	old_fs = get_fs();
3102
	set_fs(KERNEL_DS);
3103
	err = dev_ioctl(net, cmd, (void  __user __force *)&ifr);
3104
	set_fs(old_fs);
3105 3106 3107

	if (cmd == SIOCGIFMAP && !err) {
		err = copy_to_user(uifr32, &ifr, sizeof(ifr.ifr_name));
3108 3109 3110 3111 3112 3113
		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);
3114 3115 3116 3117 3118 3119
		if (err)
			err = -EFAULT;
	}
	return err;
}

3120
struct rtentry32 {
3121
	u32		rt_pad1;
3122 3123 3124
	struct sockaddr rt_dst;         /* target address               */
	struct sockaddr rt_gateway;     /* gateway addr (RTF_GATEWAY)   */
	struct sockaddr rt_genmask;     /* target network mask (IP)     */
3125 3126 3127 3128 3129 3130 3131
	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! */
3132
	/* char * */ u32 rt_dev;        /* forcing the device at add    */
3133 3134
	u32		rt_mtu;         /* per route MTU/Window         */
	u32		rt_window;      /* Window clamping              */
3135 3136 3137 3138 3139 3140 3141 3142 3143 3144 3145 3146 3147 3148 3149 3150
	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;
};

3151 3152
static int routing_ioctl(struct net *net, struct socket *sock,
			 unsigned int cmd, void __user *argp)
3153 3154 3155 3156 3157 3158 3159 3160 3161
{
	int ret;
	void *r = NULL;
	struct in6_rtmsg r6;
	struct rtentry r4;
	char devname[16];
	u32 rtdev;
	mm_segment_t old_fs = get_fs();

3162 3163
	if (sock && sock->sk && sock->sk->sk_family == AF_INET6) { /* ipv6 */
		struct in6_rtmsg32 __user *ur6 = argp;
3164
		ret = copy_from_user(&r6.rtmsg_dst, &(ur6->rtmsg_dst),
3165
			3 * sizeof(struct in6_addr));
3166 3167 3168 3169 3170 3171 3172
		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));
3173 3174 3175

		r = (void *) &r6;
	} else { /* ipv4 */
3176
		struct rtentry32 __user *ur4 = argp;
3177
		ret = copy_from_user(&r4.rt_dst, &(ur4->rt_dst),
3178
					3 * sizeof(struct sockaddr));
3179 3180 3181 3182 3183 3184
		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));
3185
		if (rtdev) {
3186
			ret |= copy_from_user(devname, compat_ptr(rtdev), 15);
3187 3188
			r4.rt_dev = (char __user __force *)devname;
			devname[15] = 0;
3189 3190 3191 3192 3193 3194 3195 3196 3197 3198 3199
		} else
			r4.rt_dev = NULL;

		r = (void *) &r4;
	}

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

3200
	set_fs(KERNEL_DS);
3201
	ret = sock_do_ioctl(net, sock, cmd, (unsigned long) r);
3202
	set_fs(old_fs);
3203 3204 3205 3206 3207 3208 3209

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
3210
 * use compatible ioctls
3211
 */
3212
static int old_bridge_ioctl(compat_ulong_t __user *argp)
3213
{
3214
	compat_ulong_t tmp;
3215

3216
	if (get_user(tmp, argp))
3217 3218 3219 3220 3221 3222
		return -EFAULT;
	if (tmp == BRCTL_GET_VERSION)
		return BRCTL_VERSION + 1;
	return -EINVAL;
}

3223 3224 3225 3226 3227 3228
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);
3229

3230
	if (cmd >= SIOCDEVPRIVATE && cmd <= (SIOCDEVPRIVATE + 15))
3231
		return compat_ifr_data_ioctl(net, cmd, argp);
3232 3233 3234 3235 3236 3237 3238 3239 3240 3241 3242

	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);
3243 3244
	case SIOCWANDEV:
		return compat_siocwandev(net, argp);
3245 3246 3247
	case SIOCGIFMAP:
	case SIOCSIFMAP:
		return compat_sioc_ifmap(net, cmd, argp);
3248 3249 3250 3251 3252 3253 3254 3255 3256 3257 3258 3259
	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);
3260 3261
	case SIOCBONDSLAVEINFOQUERY:
	case SIOCBONDINFOQUERY:
3262
	case SIOCSHWTSTAMP:
3263
	case SIOCGHWTSTAMP:
3264
		return compat_ifr_data_ioctl(net, cmd, argp);
3265 3266 3267 3268 3269 3270 3271 3272 3273 3274 3275 3276 3277 3278 3279 3280 3281 3282 3283 3284 3285 3286 3287 3288 3289 3290 3291 3292 3293 3294 3295 3296 3297 3298 3299 3300 3301 3302 3303 3304 3305 3306

	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:
3307 3308 3309 3310
	case SIOCSIFNAME:
	case SIOCGMIIPHY:
	case SIOCGMIIREG:
	case SIOCSMIIREG:
3311
		return dev_ifsioc(net, sock, cmd, argp);
3312

3313 3314 3315 3316
	case SIOCSARP:
	case SIOCGARP:
	case SIOCDARP:
	case SIOCATMARK:
3317 3318 3319
		return sock_do_ioctl(net, sock, cmd, arg);
	}

3320 3321
	return -ENOIOCTLCMD;
}
3322

3323
static long compat_sock_ioctl(struct file *file, unsigned int cmd,
3324
			      unsigned long arg)
3325 3326 3327
{
	struct socket *sock = file->private_data;
	int ret = -ENOIOCTLCMD;
3328 3329 3330 3331 3332
	struct sock *sk;
	struct net *net;

	sk = sock->sk;
	net = sock_net(sk);
3333 3334 3335 3336

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

3337 3338 3339 3340
	if (ret == -ENOIOCTLCMD &&
	    (cmd >= SIOCIWFIRST && cmd <= SIOCIWLAST))
		ret = compat_wext_handle_ioctl(net, cmd, arg);

3341 3342 3343
	if (ret == -ENOIOCTLCMD)
		ret = compat_sock_ioctl_trans(file, sock, cmd, arg);

3344 3345 3346 3347
	return ret;
}
#endif

3348 3349 3350 3351
int kernel_bind(struct socket *sock, struct sockaddr *addr, int addrlen)
{
	return sock->ops->bind(sock, addr, addrlen);
}
3352
EXPORT_SYMBOL(kernel_bind);
3353 3354 3355 3356 3357

int kernel_listen(struct socket *sock, int backlog)
{
	return sock->ops->listen(sock, backlog);
}
3358
EXPORT_SYMBOL(kernel_listen);
3359 3360 3361 3362 3363 3364 3365 3366 3367 3368 3369 3370 3371 3372

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);
3373
		*newsock = NULL;
3374 3375 3376 3377
		goto done;
	}

	(*newsock)->ops = sock->ops;
3378
	__module_get((*newsock)->ops->owner);
3379 3380 3381 3382

done:
	return err;
}
3383
EXPORT_SYMBOL(kernel_accept);
3384 3385

int kernel_connect(struct socket *sock, struct sockaddr *addr, int addrlen,
3386
		   int flags)
3387 3388 3389
{
	return sock->ops->connect(sock, addr, addrlen, flags);
}
3390
EXPORT_SYMBOL(kernel_connect);
3391 3392 3393 3394 3395 3396

int kernel_getsockname(struct socket *sock, struct sockaddr *addr,
			 int *addrlen)
{
	return sock->ops->getname(sock, addr, addrlen, 0);
}
3397
EXPORT_SYMBOL(kernel_getsockname);
3398 3399 3400 3401 3402 3403

int kernel_getpeername(struct socket *sock, struct sockaddr *addr,
			 int *addrlen)
{
	return sock->ops->getname(sock, addr, addrlen, 1);
}
3404
EXPORT_SYMBOL(kernel_getpeername);
3405 3406 3407 3408 3409

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

3414 3415 3416
	uoptval = (char __user __force *) optval;
	uoptlen = (int __user __force *) optlen;

3417 3418
	set_fs(KERNEL_DS);
	if (level == SOL_SOCKET)
3419
		err = sock_getsockopt(sock, level, optname, uoptval, uoptlen);
3420
	else
3421 3422
		err = sock->ops->getsockopt(sock, level, optname, uoptval,
					    uoptlen);
3423 3424 3425
	set_fs(oldfs);
	return err;
}
3426
EXPORT_SYMBOL(kernel_getsockopt);
3427 3428

int kernel_setsockopt(struct socket *sock, int level, int optname,
3429
			char *optval, unsigned int optlen)
3430 3431
{
	mm_segment_t oldfs = get_fs();
3432
	char __user *uoptval;
3433 3434
	int err;

3435 3436
	uoptval = (char __user __force *) optval;

3437 3438
	set_fs(KERNEL_DS);
	if (level == SOL_SOCKET)
3439
		err = sock_setsockopt(sock, level, optname, uoptval, optlen);
3440
	else
3441
		err = sock->ops->setsockopt(sock, level, optname, uoptval,
3442 3443 3444 3445
					    optlen);
	set_fs(oldfs);
	return err;
}
3446
EXPORT_SYMBOL(kernel_setsockopt);
3447 3448 3449 3450 3451 3452 3453 3454 3455

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);
}
3456
EXPORT_SYMBOL(kernel_sendpage);
3457 3458 3459 3460 3461 3462 3463 3464 3465 3466 3467 3468

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;
}
3469
EXPORT_SYMBOL(kernel_sock_ioctl);
3470

3471 3472 3473 3474 3475
int kernel_sock_shutdown(struct socket *sock, enum sock_shutdown_cmd how)
{
	return sock->ops->shutdown(sock, how);
}
EXPORT_SYMBOL(kernel_sock_shutdown);