sock.c 62.9 KB
Newer Older
L
Linus Torvalds 已提交
1 2 3 4 5 6 7 8 9
/*
 * INET		An implementation of the TCP/IP protocol suite for the LINUX
 *		operating system.  INET is implemented using the  BSD Socket
 *		interface as the means of communication with the user level.
 *
 *		Generic socket support routines. Memory allocators, socket lock/release
 *		handler for protocols to use and generic option handler.
 *
 *
10
 * Authors:	Ross Biro
L
Linus Torvalds 已提交
11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34
 *		Fred N. van Kempen, <waltje@uWalt.NL.Mugnet.ORG>
 *		Florian La Roche, <flla@stud.uni-sb.de>
 *		Alan Cox, <A.Cox@swansea.ac.uk>
 *
 * Fixes:
 *		Alan Cox	: 	Numerous verify_area() problems
 *		Alan Cox	:	Connecting on a connecting socket
 *					now returns an error for tcp.
 *		Alan Cox	:	sock->protocol is set correctly.
 *					and is not sometimes left as 0.
 *		Alan Cox	:	connect handles icmp errors on a
 *					connect properly. Unfortunately there
 *					is a restart syscall nasty there. I
 *					can't match BSD without hacking the C
 *					library. Ideas urgently sought!
 *		Alan Cox	:	Disallow bind() to addresses that are
 *					not ours - especially broadcast ones!!
 *		Alan Cox	:	Socket 1024 _IS_ ok for users. (fencepost)
 *		Alan Cox	:	sock_wfree/sock_rfree don't destroy sockets,
 *					instead they leave that for the DESTROY timer.
 *		Alan Cox	:	Clean up error flag in accept
 *		Alan Cox	:	TCP ack handling is buggy, the DESTROY timer
 *					was buggy. Put a remove_sock() in the handler
 *					for memory when we hit 0. Also altered the timer
35
 *					code. The ACK stuff can wait and needs major
L
Linus Torvalds 已提交
36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91
 *					TCP layer surgery.
 *		Alan Cox	:	Fixed TCP ack bug, removed remove sock
 *					and fixed timer/inet_bh race.
 *		Alan Cox	:	Added zapped flag for TCP
 *		Alan Cox	:	Move kfree_skb into skbuff.c and tidied up surplus code
 *		Alan Cox	:	for new sk_buff allocations wmalloc/rmalloc now call alloc_skb
 *		Alan Cox	:	kfree_s calls now are kfree_skbmem so we can track skb resources
 *		Alan Cox	:	Supports socket option broadcast now as does udp. Packet and raw need fixing.
 *		Alan Cox	:	Added RCVBUF,SNDBUF size setting. It suddenly occurred to me how easy it was so...
 *		Rick Sladkey	:	Relaxed UDP rules for matching packets.
 *		C.E.Hawkins	:	IFF_PROMISC/SIOCGHWADDR support
 *	Pauline Middelink	:	identd support
 *		Alan Cox	:	Fixed connect() taking signals I think.
 *		Alan Cox	:	SO_LINGER supported
 *		Alan Cox	:	Error reporting fixes
 *		Anonymous	:	inet_create tidied up (sk->reuse setting)
 *		Alan Cox	:	inet sockets don't set sk->type!
 *		Alan Cox	:	Split socket option code
 *		Alan Cox	:	Callbacks
 *		Alan Cox	:	Nagle flag for Charles & Johannes stuff
 *		Alex		:	Removed restriction on inet fioctl
 *		Alan Cox	:	Splitting INET from NET core
 *		Alan Cox	:	Fixed bogus SO_TYPE handling in getsockopt()
 *		Adam Caldwell	:	Missing return in SO_DONTROUTE/SO_DEBUG code
 *		Alan Cox	:	Split IP from generic code
 *		Alan Cox	:	New kfree_skbmem()
 *		Alan Cox	:	Make SO_DEBUG superuser only.
 *		Alan Cox	:	Allow anyone to clear SO_DEBUG
 *					(compatibility fix)
 *		Alan Cox	:	Added optimistic memory grabbing for AF_UNIX throughput.
 *		Alan Cox	:	Allocator for a socket is settable.
 *		Alan Cox	:	SO_ERROR includes soft errors.
 *		Alan Cox	:	Allow NULL arguments on some SO_ opts
 *		Alan Cox	: 	Generic socket allocation to make hooks
 *					easier (suggested by Craig Metz).
 *		Michael Pall	:	SO_ERROR returns positive errno again
 *              Steve Whitehouse:       Added default destructor to free
 *                                      protocol private data.
 *              Steve Whitehouse:       Added various other default routines
 *                                      common to several socket families.
 *              Chris Evans     :       Call suser() check last on F_SETOWN
 *		Jay Schulist	:	Added SO_ATTACH_FILTER and SO_DETACH_FILTER.
 *		Andi Kleen	:	Add sock_kmalloc()/sock_kfree_s()
 *		Andi Kleen	:	Fix write_space callback
 *		Chris Evans	:	Security fixes - signedness again
 *		Arnaldo C. Melo :       cleanups, use skb_queue_purge
 *
 * To Fix:
 *
 *
 *		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.
 */

92
#include <linux/capability.h>
L
Linus Torvalds 已提交
93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111
#include <linux/errno.h>
#include <linux/types.h>
#include <linux/socket.h>
#include <linux/in.h>
#include <linux/kernel.h>
#include <linux/module.h>
#include <linux/proc_fs.h>
#include <linux/seq_file.h>
#include <linux/sched.h>
#include <linux/timer.h>
#include <linux/string.h>
#include <linux/sockios.h>
#include <linux/net.h>
#include <linux/mm.h>
#include <linux/slab.h>
#include <linux/interrupt.h>
#include <linux/poll.h>
#include <linux/tcp.h>
#include <linux/init.h>
A
Al Viro 已提交
112
#include <linux/highmem.h>
L
Linus Torvalds 已提交
113 114 115 116 117 118 119

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

#include <linux/netdevice.h>
#include <net/protocol.h>
#include <linux/skbuff.h>
120
#include <net/net_namespace.h>
121
#include <net/request_sock.h>
L
Linus Torvalds 已提交
122
#include <net/sock.h>
123
#include <linux/net_tstamp.h>
L
Linus Torvalds 已提交
124 125
#include <net/xfrm.h>
#include <linux/ipsec.h>
126
#include <net/cls_cgroup.h>
L
Linus Torvalds 已提交
127 128 129 130 131 132 133

#include <linux/filter.h>

#ifdef CONFIG_INET
#include <net/tcp.h>
#endif

134 135 136 137
/*
 * Each address family might have different locking rules, so we have
 * one slock key per address family:
 */
I
Ingo Molnar 已提交
138 139 140 141 142 143 144 145
static struct lock_class_key af_family_keys[AF_MAX];
static struct lock_class_key af_family_slock_keys[AF_MAX];

/*
 * Make lock validator output more readable. (we pre-construct these
 * strings build-time, so that runtime initialization of socket
 * locks is fast):
 */
146
static const char *const af_family_key_strings[AF_MAX+1] = {
I
Ingo Molnar 已提交
147 148 149 150 151 152 153
  "sk_lock-AF_UNSPEC", "sk_lock-AF_UNIX"     , "sk_lock-AF_INET"     ,
  "sk_lock-AF_AX25"  , "sk_lock-AF_IPX"      , "sk_lock-AF_APPLETALK",
  "sk_lock-AF_NETROM", "sk_lock-AF_BRIDGE"   , "sk_lock-AF_ATMPVC"   ,
  "sk_lock-AF_X25"   , "sk_lock-AF_INET6"    , "sk_lock-AF_ROSE"     ,
  "sk_lock-AF_DECnet", "sk_lock-AF_NETBEUI"  , "sk_lock-AF_SECURITY" ,
  "sk_lock-AF_KEY"   , "sk_lock-AF_NETLINK"  , "sk_lock-AF_PACKET"   ,
  "sk_lock-AF_ASH"   , "sk_lock-AF_ECONET"   , "sk_lock-AF_ATMSVC"   ,
A
Andy Grover 已提交
154
  "sk_lock-AF_RDS"   , "sk_lock-AF_SNA"      , "sk_lock-AF_IRDA"     ,
I
Ingo Molnar 已提交
155
  "sk_lock-AF_PPPOX" , "sk_lock-AF_WANPIPE"  , "sk_lock-AF_LLC"      ,
156
  "sk_lock-27"       , "sk_lock-28"          , "sk_lock-AF_CAN"      ,
157
  "sk_lock-AF_TIPC"  , "sk_lock-AF_BLUETOOTH", "sk_lock-IUCV"        ,
R
Remi Denis-Courmont 已提交
158
  "sk_lock-AF_RXRPC" , "sk_lock-AF_ISDN"     , "sk_lock-AF_PHONET"   ,
159
  "sk_lock-AF_IEEE802154", "sk_lock-AF_CAIF" ,
R
Remi Denis-Courmont 已提交
160
  "sk_lock-AF_MAX"
I
Ingo Molnar 已提交
161
};
162
static const char *const af_family_slock_key_strings[AF_MAX+1] = {
I
Ingo Molnar 已提交
163 164 165 166 167 168 169
  "slock-AF_UNSPEC", "slock-AF_UNIX"     , "slock-AF_INET"     ,
  "slock-AF_AX25"  , "slock-AF_IPX"      , "slock-AF_APPLETALK",
  "slock-AF_NETROM", "slock-AF_BRIDGE"   , "slock-AF_ATMPVC"   ,
  "slock-AF_X25"   , "slock-AF_INET6"    , "slock-AF_ROSE"     ,
  "slock-AF_DECnet", "slock-AF_NETBEUI"  , "slock-AF_SECURITY" ,
  "slock-AF_KEY"   , "slock-AF_NETLINK"  , "slock-AF_PACKET"   ,
  "slock-AF_ASH"   , "slock-AF_ECONET"   , "slock-AF_ATMSVC"   ,
A
Andy Grover 已提交
170
  "slock-AF_RDS"   , "slock-AF_SNA"      , "slock-AF_IRDA"     ,
I
Ingo Molnar 已提交
171
  "slock-AF_PPPOX" , "slock-AF_WANPIPE"  , "slock-AF_LLC"      ,
172
  "slock-27"       , "slock-28"          , "slock-AF_CAN"      ,
173
  "slock-AF_TIPC"  , "slock-AF_BLUETOOTH", "slock-AF_IUCV"     ,
R
Remi Denis-Courmont 已提交
174
  "slock-AF_RXRPC" , "slock-AF_ISDN"     , "slock-AF_PHONET"   ,
175
  "slock-AF_IEEE802154", "slock-AF_CAIF" ,
R
Remi Denis-Courmont 已提交
176
  "slock-AF_MAX"
I
Ingo Molnar 已提交
177
};
178
static const char *const af_family_clock_key_strings[AF_MAX+1] = {
179 180 181 182 183 184 185
  "clock-AF_UNSPEC", "clock-AF_UNIX"     , "clock-AF_INET"     ,
  "clock-AF_AX25"  , "clock-AF_IPX"      , "clock-AF_APPLETALK",
  "clock-AF_NETROM", "clock-AF_BRIDGE"   , "clock-AF_ATMPVC"   ,
  "clock-AF_X25"   , "clock-AF_INET6"    , "clock-AF_ROSE"     ,
  "clock-AF_DECnet", "clock-AF_NETBEUI"  , "clock-AF_SECURITY" ,
  "clock-AF_KEY"   , "clock-AF_NETLINK"  , "clock-AF_PACKET"   ,
  "clock-AF_ASH"   , "clock-AF_ECONET"   , "clock-AF_ATMSVC"   ,
A
Andy Grover 已提交
186
  "clock-AF_RDS"   , "clock-AF_SNA"      , "clock-AF_IRDA"     ,
187
  "clock-AF_PPPOX" , "clock-AF_WANPIPE"  , "clock-AF_LLC"      ,
188
  "clock-27"       , "clock-28"          , "clock-AF_CAN"      ,
189
  "clock-AF_TIPC"  , "clock-AF_BLUETOOTH", "clock-AF_IUCV"     ,
R
Remi Denis-Courmont 已提交
190
  "clock-AF_RXRPC" , "clock-AF_ISDN"     , "clock-AF_PHONET"   ,
191
  "clock-AF_IEEE802154", "clock-AF_CAIF" ,
R
Remi Denis-Courmont 已提交
192
  "clock-AF_MAX"
193
};
194 195 196 197 198 199 200

/*
 * sk_callback_lock locking rules are per-address-family,
 * so split the lock classes by using a per-AF key:
 */
static struct lock_class_key af_callback_keys[AF_MAX];

L
Linus Torvalds 已提交
201 202 203 204 205 206 207 208 209 210 211
/* Take into consideration the size of the struct sk_buff overhead in the
 * determination of these values, since that is non-constant across
 * platforms.  This makes socket queueing behavior and performance
 * not depend upon such differences.
 */
#define _SK_MEM_PACKETS		256
#define _SK_MEM_OVERHEAD	(sizeof(struct sk_buff) + 256)
#define SK_WMEM_MAX		(_SK_MEM_OVERHEAD * _SK_MEM_PACKETS)
#define SK_RMEM_MAX		(_SK_MEM_OVERHEAD * _SK_MEM_PACKETS)

/* Run time adjustable parameters. */
212 213 214 215
__u32 sysctl_wmem_max __read_mostly = SK_WMEM_MAX;
__u32 sysctl_rmem_max __read_mostly = SK_RMEM_MAX;
__u32 sysctl_wmem_default __read_mostly = SK_WMEM_MAX;
__u32 sysctl_rmem_default __read_mostly = SK_RMEM_MAX;
L
Linus Torvalds 已提交
216 217

/* Maximal space eaten by iovec or ancilliary data plus some space */
218
int sysctl_optmem_max __read_mostly = sizeof(unsigned long)*(2*UIO_MAXIOV+512);
E
Eric Dumazet 已提交
219
EXPORT_SYMBOL(sysctl_optmem_max);
L
Linus Torvalds 已提交
220

221 222 223 224 225
#if defined(CONFIG_CGROUPS) && !defined(CONFIG_NET_CLS_CGROUP)
int net_cls_subsys_id = -1;
EXPORT_SYMBOL_GPL(net_cls_subsys_id);
#endif

L
Linus Torvalds 已提交
226 227 228 229 230 231 232 233
static int sock_set_timeout(long *timeo_p, char __user *optval, int optlen)
{
	struct timeval tv;

	if (optlen < sizeof(tv))
		return -EINVAL;
	if (copy_from_user(&tv, optval, sizeof(tv)))
		return -EFAULT;
234 235
	if (tv.tv_usec < 0 || tv.tv_usec >= USEC_PER_SEC)
		return -EDOM;
L
Linus Torvalds 已提交
236

237
	if (tv.tv_sec < 0) {
238 239
		static int warned __read_mostly;

240
		*timeo_p = 0;
241
		if (warned < 10 && net_ratelimit()) {
242 243 244
			warned++;
			printk(KERN_INFO "sock_set_timeout: `%s' (pid %d) "
			       "tries to set negative timeout\n",
245
				current->comm, task_pid_nr(current));
246
		}
247 248
		return 0;
	}
L
Linus Torvalds 已提交
249 250 251 252 253 254 255 256 257 258 259 260
	*timeo_p = MAX_SCHEDULE_TIMEOUT;
	if (tv.tv_sec == 0 && tv.tv_usec == 0)
		return 0;
	if (tv.tv_sec < (MAX_SCHEDULE_TIMEOUT/HZ - 1))
		*timeo_p = tv.tv_sec*HZ + (tv.tv_usec+(1000000/HZ-1))/(1000000/HZ);
	return 0;
}

static void sock_warn_obsolete_bsdism(const char *name)
{
	static int warned;
	static char warncomm[TASK_COMM_LEN];
261 262
	if (strcmp(warncomm, current->comm) && warned < 5) {
		strcpy(warncomm,  current->comm);
L
Linus Torvalds 已提交
263 264 265 266 267 268
		printk(KERN_WARNING "process `%s' is using obsolete "
		       "%s SO_BSDCOMPAT\n", warncomm, name);
		warned++;
	}
}

269
static void sock_disable_timestamp(struct sock *sk, int flag)
270
{
271 272 273 274 275 276
	if (sock_flag(sk, flag)) {
		sock_reset_flag(sk, flag);
		if (!sock_flag(sk, SOCK_TIMESTAMP) &&
		    !sock_flag(sk, SOCK_TIMESTAMPING_RX_SOFTWARE)) {
			net_disable_timestamp();
		}
L
Linus Torvalds 已提交
277 278 279 280
	}
}


281 282
int sock_queue_rcv_skb(struct sock *sk, struct sk_buff *skb)
{
E
Eric Dumazet 已提交
283
	int err;
284
	int skb_len;
285 286
	unsigned long flags;
	struct sk_buff_head *list = &sk->sk_receive_queue;
287

R
Rami Rosen 已提交
288
	/* Cast sk->rcvbuf to unsigned... It's pointless, but reduces
289 290 291 292
	   number of warnings when compiling with -W --ANK
	 */
	if (atomic_read(&sk->sk_rmem_alloc) + skb->truesize >=
	    (unsigned)sk->sk_rcvbuf) {
E
Eric Dumazet 已提交
293 294
		atomic_inc(&sk->sk_drops);
		return -ENOMEM;
295 296
	}

297
	err = sk_filter(sk, skb);
298
	if (err)
E
Eric Dumazet 已提交
299
		return err;
300

301
	if (!sk_rmem_schedule(sk, skb->truesize)) {
E
Eric Dumazet 已提交
302 303
		atomic_inc(&sk->sk_drops);
		return -ENOBUFS;
304 305
	}

306 307
	skb->dev = NULL;
	skb_set_owner_r(skb, sk);
308

309 310 311 312 313 314 315
	/* Cache the SKB length before we tack it onto the receive
	 * queue.  Once it is added it no longer belongs to us and
	 * may be freed by other threads of control pulling packets
	 * from the queue.
	 */
	skb_len = skb->len;

E
Eric Dumazet 已提交
316 317 318 319 320
	/* we escape from rcu protected region, make sure we dont leak
	 * a norefcounted dst
	 */
	skb_dst_force(skb);

321 322 323 324
	spin_lock_irqsave(&list->lock, flags);
	skb->dropcount = atomic_read(&sk->sk_drops);
	__skb_queue_tail(list, skb);
	spin_unlock_irqrestore(&list->lock, flags);
325 326 327

	if (!sock_flag(sk, SOCK_DEAD))
		sk->sk_data_ready(sk, skb_len);
E
Eric Dumazet 已提交
328
	return 0;
329 330 331
}
EXPORT_SYMBOL(sock_queue_rcv_skb);

332
int sk_receive_skb(struct sock *sk, struct sk_buff *skb, const int nested)
333 334 335
{
	int rc = NET_RX_SUCCESS;

336
	if (sk_filter(sk, skb))
337 338 339 340
		goto discard_and_relse;

	skb->dev = NULL;

341 342 343 344
	if (sk_rcvqueues_full(sk, skb)) {
		atomic_inc(&sk->sk_drops);
		goto discard_and_relse;
	}
345 346 347 348
	if (nested)
		bh_lock_sock_nested(sk);
	else
		bh_lock_sock(sk);
I
Ingo Molnar 已提交
349 350 351 352 353 354
	if (!sock_owned_by_user(sk)) {
		/*
		 * trylock + unlock semantics:
		 */
		mutex_acquire(&sk->sk_lock.dep_map, 0, 1, _RET_IP_);

P
Peter Zijlstra 已提交
355
		rc = sk_backlog_rcv(sk, skb);
I
Ingo Molnar 已提交
356 357

		mutex_release(&sk->sk_lock.dep_map, 1, _RET_IP_);
Z
Zhu Yi 已提交
358
	} else if (sk_add_backlog(sk, skb)) {
Z
Zhu Yi 已提交
359 360 361 362 363
		bh_unlock_sock(sk);
		atomic_inc(&sk->sk_drops);
		goto discard_and_relse;
	}

364 365 366 367 368 369 370 371 372 373
	bh_unlock_sock(sk);
out:
	sock_put(sk);
	return rc;
discard_and_relse:
	kfree_skb(skb);
	goto out;
}
EXPORT_SYMBOL(sk_receive_skb);

K
Krishna Kumar 已提交
374 375 376 377 378 379
void sk_reset_txq(struct sock *sk)
{
	sk_tx_queue_clear(sk);
}
EXPORT_SYMBOL(sk_reset_txq);

380 381
struct dst_entry *__sk_dst_check(struct sock *sk, u32 cookie)
{
E
Eric Dumazet 已提交
382
	struct dst_entry *dst = __sk_dst_get(sk);
383 384

	if (dst && dst->obsolete && dst->ops->check(dst, cookie) == NULL) {
K
Krishna Kumar 已提交
385
		sk_tx_queue_clear(sk);
E
Eric Dumazet 已提交
386
		rcu_assign_pointer(sk->sk_dst_cache, NULL);
387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408
		dst_release(dst);
		return NULL;
	}

	return dst;
}
EXPORT_SYMBOL(__sk_dst_check);

struct dst_entry *sk_dst_check(struct sock *sk, u32 cookie)
{
	struct dst_entry *dst = sk_dst_get(sk);

	if (dst && dst->obsolete && dst->ops->check(dst, cookie) == NULL) {
		sk_dst_reset(sk);
		dst_release(dst);
		return NULL;
	}

	return dst;
}
EXPORT_SYMBOL(sk_dst_check);

409 410 411 412
static int sock_bindtodevice(struct sock *sk, char __user *optval, int optlen)
{
	int ret = -ENOPROTOOPT;
#ifdef CONFIG_NETDEVICES
413
	struct net *net = sock_net(sk);
414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438
	char devname[IFNAMSIZ];
	int index;

	/* Sorry... */
	ret = -EPERM;
	if (!capable(CAP_NET_RAW))
		goto out;

	ret = -EINVAL;
	if (optlen < 0)
		goto out;

	/* Bind this socket to a particular device like "eth0",
	 * as specified in the passed interface name. If the
	 * name is "" or the option length is zero the socket
	 * is not bound.
	 */
	if (optlen > IFNAMSIZ - 1)
		optlen = IFNAMSIZ - 1;
	memset(devname, 0, sizeof(devname));

	ret = -EFAULT;
	if (copy_from_user(devname, optval, optlen))
		goto out;

439 440
	index = 0;
	if (devname[0] != '\0') {
441
		struct net_device *dev;
442

443 444 445 446 447
		rcu_read_lock();
		dev = dev_get_by_name_rcu(net, devname);
		if (dev)
			index = dev->ifindex;
		rcu_read_unlock();
448 449 450 451 452 453 454 455 456 457 458 459 460 461 462 463 464 465
		ret = -ENODEV;
		if (!dev)
			goto out;
	}

	lock_sock(sk);
	sk->sk_bound_dev_if = index;
	sk_dst_reset(sk);
	release_sock(sk);

	ret = 0;

out:
#endif

	return ret;
}

466 467 468 469 470 471 472 473
static inline void sock_valbool_flag(struct sock *sk, int bit, int valbool)
{
	if (valbool)
		sock_set_flag(sk, bit);
	else
		sock_reset_flag(sk, bit);
}

L
Linus Torvalds 已提交
474 475 476 477 478 479
/*
 *	This is meant for all protocols to use and covers goings on
 *	at the socket level. Everything here is generic.
 */

int sock_setsockopt(struct socket *sock, int level, int optname,
480
		    char __user *optval, unsigned int optlen)
L
Linus Torvalds 已提交
481
{
E
Eric Dumazet 已提交
482
	struct sock *sk = sock->sk;
L
Linus Torvalds 已提交
483 484 485 486
	int val;
	int valbool;
	struct linger ling;
	int ret = 0;
487

L
Linus Torvalds 已提交
488 489 490 491
	/*
	 *	Options without arguments
	 */

492 493 494
	if (optname == SO_BINDTODEVICE)
		return sock_bindtodevice(sk, optval, optlen);

495 496
	if (optlen < sizeof(int))
		return -EINVAL;
497

L
Linus Torvalds 已提交
498 499
	if (get_user(val, (int __user *)optval))
		return -EFAULT;
500

E
Eric Dumazet 已提交
501
	valbool = val ? 1 : 0;
L
Linus Torvalds 已提交
502 503 504

	lock_sock(sk);

E
Eric Dumazet 已提交
505
	switch (optname) {
506
	case SO_DEBUG:
E
Eric Dumazet 已提交
507
		if (val && !capable(CAP_NET_ADMIN))
508
			ret = -EACCES;
E
Eric Dumazet 已提交
509
		else
510
			sock_valbool_flag(sk, SOCK_DBG, valbool);
511 512 513 514 515
		break;
	case SO_REUSEADDR:
		sk->sk_reuse = valbool;
		break;
	case SO_TYPE:
516
	case SO_PROTOCOL:
517
	case SO_DOMAIN:
518 519 520 521
	case SO_ERROR:
		ret = -ENOPROTOOPT;
		break;
	case SO_DONTROUTE:
522
		sock_valbool_flag(sk, SOCK_LOCALROUTE, valbool);
523 524 525 526 527 528 529 530 531 532 533 534
		break;
	case SO_BROADCAST:
		sock_valbool_flag(sk, SOCK_BROADCAST, valbool);
		break;
	case SO_SNDBUF:
		/* Don't error on this BSD doesn't and if you think
		   about it this is right. Otherwise apps have to
		   play 'guess the biggest size' games. RCVBUF/SNDBUF
		   are treated in BSD as hints */

		if (val > sysctl_wmem_max)
			val = sysctl_wmem_max;
535
set_sndbuf:
536 537 538 539 540
		sk->sk_userlocks |= SOCK_SNDBUF_LOCK;
		if ((val * 2) < SOCK_MIN_SNDBUF)
			sk->sk_sndbuf = SOCK_MIN_SNDBUF;
		else
			sk->sk_sndbuf = val * 2;
L
Linus Torvalds 已提交
541

542 543 544 545 546 547
		/*
		 *	Wake up sending tasks if we
		 *	upped the value.
		 */
		sk->sk_write_space(sk);
		break;
L
Linus Torvalds 已提交
548

549 550 551 552 553 554
	case SO_SNDBUFFORCE:
		if (!capable(CAP_NET_ADMIN)) {
			ret = -EPERM;
			break;
		}
		goto set_sndbuf;
555

556 557 558 559 560
	case SO_RCVBUF:
		/* Don't error on this BSD doesn't and if you think
		   about it this is right. Otherwise apps have to
		   play 'guess the biggest size' games. RCVBUF/SNDBUF
		   are treated in BSD as hints */
561

562 563
		if (val > sysctl_rmem_max)
			val = sysctl_rmem_max;
564
set_rcvbuf:
565 566 567 568 569 570 571 572 573 574 575 576 577 578 579 580 581 582 583 584 585 586 587 588 589
		sk->sk_userlocks |= SOCK_RCVBUF_LOCK;
		/*
		 * We double it on the way in to account for
		 * "struct sk_buff" etc. overhead.   Applications
		 * assume that the SO_RCVBUF setting they make will
		 * allow that much actual data to be received on that
		 * socket.
		 *
		 * Applications are unaware that "struct sk_buff" and
		 * other overheads allocate from the receive buffer
		 * during socket buffer allocation.
		 *
		 * And after considering the possible alternatives,
		 * returning the value we actually used in getsockopt
		 * is the most desirable behavior.
		 */
		if ((val * 2) < SOCK_MIN_RCVBUF)
			sk->sk_rcvbuf = SOCK_MIN_RCVBUF;
		else
			sk->sk_rcvbuf = val * 2;
		break;

	case SO_RCVBUFFORCE:
		if (!capable(CAP_NET_ADMIN)) {
			ret = -EPERM;
L
Linus Torvalds 已提交
590
			break;
591 592
		}
		goto set_rcvbuf;
L
Linus Torvalds 已提交
593

594
	case SO_KEEPALIVE:
L
Linus Torvalds 已提交
595
#ifdef CONFIG_INET
596 597
		if (sk->sk_protocol == IPPROTO_TCP)
			tcp_set_keepalive(sk, valbool);
L
Linus Torvalds 已提交
598
#endif
599 600 601 602 603 604 605 606 607 608 609 610 611 612 613 614 615 616 617 618 619
		sock_valbool_flag(sk, SOCK_KEEPOPEN, valbool);
		break;

	case SO_OOBINLINE:
		sock_valbool_flag(sk, SOCK_URGINLINE, valbool);
		break;

	case SO_NO_CHECK:
		sk->sk_no_check = valbool;
		break;

	case SO_PRIORITY:
		if ((val >= 0 && val <= 6) || capable(CAP_NET_ADMIN))
			sk->sk_priority = val;
		else
			ret = -EPERM;
		break;

	case SO_LINGER:
		if (optlen < sizeof(ling)) {
			ret = -EINVAL;	/* 1003.1g */
L
Linus Torvalds 已提交
620
			break;
621
		}
E
Eric Dumazet 已提交
622
		if (copy_from_user(&ling, optval, sizeof(ling))) {
623
			ret = -EFAULT;
L
Linus Torvalds 已提交
624
			break;
625 626 627 628
		}
		if (!ling.l_onoff)
			sock_reset_flag(sk, SOCK_LINGER);
		else {
L
Linus Torvalds 已提交
629
#if (BITS_PER_LONG == 32)
630 631
			if ((unsigned int)ling.l_linger >= MAX_SCHEDULE_TIMEOUT/HZ)
				sk->sk_lingertime = MAX_SCHEDULE_TIMEOUT;
L
Linus Torvalds 已提交
632
			else
633 634 635 636 637 638 639 640 641 642 643 644 645 646 647 648 649 650
#endif
				sk->sk_lingertime = (unsigned int)ling.l_linger * HZ;
			sock_set_flag(sk, SOCK_LINGER);
		}
		break;

	case SO_BSDCOMPAT:
		sock_warn_obsolete_bsdism("setsockopt");
		break;

	case SO_PASSCRED:
		if (valbool)
			set_bit(SOCK_PASSCRED, &sock->flags);
		else
			clear_bit(SOCK_PASSCRED, &sock->flags);
		break;

	case SO_TIMESTAMP:
651
	case SO_TIMESTAMPNS:
652
		if (valbool)  {
653 654 655 656
			if (optname == SO_TIMESTAMP)
				sock_reset_flag(sk, SOCK_RCVTSTAMPNS);
			else
				sock_set_flag(sk, SOCK_RCVTSTAMPNS);
657
			sock_set_flag(sk, SOCK_RCVTSTAMP);
658
			sock_enable_timestamp(sk, SOCK_TIMESTAMP);
659
		} else {
660
			sock_reset_flag(sk, SOCK_RCVTSTAMP);
661 662
			sock_reset_flag(sk, SOCK_RCVTSTAMPNS);
		}
663 664
		break;

665 666
	case SO_TIMESTAMPING:
		if (val & ~SOF_TIMESTAMPING_MASK) {
667
			ret = -EINVAL;
668 669 670 671 672 673 674 675 676 677 678 679 680 681 682 683 684 685 686 687 688 689
			break;
		}
		sock_valbool_flag(sk, SOCK_TIMESTAMPING_TX_HARDWARE,
				  val & SOF_TIMESTAMPING_TX_HARDWARE);
		sock_valbool_flag(sk, SOCK_TIMESTAMPING_TX_SOFTWARE,
				  val & SOF_TIMESTAMPING_TX_SOFTWARE);
		sock_valbool_flag(sk, SOCK_TIMESTAMPING_RX_HARDWARE,
				  val & SOF_TIMESTAMPING_RX_HARDWARE);
		if (val & SOF_TIMESTAMPING_RX_SOFTWARE)
			sock_enable_timestamp(sk,
					      SOCK_TIMESTAMPING_RX_SOFTWARE);
		else
			sock_disable_timestamp(sk,
					       SOCK_TIMESTAMPING_RX_SOFTWARE);
		sock_valbool_flag(sk, SOCK_TIMESTAMPING_SOFTWARE,
				  val & SOF_TIMESTAMPING_SOFTWARE);
		sock_valbool_flag(sk, SOCK_TIMESTAMPING_SYS_HARDWARE,
				  val & SOF_TIMESTAMPING_SYS_HARDWARE);
		sock_valbool_flag(sk, SOCK_TIMESTAMPING_RAW_HARDWARE,
				  val & SOF_TIMESTAMPING_RAW_HARDWARE);
		break;

690 691 692 693 694 695 696 697 698 699 700 701 702
	case SO_RCVLOWAT:
		if (val < 0)
			val = INT_MAX;
		sk->sk_rcvlowat = val ? : 1;
		break;

	case SO_RCVTIMEO:
		ret = sock_set_timeout(&sk->sk_rcvtimeo, optval, optlen);
		break;

	case SO_SNDTIMEO:
		ret = sock_set_timeout(&sk->sk_sndtimeo, optval, optlen);
		break;
L
Linus Torvalds 已提交
703

704 705 706 707
	case SO_ATTACH_FILTER:
		ret = -EINVAL;
		if (optlen == sizeof(struct sock_fprog)) {
			struct sock_fprog fprog;
L
Linus Torvalds 已提交
708

709 710
			ret = -EFAULT;
			if (copy_from_user(&fprog, optval, sizeof(fprog)))
L
Linus Torvalds 已提交
711
				break;
712 713 714 715 716 717

			ret = sk_attach_filter(&fprog, sk);
		}
		break;

	case SO_DETACH_FILTER:
718
		ret = sk_detach_filter(sk);
719
		break;
L
Linus Torvalds 已提交
720

721 722 723 724 725 726
	case SO_PASSSEC:
		if (valbool)
			set_bit(SOCK_PASSSEC, &sock->flags);
		else
			clear_bit(SOCK_PASSSEC, &sock->flags);
		break;
727 728 729
	case SO_MARK:
		if (!capable(CAP_NET_ADMIN))
			ret = -EPERM;
E
Eric Dumazet 已提交
730
		else
731 732
			sk->sk_mark = val;
		break;
C
Catherine Zhang 已提交
733

L
Linus Torvalds 已提交
734 735
		/* We implement the SO_SNDLOWAT etc to
		   not be settable (1003.1g 5.3) */
736 737 738 739 740 741
	case SO_RXQ_OVFL:
		if (valbool)
			sock_set_flag(sk, SOCK_RXQ_OVFL);
		else
			sock_reset_flag(sk, SOCK_RXQ_OVFL);
		break;
742 743 744
	default:
		ret = -ENOPROTOOPT;
		break;
745
	}
L
Linus Torvalds 已提交
746 747 748
	release_sock(sk);
	return ret;
}
E
Eric Dumazet 已提交
749
EXPORT_SYMBOL(sock_setsockopt);
L
Linus Torvalds 已提交
750 751 752 753 754 755


int sock_getsockopt(struct socket *sock, int level, int optname,
		    char __user *optval, int __user *optlen)
{
	struct sock *sk = sock->sk;
756

757
	union {
758 759
		int val;
		struct linger ling;
L
Linus Torvalds 已提交
760 761
		struct timeval tm;
	} v;
762

763
	int lv = sizeof(int);
L
Linus Torvalds 已提交
764
	int len;
765

766
	if (get_user(len, optlen))
767
		return -EFAULT;
768
	if (len < 0)
L
Linus Torvalds 已提交
769
		return -EINVAL;
770

771
	memset(&v, 0, sizeof(v));
772

E
Eric Dumazet 已提交
773
	switch (optname) {
774 775 776 777 778 779 780 781 782 783 784 785 786 787 788 789 790 791 792 793 794 795 796 797 798 799 800 801 802 803 804 805
	case SO_DEBUG:
		v.val = sock_flag(sk, SOCK_DBG);
		break;

	case SO_DONTROUTE:
		v.val = sock_flag(sk, SOCK_LOCALROUTE);
		break;

	case SO_BROADCAST:
		v.val = !!sock_flag(sk, SOCK_BROADCAST);
		break;

	case SO_SNDBUF:
		v.val = sk->sk_sndbuf;
		break;

	case SO_RCVBUF:
		v.val = sk->sk_rcvbuf;
		break;

	case SO_REUSEADDR:
		v.val = sk->sk_reuse;
		break;

	case SO_KEEPALIVE:
		v.val = !!sock_flag(sk, SOCK_KEEPOPEN);
		break;

	case SO_TYPE:
		v.val = sk->sk_type;
		break;

806 807 808 809
	case SO_PROTOCOL:
		v.val = sk->sk_protocol;
		break;

810 811 812 813
	case SO_DOMAIN:
		v.val = sk->sk_family;
		break;

814 815
	case SO_ERROR:
		v.val = -sock_error(sk);
E
Eric Dumazet 已提交
816
		if (v.val == 0)
817 818 819 820 821 822 823 824 825 826 827 828 829 830 831 832 833 834 835 836 837 838 839 840 841 842
			v.val = xchg(&sk->sk_err_soft, 0);
		break;

	case SO_OOBINLINE:
		v.val = !!sock_flag(sk, SOCK_URGINLINE);
		break;

	case SO_NO_CHECK:
		v.val = sk->sk_no_check;
		break;

	case SO_PRIORITY:
		v.val = sk->sk_priority;
		break;

	case SO_LINGER:
		lv		= sizeof(v.ling);
		v.ling.l_onoff	= !!sock_flag(sk, SOCK_LINGER);
		v.ling.l_linger	= sk->sk_lingertime / HZ;
		break;

	case SO_BSDCOMPAT:
		sock_warn_obsolete_bsdism("getsockopt");
		break;

	case SO_TIMESTAMP:
843 844 845 846 847 848
		v.val = sock_flag(sk, SOCK_RCVTSTAMP) &&
				!sock_flag(sk, SOCK_RCVTSTAMPNS);
		break;

	case SO_TIMESTAMPNS:
		v.val = sock_flag(sk, SOCK_RCVTSTAMPNS);
849 850
		break;

851 852 853 854 855 856 857 858 859 860 861 862 863 864 865 866 867 868
	case SO_TIMESTAMPING:
		v.val = 0;
		if (sock_flag(sk, SOCK_TIMESTAMPING_TX_HARDWARE))
			v.val |= SOF_TIMESTAMPING_TX_HARDWARE;
		if (sock_flag(sk, SOCK_TIMESTAMPING_TX_SOFTWARE))
			v.val |= SOF_TIMESTAMPING_TX_SOFTWARE;
		if (sock_flag(sk, SOCK_TIMESTAMPING_RX_HARDWARE))
			v.val |= SOF_TIMESTAMPING_RX_HARDWARE;
		if (sock_flag(sk, SOCK_TIMESTAMPING_RX_SOFTWARE))
			v.val |= SOF_TIMESTAMPING_RX_SOFTWARE;
		if (sock_flag(sk, SOCK_TIMESTAMPING_SOFTWARE))
			v.val |= SOF_TIMESTAMPING_SOFTWARE;
		if (sock_flag(sk, SOCK_TIMESTAMPING_SYS_HARDWARE))
			v.val |= SOF_TIMESTAMPING_SYS_HARDWARE;
		if (sock_flag(sk, SOCK_TIMESTAMPING_RAW_HARDWARE))
			v.val |= SOF_TIMESTAMPING_RAW_HARDWARE;
		break;

869
	case SO_RCVTIMEO:
E
Eric Dumazet 已提交
870
		lv = sizeof(struct timeval);
871 872 873 874 875 876 877 878 879 880
		if (sk->sk_rcvtimeo == MAX_SCHEDULE_TIMEOUT) {
			v.tm.tv_sec = 0;
			v.tm.tv_usec = 0;
		} else {
			v.tm.tv_sec = sk->sk_rcvtimeo / HZ;
			v.tm.tv_usec = ((sk->sk_rcvtimeo % HZ) * 1000000) / HZ;
		}
		break;

	case SO_SNDTIMEO:
E
Eric Dumazet 已提交
881
		lv = sizeof(struct timeval);
882 883 884 885 886 887 888 889
		if (sk->sk_sndtimeo == MAX_SCHEDULE_TIMEOUT) {
			v.tm.tv_sec = 0;
			v.tm.tv_usec = 0;
		} else {
			v.tm.tv_sec = sk->sk_sndtimeo / HZ;
			v.tm.tv_usec = ((sk->sk_sndtimeo % HZ) * 1000000) / HZ;
		}
		break;
L
Linus Torvalds 已提交
890

891 892 893
	case SO_RCVLOWAT:
		v.val = sk->sk_rcvlowat;
		break;
L
Linus Torvalds 已提交
894

895
	case SO_SNDLOWAT:
E
Eric Dumazet 已提交
896
		v.val = 1;
897
		break;
L
Linus Torvalds 已提交
898

899 900 901
	case SO_PASSCRED:
		v.val = test_bit(SOCK_PASSCRED, &sock->flags) ? 1 : 0;
		break;
L
Linus Torvalds 已提交
902

903 904 905 906 907 908
	case SO_PEERCRED:
		if (len > sizeof(sk->sk_peercred))
			len = sizeof(sk->sk_peercred);
		if (copy_to_user(optval, &sk->sk_peercred, len))
			return -EFAULT;
		goto lenout;
L
Linus Torvalds 已提交
909

910 911 912 913 914 915 916 917 918 919 920 921
	case SO_PEERNAME:
	{
		char address[128];

		if (sock->ops->getname(sock, (struct sockaddr *)address, &lv, 2))
			return -ENOTCONN;
		if (lv < len)
			return -EINVAL;
		if (copy_to_user(optval, address, len))
			return -EFAULT;
		goto lenout;
	}
L
Linus Torvalds 已提交
922

923 924 925 926 927 928
	/* Dubious BSD thing... Probably nobody even uses it, but
	 * the UNIX standard wants it for whatever reason... -DaveM
	 */
	case SO_ACCEPTCONN:
		v.val = sk->sk_state == TCP_LISTEN;
		break;
L
Linus Torvalds 已提交
929

930 931 932
	case SO_PASSSEC:
		v.val = test_bit(SOCK_PASSSEC, &sock->flags) ? 1 : 0;
		break;
C
Catherine Zhang 已提交
933

934 935
	case SO_PEERSEC:
		return security_socket_getpeersec_stream(sock, optval, optlen, len);
L
Linus Torvalds 已提交
936

937 938 939 940
	case SO_MARK:
		v.val = sk->sk_mark;
		break;

941 942 943 944
	case SO_RXQ_OVFL:
		v.val = !!sock_flag(sk, SOCK_RXQ_OVFL);
		break;

945 946
	default:
		return -ENOPROTOOPT;
L
Linus Torvalds 已提交
947
	}
948

L
Linus Torvalds 已提交
949 950 951 952 953
	if (len > lv)
		len = lv;
	if (copy_to_user(optval, &v, len))
		return -EFAULT;
lenout:
954 955 956
	if (put_user(len, optlen))
		return -EFAULT;
	return 0;
L
Linus Torvalds 已提交
957 958
}

I
Ingo Molnar 已提交
959 960 961 962 963
/*
 * Initialize an sk_lock.
 *
 * (We also register the sk_lock with the lock validator.)
 */
D
Dave Jones 已提交
964
static inline void sock_lock_init(struct sock *sk)
I
Ingo Molnar 已提交
965
{
966 967 968 969 970
	sock_lock_init_class_and_name(sk,
			af_family_slock_key_strings[sk->sk_family],
			af_family_slock_keys + sk->sk_family,
			af_family_key_strings[sk->sk_family],
			af_family_keys + sk->sk_family);
I
Ingo Molnar 已提交
971 972
}

E
Eric Dumazet 已提交
973 974 975 976
/*
 * Copy all fields from osk to nsk but nsk->sk_refcnt must not change yet,
 * even temporarly, because of RCU lookups. sk_node should also be left as is.
 */
977 978 979 980 981
static void sock_copy(struct sock *nsk, const struct sock *osk)
{
#ifdef CONFIG_SECURITY_NETWORK
	void *sptr = nsk->sk_security;
#endif
E
Eric Dumazet 已提交
982
	BUILD_BUG_ON(offsetof(struct sock, sk_copy_start) !=
K
Krishna Kumar 已提交
983 984
		     sizeof(osk->sk_node) + sizeof(osk->sk_refcnt) +
		     sizeof(osk->sk_tx_queue_mapping));
E
Eric Dumazet 已提交
985 986
	memcpy(&nsk->sk_copy_start, &osk->sk_copy_start,
	       osk->sk_prot->obj_size - offsetof(struct sock, sk_copy_start));
987 988 989 990 991 992
#ifdef CONFIG_SECURITY_NETWORK
	nsk->sk_security = sptr;
	security_sk_clone(osk, nsk);
#endif
}

993 994
static struct sock *sk_prot_alloc(struct proto *prot, gfp_t priority,
		int family)
995 996 997 998 999
{
	struct sock *sk;
	struct kmem_cache *slab;

	slab = prot->slab;
1000 1001 1002 1003 1004 1005 1006 1007 1008 1009 1010 1011 1012 1013 1014 1015 1016
	if (slab != NULL) {
		sk = kmem_cache_alloc(slab, priority & ~__GFP_ZERO);
		if (!sk)
			return sk;
		if (priority & __GFP_ZERO) {
			/*
			 * caches using SLAB_DESTROY_BY_RCU should let
			 * sk_node.next un-modified. Special care is taken
			 * when initializing object to zero.
			 */
			if (offsetof(struct sock, sk_node.next) != 0)
				memset(sk, 0, offsetof(struct sock, sk_node.next));
			memset(&sk->sk_node.pprev, 0,
			       prot->obj_size - offsetof(struct sock,
							 sk_node.pprev));
		}
	}
1017 1018 1019
	else
		sk = kmalloc(prot->obj_size, priority);

1020
	if (sk != NULL) {
1021 1022
		kmemcheck_annotate_bitfield(sk, flags);

1023 1024 1025 1026 1027
		if (security_sk_alloc(sk, family, priority))
			goto out_free;

		if (!try_module_get(prot->owner))
			goto out_free_sec;
K
Krishna Kumar 已提交
1028
		sk_tx_queue_clear(sk);
1029 1030
	}

1031
	return sk;
1032 1033 1034 1035 1036 1037 1038 1039 1040

out_free_sec:
	security_sk_free(sk);
out_free:
	if (slab != NULL)
		kmem_cache_free(slab, sk);
	else
		kfree(sk);
	return NULL;
1041 1042 1043 1044 1045
}

static void sk_prot_free(struct proto *prot, struct sock *sk)
{
	struct kmem_cache *slab;
1046
	struct module *owner;
1047

1048
	owner = prot->owner;
1049
	slab = prot->slab;
1050 1051

	security_sk_free(sk);
1052 1053 1054 1055
	if (slab != NULL)
		kmem_cache_free(slab, sk);
	else
		kfree(sk);
1056
	module_put(owner);
1057 1058
}

1059 1060 1061 1062 1063 1064 1065 1066
#ifdef CONFIG_CGROUPS
void sock_update_classid(struct sock *sk)
{
	u32 classid = task_cls_classid(current);

	if (classid && classid != sk->sk_classid)
		sk->sk_classid = classid;
}
1067
EXPORT_SYMBOL(sock_update_classid);
1068 1069
#endif

L
Linus Torvalds 已提交
1070 1071
/**
 *	sk_alloc - All socket objects are allocated here
1072
 *	@net: the applicable net namespace
1073 1074 1075
 *	@family: protocol family
 *	@priority: for allocation (%GFP_KERNEL, %GFP_ATOMIC, etc)
 *	@prot: struct proto associated with this new sock instance
L
Linus Torvalds 已提交
1076
 */
1077
struct sock *sk_alloc(struct net *net, int family, gfp_t priority,
1078
		      struct proto *prot)
L
Linus Torvalds 已提交
1079
{
1080
	struct sock *sk;
L
Linus Torvalds 已提交
1081

1082
	sk = sk_prot_alloc(prot, priority | __GFP_ZERO, family);
L
Linus Torvalds 已提交
1083
	if (sk) {
1084 1085 1086 1087 1088 1089 1090
		sk->sk_family = family;
		/*
		 * See comment in struct sock definition to understand
		 * why we need sk_prot_creator -acme
		 */
		sk->sk_prot = sk->sk_prot_creator = prot;
		sock_lock_init(sk);
1091
		sock_net_set(sk, get_net(net));
1092
		atomic_set(&sk->sk_wmem_alloc, 1);
1093 1094

		sock_update_classid(sk);
L
Linus Torvalds 已提交
1095
	}
1096

1097
	return sk;
L
Linus Torvalds 已提交
1098
}
E
Eric Dumazet 已提交
1099
EXPORT_SYMBOL(sk_alloc);
L
Linus Torvalds 已提交
1100

1101
static void __sk_free(struct sock *sk)
L
Linus Torvalds 已提交
1102 1103 1104 1105 1106 1107
{
	struct sk_filter *filter;

	if (sk->sk_destruct)
		sk->sk_destruct(sk);

1108 1109
	filter = rcu_dereference_check(sk->sk_filter,
				       atomic_read(&sk->sk_wmem_alloc) == 0);
L
Linus Torvalds 已提交
1110
	if (filter) {
1111
		sk_filter_uncharge(sk, filter);
1112
		rcu_assign_pointer(sk->sk_filter, NULL);
L
Linus Torvalds 已提交
1113 1114
	}

1115 1116
	sock_disable_timestamp(sk, SOCK_TIMESTAMP);
	sock_disable_timestamp(sk, SOCK_TIMESTAMPING_RX_SOFTWARE);
L
Linus Torvalds 已提交
1117 1118 1119

	if (atomic_read(&sk->sk_omem_alloc))
		printk(KERN_DEBUG "%s: optmem leakage (%d bytes) detected.\n",
1120
		       __func__, atomic_read(&sk->sk_omem_alloc));
L
Linus Torvalds 已提交
1121

1122
	put_net(sock_net(sk));
1123
	sk_prot_free(sk->sk_prot_creator, sk);
L
Linus Torvalds 已提交
1124
}
1125 1126 1127 1128 1129 1130 1131 1132 1133 1134 1135

void sk_free(struct sock *sk)
{
	/*
	 * We substract one from sk_wmem_alloc and can know if
	 * some packets are still in some tx queue.
	 * If not null, sock_wfree() will call __sk_free(sk) later
	 */
	if (atomic_dec_and_test(&sk->sk_wmem_alloc))
		__sk_free(sk);
}
E
Eric Dumazet 已提交
1136
EXPORT_SYMBOL(sk_free);
L
Linus Torvalds 已提交
1137

1138 1139 1140 1141 1142 1143 1144 1145 1146 1147 1148 1149 1150 1151
/*
 * Last sock_put should drop referrence to sk->sk_net. It has already
 * been dropped in sk_change_net. Taking referrence to stopping namespace
 * is not an option.
 * Take referrence to a socket to remove it from hash _alive_ and after that
 * destroy it in the context of init_net.
 */
void sk_release_kernel(struct sock *sk)
{
	if (sk == NULL || sk->sk_socket == NULL)
		return;

	sock_hold(sk);
	sock_release(sk->sk_socket);
1152
	release_net(sock_net(sk));
1153
	sock_net_set(sk, get_net(&init_net));
1154 1155
	sock_put(sk);
}
1156
EXPORT_SYMBOL(sk_release_kernel);
1157

A
Al Viro 已提交
1158
struct sock *sk_clone(const struct sock *sk, const gfp_t priority)
1159
{
1160
	struct sock *newsk;
1161

1162
	newsk = sk_prot_alloc(sk->sk_prot, priority, sk->sk_family);
1163 1164 1165
	if (newsk != NULL) {
		struct sk_filter *filter;

1166
		sock_copy(newsk, sk);
1167 1168

		/* SANITY */
1169
		get_net(sock_net(newsk));
1170 1171 1172
		sk_node_init(&newsk->sk_node);
		sock_lock_init(newsk);
		bh_lock_sock(newsk);
1173
		newsk->sk_backlog.head	= newsk->sk_backlog.tail = NULL;
Z
Zhu Yi 已提交
1174
		newsk->sk_backlog.len = 0;
1175 1176

		atomic_set(&newsk->sk_rmem_alloc, 0);
1177 1178 1179 1180
		/*
		 * sk_wmem_alloc set to one (see sk_free() and sock_wfree())
		 */
		atomic_set(&newsk->sk_wmem_alloc, 1);
1181 1182 1183
		atomic_set(&newsk->sk_omem_alloc, 0);
		skb_queue_head_init(&newsk->sk_receive_queue);
		skb_queue_head_init(&newsk->sk_write_queue);
1184 1185 1186
#ifdef CONFIG_NET_DMA
		skb_queue_head_init(&newsk->sk_async_wait_queue);
#endif
1187

E
Eric Dumazet 已提交
1188
		spin_lock_init(&newsk->sk_dst_lock);
1189
		rwlock_init(&newsk->sk_callback_lock);
1190 1191 1192
		lockdep_set_class_and_name(&newsk->sk_callback_lock,
				af_callback_keys + newsk->sk_family,
				af_family_clock_key_strings[newsk->sk_family]);
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

		newsk->sk_dst_cache	= NULL;
		newsk->sk_wmem_queued	= 0;
		newsk->sk_forward_alloc = 0;
		newsk->sk_send_head	= NULL;
		newsk->sk_userlocks	= sk->sk_userlocks & ~SOCK_BINDPORT_LOCK;

		sock_reset_flag(newsk, SOCK_DONE);
		skb_queue_head_init(&newsk->sk_error_queue);

		filter = newsk->sk_filter;
		if (filter != NULL)
			sk_filter_charge(newsk, filter);

		if (unlikely(xfrm_sk_clone_policy(newsk))) {
			/* It is still raw copy of parent, so invalidate
			 * destructor and make plain sk_free() */
			newsk->sk_destruct = NULL;
			sk_free(newsk);
			newsk = NULL;
			goto out;
		}

		newsk->sk_err	   = 0;
		newsk->sk_priority = 0;
E
Eric Dumazet 已提交
1218 1219 1220 1221 1222
		/*
		 * Before updating sk_refcnt, we must commit prior changes to memory
		 * (Documentation/RCU/rculist_nulls.txt for details)
		 */
		smp_wmb();
1223 1224 1225 1226 1227 1228 1229 1230 1231 1232 1233 1234 1235 1236
		atomic_set(&newsk->sk_refcnt, 2);

		/*
		 * Increment the counter in the same struct proto as the master
		 * sock (sk_refcnt_debug_inc uses newsk->sk_prot->socks, that
		 * is the same as sk->sk_prot->socks, as this field was copied
		 * with memcpy).
		 *
		 * This _changes_ the previous behaviour, where
		 * tcp_create_openreq_child always was incrementing the
		 * equivalent to tcp_prot->socks (inet_sock_nr), so this have
		 * to be taken into account in all callers. -acme
		 */
		sk_refcnt_debug_inc(newsk);
1237
		sk_set_socket(newsk, NULL);
1238
		newsk->sk_wq = NULL;
1239 1240

		if (newsk->sk_prot->sockets_allocated)
1241
			percpu_counter_inc(newsk->sk_prot->sockets_allocated);
1242 1243 1244 1245

		if (sock_flag(newsk, SOCK_TIMESTAMP) ||
		    sock_flag(newsk, SOCK_TIMESTAMPING_RX_SOFTWARE))
			net_enable_timestamp();
1246 1247 1248 1249 1250 1251
	}
out:
	return newsk;
}
EXPORT_SYMBOL_GPL(sk_clone);

1252 1253 1254 1255 1256
void sk_setup_caps(struct sock *sk, struct dst_entry *dst)
{
	__sk_dst_set(sk, dst);
	sk->sk_route_caps = dst->dev->features;
	if (sk->sk_route_caps & NETIF_F_GSO)
1257
		sk->sk_route_caps |= NETIF_F_GSO_SOFTWARE;
E
Eric Dumazet 已提交
1258
	sk->sk_route_caps &= ~sk->sk_route_nocaps;
1259
	if (sk_can_gso(sk)) {
1260
		if (dst->header_len) {
1261
			sk->sk_route_caps &= ~NETIF_F_GSO_MASK;
1262
		} else {
1263
			sk->sk_route_caps |= NETIF_F_SG | NETIF_F_HW_CSUM;
1264 1265
			sk->sk_gso_max_size = dst->dev->gso_max_size;
		}
1266 1267 1268 1269
	}
}
EXPORT_SYMBOL_GPL(sk_setup_caps);

L
Linus Torvalds 已提交
1270 1271
void __init sk_init(void)
{
1272
	if (totalram_pages <= 4096) {
L
Linus Torvalds 已提交
1273 1274 1275 1276
		sysctl_wmem_max = 32767;
		sysctl_rmem_max = 32767;
		sysctl_wmem_default = 32767;
		sysctl_rmem_default = 32767;
1277
	} else if (totalram_pages >= 131072) {
L
Linus Torvalds 已提交
1278 1279 1280 1281 1282 1283 1284 1285 1286 1287
		sysctl_wmem_max = 131071;
		sysctl_rmem_max = 131071;
	}
}

/*
 *	Simple resource managers for sockets.
 */


1288 1289
/*
 * Write buffer destructor automatically called from kfree_skb.
L
Linus Torvalds 已提交
1290 1291 1292 1293
 */
void sock_wfree(struct sk_buff *skb)
{
	struct sock *sk = skb->sk;
E
Eric Dumazet 已提交
1294
	unsigned int len = skb->truesize;
L
Linus Torvalds 已提交
1295

E
Eric Dumazet 已提交
1296 1297 1298 1299 1300 1301
	if (!sock_flag(sk, SOCK_USE_WRITE_QUEUE)) {
		/*
		 * Keep a reference on sk_wmem_alloc, this will be released
		 * after sk_write_space() call
		 */
		atomic_sub(len - 1, &sk->sk_wmem_alloc);
L
Linus Torvalds 已提交
1302
		sk->sk_write_space(sk);
E
Eric Dumazet 已提交
1303 1304
		len = 1;
	}
1305
	/*
E
Eric Dumazet 已提交
1306 1307
	 * if sk_wmem_alloc reaches 0, we must finish what sk_free()
	 * could not do because of in-flight packets
1308
	 */
E
Eric Dumazet 已提交
1309
	if (atomic_sub_and_test(len, &sk->sk_wmem_alloc))
1310
		__sk_free(sk);
L
Linus Torvalds 已提交
1311
}
E
Eric Dumazet 已提交
1312
EXPORT_SYMBOL(sock_wfree);
L
Linus Torvalds 已提交
1313

1314 1315
/*
 * Read buffer destructor automatically called from kfree_skb.
L
Linus Torvalds 已提交
1316 1317 1318 1319 1320 1321
 */
void sock_rfree(struct sk_buff *skb)
{
	struct sock *sk = skb->sk;

	atomic_sub(skb->truesize, &sk->sk_rmem_alloc);
1322
	sk_mem_uncharge(skb->sk, skb->truesize);
L
Linus Torvalds 已提交
1323
}
E
Eric Dumazet 已提交
1324
EXPORT_SYMBOL(sock_rfree);
L
Linus Torvalds 已提交
1325 1326 1327 1328 1329 1330 1331 1332 1333 1334 1335


int sock_i_uid(struct sock *sk)
{
	int uid;

	read_lock(&sk->sk_callback_lock);
	uid = sk->sk_socket ? SOCK_INODE(sk->sk_socket)->i_uid : 0;
	read_unlock(&sk->sk_callback_lock);
	return uid;
}
E
Eric Dumazet 已提交
1336
EXPORT_SYMBOL(sock_i_uid);
L
Linus Torvalds 已提交
1337 1338 1339 1340 1341 1342 1343 1344 1345 1346

unsigned long sock_i_ino(struct sock *sk)
{
	unsigned long ino;

	read_lock(&sk->sk_callback_lock);
	ino = sk->sk_socket ? SOCK_INODE(sk->sk_socket)->i_ino : 0;
	read_unlock(&sk->sk_callback_lock);
	return ino;
}
E
Eric Dumazet 已提交
1347
EXPORT_SYMBOL(sock_i_ino);
L
Linus Torvalds 已提交
1348 1349 1350 1351

/*
 * Allocate a skb from the socket's send buffer.
 */
V
Victor Fusco 已提交
1352
struct sk_buff *sock_wmalloc(struct sock *sk, unsigned long size, int force,
A
Al Viro 已提交
1353
			     gfp_t priority)
L
Linus Torvalds 已提交
1354 1355
{
	if (force || atomic_read(&sk->sk_wmem_alloc) < sk->sk_sndbuf) {
E
Eric Dumazet 已提交
1356
		struct sk_buff *skb = alloc_skb(size, priority);
L
Linus Torvalds 已提交
1357 1358 1359 1360 1361 1362 1363
		if (skb) {
			skb_set_owner_w(skb, sk);
			return skb;
		}
	}
	return NULL;
}
E
Eric Dumazet 已提交
1364
EXPORT_SYMBOL(sock_wmalloc);
L
Linus Torvalds 已提交
1365 1366 1367

/*
 * Allocate a skb from the socket's receive buffer.
1368
 */
V
Victor Fusco 已提交
1369
struct sk_buff *sock_rmalloc(struct sock *sk, unsigned long size, int force,
A
Al Viro 已提交
1370
			     gfp_t priority)
L
Linus Torvalds 已提交
1371 1372 1373 1374 1375 1376 1377 1378 1379 1380 1381
{
	if (force || atomic_read(&sk->sk_rmem_alloc) < sk->sk_rcvbuf) {
		struct sk_buff *skb = alloc_skb(size, priority);
		if (skb) {
			skb_set_owner_r(skb, sk);
			return skb;
		}
	}
	return NULL;
}

1382
/*
L
Linus Torvalds 已提交
1383
 * Allocate a memory block from the socket's option memory buffer.
1384
 */
A
Al Viro 已提交
1385
void *sock_kmalloc(struct sock *sk, int size, gfp_t priority)
L
Linus Torvalds 已提交
1386 1387 1388 1389 1390
{
	if ((unsigned)size <= sysctl_optmem_max &&
	    atomic_read(&sk->sk_omem_alloc) + size < sysctl_optmem_max) {
		void *mem;
		/* First do the add, to avoid the race if kmalloc
1391
		 * might sleep.
L
Linus Torvalds 已提交
1392 1393 1394 1395 1396 1397 1398 1399 1400
		 */
		atomic_add(size, &sk->sk_omem_alloc);
		mem = kmalloc(size, priority);
		if (mem)
			return mem;
		atomic_sub(size, &sk->sk_omem_alloc);
	}
	return NULL;
}
E
Eric Dumazet 已提交
1401
EXPORT_SYMBOL(sock_kmalloc);
L
Linus Torvalds 已提交
1402 1403 1404 1405 1406 1407 1408 1409 1410

/*
 * Free an option memory block.
 */
void sock_kfree_s(struct sock *sk, void *mem, int size)
{
	kfree(mem);
	atomic_sub(size, &sk->sk_omem_alloc);
}
E
Eric Dumazet 已提交
1411
EXPORT_SYMBOL(sock_kfree_s);
L
Linus Torvalds 已提交
1412 1413 1414 1415

/* It is almost wait_for_tcp_memory minus release_sock/lock_sock.
   I think, these locks should be removed for datagram sockets.
 */
E
Eric Dumazet 已提交
1416
static long sock_wait_for_wmem(struct sock *sk, long timeo)
L
Linus Torvalds 已提交
1417 1418 1419 1420 1421 1422 1423 1424 1425 1426
{
	DEFINE_WAIT(wait);

	clear_bit(SOCK_ASYNC_NOSPACE, &sk->sk_socket->flags);
	for (;;) {
		if (!timeo)
			break;
		if (signal_pending(current))
			break;
		set_bit(SOCK_NOSPACE, &sk->sk_socket->flags);
E
Eric Dumazet 已提交
1427
		prepare_to_wait(sk_sleep(sk), &wait, TASK_INTERRUPTIBLE);
L
Linus Torvalds 已提交
1428 1429 1430 1431 1432 1433 1434 1435
		if (atomic_read(&sk->sk_wmem_alloc) < sk->sk_sndbuf)
			break;
		if (sk->sk_shutdown & SEND_SHUTDOWN)
			break;
		if (sk->sk_err)
			break;
		timeo = schedule_timeout(timeo);
	}
E
Eric Dumazet 已提交
1436
	finish_wait(sk_sleep(sk), &wait);
L
Linus Torvalds 已提交
1437 1438 1439 1440 1441 1442 1443 1444
	return timeo;
}


/*
 *	Generic send/receive buffer handlers
 */

H
Herbert Xu 已提交
1445 1446 1447
struct sk_buff *sock_alloc_send_pskb(struct sock *sk, unsigned long header_len,
				     unsigned long data_len, int noblock,
				     int *errcode)
L
Linus Torvalds 已提交
1448 1449
{
	struct sk_buff *skb;
A
Al Viro 已提交
1450
	gfp_t gfp_mask;
L
Linus Torvalds 已提交
1451 1452 1453 1454 1455 1456 1457 1458 1459 1460 1461 1462 1463 1464 1465 1466 1467 1468
	long timeo;
	int err;

	gfp_mask = sk->sk_allocation;
	if (gfp_mask & __GFP_WAIT)
		gfp_mask |= __GFP_REPEAT;

	timeo = sock_sndtimeo(sk, noblock);
	while (1) {
		err = sock_error(sk);
		if (err != 0)
			goto failure;

		err = -EPIPE;
		if (sk->sk_shutdown & SEND_SHUTDOWN)
			goto failure;

		if (atomic_read(&sk->sk_wmem_alloc) < sk->sk_sndbuf) {
1469
			skb = alloc_skb(header_len, gfp_mask);
L
Linus Torvalds 已提交
1470 1471 1472 1473 1474 1475 1476 1477 1478 1479 1480 1481 1482 1483 1484 1485 1486 1487 1488 1489 1490 1491 1492 1493 1494 1495 1496 1497 1498 1499 1500 1501 1502 1503 1504 1505 1506 1507 1508 1509 1510 1511 1512 1513 1514 1515 1516 1517 1518 1519 1520 1521 1522 1523 1524 1525 1526
			if (skb) {
				int npages;
				int i;

				/* No pages, we're done... */
				if (!data_len)
					break;

				npages = (data_len + (PAGE_SIZE - 1)) >> PAGE_SHIFT;
				skb->truesize += data_len;
				skb_shinfo(skb)->nr_frags = npages;
				for (i = 0; i < npages; i++) {
					struct page *page;
					skb_frag_t *frag;

					page = alloc_pages(sk->sk_allocation, 0);
					if (!page) {
						err = -ENOBUFS;
						skb_shinfo(skb)->nr_frags = i;
						kfree_skb(skb);
						goto failure;
					}

					frag = &skb_shinfo(skb)->frags[i];
					frag->page = page;
					frag->page_offset = 0;
					frag->size = (data_len >= PAGE_SIZE ?
						      PAGE_SIZE :
						      data_len);
					data_len -= PAGE_SIZE;
				}

				/* Full success... */
				break;
			}
			err = -ENOBUFS;
			goto failure;
		}
		set_bit(SOCK_ASYNC_NOSPACE, &sk->sk_socket->flags);
		set_bit(SOCK_NOSPACE, &sk->sk_socket->flags);
		err = -EAGAIN;
		if (!timeo)
			goto failure;
		if (signal_pending(current))
			goto interrupted;
		timeo = sock_wait_for_wmem(sk, timeo);
	}

	skb_set_owner_w(skb, sk);
	return skb;

interrupted:
	err = sock_intr_errno(timeo);
failure:
	*errcode = err;
	return NULL;
}
H
Herbert Xu 已提交
1527
EXPORT_SYMBOL(sock_alloc_send_pskb);
L
Linus Torvalds 已提交
1528

1529
struct sk_buff *sock_alloc_send_skb(struct sock *sk, unsigned long size,
L
Linus Torvalds 已提交
1530 1531 1532 1533
				    int noblock, int *errcode)
{
	return sock_alloc_send_pskb(sk, size, 0, noblock, errcode);
}
E
Eric Dumazet 已提交
1534
EXPORT_SYMBOL(sock_alloc_send_skb);
L
Linus Torvalds 已提交
1535 1536 1537 1538 1539

static void __lock_sock(struct sock *sk)
{
	DEFINE_WAIT(wait);

1540
	for (;;) {
L
Linus Torvalds 已提交
1541 1542 1543 1544 1545
		prepare_to_wait_exclusive(&sk->sk_lock.wq, &wait,
					TASK_UNINTERRUPTIBLE);
		spin_unlock_bh(&sk->sk_lock.slock);
		schedule();
		spin_lock_bh(&sk->sk_lock.slock);
1546
		if (!sock_owned_by_user(sk))
L
Linus Torvalds 已提交
1547 1548 1549 1550 1551 1552 1553 1554 1555 1556 1557 1558 1559 1560 1561 1562
			break;
	}
	finish_wait(&sk->sk_lock.wq, &wait);
}

static void __release_sock(struct sock *sk)
{
	struct sk_buff *skb = sk->sk_backlog.head;

	do {
		sk->sk_backlog.head = sk->sk_backlog.tail = NULL;
		bh_unlock_sock(sk);

		do {
			struct sk_buff *next = skb->next;

E
Eric Dumazet 已提交
1563
			WARN_ON_ONCE(skb_dst_is_noref(skb));
L
Linus Torvalds 已提交
1564
			skb->next = NULL;
P
Peter Zijlstra 已提交
1565
			sk_backlog_rcv(sk, skb);
L
Linus Torvalds 已提交
1566 1567 1568 1569 1570 1571 1572 1573 1574 1575 1576 1577 1578

			/*
			 * We are in process context here with softirqs
			 * disabled, use cond_resched_softirq() to preempt.
			 * This is safe to do because we've taken the backlog
			 * queue private:
			 */
			cond_resched_softirq();

			skb = next;
		} while (skb != NULL);

		bh_lock_sock(sk);
1579
	} while ((skb = sk->sk_backlog.head) != NULL);
Z
Zhu Yi 已提交
1580 1581 1582 1583 1584 1585

	/*
	 * Doing the zeroing here guarantee we can not loop forever
	 * while a wild producer attempts to flood us.
	 */
	sk->sk_backlog.len = 0;
L
Linus Torvalds 已提交
1586 1587 1588 1589
}

/**
 * sk_wait_data - wait for data to arrive at sk_receive_queue
1590 1591
 * @sk:    sock to wait on
 * @timeo: for how long
L
Linus Torvalds 已提交
1592 1593 1594 1595 1596 1597 1598 1599 1600 1601 1602
 *
 * Now socket state including sk->sk_err is changed only under lock,
 * hence we may omit checks after joining wait queue.
 * We check receive queue before schedule() only as optimization;
 * it is very likely that release_sock() added new data.
 */
int sk_wait_data(struct sock *sk, long *timeo)
{
	int rc;
	DEFINE_WAIT(wait);

E
Eric Dumazet 已提交
1603
	prepare_to_wait(sk_sleep(sk), &wait, TASK_INTERRUPTIBLE);
L
Linus Torvalds 已提交
1604 1605 1606
	set_bit(SOCK_ASYNC_WAITDATA, &sk->sk_socket->flags);
	rc = sk_wait_event(sk, timeo, !skb_queue_empty(&sk->sk_receive_queue));
	clear_bit(SOCK_ASYNC_WAITDATA, &sk->sk_socket->flags);
E
Eric Dumazet 已提交
1607
	finish_wait(sk_sleep(sk), &wait);
L
Linus Torvalds 已提交
1608 1609 1610 1611
	return rc;
}
EXPORT_SYMBOL(sk_wait_data);

1612 1613 1614 1615 1616 1617 1618 1619 1620 1621 1622 1623 1624 1625 1626 1627 1628 1629 1630 1631 1632 1633 1634 1635 1636 1637 1638 1639 1640
/**
 *	__sk_mem_schedule - increase sk_forward_alloc and memory_allocated
 *	@sk: socket
 *	@size: memory size to allocate
 *	@kind: allocation type
 *
 *	If kind is SK_MEM_SEND, it means wmem allocation. Otherwise it means
 *	rmem allocation. This function assumes that protocols which have
 *	memory_pressure use sk_wmem_queued as write buffer accounting.
 */
int __sk_mem_schedule(struct sock *sk, int size, int kind)
{
	struct proto *prot = sk->sk_prot;
	int amt = sk_mem_pages(size);
	int allocated;

	sk->sk_forward_alloc += amt * SK_MEM_QUANTUM;
	allocated = atomic_add_return(amt, prot->memory_allocated);

	/* Under limit. */
	if (allocated <= prot->sysctl_mem[0]) {
		if (prot->memory_pressure && *prot->memory_pressure)
			*prot->memory_pressure = 0;
		return 1;
	}

	/* Under pressure. */
	if (allocated > prot->sysctl_mem[1])
		if (prot->enter_memory_pressure)
1641
			prot->enter_memory_pressure(sk);
1642 1643 1644 1645 1646 1647 1648 1649 1650 1651 1652 1653 1654 1655 1656 1657 1658 1659 1660

	/* Over hard limit. */
	if (allocated > prot->sysctl_mem[2])
		goto suppress_allocation;

	/* guarantee minimum buffer size under pressure */
	if (kind == SK_MEM_RECV) {
		if (atomic_read(&sk->sk_rmem_alloc) < prot->sysctl_rmem[0])
			return 1;
	} else { /* SK_MEM_SEND */
		if (sk->sk_type == SOCK_STREAM) {
			if (sk->sk_wmem_queued < prot->sysctl_wmem[0])
				return 1;
		} else if (atomic_read(&sk->sk_wmem_alloc) <
			   prot->sysctl_wmem[0])
				return 1;
	}

	if (prot->memory_pressure) {
1661 1662 1663 1664 1665 1666
		int alloc;

		if (!*prot->memory_pressure)
			return 1;
		alloc = percpu_counter_read_positive(prot->sockets_allocated);
		if (prot->sysctl_mem[2] > alloc *
1667 1668 1669 1670 1671 1672 1673 1674 1675 1676 1677 1678 1679 1680 1681 1682 1683 1684 1685 1686 1687 1688 1689 1690 1691 1692 1693 1694 1695 1696 1697 1698 1699
		    sk_mem_pages(sk->sk_wmem_queued +
				 atomic_read(&sk->sk_rmem_alloc) +
				 sk->sk_forward_alloc))
			return 1;
	}

suppress_allocation:

	if (kind == SK_MEM_SEND && sk->sk_type == SOCK_STREAM) {
		sk_stream_moderate_sndbuf(sk);

		/* Fail only if socket is _under_ its sndbuf.
		 * In this case we cannot block, so that we have to fail.
		 */
		if (sk->sk_wmem_queued + size >= sk->sk_sndbuf)
			return 1;
	}

	/* Alas. Undo changes. */
	sk->sk_forward_alloc -= amt * SK_MEM_QUANTUM;
	atomic_sub(amt, prot->memory_allocated);
	return 0;
}
EXPORT_SYMBOL(__sk_mem_schedule);

/**
 *	__sk_reclaim - reclaim memory_allocated
 *	@sk: socket
 */
void __sk_mem_reclaim(struct sock *sk)
{
	struct proto *prot = sk->sk_prot;

1700
	atomic_sub(sk->sk_forward_alloc >> SK_MEM_QUANTUM_SHIFT,
1701 1702 1703 1704 1705 1706 1707 1708 1709 1710
		   prot->memory_allocated);
	sk->sk_forward_alloc &= SK_MEM_QUANTUM - 1;

	if (prot->memory_pressure && *prot->memory_pressure &&
	    (atomic_read(prot->memory_allocated) < prot->sysctl_mem[0]))
		*prot->memory_pressure = 0;
}
EXPORT_SYMBOL(__sk_mem_reclaim);


L
Linus Torvalds 已提交
1711 1712 1713 1714 1715 1716 1717 1718 1719 1720 1721
/*
 * Set of default routines for initialising struct proto_ops when
 * the protocol does not support a particular function. In certain
 * cases where it makes no sense for a protocol to have a "do nothing"
 * function, some default processing is provided.
 */

int sock_no_bind(struct socket *sock, struct sockaddr *saddr, int len)
{
	return -EOPNOTSUPP;
}
E
Eric Dumazet 已提交
1722
EXPORT_SYMBOL(sock_no_bind);
L
Linus Torvalds 已提交
1723

1724
int sock_no_connect(struct socket *sock, struct sockaddr *saddr,
L
Linus Torvalds 已提交
1725 1726 1727 1728
		    int len, int flags)
{
	return -EOPNOTSUPP;
}
E
Eric Dumazet 已提交
1729
EXPORT_SYMBOL(sock_no_connect);
L
Linus Torvalds 已提交
1730 1731 1732 1733 1734

int sock_no_socketpair(struct socket *sock1, struct socket *sock2)
{
	return -EOPNOTSUPP;
}
E
Eric Dumazet 已提交
1735
EXPORT_SYMBOL(sock_no_socketpair);
L
Linus Torvalds 已提交
1736 1737 1738 1739 1740

int sock_no_accept(struct socket *sock, struct socket *newsock, int flags)
{
	return -EOPNOTSUPP;
}
E
Eric Dumazet 已提交
1741
EXPORT_SYMBOL(sock_no_accept);
L
Linus Torvalds 已提交
1742

1743
int sock_no_getname(struct socket *sock, struct sockaddr *saddr,
L
Linus Torvalds 已提交
1744 1745 1746 1747
		    int *len, int peer)
{
	return -EOPNOTSUPP;
}
E
Eric Dumazet 已提交
1748
EXPORT_SYMBOL(sock_no_getname);
L
Linus Torvalds 已提交
1749

E
Eric Dumazet 已提交
1750
unsigned int sock_no_poll(struct file *file, struct socket *sock, poll_table *pt)
L
Linus Torvalds 已提交
1751 1752 1753
{
	return 0;
}
E
Eric Dumazet 已提交
1754
EXPORT_SYMBOL(sock_no_poll);
L
Linus Torvalds 已提交
1755 1756 1757 1758 1759

int sock_no_ioctl(struct socket *sock, unsigned int cmd, unsigned long arg)
{
	return -EOPNOTSUPP;
}
E
Eric Dumazet 已提交
1760
EXPORT_SYMBOL(sock_no_ioctl);
L
Linus Torvalds 已提交
1761 1762 1763 1764 1765

int sock_no_listen(struct socket *sock, int backlog)
{
	return -EOPNOTSUPP;
}
E
Eric Dumazet 已提交
1766
EXPORT_SYMBOL(sock_no_listen);
L
Linus Torvalds 已提交
1767 1768 1769 1770 1771

int sock_no_shutdown(struct socket *sock, int how)
{
	return -EOPNOTSUPP;
}
E
Eric Dumazet 已提交
1772
EXPORT_SYMBOL(sock_no_shutdown);
L
Linus Torvalds 已提交
1773 1774

int sock_no_setsockopt(struct socket *sock, int level, int optname,
1775
		    char __user *optval, unsigned int optlen)
L
Linus Torvalds 已提交
1776 1777 1778
{
	return -EOPNOTSUPP;
}
E
Eric Dumazet 已提交
1779
EXPORT_SYMBOL(sock_no_setsockopt);
L
Linus Torvalds 已提交
1780 1781 1782 1783 1784 1785

int sock_no_getsockopt(struct socket *sock, int level, int optname,
		    char __user *optval, int __user *optlen)
{
	return -EOPNOTSUPP;
}
E
Eric Dumazet 已提交
1786
EXPORT_SYMBOL(sock_no_getsockopt);
L
Linus Torvalds 已提交
1787 1788 1789 1790 1791 1792

int sock_no_sendmsg(struct kiocb *iocb, struct socket *sock, struct msghdr *m,
		    size_t len)
{
	return -EOPNOTSUPP;
}
E
Eric Dumazet 已提交
1793
EXPORT_SYMBOL(sock_no_sendmsg);
L
Linus Torvalds 已提交
1794 1795 1796 1797 1798 1799

int sock_no_recvmsg(struct kiocb *iocb, struct socket *sock, struct msghdr *m,
		    size_t len, int flags)
{
	return -EOPNOTSUPP;
}
E
Eric Dumazet 已提交
1800
EXPORT_SYMBOL(sock_no_recvmsg);
L
Linus Torvalds 已提交
1801 1802 1803 1804 1805 1806

int sock_no_mmap(struct file *file, struct socket *sock, struct vm_area_struct *vma)
{
	/* Mirror missing mmap method error code */
	return -ENODEV;
}
E
Eric Dumazet 已提交
1807
EXPORT_SYMBOL(sock_no_mmap);
L
Linus Torvalds 已提交
1808 1809 1810 1811 1812 1813 1814 1815 1816 1817 1818 1819 1820

ssize_t sock_no_sendpage(struct socket *sock, struct page *page, int offset, size_t size, int flags)
{
	ssize_t res;
	struct msghdr msg = {.msg_flags = flags};
	struct kvec iov;
	char *kaddr = kmap(page);
	iov.iov_base = kaddr + offset;
	iov.iov_len = size;
	res = kernel_sendmsg(sock, &msg, &iov, 1, size);
	kunmap(page);
	return res;
}
E
Eric Dumazet 已提交
1821
EXPORT_SYMBOL(sock_no_sendpage);
L
Linus Torvalds 已提交
1822 1823 1824 1825 1826 1827 1828

/*
 *	Default Socket Callbacks
 */

static void sock_def_wakeup(struct sock *sk)
{
1829 1830 1831 1832 1833 1834 1835
	struct socket_wq *wq;

	rcu_read_lock();
	wq = rcu_dereference(sk->sk_wq);
	if (wq_has_sleeper(wq))
		wake_up_interruptible_all(&wq->wait);
	rcu_read_unlock();
L
Linus Torvalds 已提交
1836 1837 1838 1839
}

static void sock_def_error_report(struct sock *sk)
{
1840 1841 1842 1843 1844 1845
	struct socket_wq *wq;

	rcu_read_lock();
	wq = rcu_dereference(sk->sk_wq);
	if (wq_has_sleeper(wq))
		wake_up_interruptible_poll(&wq->wait, POLLERR);
1846
	sk_wake_async(sk, SOCK_WAKE_IO, POLL_ERR);
1847
	rcu_read_unlock();
L
Linus Torvalds 已提交
1848 1849 1850 1851
}

static void sock_def_readable(struct sock *sk, int len)
{
1852 1853 1854 1855 1856 1857
	struct socket_wq *wq;

	rcu_read_lock();
	wq = rcu_dereference(sk->sk_wq);
	if (wq_has_sleeper(wq))
		wake_up_interruptible_sync_poll(&wq->wait, POLLIN |
1858
						POLLRDNORM | POLLRDBAND);
1859
	sk_wake_async(sk, SOCK_WAKE_WAITD, POLL_IN);
1860
	rcu_read_unlock();
L
Linus Torvalds 已提交
1861 1862 1863 1864
}

static void sock_def_write_space(struct sock *sk)
{
1865 1866 1867
	struct socket_wq *wq;

	rcu_read_lock();
L
Linus Torvalds 已提交
1868 1869 1870 1871

	/* Do not wake up a writer until he can make "significant"
	 * progress.  --DaveM
	 */
1872
	if ((atomic_read(&sk->sk_wmem_alloc) << 1) <= sk->sk_sndbuf) {
1873 1874 1875
		wq = rcu_dereference(sk->sk_wq);
		if (wq_has_sleeper(wq))
			wake_up_interruptible_sync_poll(&wq->wait, POLLOUT |
1876
						POLLWRNORM | POLLWRBAND);
L
Linus Torvalds 已提交
1877 1878 1879

		/* Should agree with poll, otherwise some programs break */
		if (sock_writeable(sk))
1880
			sk_wake_async(sk, SOCK_WAKE_SPACE, POLL_OUT);
L
Linus Torvalds 已提交
1881 1882
	}

1883
	rcu_read_unlock();
L
Linus Torvalds 已提交
1884 1885 1886 1887
}

static void sock_def_destruct(struct sock *sk)
{
J
Jesper Juhl 已提交
1888
	kfree(sk->sk_protinfo);
L
Linus Torvalds 已提交
1889 1890 1891 1892 1893 1894
}

void sk_send_sigurg(struct sock *sk)
{
	if (sk->sk_socket && sk->sk_socket->file)
		if (send_sigurg(&sk->sk_socket->file->f_owner))
1895
			sk_wake_async(sk, SOCK_WAKE_URG, POLL_PRI);
L
Linus Torvalds 已提交
1896
}
E
Eric Dumazet 已提交
1897
EXPORT_SYMBOL(sk_send_sigurg);
L
Linus Torvalds 已提交
1898 1899 1900 1901 1902 1903 1904 1905 1906 1907 1908 1909 1910 1911 1912 1913 1914 1915 1916 1917 1918

void sk_reset_timer(struct sock *sk, struct timer_list* timer,
		    unsigned long expires)
{
	if (!mod_timer(timer, expires))
		sock_hold(sk);
}
EXPORT_SYMBOL(sk_reset_timer);

void sk_stop_timer(struct sock *sk, struct timer_list* timer)
{
	if (timer_pending(timer) && del_timer(timer))
		__sock_put(sk);
}
EXPORT_SYMBOL(sk_stop_timer);

void sock_init_data(struct socket *sock, struct sock *sk)
{
	skb_queue_head_init(&sk->sk_receive_queue);
	skb_queue_head_init(&sk->sk_write_queue);
	skb_queue_head_init(&sk->sk_error_queue);
1919 1920 1921
#ifdef CONFIG_NET_DMA
	skb_queue_head_init(&sk->sk_async_wait_queue);
#endif
L
Linus Torvalds 已提交
1922 1923 1924 1925

	sk->sk_send_head	=	NULL;

	init_timer(&sk->sk_timer);
1926

L
Linus Torvalds 已提交
1927 1928 1929 1930
	sk->sk_allocation	=	GFP_KERNEL;
	sk->sk_rcvbuf		=	sysctl_rmem_default;
	sk->sk_sndbuf		=	sysctl_wmem_default;
	sk->sk_state		=	TCP_CLOSE;
1931
	sk_set_socket(sk, sock);
L
Linus Torvalds 已提交
1932 1933 1934

	sock_set_flag(sk, SOCK_ZAPPED);

1935
	if (sock) {
L
Linus Torvalds 已提交
1936
		sk->sk_type	=	sock->type;
1937
		sk->sk_wq	=	sock->wq;
L
Linus Torvalds 已提交
1938 1939
		sock->sk	=	sk;
	} else
1940
		sk->sk_wq	=	NULL;
L
Linus Torvalds 已提交
1941

E
Eric Dumazet 已提交
1942
	spin_lock_init(&sk->sk_dst_lock);
L
Linus Torvalds 已提交
1943
	rwlock_init(&sk->sk_callback_lock);
1944 1945 1946
	lockdep_set_class_and_name(&sk->sk_callback_lock,
			af_callback_keys + sk->sk_family,
			af_family_clock_key_strings[sk->sk_family]);
L
Linus Torvalds 已提交
1947 1948 1949 1950 1951 1952 1953 1954 1955 1956 1957 1958 1959 1960 1961 1962 1963 1964

	sk->sk_state_change	=	sock_def_wakeup;
	sk->sk_data_ready	=	sock_def_readable;
	sk->sk_write_space	=	sock_def_write_space;
	sk->sk_error_report	=	sock_def_error_report;
	sk->sk_destruct		=	sock_def_destruct;

	sk->sk_sndmsg_page	=	NULL;
	sk->sk_sndmsg_off	=	0;

	sk->sk_peercred.pid 	=	0;
	sk->sk_peercred.uid	=	-1;
	sk->sk_peercred.gid	=	-1;
	sk->sk_write_pending	=	0;
	sk->sk_rcvlowat		=	1;
	sk->sk_rcvtimeo		=	MAX_SCHEDULE_TIMEOUT;
	sk->sk_sndtimeo		=	MAX_SCHEDULE_TIMEOUT;

1965
	sk->sk_stamp = ktime_set(-1L, 0);
L
Linus Torvalds 已提交
1966

E
Eric Dumazet 已提交
1967 1968 1969 1970 1971
	/*
	 * Before updating sk_refcnt, we must commit prior changes to memory
	 * (Documentation/RCU/rculist_nulls.txt for details)
	 */
	smp_wmb();
L
Linus Torvalds 已提交
1972
	atomic_set(&sk->sk_refcnt, 1);
W
Wang Chen 已提交
1973
	atomic_set(&sk->sk_drops, 0);
L
Linus Torvalds 已提交
1974
}
E
Eric Dumazet 已提交
1975
EXPORT_SYMBOL(sock_init_data);
L
Linus Torvalds 已提交
1976

H
Harvey Harrison 已提交
1977
void lock_sock_nested(struct sock *sk, int subclass)
L
Linus Torvalds 已提交
1978 1979
{
	might_sleep();
I
Ingo Molnar 已提交
1980
	spin_lock_bh(&sk->sk_lock.slock);
1981
	if (sk->sk_lock.owned)
L
Linus Torvalds 已提交
1982
		__lock_sock(sk);
1983
	sk->sk_lock.owned = 1;
I
Ingo Molnar 已提交
1984 1985 1986 1987
	spin_unlock(&sk->sk_lock.slock);
	/*
	 * The sk_lock has mutex_lock() semantics here:
	 */
1988
	mutex_acquire(&sk->sk_lock.dep_map, subclass, 0, _RET_IP_);
I
Ingo Molnar 已提交
1989
	local_bh_enable();
L
Linus Torvalds 已提交
1990
}
1991
EXPORT_SYMBOL(lock_sock_nested);
L
Linus Torvalds 已提交
1992

H
Harvey Harrison 已提交
1993
void release_sock(struct sock *sk)
L
Linus Torvalds 已提交
1994
{
I
Ingo Molnar 已提交
1995 1996 1997 1998 1999 2000
	/*
	 * The sk_lock has mutex_unlock() semantics:
	 */
	mutex_release(&sk->sk_lock.dep_map, 1, _RET_IP_);

	spin_lock_bh(&sk->sk_lock.slock);
L
Linus Torvalds 已提交
2001 2002
	if (sk->sk_backlog.tail)
		__release_sock(sk);
2003
	sk->sk_lock.owned = 0;
I
Ingo Molnar 已提交
2004 2005 2006
	if (waitqueue_active(&sk->sk_lock.wq))
		wake_up(&sk->sk_lock.wq);
	spin_unlock_bh(&sk->sk_lock.slock);
L
Linus Torvalds 已提交
2007 2008 2009
}
EXPORT_SYMBOL(release_sock);

2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 2020 2021 2022 2023 2024 2025 2026 2027 2028 2029 2030 2031 2032 2033 2034 2035 2036 2037 2038 2039 2040 2041 2042
/**
 * lock_sock_fast - fast version of lock_sock
 * @sk: socket
 *
 * This version should be used for very small section, where process wont block
 * return false if fast path is taken
 *   sk_lock.slock locked, owned = 0, BH disabled
 * return true if slow path is taken
 *   sk_lock.slock unlocked, owned = 1, BH enabled
 */
bool lock_sock_fast(struct sock *sk)
{
	might_sleep();
	spin_lock_bh(&sk->sk_lock.slock);

	if (!sk->sk_lock.owned)
		/*
		 * Note : We must disable BH
		 */
		return false;

	__lock_sock(sk);
	sk->sk_lock.owned = 1;
	spin_unlock(&sk->sk_lock.slock);
	/*
	 * The sk_lock has mutex_lock() semantics here:
	 */
	mutex_acquire(&sk->sk_lock.dep_map, 0, 0, _RET_IP_);
	local_bh_enable();
	return true;
}
EXPORT_SYMBOL(lock_sock_fast);

L
Linus Torvalds 已提交
2043
int sock_get_timestamp(struct sock *sk, struct timeval __user *userstamp)
2044
{
2045
	struct timeval tv;
L
Linus Torvalds 已提交
2046
	if (!sock_flag(sk, SOCK_TIMESTAMP))
2047
		sock_enable_timestamp(sk, SOCK_TIMESTAMP);
2048 2049
	tv = ktime_to_timeval(sk->sk_stamp);
	if (tv.tv_sec == -1)
L
Linus Torvalds 已提交
2050
		return -ENOENT;
2051 2052 2053 2054 2055
	if (tv.tv_sec == 0) {
		sk->sk_stamp = ktime_get_real();
		tv = ktime_to_timeval(sk->sk_stamp);
	}
	return copy_to_user(userstamp, &tv, sizeof(tv)) ? -EFAULT : 0;
2056
}
L
Linus Torvalds 已提交
2057 2058
EXPORT_SYMBOL(sock_get_timestamp);

2059 2060 2061 2062
int sock_get_timestampns(struct sock *sk, struct timespec __user *userstamp)
{
	struct timespec ts;
	if (!sock_flag(sk, SOCK_TIMESTAMP))
2063
		sock_enable_timestamp(sk, SOCK_TIMESTAMP);
2064 2065 2066 2067 2068 2069 2070 2071 2072 2073 2074
	ts = ktime_to_timespec(sk->sk_stamp);
	if (ts.tv_sec == -1)
		return -ENOENT;
	if (ts.tv_sec == 0) {
		sk->sk_stamp = ktime_get_real();
		ts = ktime_to_timespec(sk->sk_stamp);
	}
	return copy_to_user(userstamp, &ts, sizeof(ts)) ? -EFAULT : 0;
}
EXPORT_SYMBOL(sock_get_timestampns);

2075
void sock_enable_timestamp(struct sock *sk, int flag)
2076
{
2077 2078 2079 2080 2081 2082 2083 2084 2085 2086 2087 2088
	if (!sock_flag(sk, flag)) {
		sock_set_flag(sk, flag);
		/*
		 * we just set one of the two flags which require net
		 * time stamping, but time stamping might have been on
		 * already because of the other one
		 */
		if (!sock_flag(sk,
				flag == SOCK_TIMESTAMP ?
				SOCK_TIMESTAMPING_RX_SOFTWARE :
				SOCK_TIMESTAMP))
			net_enable_timestamp();
L
Linus Torvalds 已提交
2089 2090 2091 2092 2093 2094 2095 2096 2097 2098 2099 2100 2101 2102 2103 2104 2105 2106 2107
	}
}

/*
 *	Get a socket option on an socket.
 *
 *	FIX: POSIX 1003.1g is very ambiguous here. It states that
 *	asynchronous errors should be reported by getsockopt. We assume
 *	this means if you specify SO_ERROR (otherwise whats the point of it).
 */
int sock_common_getsockopt(struct socket *sock, int level, int optname,
			   char __user *optval, int __user *optlen)
{
	struct sock *sk = sock->sk;

	return sk->sk_prot->getsockopt(sk, level, optname, optval, optlen);
}
EXPORT_SYMBOL(sock_common_getsockopt);

2108
#ifdef CONFIG_COMPAT
2109 2110
int compat_sock_common_getsockopt(struct socket *sock, int level, int optname,
				  char __user *optval, int __user *optlen)
2111 2112 2113
{
	struct sock *sk = sock->sk;

2114
	if (sk->sk_prot->compat_getsockopt != NULL)
2115 2116
		return sk->sk_prot->compat_getsockopt(sk, level, optname,
						      optval, optlen);
2117 2118 2119 2120 2121
	return sk->sk_prot->getsockopt(sk, level, optname, optval, optlen);
}
EXPORT_SYMBOL(compat_sock_common_getsockopt);
#endif

L
Linus Torvalds 已提交
2122 2123 2124 2125 2126 2127 2128 2129 2130 2131 2132 2133 2134 2135 2136 2137 2138 2139 2140
int sock_common_recvmsg(struct kiocb *iocb, struct socket *sock,
			struct msghdr *msg, size_t size, int flags)
{
	struct sock *sk = sock->sk;
	int addr_len = 0;
	int err;

	err = sk->sk_prot->recvmsg(iocb, sk, msg, size, flags & MSG_DONTWAIT,
				   flags & ~MSG_DONTWAIT, &addr_len);
	if (err >= 0)
		msg->msg_namelen = addr_len;
	return err;
}
EXPORT_SYMBOL(sock_common_recvmsg);

/*
 *	Set socket options on an inet socket.
 */
int sock_common_setsockopt(struct socket *sock, int level, int optname,
2141
			   char __user *optval, unsigned int optlen)
L
Linus Torvalds 已提交
2142 2143 2144 2145 2146 2147 2148
{
	struct sock *sk = sock->sk;

	return sk->sk_prot->setsockopt(sk, level, optname, optval, optlen);
}
EXPORT_SYMBOL(sock_common_setsockopt);

2149
#ifdef CONFIG_COMPAT
2150
int compat_sock_common_setsockopt(struct socket *sock, int level, int optname,
2151
				  char __user *optval, unsigned int optlen)
2152 2153 2154
{
	struct sock *sk = sock->sk;

2155 2156 2157
	if (sk->sk_prot->compat_setsockopt != NULL)
		return sk->sk_prot->compat_setsockopt(sk, level, optname,
						      optval, optlen);
2158 2159 2160 2161 2162
	return sk->sk_prot->setsockopt(sk, level, optname, optval, optlen);
}
EXPORT_SYMBOL(compat_sock_common_setsockopt);
#endif

L
Linus Torvalds 已提交
2163 2164 2165 2166 2167 2168 2169 2170 2171 2172 2173 2174 2175 2176 2177 2178 2179 2180 2181 2182 2183 2184 2185 2186 2187 2188 2189 2190 2191 2192 2193
void sk_common_release(struct sock *sk)
{
	if (sk->sk_prot->destroy)
		sk->sk_prot->destroy(sk);

	/*
	 * Observation: when sock_common_release is called, processes have
	 * no access to socket. But net still has.
	 * Step one, detach it from networking:
	 *
	 * A. Remove from hash tables.
	 */

	sk->sk_prot->unhash(sk);

	/*
	 * In this point socket cannot receive new packets, but it is possible
	 * that some packets are in flight because some CPU runs receiver and
	 * did hash table lookup before we unhashed socket. They will achieve
	 * receive queue and will be purged by socket destructor.
	 *
	 * Also we still have packets pending on receive queue and probably,
	 * our own packets waiting in device queues. sock_destroy will drain
	 * receive queue, but transmitted packets will delay socket destruction
	 * until the last reference will be released.
	 */

	sock_orphan(sk);

	xfrm_sk_free_policy(sk);

2194
	sk_refcnt_debug_release(sk);
L
Linus Torvalds 已提交
2195 2196 2197 2198 2199 2200 2201
	sock_put(sk);
}
EXPORT_SYMBOL(sk_common_release);

static DEFINE_RWLOCK(proto_list_lock);
static LIST_HEAD(proto_list);

2202 2203
#ifdef CONFIG_PROC_FS
#define PROTO_INUSE_NR	64	/* should be enough for the first time */
2204 2205 2206
struct prot_inuse {
	int val[PROTO_INUSE_NR];
};
2207 2208

static DECLARE_BITMAP(proto_inuse_idx, PROTO_INUSE_NR);
2209 2210 2211 2212 2213 2214 2215 2216 2217 2218 2219 2220 2221 2222 2223 2224 2225 2226 2227 2228 2229

#ifdef CONFIG_NET_NS
void sock_prot_inuse_add(struct net *net, struct proto *prot, int val)
{
	int cpu = smp_processor_id();
	per_cpu_ptr(net->core.inuse, cpu)->val[prot->inuse_idx] += val;
}
EXPORT_SYMBOL_GPL(sock_prot_inuse_add);

int sock_prot_inuse_get(struct net *net, struct proto *prot)
{
	int cpu, idx = prot->inuse_idx;
	int res = 0;

	for_each_possible_cpu(cpu)
		res += per_cpu_ptr(net->core.inuse, cpu)->val[idx];

	return res >= 0 ? res : 0;
}
EXPORT_SYMBOL_GPL(sock_prot_inuse_get);

2230
static int __net_init sock_inuse_init_net(struct net *net)
2231 2232 2233 2234 2235
{
	net->core.inuse = alloc_percpu(struct prot_inuse);
	return net->core.inuse ? 0 : -ENOMEM;
}

2236
static void __net_exit sock_inuse_exit_net(struct net *net)
2237 2238 2239 2240 2241 2242 2243 2244 2245 2246 2247 2248 2249 2250 2251 2252 2253 2254 2255
{
	free_percpu(net->core.inuse);
}

static struct pernet_operations net_inuse_ops = {
	.init = sock_inuse_init_net,
	.exit = sock_inuse_exit_net,
};

static __init int net_inuse_init(void)
{
	if (register_pernet_subsys(&net_inuse_ops))
		panic("Cannot initialize net inuse counters");

	return 0;
}

core_initcall(net_inuse_init);
#else
2256 2257
static DEFINE_PER_CPU(struct prot_inuse, prot_inuse);

2258
void sock_prot_inuse_add(struct net *net, struct proto *prot, int val)
2259 2260 2261 2262 2263
{
	__get_cpu_var(prot_inuse).val[prot->inuse_idx] += val;
}
EXPORT_SYMBOL_GPL(sock_prot_inuse_add);

2264
int sock_prot_inuse_get(struct net *net, struct proto *prot)
2265 2266 2267 2268 2269 2270 2271 2272 2273 2274
{
	int cpu, idx = prot->inuse_idx;
	int res = 0;

	for_each_possible_cpu(cpu)
		res += per_cpu(prot_inuse, cpu).val[idx];

	return res >= 0 ? res : 0;
}
EXPORT_SYMBOL_GPL(sock_prot_inuse_get);
2275
#endif
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

static void assign_proto_idx(struct proto *prot)
{
	prot->inuse_idx = find_first_zero_bit(proto_inuse_idx, PROTO_INUSE_NR);

	if (unlikely(prot->inuse_idx == PROTO_INUSE_NR - 1)) {
		printk(KERN_ERR "PROTO_INUSE_NR exhausted\n");
		return;
	}

	set_bit(prot->inuse_idx, proto_inuse_idx);
}

static void release_proto_idx(struct proto *prot)
{
	if (prot->inuse_idx != PROTO_INUSE_NR - 1)
		clear_bit(prot->inuse_idx, proto_inuse_idx);
}
#else
static inline void assign_proto_idx(struct proto *prot)
{
}

static inline void release_proto_idx(struct proto *prot)
{
}
#endif

2304 2305
int proto_register(struct proto *prot, int alloc_slab)
{
L
Linus Torvalds 已提交
2306 2307
	if (alloc_slab) {
		prot->slab = kmem_cache_create(prot->name, prot->obj_size, 0,
2308 2309
					SLAB_HWCACHE_ALIGN | prot->slab_flags,
					NULL);
L
Linus Torvalds 已提交
2310 2311 2312 2313

		if (prot->slab == NULL) {
			printk(KERN_CRIT "%s: Can't create sock SLAB cache!\n",
			       prot->name);
2314
			goto out;
L
Linus Torvalds 已提交
2315
		}
2316 2317

		if (prot->rsk_prot != NULL) {
2318
			prot->rsk_prot->slab_name = kasprintf(GFP_KERNEL, "request_sock_%s", prot->name);
2319
			if (prot->rsk_prot->slab_name == NULL)
2320 2321
				goto out_free_sock_slab;

2322
			prot->rsk_prot->slab = kmem_cache_create(prot->rsk_prot->slab_name,
2323
								 prot->rsk_prot->obj_size, 0,
2324
								 SLAB_HWCACHE_ALIGN, NULL);
2325 2326 2327 2328 2329 2330 2331

			if (prot->rsk_prot->slab == NULL) {
				printk(KERN_CRIT "%s: Can't create request sock SLAB cache!\n",
				       prot->name);
				goto out_free_request_sock_slab_name;
			}
		}
2332

2333
		if (prot->twsk_prot != NULL) {
2334
			prot->twsk_prot->twsk_slab_name = kasprintf(GFP_KERNEL, "tw_sock_%s", prot->name);
2335

2336
			if (prot->twsk_prot->twsk_slab_name == NULL)
2337 2338
				goto out_free_request_sock_slab;

2339
			prot->twsk_prot->twsk_slab =
2340
				kmem_cache_create(prot->twsk_prot->twsk_slab_name,
2341
						  prot->twsk_prot->twsk_obj_size,
2342 2343 2344
						  0,
						  SLAB_HWCACHE_ALIGN |
							prot->slab_flags,
2345
						  NULL);
2346
			if (prot->twsk_prot->twsk_slab == NULL)
2347 2348
				goto out_free_timewait_sock_slab_name;
		}
L
Linus Torvalds 已提交
2349 2350
	}

2351
	write_lock(&proto_list_lock);
L
Linus Torvalds 已提交
2352
	list_add(&prot->node, &proto_list);
2353
	assign_proto_idx(prot);
L
Linus Torvalds 已提交
2354
	write_unlock(&proto_list_lock);
2355 2356
	return 0;

2357
out_free_timewait_sock_slab_name:
2358
	kfree(prot->twsk_prot->twsk_slab_name);
2359 2360 2361 2362 2363
out_free_request_sock_slab:
	if (prot->rsk_prot && prot->rsk_prot->slab) {
		kmem_cache_destroy(prot->rsk_prot->slab);
		prot->rsk_prot->slab = NULL;
	}
2364
out_free_request_sock_slab_name:
2365 2366
	if (prot->rsk_prot)
		kfree(prot->rsk_prot->slab_name);
2367 2368 2369
out_free_sock_slab:
	kmem_cache_destroy(prot->slab);
	prot->slab = NULL;
2370 2371
out:
	return -ENOBUFS;
L
Linus Torvalds 已提交
2372 2373 2374 2375 2376 2377
}
EXPORT_SYMBOL(proto_register);

void proto_unregister(struct proto *prot)
{
	write_lock(&proto_list_lock);
2378
	release_proto_idx(prot);
2379 2380
	list_del(&prot->node);
	write_unlock(&proto_list_lock);
L
Linus Torvalds 已提交
2381 2382 2383 2384 2385 2386

	if (prot->slab != NULL) {
		kmem_cache_destroy(prot->slab);
		prot->slab = NULL;
	}

2387 2388
	if (prot->rsk_prot != NULL && prot->rsk_prot->slab != NULL) {
		kmem_cache_destroy(prot->rsk_prot->slab);
2389
		kfree(prot->rsk_prot->slab_name);
2390 2391 2392
		prot->rsk_prot->slab = NULL;
	}

2393 2394
	if (prot->twsk_prot != NULL && prot->twsk_prot->twsk_slab != NULL) {
		kmem_cache_destroy(prot->twsk_prot->twsk_slab);
2395
		kfree(prot->twsk_prot->twsk_slab_name);
2396
		prot->twsk_prot->twsk_slab = NULL;
2397
	}
L
Linus Torvalds 已提交
2398 2399 2400 2401 2402
}
EXPORT_SYMBOL(proto_unregister);

#ifdef CONFIG_PROC_FS
static void *proto_seq_start(struct seq_file *seq, loff_t *pos)
2403
	__acquires(proto_list_lock)
L
Linus Torvalds 已提交
2404 2405
{
	read_lock(&proto_list_lock);
2406
	return seq_list_start_head(&proto_list, *pos);
L
Linus Torvalds 已提交
2407 2408 2409 2410
}

static void *proto_seq_next(struct seq_file *seq, void *v, loff_t *pos)
{
2411
	return seq_list_next(v, &proto_list, pos);
L
Linus Torvalds 已提交
2412 2413 2414
}

static void proto_seq_stop(struct seq_file *seq, void *v)
2415
	__releases(proto_list_lock)
L
Linus Torvalds 已提交
2416 2417 2418 2419 2420 2421 2422 2423 2424 2425 2426 2427 2428 2429 2430
{
	read_unlock(&proto_list_lock);
}

static char proto_method_implemented(const void *method)
{
	return method == NULL ? 'n' : 'y';
}

static void proto_seq_printf(struct seq_file *seq, struct proto *proto)
{
	seq_printf(seq, "%-9s %4u %6d  %6d   %-3s %6u   %-3s  %-10s "
			"%2c %2c %2c %2c %2c %2c %2c %2c %2c %2c %2c %2c %2c %2c %2c %2c %2c %2c %2c\n",
		   proto->name,
		   proto->obj_size,
2431
		   sock_prot_inuse_get(seq_file_net(seq), proto),
L
Linus Torvalds 已提交
2432 2433 2434 2435 2436 2437 2438 2439 2440 2441 2442 2443 2444 2445 2446 2447 2448 2449 2450 2451 2452 2453 2454 2455 2456 2457 2458 2459
		   proto->memory_allocated != NULL ? atomic_read(proto->memory_allocated) : -1,
		   proto->memory_pressure != NULL ? *proto->memory_pressure ? "yes" : "no" : "NI",
		   proto->max_header,
		   proto->slab == NULL ? "no" : "yes",
		   module_name(proto->owner),
		   proto_method_implemented(proto->close),
		   proto_method_implemented(proto->connect),
		   proto_method_implemented(proto->disconnect),
		   proto_method_implemented(proto->accept),
		   proto_method_implemented(proto->ioctl),
		   proto_method_implemented(proto->init),
		   proto_method_implemented(proto->destroy),
		   proto_method_implemented(proto->shutdown),
		   proto_method_implemented(proto->setsockopt),
		   proto_method_implemented(proto->getsockopt),
		   proto_method_implemented(proto->sendmsg),
		   proto_method_implemented(proto->recvmsg),
		   proto_method_implemented(proto->sendpage),
		   proto_method_implemented(proto->bind),
		   proto_method_implemented(proto->backlog_rcv),
		   proto_method_implemented(proto->hash),
		   proto_method_implemented(proto->unhash),
		   proto_method_implemented(proto->get_port),
		   proto_method_implemented(proto->enter_memory_pressure));
}

static int proto_seq_show(struct seq_file *seq, void *v)
{
2460
	if (v == &proto_list)
L
Linus Torvalds 已提交
2461 2462 2463 2464 2465 2466 2467 2468 2469 2470 2471
		seq_printf(seq, "%-9s %-4s %-8s %-6s %-5s %-7s %-4s %-10s %s",
			   "protocol",
			   "size",
			   "sockets",
			   "memory",
			   "press",
			   "maxhdr",
			   "slab",
			   "module",
			   "cl co di ac io in de sh ss gs se re sp bi br ha uh gp em\n");
	else
2472
		proto_seq_printf(seq, list_entry(v, struct proto, node));
L
Linus Torvalds 已提交
2473 2474 2475
	return 0;
}

2476
static const struct seq_operations proto_seq_ops = {
L
Linus Torvalds 已提交
2477 2478 2479 2480 2481 2482 2483 2484
	.start  = proto_seq_start,
	.next   = proto_seq_next,
	.stop   = proto_seq_stop,
	.show   = proto_seq_show,
};

static int proto_seq_open(struct inode *inode, struct file *file)
{
2485 2486
	return seq_open_net(inode, file, &proto_seq_ops,
			    sizeof(struct seq_net_private));
L
Linus Torvalds 已提交
2487 2488
}

2489
static const struct file_operations proto_seq_fops = {
L
Linus Torvalds 已提交
2490 2491 2492 2493
	.owner		= THIS_MODULE,
	.open		= proto_seq_open,
	.read		= seq_read,
	.llseek		= seq_lseek,
2494 2495 2496 2497 2498 2499 2500 2501 2502 2503 2504 2505 2506 2507 2508 2509 2510 2511 2512 2513
	.release	= seq_release_net,
};

static __net_init int proto_init_net(struct net *net)
{
	if (!proc_net_fops_create(net, "protocols", S_IRUGO, &proto_seq_fops))
		return -ENOMEM;

	return 0;
}

static __net_exit void proto_exit_net(struct net *net)
{
	proc_net_remove(net, "protocols");
}


static __net_initdata struct pernet_operations proto_net_ops = {
	.init = proto_init_net,
	.exit = proto_exit_net,
L
Linus Torvalds 已提交
2514 2515 2516 2517
};

static int __init proto_init(void)
{
2518
	return register_pernet_subsys(&proto_net_ops);
L
Linus Torvalds 已提交
2519 2520 2521 2522 2523
}

subsys_initcall(proto_init);

#endif /* PROC_FS */