socket.c 54.6 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>
66
#include <linux/rcupdate.h>
L
Linus Torvalds 已提交
67 68 69
#include <linux/netdevice.h>
#include <linux/proc_fs.h>
#include <linux/seq_file.h>
A
Arjan van de Ven 已提交
70
#include <linux/mutex.h>
L
Linus Torvalds 已提交
71 72
#include <linux/wanrouter.h>
#include <linux/if_bridge.h>
73 74
#include <linux/if_frad.h>
#include <linux/if_vlan.h>
L
Linus Torvalds 已提交
75 76 77 78 79 80 81 82 83 84
#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>
85
#include <linux/audit.h>
86
#include <linux/wireless.h>
L
Linus Torvalds 已提交
87 88 89 90 91 92 93 94 95 96

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

#include <net/compat.h>

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

static int sock_no_open(struct inode *irrelevant, struct file *dontcare);
97 98 99 100
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);
101
static int sock_mmap(struct file *file, struct vm_area_struct *vma);
L
Linus Torvalds 已提交
102 103 104 105

static int sock_close(struct inode *inode, struct file *file);
static unsigned int sock_poll(struct file *file,
			      struct poll_table_struct *wait);
106
static long sock_ioctl(struct file *file, unsigned int cmd, unsigned long arg);
107 108
#ifdef CONFIG_COMPAT
static long compat_sock_ioctl(struct file *file,
109
			      unsigned int cmd, unsigned long arg);
110
#endif
L
Linus Torvalds 已提交
111 112 113 114 115 116 117 118 119
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);

/*
 *	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.
 */

120
static const struct file_operations socket_file_ops = {
L
Linus Torvalds 已提交
121 122 123 124 125 126
	.owner =	THIS_MODULE,
	.llseek =	no_llseek,
	.aio_read =	sock_aio_read,
	.aio_write =	sock_aio_write,
	.poll =		sock_poll,
	.unlocked_ioctl = sock_ioctl,
127 128 129
#ifdef CONFIG_COMPAT
	.compat_ioctl = compat_sock_ioctl,
#endif
L
Linus Torvalds 已提交
130 131 132 133
	.mmap =		sock_mmap,
	.open =		sock_no_open,	/* special open code to disallow open via /proc */
	.release =	sock_close,
	.fasync =	sock_fasync,
134 135
	.sendpage =	sock_sendpage,
	.splice_write = generic_splice_sendpage,
L
Linus Torvalds 已提交
136 137 138 139 140 141 142
};

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

static DEFINE_SPINLOCK(net_family_lock);
143
static const struct net_proto_family *net_families[NPROTO] __read_mostly;
L
Linus Torvalds 已提交
144 145 146 147 148 149 150 151

/*
 *	Statistics counters of the socket lists
 */

static DEFINE_PER_CPU(int, sockets_in_use) = 0;

/*
152 153 154
 * 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
155 156
 */

157
#define MAX_SOCK_ADDR	128		/* 108 for Unix domain -
L
Linus Torvalds 已提交
158 159
					   16 for IP, 16 for IPX,
					   24 for IPv6,
160
					   about 80 for AX.25
L
Linus Torvalds 已提交
161 162
					   must be at least one bigger than
					   the AF_UNIX size (see net/unix/af_unix.c
163
					   :unix_mkname()).
L
Linus Torvalds 已提交
164
					 */
165

L
Linus Torvalds 已提交
166 167 168 169 170 171 172 173 174 175 176 177 178
/**
 *	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.
 */

int move_addr_to_kernel(void __user *uaddr, int ulen, void *kaddr)
{
179
	if (ulen < 0 || ulen > MAX_SOCK_ADDR)
L
Linus Torvalds 已提交
180
		return -EINVAL;
181
	if (ulen == 0)
L
Linus Torvalds 已提交
182
		return 0;
183
	if (copy_from_user(kaddr, uaddr, ulen))
L
Linus Torvalds 已提交
184
		return -EFAULT;
185
	return audit_sockaddr(ulen, kaddr);
L
Linus Torvalds 已提交
186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203
}

/**
 *	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.
 */
204 205 206

int move_addr_to_user(void *kaddr, int klen, void __user *uaddr,
		      int __user *ulen)
L
Linus Torvalds 已提交
207 208 209 210
{
	int err;
	int len;

211 212
	err = get_user(len, ulen);
	if (err)
L
Linus Torvalds 已提交
213
		return err;
214 215 216
	if (len > klen)
		len = klen;
	if (len < 0 || len > MAX_SOCK_ADDR)
L
Linus Torvalds 已提交
217
		return -EINVAL;
218
	if (len) {
S
Steve Grubb 已提交
219 220
		if (audit_sockaddr(klen, kaddr))
			return -ENOMEM;
221
		if (copy_to_user(uaddr, kaddr, len))
L
Linus Torvalds 已提交
222 223 224
			return -EFAULT;
	}
	/*
225 226
	 *      "fromlen shall refer to the value before truncation.."
	 *                      1003.1g
L
Linus Torvalds 已提交
227 228 229 230 231 232
	 */
	return __put_user(klen, ulen);
}

#define SOCKFS_MAGIC 0x534F434B

233
static struct kmem_cache *sock_inode_cachep __read_mostly;
L
Linus Torvalds 已提交
234 235 236 237

static struct inode *sock_alloc_inode(struct super_block *sb)
{
	struct socket_alloc *ei;
238

239
	ei = kmem_cache_alloc(sock_inode_cachep, GFP_KERNEL);
L
Linus Torvalds 已提交
240 241 242
	if (!ei)
		return NULL;
	init_waitqueue_head(&ei->socket.wait);
243

L
Linus Torvalds 已提交
244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259
	ei->socket.fasync_list = NULL;
	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)
{
	kmem_cache_free(sock_inode_cachep,
			container_of(inode, struct socket_alloc, vfs_inode));
}

260
static void init_once(void *foo, struct kmem_cache *cachep, unsigned long flags)
L
Linus Torvalds 已提交
261
{
262
	struct socket_alloc *ei = (struct socket_alloc *)foo;
L
Linus Torvalds 已提交
263

C
Christoph Lameter 已提交
264
	inode_init_once(&ei->vfs_inode);
L
Linus Torvalds 已提交
265
}
266

L
Linus Torvalds 已提交
267 268 269
static int init_inodecache(void)
{
	sock_inode_cachep = kmem_cache_create("sock_inode_cache",
270 271 272 273 274 275 276
					      sizeof(struct socket_alloc),
					      0,
					      (SLAB_HWCACHE_ALIGN |
					       SLAB_RECLAIM_ACCOUNT |
					       SLAB_MEM_SPREAD),
					      init_once,
					      NULL);
L
Linus Torvalds 已提交
277 278 279 280 281 282 283 284 285 286 287
	if (sock_inode_cachep == NULL)
		return -ENOMEM;
	return 0;
}

static struct super_operations sockfs_ops = {
	.alloc_inode =	sock_alloc_inode,
	.destroy_inode =sock_destroy_inode,
	.statfs =	simple_statfs,
};

288
static int sockfs_get_sb(struct file_system_type *fs_type,
289 290
			 int flags, const char *dev_name, void *data,
			 struct vfsmount *mnt)
L
Linus Torvalds 已提交
291
{
292 293
	return get_sb_pseudo(fs_type, "socket:", &sockfs_ops, SOCKFS_MAGIC,
			     mnt);
L
Linus Torvalds 已提交
294 295
}

296
static struct vfsmount *sock_mnt __read_mostly;
L
Linus Torvalds 已提交
297 298 299 300 301 302

static struct file_system_type sock_fs_type = {
	.name =		"sockfs",
	.get_sb =	sockfs_get_sb,
	.kill_sb =	kill_anon_super,
};
303

L
Linus Torvalds 已提交
304 305
static int sockfs_delete_dentry(struct dentry *dentry)
{
306 307 308 309 310 311 312 313
	/*
	 * At creation time, we pretended this dentry was hashed
	 * (by clearing DCACHE_UNHASHED bit in d_flags)
	 * At delete time, we restore the truth : not hashed.
	 * (so that dput() can proceed correctly)
	 */
	dentry->d_flags |= DCACHE_UNHASHED;
	return 0;
L
Linus Torvalds 已提交
314
}
315 316 317 318 319 320 321 322 323 324

/*
 * 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);
}

L
Linus Torvalds 已提交
325
static struct dentry_operations sockfs_dentry_operations = {
326
	.d_delete = sockfs_delete_dentry,
327
	.d_dname  = sockfs_dname,
L
Linus Torvalds 已提交
328 329 330 331 332
};

/*
 *	Obtains the first available file descriptor and sets it up for use.
 *
333 334
 *	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
335 336 337 338 339 340 341 342 343 344 345 346
 *	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.
 */

347
static int sock_alloc_fd(struct file **filep)
L
Linus Torvalds 已提交
348 349 350 351
{
	int fd;

	fd = get_unused_fd();
352
	if (likely(fd >= 0)) {
L
Linus Torvalds 已提交
353 354
		struct file *file = get_empty_filp();

355 356
		*filep = file;
		if (unlikely(!file)) {
L
Linus Torvalds 已提交
357
			put_unused_fd(fd);
358
			return -ENFILE;
L
Linus Torvalds 已提交
359
		}
360 361 362 363
	} else
		*filep = NULL;
	return fd;
}
L
Linus Torvalds 已提交
364

365 366
static int sock_attach_fd(struct socket *sock, struct file *file)
{
367
	struct qstr name = { .name = "" };
368

369
	file->f_path.dentry = d_alloc(sock_mnt->mnt_sb->s_root, &name);
J
Josef Sipek 已提交
370
	if (unlikely(!file->f_path.dentry))
371 372
		return -ENOMEM;

J
Josef Sipek 已提交
373
	file->f_path.dentry->d_op = &sockfs_dentry_operations;
374 375 376 377 378
	/*
	 * We dont want to push this dentry into global dentry hash table.
	 * We pretend dentry is already hashed, by unsetting DCACHE_UNHASHED
	 * This permits a working /proc/$pid/fd/XXX on sockets
	 */
J
Josef Sipek 已提交
379 380 381 382
	file->f_path.dentry->d_flags &= ~DCACHE_UNHASHED;
	d_instantiate(file->f_path.dentry, SOCK_INODE(sock));
	file->f_path.mnt = mntget(sock_mnt);
	file->f_mapping = file->f_path.dentry->d_inode->i_mapping;
383 384 385 386 387 388 389

	sock->file = file;
	file->f_op = SOCK_INODE(sock)->i_fop = &socket_file_ops;
	file->f_mode = FMODE_READ | FMODE_WRITE;
	file->f_flags = O_RDWR;
	file->f_pos = 0;
	file->private_data = sock;
L
Linus Torvalds 已提交
390

391 392 393 394 395 396 397 398 399 400 401 402 403
	return 0;
}

int sock_map_fd(struct socket *sock)
{
	struct file *newfile;
	int fd = sock_alloc_fd(&newfile);

	if (likely(fd >= 0)) {
		int err = sock_attach_fd(sock, newfile);

		if (unlikely(err < 0)) {
			put_filp(newfile);
L
Linus Torvalds 已提交
404
			put_unused_fd(fd);
405
			return err;
L
Linus Torvalds 已提交
406
		}
407
		fd_install(fd, newfile);
L
Linus Torvalds 已提交
408 409 410 411
	}
	return fd;
}

412 413 414 415 416
static struct socket *sock_from_file(struct file *file, int *err)
{
	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
}

L
Linus Torvalds 已提交
421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438
/**
 *	sockfd_lookup	- 	Go from a file number to its socket slot
 *	@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;

439 440
	file = fget(fd);
	if (!file) {
L
Linus Torvalds 已提交
441 442 443
		*err = -EBADF;
		return NULL;
	}
444

445 446
	sock = sock_from_file(file, err);
	if (!sock)
L
Linus Torvalds 已提交
447
		fput(file);
448 449
	return sock;
}
L
Linus Torvalds 已提交
450

451 452 453 454 455
static struct socket *sockfd_lookup_light(int fd, int *err, int *fput_needed)
{
	struct file *file;
	struct socket *sock;

456
	*err = -EBADF;
457 458 459 460 461 462
	file = fget_light(fd, fput_needed);
	if (file) {
		sock = sock_from_file(file, err);
		if (sock)
			return sock;
		fput_light(file, *fput_needed);
L
Linus Torvalds 已提交
463
	}
464
	return NULL;
L
Linus Torvalds 已提交
465 466 467 468
}

/**
 *	sock_alloc	-	allocate a socket
469
 *
L
Linus Torvalds 已提交
470 471 472 473 474 475 476
 *	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)
{
477 478
	struct inode *inode;
	struct socket *sock;
L
Linus Torvalds 已提交
479 480 481 482 483 484 485

	inode = new_inode(sock_mnt->mnt_sb);
	if (!inode)
		return NULL;

	sock = SOCKET_I(inode);

486
	inode->i_mode = S_IFSOCK | S_IRWXUGO;
L
Linus Torvalds 已提交
487 488 489 490 491 492 493 494 495 496 497 498 499
	inode->i_uid = current->fsuid;
	inode->i_gid = current->fsgid;

	get_cpu_var(sockets_in_use)++;
	put_cpu_var(sockets_in_use);
	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.
 */
500

L
Linus Torvalds 已提交
501 502 503 504 505
static int sock_no_open(struct inode *irrelevant, struct file *dontcare)
{
	return -ENXIO;
}

506
const struct file_operations bad_sock_fops = {
L
Linus Torvalds 已提交
507 508 509 510 511 512 513 514 515 516
	.owner = THIS_MODULE,
	.open = sock_no_open,
};

/**
 *	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
517
 *	an inode not a file.
L
Linus Torvalds 已提交
518
 */
519

L
Linus Torvalds 已提交
520 521 522 523 524 525 526 527 528 529 530 531 532 533 534 535 536 537 538
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);
	}

	if (sock->fasync_list)
		printk(KERN_ERR "sock_release: fasync list not empty!\n");

	get_cpu_var(sockets_in_use)--;
	put_cpu_var(sockets_in_use);
	if (!sock->file) {
		iput(SOCK_INODE(sock));
		return;
	}
539
	sock->file = NULL;
L
Linus Torvalds 已提交
540 541
}

542
static inline int __sock_sendmsg(struct kiocb *iocb, struct socket *sock,
L
Linus Torvalds 已提交
543 544 545 546 547 548 549 550 551 552 553 554 555 556 557 558 559 560 561 562 563 564 565 566 567 568 569 570 571 572 573 574 575 576 577 578 579 580 581 582 583 584
				 struct msghdr *msg, size_t size)
{
	struct sock_iocb *si = kiocb_to_siocb(iocb);
	int err;

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

	err = security_socket_sendmsg(sock, msg, size);
	if (err)
		return err;

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

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;
}

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
	 */
585
	msg->msg_iov = (struct iovec *)vec;
L
Linus Torvalds 已提交
586 587 588 589 590 591
	msg->msg_iovlen = num;
	result = sock_sendmsg(sock, msg, size);
	set_fs(oldfs);
	return result;
}

592 593 594 595 596 597 598 599 600 601 602 603 604 605 606 607 608 609 610 611 612 613 614 615 616 617 618 619 620
/*
 * 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)
{
	ktime_t kt = skb->tstamp;

	if (!sock_flag(sk, SOCK_RCVTSTAMPNS)) {
		struct timeval tv;
		/* Race occurred between timestamp enabling and packet
		   receiving.  Fill in the current time for now. */
		if (kt.tv64 == 0)
			kt = ktime_get_real();
		skb->tstamp = kt;
		tv = ktime_to_timeval(kt);
		put_cmsg(msg, SOL_SOCKET, SCM_TIMESTAMP, sizeof(tv), &tv);
	} else {
		struct timespec ts;
		/* Race occurred between timestamp enabling and packet
		   receiving.  Fill in the current time for now. */
		if (kt.tv64 == 0)
			kt = ktime_get_real();
		skb->tstamp = kt;
		ts = ktime_to_timespec(kt);
		put_cmsg(msg, SOL_SOCKET, SCM_TIMESTAMPNS, sizeof(ts), &ts);
	}
}

621 622
EXPORT_SYMBOL_GPL(__sock_recv_timestamp);

623
static inline int __sock_recvmsg(struct kiocb *iocb, struct socket *sock,
L
Linus Torvalds 已提交
624 625 626 627 628 629 630 631 632 633 634 635 636 637 638 639 640 641
				 struct msghdr *msg, size_t size, int flags)
{
	int err;
	struct sock_iocb *si = kiocb_to_siocb(iocb);

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

	err = security_socket_recvmsg(sock, msg, size, flags);
	if (err)
		return err;

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

642
int sock_recvmsg(struct socket *sock, struct msghdr *msg,
L
Linus Torvalds 已提交
643 644 645 646 647 648
		 size_t size, int flags)
{
	struct kiocb iocb;
	struct sock_iocb siocb;
	int ret;

649
	init_sync_kiocb(&iocb, NULL);
L
Linus Torvalds 已提交
650 651 652 653 654 655 656
	iocb.private = &siocb;
	ret = __sock_recvmsg(&iocb, sock, msg, size, flags);
	if (-EIOCBQUEUED == ret)
		ret = wait_on_sync_kiocb(&iocb);
	return ret;
}

657 658
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
659 660 661 662 663 664 665 666 667
{
	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
	 */
668
	msg->msg_iov = (struct iovec *)vec, msg->msg_iovlen = num;
L
Linus Torvalds 已提交
669 670 671 672 673 674 675 676 677 678
	result = sock_recvmsg(sock, msg, size, flags);
	set_fs(oldfs);
	return result;
}

static void sock_aio_dtor(struct kiocb *iocb)
{
	kfree(iocb->private);
}

679 680
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
681 682 683 684
{
	struct socket *sock;
	int flags;

685 686 687 688 689 690 691 692
	sock = file->private_data;

	flags = !(file->f_flags & O_NONBLOCK) ? 0 : MSG_DONTWAIT;
	if (more)
		flags |= MSG_MORE;

	return sock->ops->sendpage(sock, page, offset, size, flags);
}
L
Linus Torvalds 已提交
693

694
static struct sock_iocb *alloc_sock_iocb(struct kiocb *iocb,
695
					 struct sock_iocb *siocb)
696 697 698 699 700
{
	if (!is_sync_kiocb(iocb)) {
		siocb = kmalloc(sizeof(*siocb), GFP_KERNEL);
		if (!siocb)
			return NULL;
L
Linus Torvalds 已提交
701 702 703
		iocb->ki_dtor = sock_aio_dtor;
	}

704 705 706
	siocb->kiocb = iocb;
	iocb->private = siocb;
	return siocb;
L
Linus Torvalds 已提交
707 708
}

709
static ssize_t do_sock_read(struct msghdr *msg, struct kiocb *iocb,
710 711
		struct file *file, const struct iovec *iov,
		unsigned long nr_segs)
712 713 714 715
{
	struct socket *sock = file->private_data;
	size_t size = 0;
	int i;
L
Linus Torvalds 已提交
716

717 718
	for (i = 0; i < nr_segs; i++)
		size += iov[i].iov_len;
L
Linus Torvalds 已提交
719

720 721 722 723
	msg->msg_name = NULL;
	msg->msg_namelen = 0;
	msg->msg_control = NULL;
	msg->msg_controllen = 0;
724
	msg->msg_iov = (struct iovec *)iov;
725 726 727 728 729 730
	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);
}

731 732
static ssize_t sock_aio_read(struct kiocb *iocb, const struct iovec *iov,
				unsigned long nr_segs, loff_t pos)
733 734 735
{
	struct sock_iocb siocb, *x;

L
Linus Torvalds 已提交
736 737
	if (pos != 0)
		return -ESPIPE;
738 739

	if (iocb->ki_left == 0)	/* Match SYS5 behaviour */
L
Linus Torvalds 已提交
740 741
		return 0;

742 743

	x = alloc_sock_iocb(iocb, &siocb);
744 745
	if (!x)
		return -ENOMEM;
746
	return do_sock_read(&x->async_msg, iocb, iocb->ki_filp, iov, nr_segs);
L
Linus Torvalds 已提交
747 748
}

749
static ssize_t do_sock_write(struct msghdr *msg, struct kiocb *iocb,
750 751
			struct file *file, const struct iovec *iov,
			unsigned long nr_segs)
L
Linus Torvalds 已提交
752
{
753 754 755
	struct socket *sock = file->private_data;
	size_t size = 0;
	int i;
L
Linus Torvalds 已提交
756

757 758
	for (i = 0; i < nr_segs; i++)
		size += iov[i].iov_len;
L
Linus Torvalds 已提交
759

760 761 762 763
	msg->msg_name = NULL;
	msg->msg_namelen = 0;
	msg->msg_control = NULL;
	msg->msg_controllen = 0;
764
	msg->msg_iov = (struct iovec *)iov;
765 766 767 768
	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
769

770
	return __sock_sendmsg(iocb, sock, msg, size);
L
Linus Torvalds 已提交
771 772
}

773 774
static ssize_t sock_aio_write(struct kiocb *iocb, const struct iovec *iov,
			  unsigned long nr_segs, loff_t pos)
775 776
{
	struct sock_iocb siocb, *x;
L
Linus Torvalds 已提交
777

778 779
	if (pos != 0)
		return -ESPIPE;
780 781

	if (iocb->ki_left == 0)	/* Match SYS5 behaviour */
782
		return 0;
L
Linus Torvalds 已提交
783

784
	x = alloc_sock_iocb(iocb, &siocb);
785 786
	if (!x)
		return -ENOMEM;
L
Linus Torvalds 已提交
787

788
	return do_sock_write(&x->async_msg, iocb, iocb->ki_filp, iov, nr_segs);
L
Linus Torvalds 已提交
789 790 791 792 793 794 795
}

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

A
Arjan van de Ven 已提交
796
static DEFINE_MUTEX(br_ioctl_mutex);
797
static int (*br_ioctl_hook) (unsigned int cmd, void __user *arg) = NULL;
L
Linus Torvalds 已提交
798

799
void brioctl_set(int (*hook) (unsigned int, void __user *))
L
Linus Torvalds 已提交
800
{
A
Arjan van de Ven 已提交
801
	mutex_lock(&br_ioctl_mutex);
L
Linus Torvalds 已提交
802
	br_ioctl_hook = hook;
A
Arjan van de Ven 已提交
803
	mutex_unlock(&br_ioctl_mutex);
L
Linus Torvalds 已提交
804
}
805

L
Linus Torvalds 已提交
806 807
EXPORT_SYMBOL(brioctl_set);

A
Arjan van de Ven 已提交
808
static DEFINE_MUTEX(vlan_ioctl_mutex);
809
static int (*vlan_ioctl_hook) (void __user *arg);
L
Linus Torvalds 已提交
810

811
void vlan_ioctl_set(int (*hook) (void __user *))
L
Linus Torvalds 已提交
812
{
A
Arjan van de Ven 已提交
813
	mutex_lock(&vlan_ioctl_mutex);
L
Linus Torvalds 已提交
814
	vlan_ioctl_hook = hook;
A
Arjan van de Ven 已提交
815
	mutex_unlock(&vlan_ioctl_mutex);
L
Linus Torvalds 已提交
816
}
817

L
Linus Torvalds 已提交
818 819
EXPORT_SYMBOL(vlan_ioctl_set);

A
Arjan van de Ven 已提交
820
static DEFINE_MUTEX(dlci_ioctl_mutex);
821
static int (*dlci_ioctl_hook) (unsigned int, void __user *);
L
Linus Torvalds 已提交
822

823
void dlci_ioctl_set(int (*hook) (unsigned int, void __user *))
L
Linus Torvalds 已提交
824
{
A
Arjan van de Ven 已提交
825
	mutex_lock(&dlci_ioctl_mutex);
L
Linus Torvalds 已提交
826
	dlci_ioctl_hook = hook;
A
Arjan van de Ven 已提交
827
	mutex_unlock(&dlci_ioctl_mutex);
L
Linus Torvalds 已提交
828
}
829

L
Linus Torvalds 已提交
830 831 832 833 834 835 836 837 838 839 840 841 842
EXPORT_SYMBOL(dlci_ioctl_set);

/*
 *	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;
	void __user *argp = (void __user *)arg;
	int pid, err;

843
	sock = file->private_data;
L
Linus Torvalds 已提交
844 845 846
	if (cmd >= SIOCDEVPRIVATE && cmd <= (SIOCDEVPRIVATE + 15)) {
		err = dev_ioctl(cmd, argp);
	} else
847
#ifdef CONFIG_WIRELESS_EXT
L
Linus Torvalds 已提交
848 849 850
	if (cmd >= SIOCIWFIRST && cmd <= SIOCIWLAST) {
		err = dev_ioctl(cmd, argp);
	} else
851 852
#endif				/* CONFIG_WIRELESS_EXT */
		switch (cmd) {
L
Linus Torvalds 已提交
853 854 855 856 857 858 859 860 861
		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:
862
			err = put_user(f_getown(sock->file),
863
				       (int __user *)argp);
L
Linus Torvalds 已提交
864 865 866 867 868 869 870 871 872
			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
873
			mutex_lock(&br_ioctl_mutex);
874
			if (br_ioctl_hook)
L
Linus Torvalds 已提交
875
				err = br_ioctl_hook(cmd, argp);
A
Arjan van de Ven 已提交
876
			mutex_unlock(&br_ioctl_mutex);
L
Linus Torvalds 已提交
877 878 879 880 881 882 883
			break;
		case SIOCGIFVLAN:
		case SIOCSIFVLAN:
			err = -ENOPKG;
			if (!vlan_ioctl_hook)
				request_module("8021q");

A
Arjan van de Ven 已提交
884
			mutex_lock(&vlan_ioctl_mutex);
L
Linus Torvalds 已提交
885 886
			if (vlan_ioctl_hook)
				err = vlan_ioctl_hook(argp);
A
Arjan van de Ven 已提交
887
			mutex_unlock(&vlan_ioctl_mutex);
L
Linus Torvalds 已提交
888 889 890 891 892 893 894 895
			break;
		case SIOCADDDLCI:
		case SIOCDELDLCI:
			err = -ENOPKG;
			if (!dlci_ioctl_hook)
				request_module("dlci");

			if (dlci_ioctl_hook) {
A
Arjan van de Ven 已提交
896
				mutex_lock(&dlci_ioctl_mutex);
L
Linus Torvalds 已提交
897
				err = dlci_ioctl_hook(cmd, argp);
A
Arjan van de Ven 已提交
898
				mutex_unlock(&dlci_ioctl_mutex);
L
Linus Torvalds 已提交
899 900 901 902
			}
			break;
		default:
			err = sock->ops->ioctl(sock, cmd, arg);
903 904 905 906 907 908 909

			/*
			 * If this ioctl is unknown try to hand it down
			 * to the NIC driver.
			 */
			if (err == -ENOIOCTLCMD)
				err = dev_ioctl(cmd, argp);
L
Linus Torvalds 已提交
910
			break;
911
		}
L
Linus Torvalds 已提交
912 913 914 915 916 917 918
	return err;
}

int sock_create_lite(int family, int type, int protocol, struct socket **res)
{
	int err;
	struct socket *sock = NULL;
919

L
Linus Torvalds 已提交
920 921 922 923 924 925 926 927 928 929 930
	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
931 932 933 934
	err = security_socket_post_create(sock, family, type, protocol, 1);
	if (err)
		goto out_release;

L
Linus Torvalds 已提交
935 936 937
out:
	*res = sock;
	return err;
V
Venkat Yekkirala 已提交
938 939 940 941
out_release:
	sock_release(sock);
	sock = NULL;
	goto out;
L
Linus Torvalds 已提交
942 943 944
}

/* No kernel lock held - perfect */
945
static unsigned int sock_poll(struct file *file, poll_table *wait)
L
Linus Torvalds 已提交
946 947 948 949
{
	struct socket *sock;

	/*
950
	 *      We can't return errors to poll, so it's either yes or no.
L
Linus Torvalds 已提交
951
	 */
952
	sock = file->private_data;
L
Linus Torvalds 已提交
953 954 955
	return sock->ops->poll(file, sock, wait);
}

956
static int sock_mmap(struct file *file, struct vm_area_struct *vma)
L
Linus Torvalds 已提交
957
{
958
	struct socket *sock = file->private_data;
L
Linus Torvalds 已提交
959 960 961 962

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

963
static int sock_close(struct inode *inode, struct file *filp)
L
Linus Torvalds 已提交
964 965
{
	/*
966 967
	 *      It was possible the inode is NULL we were
	 *      closing an unfinished socket.
L
Linus Torvalds 已提交
968 969
	 */

970
	if (!inode) {
L
Linus Torvalds 已提交
971 972 973 974 975 976 977 978 979 980 981 982 983 984 985 986 987 988 989 990 991 992 993 994 995
		printk(KERN_DEBUG "sock_close: NULL inode\n");
		return 0;
	}
	sock_fasync(-1, filp, 0);
	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)
 *	   or under socket lock.
 *	3. fasync_list can be used from softirq context, so that
 *	   modification under socket lock have to be enhanced with
 *	   write_lock_bh(&sk->sk_callback_lock).
 *							--ANK (990710)
 */

static int sock_fasync(int fd, struct file *filp, int on)
{
996
	struct fasync_struct *fa, *fna = NULL, **prev;
L
Linus Torvalds 已提交
997 998 999
	struct socket *sock;
	struct sock *sk;

1000
	if (on) {
1001
		fna = kmalloc(sizeof(struct fasync_struct), GFP_KERNEL);
1002
		if (fna == NULL)
L
Linus Torvalds 已提交
1003 1004 1005
			return -ENOMEM;
	}

1006
	sock = filp->private_data;
L
Linus Torvalds 已提交
1007

1008 1009
	sk = sock->sk;
	if (sk == NULL) {
L
Linus Torvalds 已提交
1010 1011 1012 1013 1014 1015
		kfree(fna);
		return -EINVAL;
	}

	lock_sock(sk);

1016
	prev = &(sock->fasync_list);
L
Linus Torvalds 已提交
1017

1018 1019
	for (fa = *prev; fa != NULL; prev = &fa->fa_next, fa = *prev)
		if (fa->fa_file == filp)
L
Linus Torvalds 已提交
1020 1021
			break;

1022 1023
	if (on) {
		if (fa != NULL) {
L
Linus Torvalds 已提交
1024
			write_lock_bh(&sk->sk_callback_lock);
1025
			fa->fa_fd = fd;
L
Linus Torvalds 已提交
1026 1027 1028 1029 1030
			write_unlock_bh(&sk->sk_callback_lock);

			kfree(fna);
			goto out;
		}
1031 1032 1033 1034
		fna->fa_file = filp;
		fna->fa_fd = fd;
		fna->magic = FASYNC_MAGIC;
		fna->fa_next = sock->fasync_list;
L
Linus Torvalds 已提交
1035
		write_lock_bh(&sk->sk_callback_lock);
1036
		sock->fasync_list = fna;
L
Linus Torvalds 已提交
1037
		write_unlock_bh(&sk->sk_callback_lock);
1038 1039
	} else {
		if (fa != NULL) {
L
Linus Torvalds 已提交
1040
			write_lock_bh(&sk->sk_callback_lock);
1041
			*prev = fa->fa_next;
L
Linus Torvalds 已提交
1042 1043 1044 1045 1046 1047 1048 1049 1050 1051 1052 1053 1054 1055 1056 1057
			write_unlock_bh(&sk->sk_callback_lock);
			kfree(fa);
		}
	}

out:
	release_sock(sock->sk);
	return 0;
}

/* This function may be called only under socket lock or callback_lock */

int sock_wake_async(struct socket *sock, int how, int band)
{
	if (!sock || !sock->fasync_list)
		return -1;
1058
	switch (how) {
L
Linus Torvalds 已提交
1059
	case 1:
1060

L
Linus Torvalds 已提交
1061 1062 1063 1064 1065 1066 1067 1068
		if (test_bit(SOCK_ASYNC_WAITDATA, &sock->flags))
			break;
		goto call_kill;
	case 2:
		if (!test_and_clear_bit(SOCK_ASYNC_NOSPACE, &sock->flags))
			break;
		/* fall through */
	case 0:
1069
call_kill:
L
Linus Torvalds 已提交
1070 1071 1072 1073 1074 1075 1076 1077
		__kill_fasync(sock->fasync_list, SIGIO, band);
		break;
	case 3:
		__kill_fasync(sock->fasync_list, SIGURG, band);
	}
	return 0;
}

1078 1079
static int __sock_create(int family, int type, int protocol,
			 struct socket **res, int kern)
L
Linus Torvalds 已提交
1080 1081 1082
{
	int err;
	struct socket *sock;
1083
	const struct net_proto_family *pf;
L
Linus Torvalds 已提交
1084 1085

	/*
1086
	 *      Check protocol is in range
L
Linus Torvalds 已提交
1087 1088 1089 1090 1091 1092 1093 1094 1095 1096 1097 1098
	 */
	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) {
1099
		static int warned;
L
Linus Torvalds 已提交
1100 1101
		if (!warned) {
			warned = 1;
1102 1103
			printk(KERN_INFO "%s uses obsolete (PF_INET,SOCK_PACKET)\n",
			       current->comm);
L
Linus Torvalds 已提交
1104 1105 1106 1107 1108 1109 1110
		}
		family = PF_PACKET;
	}

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

1112 1113 1114 1115 1116 1117 1118 1119 1120 1121 1122 1123 1124 1125 1126
	/*
	 *	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) {
		if (net_ratelimit())
			printk(KERN_WARNING "socket: no more sockets\n");
		return -ENFILE;	/* Not exactly a match, but its the
				   closest posix thing */
	}

	sock->type = type;

L
Linus Torvalds 已提交
1127
#if defined(CONFIG_KMOD)
1128 1129 1130
	/* 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
1131 1132 1133
	 * requested real, full-featured networking support upon configuration.
	 * Otherwise module support will break!
	 */
1134
	if (net_families[family] == NULL)
1135
		request_module("net-pf-%d", family);
L
Linus Torvalds 已提交
1136 1137
#endif

1138 1139 1140 1141 1142
	rcu_read_lock();
	pf = rcu_dereference(net_families[family]);
	err = -EAFNOSUPPORT;
	if (!pf)
		goto out_release;
L
Linus Torvalds 已提交
1143 1144 1145 1146 1147

	/*
	 * We will call the ->create function, that possibly is in a loadable
	 * module, so we have to bump that loadable module refcnt first.
	 */
1148
	if (!try_module_get(pf->owner))
L
Linus Torvalds 已提交
1149 1150
		goto out_release;

1151 1152 1153 1154 1155
	/* Now protected by module ref count */
	rcu_read_unlock();

	err = pf->create(sock, protocol);
	if (err < 0)
L
Linus Torvalds 已提交
1156
		goto out_module_put;
1157

L
Linus Torvalds 已提交
1158 1159 1160 1161
	/*
	 * Now to bump the refcnt of the [loadable] module that owns this
	 * socket at sock_release time we decrement its refcnt.
	 */
1162 1163 1164
	if (!try_module_get(sock->ops->owner))
		goto out_module_busy;

L
Linus Torvalds 已提交
1165 1166 1167 1168
	/*
	 * Now that we're done with the ->create function, the [loadable]
	 * module can have its refcnt decremented
	 */
1169
	module_put(pf->owner);
V
Venkat Yekkirala 已提交
1170 1171 1172
	err = security_socket_post_create(sock, family, type, protocol, kern);
	if (err)
		goto out_release;
1173
	*res = sock;
L
Linus Torvalds 已提交
1174

1175 1176 1177 1178
	return 0;

out_module_busy:
	err = -EAFNOSUPPORT;
L
Linus Torvalds 已提交
1179
out_module_put:
1180 1181 1182
	sock->ops = NULL;
	module_put(pf->owner);
out_sock_release:
L
Linus Torvalds 已提交
1183
	sock_release(sock);
1184 1185 1186 1187 1188
	return err;

out_release:
	rcu_read_unlock();
	goto out_sock_release;
L
Linus Torvalds 已提交
1189 1190 1191 1192 1193 1194 1195 1196 1197 1198 1199 1200 1201 1202 1203 1204 1205 1206 1207 1208 1209 1210 1211 1212 1213 1214 1215 1216 1217 1218 1219 1220 1221 1222 1223 1224 1225 1226
}

int sock_create(int family, int type, int protocol, struct socket **res)
{
	return __sock_create(family, type, protocol, res, 0);
}

int sock_create_kern(int family, int type, int protocol, struct socket **res)
{
	return __sock_create(family, type, protocol, res, 1);
}

asmlinkage long sys_socket(int family, int type, int protocol)
{
	int retval;
	struct socket *sock;

	retval = sock_create(family, type, protocol, &sock);
	if (retval < 0)
		goto out;

	retval = sock_map_fd(sock);
	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.
 */

1227 1228
asmlinkage long sys_socketpair(int family, int type, int protocol,
			       int __user *usockvec)
L
Linus Torvalds 已提交
1229 1230 1231
{
	struct socket *sock1, *sock2;
	int fd1, fd2, err;
A
Al Viro 已提交
1232
	struct file *newfile1, *newfile2;
L
Linus Torvalds 已提交
1233 1234 1235 1236 1237 1238 1239 1240 1241 1242 1243 1244 1245 1246 1247

	/*
	 * 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);
1248
	if (err < 0)
L
Linus Torvalds 已提交
1249 1250
		goto out_release_both;

A
Al Viro 已提交
1251 1252 1253
	fd1 = sock_alloc_fd(&newfile1);
	if (unlikely(fd1 < 0))
		goto out_release_both;
L
Linus Torvalds 已提交
1254

A
Al Viro 已提交
1255 1256 1257 1258
	fd2 = sock_alloc_fd(&newfile2);
	if (unlikely(fd2 < 0)) {
		put_filp(newfile1);
		put_unused_fd(fd1);
L
Linus Torvalds 已提交
1259
		goto out_release_both;
A
Al Viro 已提交
1260
	}
L
Linus Torvalds 已提交
1261

A
Al Viro 已提交
1262 1263 1264 1265 1266 1267 1268 1269 1270 1271 1272 1273 1274 1275 1276 1277 1278
	err = sock_attach_fd(sock1, newfile1);
	if (unlikely(err < 0)) {
		goto out_fd2;
	}

	err = sock_attach_fd(sock2, newfile2);
	if (unlikely(err < 0)) {
		fput(newfile1);
		goto out_fd1;
	}

	err = audit_fd_pair(fd1, fd2);
	if (err < 0) {
		fput(newfile1);
		fput(newfile2);
		goto out_fd;
	}
L
Linus Torvalds 已提交
1279

A
Al Viro 已提交
1280 1281
	fd_install(fd1, newfile1);
	fd_install(fd2, newfile2);
L
Linus Torvalds 已提交
1282 1283 1284 1285
	/* fd1 and fd2 may be already another descriptors.
	 * Not kernel problem.
	 */

1286
	err = put_user(fd1, &usockvec[0]);
L
Linus Torvalds 已提交
1287 1288 1289 1290 1291 1292 1293 1294 1295 1296
	if (!err)
		err = put_user(fd2, &usockvec[1]);
	if (!err)
		return 0;

	sys_close(fd2);
	sys_close(fd1);
	return err;

out_release_both:
1297
	sock_release(sock2);
L
Linus Torvalds 已提交
1298
out_release_1:
1299
	sock_release(sock1);
L
Linus Torvalds 已提交
1300 1301
out:
	return err;
A
Al Viro 已提交
1302 1303 1304 1305 1306 1307 1308 1309 1310 1311 1312

out_fd2:
	put_filp(newfile1);
	sock_release(sock1);
out_fd1:
	put_filp(newfile2);
	sock_release(sock2);
out_fd:
	put_unused_fd(fd1);
	put_unused_fd(fd2);
	goto out;
L
Linus Torvalds 已提交
1313 1314 1315 1316 1317 1318 1319 1320 1321 1322 1323 1324 1325 1326
}

/*
 *	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).
 */

asmlinkage long sys_bind(int fd, struct sockaddr __user *umyaddr, int addrlen)
{
	struct socket *sock;
	char address[MAX_SOCK_ADDR];
1327
	int err, fput_needed;
L
Linus Torvalds 已提交
1328

1329
	sock = sockfd_lookup_light(fd, &err, &fput_needed);
1330
	if (sock) {
1331 1332 1333 1334 1335
		err = move_addr_to_kernel(umyaddr, addrlen, address);
		if (err >= 0) {
			err = security_socket_bind(sock,
						   (struct sockaddr *)address,
						   addrlen);
1336 1337
			if (!err)
				err = sock->ops->bind(sock,
1338 1339
						      (struct sockaddr *)
						      address, addrlen);
L
Linus Torvalds 已提交
1340
		}
1341
		fput_light(sock->file, fput_needed);
1342
	}
L
Linus Torvalds 已提交
1343 1344 1345 1346 1347 1348 1349 1350 1351
	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.
 */

1352
int sysctl_somaxconn __read_mostly = SOMAXCONN;
L
Linus Torvalds 已提交
1353 1354 1355 1356

asmlinkage long sys_listen(int fd, int backlog)
{
	struct socket *sock;
1357
	int err, fput_needed;
1358 1359 1360 1361

	sock = sockfd_lookup_light(fd, &err, &fput_needed);
	if (sock) {
		if ((unsigned)backlog > sysctl_somaxconn)
L
Linus Torvalds 已提交
1362 1363 1364
			backlog = sysctl_somaxconn;

		err = security_socket_listen(sock, backlog);
1365 1366
		if (!err)
			err = sock->ops->listen(sock, backlog);
L
Linus Torvalds 已提交
1367

1368
		fput_light(sock->file, fput_needed);
L
Linus Torvalds 已提交
1369 1370 1371 1372 1373 1374 1375 1376 1377 1378 1379 1380 1381 1382 1383 1384
	}
	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.
 */

1385 1386
asmlinkage long sys_accept(int fd, struct sockaddr __user *upeer_sockaddr,
			   int __user *upeer_addrlen)
L
Linus Torvalds 已提交
1387 1388
{
	struct socket *sock, *newsock;
1389
	struct file *newfile;
1390
	int err, len, newfd, fput_needed;
L
Linus Torvalds 已提交
1391 1392
	char address[MAX_SOCK_ADDR];

1393
	sock = sockfd_lookup_light(fd, &err, &fput_needed);
L
Linus Torvalds 已提交
1394 1395 1396 1397
	if (!sock)
		goto out;

	err = -ENFILE;
1398
	if (!(newsock = sock_alloc()))
L
Linus Torvalds 已提交
1399 1400 1401 1402 1403 1404 1405 1406 1407 1408 1409
		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);

1410 1411 1412
	newfd = sock_alloc_fd(&newfile);
	if (unlikely(newfd < 0)) {
		err = newfd;
1413 1414
		sock_release(newsock);
		goto out_put;
1415 1416 1417 1418
	}

	err = sock_attach_fd(newsock, newfile);
	if (err < 0)
1419
		goto out_fd_simple;
1420

1421 1422
	err = security_socket_accept(sock, newsock);
	if (err)
1423
		goto out_fd;
1424

L
Linus Torvalds 已提交
1425 1426
	err = sock->ops->accept(sock, newsock, sock->file->f_flags);
	if (err < 0)
1427
		goto out_fd;
L
Linus Torvalds 已提交
1428 1429

	if (upeer_sockaddr) {
1430 1431
		if (newsock->ops->getname(newsock, (struct sockaddr *)address,
					  &len, 2) < 0) {
L
Linus Torvalds 已提交
1432
			err = -ECONNABORTED;
1433
			goto out_fd;
L
Linus Torvalds 已提交
1434
		}
1435 1436
		err = move_addr_to_user(address, len, upeer_sockaddr,
					upeer_addrlen);
L
Linus Torvalds 已提交
1437
		if (err < 0)
1438
			goto out_fd;
L
Linus Torvalds 已提交
1439 1440 1441 1442
	}

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

1443 1444
	fd_install(newfd, newfile);
	err = newfd;
L
Linus Torvalds 已提交
1445 1446 1447 1448

	security_socket_post_accept(sock, newsock);

out_put:
1449
	fput_light(sock->file, fput_needed);
L
Linus Torvalds 已提交
1450 1451
out:
	return err;
1452 1453 1454 1455 1456
out_fd_simple:
	sock_release(newsock);
	put_filp(newfile);
	put_unused_fd(newfd);
	goto out_put;
1457
out_fd:
1458
	fput(newfile);
1459
	put_unused_fd(newfd);
L
Linus Torvalds 已提交
1460 1461 1462 1463 1464 1465 1466 1467 1468 1469 1470 1471 1472 1473 1474
	goto out_put;
}

/*
 *	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.
 */

1475 1476
asmlinkage long sys_connect(int fd, struct sockaddr __user *uservaddr,
			    int addrlen)
L
Linus Torvalds 已提交
1477 1478 1479
{
	struct socket *sock;
	char address[MAX_SOCK_ADDR];
1480
	int err, fput_needed;
L
Linus Torvalds 已提交
1481

1482
	sock = sockfd_lookup_light(fd, &err, &fput_needed);
L
Linus Torvalds 已提交
1483 1484 1485 1486 1487 1488
	if (!sock)
		goto out;
	err = move_addr_to_kernel(uservaddr, addrlen, address);
	if (err < 0)
		goto out_put;

1489 1490
	err =
	    security_socket_connect(sock, (struct sockaddr *)address, addrlen);
L
Linus Torvalds 已提交
1491 1492 1493
	if (err)
		goto out_put;

1494
	err = sock->ops->connect(sock, (struct sockaddr *)address, addrlen,
L
Linus Torvalds 已提交
1495 1496
				 sock->file->f_flags);
out_put:
1497
	fput_light(sock->file, fput_needed);
L
Linus Torvalds 已提交
1498 1499 1500 1501 1502 1503 1504 1505 1506
out:
	return err;
}

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

1507 1508
asmlinkage long sys_getsockname(int fd, struct sockaddr __user *usockaddr,
				int __user *usockaddr_len)
L
Linus Torvalds 已提交
1509 1510 1511
{
	struct socket *sock;
	char address[MAX_SOCK_ADDR];
1512
	int len, err, fput_needed;
1513

1514
	sock = sockfd_lookup_light(fd, &err, &fput_needed);
L
Linus Torvalds 已提交
1515 1516 1517 1518 1519 1520 1521 1522 1523 1524 1525 1526 1527
	if (!sock)
		goto out;

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

	err = sock->ops->getname(sock, (struct sockaddr *)address, &len, 0);
	if (err)
		goto out_put;
	err = move_addr_to_user(address, len, usockaddr, usockaddr_len);

out_put:
1528
	fput_light(sock->file, fput_needed);
L
Linus Torvalds 已提交
1529 1530 1531 1532 1533 1534 1535 1536 1537
out:
	return err;
}

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

1538 1539
asmlinkage long sys_getpeername(int fd, struct sockaddr __user *usockaddr,
				int __user *usockaddr_len)
L
Linus Torvalds 已提交
1540 1541 1542
{
	struct socket *sock;
	char address[MAX_SOCK_ADDR];
1543
	int len, err, fput_needed;
L
Linus Torvalds 已提交
1544

1545 1546
	sock = sockfd_lookup_light(fd, &err, &fput_needed);
	if (sock != NULL) {
L
Linus Torvalds 已提交
1547 1548
		err = security_socket_getpeername(sock);
		if (err) {
1549
			fput_light(sock->file, fput_needed);
L
Linus Torvalds 已提交
1550 1551 1552
			return err;
		}

1553 1554 1555
		err =
		    sock->ops->getname(sock, (struct sockaddr *)address, &len,
				       1);
L
Linus Torvalds 已提交
1556
		if (!err)
1557 1558
			err = move_addr_to_user(address, len, usockaddr,
						usockaddr_len);
1559
		fput_light(sock->file, fput_needed);
L
Linus Torvalds 已提交
1560 1561 1562 1563 1564 1565 1566 1567 1568 1569
	}
	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.
 */

1570 1571 1572
asmlinkage long sys_sendto(int fd, void __user *buff, size_t len,
			   unsigned flags, struct sockaddr __user *addr,
			   int addr_len)
L
Linus Torvalds 已提交
1573 1574 1575 1576 1577 1578
{
	struct socket *sock;
	char address[MAX_SOCK_ADDR];
	int err;
	struct msghdr msg;
	struct iovec iov;
1579 1580 1581 1582
	int fput_needed;
	struct file *sock_file;

	sock_file = fget_light(fd, &fput_needed);
1583
	err = -EBADF;
1584
	if (!sock_file)
1585
		goto out;
1586 1587

	sock = sock_from_file(sock_file, &err);
L
Linus Torvalds 已提交
1588
	if (!sock)
1589
		goto out_put;
1590 1591 1592 1593 1594 1595 1596 1597
	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;
1598
	if (addr) {
L
Linus Torvalds 已提交
1599 1600 1601
		err = move_addr_to_kernel(addr, addr_len, address);
		if (err < 0)
			goto out_put;
1602 1603
		msg.msg_name = address;
		msg.msg_namelen = addr_len;
L
Linus Torvalds 已提交
1604 1605 1606 1607 1608 1609
	}
	if (sock->file->f_flags & O_NONBLOCK)
		flags |= MSG_DONTWAIT;
	msg.msg_flags = flags;
	err = sock_sendmsg(sock, &msg, len);

1610
out_put:
1611
	fput_light(sock_file, fput_needed);
1612
out:
L
Linus Torvalds 已提交
1613 1614 1615 1616
	return err;
}

/*
1617
 *	Send a datagram down a socket.
L
Linus Torvalds 已提交
1618 1619
 */

1620
asmlinkage long sys_send(int fd, void __user *buff, size_t len, unsigned flags)
L
Linus Torvalds 已提交
1621 1622 1623 1624 1625
{
	return sys_sendto(fd, buff, len, flags, NULL, 0);
}

/*
1626
 *	Receive a frame from the socket and optionally record the address of the
L
Linus Torvalds 已提交
1627 1628 1629 1630
 *	sender. We verify the buffers are writable and if needed move the
 *	sender address from kernel to user space.
 */

1631 1632 1633
asmlinkage long sys_recvfrom(int fd, void __user *ubuf, size_t size,
			     unsigned flags, struct sockaddr __user *addr,
			     int __user *addr_len)
L
Linus Torvalds 已提交
1634 1635 1636 1637 1638
{
	struct socket *sock;
	struct iovec iov;
	struct msghdr msg;
	char address[MAX_SOCK_ADDR];
1639
	int err, err2;
1640 1641 1642 1643
	struct file *sock_file;
	int fput_needed;

	sock_file = fget_light(fd, &fput_needed);
1644
	err = -EBADF;
1645
	if (!sock_file)
1646
		goto out;
L
Linus Torvalds 已提交
1647

1648
	sock = sock_from_file(sock_file, &err);
L
Linus Torvalds 已提交
1649
	if (!sock)
1650
		goto out_put;
L
Linus Torvalds 已提交
1651

1652 1653 1654 1655 1656 1657 1658 1659
	msg.msg_control = NULL;
	msg.msg_controllen = 0;
	msg.msg_iovlen = 1;
	msg.msg_iov = &iov;
	iov.iov_len = size;
	iov.iov_base = ubuf;
	msg.msg_name = address;
	msg.msg_namelen = MAX_SOCK_ADDR;
L
Linus Torvalds 已提交
1660 1661
	if (sock->file->f_flags & O_NONBLOCK)
		flags |= MSG_DONTWAIT;
1662
	err = sock_recvmsg(sock, &msg, size, flags);
L
Linus Torvalds 已提交
1663

1664 1665 1666 1667
	if (err >= 0 && addr != NULL) {
		err2 = move_addr_to_user(address, msg.msg_namelen, addr, addr_len);
		if (err2 < 0)
			err = err2;
L
Linus Torvalds 已提交
1668
	}
1669
out_put:
1670
	fput_light(sock_file, fput_needed);
1671
out:
L
Linus Torvalds 已提交
1672 1673 1674 1675
	return err;
}

/*
1676
 *	Receive a datagram from a socket.
L
Linus Torvalds 已提交
1677 1678
 */

1679 1680
asmlinkage long sys_recv(int fd, void __user *ubuf, size_t size,
			 unsigned flags)
L
Linus Torvalds 已提交
1681 1682 1683 1684 1685 1686 1687 1688 1689
{
	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.
 */

1690 1691
asmlinkage long sys_setsockopt(int fd, int level, int optname,
			       char __user *optval, int optlen)
L
Linus Torvalds 已提交
1692
{
1693
	int err, fput_needed;
L
Linus Torvalds 已提交
1694 1695 1696 1697
	struct socket *sock;

	if (optlen < 0)
		return -EINVAL;
1698 1699 1700 1701

	sock = sockfd_lookup_light(fd, &err, &fput_needed);
	if (sock != NULL) {
		err = security_socket_setsockopt(sock, level, optname);
1702 1703
		if (err)
			goto out_put;
L
Linus Torvalds 已提交
1704 1705

		if (level == SOL_SOCKET)
1706 1707 1708
			err =
			    sock_setsockopt(sock, level, optname, optval,
					    optlen);
L
Linus Torvalds 已提交
1709
		else
1710 1711 1712
			err =
			    sock->ops->setsockopt(sock, level, optname, optval,
						  optlen);
1713 1714
out_put:
		fput_light(sock->file, fput_needed);
L
Linus Torvalds 已提交
1715 1716 1717 1718 1719 1720 1721 1722 1723
	}
	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.
 */

1724 1725
asmlinkage long sys_getsockopt(int fd, int level, int optname,
			       char __user *optval, int __user *optlen)
L
Linus Torvalds 已提交
1726
{
1727
	int err, fput_needed;
L
Linus Torvalds 已提交
1728 1729
	struct socket *sock;

1730 1731
	sock = sockfd_lookup_light(fd, &err, &fput_needed);
	if (sock != NULL) {
1732 1733 1734
		err = security_socket_getsockopt(sock, level, optname);
		if (err)
			goto out_put;
L
Linus Torvalds 已提交
1735 1736

		if (level == SOL_SOCKET)
1737 1738 1739
			err =
			    sock_getsockopt(sock, level, optname, optval,
					    optlen);
L
Linus Torvalds 已提交
1740
		else
1741 1742 1743
			err =
			    sock->ops->getsockopt(sock, level, optname, optval,
						  optlen);
1744 1745
out_put:
		fput_light(sock->file, fput_needed);
L
Linus Torvalds 已提交
1746 1747 1748 1749 1750 1751 1752 1753 1754 1755
	}
	return err;
}

/*
 *	Shutdown a socket.
 */

asmlinkage long sys_shutdown(int fd, int how)
{
1756
	int err, fput_needed;
L
Linus Torvalds 已提交
1757 1758
	struct socket *sock;

1759 1760
	sock = sockfd_lookup_light(fd, &err, &fput_needed);
	if (sock != NULL) {
L
Linus Torvalds 已提交
1761
		err = security_socket_shutdown(sock, how);
1762 1763 1764
		if (!err)
			err = sock->ops->shutdown(sock, how);
		fput_light(sock->file, fput_needed);
L
Linus Torvalds 已提交
1765 1766 1767 1768
	}
	return err;
}

1769
/* A couple of helpful macros for getting the address of the 32/64 bit
L
Linus Torvalds 已提交
1770 1771 1772 1773 1774 1775 1776 1777 1778 1779 1780 1781
 * 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)

/*
 *	BSD sendmsg interface
 */

asmlinkage long sys_sendmsg(int fd, struct msghdr __user *msg, unsigned flags)
{
1782 1783
	struct compat_msghdr __user *msg_compat =
	    (struct compat_msghdr __user *)msg;
L
Linus Torvalds 已提交
1784 1785 1786
	struct socket *sock;
	char address[MAX_SOCK_ADDR];
	struct iovec iovstack[UIO_FASTIOV], *iov = iovstack;
1787
	unsigned char ctl[sizeof(struct cmsghdr) + 20]
1788 1789
	    __attribute__ ((aligned(sizeof(__kernel_size_t))));
	/* 20 is size of ipv6_pktinfo */
L
Linus Torvalds 已提交
1790 1791 1792
	unsigned char *ctl_buf = ctl;
	struct msghdr msg_sys;
	int err, ctl_len, iov_size, total_len;
1793
	int fput_needed;
1794

L
Linus Torvalds 已提交
1795 1796 1797 1798
	err = -EFAULT;
	if (MSG_CMSG_COMPAT & flags) {
		if (get_compat_msghdr(&msg_sys, msg_compat))
			return -EFAULT;
1799 1800
	}
	else if (copy_from_user(&msg_sys, msg, sizeof(struct msghdr)))
L
Linus Torvalds 已提交
1801 1802
		return -EFAULT;

1803
	sock = sockfd_lookup_light(fd, &err, &fput_needed);
1804
	if (!sock)
L
Linus Torvalds 已提交
1805 1806 1807 1808 1809 1810 1811
		goto out;

	/* do not move before msg_sys is valid */
	err = -EMSGSIZE;
	if (msg_sys.msg_iovlen > UIO_MAXIOV)
		goto out_put;

1812
	/* Check whether to allocate the iovec area */
L
Linus Torvalds 已提交
1813 1814 1815 1816 1817 1818 1819 1820 1821 1822 1823 1824 1825
	err = -ENOMEM;
	iov_size = msg_sys.msg_iovlen * sizeof(struct iovec);
	if (msg_sys.msg_iovlen > UIO_FASTIOV) {
		iov = sock_kmalloc(sock->sk, iov_size, GFP_KERNEL);
		if (!iov)
			goto out_put;
	}

	/* This will also move the address data into kernel space */
	if (MSG_CMSG_COMPAT & flags) {
		err = verify_compat_iovec(&msg_sys, iov, address, VERIFY_READ);
	} else
		err = verify_iovec(&msg_sys, iov, address, VERIFY_READ);
1826
	if (err < 0)
L
Linus Torvalds 已提交
1827 1828 1829 1830 1831 1832 1833
		goto out_freeiov;
	total_len = err;

	err = -ENOBUFS;

	if (msg_sys.msg_controllen > INT_MAX)
		goto out_freeiov;
1834
	ctl_len = msg_sys.msg_controllen;
L
Linus Torvalds 已提交
1835
	if ((MSG_CMSG_COMPAT & flags) && ctl_len) {
1836 1837 1838
		err =
		    cmsghdr_from_user_compat_to_kern(&msg_sys, sock->sk, ctl,
						     sizeof(ctl));
L
Linus Torvalds 已提交
1839 1840 1841
		if (err)
			goto out_freeiov;
		ctl_buf = msg_sys.msg_control;
A
Al Viro 已提交
1842
		ctl_len = msg_sys.msg_controllen;
L
Linus Torvalds 已提交
1843
	} else if (ctl_len) {
1844
		if (ctl_len > sizeof(ctl)) {
L
Linus Torvalds 已提交
1845
			ctl_buf = sock_kmalloc(sock->sk, ctl_len, GFP_KERNEL);
1846
			if (ctl_buf == NULL)
L
Linus Torvalds 已提交
1847 1848 1849 1850 1851 1852 1853 1854
				goto out_freeiov;
		}
		err = -EFAULT;
		/*
		 * Careful! Before this, msg_sys.msg_control contains a user pointer.
		 * Afterwards, it will be a kernel pointer. Thus the compiler-assisted
		 * checking falls down on this.
		 */
1855 1856
		if (copy_from_user(ctl_buf, (void __user *)msg_sys.msg_control,
				   ctl_len))
L
Linus Torvalds 已提交
1857 1858 1859 1860 1861 1862 1863 1864 1865 1866
			goto out_freectl;
		msg_sys.msg_control = ctl_buf;
	}
	msg_sys.msg_flags = flags;

	if (sock->file->f_flags & O_NONBLOCK)
		msg_sys.msg_flags |= MSG_DONTWAIT;
	err = sock_sendmsg(sock, &msg_sys, total_len);

out_freectl:
1867
	if (ctl_buf != ctl)
L
Linus Torvalds 已提交
1868 1869 1870 1871 1872
		sock_kfree_s(sock->sk, ctl_buf, ctl_len);
out_freeiov:
	if (iov != iovstack)
		sock_kfree_s(sock->sk, iov, iov_size);
out_put:
1873
	fput_light(sock->file, fput_needed);
1874
out:
L
Linus Torvalds 已提交
1875 1876 1877 1878 1879 1880 1881
	return err;
}

/*
 *	BSD recvmsg interface
 */

1882 1883
asmlinkage long sys_recvmsg(int fd, struct msghdr __user *msg,
			    unsigned int flags)
L
Linus Torvalds 已提交
1884
{
1885 1886
	struct compat_msghdr __user *msg_compat =
	    (struct compat_msghdr __user *)msg;
L
Linus Torvalds 已提交
1887 1888
	struct socket *sock;
	struct iovec iovstack[UIO_FASTIOV];
1889
	struct iovec *iov = iovstack;
L
Linus Torvalds 已提交
1890 1891 1892
	struct msghdr msg_sys;
	unsigned long cmsg_ptr;
	int err, iov_size, total_len, len;
1893
	int fput_needed;
L
Linus Torvalds 已提交
1894 1895 1896 1897 1898 1899 1900

	/* kernel mode address */
	char addr[MAX_SOCK_ADDR];

	/* user mode address pointers */
	struct sockaddr __user *uaddr;
	int __user *uaddr_len;
1901

L
Linus Torvalds 已提交
1902 1903 1904
	if (MSG_CMSG_COMPAT & flags) {
		if (get_compat_msghdr(&msg_sys, msg_compat))
			return -EFAULT;
1905 1906 1907
	}
	else if (copy_from_user(&msg_sys, msg, sizeof(struct msghdr)))
		return -EFAULT;
L
Linus Torvalds 已提交
1908

1909
	sock = sockfd_lookup_light(fd, &err, &fput_needed);
L
Linus Torvalds 已提交
1910 1911 1912 1913 1914 1915
	if (!sock)
		goto out;

	err = -EMSGSIZE;
	if (msg_sys.msg_iovlen > UIO_MAXIOV)
		goto out_put;
1916 1917

	/* Check whether to allocate the iovec area */
L
Linus Torvalds 已提交
1918 1919 1920 1921 1922 1923 1924 1925 1926
	err = -ENOMEM;
	iov_size = msg_sys.msg_iovlen * sizeof(struct iovec);
	if (msg_sys.msg_iovlen > UIO_FASTIOV) {
		iov = sock_kmalloc(sock->sk, iov_size, GFP_KERNEL);
		if (!iov)
			goto out_put;
	}

	/*
1927 1928
	 *      Save the user-mode address (verify_iovec will change the
	 *      kernel msghdr to use the kernel address space)
L
Linus Torvalds 已提交
1929
	 */
1930 1931

	uaddr = (void __user *)msg_sys.msg_name;
L
Linus Torvalds 已提交
1932 1933 1934 1935 1936 1937 1938
	uaddr_len = COMPAT_NAMELEN(msg);
	if (MSG_CMSG_COMPAT & flags) {
		err = verify_compat_iovec(&msg_sys, iov, addr, VERIFY_WRITE);
	} else
		err = verify_iovec(&msg_sys, iov, addr, VERIFY_WRITE);
	if (err < 0)
		goto out_freeiov;
1939
	total_len = err;
L
Linus Torvalds 已提交
1940 1941 1942 1943 1944

	cmsg_ptr = (unsigned long)msg_sys.msg_control;
	msg_sys.msg_flags = 0;
	if (MSG_CMSG_COMPAT & flags)
		msg_sys.msg_flags = MSG_CMSG_COMPAT;
1945

L
Linus Torvalds 已提交
1946 1947 1948 1949 1950 1951 1952 1953
	if (sock->file->f_flags & O_NONBLOCK)
		flags |= MSG_DONTWAIT;
	err = sock_recvmsg(sock, &msg_sys, total_len, flags);
	if (err < 0)
		goto out_freeiov;
	len = err;

	if (uaddr != NULL) {
1954 1955
		err = move_addr_to_user(addr, msg_sys.msg_namelen, uaddr,
					uaddr_len);
L
Linus Torvalds 已提交
1956 1957 1958
		if (err < 0)
			goto out_freeiov;
	}
1959 1960
	err = __put_user((msg_sys.msg_flags & ~MSG_CMSG_COMPAT),
			 COMPAT_FLAGS(msg));
L
Linus Torvalds 已提交
1961 1962 1963
	if (err)
		goto out_freeiov;
	if (MSG_CMSG_COMPAT & flags)
1964
		err = __put_user((unsigned long)msg_sys.msg_control - cmsg_ptr,
L
Linus Torvalds 已提交
1965 1966
				 &msg_compat->msg_controllen);
	else
1967
		err = __put_user((unsigned long)msg_sys.msg_control - cmsg_ptr,
L
Linus Torvalds 已提交
1968 1969 1970 1971 1972 1973 1974 1975 1976
				 &msg->msg_controllen);
	if (err)
		goto out_freeiov;
	err = len;

out_freeiov:
	if (iov != iovstack)
		sock_kfree_s(sock->sk, iov, iov_size);
out_put:
1977
	fput_light(sock->file, fput_needed);
L
Linus Torvalds 已提交
1978 1979 1980 1981 1982 1983 1984 1985
out:
	return err;
}

#ifdef __ARCH_WANT_SYS_SOCKETCALL

/* Argument list sizes for sys_socketcall */
#define AL(x) ((x) * sizeof(unsigned long))
1986 1987 1988 1989 1990 1991
static const unsigned char nargs[18]={
	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)
};

L
Linus Torvalds 已提交
1992 1993 1994
#undef AL

/*
1995
 *	System call vectors.
L
Linus Torvalds 已提交
1996 1997 1998
 *
 *	Argument checking cleaned up. Saved 20% in size.
 *  This function doesn't need to set the kernel lock because
1999
 *  it is set by the callees.
L
Linus Torvalds 已提交
2000 2001 2002 2003 2004
 */

asmlinkage long sys_socketcall(int call, unsigned long __user *args)
{
	unsigned long a[6];
2005
	unsigned long a0, a1;
L
Linus Torvalds 已提交
2006 2007
	int err;

2008
	if (call < 1 || call > SYS_RECVMSG)
L
Linus Torvalds 已提交
2009 2010 2011 2012 2013
		return -EINVAL;

	/* copy_from_user should be SMP safe. */
	if (copy_from_user(a, args, nargs[call]))
		return -EFAULT;
2014

2015
	err = audit_socketcall(nargs[call] / sizeof(unsigned long), a);
2016 2017 2018
	if (err)
		return err;

2019 2020 2021 2022 2023 2024 2025 2026 2027 2028 2029 2030 2031 2032 2033 2034 2035 2036 2037 2038 2039 2040 2041 2042 2043 2044 2045 2046 2047 2048 2049 2050 2051 2052 2053 2054 2055 2056 2057 2058 2059 2060 2061 2062 2063 2064 2065 2066 2067 2068 2069 2070 2071 2072 2073 2074 2075 2076 2077 2078 2079 2080 2081 2082 2083 2084 2085 2086 2087
	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:
		err =
		    sys_accept(a0, (struct sockaddr __user *)a1,
			       (int __user *)a[2]);
		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;
	case SYS_RECVMSG:
		err = sys_recvmsg(a0, (struct msghdr __user *)a1, a[2]);
		break;
	default:
		err = -EINVAL;
		break;
L
Linus Torvalds 已提交
2088 2089 2090 2091
	}
	return err;
}

2092
#endif				/* __ARCH_WANT_SYS_SOCKETCALL */
L
Linus Torvalds 已提交
2093

2094 2095 2096 2097
/**
 *	sock_register - add a socket protocol handler
 *	@ops: description of protocol
 *
L
Linus Torvalds 已提交
2098 2099
 *	This function is called by a protocol handler that wants to
 *	advertise its address family, and have it linked into the
2100 2101
 *	socket interface. The value ops->family coresponds to the
 *	socket system call protocol family.
L
Linus Torvalds 已提交
2102
 */
2103
int sock_register(const struct net_proto_family *ops)
L
Linus Torvalds 已提交
2104 2105 2106 2107
{
	int err;

	if (ops->family >= NPROTO) {
2108 2109
		printk(KERN_CRIT "protocol %d >= NPROTO(%d)\n", ops->family,
		       NPROTO);
L
Linus Torvalds 已提交
2110 2111
		return -ENOBUFS;
	}
2112 2113 2114 2115 2116

	spin_lock(&net_family_lock);
	if (net_families[ops->family])
		err = -EEXIST;
	else {
2117
		net_families[ops->family] = ops;
L
Linus Torvalds 已提交
2118 2119
		err = 0;
	}
2120 2121
	spin_unlock(&net_family_lock);

2122
	printk(KERN_INFO "NET: Registered protocol family %d\n", ops->family);
L
Linus Torvalds 已提交
2123 2124 2125
	return err;
}

2126 2127 2128 2129
/**
 *	sock_unregister - remove a protocol handler
 *	@family: protocol family to remove
 *
L
Linus Torvalds 已提交
2130 2131
 *	This function is called by a protocol handler that wants to
 *	remove its address family, and have it unlinked from the
2132 2133 2134 2135 2136 2137
 *	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
2138
 */
2139
void sock_unregister(int family)
L
Linus Torvalds 已提交
2140
{
2141
	BUG_ON(family < 0 || family >= NPROTO);
L
Linus Torvalds 已提交
2142

2143
	spin_lock(&net_family_lock);
2144
	net_families[family] = NULL;
2145 2146 2147 2148
	spin_unlock(&net_family_lock);

	synchronize_rcu();

2149
	printk(KERN_INFO "NET: Unregistered protocol family %d\n", family);
L
Linus Torvalds 已提交
2150 2151
}

2152
static int __init sock_init(void)
L
Linus Torvalds 已提交
2153 2154
{
	/*
2155
	 *      Initialize sock SLAB cache.
L
Linus Torvalds 已提交
2156
	 */
2157

L
Linus Torvalds 已提交
2158 2159 2160
	sk_init();

	/*
2161
	 *      Initialize skbuff SLAB cache
L
Linus Torvalds 已提交
2162 2163 2164 2165
	 */
	skb_init();

	/*
2166
	 *      Initialize the protocols module.
L
Linus Torvalds 已提交
2167 2168 2169 2170 2171
	 */

	init_inodecache();
	register_filesystem(&sock_fs_type);
	sock_mnt = kern_mount(&sock_fs_type);
2172 2173

	/* The real protocol initialization is performed in later initcalls.
L
Linus Torvalds 已提交
2174 2175 2176 2177 2178
	 */

#ifdef CONFIG_NETFILTER
	netfilter_init();
#endif
2179 2180

	return 0;
L
Linus Torvalds 已提交
2181 2182
}

2183 2184
core_initcall(sock_init);	/* early initcall */

L
Linus Torvalds 已提交
2185 2186 2187 2188 2189 2190
#ifdef CONFIG_PROC_FS
void socket_seq_show(struct seq_file *seq)
{
	int cpu;
	int counter = 0;

2191
	for_each_possible_cpu(cpu)
2192
	    counter += per_cpu(sockets_in_use, cpu);
L
Linus Torvalds 已提交
2193 2194 2195 2196 2197 2198 2199

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

	seq_printf(seq, "sockets: used %d\n", counter);
}
2200
#endif				/* CONFIG_PROC_FS */
L
Linus Torvalds 已提交
2201

2202 2203
#ifdef CONFIG_COMPAT
static long compat_sock_ioctl(struct file *file, unsigned cmd,
2204
			      unsigned long arg)
2205 2206 2207 2208 2209 2210 2211 2212 2213 2214 2215
{
	struct socket *sock = file->private_data;
	int ret = -ENOIOCTLCMD;

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

	return ret;
}
#endif

2216 2217 2218 2219 2220 2221 2222 2223 2224 2225 2226 2227 2228 2229 2230 2231 2232 2233 2234 2235 2236 2237 2238 2239 2240 2241 2242 2243 2244 2245 2246 2247 2248
int kernel_bind(struct socket *sock, struct sockaddr *addr, int addrlen)
{
	return sock->ops->bind(sock, addr, addrlen);
}

int kernel_listen(struct socket *sock, int backlog)
{
	return sock->ops->listen(sock, backlog);
}

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);
		goto done;
	}

	(*newsock)->ops = sock->ops;

done:
	return err;
}

int kernel_connect(struct socket *sock, struct sockaddr *addr, int addrlen,
2249
		   int flags)
2250 2251 2252 2253 2254 2255 2256 2257 2258 2259 2260 2261 2262 2263 2264 2265 2266 2267 2268 2269 2270 2271 2272 2273 2274 2275 2276 2277 2278 2279 2280 2281 2282 2283 2284 2285 2286 2287 2288 2289 2290 2291 2292 2293 2294 2295 2296 2297 2298 2299 2300 2301 2302 2303 2304 2305 2306 2307 2308 2309 2310 2311 2312 2313 2314 2315 2316 2317 2318
{
	return sock->ops->connect(sock, addr, addrlen, flags);
}

int kernel_getsockname(struct socket *sock, struct sockaddr *addr,
			 int *addrlen)
{
	return sock->ops->getname(sock, addr, addrlen, 0);
}

int kernel_getpeername(struct socket *sock, struct sockaddr *addr,
			 int *addrlen)
{
	return sock->ops->getname(sock, addr, addrlen, 1);
}

int kernel_getsockopt(struct socket *sock, int level, int optname,
			char *optval, int *optlen)
{
	mm_segment_t oldfs = get_fs();
	int err;

	set_fs(KERNEL_DS);
	if (level == SOL_SOCKET)
		err = sock_getsockopt(sock, level, optname, optval, optlen);
	else
		err = sock->ops->getsockopt(sock, level, optname, optval,
					    optlen);
	set_fs(oldfs);
	return err;
}

int kernel_setsockopt(struct socket *sock, int level, int optname,
			char *optval, int optlen)
{
	mm_segment_t oldfs = get_fs();
	int err;

	set_fs(KERNEL_DS);
	if (level == SOL_SOCKET)
		err = sock_setsockopt(sock, level, optname, optval, optlen);
	else
		err = sock->ops->setsockopt(sock, level, optname, optval,
					    optlen);
	set_fs(oldfs);
	return err;
}

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);
}

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;
}

L
Linus Torvalds 已提交
2319 2320 2321 2322 2323 2324 2325 2326 2327 2328 2329 2330 2331 2332 2333 2334
/* ABI emulation layers need these two */
EXPORT_SYMBOL(move_addr_to_kernel);
EXPORT_SYMBOL(move_addr_to_user);
EXPORT_SYMBOL(sock_create);
EXPORT_SYMBOL(sock_create_kern);
EXPORT_SYMBOL(sock_create_lite);
EXPORT_SYMBOL(sock_map_fd);
EXPORT_SYMBOL(sock_recvmsg);
EXPORT_SYMBOL(sock_register);
EXPORT_SYMBOL(sock_release);
EXPORT_SYMBOL(sock_sendmsg);
EXPORT_SYMBOL(sock_unregister);
EXPORT_SYMBOL(sock_wake_async);
EXPORT_SYMBOL(sockfd_lookup);
EXPORT_SYMBOL(kernel_sendmsg);
EXPORT_SYMBOL(kernel_recvmsg);
2335 2336 2337 2338 2339 2340 2341 2342 2343 2344
EXPORT_SYMBOL(kernel_bind);
EXPORT_SYMBOL(kernel_listen);
EXPORT_SYMBOL(kernel_accept);
EXPORT_SYMBOL(kernel_connect);
EXPORT_SYMBOL(kernel_getsockname);
EXPORT_SYMBOL(kernel_getpeername);
EXPORT_SYMBOL(kernel_getsockopt);
EXPORT_SYMBOL(kernel_setsockopt);
EXPORT_SYMBOL(kernel_sendpage);
EXPORT_SYMBOL(kernel_sock_ioctl);