dev.c 147.6 KB
Newer Older
L
Linus Torvalds 已提交
1 2 3 4 5 6 7 8 9
/*
 * 	NET3	Protocol independent device support routines.
 *
 *		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.
 *
 *	Derived from the non IP parts of dev.c 1.0.19
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 35 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
 *				Fred N. van Kempen, <waltje@uWalt.NL.Mugnet.ORG>
 *				Mark Evans, <evansmp@uhura.aston.ac.uk>
 *
 *	Additional Authors:
 *		Florian la Roche <rzsfl@rz.uni-sb.de>
 *		Alan Cox <gw4pts@gw4pts.ampr.org>
 *		David Hinds <dahinds@users.sourceforge.net>
 *		Alexey Kuznetsov <kuznet@ms2.inr.ac.ru>
 *		Adam Sulmicki <adam@cfar.umd.edu>
 *              Pekka Riikonen <priikone@poesidon.pspt.fi>
 *
 *	Changes:
 *              D.J. Barrow     :       Fixed bug where dev->refcnt gets set
 *              			to 2 if register_netdev gets called
 *              			before net_dev_init & also removed a
 *              			few lines of code in the process.
 *		Alan Cox	:	device private ioctl copies fields back.
 *		Alan Cox	:	Transmit queue code does relevant
 *					stunts to keep the queue safe.
 *		Alan Cox	:	Fixed double lock.
 *		Alan Cox	:	Fixed promisc NULL pointer trap
 *		????????	:	Support the full private ioctl range
 *		Alan Cox	:	Moved ioctl permission check into
 *					drivers
 *		Tim Kordas	:	SIOCADDMULTI/SIOCDELMULTI
 *		Alan Cox	:	100 backlog just doesn't cut it when
 *					you start doing multicast video 8)
 *		Alan Cox	:	Rewrote net_bh and list manager.
 *		Alan Cox	: 	Fix ETH_P_ALL echoback lengths.
 *		Alan Cox	:	Took out transmit every packet pass
 *					Saved a few bytes in the ioctl handler
 *		Alan Cox	:	Network driver sets packet type before
 *					calling netif_rx. Saves a function
 *					call a packet.
 *		Alan Cox	:	Hashed net_bh()
 *		Richard Kooijman:	Timestamp fixes.
 *		Alan Cox	:	Wrong field in SIOCGIFDSTADDR
 *		Alan Cox	:	Device lock protection.
 *		Alan Cox	: 	Fixed nasty side effect of device close
 *					changes.
 *		Rudi Cilibrasi	:	Pass the right thing to
 *					set_mac_address()
 *		Dave Miller	:	32bit quantity for the device lock to
 *					make it work out on a Sparc.
 *		Bjorn Ekwall	:	Added KERNELD hack.
 *		Alan Cox	:	Cleaned up the backlog initialise.
 *		Craig Metz	:	SIOCGIFCONF fix if space for under
 *					1 device.
 *	    Thomas Bogendoerfer :	Return ENODEV for dev_open, if there
 *					is no device open function.
 *		Andi Kleen	:	Fix error reporting for SIOCGIFCONF
 *	    Michael Chastain	:	Fix signed/unsigned for SIOCGIFCONF
 *		Cyrus Durgin	:	Cleaned for KMOD
 *		Adam Sulmicki   :	Bug Fix : Network Device Unload
 *					A network device unload needs to purge
 *					the backlog queue.
 *	Paul Rusty Russell	:	SIOCSIFNAME
 *              Pekka Riikonen  :	Netdev boot-time settings code
 *              Andrew Morton   :       Make unregister_netdevice wait
 *              			indefinitely on dev->refcnt
 * 		J Hadi Salim	:	- Backlog queue sampling
 *				        - netif_rx() feedback
 */

#include <asm/uaccess.h>
#include <asm/system.h>
#include <linux/bitops.h>
78
#include <linux/capability.h>
L
Linus Torvalds 已提交
79 80 81
#include <linux/cpu.h>
#include <linux/types.h>
#include <linux/kernel.h>
82
#include <linux/hash.h>
83
#include <linux/slab.h>
L
Linus Torvalds 已提交
84
#include <linux/sched.h>
A
Arjan van de Ven 已提交
85
#include <linux/mutex.h>
L
Linus Torvalds 已提交
86 87 88 89 90 91 92 93 94
#include <linux/string.h>
#include <linux/mm.h>
#include <linux/socket.h>
#include <linux/sockios.h>
#include <linux/errno.h>
#include <linux/interrupt.h>
#include <linux/if_ether.h>
#include <linux/netdevice.h>
#include <linux/etherdevice.h>
95
#include <linux/ethtool.h>
L
Linus Torvalds 已提交
96 97
#include <linux/notifier.h>
#include <linux/skbuff.h>
98
#include <net/net_namespace.h>
L
Linus Torvalds 已提交
99 100 101 102 103 104 105 106
#include <net/sock.h>
#include <linux/rtnetlink.h>
#include <linux/proc_fs.h>
#include <linux/seq_file.h>
#include <linux/stat.h>
#include <net/dst.h>
#include <net/pkt_sched.h>
#include <net/checksum.h>
107
#include <net/xfrm.h>
L
Linus Torvalds 已提交
108 109 110 111 112 113 114
#include <linux/highmem.h>
#include <linux/init.h>
#include <linux/kmod.h>
#include <linux/module.h>
#include <linux/netpoll.h>
#include <linux/rcupdate.h>
#include <linux/delay.h>
115
#include <net/wext.h>
L
Linus Torvalds 已提交
116 117
#include <net/iw_handler.h>
#include <asm/current.h>
S
Steve Grubb 已提交
118
#include <linux/audit.h>
119
#include <linux/dmaengine.h>
120
#include <linux/err.h>
121
#include <linux/ctype.h>
122
#include <linux/if_arp.h>
123
#include <linux/if_vlan.h>
124
#include <linux/ip.h>
125
#include <net/ip.h>
126 127
#include <linux/ipv6.h>
#include <linux/in.h>
D
David S. Miller 已提交
128 129
#include <linux/jhash.h>
#include <linux/random.h>
130
#include <trace/events/napi.h>
131
#include <linux/pci.h>
L
Linus Torvalds 已提交
132

133 134
#include "net-sysfs.h"

135 136 137
/* Instead of increasing this, you should create a hash table. */
#define MAX_GRO_SKBS 8

H
Herbert Xu 已提交
138 139 140
/* This should be increased if a protocol with a bigger head is added. */
#define GRO_MAX_HEAD (MAX_HEADER + 128)

L
Linus Torvalds 已提交
141 142 143 144 145 146 147 148 149 150 151
/*
 *	The list of packet types we will receive (as opposed to discard)
 *	and the routines to invoke.
 *
 *	Why 16. Because with 16 the only overlap we get on a hash of the
 *	low nibble of the protocol value is RARP/SNAP/X.25.
 *
 *      NOTE:  That is no longer true with the addition of VLAN tags.  Not
 *             sure which should go first, but I bet it won't make much
 *             difference if we are running VLANs.  The good news is that
 *             this protocol won't be in the list unless compiled in, so
S
Stephen Hemminger 已提交
152
 *             the average user (w/out VLANs) will not be adversely affected.
L
Linus Torvalds 已提交
153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168
 *             --BLG
 *
 *		0800	IP
 *		8100    802.1Q VLAN
 *		0001	802.3
 *		0002	AX.25
 *		0004	802.2
 *		8035	RARP
 *		0005	SNAP
 *		0805	X.25
 *		0806	ARP
 *		8137	IPX
 *		0009	Localtalk
 *		86DD	IPv6
 */

169 170 171
#define PTYPE_HASH_SIZE	(16)
#define PTYPE_HASH_MASK	(PTYPE_HASH_SIZE - 1)

L
Linus Torvalds 已提交
172
static DEFINE_SPINLOCK(ptype_lock);
173
static struct list_head ptype_base[PTYPE_HASH_SIZE] __read_mostly;
174
static struct list_head ptype_all __read_mostly;	/* Taps */
L
Linus Torvalds 已提交
175 176

/*
177
 * The @dev_base_head list is protected by @dev_base_lock and the rtnl
L
Linus Torvalds 已提交
178 179
 * semaphore.
 *
180
 * Pure readers hold dev_base_lock for reading, or rcu_read_lock()
L
Linus Torvalds 已提交
181 182
 *
 * Writers must hold the rtnl semaphore while they loop through the
183
 * dev_base_head list, and hold dev_base_lock for writing when they do the
L
Linus Torvalds 已提交
184 185 186 187 188 189 190 191 192 193 194 195 196 197
 * actual updates.  This allows pure readers to access the list even
 * while a writer is preparing to update it.
 *
 * To put it another way, dev_base_lock is held for writing only to
 * protect against pure readers; the rtnl semaphore provides the
 * protection against other writers.
 *
 * See, for example usages, register_netdevice() and
 * unregister_netdevice(), which must be called with the rtnl
 * semaphore held.
 */
DEFINE_RWLOCK(dev_base_lock);
EXPORT_SYMBOL(dev_base_lock);

198
static inline struct hlist_head *dev_name_hash(struct net *net, const char *name)
L
Linus Torvalds 已提交
199 200
{
	unsigned hash = full_name_hash(name, strnlen(name, IFNAMSIZ));
201
	return &net->dev_name_head[hash_32(hash, NETDEV_HASHBITS)];
L
Linus Torvalds 已提交
202 203
}

204
static inline struct hlist_head *dev_index_hash(struct net *net, int ifindex)
L
Linus Torvalds 已提交
205
{
206
	return &net->dev_index_head[ifindex & (NETDEV_HASHENTRIES - 1)];
L
Linus Torvalds 已提交
207 208
}

E
Eric Dumazet 已提交
209
static inline void rps_lock(struct softnet_data *sd)
210 211
{
#ifdef CONFIG_RPS
E
Eric Dumazet 已提交
212
	spin_lock(&sd->input_pkt_queue.lock);
213 214 215
#endif
}

E
Eric Dumazet 已提交
216
static inline void rps_unlock(struct softnet_data *sd)
217 218
{
#ifdef CONFIG_RPS
E
Eric Dumazet 已提交
219
	spin_unlock(&sd->input_pkt_queue.lock);
220 221 222
#endif
}

223 224 225
/* Device list insertion */
static int list_netdevice(struct net_device *dev)
{
226
	struct net *net = dev_net(dev);
227 228 229 230

	ASSERT_RTNL();

	write_lock_bh(&dev_base_lock);
231
	list_add_tail_rcu(&dev->dev_list, &net->dev_base_head);
232
	hlist_add_head_rcu(&dev->name_hlist, dev_name_hash(net, dev->name));
233 234
	hlist_add_head_rcu(&dev->index_hlist,
			   dev_index_hash(net, dev->ifindex));
235 236 237 238
	write_unlock_bh(&dev_base_lock);
	return 0;
}

239 240 241
/* Device list removal
 * caller must respect a RCU grace period before freeing/reusing dev
 */
242 243 244 245 246 247
static void unlist_netdevice(struct net_device *dev)
{
	ASSERT_RTNL();

	/* Unlink dev from the device chain */
	write_lock_bh(&dev_base_lock);
248
	list_del_rcu(&dev->dev_list);
249
	hlist_del_rcu(&dev->name_hlist);
250
	hlist_del_rcu(&dev->index_hlist);
251 252 253
	write_unlock_bh(&dev_base_lock);
}

L
Linus Torvalds 已提交
254 255 256 257
/*
 *	Our notifier list
 */

258
static RAW_NOTIFIER_HEAD(netdev_chain);
L
Linus Torvalds 已提交
259 260 261 262 263

/*
 *	Device drivers call our routines to queue packets here. We empty the
 *	queue in the local softnet handler.
 */
264

265
DEFINE_PER_CPU_ALIGNED(struct softnet_data, softnet_data);
E
Eric Dumazet 已提交
266
EXPORT_PER_CPU_SYMBOL(softnet_data);
L
Linus Torvalds 已提交
267

268
#ifdef CONFIG_LOCKDEP
269
/*
270
 * register_netdevice() inits txq->_xmit_lock and sets lockdep class
271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286
 * according to dev->type
 */
static const unsigned short netdev_lock_type[] =
	{ARPHRD_NETROM, ARPHRD_ETHER, ARPHRD_EETHER, ARPHRD_AX25,
	 ARPHRD_PRONET, ARPHRD_CHAOS, ARPHRD_IEEE802, ARPHRD_ARCNET,
	 ARPHRD_APPLETLK, ARPHRD_DLCI, ARPHRD_ATM, ARPHRD_METRICOM,
	 ARPHRD_IEEE1394, ARPHRD_EUI64, ARPHRD_INFINIBAND, ARPHRD_SLIP,
	 ARPHRD_CSLIP, ARPHRD_SLIP6, ARPHRD_CSLIP6, ARPHRD_RSRVD,
	 ARPHRD_ADAPT, ARPHRD_ROSE, ARPHRD_X25, ARPHRD_HWX25,
	 ARPHRD_PPP, ARPHRD_CISCO, ARPHRD_LAPB, ARPHRD_DDCMP,
	 ARPHRD_RAWHDLC, ARPHRD_TUNNEL, ARPHRD_TUNNEL6, ARPHRD_FRAD,
	 ARPHRD_SKIP, ARPHRD_LOOPBACK, ARPHRD_LOCALTLK, ARPHRD_FDDI,
	 ARPHRD_BIF, ARPHRD_SIT, ARPHRD_IPDDP, ARPHRD_IPGRE,
	 ARPHRD_PIMREG, ARPHRD_HIPPI, ARPHRD_ASH, ARPHRD_ECONET,
	 ARPHRD_IRDA, ARPHRD_FCPP, ARPHRD_FCAL, ARPHRD_FCPL,
	 ARPHRD_FCFABRIC, ARPHRD_IEEE802_TR, ARPHRD_IEEE80211,
287
	 ARPHRD_IEEE80211_PRISM, ARPHRD_IEEE80211_RADIOTAP, ARPHRD_PHONET,
288
	 ARPHRD_PHONET_PIPE, ARPHRD_IEEE802154,
289
	 ARPHRD_VOID, ARPHRD_NONE};
290

291
static const char *const netdev_lock_name[] =
292 293 294 295 296 297 298 299 300 301 302 303 304
	{"_xmit_NETROM", "_xmit_ETHER", "_xmit_EETHER", "_xmit_AX25",
	 "_xmit_PRONET", "_xmit_CHAOS", "_xmit_IEEE802", "_xmit_ARCNET",
	 "_xmit_APPLETLK", "_xmit_DLCI", "_xmit_ATM", "_xmit_METRICOM",
	 "_xmit_IEEE1394", "_xmit_EUI64", "_xmit_INFINIBAND", "_xmit_SLIP",
	 "_xmit_CSLIP", "_xmit_SLIP6", "_xmit_CSLIP6", "_xmit_RSRVD",
	 "_xmit_ADAPT", "_xmit_ROSE", "_xmit_X25", "_xmit_HWX25",
	 "_xmit_PPP", "_xmit_CISCO", "_xmit_LAPB", "_xmit_DDCMP",
	 "_xmit_RAWHDLC", "_xmit_TUNNEL", "_xmit_TUNNEL6", "_xmit_FRAD",
	 "_xmit_SKIP", "_xmit_LOOPBACK", "_xmit_LOCALTLK", "_xmit_FDDI",
	 "_xmit_BIF", "_xmit_SIT", "_xmit_IPDDP", "_xmit_IPGRE",
	 "_xmit_PIMREG", "_xmit_HIPPI", "_xmit_ASH", "_xmit_ECONET",
	 "_xmit_IRDA", "_xmit_FCPP", "_xmit_FCAL", "_xmit_FCPL",
	 "_xmit_FCFABRIC", "_xmit_IEEE802_TR", "_xmit_IEEE80211",
305
	 "_xmit_IEEE80211_PRISM", "_xmit_IEEE80211_RADIOTAP", "_xmit_PHONET",
306
	 "_xmit_PHONET_PIPE", "_xmit_IEEE802154",
307
	 "_xmit_VOID", "_xmit_NONE"};
308 309

static struct lock_class_key netdev_xmit_lock_key[ARRAY_SIZE(netdev_lock_type)];
310
static struct lock_class_key netdev_addr_lock_key[ARRAY_SIZE(netdev_lock_type)];
311 312 313 314 315 316 317 318 319 320 321 322

static inline unsigned short netdev_lock_pos(unsigned short dev_type)
{
	int i;

	for (i = 0; i < ARRAY_SIZE(netdev_lock_type); i++)
		if (netdev_lock_type[i] == dev_type)
			return i;
	/* the last key is used by default */
	return ARRAY_SIZE(netdev_lock_type) - 1;
}

323 324
static inline void netdev_set_xmit_lockdep_class(spinlock_t *lock,
						 unsigned short dev_type)
325 326 327 328 329 330 331
{
	int i;

	i = netdev_lock_pos(dev_type);
	lockdep_set_class_and_name(lock, &netdev_xmit_lock_key[i],
				   netdev_lock_name[i]);
}
332 333 334 335 336 337 338 339 340 341

static inline void netdev_set_addr_lockdep_class(struct net_device *dev)
{
	int i;

	i = netdev_lock_pos(dev->type);
	lockdep_set_class_and_name(&dev->addr_list_lock,
				   &netdev_addr_lock_key[i],
				   netdev_lock_name[i]);
}
342
#else
343 344 345 346 347
static inline void netdev_set_xmit_lockdep_class(spinlock_t *lock,
						 unsigned short dev_type)
{
}
static inline void netdev_set_addr_lockdep_class(struct net_device *dev)
348 349 350
{
}
#endif
L
Linus Torvalds 已提交
351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381

/*******************************************************************************

		Protocol management and registration routines

*******************************************************************************/

/*
 *	Add a protocol ID to the list. Now that the input handler is
 *	smarter we can dispense with all the messy stuff that used to be
 *	here.
 *
 *	BEWARE!!! Protocol handlers, mangling input packets,
 *	MUST BE last in hash buckets and checking protocol handlers
 *	MUST start from promiscuous ptype_all chain in net_bh.
 *	It is true now, do not change it.
 *	Explanation follows: if protocol handler, mangling packet, will
 *	be the first on list, it is not able to sense, that packet
 *	is cloned and should be copied-on-write, so that it will
 *	change it and subsequent readers will get broken packet.
 *							--ANK (980803)
 */

/**
 *	dev_add_pack - add packet handler
 *	@pt: packet type declaration
 *
 *	Add a protocol handler to the networking stack. The passed &packet_type
 *	is linked into kernel lists and may not be freed until it has been
 *	removed from the kernel lists.
 *
382
 *	This call does not sleep therefore it can not
L
Linus Torvalds 已提交
383 384 385 386 387 388 389 390 391
 *	guarantee all CPU's that are in middle of receiving packets
 *	will see the new packet type (until the next received packet).
 */

void dev_add_pack(struct packet_type *pt)
{
	int hash;

	spin_lock_bh(&ptype_lock);
S
Stephen Hemminger 已提交
392
	if (pt->type == htons(ETH_P_ALL))
L
Linus Torvalds 已提交
393
		list_add_rcu(&pt->list, &ptype_all);
S
Stephen Hemminger 已提交
394
	else {
395
		hash = ntohs(pt->type) & PTYPE_HASH_MASK;
L
Linus Torvalds 已提交
396 397 398 399
		list_add_rcu(&pt->list, &ptype_base[hash]);
	}
	spin_unlock_bh(&ptype_lock);
}
E
Eric Dumazet 已提交
400
EXPORT_SYMBOL(dev_add_pack);
L
Linus Torvalds 已提交
401 402 403 404 405 406 407 408

/**
 *	__dev_remove_pack	 - remove packet handler
 *	@pt: packet type declaration
 *
 *	Remove a protocol handler that was previously added to the kernel
 *	protocol handlers by dev_add_pack(). The passed &packet_type is removed
 *	from the kernel lists and can be freed or reused once this function
409
 *	returns.
L
Linus Torvalds 已提交
410 411 412 413 414 415 416 417 418 419 420 421
 *
 *      The packet type might still be in use by receivers
 *	and must not be freed until after all the CPU's have gone
 *	through a quiescent state.
 */
void __dev_remove_pack(struct packet_type *pt)
{
	struct list_head *head;
	struct packet_type *pt1;

	spin_lock_bh(&ptype_lock);

S
Stephen Hemminger 已提交
422
	if (pt->type == htons(ETH_P_ALL))
L
Linus Torvalds 已提交
423
		head = &ptype_all;
S
Stephen Hemminger 已提交
424
	else
425
		head = &ptype_base[ntohs(pt->type) & PTYPE_HASH_MASK];
L
Linus Torvalds 已提交
426 427 428 429 430 431 432 433 434 435 436 437

	list_for_each_entry(pt1, head, list) {
		if (pt == pt1) {
			list_del_rcu(&pt->list);
			goto out;
		}
	}

	printk(KERN_WARNING "dev_remove_pack: %p not found.\n", pt);
out:
	spin_unlock_bh(&ptype_lock);
}
E
Eric Dumazet 已提交
438 439
EXPORT_SYMBOL(__dev_remove_pack);

L
Linus Torvalds 已提交
440 441 442 443 444 445 446 447 448 449 450 451 452 453 454
/**
 *	dev_remove_pack	 - remove packet handler
 *	@pt: packet type declaration
 *
 *	Remove a protocol handler that was previously added to the kernel
 *	protocol handlers by dev_add_pack(). The passed &packet_type is removed
 *	from the kernel lists and can be freed or reused once this function
 *	returns.
 *
 *	This call sleeps to guarantee that no CPU is looking at the packet
 *	type after return.
 */
void dev_remove_pack(struct packet_type *pt)
{
	__dev_remove_pack(pt);
455

L
Linus Torvalds 已提交
456 457
	synchronize_net();
}
E
Eric Dumazet 已提交
458
EXPORT_SYMBOL(dev_remove_pack);
L
Linus Torvalds 已提交
459 460 461 462 463 464 465 466 467 468 469 470 471 472 473 474 475 476 477 478 479 480 481 482 483 484 485 486

/******************************************************************************

		      Device Boot-time Settings Routines

*******************************************************************************/

/* Boot time configuration table */
static struct netdev_boot_setup dev_boot_setup[NETDEV_BOOT_SETUP_MAX];

/**
 *	netdev_boot_setup_add	- add new setup entry
 *	@name: name of the device
 *	@map: configured settings for the device
 *
 *	Adds new setup entry to the dev_boot_setup list.  The function
 *	returns 0 on error and 1 on success.  This is a generic routine to
 *	all netdevices.
 */
static int netdev_boot_setup_add(char *name, struct ifmap *map)
{
	struct netdev_boot_setup *s;
	int i;

	s = dev_boot_setup;
	for (i = 0; i < NETDEV_BOOT_SETUP_MAX; i++) {
		if (s[i].name[0] == '\0' || s[i].name[0] == ' ') {
			memset(s[i].name, 0, sizeof(s[i].name));
487
			strlcpy(s[i].name, name, IFNAMSIZ);
L
Linus Torvalds 已提交
488 489 490 491 492 493 494 495 496 497 498 499 500 501 502 503 504 505 506 507 508 509 510 511
			memcpy(&s[i].map, map, sizeof(s[i].map));
			break;
		}
	}

	return i >= NETDEV_BOOT_SETUP_MAX ? 0 : 1;
}

/**
 *	netdev_boot_setup_check	- check boot time settings
 *	@dev: the netdevice
 *
 * 	Check boot time settings for the device.
 *	The found settings are set for the device to be used
 *	later in the device probing.
 *	Returns 0 if no settings found, 1 if they are.
 */
int netdev_boot_setup_check(struct net_device *dev)
{
	struct netdev_boot_setup *s = dev_boot_setup;
	int i;

	for (i = 0; i < NETDEV_BOOT_SETUP_MAX; i++) {
		if (s[i].name[0] != '\0' && s[i].name[0] != ' ' &&
512
		    !strcmp(dev->name, s[i].name)) {
L
Linus Torvalds 已提交
513 514 515 516 517 518 519 520 521
			dev->irq 	= s[i].map.irq;
			dev->base_addr 	= s[i].map.base_addr;
			dev->mem_start 	= s[i].map.mem_start;
			dev->mem_end 	= s[i].map.mem_end;
			return 1;
		}
	}
	return 0;
}
E
Eric Dumazet 已提交
522
EXPORT_SYMBOL(netdev_boot_setup_check);
L
Linus Torvalds 已提交
523 524 525 526 527 528 529 530 531 532 533 534 535 536 537 538 539 540 541 542 543 544 545 546


/**
 *	netdev_boot_base	- get address from boot time settings
 *	@prefix: prefix for network device
 *	@unit: id for network device
 *
 * 	Check boot time settings for the base address of device.
 *	The found settings are set for the device to be used
 *	later in the device probing.
 *	Returns 0 if no settings found.
 */
unsigned long netdev_boot_base(const char *prefix, int unit)
{
	const struct netdev_boot_setup *s = dev_boot_setup;
	char name[IFNAMSIZ];
	int i;

	sprintf(name, "%s%d", prefix, unit);

	/*
	 * If device already registered then return base of 1
	 * to indicate not to probe for this interface
	 */
547
	if (__dev_get_by_name(&init_net, name))
L
Linus Torvalds 已提交
548 549 550 551 552 553 554 555 556 557 558 559 560 561 562 563 564 565 566 567 568 569 570 571 572 573 574 575 576 577 578 579 580 581 582 583 584 585 586 587 588 589 590 591 592
		return 1;

	for (i = 0; i < NETDEV_BOOT_SETUP_MAX; i++)
		if (!strcmp(name, s[i].name))
			return s[i].map.base_addr;
	return 0;
}

/*
 * Saves at boot time configured settings for any netdevice.
 */
int __init netdev_boot_setup(char *str)
{
	int ints[5];
	struct ifmap map;

	str = get_options(str, ARRAY_SIZE(ints), ints);
	if (!str || !*str)
		return 0;

	/* Save settings */
	memset(&map, 0, sizeof(map));
	if (ints[0] > 0)
		map.irq = ints[1];
	if (ints[0] > 1)
		map.base_addr = ints[2];
	if (ints[0] > 2)
		map.mem_start = ints[3];
	if (ints[0] > 3)
		map.mem_end = ints[4];

	/* Add new entry to the list */
	return netdev_boot_setup_add(str, &map);
}

__setup("netdev=", netdev_boot_setup);

/*******************************************************************************

			    Device Interface Subroutines

*******************************************************************************/

/**
 *	__dev_get_by_name	- find a device by its name
593
 *	@net: the applicable net namespace
L
Linus Torvalds 已提交
594 595 596 597 598 599 600 601 602
 *	@name: name to find
 *
 *	Find an interface by name. Must be called under RTNL semaphore
 *	or @dev_base_lock. If the name is found a pointer to the device
 *	is returned. If the name is not found then %NULL is returned. The
 *	reference counters are not incremented so the caller must be
 *	careful with locks.
 */

603
struct net_device *__dev_get_by_name(struct net *net, const char *name)
L
Linus Torvalds 已提交
604 605
{
	struct hlist_node *p;
E
Eric Dumazet 已提交
606 607
	struct net_device *dev;
	struct hlist_head *head = dev_name_hash(net, name);
L
Linus Torvalds 已提交
608

E
Eric Dumazet 已提交
609
	hlist_for_each_entry(dev, p, head, name_hlist)
L
Linus Torvalds 已提交
610 611
		if (!strncmp(dev->name, name, IFNAMSIZ))
			return dev;
E
Eric Dumazet 已提交
612

L
Linus Torvalds 已提交
613 614
	return NULL;
}
E
Eric Dumazet 已提交
615
EXPORT_SYMBOL(__dev_get_by_name);
L
Linus Torvalds 已提交
616

617 618 619 620 621 622 623 624 625 626 627 628 629 630 631 632 633 634 635 636 637 638 639 640 641 642
/**
 *	dev_get_by_name_rcu	- find a device by its name
 *	@net: the applicable net namespace
 *	@name: name to find
 *
 *	Find an interface by name.
 *	If the name is found a pointer to the device is returned.
 * 	If the name is not found then %NULL is returned.
 *	The reference counters are not incremented so the caller must be
 *	careful with locks. The caller must hold RCU lock.
 */

struct net_device *dev_get_by_name_rcu(struct net *net, const char *name)
{
	struct hlist_node *p;
	struct net_device *dev;
	struct hlist_head *head = dev_name_hash(net, name);

	hlist_for_each_entry_rcu(dev, p, head, name_hlist)
		if (!strncmp(dev->name, name, IFNAMSIZ))
			return dev;

	return NULL;
}
EXPORT_SYMBOL(dev_get_by_name_rcu);

L
Linus Torvalds 已提交
643 644
/**
 *	dev_get_by_name		- find a device by its name
645
 *	@net: the applicable net namespace
L
Linus Torvalds 已提交
646 647 648 649 650 651 652 653 654
 *	@name: name to find
 *
 *	Find an interface by name. This can be called from any
 *	context and does its own locking. The returned handle has
 *	the usage count incremented and the caller must use dev_put() to
 *	release it when it is no longer needed. %NULL is returned if no
 *	matching device is found.
 */

655
struct net_device *dev_get_by_name(struct net *net, const char *name)
L
Linus Torvalds 已提交
656 657 658
{
	struct net_device *dev;

659 660
	rcu_read_lock();
	dev = dev_get_by_name_rcu(net, name);
L
Linus Torvalds 已提交
661 662
	if (dev)
		dev_hold(dev);
663
	rcu_read_unlock();
L
Linus Torvalds 已提交
664 665
	return dev;
}
E
Eric Dumazet 已提交
666
EXPORT_SYMBOL(dev_get_by_name);
L
Linus Torvalds 已提交
667 668 669

/**
 *	__dev_get_by_index - find a device by its ifindex
670
 *	@net: the applicable net namespace
L
Linus Torvalds 已提交
671 672 673 674 675 676 677 678 679
 *	@ifindex: index of device
 *
 *	Search for an interface by index. Returns %NULL if the device
 *	is not found or a pointer to the device. The device has not
 *	had its reference counter increased so the caller must be careful
 *	about locking. The caller must hold either the RTNL semaphore
 *	or @dev_base_lock.
 */

680
struct net_device *__dev_get_by_index(struct net *net, int ifindex)
L
Linus Torvalds 已提交
681 682
{
	struct hlist_node *p;
E
Eric Dumazet 已提交
683 684
	struct net_device *dev;
	struct hlist_head *head = dev_index_hash(net, ifindex);
L
Linus Torvalds 已提交
685

E
Eric Dumazet 已提交
686
	hlist_for_each_entry(dev, p, head, index_hlist)
L
Linus Torvalds 已提交
687 688
		if (dev->ifindex == ifindex)
			return dev;
E
Eric Dumazet 已提交
689

L
Linus Torvalds 已提交
690 691
	return NULL;
}
E
Eric Dumazet 已提交
692
EXPORT_SYMBOL(__dev_get_by_index);
L
Linus Torvalds 已提交
693

694 695 696 697 698 699 700 701 702 703 704 705 706 707 708 709 710 711 712 713 714 715 716 717 718
/**
 *	dev_get_by_index_rcu - find a device by its ifindex
 *	@net: the applicable net namespace
 *	@ifindex: index of device
 *
 *	Search for an interface by index. Returns %NULL if the device
 *	is not found or a pointer to the device. The device has not
 *	had its reference counter increased so the caller must be careful
 *	about locking. The caller must hold RCU lock.
 */

struct net_device *dev_get_by_index_rcu(struct net *net, int ifindex)
{
	struct hlist_node *p;
	struct net_device *dev;
	struct hlist_head *head = dev_index_hash(net, ifindex);

	hlist_for_each_entry_rcu(dev, p, head, index_hlist)
		if (dev->ifindex == ifindex)
			return dev;

	return NULL;
}
EXPORT_SYMBOL(dev_get_by_index_rcu);

L
Linus Torvalds 已提交
719 720 721

/**
 *	dev_get_by_index - find a device by its ifindex
722
 *	@net: the applicable net namespace
L
Linus Torvalds 已提交
723 724 725 726 727 728 729 730
 *	@ifindex: index of device
 *
 *	Search for an interface by index. Returns NULL if the device
 *	is not found or a pointer to the device. The device returned has
 *	had a reference added and the pointer is safe until the user calls
 *	dev_put to indicate they have finished with it.
 */

731
struct net_device *dev_get_by_index(struct net *net, int ifindex)
L
Linus Torvalds 已提交
732 733 734
{
	struct net_device *dev;

735 736
	rcu_read_lock();
	dev = dev_get_by_index_rcu(net, ifindex);
L
Linus Torvalds 已提交
737 738
	if (dev)
		dev_hold(dev);
739
	rcu_read_unlock();
L
Linus Torvalds 已提交
740 741
	return dev;
}
E
Eric Dumazet 已提交
742
EXPORT_SYMBOL(dev_get_by_index);
L
Linus Torvalds 已提交
743 744 745

/**
 *	dev_getbyhwaddr - find a device by its hardware address
746
 *	@net: the applicable net namespace
L
Linus Torvalds 已提交
747 748 749 750 751 752 753 754 755 756 757 758
 *	@type: media type of device
 *	@ha: hardware address
 *
 *	Search for an interface by MAC address. Returns NULL if the device
 *	is not found or a pointer to the device. The caller must hold the
 *	rtnl semaphore. The returned device has not had its ref count increased
 *	and the caller must therefore be careful about locking
 *
 *	BUGS:
 *	If the API was consistent this would be __dev_get_by_hwaddr
 */

759
struct net_device *dev_getbyhwaddr(struct net *net, unsigned short type, char *ha)
L
Linus Torvalds 已提交
760 761 762 763 764
{
	struct net_device *dev;

	ASSERT_RTNL();

765
	for_each_netdev(net, dev)
L
Linus Torvalds 已提交
766 767
		if (dev->type == type &&
		    !memcmp(dev->dev_addr, ha, dev->addr_len))
768 769 770
			return dev;

	return NULL;
L
Linus Torvalds 已提交
771
}
772 773
EXPORT_SYMBOL(dev_getbyhwaddr);

774
struct net_device *__dev_getfirstbyhwtype(struct net *net, unsigned short type)
L
Linus Torvalds 已提交
775 776 777
{
	struct net_device *dev;

778
	ASSERT_RTNL();
779
	for_each_netdev(net, dev)
780
		if (dev->type == type)
781 782 783
			return dev;

	return NULL;
784 785 786
}
EXPORT_SYMBOL(__dev_getfirstbyhwtype);

787
struct net_device *dev_getfirstbyhwtype(struct net *net, unsigned short type)
788
{
789
	struct net_device *dev, *ret = NULL;
790

791 792 793 794 795 796 797 798 799
	rcu_read_lock();
	for_each_netdev_rcu(net, dev)
		if (dev->type == type) {
			dev_hold(dev);
			ret = dev;
			break;
		}
	rcu_read_unlock();
	return ret;
L
Linus Torvalds 已提交
800 801 802 803
}
EXPORT_SYMBOL(dev_getfirstbyhwtype);

/**
E
Eric Dumazet 已提交
804
 *	dev_get_by_flags_rcu - find any device with given flags
805
 *	@net: the applicable net namespace
L
Linus Torvalds 已提交
806 807 808 809
 *	@if_flags: IFF_* values
 *	@mask: bitmask of bits in if_flags to check
 *
 *	Search for any interface with the given flags. Returns NULL if a device
E
Eric Dumazet 已提交
810 811
 *	is not found or a pointer to the device. Must be called inside
 *	rcu_read_lock(), and result refcount is unchanged.
L
Linus Torvalds 已提交
812 813
 */

E
Eric Dumazet 已提交
814
struct net_device *dev_get_by_flags_rcu(struct net *net, unsigned short if_flags,
E
Eric Dumazet 已提交
815
				    unsigned short mask)
L
Linus Torvalds 已提交
816
{
817
	struct net_device *dev, *ret;
L
Linus Torvalds 已提交
818

819
	ret = NULL;
820
	for_each_netdev_rcu(net, dev) {
L
Linus Torvalds 已提交
821
		if (((dev->flags ^ if_flags) & mask) == 0) {
822
			ret = dev;
L
Linus Torvalds 已提交
823 824 825
			break;
		}
	}
826
	return ret;
L
Linus Torvalds 已提交
827
}
E
Eric Dumazet 已提交
828
EXPORT_SYMBOL(dev_get_by_flags_rcu);
L
Linus Torvalds 已提交
829 830 831 832 833 834

/**
 *	dev_valid_name - check if name is okay for network device
 *	@name: name string
 *
 *	Network device names need to be valid file names to
835 836
 *	to allow sysfs to work.  We also disallow any kind of
 *	whitespace.
L
Linus Torvalds 已提交
837
 */
838
int dev_valid_name(const char *name)
L
Linus Torvalds 已提交
839
{
840 841
	if (*name == '\0')
		return 0;
842 843
	if (strlen(name) >= IFNAMSIZ)
		return 0;
844 845 846 847 848 849 850 851 852
	if (!strcmp(name, ".") || !strcmp(name, ".."))
		return 0;

	while (*name) {
		if (*name == '/' || isspace(*name))
			return 0;
		name++;
	}
	return 1;
L
Linus Torvalds 已提交
853
}
E
Eric Dumazet 已提交
854
EXPORT_SYMBOL(dev_valid_name);
L
Linus Torvalds 已提交
855 856

/**
857 858
 *	__dev_alloc_name - allocate a name for a device
 *	@net: network namespace to allocate the device name in
L
Linus Torvalds 已提交
859
 *	@name: name format string
860
 *	@buf:  scratch buffer and result name string
L
Linus Torvalds 已提交
861 862
 *
 *	Passed a format string - eg "lt%d" it will try and find a suitable
S
Stephen Hemminger 已提交
863 864 865 866 867 868
 *	id. It scans list of devices to build up a free map, then chooses
 *	the first empty slot. The caller must hold the dev_base or rtnl lock
 *	while allocating the name and adding the device in order to avoid
 *	duplicates.
 *	Limited to bits_per_byte * page size devices (ie 32K on most platforms).
 *	Returns the number of the unit assigned or a negative errno code.
L
Linus Torvalds 已提交
869 870
 */

871
static int __dev_alloc_name(struct net *net, const char *name, char *buf)
L
Linus Torvalds 已提交
872 873 874 875
{
	int i = 0;
	const char *p;
	const int max_netdevices = 8*PAGE_SIZE;
S
Stephen Hemminger 已提交
876
	unsigned long *inuse;
L
Linus Torvalds 已提交
877 878 879 880 881 882 883 884 885 886 887 888 889
	struct net_device *d;

	p = strnchr(name, IFNAMSIZ-1, '%');
	if (p) {
		/*
		 * Verify the string as this thing may have come from
		 * the user.  There must be either one "%d" and no other "%"
		 * characters.
		 */
		if (p[1] != 'd' || strchr(p + 2, '%'))
			return -EINVAL;

		/* Use one page as a bit array of possible slots */
S
Stephen Hemminger 已提交
890
		inuse = (unsigned long *) get_zeroed_page(GFP_ATOMIC);
L
Linus Torvalds 已提交
891 892 893
		if (!inuse)
			return -ENOMEM;

894
		for_each_netdev(net, d) {
L
Linus Torvalds 已提交
895 896 897 898 899 900
			if (!sscanf(d->name, name, &i))
				continue;
			if (i < 0 || i >= max_netdevices)
				continue;

			/*  avoid cases where sscanf is not exact inverse of printf */
901
			snprintf(buf, IFNAMSIZ, name, i);
L
Linus Torvalds 已提交
902 903 904 905 906 907 908 909
			if (!strncmp(buf, d->name, IFNAMSIZ))
				set_bit(i, inuse);
		}

		i = find_first_zero_bit(inuse, max_netdevices);
		free_page((unsigned long) inuse);
	}

910 911
	if (buf != name)
		snprintf(buf, IFNAMSIZ, name, i);
912
	if (!__dev_get_by_name(net, buf))
L
Linus Torvalds 已提交
913 914 915 916 917 918 919 920 921
		return i;

	/* It is possible to run out of possible slots
	 * when the name is long and there isn't enough space left
	 * for the digits, or if all bits are used.
	 */
	return -ENFILE;
}

922 923 924 925 926 927 928 929 930 931 932 933 934 935 936 937 938 939 940 941
/**
 *	dev_alloc_name - allocate a name for a device
 *	@dev: device
 *	@name: name format string
 *
 *	Passed a format string - eg "lt%d" it will try and find a suitable
 *	id. It scans list of devices to build up a free map, then chooses
 *	the first empty slot. The caller must hold the dev_base or rtnl lock
 *	while allocating the name and adding the device in order to avoid
 *	duplicates.
 *	Limited to bits_per_byte * page size devices (ie 32K on most platforms).
 *	Returns the number of the unit assigned or a negative errno code.
 */

int dev_alloc_name(struct net_device *dev, const char *name)
{
	char buf[IFNAMSIZ];
	struct net *net;
	int ret;

942 943
	BUG_ON(!dev_net(dev));
	net = dev_net(dev);
944 945 946 947 948
	ret = __dev_alloc_name(net, name, buf);
	if (ret >= 0)
		strlcpy(dev->name, buf, IFNAMSIZ);
	return ret;
}
E
Eric Dumazet 已提交
949
EXPORT_SYMBOL(dev_alloc_name);
950

951
static int dev_get_valid_name(struct net_device *dev, const char *name, bool fmt)
952
{
953 954 955 956 957
	struct net *net;

	BUG_ON(!dev_net(dev));
	net = dev_net(dev);

958 959 960 961
	if (!dev_valid_name(name))
		return -EINVAL;

	if (fmt && strchr(name, '%'))
962
		return dev_alloc_name(dev, name);
963 964
	else if (__dev_get_by_name(net, name))
		return -EEXIST;
965 966
	else if (dev->name != name)
		strlcpy(dev->name, name, IFNAMSIZ);
967 968 969

	return 0;
}
L
Linus Torvalds 已提交
970 971 972 973 974 975 976 977 978

/**
 *	dev_change_name - change name of a device
 *	@dev: device
 *	@newname: name (or format string) must be at least IFNAMSIZ
 *
 *	Change name of a device, can pass format strings "eth%d".
 *	for wildcarding.
 */
979
int dev_change_name(struct net_device *dev, const char *newname)
L
Linus Torvalds 已提交
980
{
981
	char oldname[IFNAMSIZ];
L
Linus Torvalds 已提交
982
	int err = 0;
983
	int ret;
984
	struct net *net;
L
Linus Torvalds 已提交
985 986

	ASSERT_RTNL();
987
	BUG_ON(!dev_net(dev));
L
Linus Torvalds 已提交
988

989
	net = dev_net(dev);
L
Linus Torvalds 已提交
990 991 992
	if (dev->flags & IFF_UP)
		return -EBUSY;

993 994 995
	if (strncmp(newname, dev->name, IFNAMSIZ) == 0)
		return 0;

996 997
	memcpy(oldname, dev->name, IFNAMSIZ);

998
	err = dev_get_valid_name(dev, newname, 1);
999 1000
	if (err < 0)
		return err;
L
Linus Torvalds 已提交
1001

1002
rollback:
1003 1004 1005 1006
	ret = device_rename(&dev->dev, dev->name);
	if (ret) {
		memcpy(dev->name, oldname, IFNAMSIZ);
		return ret;
1007
	}
1008 1009

	write_lock_bh(&dev_base_lock);
1010
	hlist_del(&dev->name_hlist);
1011 1012 1013 1014 1015 1016
	write_unlock_bh(&dev_base_lock);

	synchronize_rcu();

	write_lock_bh(&dev_base_lock);
	hlist_add_head_rcu(&dev->name_hlist, dev_name_hash(net, dev->name));
1017 1018
	write_unlock_bh(&dev_base_lock);

1019
	ret = call_netdevice_notifiers(NETDEV_CHANGENAME, dev);
1020 1021 1022
	ret = notifier_to_errno(ret);

	if (ret) {
1023 1024
		/* err >= 0 after dev_alloc_name() or stores the first errno */
		if (err >= 0) {
1025 1026 1027
			err = ret;
			memcpy(dev->name, oldname, IFNAMSIZ);
			goto rollback;
1028 1029 1030 1031
		} else {
			printk(KERN_ERR
			       "%s: name change rollback failed: %d.\n",
			       dev->name, ret);
1032 1033
		}
	}
L
Linus Torvalds 已提交
1034 1035 1036 1037

	return err;
}

1038 1039 1040 1041
/**
 *	dev_set_alias - change ifalias of a device
 *	@dev: device
 *	@alias: name up to IFALIASZ
1042
 *	@len: limit of bytes to copy from info
1043 1044 1045 1046 1047 1048 1049 1050 1051 1052
 *
 *	Set ifalias for a device,
 */
int dev_set_alias(struct net_device *dev, const char *alias, size_t len)
{
	ASSERT_RTNL();

	if (len >= IFALIASZ)
		return -EINVAL;

1053 1054 1055 1056 1057 1058 1059 1060
	if (!len) {
		if (dev->ifalias) {
			kfree(dev->ifalias);
			dev->ifalias = NULL;
		}
		return 0;
	}

E
Eric Dumazet 已提交
1061
	dev->ifalias = krealloc(dev->ifalias, len + 1, GFP_KERNEL);
1062 1063 1064 1065 1066 1067 1068 1069
	if (!dev->ifalias)
		return -ENOMEM;

	strlcpy(dev->ifalias, alias, len+1);
	return len;
}


1070
/**
S
Stephen Hemminger 已提交
1071
 *	netdev_features_change - device changes features
1072 1073 1074 1075 1076 1077
 *	@dev: device to cause notification
 *
 *	Called to indicate a device has changed features.
 */
void netdev_features_change(struct net_device *dev)
{
1078
	call_netdevice_notifiers(NETDEV_FEAT_CHANGE, dev);
1079 1080 1081
}
EXPORT_SYMBOL(netdev_features_change);

L
Linus Torvalds 已提交
1082 1083 1084 1085 1086 1087 1088 1089 1090 1091 1092
/**
 *	netdev_state_change - device changes state
 *	@dev: device to cause notification
 *
 *	Called to indicate a device has changed state. This function calls
 *	the notifier chains for netdev_chain and sends a NEWLINK message
 *	to the routing socket.
 */
void netdev_state_change(struct net_device *dev)
{
	if (dev->flags & IFF_UP) {
1093
		call_netdevice_notifiers(NETDEV_CHANGE, dev);
L
Linus Torvalds 已提交
1094 1095 1096
		rtmsg_ifinfo(RTM_NEWLINK, dev, 0);
	}
}
E
Eric Dumazet 已提交
1097
EXPORT_SYMBOL(netdev_state_change);
L
Linus Torvalds 已提交
1098

1099
int netdev_bonding_change(struct net_device *dev, unsigned long event)
1100
{
1101
	return call_netdevice_notifiers(event, dev);
1102 1103 1104
}
EXPORT_SYMBOL(netdev_bonding_change);

L
Linus Torvalds 已提交
1105 1106
/**
 *	dev_load 	- load a network module
1107
 *	@net: the applicable net namespace
L
Linus Torvalds 已提交
1108 1109 1110 1111 1112 1113 1114
 *	@name: name of interface
 *
 *	If a network interface is not present and the process has suitable
 *	privileges this function loads the module. If module loading is not
 *	available in this kernel then it becomes a nop.
 */

1115
void dev_load(struct net *net, const char *name)
L
Linus Torvalds 已提交
1116
{
1117
	struct net_device *dev;
L
Linus Torvalds 已提交
1118

1119 1120 1121
	rcu_read_lock();
	dev = dev_get_by_name_rcu(net, name);
	rcu_read_unlock();
L
Linus Torvalds 已提交
1122

1123
	if (!dev && capable(CAP_NET_ADMIN))
L
Linus Torvalds 已提交
1124 1125
		request_module("%s", name);
}
E
Eric Dumazet 已提交
1126
EXPORT_SYMBOL(dev_load);
L
Linus Torvalds 已提交
1127

1128
static int __dev_open(struct net_device *dev)
L
Linus Torvalds 已提交
1129
{
1130
	const struct net_device_ops *ops = dev->netdev_ops;
1131
	int ret;
L
Linus Torvalds 已提交
1132

1133 1134
	ASSERT_RTNL();

L
Linus Torvalds 已提交
1135 1136 1137 1138 1139 1140
	/*
	 *	Is it even present?
	 */
	if (!netif_device_present(dev))
		return -ENODEV;

1141 1142 1143 1144 1145
	ret = call_netdevice_notifiers(NETDEV_PRE_UP, dev);
	ret = notifier_to_errno(ret);
	if (ret)
		return ret;

L
Linus Torvalds 已提交
1146 1147 1148 1149
	/*
	 *	Call device private open method
	 */
	set_bit(__LINK_STATE_START, &dev->state);
1150

1151 1152
	if (ops->ndo_validate_addr)
		ret = ops->ndo_validate_addr(dev);
1153

1154 1155
	if (!ret && ops->ndo_open)
		ret = ops->ndo_open(dev);
L
Linus Torvalds 已提交
1156

1157
	/*
L
Linus Torvalds 已提交
1158 1159 1160
	 *	If it went open OK then:
	 */

1161 1162 1163
	if (ret)
		clear_bit(__LINK_STATE_START, &dev->state);
	else {
L
Linus Torvalds 已提交
1164 1165 1166 1167 1168
		/*
		 *	Set the flags.
		 */
		dev->flags |= IFF_UP;

1169 1170 1171
		/*
		 *	Enable NET_DMA
		 */
1172
		net_dmaengine_get();
1173

L
Linus Torvalds 已提交
1174 1175 1176
		/*
		 *	Initialize multicasting status
		 */
1177
		dev_set_rx_mode(dev);
L
Linus Torvalds 已提交
1178 1179 1180 1181 1182 1183

		/*
		 *	Wakeup transmit queue engine
		 */
		dev_activate(dev);
	}
1184

L
Linus Torvalds 已提交
1185 1186 1187 1188
	return ret;
}

/**
1189 1190
 *	dev_open	- prepare an interface for use.
 *	@dev:	device to open
L
Linus Torvalds 已提交
1191
 *
1192 1193 1194 1195 1196 1197 1198
 *	Takes a device from down to up state. The device's private open
 *	function is invoked and then the multicast lists are loaded. Finally
 *	the device is moved into the up state and a %NETDEV_UP message is
 *	sent to the netdev notifier chain.
 *
 *	Calling this function on an active interface is a nop. On a failure
 *	a negative errno code is returned.
L
Linus Torvalds 已提交
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 1227
int dev_open(struct net_device *dev)
{
	int ret;

	/*
	 *	Is it already up?
	 */
	if (dev->flags & IFF_UP)
		return 0;

	/*
	 *	Open device
	 */
	ret = __dev_open(dev);
	if (ret < 0)
		return ret;

	/*
	 *	... and announce new interface.
	 */
	rtmsg_ifinfo(RTM_NEWLINK, dev, IFF_UP|IFF_RUNNING);
	call_netdevice_notifiers(NETDEV_UP, dev);

	return ret;
}
EXPORT_SYMBOL(dev_open);

static int __dev_close(struct net_device *dev)
L
Linus Torvalds 已提交
1228
{
1229
	const struct net_device_ops *ops = dev->netdev_ops;
1230

1231
	ASSERT_RTNL();
1232 1233
	might_sleep();

L
Linus Torvalds 已提交
1234 1235 1236 1237
	/*
	 *	Tell people we are going down, so that they can
	 *	prepare to death, when device is still operating.
	 */
1238
	call_netdevice_notifiers(NETDEV_GOING_DOWN, dev);
L
Linus Torvalds 已提交
1239 1240 1241 1242

	clear_bit(__LINK_STATE_START, &dev->state);

	/* Synchronize to scheduled poll. We cannot touch poll list,
1243 1244 1245 1246 1247
	 * it can be even on different cpu. So just clear netif_running().
	 *
	 * dev->stop() will invoke napi_disable() on all of it's
	 * napi_struct instances on this device.
	 */
L
Linus Torvalds 已提交
1248 1249
	smp_mb__after_clear_bit(); /* Commit netif_running(). */

1250 1251
	dev_deactivate(dev);

L
Linus Torvalds 已提交
1252 1253 1254 1255 1256 1257 1258
	/*
	 *	Call the device specific close. This cannot fail.
	 *	Only if device is UP
	 *
	 *	We allow it to be called even after a DETACH hot-plug
	 *	event.
	 */
1259 1260
	if (ops->ndo_stop)
		ops->ndo_stop(dev);
L
Linus Torvalds 已提交
1261 1262 1263 1264 1265 1266 1267 1268

	/*
	 *	Device is now down.
	 */

	dev->flags &= ~IFF_UP;

	/*
1269
	 *	Shutdown NET_DMA
L
Linus Torvalds 已提交
1270
	 */
1271 1272 1273 1274 1275 1276 1277 1278 1279 1280 1281 1282 1283 1284 1285 1286 1287 1288 1289 1290
	net_dmaengine_put();

	return 0;
}

/**
 *	dev_close - shutdown an interface.
 *	@dev: device to shutdown
 *
 *	This function moves an active device into down state. A
 *	%NETDEV_GOING_DOWN is sent to the netdev notifier chain. The device
 *	is then deactivated and finally a %NETDEV_DOWN is sent to the notifier
 *	chain.
 */
int dev_close(struct net_device *dev)
{
	if (!(dev->flags & IFF_UP))
		return 0;

	__dev_close(dev);
L
Linus Torvalds 已提交
1291

1292
	/*
1293
	 * Tell people we are down
1294
	 */
1295 1296
	rtmsg_ifinfo(RTM_NEWLINK, dev, IFF_UP|IFF_RUNNING);
	call_netdevice_notifiers(NETDEV_DOWN, dev);
1297

L
Linus Torvalds 已提交
1298 1299
	return 0;
}
E
Eric Dumazet 已提交
1300
EXPORT_SYMBOL(dev_close);
L
Linus Torvalds 已提交
1301 1302


1303 1304 1305 1306 1307 1308 1309 1310 1311 1312 1313 1314 1315 1316 1317 1318 1319 1320 1321 1322 1323 1324 1325
/**
 *	dev_disable_lro - disable Large Receive Offload on a device
 *	@dev: device
 *
 *	Disable Large Receive Offload (LRO) on a net device.  Must be
 *	called under RTNL.  This is needed if received packets may be
 *	forwarded to another interface.
 */
void dev_disable_lro(struct net_device *dev)
{
	if (dev->ethtool_ops && dev->ethtool_ops->get_flags &&
	    dev->ethtool_ops->set_flags) {
		u32 flags = dev->ethtool_ops->get_flags(dev);
		if (flags & ETH_FLAG_LRO) {
			flags &= ~ETH_FLAG_LRO;
			dev->ethtool_ops->set_flags(dev, flags);
		}
	}
	WARN_ON(dev->features & NETIF_F_LRO);
}
EXPORT_SYMBOL(dev_disable_lro);


1326 1327
static int dev_boot_phase = 1;

L
Linus Torvalds 已提交
1328 1329 1330 1331 1332 1333 1334 1335 1336 1337 1338 1339 1340 1341 1342
/*
 *	Device change register/unregister. These are not inline or static
 *	as we export them to the world.
 */

/**
 *	register_netdevice_notifier - register a network notifier block
 *	@nb: notifier
 *
 *	Register a notifier to be called when network device events occur.
 *	The notifier passed is linked into the kernel structures and must
 *	not be reused until it has been unregistered. A negative errno code
 *	is returned on a failure.
 *
 * 	When registered all registration and up events are replayed
1343
 *	to the new notifier to allow device to have a race free
L
Linus Torvalds 已提交
1344 1345 1346 1347 1348 1349
 *	view of the network device list.
 */

int register_netdevice_notifier(struct notifier_block *nb)
{
	struct net_device *dev;
1350
	struct net_device *last;
1351
	struct net *net;
L
Linus Torvalds 已提交
1352 1353 1354
	int err;

	rtnl_lock();
1355
	err = raw_notifier_chain_register(&netdev_chain, nb);
1356 1357
	if (err)
		goto unlock;
1358 1359 1360 1361 1362 1363 1364 1365 1366 1367 1368
	if (dev_boot_phase)
		goto unlock;
	for_each_net(net) {
		for_each_netdev(net, dev) {
			err = nb->notifier_call(nb, NETDEV_REGISTER, dev);
			err = notifier_to_errno(err);
			if (err)
				goto rollback;

			if (!(dev->flags & IFF_UP))
				continue;
L
Linus Torvalds 已提交
1369

1370 1371
			nb->notifier_call(nb, NETDEV_UP, dev);
		}
L
Linus Torvalds 已提交
1372
	}
1373 1374

unlock:
L
Linus Torvalds 已提交
1375 1376
	rtnl_unlock();
	return err;
1377 1378 1379

rollback:
	last = dev;
1380 1381 1382 1383
	for_each_net(net) {
		for_each_netdev(net, dev) {
			if (dev == last)
				break;
1384

1385 1386 1387 1388 1389
			if (dev->flags & IFF_UP) {
				nb->notifier_call(nb, NETDEV_GOING_DOWN, dev);
				nb->notifier_call(nb, NETDEV_DOWN, dev);
			}
			nb->notifier_call(nb, NETDEV_UNREGISTER, dev);
1390
			nb->notifier_call(nb, NETDEV_UNREGISTER_BATCH, dev);
1391 1392
		}
	}
1393 1394

	raw_notifier_chain_unregister(&netdev_chain, nb);
1395
	goto unlock;
L
Linus Torvalds 已提交
1396
}
E
Eric Dumazet 已提交
1397
EXPORT_SYMBOL(register_netdevice_notifier);
L
Linus Torvalds 已提交
1398 1399 1400 1401 1402 1403 1404 1405 1406 1407 1408 1409 1410

/**
 *	unregister_netdevice_notifier - unregister a network notifier block
 *	@nb: notifier
 *
 *	Unregister a notifier previously registered by
 *	register_netdevice_notifier(). The notifier is unlinked into the
 *	kernel structures and may then be reused. A negative errno code
 *	is returned on a failure.
 */

int unregister_netdevice_notifier(struct notifier_block *nb)
{
1411 1412 1413
	int err;

	rtnl_lock();
1414
	err = raw_notifier_chain_unregister(&netdev_chain, nb);
1415 1416
	rtnl_unlock();
	return err;
L
Linus Torvalds 已提交
1417
}
E
Eric Dumazet 已提交
1418
EXPORT_SYMBOL(unregister_netdevice_notifier);
L
Linus Torvalds 已提交
1419 1420 1421 1422

/**
 *	call_netdevice_notifiers - call all network notifier blocks
 *      @val: value passed unmodified to notifier function
1423
 *      @dev: net_device pointer passed unmodified to notifier function
L
Linus Torvalds 已提交
1424 1425
 *
 *	Call all network notifier blocks.  Parameters and return value
1426
 *	are as for raw_notifier_call_chain().
L
Linus Torvalds 已提交
1427 1428
 */

1429
int call_netdevice_notifiers(unsigned long val, struct net_device *dev)
L
Linus Torvalds 已提交
1430
{
1431
	ASSERT_RTNL();
1432
	return raw_notifier_call_chain(&netdev_chain, val, dev);
L
Linus Torvalds 已提交
1433 1434 1435 1436 1437 1438 1439 1440 1441
}

/* When > 0 there are consumers of rx skb time stamps */
static atomic_t netstamp_needed = ATOMIC_INIT(0);

void net_enable_timestamp(void)
{
	atomic_inc(&netstamp_needed);
}
E
Eric Dumazet 已提交
1442
EXPORT_SYMBOL(net_enable_timestamp);
L
Linus Torvalds 已提交
1443 1444 1445 1446 1447

void net_disable_timestamp(void)
{
	atomic_dec(&netstamp_needed);
}
E
Eric Dumazet 已提交
1448
EXPORT_SYMBOL(net_disable_timestamp);
L
Linus Torvalds 已提交
1449

E
Eric Dumazet 已提交
1450
static inline void net_timestamp_set(struct sk_buff *skb)
L
Linus Torvalds 已提交
1451 1452
{
	if (atomic_read(&netstamp_needed))
1453
		__net_timestamp(skb);
1454 1455
	else
		skb->tstamp.tv64 = 0;
L
Linus Torvalds 已提交
1456 1457
}

E
Eric Dumazet 已提交
1458 1459 1460 1461 1462 1463
static inline void net_timestamp_check(struct sk_buff *skb)
{
	if (!skb->tstamp.tv64 && atomic_read(&netstamp_needed))
		__net_timestamp(skb);
}

1464 1465 1466 1467 1468 1469 1470 1471
/**
 * dev_forward_skb - loopback an skb to another netif
 *
 * @dev: destination network device
 * @skb: buffer to forward
 *
 * return values:
 *	NET_RX_SUCCESS	(no congestion)
1472
 *	NET_RX_DROP     (packet was dropped, but freed)
1473 1474 1475 1476 1477 1478 1479 1480 1481 1482 1483 1484
 *
 * dev_forward_skb can be used for injecting an skb from the
 * start_xmit function of one device into the receive queue
 * of another device.
 *
 * The receiving device may be in another namespace, so
 * we have to clear all information in the skb that could
 * impact namespace isolation.
 */
int dev_forward_skb(struct net_device *dev, struct sk_buff *skb)
{
	skb_orphan(skb);
1485
	nf_reset(skb);
1486

1487 1488 1489
	if (!(dev->flags & IFF_UP) ||
	    (skb->len > (dev->mtu + dev->hard_header_len))) {
		kfree_skb(skb);
1490
		return NET_RX_DROP;
1491
	}
1492
	skb_set_dev(skb, dev);
1493 1494 1495 1496 1497 1498 1499
	skb->tstamp.tv64 = 0;
	skb->pkt_type = PACKET_HOST;
	skb->protocol = eth_type_trans(skb, dev);
	return netif_rx(skb);
}
EXPORT_SYMBOL_GPL(dev_forward_skb);

L
Linus Torvalds 已提交
1500 1501 1502 1503 1504
/*
 *	Support routine. Sends outgoing frames to any network
 *	taps currently in use.
 */

1505
static void dev_queue_xmit_nit(struct sk_buff *skb, struct net_device *dev)
L
Linus Torvalds 已提交
1506 1507
{
	struct packet_type *ptype;
1508

1509 1510
#ifdef CONFIG_NET_CLS_ACT
	if (!(skb->tstamp.tv64 && (G_TC_FROM(skb->tc_verd) & AT_INGRESS)))
E
Eric Dumazet 已提交
1511
		net_timestamp_set(skb);
1512
#else
E
Eric Dumazet 已提交
1513
	net_timestamp_set(skb);
1514
#endif
L
Linus Torvalds 已提交
1515 1516 1517 1518 1519 1520 1521 1522 1523

	rcu_read_lock();
	list_for_each_entry_rcu(ptype, &ptype_all, list) {
		/* Never send packets back to the socket
		 * they originated from - MvS (miquels@drinkel.ow.org)
		 */
		if ((ptype->dev == dev || !ptype->dev) &&
		    (ptype->af_packet_priv == NULL ||
		     (struct sock *)ptype->af_packet_priv != skb->sk)) {
E
Eric Dumazet 已提交
1524
			struct sk_buff *skb2 = skb_clone(skb, GFP_ATOMIC);
L
Linus Torvalds 已提交
1525 1526 1527 1528 1529 1530 1531
			if (!skb2)
				break;

			/* skb->nh should be correctly
			   set by sender, so that the second statement is
			   just protection against buggy protocols.
			 */
1532
			skb_reset_mac_header(skb2);
L
Linus Torvalds 已提交
1533

1534
			if (skb_network_header(skb2) < skb2->data ||
1535
			    skb2->network_header > skb2->tail) {
L
Linus Torvalds 已提交
1536 1537 1538
				if (net_ratelimit())
					printk(KERN_CRIT "protocol %04x is "
					       "buggy, dev %s\n",
1539 1540
					       ntohs(skb2->protocol),
					       dev->name);
1541
				skb_reset_network_header(skb2);
L
Linus Torvalds 已提交
1542 1543
			}

1544
			skb2->transport_header = skb2->network_header;
L
Linus Torvalds 已提交
1545
			skb2->pkt_type = PACKET_OUTGOING;
D
David S. Miller 已提交
1546
			ptype->func(skb2, skb->dev, ptype, skb->dev);
L
Linus Torvalds 已提交
1547 1548 1549 1550 1551
		}
	}
	rcu_read_unlock();
}

1552 1553 1554 1555 1556 1557 1558 1559 1560 1561 1562 1563 1564 1565 1566 1567 1568 1569
/*
 * Routine to help set real_num_tx_queues. To avoid skbs mapped to queues
 * greater then real_num_tx_queues stale skbs on the qdisc must be flushed.
 */
void netif_set_real_num_tx_queues(struct net_device *dev, unsigned int txq)
{
	unsigned int real_num = dev->real_num_tx_queues;

	if (unlikely(txq > dev->num_tx_queues))
		;
	else if (txq > real_num)
		dev->real_num_tx_queues = txq;
	else if (txq < real_num) {
		dev->real_num_tx_queues = txq;
		qdisc_reset_all_tx_gt(dev, txq);
	}
}
EXPORT_SYMBOL(netif_set_real_num_tx_queues);
1570

1571
static inline void __netif_reschedule(struct Qdisc *q)
1572
{
1573 1574
	struct softnet_data *sd;
	unsigned long flags;
1575

1576 1577
	local_irq_save(flags);
	sd = &__get_cpu_var(softnet_data);
1578 1579 1580
	q->next_sched = NULL;
	*sd->output_queue_tailp = q;
	sd->output_queue_tailp = &q->next_sched;
1581 1582 1583 1584 1585 1586 1587 1588
	raise_softirq_irqoff(NET_TX_SOFTIRQ);
	local_irq_restore(flags);
}

void __netif_schedule(struct Qdisc *q)
{
	if (!test_and_set_bit(__QDISC_STATE_SCHED, &q->state))
		__netif_reschedule(q);
1589 1590 1591
}
EXPORT_SYMBOL(__netif_schedule);

1592
void dev_kfree_skb_irq(struct sk_buff *skb)
1593
{
1594
	if (atomic_dec_and_test(&skb->users)) {
1595 1596
		struct softnet_data *sd;
		unsigned long flags;
1597

1598 1599 1600 1601 1602 1603 1604
		local_irq_save(flags);
		sd = &__get_cpu_var(softnet_data);
		skb->next = sd->completion_queue;
		sd->completion_queue = skb;
		raise_softirq_irqoff(NET_TX_SOFTIRQ);
		local_irq_restore(flags);
	}
1605
}
1606
EXPORT_SYMBOL(dev_kfree_skb_irq);
1607 1608 1609 1610 1611 1612 1613 1614 1615 1616 1617

void dev_kfree_skb_any(struct sk_buff *skb)
{
	if (in_irq() || irqs_disabled())
		dev_kfree_skb_irq(skb);
	else
		dev_kfree_skb(skb);
}
EXPORT_SYMBOL(dev_kfree_skb_any);


1618 1619 1620 1621 1622 1623
/**
 * netif_device_detach - mark device as removed
 * @dev: network device
 *
 * Mark device as removed from system and therefore no longer available.
 */
1624 1625 1626 1627
void netif_device_detach(struct net_device *dev)
{
	if (test_and_clear_bit(__LINK_STATE_PRESENT, &dev->state) &&
	    netif_running(dev)) {
1628
		netif_tx_stop_all_queues(dev);
1629 1630 1631 1632
	}
}
EXPORT_SYMBOL(netif_device_detach);

1633 1634 1635 1636 1637 1638
/**
 * netif_device_attach - mark device as attached
 * @dev: network device
 *
 * Mark device as attached from system and restart if needed.
 */
1639 1640 1641 1642
void netif_device_attach(struct net_device *dev)
{
	if (!test_and_set_bit(__LINK_STATE_PRESENT, &dev->state) &&
	    netif_running(dev)) {
1643
		netif_tx_wake_all_queues(dev);
1644
		__netdev_watchdog_up(dev);
1645 1646 1647 1648
	}
}
EXPORT_SYMBOL(netif_device_attach);

1649 1650 1651 1652 1653 1654
static bool can_checksum_protocol(unsigned long features, __be16 protocol)
{
	return ((features & NETIF_F_GEN_CSUM) ||
		((features & NETIF_F_IP_CSUM) &&
		 protocol == htons(ETH_P_IP)) ||
		((features & NETIF_F_IPV6_CSUM) &&
1655 1656 1657
		 protocol == htons(ETH_P_IPV6)) ||
		((features & NETIF_F_FCOE_CRC) &&
		 protocol == htons(ETH_P_FCOE)));
1658 1659 1660 1661 1662 1663 1664 1665 1666 1667 1668 1669 1670 1671 1672 1673
}

static bool dev_can_checksum(struct net_device *dev, struct sk_buff *skb)
{
	if (can_checksum_protocol(dev->features, skb->protocol))
		return true;

	if (skb->protocol == htons(ETH_P_8021Q)) {
		struct vlan_ethhdr *veh = (struct vlan_ethhdr *)skb->data;
		if (can_checksum_protocol(dev->features & dev->vlan_features,
					  veh->h_vlan_encapsulated_proto))
			return true;
	}

	return false;
}
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 1700 1701 1702 1703 1704
/**
 * skb_dev_set -- assign a new device to a buffer
 * @skb: buffer for the new device
 * @dev: network device
 *
 * If an skb is owned by a device already, we have to reset
 * all data private to the namespace a device belongs to
 * before assigning it a new device.
 */
#ifdef CONFIG_NET_NS
void skb_set_dev(struct sk_buff *skb, struct net_device *dev)
{
	skb_dst_drop(skb);
	if (skb->dev && !net_eq(dev_net(skb->dev), dev_net(dev))) {
		secpath_reset(skb);
		nf_reset(skb);
		skb_init_secmark(skb);
		skb->mark = 0;
		skb->priority = 0;
		skb->nf_trace = 0;
		skb->ipvs_property = 0;
#ifdef CONFIG_NET_SCHED
		skb->tc_index = 0;
#endif
	}
	skb->dev = dev;
}
EXPORT_SYMBOL(skb_set_dev);
#endif /* CONFIG_NET_NS */

L
Linus Torvalds 已提交
1705 1706 1707 1708
/*
 * Invalidate hardware checksum when packet is to be mangled, and
 * complete checksum manually on outgoing path.
 */
1709
int skb_checksum_help(struct sk_buff *skb)
L
Linus Torvalds 已提交
1710
{
1711
	__wsum csum;
1712
	int ret = 0, offset;
L
Linus Torvalds 已提交
1713

1714
	if (skb->ip_summed == CHECKSUM_COMPLETE)
1715 1716 1717 1718 1719
		goto out_set_summed;

	if (unlikely(skb_shinfo(skb)->gso_size)) {
		/* Let GSO fix up the checksum. */
		goto out_set_summed;
L
Linus Torvalds 已提交
1720 1721
	}

1722 1723 1724 1725 1726 1727 1728 1729 1730
	offset = skb->csum_start - skb_headroom(skb);
	BUG_ON(offset >= skb_headlen(skb));
	csum = skb_checksum(skb, offset, skb->len - offset, 0);

	offset += skb->csum_offset;
	BUG_ON(offset + sizeof(__sum16) > skb_headlen(skb));

	if (skb_cloned(skb) &&
	    !skb_clone_writable(skb, offset + sizeof(__sum16))) {
L
Linus Torvalds 已提交
1731 1732 1733 1734 1735
		ret = pskb_expand_head(skb, 0, 0, GFP_ATOMIC);
		if (ret)
			goto out;
	}

1736
	*(__sum16 *)(skb->data + offset) = csum_fold(csum);
1737
out_set_summed:
L
Linus Torvalds 已提交
1738
	skb->ip_summed = CHECKSUM_NONE;
1739
out:
L
Linus Torvalds 已提交
1740 1741
	return ret;
}
E
Eric Dumazet 已提交
1742
EXPORT_SYMBOL(skb_checksum_help);
L
Linus Torvalds 已提交
1743

1744 1745 1746
/**
 *	skb_gso_segment - Perform segmentation on skb.
 *	@skb: buffer to segment
1747
 *	@features: features for the output path (see dev->features)
1748 1749
 *
 *	This function segments the given skb and returns a list of segments.
1750 1751 1752
 *
 *	It may return NULL if the skb requires no segmentation.  This is
 *	only possible when GSO is used for verifying header integrity.
1753
 */
1754
struct sk_buff *skb_gso_segment(struct sk_buff *skb, int features)
1755 1756 1757
{
	struct sk_buff *segs = ERR_PTR(-EPROTONOSUPPORT);
	struct packet_type *ptype;
A
Al Viro 已提交
1758
	__be16 type = skb->protocol;
1759
	int err;
1760

1761
	skb_reset_mac_header(skb);
1762
	skb->mac_len = skb->network_header - skb->mac_header;
1763 1764
	__skb_pull(skb, skb->mac_len);

1765 1766 1767 1768 1769 1770 1771 1772 1773 1774 1775 1776 1777
	if (unlikely(skb->ip_summed != CHECKSUM_PARTIAL)) {
		struct net_device *dev = skb->dev;
		struct ethtool_drvinfo info = {};

		if (dev && dev->ethtool_ops && dev->ethtool_ops->get_drvinfo)
			dev->ethtool_ops->get_drvinfo(dev, &info);

		WARN(1, "%s: caps=(0x%lx, 0x%lx) len=%d data_len=%d "
			"ip_summed=%d",
		     info.driver, dev ? dev->features : 0L,
		     skb->sk ? skb->sk->sk_route_caps : 0L,
		     skb->len, skb->data_len, skb->ip_summed);

1778 1779 1780 1781 1782
		if (skb_header_cloned(skb) &&
		    (err = pskb_expand_head(skb, 0, 0, GFP_ATOMIC)))
			return ERR_PTR(err);
	}

1783
	rcu_read_lock();
1784 1785
	list_for_each_entry_rcu(ptype,
			&ptype_base[ntohs(type) & PTYPE_HASH_MASK], list) {
1786
		if (ptype->type == type && !ptype->dev && ptype->gso_segment) {
1787
			if (unlikely(skb->ip_summed != CHECKSUM_PARTIAL)) {
1788 1789 1790 1791
				err = ptype->gso_send_check(skb);
				segs = ERR_PTR(err);
				if (err || skb_gso_ok(skb, features))
					break;
1792 1793
				__skb_push(skb, (skb->data -
						 skb_network_header(skb)));
1794
			}
1795
			segs = ptype->gso_segment(skb, features);
1796 1797 1798 1799 1800
			break;
		}
	}
	rcu_read_unlock();

1801
	__skb_push(skb, skb->data - skb_mac_header(skb));
1802

1803 1804 1805 1806
	return segs;
}
EXPORT_SYMBOL(skb_gso_segment);

1807 1808 1809 1810 1811
/* Take action when hardware reception checksum errors are detected. */
#ifdef CONFIG_BUG
void netdev_rx_csum_fault(struct net_device *dev)
{
	if (net_ratelimit()) {
1812
		printk(KERN_ERR "%s: hw csum failure.\n",
1813
			dev ? dev->name : "<unknown>");
1814 1815 1816 1817 1818 1819
		dump_stack();
	}
}
EXPORT_SYMBOL(netdev_rx_csum_fault);
#endif

L
Linus Torvalds 已提交
1820 1821 1822 1823 1824
/* Actually, we should eliminate this check as soon as we know, that:
 * 1. IOMMU is present and allows to map all the memory.
 * 2. No high memory really exists on this machine.
 */

E
Eric Dumazet 已提交
1825
static int illegal_highdma(struct net_device *dev, struct sk_buff *skb)
L
Linus Torvalds 已提交
1826
{
1827
#ifdef CONFIG_HIGHMEM
L
Linus Torvalds 已提交
1828
	int i;
1829 1830 1831 1832 1833
	if (!(dev->features & NETIF_F_HIGHDMA)) {
		for (i = 0; i < skb_shinfo(skb)->nr_frags; i++)
			if (PageHighMem(skb_shinfo(skb)->frags[i].page))
				return 1;
	}
L
Linus Torvalds 已提交
1834

1835 1836
	if (PCI_DMA_BUS_IS_PHYS) {
		struct device *pdev = dev->dev.parent;
L
Linus Torvalds 已提交
1837

E
Eric Dumazet 已提交
1838 1839
		if (!pdev)
			return 0;
1840 1841 1842 1843 1844 1845
		for (i = 0; i < skb_shinfo(skb)->nr_frags; i++) {
			dma_addr_t addr = page_to_phys(skb_shinfo(skb)->frags[i].page);
			if (!pdev->dma_mask || addr + PAGE_SIZE - 1 > *pdev->dma_mask)
				return 1;
		}
	}
1846
#endif
L
Linus Torvalds 已提交
1847 1848 1849
	return 0;
}

1850 1851 1852 1853 1854 1855 1856 1857 1858 1859 1860 1861 1862 1863 1864 1865 1866 1867 1868 1869 1870 1871 1872 1873 1874 1875 1876 1877 1878 1879 1880 1881 1882 1883
struct dev_gso_cb {
	void (*destructor)(struct sk_buff *skb);
};

#define DEV_GSO_CB(skb) ((struct dev_gso_cb *)(skb)->cb)

static void dev_gso_skb_destructor(struct sk_buff *skb)
{
	struct dev_gso_cb *cb;

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

		skb->next = nskb->next;
		nskb->next = NULL;
		kfree_skb(nskb);
	} while (skb->next);

	cb = DEV_GSO_CB(skb);
	if (cb->destructor)
		cb->destructor(skb);
}

/**
 *	dev_gso_segment - Perform emulated hardware segmentation on skb.
 *	@skb: buffer to segment
 *
 *	This function segments the given skb and stores the list of segments
 *	in skb->next.
 */
static int dev_gso_segment(struct sk_buff *skb)
{
	struct net_device *dev = skb->dev;
	struct sk_buff *segs;
1884 1885 1886 1887 1888 1889 1890 1891
	int features = dev->features & ~(illegal_highdma(dev, skb) ?
					 NETIF_F_SG : 0);

	segs = skb_gso_segment(skb, features);

	/* Verifying header integrity only. */
	if (!segs)
		return 0;
1892

1893
	if (IS_ERR(segs))
1894 1895 1896 1897 1898 1899 1900 1901 1902
		return PTR_ERR(segs);

	skb->next = segs;
	DEV_GSO_CB(skb)->destructor = skb->destructor;
	skb->destructor = dev_gso_skb_destructor;

	return 0;
}

E
Eric Dumazet 已提交
1903 1904
/*
 * Try to orphan skb early, right before transmission by the device.
1905 1906
 * We cannot orphan skb if tx timestamp is requested or the sk-reference
 * is needed on driver level for other reasons, e.g. see net/can/raw.c
E
Eric Dumazet 已提交
1907 1908 1909
 */
static inline void skb_orphan_try(struct sk_buff *skb)
{
1910 1911
	struct sock *sk = skb->sk;

1912
	if (sk && !skb_shinfo(skb)->tx_flags) {
1913 1914 1915 1916 1917
		/* skb_tx_hash() wont be able to get sk.
		 * We copy sk_hash into skb->rxhash
		 */
		if (!skb->rxhash)
			skb->rxhash = sk->sk_hash;
E
Eric Dumazet 已提交
1918
		skb_orphan(skb);
1919
	}
E
Eric Dumazet 已提交
1920 1921
}

1922 1923 1924 1925 1926 1927 1928 1929 1930 1931 1932 1933 1934 1935 1936 1937
/*
 * Returns true if either:
 *	1. skb has frag_list and the device doesn't support FRAGLIST, or
 *	2. skb is fragmented and the device does not support SG, or if
 *	   at least one of fragments is in highmem and device does not
 *	   support DMA from it.
 */
static inline int skb_needs_linearize(struct sk_buff *skb,
				      struct net_device *dev)
{
	return skb_is_nonlinear(skb) &&
	       ((skb_has_frags(skb) && !(dev->features & NETIF_F_FRAGLIST)) ||
	        (skb_shinfo(skb)->nr_frags && (!(dev->features & NETIF_F_SG) ||
					      illegal_highdma(dev, skb))));
}

1938 1939
int dev_hard_start_xmit(struct sk_buff *skb, struct net_device *dev,
			struct netdev_queue *txq)
1940
{
1941
	const struct net_device_ops *ops = dev->netdev_ops;
1942
	int rc = NETDEV_TX_OK;
1943

1944
	if (likely(!skb->next)) {
S
Stephen Hemminger 已提交
1945
		if (!list_empty(&ptype_all))
1946 1947
			dev_queue_xmit_nit(skb, dev);

1948 1949 1950 1951
		/*
		 * If device doesnt need skb->dst, release it right now while
		 * its hot in this cpu cache
		 */
E
Eric Dumazet 已提交
1952 1953 1954
		if (dev->priv_flags & IFF_XMIT_DST_RELEASE)
			skb_dst_drop(skb);

E
Eric Dumazet 已提交
1955
		skb_orphan_try(skb);
1956 1957 1958 1959 1960 1961

		if (netif_needs_gso(dev, skb)) {
			if (unlikely(dev_gso_segment(skb)))
				goto out_kfree_skb;
			if (skb->next)
				goto gso;
1962 1963 1964 1965 1966 1967 1968 1969 1970 1971 1972 1973 1974 1975 1976 1977
		} else {
			if (skb_needs_linearize(skb, dev) &&
			    __skb_linearize(skb))
				goto out_kfree_skb;

			/* If packet is not checksummed and device does not
			 * support checksumming for this protocol, complete
			 * checksumming here.
			 */
			if (skb->ip_summed == CHECKSUM_PARTIAL) {
				skb_set_transport_header(skb, skb->csum_start -
					      skb_headroom(skb));
				if (!dev_can_checksum(dev, skb) &&
				     skb_checksum_help(skb))
					goto out_kfree_skb;
			}
1978 1979
		}

1980
		rc = ops->ndo_start_xmit(skb, dev);
1981
		if (rc == NETDEV_TX_OK)
E
Eric Dumazet 已提交
1982
			txq_trans_update(txq);
1983
		return rc;
1984 1985
	}

1986
gso:
1987 1988 1989 1990 1991
	do {
		struct sk_buff *nskb = skb->next;

		skb->next = nskb->next;
		nskb->next = NULL;
1992 1993 1994 1995 1996 1997 1998 1999

		/*
		 * If device doesnt need nskb->dst, release it right now while
		 * its hot in this cpu cache
		 */
		if (dev->priv_flags & IFF_XMIT_DST_RELEASE)
			skb_dst_drop(nskb);

2000
		rc = ops->ndo_start_xmit(nskb, dev);
2001
		if (unlikely(rc != NETDEV_TX_OK)) {
2002 2003
			if (rc & ~NETDEV_TX_MASK)
				goto out_kfree_gso_skb;
2004
			nskb->next = skb->next;
2005 2006 2007
			skb->next = nskb;
			return rc;
		}
E
Eric Dumazet 已提交
2008
		txq_trans_update(txq);
2009
		if (unlikely(netif_tx_queue_stopped(txq) && skb->next))
2010
			return NETDEV_TX_BUSY;
2011
	} while (skb->next);
2012

2013 2014 2015
out_kfree_gso_skb:
	if (likely(skb->next == NULL))
		skb->destructor = DEV_GSO_CB(skb)->destructor;
2016 2017
out_kfree_skb:
	kfree_skb(skb);
2018
	return rc;
2019 2020
}

T
Tom Herbert 已提交
2021
static u32 hashrnd __read_mostly;
D
David S. Miller 已提交
2022

2023
u16 skb_tx_hash(const struct net_device *dev, const struct sk_buff *skb)
2024
{
2025
	u32 hash;
D
David S. Miller 已提交
2026

2027 2028
	if (skb_rx_queue_recorded(skb)) {
		hash = skb_get_rx_queue(skb);
E
Eric Dumazet 已提交
2029
		while (unlikely(hash >= dev->real_num_tx_queues))
2030 2031 2032
			hash -= dev->real_num_tx_queues;
		return hash;
	}
2033 2034

	if (skb->sk && skb->sk->sk_hash)
2035
		hash = skb->sk->sk_hash;
2036
	else
2037
		hash = (__force u16) skb->protocol ^ skb->rxhash;
T
Tom Herbert 已提交
2038
	hash = jhash_1word(hash, hashrnd);
D
David S. Miller 已提交
2039 2040

	return (u16) (((u64) hash * dev->real_num_tx_queues) >> 32);
2041
}
2042
EXPORT_SYMBOL(skb_tx_hash);
2043

2044 2045 2046 2047
static inline u16 dev_cap_txqueue(struct net_device *dev, u16 queue_index)
{
	if (unlikely(queue_index >= dev->real_num_tx_queues)) {
		if (net_ratelimit()) {
E
Eric Dumazet 已提交
2048 2049 2050
			pr_warning("%s selects TX queue %d, but "
				"real number of TX queues is %d\n",
				dev->name, queue_index, dev->real_num_tx_queues);
2051 2052 2053 2054 2055 2056
		}
		return 0;
	}
	return queue_index;
}

2057 2058 2059
static struct netdev_queue *dev_pick_tx(struct net_device *dev,
					struct sk_buff *skb)
{
2060
	int queue_index;
2061 2062
	struct sock *sk = skb->sk;

2063 2064
	queue_index = sk_tx_queue_get(sk);
	if (queue_index < 0) {
2065 2066 2067 2068
		const struct net_device_ops *ops = dev->netdev_ops;

		if (ops->ndo_select_queue) {
			queue_index = ops->ndo_select_queue(dev, skb);
2069
			queue_index = dev_cap_txqueue(dev, queue_index);
2070 2071 2072 2073
		} else {
			queue_index = 0;
			if (dev->real_num_tx_queues > 1)
				queue_index = skb_tx_hash(dev, skb);
2074

E
Eric Dumazet 已提交
2075
			if (sk) {
2076
				struct dst_entry *dst = rcu_dereference_check(sk->sk_dst_cache, 1);
E
Eric Dumazet 已提交
2077 2078 2079 2080

				if (dst && skb_dst(skb) == dst)
					sk_tx_queue_set(sk, queue_index);
			}
2081 2082
		}
	}
2083

2084 2085
	skb_set_queue_mapping(skb, queue_index);
	return netdev_get_tx_queue(dev, queue_index);
2086 2087
}

2088 2089 2090 2091 2092
static inline int __dev_xmit_skb(struct sk_buff *skb, struct Qdisc *q,
				 struct net_device *dev,
				 struct netdev_queue *txq)
{
	spinlock_t *root_lock = qdisc_lock(q);
2093
	bool contended = qdisc_is_running(q);
2094 2095
	int rc;

2096 2097 2098 2099 2100 2101 2102 2103 2104
	/*
	 * Heuristic to force contended enqueues to serialize on a
	 * separate lock before trying to get qdisc main lock.
	 * This permits __QDISC_STATE_RUNNING owner to get the lock more often
	 * and dequeue packets faster.
	 */
	if (unlikely(contended))
		spin_lock(&q->busylock);

2105 2106 2107 2108 2109
	spin_lock(root_lock);
	if (unlikely(test_bit(__QDISC_STATE_DEACTIVATED, &q->state))) {
		kfree_skb(skb);
		rc = NET_XMIT_DROP;
	} else if ((q->flags & TCQ_F_CAN_BYPASS) && !qdisc_qlen(q) &&
2110
		   qdisc_run_begin(q)) {
2111 2112 2113 2114 2115
		/*
		 * This is a work-conserving queue; there are no old skbs
		 * waiting to be sent out; and the qdisc is not running -
		 * xmit the skb directly.
		 */
E
Eric Dumazet 已提交
2116 2117
		if (!(dev->priv_flags & IFF_XMIT_DST_RELEASE))
			skb_dst_force(skb);
2118
		__qdisc_update_bstats(q, skb->len);
2119 2120 2121 2122 2123
		if (sch_direct_xmit(skb, q, dev, txq, root_lock)) {
			if (unlikely(contended)) {
				spin_unlock(&q->busylock);
				contended = false;
			}
2124
			__qdisc_run(q);
2125
		} else
2126
			qdisc_run_end(q);
2127 2128 2129

		rc = NET_XMIT_SUCCESS;
	} else {
E
Eric Dumazet 已提交
2130
		skb_dst_force(skb);
2131
		rc = qdisc_enqueue_root(skb, q);
2132 2133 2134 2135 2136 2137 2138
		if (qdisc_run_begin(q)) {
			if (unlikely(contended)) {
				spin_unlock(&q->busylock);
				contended = false;
			}
			__qdisc_run(q);
		}
2139 2140
	}
	spin_unlock(root_lock);
2141 2142
	if (unlikely(contended))
		spin_unlock(&q->busylock);
2143 2144 2145
	return rc;
}

2146 2147 2148 2149 2150 2151 2152 2153 2154 2155 2156 2157 2158 2159 2160 2161 2162 2163 2164 2165 2166 2167 2168 2169 2170
/**
 *	dev_queue_xmit - transmit a buffer
 *	@skb: buffer to transmit
 *
 *	Queue a buffer for transmission to a network device. The caller must
 *	have set the device and priority and built the buffer before calling
 *	this function. The function can be called from an interrupt.
 *
 *	A negative errno code is returned on a failure. A success does not
 *	guarantee the frame will be transmitted as it may be dropped due
 *	to congestion or traffic shaping.
 *
 * -----------------------------------------------------------------------------------
 *      I notice this method can also return errors from the queue disciplines,
 *      including NET_XMIT_DROP, which is a positive value.  So, errors can also
 *      be positive.
 *
 *      Regardless of the return value, the skb is consumed, so it is currently
 *      difficult to retry a send to this method.  (You can bump the ref count
 *      before sending to hold a reference for retry if you are careful.)
 *
 *      When calling this method, interrupts MUST be enabled.  This is because
 *      the BH enable code must have IRQs enabled so that it will not deadlock.
 *          --BLG
 */
L
Linus Torvalds 已提交
2171 2172 2173
int dev_queue_xmit(struct sk_buff *skb)
{
	struct net_device *dev = skb->dev;
2174
	struct netdev_queue *txq;
L
Linus Torvalds 已提交
2175 2176 2177
	struct Qdisc *q;
	int rc = -ENOMEM;

2178 2179
	/* Disable soft irqs for various locks below. Also
	 * stops preemption for RCU.
L
Linus Torvalds 已提交
2180
	 */
2181
	rcu_read_lock_bh();
L
Linus Torvalds 已提交
2182

2183
	txq = dev_pick_tx(dev, skb);
2184
	q = rcu_dereference_bh(txq->qdisc);
2185

L
Linus Torvalds 已提交
2186
#ifdef CONFIG_NET_CLS_ACT
E
Eric Dumazet 已提交
2187
	skb->tc_verd = SET_TC_AT(skb->tc_verd, AT_EGRESS);
L
Linus Torvalds 已提交
2188 2189
#endif
	if (q->enqueue) {
2190
		rc = __dev_xmit_skb(skb, q, dev, txq);
2191
		goto out;
L
Linus Torvalds 已提交
2192 2193 2194 2195 2196
	}

	/* The device has no queue. Common case for software devices:
	   loopback, all the sorts of tunnels...

H
Herbert Xu 已提交
2197 2198
	   Really, it is unlikely that netif_tx_lock protection is necessary
	   here.  (f.e. loopback and IP tunnels are clean ignoring statistics
L
Linus Torvalds 已提交
2199 2200 2201 2202 2203 2204 2205 2206 2207 2208
	   counters.)
	   However, it is possible, that they rely on protection
	   made by us here.

	   Check this and shot the lock. It is not prone from deadlocks.
	   Either shot noqueue qdisc, it is even simpler 8)
	 */
	if (dev->flags & IFF_UP) {
		int cpu = smp_processor_id(); /* ok because BHs are off */

2209
		if (txq->xmit_lock_owner != cpu) {
L
Linus Torvalds 已提交
2210

2211
			HARD_TX_LOCK(dev, txq, cpu);
L
Linus Torvalds 已提交
2212

2213
			if (!netif_tx_queue_stopped(txq)) {
2214 2215
				rc = dev_hard_start_xmit(skb, dev, txq);
				if (dev_xmit_complete(rc)) {
2216
					HARD_TX_UNLOCK(dev, txq);
L
Linus Torvalds 已提交
2217 2218 2219
					goto out;
				}
			}
2220
			HARD_TX_UNLOCK(dev, txq);
L
Linus Torvalds 已提交
2221 2222 2223 2224 2225 2226 2227 2228 2229 2230 2231 2232 2233
			if (net_ratelimit())
				printk(KERN_CRIT "Virtual device %s asks to "
				       "queue packet!\n", dev->name);
		} else {
			/* Recursion is detected! It is possible,
			 * unfortunately */
			if (net_ratelimit())
				printk(KERN_CRIT "Dead loop on virtual device "
				       "%s, fix it urgently!\n", dev->name);
		}
	}

	rc = -ENETDOWN;
2234
	rcu_read_unlock_bh();
L
Linus Torvalds 已提交
2235 2236 2237 2238

	kfree_skb(skb);
	return rc;
out:
2239
	rcu_read_unlock_bh();
L
Linus Torvalds 已提交
2240 2241
	return rc;
}
E
Eric Dumazet 已提交
2242
EXPORT_SYMBOL(dev_queue_xmit);
L
Linus Torvalds 已提交
2243 2244 2245 2246 2247 2248


/*=======================================================================
			Receiver routines
  =======================================================================*/

2249
int netdev_max_backlog __read_mostly = 1000;
E
Eric Dumazet 已提交
2250
int netdev_tstamp_prequeue __read_mostly = 1;
2251 2252
int netdev_budget __read_mostly = 300;
int weight_p __read_mostly = 64;            /* old backlog weight */
L
Linus Torvalds 已提交
2253

E
Eric Dumazet 已提交
2254 2255 2256 2257 2258 2259 2260 2261
/* Called with irq disabled */
static inline void ____napi_schedule(struct softnet_data *sd,
				     struct napi_struct *napi)
{
	list_add_tail(&napi->poll_list, &sd->poll_list);
	__raise_softirq_irqoff(NET_RX_SOFTIRQ);
}

T
Tom Herbert 已提交
2262
/*
2263 2264 2265
 * __skb_get_rxhash: calculate a flow hash based on src/dst addresses
 * and src/dst port numbers. Returns a non-zero hash number on success
 * and 0 on failure.
T
Tom Herbert 已提交
2266
 */
2267
__u32 __skb_get_rxhash(struct sk_buff *skb)
T
Tom Herbert 已提交
2268
{
2269
	int nhoff, hash = 0;
T
Tom Herbert 已提交
2270 2271 2272
	struct ipv6hdr *ip6;
	struct iphdr *ip;
	u8 ip_proto;
C
Changli Gao 已提交
2273 2274 2275 2276 2277
	u32 addr1, addr2, ihl;
	union {
		u32 v32;
		u16 v16[2];
	} ports;
T
Tom Herbert 已提交
2278

2279
	nhoff = skb_network_offset(skb);
T
Tom Herbert 已提交
2280 2281 2282

	switch (skb->protocol) {
	case __constant_htons(ETH_P_IP):
2283
		if (!pskb_may_pull(skb, sizeof(*ip) + nhoff))
T
Tom Herbert 已提交
2284 2285
			goto done;

2286
		ip = (struct iphdr *) skb->data + nhoff;
T
Tom Herbert 已提交
2287
		ip_proto = ip->protocol;
E
Eric Dumazet 已提交
2288 2289
		addr1 = (__force u32) ip->saddr;
		addr2 = (__force u32) ip->daddr;
T
Tom Herbert 已提交
2290 2291 2292
		ihl = ip->ihl;
		break;
	case __constant_htons(ETH_P_IPV6):
2293
		if (!pskb_may_pull(skb, sizeof(*ip6) + nhoff))
T
Tom Herbert 已提交
2294 2295
			goto done;

2296
		ip6 = (struct ipv6hdr *) skb->data + nhoff;
T
Tom Herbert 已提交
2297
		ip_proto = ip6->nexthdr;
E
Eric Dumazet 已提交
2298 2299
		addr1 = (__force u32) ip6->saddr.s6_addr32[3];
		addr2 = (__force u32) ip6->daddr.s6_addr32[3];
T
Tom Herbert 已提交
2300 2301 2302 2303 2304
		ihl = (40 >> 2);
		break;
	default:
		goto done;
	}
2305

T
Tom Herbert 已提交
2306 2307 2308 2309 2310 2311 2312 2313
	switch (ip_proto) {
	case IPPROTO_TCP:
	case IPPROTO_UDP:
	case IPPROTO_DCCP:
	case IPPROTO_ESP:
	case IPPROTO_AH:
	case IPPROTO_SCTP:
	case IPPROTO_UDPLITE:
2314 2315 2316
		if (pskb_may_pull(skb, (ihl * 4) + 4 + nhoff)) {
			ports.v32 = * (__force u32 *) (skb->data + nhoff +
						       (ihl * 4));
C
Changli Gao 已提交
2317 2318 2319
			if (ports.v16[1] < ports.v16[0])
				swap(ports.v16[0], ports.v16[1]);
			break;
E
Eric Dumazet 已提交
2320
		}
T
Tom Herbert 已提交
2321
	default:
C
Changli Gao 已提交
2322
		ports.v32 = 0;
T
Tom Herbert 已提交
2323 2324 2325
		break;
	}

E
Eric Dumazet 已提交
2326 2327 2328
	/* get a consistent hash (same value on both flow directions) */
	if (addr2 < addr1)
		swap(addr1, addr2);
T
Tom Herbert 已提交
2329

2330 2331 2332 2333 2334 2335 2336 2337 2338 2339 2340 2341 2342 2343 2344 2345 2346 2347 2348 2349 2350 2351 2352 2353 2354 2355 2356 2357 2358 2359 2360 2361 2362 2363 2364 2365 2366 2367 2368 2369 2370 2371 2372 2373 2374 2375 2376 2377
	hash = jhash_3words(addr1, addr2, ports.v32, hashrnd);
	if (!hash)
		hash = 1;

done:
	return hash;
}
EXPORT_SYMBOL(__skb_get_rxhash);

#ifdef CONFIG_RPS

/* One global table that all flow-based protocols share. */
struct rps_sock_flow_table *rps_sock_flow_table __read_mostly;
EXPORT_SYMBOL(rps_sock_flow_table);

/*
 * get_rps_cpu is called from netif_receive_skb and returns the target
 * CPU from the RPS map of the receiving queue for a given skb.
 * rcu_read_lock must be held on entry.
 */
static int get_rps_cpu(struct net_device *dev, struct sk_buff *skb,
		       struct rps_dev_flow **rflowp)
{
	struct netdev_rx_queue *rxqueue;
	struct rps_map *map;
	struct rps_dev_flow_table *flow_table;
	struct rps_sock_flow_table *sock_flow_table;
	int cpu = -1;
	u16 tcpu;

	if (skb_rx_queue_recorded(skb)) {
		u16 index = skb_get_rx_queue(skb);
		if (unlikely(index >= dev->num_rx_queues)) {
			WARN_ONCE(dev->num_rx_queues > 1, "%s received packet "
				"on queue %u, but number of RX queues is %u\n",
				dev->name, index, dev->num_rx_queues);
			goto done;
		}
		rxqueue = dev->_rx + index;
	} else
		rxqueue = dev->_rx;

	if (!rxqueue->rps_map && !rxqueue->rps_flow_table)
		goto done;

	if (!skb_get_rxhash(skb))
		goto done;

T
Tom Herbert 已提交
2378 2379 2380 2381 2382 2383 2384 2385 2386 2387 2388 2389 2390 2391 2392 2393 2394 2395 2396 2397 2398 2399 2400 2401 2402 2403 2404 2405 2406 2407 2408 2409 2410 2411 2412 2413 2414 2415 2416
	flow_table = rcu_dereference(rxqueue->rps_flow_table);
	sock_flow_table = rcu_dereference(rps_sock_flow_table);
	if (flow_table && sock_flow_table) {
		u16 next_cpu;
		struct rps_dev_flow *rflow;

		rflow = &flow_table->flows[skb->rxhash & flow_table->mask];
		tcpu = rflow->cpu;

		next_cpu = sock_flow_table->ents[skb->rxhash &
		    sock_flow_table->mask];

		/*
		 * If the desired CPU (where last recvmsg was done) is
		 * different from current CPU (one in the rx-queue flow
		 * table entry), switch if one of the following holds:
		 *   - Current CPU is unset (equal to RPS_NO_CPU).
		 *   - Current CPU is offline.
		 *   - The current CPU's queue tail has advanced beyond the
		 *     last packet that was enqueued using this table entry.
		 *     This guarantees that all previous packets for the flow
		 *     have been dequeued, thus preserving in order delivery.
		 */
		if (unlikely(tcpu != next_cpu) &&
		    (tcpu == RPS_NO_CPU || !cpu_online(tcpu) ||
		     ((int)(per_cpu(softnet_data, tcpu).input_queue_head -
		      rflow->last_qtail)) >= 0)) {
			tcpu = rflow->cpu = next_cpu;
			if (tcpu != RPS_NO_CPU)
				rflow->last_qtail = per_cpu(softnet_data,
				    tcpu).input_queue_head;
		}
		if (tcpu != RPS_NO_CPU && cpu_online(tcpu)) {
			*rflowp = rflow;
			cpu = tcpu;
			goto done;
		}
	}

T
Tom Herbert 已提交
2417 2418
	map = rcu_dereference(rxqueue->rps_map);
	if (map) {
T
Tom Herbert 已提交
2419
		tcpu = map->cpus[((u64) skb->rxhash * map->len) >> 32];
T
Tom Herbert 已提交
2420 2421 2422 2423 2424 2425 2426 2427 2428 2429 2430 2431

		if (cpu_online(tcpu)) {
			cpu = tcpu;
			goto done;
		}
	}

done:
	return cpu;
}

/* Called from hardirq (IPI) context */
E
Eric Dumazet 已提交
2432
static void rps_trigger_softirq(void *data)
T
Tom Herbert 已提交
2433
{
E
Eric Dumazet 已提交
2434 2435
	struct softnet_data *sd = data;

E
Eric Dumazet 已提交
2436
	____napi_schedule(sd, &sd->backlog);
C
Changli Gao 已提交
2437
	sd->received_rps++;
T
Tom Herbert 已提交
2438
}
E
Eric Dumazet 已提交
2439

T
Tom Herbert 已提交
2440
#endif /* CONFIG_RPS */
T
Tom Herbert 已提交
2441

E
Eric Dumazet 已提交
2442 2443 2444 2445 2446 2447 2448 2449 2450 2451 2452 2453 2454 2455 2456 2457 2458 2459 2460 2461 2462
/*
 * Check if this softnet_data structure is another cpu one
 * If yes, queue it to our IPI list and return 1
 * If no, return 0
 */
static int rps_ipi_queued(struct softnet_data *sd)
{
#ifdef CONFIG_RPS
	struct softnet_data *mysd = &__get_cpu_var(softnet_data);

	if (sd != mysd) {
		sd->rps_ipi_next = mysd->rps_ipi_list;
		mysd->rps_ipi_list = sd;

		__raise_softirq_irqoff(NET_RX_SOFTIRQ);
		return 1;
	}
#endif /* CONFIG_RPS */
	return 0;
}

T
Tom Herbert 已提交
2463 2464 2465 2466
/*
 * enqueue_to_backlog is called to queue an skb to a per CPU backlog
 * queue (may be a remote CPU queue).
 */
T
Tom Herbert 已提交
2467 2468
static int enqueue_to_backlog(struct sk_buff *skb, int cpu,
			      unsigned int *qtail)
T
Tom Herbert 已提交
2469
{
E
Eric Dumazet 已提交
2470
	struct softnet_data *sd;
T
Tom Herbert 已提交
2471 2472
	unsigned long flags;

E
Eric Dumazet 已提交
2473
	sd = &per_cpu(softnet_data, cpu);
T
Tom Herbert 已提交
2474 2475 2476

	local_irq_save(flags);

E
Eric Dumazet 已提交
2477
	rps_lock(sd);
2478 2479
	if (skb_queue_len(&sd->input_pkt_queue) <= netdev_max_backlog) {
		if (skb_queue_len(&sd->input_pkt_queue)) {
T
Tom Herbert 已提交
2480
enqueue:
E
Eric Dumazet 已提交
2481
			__skb_queue_tail(&sd->input_pkt_queue, skb);
2482
			input_queue_tail_incr_save(sd, qtail);
E
Eric Dumazet 已提交
2483
			rps_unlock(sd);
2484
			local_irq_restore(flags);
T
Tom Herbert 已提交
2485 2486 2487
			return NET_RX_SUCCESS;
		}

2488 2489 2490 2491
		/* Schedule NAPI for backlog device
		 * We can use non atomic operation since we own the queue lock
		 */
		if (!__test_and_set_bit(NAPI_STATE_SCHED, &sd->backlog.state)) {
E
Eric Dumazet 已提交
2492
			if (!rps_ipi_queued(sd))
E
Eric Dumazet 已提交
2493
				____napi_schedule(sd, &sd->backlog);
T
Tom Herbert 已提交
2494 2495 2496 2497
		}
		goto enqueue;
	}

C
Changli Gao 已提交
2498
	sd->dropped++;
E
Eric Dumazet 已提交
2499
	rps_unlock(sd);
T
Tom Herbert 已提交
2500 2501 2502 2503 2504 2505

	local_irq_restore(flags);

	kfree_skb(skb);
	return NET_RX_DROP;
}
L
Linus Torvalds 已提交
2506 2507 2508 2509 2510 2511 2512 2513 2514 2515 2516 2517 2518 2519 2520 2521 2522 2523

/**
 *	netif_rx	-	post buffer to the network code
 *	@skb: buffer to post
 *
 *	This function receives a packet from a device driver and queues it for
 *	the upper (protocol) levels to process.  It always succeeds. The buffer
 *	may be dropped during processing for congestion control or by the
 *	protocol layers.
 *
 *	return values:
 *	NET_RX_SUCCESS	(no congestion)
 *	NET_RX_DROP     (packet was dropped)
 *
 */

int netif_rx(struct sk_buff *skb)
{
2524
	int ret;
L
Linus Torvalds 已提交
2525 2526 2527 2528 2529

	/* if netpoll wants it, pretend we never saw it */
	if (netpoll_rx(skb))
		return NET_RX_DROP;

E
Eric Dumazet 已提交
2530 2531
	if (netdev_tstamp_prequeue)
		net_timestamp_check(skb);
L
Linus Torvalds 已提交
2532

E
Eric Dumazet 已提交
2533
#ifdef CONFIG_RPS
2534
	{
T
Tom Herbert 已提交
2535
		struct rps_dev_flow voidflow, *rflow = &voidflow;
2536 2537
		int cpu;

2538
		preempt_disable();
2539
		rcu_read_lock();
T
Tom Herbert 已提交
2540 2541

		cpu = get_rps_cpu(skb->dev, skb, &rflow);
2542 2543
		if (cpu < 0)
			cpu = smp_processor_id();
T
Tom Herbert 已提交
2544 2545 2546

		ret = enqueue_to_backlog(skb, cpu, &rflow->last_qtail);

2547
		rcu_read_unlock();
2548
		preempt_enable();
2549
	}
2550
#else
T
Tom Herbert 已提交
2551 2552 2553 2554 2555
	{
		unsigned int qtail;
		ret = enqueue_to_backlog(skb, get_cpu(), &qtail);
		put_cpu();
	}
2556
#endif
2557
	return ret;
L
Linus Torvalds 已提交
2558
}
E
Eric Dumazet 已提交
2559
EXPORT_SYMBOL(netif_rx);
L
Linus Torvalds 已提交
2560 2561 2562 2563 2564 2565 2566 2567 2568 2569 2570 2571 2572 2573 2574 2575 2576 2577 2578 2579 2580 2581 2582 2583 2584 2585 2586 2587 2588 2589 2590

int netif_rx_ni(struct sk_buff *skb)
{
	int err;

	preempt_disable();
	err = netif_rx(skb);
	if (local_softirq_pending())
		do_softirq();
	preempt_enable();

	return err;
}
EXPORT_SYMBOL(netif_rx_ni);

static void net_tx_action(struct softirq_action *h)
{
	struct softnet_data *sd = &__get_cpu_var(softnet_data);

	if (sd->completion_queue) {
		struct sk_buff *clist;

		local_irq_disable();
		clist = sd->completion_queue;
		sd->completion_queue = NULL;
		local_irq_enable();

		while (clist) {
			struct sk_buff *skb = clist;
			clist = clist->next;

2591
			WARN_ON(atomic_read(&skb->users));
L
Linus Torvalds 已提交
2592 2593 2594 2595 2596
			__kfree_skb(skb);
		}
	}

	if (sd->output_queue) {
2597
		struct Qdisc *head;
L
Linus Torvalds 已提交
2598 2599 2600 2601

		local_irq_disable();
		head = sd->output_queue;
		sd->output_queue = NULL;
2602
		sd->output_queue_tailp = &sd->output_queue;
L
Linus Torvalds 已提交
2603 2604 2605
		local_irq_enable();

		while (head) {
2606 2607 2608
			struct Qdisc *q = head;
			spinlock_t *root_lock;

L
Linus Torvalds 已提交
2609 2610
			head = head->next_sched;

2611
			root_lock = qdisc_lock(q);
2612
			if (spin_trylock(root_lock)) {
2613 2614 2615
				smp_mb__before_clear_bit();
				clear_bit(__QDISC_STATE_SCHED,
					  &q->state);
2616 2617
				qdisc_run(q);
				spin_unlock(root_lock);
L
Linus Torvalds 已提交
2618
			} else {
2619
				if (!test_bit(__QDISC_STATE_DEACTIVATED,
2620
					      &q->state)) {
2621
					__netif_reschedule(q);
2622 2623 2624 2625 2626
				} else {
					smp_mb__before_clear_bit();
					clear_bit(__QDISC_STATE_SCHED,
						  &q->state);
				}
L
Linus Torvalds 已提交
2627 2628 2629 2630 2631
			}
		}
	}
}

2632 2633 2634
static inline int deliver_skb(struct sk_buff *skb,
			      struct packet_type *pt_prev,
			      struct net_device *orig_dev)
L
Linus Torvalds 已提交
2635 2636
{
	atomic_inc(&skb->users);
D
David S. Miller 已提交
2637
	return pt_prev->func(skb, skb->dev, pt_prev, orig_dev);
L
Linus Torvalds 已提交
2638 2639
}

2640 2641
#if (defined(CONFIG_BRIDGE) || defined(CONFIG_BRIDGE_MODULE)) && \
    (defined(CONFIG_ATM_LANE) || defined(CONFIG_ATM_LANE_MODULE))
2642 2643 2644
/* This hook is defined here for ATM LANE */
int (*br_fdb_test_addr_hook)(struct net_device *dev,
			     unsigned char *addr) __read_mostly;
2645
EXPORT_SYMBOL_GPL(br_fdb_test_addr_hook);
2646
#endif
L
Linus Torvalds 已提交
2647 2648 2649 2650 2651 2652

#ifdef CONFIG_NET_CLS_ACT
/* TODO: Maybe we should just force sch_ingress to be compiled in
 * when CONFIG_NET_CLS_ACT is? otherwise some useless instructions
 * a compare and 2 stores extra right now if we dont have it on
 * but have CONFIG_NET_CLS_ACT
2653
 * NOTE: This doesnt stop any functionality; if you dont have
L
Linus Torvalds 已提交
2654 2655 2656
 * the ingress scheduler, you just cant add policies on ingress.
 *
 */
2657
static int ing_filter(struct sk_buff *skb)
L
Linus Torvalds 已提交
2658 2659
{
	struct net_device *dev = skb->dev;
2660
	u32 ttl = G_TC_RTTL(skb->tc_verd);
2661 2662 2663
	struct netdev_queue *rxq;
	int result = TC_ACT_OK;
	struct Qdisc *q;
2664

2665 2666 2667 2668
	if (unlikely(MAX_RED_LOOP < ttl++)) {
		if (net_ratelimit())
			pr_warning( "Redir loop detected Dropping packet (%d->%d)\n",
			       skb->skb_iif, dev->ifindex);
2669 2670
		return TC_ACT_SHOT;
	}
L
Linus Torvalds 已提交
2671

2672 2673
	skb->tc_verd = SET_TC_RTTL(skb->tc_verd, ttl);
	skb->tc_verd = SET_TC_AT(skb->tc_verd, AT_INGRESS);
L
Linus Torvalds 已提交
2674

2675 2676
	rxq = &dev->rx_queue;

2677
	q = rxq->qdisc;
2678
	if (q != &noop_qdisc) {
2679
		spin_lock(qdisc_lock(q));
2680 2681
		if (likely(!test_bit(__QDISC_STATE_DEACTIVATED, &q->state)))
			result = qdisc_enqueue_root(skb, q);
2682 2683
		spin_unlock(qdisc_lock(q));
	}
2684 2685 2686

	return result;
}
2687

2688 2689 2690 2691
static inline struct sk_buff *handle_ing(struct sk_buff *skb,
					 struct packet_type **pt_prev,
					 int *ret, struct net_device *orig_dev)
{
2692
	if (skb->dev->rx_queue.qdisc == &noop_qdisc)
2693
		goto out;
L
Linus Torvalds 已提交
2694

2695 2696 2697
	if (*pt_prev) {
		*ret = deliver_skb(skb, *pt_prev, orig_dev);
		*pt_prev = NULL;
L
Linus Torvalds 已提交
2698 2699
	}

2700 2701 2702 2703 2704 2705 2706 2707 2708 2709
	switch (ing_filter(skb)) {
	case TC_ACT_SHOT:
	case TC_ACT_STOLEN:
		kfree_skb(skb);
		return NULL;
	}

out:
	skb->tc_verd = 0;
	return skb;
L
Linus Torvalds 已提交
2710 2711 2712
}
#endif

2713 2714 2715 2716 2717 2718 2719 2720 2721 2722 2723 2724 2725 2726 2727 2728 2729 2730 2731 2732 2733 2734 2735 2736 2737 2738 2739
/*
 * 	netif_nit_deliver - deliver received packets to network taps
 * 	@skb: buffer
 *
 * 	This function is used to deliver incoming packets to network
 * 	taps. It should be used when the normal netif_receive_skb path
 * 	is bypassed, for example because of VLAN acceleration.
 */
void netif_nit_deliver(struct sk_buff *skb)
{
	struct packet_type *ptype;

	if (list_empty(&ptype_all))
		return;

	skb_reset_network_header(skb);
	skb_reset_transport_header(skb);
	skb->mac_len = skb->network_header - skb->mac_header;

	rcu_read_lock();
	list_for_each_entry_rcu(ptype, &ptype_all, list) {
		if (!ptype->dev || ptype->dev == skb->dev)
			deliver_skb(skb, ptype, skb->dev);
	}
	rcu_read_unlock();
}

2740 2741 2742 2743
/**
 *	netdev_rx_handler_register - register receive handler
 *	@dev: device to register a handler for
 *	@rx_handler: receive handler to register
J
Jiri Pirko 已提交
2744
 *	@rx_handler_data: data pointer that is used by rx handler
2745 2746 2747 2748 2749 2750 2751 2752
 *
 *	Register a receive hander for a device. This handler will then be
 *	called from __netif_receive_skb. A negative errno code is returned
 *	on a failure.
 *
 *	The caller must hold the rtnl_mutex.
 */
int netdev_rx_handler_register(struct net_device *dev,
J
Jiri Pirko 已提交
2753 2754
			       rx_handler_func_t *rx_handler,
			       void *rx_handler_data)
2755 2756 2757 2758 2759 2760
{
	ASSERT_RTNL();

	if (dev->rx_handler)
		return -EBUSY;

J
Jiri Pirko 已提交
2761
	rcu_assign_pointer(dev->rx_handler_data, rx_handler_data);
2762 2763 2764 2765 2766 2767 2768 2769 2770 2771 2772 2773 2774 2775 2776 2777 2778 2779 2780
	rcu_assign_pointer(dev->rx_handler, rx_handler);

	return 0;
}
EXPORT_SYMBOL_GPL(netdev_rx_handler_register);

/**
 *	netdev_rx_handler_unregister - unregister receive handler
 *	@dev: device to unregister a handler from
 *
 *	Unregister a receive hander from a device.
 *
 *	The caller must hold the rtnl_mutex.
 */
void netdev_rx_handler_unregister(struct net_device *dev)
{

	ASSERT_RTNL();
	rcu_assign_pointer(dev->rx_handler, NULL);
J
Jiri Pirko 已提交
2781
	rcu_assign_pointer(dev->rx_handler_data, NULL);
2782 2783 2784
}
EXPORT_SYMBOL_GPL(netdev_rx_handler_unregister);

2785 2786 2787 2788 2789 2790 2791 2792 2793 2794 2795 2796 2797 2798 2799 2800 2801 2802 2803 2804 2805
static inline void skb_bond_set_mac_by_master(struct sk_buff *skb,
					      struct net_device *master)
{
	if (skb->pkt_type == PACKET_HOST) {
		u16 *dest = (u16 *) eth_hdr(skb)->h_dest;

		memcpy(dest, master->dev_addr, ETH_ALEN);
	}
}

/* On bonding slaves other than the currently active slave, suppress
 * duplicates except for 802.3ad ETH_P_SLOW, alb non-mcast/bcast, and
 * ARP on active-backup slaves with arp_validate enabled.
 */
int __skb_bond_should_drop(struct sk_buff *skb, struct net_device *master)
{
	struct net_device *dev = skb->dev;

	if (master->priv_flags & IFF_MASTER_ARPMON)
		dev->last_rx = jiffies;

2806 2807
	if ((master->priv_flags & IFF_MASTER_ALB) &&
	    (master->priv_flags & IFF_BRIDGE_PORT)) {
2808 2809 2810 2811 2812 2813 2814 2815 2816 2817 2818 2819 2820 2821 2822 2823 2824 2825 2826 2827 2828 2829 2830 2831 2832 2833 2834
		/* Do address unmangle. The local destination address
		 * will be always the one master has. Provides the right
		 * functionality in a bridge.
		 */
		skb_bond_set_mac_by_master(skb, master);
	}

	if (dev->priv_flags & IFF_SLAVE_INACTIVE) {
		if ((dev->priv_flags & IFF_SLAVE_NEEDARP) &&
		    skb->protocol == __cpu_to_be16(ETH_P_ARP))
			return 0;

		if (master->priv_flags & IFF_MASTER_ALB) {
			if (skb->pkt_type != PACKET_BROADCAST &&
			    skb->pkt_type != PACKET_MULTICAST)
				return 0;
		}
		if (master->priv_flags & IFF_MASTER_8023AD &&
		    skb->protocol == __cpu_to_be16(ETH_P_SLOW))
			return 0;

		return 1;
	}
	return 0;
}
EXPORT_SYMBOL(__skb_bond_should_drop);

2835
static int __netif_receive_skb(struct sk_buff *skb)
L
Linus Torvalds 已提交
2836 2837
{
	struct packet_type *ptype, *pt_prev;
2838
	rx_handler_func_t *rx_handler;
D
David S. Miller 已提交
2839
	struct net_device *orig_dev;
2840
	struct net_device *master;
2841
	struct net_device *null_or_orig;
2842
	struct net_device *orig_or_bond;
L
Linus Torvalds 已提交
2843
	int ret = NET_RX_DROP;
A
Al Viro 已提交
2844
	__be16 type;
L
Linus Torvalds 已提交
2845

E
Eric Dumazet 已提交
2846 2847
	if (!netdev_tstamp_prequeue)
		net_timestamp_check(skb);
2848

E
Eric Dumazet 已提交
2849
	if (vlan_tx_tag_present(skb) && vlan_hwaccel_do_receive(skb))
2850 2851
		return NET_RX_SUCCESS;

L
Linus Torvalds 已提交
2852
	/* if we've gotten here through NAPI, check netpoll */
2853
	if (netpoll_receive_skb(skb))
L
Linus Torvalds 已提交
2854 2855
		return NET_RX_DROP;

2856 2857
	if (!skb->skb_iif)
		skb->skb_iif = skb->dev->ifindex;
2858

2859 2860 2861 2862 2863 2864 2865 2866
	/*
	 * bonding note: skbs received on inactive slaves should only
	 * be delivered to pkt handlers that are exact matches.  Also
	 * the deliver_no_wcard flag will be set.  If packet handlers
	 * are sensitive to duplicate packets these skbs will need to
	 * be dropped at the handler.  The vlan accel path may have
	 * already set the deliver_no_wcard flag.
	 */
2867
	null_or_orig = NULL;
J
Joe Eykholt 已提交
2868
	orig_dev = skb->dev;
2869
	master = ACCESS_ONCE(orig_dev->master);
2870 2871 2872 2873 2874
	if (skb->deliver_no_wcard)
		null_or_orig = orig_dev;
	else if (master) {
		if (skb_bond_should_drop(skb, master)) {
			skb->deliver_no_wcard = 1;
2875
			null_or_orig = orig_dev; /* deliver only exact match */
2876
		} else
2877
			skb->dev = master;
J
Joe Eykholt 已提交
2878
	}
2879

2880
	__this_cpu_inc(softnet_data.processed);
2881
	skb_reset_network_header(skb);
2882
	skb_reset_transport_header(skb);
2883
	skb->mac_len = skb->network_header - skb->mac_header;
L
Linus Torvalds 已提交
2884 2885 2886 2887 2888 2889 2890 2891 2892 2893 2894 2895 2896

	pt_prev = NULL;

	rcu_read_lock();

#ifdef CONFIG_NET_CLS_ACT
	if (skb->tc_verd & TC_NCLS) {
		skb->tc_verd = CLR_TC_NCLS(skb->tc_verd);
		goto ncls;
	}
#endif

	list_for_each_entry_rcu(ptype, &ptype_all, list) {
2897 2898
		if (ptype->dev == null_or_orig || ptype->dev == skb->dev ||
		    ptype->dev == orig_dev) {
2899
			if (pt_prev)
D
David S. Miller 已提交
2900
				ret = deliver_skb(skb, pt_prev, orig_dev);
L
Linus Torvalds 已提交
2901 2902 2903 2904 2905
			pt_prev = ptype;
		}
	}

#ifdef CONFIG_NET_CLS_ACT
2906 2907
	skb = handle_ing(skb, &pt_prev, &ret, orig_dev);
	if (!skb)
L
Linus Torvalds 已提交
2908 2909 2910 2911
		goto out;
ncls:
#endif

2912 2913 2914 2915 2916 2917 2918 2919 2920 2921 2922
	/* Handle special case of bridge or macvlan */
	rx_handler = rcu_dereference(skb->dev->rx_handler);
	if (rx_handler) {
		if (pt_prev) {
			ret = deliver_skb(skb, pt_prev, orig_dev);
			pt_prev = NULL;
		}
		skb = rx_handler(skb);
		if (!skb)
			goto out;
	}
L
Linus Torvalds 已提交
2923

2924 2925 2926 2927 2928 2929
	/*
	 * Make sure frames received on VLAN interfaces stacked on
	 * bonding interfaces still make their way to any base bonding
	 * device that may have registered for a specific ptype.  The
	 * handler may have to adjust skb->dev and orig_dev.
	 */
2930
	orig_or_bond = orig_dev;
2931 2932
	if ((skb->dev->priv_flags & IFF_802_1Q_VLAN) &&
	    (vlan_dev_real_dev(skb->dev)->priv_flags & IFF_BONDING)) {
2933
		orig_or_bond = vlan_dev_real_dev(skb->dev);
2934 2935
	}

L
Linus Torvalds 已提交
2936
	type = skb->protocol;
2937 2938
	list_for_each_entry_rcu(ptype,
			&ptype_base[ntohs(type) & PTYPE_HASH_MASK], list) {
2939
		if (ptype->type == type && (ptype->dev == null_or_orig ||
2940
		     ptype->dev == skb->dev || ptype->dev == orig_dev ||
2941
		     ptype->dev == orig_or_bond)) {
2942
			if (pt_prev)
D
David S. Miller 已提交
2943
				ret = deliver_skb(skb, pt_prev, orig_dev);
L
Linus Torvalds 已提交
2944 2945 2946 2947 2948
			pt_prev = ptype;
		}
	}

	if (pt_prev) {
D
David S. Miller 已提交
2949
		ret = pt_prev->func(skb, skb->dev, pt_prev, orig_dev);
L
Linus Torvalds 已提交
2950 2951 2952 2953 2954 2955 2956 2957 2958 2959 2960 2961
	} else {
		kfree_skb(skb);
		/* Jamal, now you will not able to escape explaining
		 * me how you were going to use this. :-)
		 */
		ret = NET_RX_DROP;
	}

out:
	rcu_read_unlock();
	return ret;
}
T
Tom Herbert 已提交
2962 2963 2964 2965 2966 2967 2968 2969 2970 2971 2972 2973 2974 2975 2976 2977 2978 2979

/**
 *	netif_receive_skb - process receive buffer from network
 *	@skb: buffer to process
 *
 *	netif_receive_skb() is the main receive data processing function.
 *	It always succeeds. The buffer may be dropped during processing
 *	for congestion control or by the protocol layers.
 *
 *	This function may only be called from softirq context and interrupts
 *	should be enabled.
 *
 *	Return values (usually ignored):
 *	NET_RX_SUCCESS: no congestion
 *	NET_RX_DROP: packet was dropped
 */
int netif_receive_skb(struct sk_buff *skb)
{
E
Eric Dumazet 已提交
2980 2981 2982
	if (netdev_tstamp_prequeue)
		net_timestamp_check(skb);

2983 2984 2985
	if (skb_defer_rx_timestamp(skb))
		return NET_RX_SUCCESS;

E
Eric Dumazet 已提交
2986
#ifdef CONFIG_RPS
E
Eric Dumazet 已提交
2987 2988 2989
	{
		struct rps_dev_flow voidflow, *rflow = &voidflow;
		int cpu, ret;
T
Tom Herbert 已提交
2990

E
Eric Dumazet 已提交
2991 2992 2993
		rcu_read_lock();

		cpu = get_rps_cpu(skb->dev, skb, &rflow);
T
Tom Herbert 已提交
2994

E
Eric Dumazet 已提交
2995 2996 2997 2998 2999 3000 3001
		if (cpu >= 0) {
			ret = enqueue_to_backlog(skb, cpu, &rflow->last_qtail);
			rcu_read_unlock();
		} else {
			rcu_read_unlock();
			ret = __netif_receive_skb(skb);
		}
T
Tom Herbert 已提交
3002

E
Eric Dumazet 已提交
3003
		return ret;
T
Tom Herbert 已提交
3004
	}
3005 3006 3007
#else
	return __netif_receive_skb(skb);
#endif
T
Tom Herbert 已提交
3008
}
E
Eric Dumazet 已提交
3009
EXPORT_SYMBOL(netif_receive_skb);
L
Linus Torvalds 已提交
3010

E
Eric Dumazet 已提交
3011 3012 3013
/* Network device is going away, flush any packets still pending
 * Called with irqs disabled.
 */
3014
static void flush_backlog(void *arg)
3015
{
3016
	struct net_device *dev = arg;
E
Eric Dumazet 已提交
3017
	struct softnet_data *sd = &__get_cpu_var(softnet_data);
3018 3019
	struct sk_buff *skb, *tmp;

E
Eric Dumazet 已提交
3020
	rps_lock(sd);
3021
	skb_queue_walk_safe(&sd->input_pkt_queue, skb, tmp) {
3022
		if (skb->dev == dev) {
E
Eric Dumazet 已提交
3023
			__skb_unlink(skb, &sd->input_pkt_queue);
3024
			kfree_skb(skb);
3025
			input_queue_head_incr(sd);
3026
		}
3027
	}
E
Eric Dumazet 已提交
3028
	rps_unlock(sd);
3029 3030 3031 3032 3033

	skb_queue_walk_safe(&sd->process_queue, skb, tmp) {
		if (skb->dev == dev) {
			__skb_unlink(skb, &sd->process_queue);
			kfree_skb(skb);
3034
			input_queue_head_incr(sd);
3035 3036
		}
	}
3037 3038
}

3039 3040 3041 3042 3043 3044 3045
static int napi_gro_complete(struct sk_buff *skb)
{
	struct packet_type *ptype;
	__be16 type = skb->protocol;
	struct list_head *head = &ptype_base[ntohs(type) & PTYPE_HASH_MASK];
	int err = -ENOENT;

3046 3047
	if (NAPI_GRO_CB(skb)->count == 1) {
		skb_shinfo(skb)->gso_size = 0;
3048
		goto out;
3049
	}
3050 3051 3052 3053 3054 3055 3056 3057 3058 3059 3060 3061 3062 3063 3064 3065 3066 3067 3068 3069 3070

	rcu_read_lock();
	list_for_each_entry_rcu(ptype, head, list) {
		if (ptype->type != type || ptype->dev || !ptype->gro_complete)
			continue;

		err = ptype->gro_complete(skb);
		break;
	}
	rcu_read_unlock();

	if (err) {
		WARN_ON(&ptype->list == head);
		kfree_skb(skb);
		return NET_RX_SUCCESS;
	}

out:
	return netif_receive_skb(skb);
}

3071
static void napi_gro_flush(struct napi_struct *napi)
3072 3073 3074 3075 3076 3077 3078 3079 3080
{
	struct sk_buff *skb, *next;

	for (skb = napi->gro_list; skb; skb = next) {
		next = skb->next;
		skb->next = NULL;
		napi_gro_complete(skb);
	}

3081
	napi->gro_count = 0;
3082 3083 3084
	napi->gro_list = NULL;
}

3085
enum gro_result dev_gro_receive(struct napi_struct *napi, struct sk_buff *skb)
3086 3087 3088 3089 3090
{
	struct sk_buff **pp = NULL;
	struct packet_type *ptype;
	__be16 type = skb->protocol;
	struct list_head *head = &ptype_base[ntohs(type) & PTYPE_HASH_MASK];
H
Herbert Xu 已提交
3091
	int same_flow;
3092
	int mac_len;
3093
	enum gro_result ret;
3094

3095
	if (!(skb->dev->features & NETIF_F_GRO) || netpoll_rx_on(skb))
3096 3097
		goto normal;

3098
	if (skb_is_gso(skb) || skb_has_frags(skb))
3099 3100
		goto normal;

3101 3102 3103 3104 3105
	rcu_read_lock();
	list_for_each_entry_rcu(ptype, head, list) {
		if (ptype->type != type || ptype->dev || !ptype->gro_receive)
			continue;

3106
		skb_set_network_header(skb, skb_gro_offset(skb));
3107 3108 3109 3110
		mac_len = skb->network_header - skb->mac_header;
		skb->mac_len = mac_len;
		NAPI_GRO_CB(skb)->same_flow = 0;
		NAPI_GRO_CB(skb)->flush = 0;
H
Herbert Xu 已提交
3111
		NAPI_GRO_CB(skb)->free = 0;
3112 3113 3114 3115 3116 3117 3118 3119 3120

		pp = ptype->gro_receive(&napi->gro_list, skb);
		break;
	}
	rcu_read_unlock();

	if (&ptype->list == head)
		goto normal;

H
Herbert Xu 已提交
3121
	same_flow = NAPI_GRO_CB(skb)->same_flow;
3122
	ret = NAPI_GRO_CB(skb)->free ? GRO_MERGED_FREE : GRO_MERGED;
H
Herbert Xu 已提交
3123

3124 3125 3126 3127 3128 3129
	if (pp) {
		struct sk_buff *nskb = *pp;

		*pp = nskb->next;
		nskb->next = NULL;
		napi_gro_complete(nskb);
3130
		napi->gro_count--;
3131 3132
	}

H
Herbert Xu 已提交
3133
	if (same_flow)
3134 3135
		goto ok;

3136
	if (NAPI_GRO_CB(skb)->flush || napi->gro_count >= MAX_GRO_SKBS)
3137 3138
		goto normal;

3139
	napi->gro_count++;
3140
	NAPI_GRO_CB(skb)->count = 1;
3141
	skb_shinfo(skb)->gso_size = skb_gro_len(skb);
3142 3143
	skb->next = napi->gro_list;
	napi->gro_list = skb;
3144
	ret = GRO_HELD;
3145

3146
pull:
H
Herbert Xu 已提交
3147 3148 3149 3150 3151 3152 3153 3154 3155 3156 3157 3158 3159 3160 3161 3162 3163 3164 3165
	if (skb_headlen(skb) < skb_gro_offset(skb)) {
		int grow = skb_gro_offset(skb) - skb_headlen(skb);

		BUG_ON(skb->end - skb->tail < grow);

		memcpy(skb_tail_pointer(skb), NAPI_GRO_CB(skb)->frag0, grow);

		skb->tail += grow;
		skb->data_len -= grow;

		skb_shinfo(skb)->frags[0].page_offset += grow;
		skb_shinfo(skb)->frags[0].size -= grow;

		if (unlikely(!skb_shinfo(skb)->frags[0].size)) {
			put_page(skb_shinfo(skb)->frags[0].page);
			memmove(skb_shinfo(skb)->frags,
				skb_shinfo(skb)->frags + 1,
				--skb_shinfo(skb)->nr_frags);
		}
3166 3167
	}

3168
ok:
3169
	return ret;
3170 3171

normal:
3172 3173
	ret = GRO_NORMAL;
	goto pull;
H
Herbert Xu 已提交
3174
}
3175 3176
EXPORT_SYMBOL(dev_gro_receive);

3177 3178
static gro_result_t
__napi_gro_receive(struct napi_struct *napi, struct sk_buff *skb)
3179 3180 3181 3182
{
	struct sk_buff *p;

	for (p = napi->gro_list; p; p = p->next) {
3183 3184 3185 3186
		NAPI_GRO_CB(p)->same_flow =
			(p->dev == skb->dev) &&
			!compare_ether_header(skb_mac_header(p),
					      skb_gro_mac_header(skb));
3187 3188 3189 3190 3191
		NAPI_GRO_CB(p)->flush = 0;
	}

	return dev_gro_receive(napi, skb);
}
H
Herbert Xu 已提交
3192

3193
gro_result_t napi_skb_finish(gro_result_t ret, struct sk_buff *skb)
H
Herbert Xu 已提交
3194
{
3195 3196
	switch (ret) {
	case GRO_NORMAL:
3197 3198 3199
		if (netif_receive_skb(skb))
			ret = GRO_DROP;
		break;
H
Herbert Xu 已提交
3200

3201 3202
	case GRO_DROP:
	case GRO_MERGED_FREE:
H
Herbert Xu 已提交
3203 3204
		kfree_skb(skb);
		break;
3205 3206 3207 3208

	case GRO_HELD:
	case GRO_MERGED:
		break;
H
Herbert Xu 已提交
3209 3210
	}

3211
	return ret;
3212 3213 3214
}
EXPORT_SYMBOL(napi_skb_finish);

3215 3216 3217 3218
void skb_gro_reset_offset(struct sk_buff *skb)
{
	NAPI_GRO_CB(skb)->data_offset = 0;
	NAPI_GRO_CB(skb)->frag0 = NULL;
3219
	NAPI_GRO_CB(skb)->frag0_len = 0;
3220

3221
	if (skb->mac_header == skb->tail &&
3222
	    !PageHighMem(skb_shinfo(skb)->frags[0].page)) {
3223 3224 3225
		NAPI_GRO_CB(skb)->frag0 =
			page_address(skb_shinfo(skb)->frags[0].page) +
			skb_shinfo(skb)->frags[0].page_offset;
3226 3227
		NAPI_GRO_CB(skb)->frag0_len = skb_shinfo(skb)->frags[0].size;
	}
3228 3229 3230
}
EXPORT_SYMBOL(skb_gro_reset_offset);

3231
gro_result_t napi_gro_receive(struct napi_struct *napi, struct sk_buff *skb)
3232
{
3233 3234
	skb_gro_reset_offset(skb);

3235
	return napi_skb_finish(__napi_gro_receive(napi, skb), skb);
3236 3237 3238
}
EXPORT_SYMBOL(napi_gro_receive);

3239 3240 3241 3242 3243 3244 3245 3246 3247
void napi_reuse_skb(struct napi_struct *napi, struct sk_buff *skb)
{
	__skb_pull(skb, skb_headlen(skb));
	skb_reserve(skb, NET_IP_ALIGN - skb_headroom(skb));

	napi->skb = skb;
}
EXPORT_SYMBOL(napi_reuse_skb);

3248
struct sk_buff *napi_get_frags(struct napi_struct *napi)
H
Herbert Xu 已提交
3249 3250 3251 3252
{
	struct sk_buff *skb = napi->skb;

	if (!skb) {
3253 3254 3255
		skb = netdev_alloc_skb_ip_align(napi->dev, GRO_MAX_HEAD);
		if (skb)
			napi->skb = skb;
3256
	}
3257 3258
	return skb;
}
3259
EXPORT_SYMBOL(napi_get_frags);
3260

3261 3262
gro_result_t napi_frags_finish(struct napi_struct *napi, struct sk_buff *skb,
			       gro_result_t ret)
3263
{
3264 3265
	switch (ret) {
	case GRO_NORMAL:
3266
	case GRO_HELD:
3267
		skb->protocol = eth_type_trans(skb, skb->dev);
3268

3269 3270 3271 3272
		if (ret == GRO_HELD)
			skb_gro_pull(skb, -ETH_HLEN);
		else if (netif_receive_skb(skb))
			ret = GRO_DROP;
3273
		break;
H
Herbert Xu 已提交
3274

3275 3276 3277 3278
	case GRO_DROP:
	case GRO_MERGED_FREE:
		napi_reuse_skb(napi, skb);
		break;
3279 3280 3281

	case GRO_MERGED:
		break;
3282
	}
H
Herbert Xu 已提交
3283

3284
	return ret;
H
Herbert Xu 已提交
3285
}
3286 3287
EXPORT_SYMBOL(napi_frags_finish);

3288 3289 3290 3291
struct sk_buff *napi_frags_skb(struct napi_struct *napi)
{
	struct sk_buff *skb = napi->skb;
	struct ethhdr *eth;
3292 3293
	unsigned int hlen;
	unsigned int off;
3294 3295 3296 3297 3298 3299

	napi->skb = NULL;

	skb_reset_mac_header(skb);
	skb_gro_reset_offset(skb);

3300 3301 3302 3303 3304 3305 3306 3307 3308 3309
	off = skb_gro_offset(skb);
	hlen = off + sizeof(*eth);
	eth = skb_gro_header_fast(skb, off);
	if (skb_gro_header_hard(skb, hlen)) {
		eth = skb_gro_header_slow(skb, hlen, off);
		if (unlikely(!eth)) {
			napi_reuse_skb(napi, skb);
			skb = NULL;
			goto out;
		}
3310 3311 3312 3313 3314 3315 3316 3317 3318 3319 3320 3321 3322 3323 3324
	}

	skb_gro_pull(skb, sizeof(*eth));

	/*
	 * This works because the only protocols we care about don't require
	 * special handling.  We'll fix it up properly at the end.
	 */
	skb->protocol = eth->h_proto;

out:
	return skb;
}
EXPORT_SYMBOL(napi_frags_skb);

3325
gro_result_t napi_gro_frags(struct napi_struct *napi)
3326
{
3327
	struct sk_buff *skb = napi_frags_skb(napi);
3328 3329

	if (!skb)
3330
		return GRO_DROP;
3331 3332 3333

	return napi_frags_finish(napi, skb, __napi_gro_receive(napi, skb));
}
H
Herbert Xu 已提交
3334 3335
EXPORT_SYMBOL(napi_gro_frags);

3336 3337 3338 3339 3340 3341 3342 3343 3344 3345 3346 3347 3348 3349 3350 3351 3352 3353 3354 3355 3356 3357 3358 3359 3360 3361 3362 3363
/*
 * net_rps_action sends any pending IPI's for rps.
 * Note: called with local irq disabled, but exits with local irq enabled.
 */
static void net_rps_action_and_irq_enable(struct softnet_data *sd)
{
#ifdef CONFIG_RPS
	struct softnet_data *remsd = sd->rps_ipi_list;

	if (remsd) {
		sd->rps_ipi_list = NULL;

		local_irq_enable();

		/* Send pending IPI's to kick RPS processing on remote cpus. */
		while (remsd) {
			struct softnet_data *next = remsd->rps_ipi_next;

			if (cpu_online(remsd->cpu))
				__smp_call_function_single(remsd->cpu,
							   &remsd->csd, 0);
			remsd = next;
		}
	} else
#endif
		local_irq_enable();
}

3364
static int process_backlog(struct napi_struct *napi, int quota)
L
Linus Torvalds 已提交
3365 3366
{
	int work = 0;
E
Eric Dumazet 已提交
3367
	struct softnet_data *sd = container_of(napi, struct softnet_data, backlog);
L
Linus Torvalds 已提交
3368

3369 3370 3371 3372 3373 3374 3375 3376 3377
#ifdef CONFIG_RPS
	/* Check if we have pending ipi, its better to send them now,
	 * not waiting net_rx_action() end.
	 */
	if (sd->rps_ipi_list) {
		local_irq_disable();
		net_rps_action_and_irq_enable(sd);
	}
#endif
3378
	napi->weight = weight_p;
3379 3380
	local_irq_disable();
	while (work < quota) {
L
Linus Torvalds 已提交
3381
		struct sk_buff *skb;
3382 3383 3384 3385 3386 3387
		unsigned int qlen;

		while ((skb = __skb_dequeue(&sd->process_queue))) {
			local_irq_enable();
			__netif_receive_skb(skb);
			local_irq_disable();
3388 3389 3390 3391 3392
			input_queue_head_incr(sd);
			if (++work >= quota) {
				local_irq_enable();
				return work;
			}
3393
		}
L
Linus Torvalds 已提交
3394

E
Eric Dumazet 已提交
3395
		rps_lock(sd);
3396
		qlen = skb_queue_len(&sd->input_pkt_queue);
3397
		if (qlen)
3398 3399
			skb_queue_splice_tail_init(&sd->input_pkt_queue,
						   &sd->process_queue);
3400

3401
		if (qlen < quota - work) {
E
Eric Dumazet 已提交
3402 3403 3404 3405 3406 3407 3408 3409 3410 3411
			/*
			 * Inline a custom version of __napi_complete().
			 * only current cpu owns and manipulates this napi,
			 * and NAPI_STATE_SCHED is the only possible flag set on backlog.
			 * we can use a plain write instead of clear_bit(),
			 * and we dont need an smp_mb() memory barrier.
			 */
			list_del(&napi->poll_list);
			napi->state = 0;

3412
			quota = work + qlen;
3413
		}
E
Eric Dumazet 已提交
3414
		rps_unlock(sd);
3415 3416
	}
	local_irq_enable();
L
Linus Torvalds 已提交
3417

3418 3419
	return work;
}
L
Linus Torvalds 已提交
3420

3421 3422
/**
 * __napi_schedule - schedule for receive
3423
 * @n: entry to schedule
3424 3425 3426
 *
 * The entry's receive function will be scheduled to run
 */
H
Harvey Harrison 已提交
3427
void __napi_schedule(struct napi_struct *n)
3428 3429
{
	unsigned long flags;
L
Linus Torvalds 已提交
3430

3431
	local_irq_save(flags);
E
Eric Dumazet 已提交
3432
	____napi_schedule(&__get_cpu_var(softnet_data), n);
3433
	local_irq_restore(flags);
L
Linus Torvalds 已提交
3434
}
3435 3436
EXPORT_SYMBOL(__napi_schedule);

3437 3438 3439 3440 3441 3442 3443 3444 3445 3446 3447 3448 3449 3450 3451 3452 3453 3454 3455 3456 3457 3458 3459 3460 3461 3462 3463 3464 3465 3466 3467 3468 3469
void __napi_complete(struct napi_struct *n)
{
	BUG_ON(!test_bit(NAPI_STATE_SCHED, &n->state));
	BUG_ON(n->gro_list);

	list_del(&n->poll_list);
	smp_mb__before_clear_bit();
	clear_bit(NAPI_STATE_SCHED, &n->state);
}
EXPORT_SYMBOL(__napi_complete);

void napi_complete(struct napi_struct *n)
{
	unsigned long flags;

	/*
	 * don't let napi dequeue from the cpu poll list
	 * just in case its running on a different cpu
	 */
	if (unlikely(test_bit(NAPI_STATE_NPSVC, &n->state)))
		return;

	napi_gro_flush(n);
	local_irq_save(flags);
	__napi_complete(n);
	local_irq_restore(flags);
}
EXPORT_SYMBOL(napi_complete);

void netif_napi_add(struct net_device *dev, struct napi_struct *napi,
		    int (*poll)(struct napi_struct *, int), int weight)
{
	INIT_LIST_HEAD(&napi->poll_list);
3470
	napi->gro_count = 0;
3471
	napi->gro_list = NULL;
H
Herbert Xu 已提交
3472
	napi->skb = NULL;
3473 3474 3475 3476
	napi->poll = poll;
	napi->weight = weight;
	list_add(&napi->dev_list, &dev->napi_list);
	napi->dev = dev;
H
Herbert Xu 已提交
3477
#ifdef CONFIG_NETPOLL
3478 3479 3480 3481 3482 3483 3484 3485 3486 3487 3488
	spin_lock_init(&napi->poll_lock);
	napi->poll_owner = -1;
#endif
	set_bit(NAPI_STATE_SCHED, &napi->state);
}
EXPORT_SYMBOL(netif_napi_add);

void netif_napi_del(struct napi_struct *napi)
{
	struct sk_buff *skb, *next;

3489
	list_del_init(&napi->dev_list);
3490
	napi_free_frags(napi);
3491 3492 3493 3494 3495 3496 3497 3498

	for (skb = napi->gro_list; skb; skb = next) {
		next = skb->next;
		skb->next = NULL;
		kfree_skb(skb);
	}

	napi->gro_list = NULL;
3499
	napi->gro_count = 0;
3500 3501 3502
}
EXPORT_SYMBOL(netif_napi_del);

L
Linus Torvalds 已提交
3503 3504
static void net_rx_action(struct softirq_action *h)
{
3505
	struct softnet_data *sd = &__get_cpu_var(softnet_data);
3506
	unsigned long time_limit = jiffies + 2;
3507
	int budget = netdev_budget;
3508 3509
	void *have;

L
Linus Torvalds 已提交
3510 3511
	local_irq_disable();

3512
	while (!list_empty(&sd->poll_list)) {
3513 3514
		struct napi_struct *n;
		int work, weight;
L
Linus Torvalds 已提交
3515

3516
		/* If softirq window is exhuasted then punt.
3517 3518
		 * Allow this to run for 2 jiffies since which will allow
		 * an average latency of 1.5/HZ.
3519
		 */
3520
		if (unlikely(budget <= 0 || time_after(jiffies, time_limit)))
L
Linus Torvalds 已提交
3521 3522 3523 3524
			goto softnet_break;

		local_irq_enable();

3525 3526 3527 3528 3529
		/* Even though interrupts have been re-enabled, this
		 * access is safe because interrupts can only add new
		 * entries to the tail of this list, and only ->poll()
		 * calls can remove this head entry from the list.
		 */
3530
		n = list_first_entry(&sd->poll_list, struct napi_struct, poll_list);
L
Linus Torvalds 已提交
3531

3532 3533 3534 3535
		have = netpoll_poll_lock(n);

		weight = n->weight;

3536 3537 3538 3539 3540 3541 3542
		/* This NAPI_STATE_SCHED test is for avoiding a race
		 * with netpoll's poll_napi().  Only the entity which
		 * obtains the lock and sees NAPI_STATE_SCHED set will
		 * actually make the ->poll() call.  Therefore we avoid
		 * accidently calling ->poll() when NAPI is not scheduled.
		 */
		work = 0;
3543
		if (test_bit(NAPI_STATE_SCHED, &n->state)) {
3544
			work = n->poll(n, weight);
3545 3546
			trace_napi_poll(n);
		}
3547 3548 3549 3550 3551 3552 3553 3554 3555 3556 3557 3558

		WARN_ON_ONCE(work > weight);

		budget -= work;

		local_irq_disable();

		/* Drivers must not modify the NAPI state if they
		 * consume the entire weight.  In such cases this code
		 * still "owns" the NAPI instance and therefore can
		 * move the instance around on the list at-will.
		 */
3559
		if (unlikely(work == weight)) {
3560 3561 3562 3563 3564
			if (unlikely(napi_disable_pending(n))) {
				local_irq_enable();
				napi_complete(n);
				local_irq_disable();
			} else
3565
				list_move_tail(&n->poll_list, &sd->poll_list);
3566
		}
3567 3568

		netpoll_poll_unlock(have);
L
Linus Torvalds 已提交
3569 3570
	}
out:
3571
	net_rps_action_and_irq_enable(sd);
T
Tom Herbert 已提交
3572

3573 3574 3575 3576 3577
#ifdef CONFIG_NET_DMA
	/*
	 * There may not be any more sk_buffs coming right now, so push
	 * any pending DMA copies to hardware
	 */
3578
	dma_issue_pending_all();
3579
#endif
3580

L
Linus Torvalds 已提交
3581 3582 3583
	return;

softnet_break:
C
Changli Gao 已提交
3584
	sd->time_squeeze++;
L
Linus Torvalds 已提交
3585 3586 3587 3588
	__raise_softirq_irqoff(NET_RX_SOFTIRQ);
	goto out;
}

E
Eric Dumazet 已提交
3589
static gifconf_func_t *gifconf_list[NPROTO];
L
Linus Torvalds 已提交
3590 3591 3592 3593 3594 3595 3596 3597 3598 3599

/**
 *	register_gifconf	-	register a SIOCGIF handler
 *	@family: Address family
 *	@gifconf: Function handler
 *
 *	Register protocol dependent address dumping routines. The handler
 *	that is passed must not be freed or reused until it has been replaced
 *	by another handler.
 */
E
Eric Dumazet 已提交
3600
int register_gifconf(unsigned int family, gifconf_func_t *gifconf)
L
Linus Torvalds 已提交
3601 3602 3603 3604 3605 3606
{
	if (family >= NPROTO)
		return -EINVAL;
	gifconf_list[family] = gifconf;
	return 0;
}
E
Eric Dumazet 已提交
3607
EXPORT_SYMBOL(register_gifconf);
L
Linus Torvalds 已提交
3608 3609 3610 3611 3612 3613 3614 3615 3616 3617 3618 3619 3620


/*
 *	Map an interface index to its name (SIOCGIFNAME)
 */

/*
 *	We need this ioctl for efficient implementation of the
 *	if_indextoname() function required by the IPv6 API.  Without
 *	it, we would have to search all the interfaces to find a
 *	match.  --pb
 */

3621
static int dev_ifname(struct net *net, struct ifreq __user *arg)
L
Linus Torvalds 已提交
3622 3623 3624 3625 3626 3627 3628 3629 3630 3631 3632
{
	struct net_device *dev;
	struct ifreq ifr;

	/*
	 *	Fetch the caller's info block.
	 */

	if (copy_from_user(&ifr, arg, sizeof(struct ifreq)))
		return -EFAULT;

3633 3634
	rcu_read_lock();
	dev = dev_get_by_index_rcu(net, ifr.ifr_ifindex);
L
Linus Torvalds 已提交
3635
	if (!dev) {
3636
		rcu_read_unlock();
L
Linus Torvalds 已提交
3637 3638 3639 3640
		return -ENODEV;
	}

	strcpy(ifr.ifr_name, dev->name);
3641
	rcu_read_unlock();
L
Linus Torvalds 已提交
3642 3643 3644 3645 3646 3647 3648 3649 3650 3651 3652 3653

	if (copy_to_user(arg, &ifr, sizeof(struct ifreq)))
		return -EFAULT;
	return 0;
}

/*
 *	Perform a SIOCGIFCONF call. This structure will change
 *	size eventually, and there is nothing I can do about it.
 *	Thus we will need a 'compatibility mode'.
 */

3654
static int dev_ifconf(struct net *net, char __user *arg)
L
Linus Torvalds 已提交
3655 3656 3657 3658 3659 3660 3661 3662 3663 3664 3665 3666 3667 3668 3669 3670 3671 3672 3673 3674 3675 3676 3677
{
	struct ifconf ifc;
	struct net_device *dev;
	char __user *pos;
	int len;
	int total;
	int i;

	/*
	 *	Fetch the caller's info block.
	 */

	if (copy_from_user(&ifc, arg, sizeof(struct ifconf)))
		return -EFAULT;

	pos = ifc.ifc_buf;
	len = ifc.ifc_len;

	/*
	 *	Loop over the interfaces, and write an info block for each.
	 */

	total = 0;
3678
	for_each_netdev(net, dev) {
L
Linus Torvalds 已提交
3679 3680 3681 3682 3683 3684 3685 3686 3687 3688 3689 3690 3691
		for (i = 0; i < NPROTO; i++) {
			if (gifconf_list[i]) {
				int done;
				if (!pos)
					done = gifconf_list[i](dev, NULL, 0);
				else
					done = gifconf_list[i](dev, pos + total,
							       len - total);
				if (done < 0)
					return -EFAULT;
				total += done;
			}
		}
3692
	}
L
Linus Torvalds 已提交
3693 3694 3695 3696 3697 3698 3699 3700 3701 3702 3703 3704 3705 3706 3707 3708 3709

	/*
	 *	All done.  Write the updated control block back to the caller.
	 */
	ifc.ifc_len = total;

	/*
	 * 	Both BSD and Solaris return 0 here, so we do too.
	 */
	return copy_to_user(arg, &ifc, sizeof(struct ifconf)) ? -EFAULT : 0;
}

#ifdef CONFIG_PROC_FS
/*
 *	This is invoked by the /proc filesystem handler to display a device
 *	in detail.
 */
3710
void *dev_seq_start(struct seq_file *seq, loff_t *pos)
3711
	__acquires(RCU)
L
Linus Torvalds 已提交
3712
{
3713
	struct net *net = seq_file_net(seq);
3714
	loff_t off;
L
Linus Torvalds 已提交
3715 3716
	struct net_device *dev;

3717
	rcu_read_lock();
3718 3719
	if (!*pos)
		return SEQ_START_TOKEN;
L
Linus Torvalds 已提交
3720

3721
	off = 1;
3722
	for_each_netdev_rcu(net, dev)
3723 3724
		if (off++ == *pos)
			return dev;
L
Linus Torvalds 已提交
3725

3726
	return NULL;
L
Linus Torvalds 已提交
3727 3728 3729 3730
}

void *dev_seq_next(struct seq_file *seq, void *v, loff_t *pos)
{
3731 3732 3733 3734
	struct net_device *dev = (v == SEQ_START_TOKEN) ?
				  first_net_device(seq_file_net(seq)) :
				  next_net_device((struct net_device *)v);

L
Linus Torvalds 已提交
3735
	++*pos;
3736
	return rcu_dereference(dev);
L
Linus Torvalds 已提交
3737 3738 3739
}

void dev_seq_stop(struct seq_file *seq, void *v)
3740
	__releases(RCU)
L
Linus Torvalds 已提交
3741
{
3742
	rcu_read_unlock();
L
Linus Torvalds 已提交
3743 3744 3745 3746
}

static void dev_seq_printf_stats(struct seq_file *seq, struct net_device *dev)
{
3747 3748
	struct rtnl_link_stats64 temp;
	const struct rtnl_link_stats64 *stats = dev_get_stats(dev, &temp);
L
Linus Torvalds 已提交
3749

3750 3751
	seq_printf(seq, "%6s: %7llu %7llu %4llu %4llu %4llu %5llu %10llu %9llu "
		   "%8llu %7llu %4llu %4llu %4llu %5llu %7llu %10llu\n",
3752 3753 3754 3755 3756 3757 3758 3759 3760 3761 3762 3763 3764 3765 3766
		   dev->name, stats->rx_bytes, stats->rx_packets,
		   stats->rx_errors,
		   stats->rx_dropped + stats->rx_missed_errors,
		   stats->rx_fifo_errors,
		   stats->rx_length_errors + stats->rx_over_errors +
		    stats->rx_crc_errors + stats->rx_frame_errors,
		   stats->rx_compressed, stats->multicast,
		   stats->tx_bytes, stats->tx_packets,
		   stats->tx_errors, stats->tx_dropped,
		   stats->tx_fifo_errors, stats->collisions,
		   stats->tx_carrier_errors +
		    stats->tx_aborted_errors +
		    stats->tx_window_errors +
		    stats->tx_heartbeat_errors,
		   stats->tx_compressed);
L
Linus Torvalds 已提交
3767 3768 3769 3770 3771 3772 3773 3774 3775 3776 3777 3778 3779 3780 3781 3782 3783 3784 3785
}

/*
 *	Called from the PROCfs module. This now uses the new arbitrary sized
 *	/proc/net interface to create /proc/net/dev
 */
static int dev_seq_show(struct seq_file *seq, void *v)
{
	if (v == SEQ_START_TOKEN)
		seq_puts(seq, "Inter-|   Receive                            "
			      "                    |  Transmit\n"
			      " face |bytes    packets errs drop fifo frame "
			      "compressed multicast|bytes    packets errs "
			      "drop fifo colls carrier compressed\n");
	else
		dev_seq_printf_stats(seq, v);
	return 0;
}

C
Changli Gao 已提交
3786
static struct softnet_data *softnet_get_online(loff_t *pos)
L
Linus Torvalds 已提交
3787
{
C
Changli Gao 已提交
3788
	struct softnet_data *sd = NULL;
L
Linus Torvalds 已提交
3789

3790
	while (*pos < nr_cpu_ids)
3791
		if (cpu_online(*pos)) {
C
Changli Gao 已提交
3792
			sd = &per_cpu(softnet_data, *pos);
L
Linus Torvalds 已提交
3793 3794 3795
			break;
		} else
			++*pos;
C
Changli Gao 已提交
3796
	return sd;
L
Linus Torvalds 已提交
3797 3798 3799 3800 3801 3802 3803 3804 3805 3806 3807 3808 3809 3810 3811 3812 3813 3814 3815
}

static void *softnet_seq_start(struct seq_file *seq, loff_t *pos)
{
	return softnet_get_online(pos);
}

static void *softnet_seq_next(struct seq_file *seq, void *v, loff_t *pos)
{
	++*pos;
	return softnet_get_online(pos);
}

static void softnet_seq_stop(struct seq_file *seq, void *v)
{
}

static int softnet_seq_show(struct seq_file *seq, void *v)
{
C
Changli Gao 已提交
3816
	struct softnet_data *sd = v;
L
Linus Torvalds 已提交
3817

T
Tom Herbert 已提交
3818
	seq_printf(seq, "%08x %08x %08x %08x %08x %08x %08x %08x %08x %08x\n",
C
Changli Gao 已提交
3819
		   sd->processed, sd->dropped, sd->time_squeeze, 0,
3820
		   0, 0, 0, 0, /* was fastroute */
C
Changli Gao 已提交
3821
		   sd->cpu_collision, sd->received_rps);
L
Linus Torvalds 已提交
3822 3823 3824
	return 0;
}

3825
static const struct seq_operations dev_seq_ops = {
L
Linus Torvalds 已提交
3826 3827 3828 3829 3830 3831 3832 3833
	.start = dev_seq_start,
	.next  = dev_seq_next,
	.stop  = dev_seq_stop,
	.show  = dev_seq_show,
};

static int dev_seq_open(struct inode *inode, struct file *file)
{
3834 3835
	return seq_open_net(inode, file, &dev_seq_ops,
			    sizeof(struct seq_net_private));
L
Linus Torvalds 已提交
3836 3837
}

3838
static const struct file_operations dev_seq_fops = {
L
Linus Torvalds 已提交
3839 3840 3841 3842
	.owner	 = THIS_MODULE,
	.open    = dev_seq_open,
	.read    = seq_read,
	.llseek  = seq_lseek,
3843
	.release = seq_release_net,
L
Linus Torvalds 已提交
3844 3845
};

3846
static const struct seq_operations softnet_seq_ops = {
L
Linus Torvalds 已提交
3847 3848 3849 3850 3851 3852 3853 3854 3855 3856 3857
	.start = softnet_seq_start,
	.next  = softnet_seq_next,
	.stop  = softnet_seq_stop,
	.show  = softnet_seq_show,
};

static int softnet_seq_open(struct inode *inode, struct file *file)
{
	return seq_open(file, &softnet_seq_ops);
}

3858
static const struct file_operations softnet_seq_fops = {
L
Linus Torvalds 已提交
3859 3860 3861 3862 3863 3864 3865
	.owner	 = THIS_MODULE,
	.open    = softnet_seq_open,
	.read    = seq_read,
	.llseek  = seq_lseek,
	.release = seq_release,
};

3866 3867 3868 3869 3870 3871 3872 3873 3874 3875 3876 3877
static void *ptype_get_idx(loff_t pos)
{
	struct packet_type *pt = NULL;
	loff_t i = 0;
	int t;

	list_for_each_entry_rcu(pt, &ptype_all, list) {
		if (i == pos)
			return pt;
		++i;
	}

3878
	for (t = 0; t < PTYPE_HASH_SIZE; t++) {
3879 3880 3881 3882 3883 3884 3885 3886 3887 3888
		list_for_each_entry_rcu(pt, &ptype_base[t], list) {
			if (i == pos)
				return pt;
			++i;
		}
	}
	return NULL;
}

static void *ptype_seq_start(struct seq_file *seq, loff_t *pos)
3889
	__acquires(RCU)
3890 3891 3892 3893 3894 3895 3896 3897 3898 3899 3900 3901 3902 3903 3904 3905 3906 3907 3908 3909 3910 3911 3912
{
	rcu_read_lock();
	return *pos ? ptype_get_idx(*pos - 1) : SEQ_START_TOKEN;
}

static void *ptype_seq_next(struct seq_file *seq, void *v, loff_t *pos)
{
	struct packet_type *pt;
	struct list_head *nxt;
	int hash;

	++*pos;
	if (v == SEQ_START_TOKEN)
		return ptype_get_idx(0);

	pt = v;
	nxt = pt->list.next;
	if (pt->type == htons(ETH_P_ALL)) {
		if (nxt != &ptype_all)
			goto found;
		hash = 0;
		nxt = ptype_base[0].next;
	} else
3913
		hash = ntohs(pt->type) & PTYPE_HASH_MASK;
3914 3915

	while (nxt == &ptype_base[hash]) {
3916
		if (++hash >= PTYPE_HASH_SIZE)
3917 3918 3919 3920 3921 3922 3923 3924
			return NULL;
		nxt = ptype_base[hash].next;
	}
found:
	return list_entry(nxt, struct packet_type, list);
}

static void ptype_seq_stop(struct seq_file *seq, void *v)
3925
	__releases(RCU)
3926 3927 3928 3929 3930 3931 3932 3933 3934 3935
{
	rcu_read_unlock();
}

static int ptype_seq_show(struct seq_file *seq, void *v)
{
	struct packet_type *pt = v;

	if (v == SEQ_START_TOKEN)
		seq_puts(seq, "Type Device      Function\n");
3936
	else if (pt->dev == NULL || dev_net(pt->dev) == seq_file_net(seq)) {
3937 3938 3939 3940 3941
		if (pt->type == htons(ETH_P_ALL))
			seq_puts(seq, "ALL ");
		else
			seq_printf(seq, "%04x", ntohs(pt->type));

A
Alexey Dobriyan 已提交
3942 3943
		seq_printf(seq, " %-8s %pF\n",
			   pt->dev ? pt->dev->name : "", pt->func);
3944 3945 3946 3947 3948 3949 3950 3951 3952 3953 3954 3955 3956 3957
	}

	return 0;
}

static const struct seq_operations ptype_seq_ops = {
	.start = ptype_seq_start,
	.next  = ptype_seq_next,
	.stop  = ptype_seq_stop,
	.show  = ptype_seq_show,
};

static int ptype_seq_open(struct inode *inode, struct file *file)
{
3958 3959
	return seq_open_net(inode, file, &ptype_seq_ops,
			sizeof(struct seq_net_private));
3960 3961 3962 3963 3964 3965 3966
}

static const struct file_operations ptype_seq_fops = {
	.owner	 = THIS_MODULE,
	.open    = ptype_seq_open,
	.read    = seq_read,
	.llseek  = seq_lseek,
3967
	.release = seq_release_net,
3968 3969 3970
};


3971
static int __net_init dev_proc_net_init(struct net *net)
L
Linus Torvalds 已提交
3972 3973 3974
{
	int rc = -ENOMEM;

3975
	if (!proc_net_fops_create(net, "dev", S_IRUGO, &dev_seq_fops))
L
Linus Torvalds 已提交
3976
		goto out;
3977
	if (!proc_net_fops_create(net, "softnet_stat", S_IRUGO, &softnet_seq_fops))
L
Linus Torvalds 已提交
3978
		goto out_dev;
3979
	if (!proc_net_fops_create(net, "ptype", S_IRUGO, &ptype_seq_fops))
3980
		goto out_softnet;
3981

3982
	if (wext_proc_init(net))
3983
		goto out_ptype;
L
Linus Torvalds 已提交
3984 3985 3986
	rc = 0;
out:
	return rc;
3987
out_ptype:
3988
	proc_net_remove(net, "ptype");
L
Linus Torvalds 已提交
3989
out_softnet:
3990
	proc_net_remove(net, "softnet_stat");
L
Linus Torvalds 已提交
3991
out_dev:
3992
	proc_net_remove(net, "dev");
L
Linus Torvalds 已提交
3993 3994
	goto out;
}
3995

3996
static void __net_exit dev_proc_net_exit(struct net *net)
3997 3998 3999 4000 4001 4002 4003 4004
{
	wext_proc_exit(net);

	proc_net_remove(net, "ptype");
	proc_net_remove(net, "softnet_stat");
	proc_net_remove(net, "dev");
}

4005
static struct pernet_operations __net_initdata dev_proc_ops = {
4006 4007 4008 4009 4010 4011 4012 4013
	.init = dev_proc_net_init,
	.exit = dev_proc_net_exit,
};

static int __init dev_proc_init(void)
{
	return register_pernet_subsys(&dev_proc_ops);
}
L
Linus Torvalds 已提交
4014 4015 4016 4017 4018 4019 4020 4021 4022 4023 4024 4025 4026 4027 4028 4029 4030 4031 4032 4033 4034 4035 4036 4037 4038 4039 4040 4041 4042
#else
#define dev_proc_init() 0
#endif	/* CONFIG_PROC_FS */


/**
 *	netdev_set_master	-	set up master/slave pair
 *	@slave: slave device
 *	@master: new master device
 *
 *	Changes the master device of the slave. Pass %NULL to break the
 *	bonding. The caller must hold the RTNL semaphore. On a failure
 *	a negative errno code is returned. On success the reference counts
 *	are adjusted, %RTM_NEWLINK is sent to the routing socket and the
 *	function returns zero.
 */
int netdev_set_master(struct net_device *slave, struct net_device *master)
{
	struct net_device *old = slave->master;

	ASSERT_RTNL();

	if (master) {
		if (old)
			return -EBUSY;
		dev_hold(master);
	}

	slave->master = master;
4043

E
Eric Dumazet 已提交
4044 4045
	if (old) {
		synchronize_net();
L
Linus Torvalds 已提交
4046
		dev_put(old);
E
Eric Dumazet 已提交
4047
	}
L
Linus Torvalds 已提交
4048 4049 4050 4051 4052 4053 4054 4055
	if (master)
		slave->flags |= IFF_SLAVE;
	else
		slave->flags &= ~IFF_SLAVE;

	rtmsg_ifinfo(RTM_NEWLINK, slave, IFF_SLAVE);
	return 0;
}
E
Eric Dumazet 已提交
4056
EXPORT_SYMBOL(netdev_set_master);
L
Linus Torvalds 已提交
4057

4058 4059
static void dev_change_rx_flags(struct net_device *dev, int flags)
{
4060 4061 4062 4063
	const struct net_device_ops *ops = dev->netdev_ops;

	if ((dev->flags & IFF_UP) && ops->ndo_change_rx_flags)
		ops->ndo_change_rx_flags(dev, flags);
4064 4065
}

4066
static int __dev_set_promiscuity(struct net_device *dev, int inc)
L
Linus Torvalds 已提交
4067 4068
{
	unsigned short old_flags = dev->flags;
4069 4070
	uid_t uid;
	gid_t gid;
L
Linus Torvalds 已提交
4071

4072 4073
	ASSERT_RTNL();

4074 4075 4076 4077 4078 4079 4080 4081 4082 4083 4084 4085 4086 4087 4088 4089 4090
	dev->flags |= IFF_PROMISC;
	dev->promiscuity += inc;
	if (dev->promiscuity == 0) {
		/*
		 * Avoid overflow.
		 * If inc causes overflow, untouch promisc and return error.
		 */
		if (inc < 0)
			dev->flags &= ~IFF_PROMISC;
		else {
			dev->promiscuity -= inc;
			printk(KERN_WARNING "%s: promiscuity touches roof, "
				"set promiscuity failed, promiscuity feature "
				"of device might be broken.\n", dev->name);
			return -EOVERFLOW;
		}
	}
4091
	if (dev->flags != old_flags) {
L
Linus Torvalds 已提交
4092 4093
		printk(KERN_INFO "device %s %s promiscuous mode\n",
		       dev->name, (dev->flags & IFF_PROMISC) ? "entered" :
4094
							       "left");
4095 4096
		if (audit_enabled) {
			current_uid_gid(&uid, &gid);
4097 4098 4099 4100 4101 4102
			audit_log(current->audit_context, GFP_ATOMIC,
				AUDIT_ANOM_PROMISCUOUS,
				"dev=%s prom=%d old_prom=%d auid=%u uid=%u gid=%u ses=%u",
				dev->name, (dev->flags & IFF_PROMISC),
				(old_flags & IFF_PROMISC),
				audit_get_loginuid(current),
4103
				uid, gid,
4104
				audit_get_sessionid(current));
4105
		}
4106

4107
		dev_change_rx_flags(dev, IFF_PROMISC);
L
Linus Torvalds 已提交
4108
	}
4109
	return 0;
L
Linus Torvalds 已提交
4110 4111
}

4112 4113 4114 4115 4116 4117 4118 4119 4120
/**
 *	dev_set_promiscuity	- update promiscuity count on a device
 *	@dev: device
 *	@inc: modifier
 *
 *	Add or remove promiscuity from a device. While the count in the device
 *	remains above zero the interface remains promiscuous. Once it hits zero
 *	the device reverts back to normal filtering operation. A negative inc
 *	value is used to drop promiscuity on the device.
4121
 *	Return 0 if successful or a negative errno code on error.
4122
 */
4123
int dev_set_promiscuity(struct net_device *dev, int inc)
4124 4125
{
	unsigned short old_flags = dev->flags;
4126
	int err;
4127

4128
	err = __dev_set_promiscuity(dev, inc);
4129
	if (err < 0)
4130
		return err;
4131 4132
	if (dev->flags != old_flags)
		dev_set_rx_mode(dev);
4133
	return err;
4134
}
E
Eric Dumazet 已提交
4135
EXPORT_SYMBOL(dev_set_promiscuity);
4136

L
Linus Torvalds 已提交
4137 4138 4139 4140 4141 4142 4143 4144 4145 4146
/**
 *	dev_set_allmulti	- update allmulti count on a device
 *	@dev: device
 *	@inc: modifier
 *
 *	Add or remove reception of all multicast frames to a device. While the
 *	count in the device remains above zero the interface remains listening
 *	to all interfaces. Once it hits zero the device reverts back to normal
 *	filtering operation. A negative @inc value is used to drop the counter
 *	when releasing a resource needing all multicasts.
4147
 *	Return 0 if successful or a negative errno code on error.
L
Linus Torvalds 已提交
4148 4149
 */

4150
int dev_set_allmulti(struct net_device *dev, int inc)
L
Linus Torvalds 已提交
4151 4152 4153
{
	unsigned short old_flags = dev->flags;

4154 4155
	ASSERT_RTNL();

L
Linus Torvalds 已提交
4156
	dev->flags |= IFF_ALLMULTI;
4157 4158 4159 4160 4161 4162 4163 4164 4165 4166 4167 4168 4169 4170 4171 4172
	dev->allmulti += inc;
	if (dev->allmulti == 0) {
		/*
		 * Avoid overflow.
		 * If inc causes overflow, untouch allmulti and return error.
		 */
		if (inc < 0)
			dev->flags &= ~IFF_ALLMULTI;
		else {
			dev->allmulti -= inc;
			printk(KERN_WARNING "%s: allmulti touches roof, "
				"set allmulti failed, allmulti feature of "
				"device might be broken.\n", dev->name);
			return -EOVERFLOW;
		}
	}
4173
	if (dev->flags ^ old_flags) {
4174
		dev_change_rx_flags(dev, IFF_ALLMULTI);
4175
		dev_set_rx_mode(dev);
4176
	}
4177
	return 0;
4178
}
E
Eric Dumazet 已提交
4179
EXPORT_SYMBOL(dev_set_allmulti);
4180 4181 4182 4183

/*
 *	Upload unicast and multicast address lists to device and
 *	configure RX filtering. When the device doesn't support unicast
J
Joe Perches 已提交
4184
 *	filtering it is put in promiscuous mode while unicast addresses
4185 4186 4187 4188
 *	are present.
 */
void __dev_set_rx_mode(struct net_device *dev)
{
4189 4190
	const struct net_device_ops *ops = dev->netdev_ops;

4191 4192 4193 4194 4195
	/* dev_open will call this function so the list will stay sane. */
	if (!(dev->flags&IFF_UP))
		return;

	if (!netif_device_present(dev))
4196
		return;
4197

4198 4199
	if (ops->ndo_set_rx_mode)
		ops->ndo_set_rx_mode(dev);
4200 4201 4202 4203
	else {
		/* Unicast addresses changes may only happen under the rtnl,
		 * therefore calling __dev_set_promiscuity here is safe.
		 */
4204
		if (!netdev_uc_empty(dev) && !dev->uc_promisc) {
4205 4206
			__dev_set_promiscuity(dev, 1);
			dev->uc_promisc = 1;
4207
		} else if (netdev_uc_empty(dev) && dev->uc_promisc) {
4208 4209 4210 4211
			__dev_set_promiscuity(dev, -1);
			dev->uc_promisc = 0;
		}

4212 4213
		if (ops->ndo_set_multicast_list)
			ops->ndo_set_multicast_list(dev);
4214 4215 4216 4217 4218
	}
}

void dev_set_rx_mode(struct net_device *dev)
{
4219
	netif_addr_lock_bh(dev);
4220
	__dev_set_rx_mode(dev);
4221
	netif_addr_unlock_bh(dev);
L
Linus Torvalds 已提交
4222 4223
}

4224 4225 4226 4227 4228 4229
/**
 *	dev_get_flags - get flags reported to userspace
 *	@dev: device
 *
 *	Get the combination of flag bits exported through APIs to userspace.
 */
L
Linus Torvalds 已提交
4230 4231 4232 4233 4234 4235
unsigned dev_get_flags(const struct net_device *dev)
{
	unsigned flags;

	flags = (dev->flags & ~(IFF_PROMISC |
				IFF_ALLMULTI |
S
Stefan Rompf 已提交
4236 4237 4238
				IFF_RUNNING |
				IFF_LOWER_UP |
				IFF_DORMANT)) |
L
Linus Torvalds 已提交
4239 4240 4241
		(dev->gflags & (IFF_PROMISC |
				IFF_ALLMULTI));

S
Stefan Rompf 已提交
4242 4243 4244 4245 4246 4247 4248 4249
	if (netif_running(dev)) {
		if (netif_oper_up(dev))
			flags |= IFF_RUNNING;
		if (netif_carrier_ok(dev))
			flags |= IFF_LOWER_UP;
		if (netif_dormant(dev))
			flags |= IFF_DORMANT;
	}
L
Linus Torvalds 已提交
4250 4251 4252

	return flags;
}
E
Eric Dumazet 已提交
4253
EXPORT_SYMBOL(dev_get_flags);
L
Linus Torvalds 已提交
4254

4255
int __dev_change_flags(struct net_device *dev, unsigned int flags)
L
Linus Torvalds 已提交
4256 4257
{
	int old_flags = dev->flags;
4258
	int ret;
L
Linus Torvalds 已提交
4259

4260 4261
	ASSERT_RTNL();

L
Linus Torvalds 已提交
4262 4263 4264 4265 4266 4267 4268 4269 4270 4271 4272 4273 4274 4275
	/*
	 *	Set the flags on our device.
	 */

	dev->flags = (flags & (IFF_DEBUG | IFF_NOTRAILERS | IFF_NOARP |
			       IFF_DYNAMIC | IFF_MULTICAST | IFF_PORTSEL |
			       IFF_AUTOMEDIA)) |
		     (dev->flags & (IFF_UP | IFF_VOLATILE | IFF_PROMISC |
				    IFF_ALLMULTI));

	/*
	 *	Load in the correct multicast list now the flags have changed.
	 */

4276 4277
	if ((old_flags ^ flags) & IFF_MULTICAST)
		dev_change_rx_flags(dev, IFF_MULTICAST);
4278

4279
	dev_set_rx_mode(dev);
L
Linus Torvalds 已提交
4280 4281 4282 4283 4284 4285 4286 4287 4288

	/*
	 *	Have we downed the interface. We handle IFF_UP ourselves
	 *	according to user attempts to set it, rather than blindly
	 *	setting it.
	 */

	ret = 0;
	if ((old_flags ^ flags) & IFF_UP) {	/* Bit is different  ? */
4289
		ret = ((old_flags & IFF_UP) ? __dev_close : __dev_open)(dev);
L
Linus Torvalds 已提交
4290 4291

		if (!ret)
4292
			dev_set_rx_mode(dev);
L
Linus Torvalds 已提交
4293 4294 4295
	}

	if ((flags ^ dev->gflags) & IFF_PROMISC) {
E
Eric Dumazet 已提交
4296 4297
		int inc = (flags & IFF_PROMISC) ? 1 : -1;

L
Linus Torvalds 已提交
4298 4299 4300 4301 4302 4303 4304 4305 4306
		dev->gflags ^= IFF_PROMISC;
		dev_set_promiscuity(dev, inc);
	}

	/* NOTE: order of synchronization of IFF_PROMISC and IFF_ALLMULTI
	   is important. Some (broken) drivers set IFF_PROMISC, when
	   IFF_ALLMULTI is requested not asking us and not reporting.
	 */
	if ((flags ^ dev->gflags) & IFF_ALLMULTI) {
E
Eric Dumazet 已提交
4307 4308
		int inc = (flags & IFF_ALLMULTI) ? 1 : -1;

L
Linus Torvalds 已提交
4309 4310 4311 4312
		dev->gflags ^= IFF_ALLMULTI;
		dev_set_allmulti(dev, inc);
	}

4313 4314 4315 4316 4317 4318 4319 4320 4321 4322 4323 4324 4325 4326 4327 4328 4329 4330 4331 4332 4333 4334 4335 4336 4337 4338 4339 4340 4341 4342 4343 4344 4345 4346 4347 4348 4349
	return ret;
}

void __dev_notify_flags(struct net_device *dev, unsigned int old_flags)
{
	unsigned int changes = dev->flags ^ old_flags;

	if (changes & IFF_UP) {
		if (dev->flags & IFF_UP)
			call_netdevice_notifiers(NETDEV_UP, dev);
		else
			call_netdevice_notifiers(NETDEV_DOWN, dev);
	}

	if (dev->flags & IFF_UP &&
	    (changes & ~(IFF_UP | IFF_PROMISC | IFF_ALLMULTI | IFF_VOLATILE)))
		call_netdevice_notifiers(NETDEV_CHANGE, dev);
}

/**
 *	dev_change_flags - change device settings
 *	@dev: device
 *	@flags: device state flags
 *
 *	Change settings on device based state flags. The flags are
 *	in the userspace exported format.
 */
int dev_change_flags(struct net_device *dev, unsigned flags)
{
	int ret, changes;
	int old_flags = dev->flags;

	ret = __dev_change_flags(dev, flags);
	if (ret < 0)
		return ret;

	changes = old_flags ^ dev->flags;
4350 4351
	if (changes)
		rtmsg_ifinfo(RTM_NEWLINK, dev, changes);
L
Linus Torvalds 已提交
4352

4353
	__dev_notify_flags(dev, old_flags);
L
Linus Torvalds 已提交
4354 4355
	return ret;
}
E
Eric Dumazet 已提交
4356
EXPORT_SYMBOL(dev_change_flags);
L
Linus Torvalds 已提交
4357

4358 4359 4360 4361 4362 4363 4364
/**
 *	dev_set_mtu - Change maximum transfer unit
 *	@dev: device
 *	@new_mtu: new transfer unit
 *
 *	Change the maximum transfer size of the network device.
 */
L
Linus Torvalds 已提交
4365 4366
int dev_set_mtu(struct net_device *dev, int new_mtu)
{
4367
	const struct net_device_ops *ops = dev->netdev_ops;
L
Linus Torvalds 已提交
4368 4369 4370 4371 4372 4373 4374 4375 4376 4377 4378 4379 4380
	int err;

	if (new_mtu == dev->mtu)
		return 0;

	/*	MTU must be positive.	 */
	if (new_mtu < 0)
		return -EINVAL;

	if (!netif_device_present(dev))
		return -ENODEV;

	err = 0;
4381 4382
	if (ops->ndo_change_mtu)
		err = ops->ndo_change_mtu(dev, new_mtu);
L
Linus Torvalds 已提交
4383 4384
	else
		dev->mtu = new_mtu;
4385

L
Linus Torvalds 已提交
4386
	if (!err && dev->flags & IFF_UP)
4387
		call_netdevice_notifiers(NETDEV_CHANGEMTU, dev);
L
Linus Torvalds 已提交
4388 4389
	return err;
}
E
Eric Dumazet 已提交
4390
EXPORT_SYMBOL(dev_set_mtu);
L
Linus Torvalds 已提交
4391

4392 4393 4394 4395 4396 4397 4398
/**
 *	dev_set_mac_address - Change Media Access Control Address
 *	@dev: device
 *	@sa: new address
 *
 *	Change the hardware (MAC) address of the device
 */
L
Linus Torvalds 已提交
4399 4400
int dev_set_mac_address(struct net_device *dev, struct sockaddr *sa)
{
4401
	const struct net_device_ops *ops = dev->netdev_ops;
L
Linus Torvalds 已提交
4402 4403
	int err;

4404
	if (!ops->ndo_set_mac_address)
L
Linus Torvalds 已提交
4405 4406 4407 4408 4409
		return -EOPNOTSUPP;
	if (sa->sa_family != dev->type)
		return -EINVAL;
	if (!netif_device_present(dev))
		return -ENODEV;
4410
	err = ops->ndo_set_mac_address(dev, sa);
L
Linus Torvalds 已提交
4411
	if (!err)
4412
		call_netdevice_notifiers(NETDEV_CHANGEADDR, dev);
L
Linus Torvalds 已提交
4413 4414
	return err;
}
E
Eric Dumazet 已提交
4415
EXPORT_SYMBOL(dev_set_mac_address);
L
Linus Torvalds 已提交
4416 4417

/*
E
Eric Dumazet 已提交
4418
 *	Perform the SIOCxIFxxx calls, inside rcu_read_lock()
L
Linus Torvalds 已提交
4419
 */
4420
static int dev_ifsioc_locked(struct net *net, struct ifreq *ifr, unsigned int cmd)
L
Linus Torvalds 已提交
4421 4422
{
	int err;
E
Eric Dumazet 已提交
4423
	struct net_device *dev = dev_get_by_name_rcu(net, ifr->ifr_name);
L
Linus Torvalds 已提交
4424 4425 4426 4427 4428

	if (!dev)
		return -ENODEV;

	switch (cmd) {
E
Eric Dumazet 已提交
4429 4430 4431
	case SIOCGIFFLAGS:	/* Get interface flags */
		ifr->ifr_flags = (short) dev_get_flags(dev);
		return 0;
L
Linus Torvalds 已提交
4432

E
Eric Dumazet 已提交
4433 4434 4435 4436
	case SIOCGIFMETRIC:	/* Get the metric on the interface
				   (currently unused) */
		ifr->ifr_metric = 0;
		return 0;
L
Linus Torvalds 已提交
4437

E
Eric Dumazet 已提交
4438 4439 4440
	case SIOCGIFMTU:	/* Get the MTU of a device */
		ifr->ifr_mtu = dev->mtu;
		return 0;
L
Linus Torvalds 已提交
4441

E
Eric Dumazet 已提交
4442 4443 4444 4445 4446 4447 4448 4449
	case SIOCGIFHWADDR:
		if (!dev->addr_len)
			memset(ifr->ifr_hwaddr.sa_data, 0, sizeof ifr->ifr_hwaddr.sa_data);
		else
			memcpy(ifr->ifr_hwaddr.sa_data, dev->dev_addr,
			       min(sizeof ifr->ifr_hwaddr.sa_data, (size_t) dev->addr_len));
		ifr->ifr_hwaddr.sa_family = dev->type;
		return 0;
L
Linus Torvalds 已提交
4450

E
Eric Dumazet 已提交
4451 4452 4453
	case SIOCGIFSLAVE:
		err = -EINVAL;
		break;
4454

E
Eric Dumazet 已提交
4455 4456 4457 4458 4459 4460 4461 4462
	case SIOCGIFMAP:
		ifr->ifr_map.mem_start = dev->mem_start;
		ifr->ifr_map.mem_end   = dev->mem_end;
		ifr->ifr_map.base_addr = dev->base_addr;
		ifr->ifr_map.irq       = dev->irq;
		ifr->ifr_map.dma       = dev->dma;
		ifr->ifr_map.port      = dev->if_port;
		return 0;
4463

E
Eric Dumazet 已提交
4464 4465 4466
	case SIOCGIFINDEX:
		ifr->ifr_ifindex = dev->ifindex;
		return 0;
4467

E
Eric Dumazet 已提交
4468 4469 4470
	case SIOCGIFTXQLEN:
		ifr->ifr_qlen = dev->tx_queue_len;
		return 0;
4471

E
Eric Dumazet 已提交
4472 4473 4474 4475 4476 4477 4478
	default:
		/* dev_ioctl() should ensure this case
		 * is never reached
		 */
		WARN_ON(1);
		err = -EINVAL;
		break;
4479 4480 4481 4482 4483 4484 4485 4486 4487 4488 4489 4490

	}
	return err;
}

/*
 *	Perform the SIOCxIFxxx calls, inside rtnl_lock()
 */
static int dev_ifsioc(struct net *net, struct ifreq *ifr, unsigned int cmd)
{
	int err;
	struct net_device *dev = __dev_get_by_name(net, ifr->ifr_name);
J
Jarek Poplawski 已提交
4491
	const struct net_device_ops *ops;
4492 4493 4494 4495

	if (!dev)
		return -ENODEV;

J
Jarek Poplawski 已提交
4496 4497
	ops = dev->netdev_ops;

4498
	switch (cmd) {
E
Eric Dumazet 已提交
4499 4500
	case SIOCSIFFLAGS:	/* Set interface flags */
		return dev_change_flags(dev, ifr->ifr_flags);
4501

E
Eric Dumazet 已提交
4502 4503 4504
	case SIOCSIFMETRIC:	/* Set the metric on the interface
				   (currently unused) */
		return -EOPNOTSUPP;
4505

E
Eric Dumazet 已提交
4506 4507
	case SIOCSIFMTU:	/* Set the MTU of a device */
		return dev_set_mtu(dev, ifr->ifr_mtu);
L
Linus Torvalds 已提交
4508

E
Eric Dumazet 已提交
4509 4510
	case SIOCSIFHWADDR:
		return dev_set_mac_address(dev, &ifr->ifr_hwaddr);
L
Linus Torvalds 已提交
4511

E
Eric Dumazet 已提交
4512 4513 4514 4515 4516 4517 4518
	case SIOCSIFHWBROADCAST:
		if (ifr->ifr_hwaddr.sa_family != dev->type)
			return -EINVAL;
		memcpy(dev->broadcast, ifr->ifr_hwaddr.sa_data,
		       min(sizeof ifr->ifr_hwaddr.sa_data, (size_t) dev->addr_len));
		call_netdevice_notifiers(NETDEV_CHANGEADDR, dev);
		return 0;
L
Linus Torvalds 已提交
4519

E
Eric Dumazet 已提交
4520 4521
	case SIOCSIFMAP:
		if (ops->ndo_set_config) {
L
Linus Torvalds 已提交
4522 4523
			if (!netif_device_present(dev))
				return -ENODEV;
E
Eric Dumazet 已提交
4524 4525 4526
			return ops->ndo_set_config(dev, &ifr->ifr_map);
		}
		return -EOPNOTSUPP;
L
Linus Torvalds 已提交
4527

E
Eric Dumazet 已提交
4528 4529 4530 4531 4532 4533
	case SIOCADDMULTI:
		if ((!ops->ndo_set_multicast_list && !ops->ndo_set_rx_mode) ||
		    ifr->ifr_hwaddr.sa_family != AF_UNSPEC)
			return -EINVAL;
		if (!netif_device_present(dev))
			return -ENODEV;
4534
		return dev_mc_add_global(dev, ifr->ifr_hwaddr.sa_data);
E
Eric Dumazet 已提交
4535 4536 4537 4538 4539 4540 4541

	case SIOCDELMULTI:
		if ((!ops->ndo_set_multicast_list && !ops->ndo_set_rx_mode) ||
		    ifr->ifr_hwaddr.sa_family != AF_UNSPEC)
			return -EINVAL;
		if (!netif_device_present(dev))
			return -ENODEV;
4542
		return dev_mc_del_global(dev, ifr->ifr_hwaddr.sa_data);
L
Linus Torvalds 已提交
4543

E
Eric Dumazet 已提交
4544 4545 4546 4547 4548
	case SIOCSIFTXQLEN:
		if (ifr->ifr_qlen < 0)
			return -EINVAL;
		dev->tx_queue_len = ifr->ifr_qlen;
		return 0;
L
Linus Torvalds 已提交
4549

E
Eric Dumazet 已提交
4550 4551 4552
	case SIOCSIFNAME:
		ifr->ifr_newname[IFNAMSIZ-1] = '\0';
		return dev_change_name(dev, ifr->ifr_newname);
L
Linus Torvalds 已提交
4553

E
Eric Dumazet 已提交
4554 4555 4556 4557 4558 4559 4560 4561 4562 4563 4564 4565 4566 4567 4568 4569 4570 4571 4572 4573 4574 4575 4576 4577 4578 4579 4580 4581
	/*
	 *	Unknown or private ioctl
	 */
	default:
		if ((cmd >= SIOCDEVPRIVATE &&
		    cmd <= SIOCDEVPRIVATE + 15) ||
		    cmd == SIOCBONDENSLAVE ||
		    cmd == SIOCBONDRELEASE ||
		    cmd == SIOCBONDSETHWADDR ||
		    cmd == SIOCBONDSLAVEINFOQUERY ||
		    cmd == SIOCBONDINFOQUERY ||
		    cmd == SIOCBONDCHANGEACTIVE ||
		    cmd == SIOCGMIIPHY ||
		    cmd == SIOCGMIIREG ||
		    cmd == SIOCSMIIREG ||
		    cmd == SIOCBRADDIF ||
		    cmd == SIOCBRDELIF ||
		    cmd == SIOCSHWTSTAMP ||
		    cmd == SIOCWANDEV) {
			err = -EOPNOTSUPP;
			if (ops->ndo_do_ioctl) {
				if (netif_device_present(dev))
					err = ops->ndo_do_ioctl(dev, ifr, cmd);
				else
					err = -ENODEV;
			}
		} else
			err = -EINVAL;
L
Linus Torvalds 已提交
4582 4583 4584 4585 4586 4587 4588 4589 4590 4591 4592 4593

	}
	return err;
}

/*
 *	This function handles all "interface"-type I/O control requests. The actual
 *	'doing' part of this is dev_ifsioc above.
 */

/**
 *	dev_ioctl	-	network device ioctl
4594
 *	@net: the applicable net namespace
L
Linus Torvalds 已提交
4595 4596 4597 4598 4599 4600 4601 4602 4603
 *	@cmd: command to issue
 *	@arg: pointer to a struct ifreq in user space
 *
 *	Issue ioctl functions to devices. This is normally called by the
 *	user space syscall interfaces but can sometimes be useful for
 *	other purposes. The return value is the return from the syscall if
 *	positive or a negative errno code on error.
 */

4604
int dev_ioctl(struct net *net, unsigned int cmd, void __user *arg)
L
Linus Torvalds 已提交
4605 4606 4607 4608 4609 4610 4611 4612 4613 4614 4615
{
	struct ifreq ifr;
	int ret;
	char *colon;

	/* One special case: SIOCGIFCONF takes ifconf argument
	   and requires shared lock, because it sleeps writing
	   to user space.
	 */

	if (cmd == SIOCGIFCONF) {
4616
		rtnl_lock();
4617
		ret = dev_ifconf(net, (char __user *) arg);
4618
		rtnl_unlock();
L
Linus Torvalds 已提交
4619 4620 4621
		return ret;
	}
	if (cmd == SIOCGIFNAME)
4622
		return dev_ifname(net, (struct ifreq __user *)arg);
L
Linus Torvalds 已提交
4623 4624 4625 4626 4627 4628 4629 4630 4631 4632 4633 4634 4635 4636 4637

	if (copy_from_user(&ifr, arg, sizeof(struct ifreq)))
		return -EFAULT;

	ifr.ifr_name[IFNAMSIZ-1] = 0;

	colon = strchr(ifr.ifr_name, ':');
	if (colon)
		*colon = 0;

	/*
	 *	See which interface the caller is talking about.
	 */

	switch (cmd) {
E
Eric Dumazet 已提交
4638 4639 4640 4641 4642 4643 4644 4645 4646 4647 4648 4649 4650 4651 4652
	/*
	 *	These ioctl calls:
	 *	- can be done by all.
	 *	- atomic and do not require locking.
	 *	- return a value
	 */
	case SIOCGIFFLAGS:
	case SIOCGIFMETRIC:
	case SIOCGIFMTU:
	case SIOCGIFHWADDR:
	case SIOCGIFSLAVE:
	case SIOCGIFMAP:
	case SIOCGIFINDEX:
	case SIOCGIFTXQLEN:
		dev_load(net, ifr.ifr_name);
E
Eric Dumazet 已提交
4653
		rcu_read_lock();
E
Eric Dumazet 已提交
4654
		ret = dev_ifsioc_locked(net, &ifr, cmd);
E
Eric Dumazet 已提交
4655
		rcu_read_unlock();
E
Eric Dumazet 已提交
4656 4657 4658 4659 4660 4661 4662 4663
		if (!ret) {
			if (colon)
				*colon = ':';
			if (copy_to_user(arg, &ifr,
					 sizeof(struct ifreq)))
				ret = -EFAULT;
		}
		return ret;
L
Linus Torvalds 已提交
4664

E
Eric Dumazet 已提交
4665 4666 4667 4668 4669 4670 4671 4672 4673 4674 4675 4676 4677
	case SIOCETHTOOL:
		dev_load(net, ifr.ifr_name);
		rtnl_lock();
		ret = dev_ethtool(net, &ifr);
		rtnl_unlock();
		if (!ret) {
			if (colon)
				*colon = ':';
			if (copy_to_user(arg, &ifr,
					 sizeof(struct ifreq)))
				ret = -EFAULT;
		}
		return ret;
L
Linus Torvalds 已提交
4678

E
Eric Dumazet 已提交
4679 4680 4681 4682 4683 4684 4685 4686 4687 4688 4689 4690 4691 4692 4693 4694 4695 4696 4697 4698 4699 4700 4701
	/*
	 *	These ioctl calls:
	 *	- require superuser power.
	 *	- require strict serialization.
	 *	- return a value
	 */
	case SIOCGMIIPHY:
	case SIOCGMIIREG:
	case SIOCSIFNAME:
		if (!capable(CAP_NET_ADMIN))
			return -EPERM;
		dev_load(net, ifr.ifr_name);
		rtnl_lock();
		ret = dev_ifsioc(net, &ifr, cmd);
		rtnl_unlock();
		if (!ret) {
			if (colon)
				*colon = ':';
			if (copy_to_user(arg, &ifr,
					 sizeof(struct ifreq)))
				ret = -EFAULT;
		}
		return ret;
L
Linus Torvalds 已提交
4702

E
Eric Dumazet 已提交
4703 4704 4705 4706 4707 4708 4709 4710 4711 4712 4713 4714 4715 4716 4717 4718 4719 4720 4721 4722 4723 4724 4725 4726 4727 4728 4729 4730 4731 4732 4733 4734 4735 4736 4737 4738 4739 4740 4741 4742 4743 4744 4745 4746 4747 4748 4749 4750 4751 4752 4753
	/*
	 *	These ioctl calls:
	 *	- require superuser power.
	 *	- require strict serialization.
	 *	- do not return a value
	 */
	case SIOCSIFFLAGS:
	case SIOCSIFMETRIC:
	case SIOCSIFMTU:
	case SIOCSIFMAP:
	case SIOCSIFHWADDR:
	case SIOCSIFSLAVE:
	case SIOCADDMULTI:
	case SIOCDELMULTI:
	case SIOCSIFHWBROADCAST:
	case SIOCSIFTXQLEN:
	case SIOCSMIIREG:
	case SIOCBONDENSLAVE:
	case SIOCBONDRELEASE:
	case SIOCBONDSETHWADDR:
	case SIOCBONDCHANGEACTIVE:
	case SIOCBRADDIF:
	case SIOCBRDELIF:
	case SIOCSHWTSTAMP:
		if (!capable(CAP_NET_ADMIN))
			return -EPERM;
		/* fall through */
	case SIOCBONDSLAVEINFOQUERY:
	case SIOCBONDINFOQUERY:
		dev_load(net, ifr.ifr_name);
		rtnl_lock();
		ret = dev_ifsioc(net, &ifr, cmd);
		rtnl_unlock();
		return ret;

	case SIOCGIFMEM:
		/* Get the per device memory space. We can add this but
		 * currently do not support it */
	case SIOCSIFMEM:
		/* Set the per device memory buffer space.
		 * Not applicable in our case */
	case SIOCSIFLINK:
		return -EINVAL;

	/*
	 *	Unknown or private ioctl.
	 */
	default:
		if (cmd == SIOCWANDEV ||
		    (cmd >= SIOCDEVPRIVATE &&
		     cmd <= SIOCDEVPRIVATE + 15)) {
4754
			dev_load(net, ifr.ifr_name);
L
Linus Torvalds 已提交
4755
			rtnl_lock();
4756
			ret = dev_ifsioc(net, &ifr, cmd);
L
Linus Torvalds 已提交
4757
			rtnl_unlock();
E
Eric Dumazet 已提交
4758 4759 4760
			if (!ret && copy_to_user(arg, &ifr,
						 sizeof(struct ifreq)))
				ret = -EFAULT;
L
Linus Torvalds 已提交
4761
			return ret;
E
Eric Dumazet 已提交
4762 4763 4764 4765 4766
		}
		/* Take care of Wireless Extensions */
		if (cmd >= SIOCIWFIRST && cmd <= SIOCIWLAST)
			return wext_handle_ioctl(net, &ifr, cmd, arg);
		return -EINVAL;
L
Linus Torvalds 已提交
4767 4768 4769 4770 4771 4772
	}
}


/**
 *	dev_new_index	-	allocate an ifindex
4773
 *	@net: the applicable net namespace
L
Linus Torvalds 已提交
4774 4775 4776 4777 4778
 *
 *	Returns a suitable unique value for a new device interface
 *	number.  The caller must hold the rtnl semaphore or the
 *	dev_base_lock to be sure it remains unique.
 */
4779
static int dev_new_index(struct net *net)
L
Linus Torvalds 已提交
4780 4781 4782 4783 4784
{
	static int ifindex;
	for (;;) {
		if (++ifindex <= 0)
			ifindex = 1;
4785
		if (!__dev_get_by_index(net, ifindex))
L
Linus Torvalds 已提交
4786 4787 4788 4789 4790
			return ifindex;
	}
}

/* Delayed registration/unregisteration */
4791
static LIST_HEAD(net_todo_list);
L
Linus Torvalds 已提交
4792

4793
static void net_set_todo(struct net_device *dev)
L
Linus Torvalds 已提交
4794 4795 4796 4797
{
	list_add_tail(&dev->todo_list, &net_todo_list);
}

4798
static void rollback_registered_many(struct list_head *head)
4799
{
4800
	struct net_device *dev, *tmp;
4801

4802 4803 4804
	BUG_ON(dev_boot_phase);
	ASSERT_RTNL();

4805
	list_for_each_entry_safe(dev, tmp, head, unreg_list) {
4806
		/* Some devices call without registering
4807 4808
		 * for initialization unwind. Remove those
		 * devices and proceed with the remaining.
4809 4810 4811 4812
		 */
		if (dev->reg_state == NETREG_UNINITIALIZED) {
			pr_debug("unregister_netdevice: device %s/%p never "
				 "was registered\n", dev->name, dev);
4813

4814
			WARN_ON(1);
4815 4816
			list_del(&dev->unreg_list);
			continue;
4817
		}
4818

4819
		BUG_ON(dev->reg_state != NETREG_REGISTERED);
4820

4821 4822
		/* If device is running, close it first. */
		dev_close(dev);
4823

4824 4825
		/* And unlink it from device chain. */
		unlist_netdevice(dev);
4826

4827 4828
		dev->reg_state = NETREG_UNREGISTERING;
	}
4829 4830 4831

	synchronize_net();

4832 4833 4834
	list_for_each_entry(dev, head, unreg_list) {
		/* Shutdown queueing discipline. */
		dev_shutdown(dev);
4835 4836


4837 4838 4839 4840
		/* Notify protocols, that we are about to destroy
		   this device. They should clean all the things.
		*/
		call_netdevice_notifiers(NETDEV_UNREGISTER, dev);
4841

4842 4843 4844 4845
		if (!dev->rtnl_link_ops ||
		    dev->rtnl_link_state == RTNL_LINK_INITIALIZED)
			rtmsg_ifinfo(RTM_DELLINK, dev, ~0U);

4846 4847 4848
		/*
		 *	Flush the unicast and multicast chains
		 */
4849
		dev_uc_flush(dev);
4850
		dev_mc_flush(dev);
4851

4852 4853
		if (dev->netdev_ops->ndo_uninit)
			dev->netdev_ops->ndo_uninit(dev);
4854

4855 4856
		/* Notifier chain MUST detach us from master device. */
		WARN_ON(dev->master);
4857

4858 4859 4860
		/* Remove entries from kobject tree */
		netdev_unregister_kobject(dev);
	}
4861

4862
	/* Process any work delayed until the end of the batch */
4863
	dev = list_first_entry(head, struct net_device, unreg_list);
4864
	call_netdevice_notifiers(NETDEV_UNREGISTER_BATCH, dev);
4865

4866
	synchronize_net();
4867

4868
	list_for_each_entry(dev, head, unreg_list)
4869 4870 4871 4872 4873 4874 4875 4876 4877
		dev_put(dev);
}

static void rollback_registered(struct net_device *dev)
{
	LIST_HEAD(single);

	list_add(&dev->unreg_list, &single);
	rollback_registered_many(&single);
4878 4879
}

4880 4881 4882
static void __netdev_init_queue_locks_one(struct net_device *dev,
					  struct netdev_queue *dev_queue,
					  void *_unused)
4883 4884
{
	spin_lock_init(&dev_queue->_xmit_lock);
4885
	netdev_set_xmit_lockdep_class(&dev_queue->_xmit_lock, dev->type);
4886 4887 4888 4889 4890
	dev_queue->xmit_lock_owner = -1;
}

static void netdev_init_queue_locks(struct net_device *dev)
{
4891 4892
	netdev_for_each_tx_queue(dev, __netdev_init_queue_locks_one, NULL);
	__netdev_init_queue_locks_one(dev, &dev->rx_queue, NULL);
4893 4894
}

4895 4896 4897 4898 4899 4900 4901 4902 4903 4904 4905 4906 4907 4908 4909 4910 4911 4912 4913 4914 4915 4916 4917 4918 4919 4920 4921 4922 4923 4924 4925 4926 4927 4928 4929 4930 4931 4932 4933 4934
unsigned long netdev_fix_features(unsigned long features, const char *name)
{
	/* Fix illegal SG+CSUM combinations. */
	if ((features & NETIF_F_SG) &&
	    !(features & NETIF_F_ALL_CSUM)) {
		if (name)
			printk(KERN_NOTICE "%s: Dropping NETIF_F_SG since no "
			       "checksum feature.\n", name);
		features &= ~NETIF_F_SG;
	}

	/* TSO requires that SG is present as well. */
	if ((features & NETIF_F_TSO) && !(features & NETIF_F_SG)) {
		if (name)
			printk(KERN_NOTICE "%s: Dropping NETIF_F_TSO since no "
			       "SG feature.\n", name);
		features &= ~NETIF_F_TSO;
	}

	if (features & NETIF_F_UFO) {
		if (!(features & NETIF_F_GEN_CSUM)) {
			if (name)
				printk(KERN_ERR "%s: Dropping NETIF_F_UFO "
				       "since no NETIF_F_HW_CSUM feature.\n",
				       name);
			features &= ~NETIF_F_UFO;
		}

		if (!(features & NETIF_F_SG)) {
			if (name)
				printk(KERN_ERR "%s: Dropping NETIF_F_UFO "
				       "since no NETIF_F_SG feature.\n", name);
			features &= ~NETIF_F_UFO;
		}
	}

	return features;
}
EXPORT_SYMBOL(netdev_fix_features);

4935 4936 4937 4938 4939 4940 4941 4942 4943 4944 4945 4946 4947 4948 4949 4950 4951 4952 4953 4954 4955 4956 4957 4958 4959 4960 4961
/**
 *	netif_stacked_transfer_operstate -	transfer operstate
 *	@rootdev: the root or lower level device to transfer state from
 *	@dev: the device to transfer operstate to
 *
 *	Transfer operational state from root to device. This is normally
 *	called when a stacking relationship exists between the root
 *	device and the device(a leaf device).
 */
void netif_stacked_transfer_operstate(const struct net_device *rootdev,
					struct net_device *dev)
{
	if (rootdev->operstate == IF_OPER_DORMANT)
		netif_dormant_on(dev);
	else
		netif_dormant_off(dev);

	if (netif_carrier_ok(rootdev)) {
		if (!netif_carrier_ok(dev))
			netif_carrier_on(dev);
	} else {
		if (netif_carrier_ok(dev))
			netif_carrier_off(dev);
	}
}
EXPORT_SYMBOL(netif_stacked_transfer_operstate);

L
Linus Torvalds 已提交
4962 4963 4964 4965 4966 4967 4968 4969 4970 4971 4972 4973 4974 4975 4976 4977 4978 4979 4980 4981
/**
 *	register_netdevice	- register a network device
 *	@dev: device to register
 *
 *	Take a completed network device structure and add it to the kernel
 *	interfaces. A %NETDEV_REGISTER message is sent to the netdev notifier
 *	chain. 0 is returned on success. A negative errno code is returned
 *	on a failure to set up the device, or if the name is a duplicate.
 *
 *	Callers must hold the rtnl semaphore. You may want
 *	register_netdev() instead of this.
 *
 *	BUGS:
 *	The locking appears insufficient to guarantee two parallel registers
 *	will not get the same name.
 */

int register_netdevice(struct net_device *dev)
{
	int ret;
4982
	struct net *net = dev_net(dev);
L
Linus Torvalds 已提交
4983 4984 4985 4986

	BUG_ON(dev_boot_phase);
	ASSERT_RTNL();

4987 4988
	might_sleep();

L
Linus Torvalds 已提交
4989 4990
	/* When net_device's are persistent, this will be fatal. */
	BUG_ON(dev->reg_state != NETREG_UNINITIALIZED);
4991
	BUG_ON(!net);
L
Linus Torvalds 已提交
4992

4993
	spin_lock_init(&dev->addr_list_lock);
4994
	netdev_set_addr_lockdep_class(dev);
4995
	netdev_init_queue_locks(dev);
L
Linus Torvalds 已提交
4996 4997 4998

	dev->iflink = -1;

E
Eric Dumazet 已提交
4999
#ifdef CONFIG_RPS
T
Tom Herbert 已提交
5000 5001 5002 5003 5004 5005 5006 5007 5008 5009 5010 5011 5012 5013 5014 5015
	if (!dev->num_rx_queues) {
		/*
		 * Allocate a single RX queue if driver never called
		 * alloc_netdev_mq
		 */

		dev->_rx = kzalloc(sizeof(struct netdev_rx_queue), GFP_KERNEL);
		if (!dev->_rx) {
			ret = -ENOMEM;
			goto out;
		}

		dev->_rx->first = dev->_rx;
		atomic_set(&dev->_rx->count, 1);
		dev->num_rx_queues = 1;
	}
E
Eric Dumazet 已提交
5016
#endif
L
Linus Torvalds 已提交
5017
	/* Init, if this function is available */
5018 5019
	if (dev->netdev_ops->ndo_init) {
		ret = dev->netdev_ops->ndo_init(dev);
L
Linus Torvalds 已提交
5020 5021 5022
		if (ret) {
			if (ret > 0)
				ret = -EIO;
5023
			goto out;
L
Linus Torvalds 已提交
5024 5025
		}
	}
5026

5027
	ret = dev_get_valid_name(dev, dev->name, 0);
5028
	if (ret)
5029
		goto err_uninit;
L
Linus Torvalds 已提交
5030

5031
	dev->ifindex = dev_new_index(net);
L
Linus Torvalds 已提交
5032 5033 5034
	if (dev->iflink == -1)
		dev->iflink = dev->ifindex;

5035 5036 5037 5038 5039 5040 5041 5042 5043 5044 5045 5046 5047 5048 5049
	/* Fix illegal checksum combinations */
	if ((dev->features & NETIF_F_HW_CSUM) &&
	    (dev->features & (NETIF_F_IP_CSUM|NETIF_F_IPV6_CSUM))) {
		printk(KERN_NOTICE "%s: mixed HW and IP checksum settings.\n",
		       dev->name);
		dev->features &= ~(NETIF_F_IP_CSUM|NETIF_F_IPV6_CSUM);
	}

	if ((dev->features & NETIF_F_NO_CSUM) &&
	    (dev->features & (NETIF_F_HW_CSUM|NETIF_F_IP_CSUM|NETIF_F_IPV6_CSUM))) {
		printk(KERN_NOTICE "%s: mixed no checksumming and other settings.\n",
		       dev->name);
		dev->features &= ~(NETIF_F_IP_CSUM|NETIF_F_IPV6_CSUM|NETIF_F_HW_CSUM);
	}

5050
	dev->features = netdev_fix_features(dev->features, dev->name);
L
Linus Torvalds 已提交
5051

5052 5053 5054 5055
	/* Enable software GSO if SG is supported. */
	if (dev->features & NETIF_F_SG)
		dev->features |= NETIF_F_GSO;

5056 5057 5058 5059 5060
	ret = call_netdevice_notifiers(NETDEV_POST_INIT, dev);
	ret = notifier_to_errno(ret);
	if (ret)
		goto err_uninit;

5061
	ret = netdev_register_kobject(dev);
5062
	if (ret)
5063
		goto err_uninit;
5064 5065
	dev->reg_state = NETREG_REGISTERED;

L
Linus Torvalds 已提交
5066 5067 5068 5069 5070 5071 5072 5073 5074
	/*
	 *	Default initial state at registry is that the
	 *	device is present.
	 */

	set_bit(__LINK_STATE_PRESENT, &dev->state);

	dev_init_scheduler(dev);
	dev_hold(dev);
5075
	list_netdevice(dev);
L
Linus Torvalds 已提交
5076 5077

	/* Notify protocols, that a new device appeared. */
5078
	ret = call_netdevice_notifiers(NETDEV_REGISTER, dev);
5079
	ret = notifier_to_errno(ret);
5080 5081 5082 5083
	if (ret) {
		rollback_registered(dev);
		dev->reg_state = NETREG_UNREGISTERED;
	}
5084 5085 5086 5087
	/*
	 *	Prevent userspace races by waiting until the network
	 *	device is fully setup before sending notifications.
	 */
5088 5089 5090
	if (!dev->rtnl_link_ops ||
	    dev->rtnl_link_state == RTNL_LINK_INITIALIZED)
		rtmsg_ifinfo(RTM_NEWLINK, dev, ~0U);
L
Linus Torvalds 已提交
5091 5092 5093

out:
	return ret;
5094 5095

err_uninit:
5096 5097
	if (dev->netdev_ops->ndo_uninit)
		dev->netdev_ops->ndo_uninit(dev);
5098
	goto out;
L
Linus Torvalds 已提交
5099
}
E
Eric Dumazet 已提交
5100
EXPORT_SYMBOL(register_netdevice);
L
Linus Torvalds 已提交
5101

5102 5103 5104 5105 5106 5107 5108 5109 5110 5111 5112 5113 5114 5115 5116 5117 5118 5119 5120 5121 5122 5123 5124 5125 5126 5127 5128 5129 5130 5131 5132 5133 5134 5135 5136 5137 5138 5139 5140
/**
 *	init_dummy_netdev	- init a dummy network device for NAPI
 *	@dev: device to init
 *
 *	This takes a network device structure and initialize the minimum
 *	amount of fields so it can be used to schedule NAPI polls without
 *	registering a full blown interface. This is to be used by drivers
 *	that need to tie several hardware interfaces to a single NAPI
 *	poll scheduler due to HW limitations.
 */
int init_dummy_netdev(struct net_device *dev)
{
	/* Clear everything. Note we don't initialize spinlocks
	 * are they aren't supposed to be taken by any of the
	 * NAPI code and this dummy netdev is supposed to be
	 * only ever used for NAPI polls
	 */
	memset(dev, 0, sizeof(struct net_device));

	/* make sure we BUG if trying to hit standard
	 * register/unregister code path
	 */
	dev->reg_state = NETREG_DUMMY;

	/* initialize the ref count */
	atomic_set(&dev->refcnt, 1);

	/* NAPI wants this */
	INIT_LIST_HEAD(&dev->napi_list);

	/* a dummy interface is started by default */
	set_bit(__LINK_STATE_PRESENT, &dev->state);
	set_bit(__LINK_STATE_START, &dev->state);

	return 0;
}
EXPORT_SYMBOL_GPL(init_dummy_netdev);


L
Linus Torvalds 已提交
5141 5142 5143 5144 5145 5146 5147 5148 5149
/**
 *	register_netdev	- register a network device
 *	@dev: device to register
 *
 *	Take a completed network device structure and add it to the kernel
 *	interfaces. A %NETDEV_REGISTER message is sent to the netdev notifier
 *	chain. 0 is returned on success. A negative errno code is returned
 *	on a failure to set up the device, or if the name is a duplicate.
 *
5150
 *	This is a wrapper around register_netdevice that takes the rtnl semaphore
L
Linus Torvalds 已提交
5151 5152 5153 5154 5155 5156 5157 5158 5159 5160 5161 5162 5163 5164 5165 5166 5167 5168
 *	and expands the device name if you passed a format string to
 *	alloc_netdev.
 */
int register_netdev(struct net_device *dev)
{
	int err;

	rtnl_lock();

	/*
	 * If the name is a format string the caller wants us to do a
	 * name allocation.
	 */
	if (strchr(dev->name, '%')) {
		err = dev_alloc_name(dev, dev->name);
		if (err < 0)
			goto out;
	}
5169

L
Linus Torvalds 已提交
5170 5171 5172 5173 5174 5175 5176 5177 5178 5179 5180 5181 5182 5183 5184 5185
	err = register_netdevice(dev);
out:
	rtnl_unlock();
	return err;
}
EXPORT_SYMBOL(register_netdev);

/*
 * netdev_wait_allrefs - wait until all references are gone.
 *
 * This is called when unregistering network devices.
 *
 * Any protocol or device that holds a reference should register
 * for netdevice notification, and cleanup and put back the
 * reference if they receive an UNREGISTER event.
 * We can get stuck here if buggy protocols don't correctly
5186
 * call dev_put.
L
Linus Torvalds 已提交
5187 5188 5189 5190 5191
 */
static void netdev_wait_allrefs(struct net_device *dev)
{
	unsigned long rebroadcast_time, warning_time;

5192 5193
	linkwatch_forget_dev(dev);

L
Linus Torvalds 已提交
5194 5195 5196
	rebroadcast_time = warning_time = jiffies;
	while (atomic_read(&dev->refcnt) != 0) {
		if (time_after(jiffies, rebroadcast_time + 1 * HZ)) {
5197
			rtnl_lock();
L
Linus Torvalds 已提交
5198 5199

			/* Rebroadcast unregister notification */
5200
			call_netdevice_notifiers(NETDEV_UNREGISTER, dev);
5201
			/* don't resend NETDEV_UNREGISTER_BATCH, _BATCH users
5202
			 * should have already handle it the first time */
L
Linus Torvalds 已提交
5203 5204 5205 5206 5207 5208 5209 5210 5211 5212 5213 5214

			if (test_bit(__LINK_STATE_LINKWATCH_PENDING,
				     &dev->state)) {
				/* We must not have linkwatch events
				 * pending on unregister. If this
				 * happens, we simply run the queue
				 * unscheduled, resulting in a noop
				 * for this device.
				 */
				linkwatch_run_queue();
			}

5215
			__rtnl_unlock();
L
Linus Torvalds 已提交
5216 5217 5218 5219 5220 5221 5222 5223 5224 5225 5226 5227 5228 5229 5230 5231 5232 5233 5234 5235 5236 5237 5238 5239 5240 5241 5242 5243 5244 5245

			rebroadcast_time = jiffies;
		}

		msleep(250);

		if (time_after(jiffies, warning_time + 10 * HZ)) {
			printk(KERN_EMERG "unregister_netdevice: "
			       "waiting for %s to become free. Usage "
			       "count = %d\n",
			       dev->name, atomic_read(&dev->refcnt));
			warning_time = jiffies;
		}
	}
}

/* The sequence is:
 *
 *	rtnl_lock();
 *	...
 *	register_netdevice(x1);
 *	register_netdevice(x2);
 *	...
 *	unregister_netdevice(y1);
 *	unregister_netdevice(y2);
 *      ...
 *	rtnl_unlock();
 *	free_netdev(y1);
 *	free_netdev(y2);
 *
H
Herbert Xu 已提交
5246
 * We are invoked by rtnl_unlock().
L
Linus Torvalds 已提交
5247
 * This allows us to deal with problems:
5248
 * 1) We can delete sysfs objects which invoke hotplug
L
Linus Torvalds 已提交
5249 5250 5251
 *    without deadlocking with linkwatch via keventd.
 * 2) Since we run with the RTNL semaphore not held, we can sleep
 *    safely in order to wait for the netdev refcnt to drop to zero.
H
Herbert Xu 已提交
5252 5253 5254
 *
 * We must not return until all unregister events added during
 * the interval the lock was held have been completed.
L
Linus Torvalds 已提交
5255 5256 5257
 */
void netdev_run_todo(void)
{
5258
	struct list_head list;
L
Linus Torvalds 已提交
5259 5260

	/* Snapshot list, allow later requests */
5261
	list_replace_init(&net_todo_list, &list);
H
Herbert Xu 已提交
5262 5263

	__rtnl_unlock();
5264

L
Linus Torvalds 已提交
5265 5266
	while (!list_empty(&list)) {
		struct net_device *dev
5267
			= list_first_entry(&list, struct net_device, todo_list);
L
Linus Torvalds 已提交
5268 5269
		list_del(&dev->todo_list);

5270 5271 5272 5273 5274 5275
		if (unlikely(dev->reg_state != NETREG_UNREGISTERING)) {
			printk(KERN_ERR "network todo '%s' but state %d\n",
			       dev->name, dev->reg_state);
			dump_stack();
			continue;
		}
L
Linus Torvalds 已提交
5276

5277
		dev->reg_state = NETREG_UNREGISTERED;
L
Linus Torvalds 已提交
5278

5279
		on_each_cpu(flush_backlog, dev, 1);
5280

5281
		netdev_wait_allrefs(dev);
L
Linus Torvalds 已提交
5282

5283 5284
		/* paranoia */
		BUG_ON(atomic_read(&dev->refcnt));
5285 5286 5287
		WARN_ON(dev->ip_ptr);
		WARN_ON(dev->ip6_ptr);
		WARN_ON(dev->dn_ptr);
L
Linus Torvalds 已提交
5288

5289 5290
		if (dev->destructor)
			dev->destructor(dev);
5291 5292 5293

		/* Free network device */
		kobject_put(&dev->dev.kobj);
L
Linus Torvalds 已提交
5294 5295 5296
	}
}

5297 5298 5299
/**
 *	dev_txq_stats_fold - fold tx_queues stats
 *	@dev: device to get statistics from
5300
 *	@stats: struct rtnl_link_stats64 to hold results
5301 5302
 */
void dev_txq_stats_fold(const struct net_device *dev,
5303
			struct rtnl_link_stats64 *stats)
5304
{
E
Eric Dumazet 已提交
5305
	u64 tx_bytes = 0, tx_packets = 0, tx_dropped = 0;
5306 5307 5308 5309 5310
	unsigned int i;
	struct netdev_queue *txq;

	for (i = 0; i < dev->num_tx_queues; i++) {
		txq = netdev_get_tx_queue(dev, i);
E
Eric Dumazet 已提交
5311
		spin_lock_bh(&txq->_xmit_lock);
5312 5313 5314
		tx_bytes   += txq->tx_bytes;
		tx_packets += txq->tx_packets;
		tx_dropped += txq->tx_dropped;
E
Eric Dumazet 已提交
5315
		spin_unlock_bh(&txq->_xmit_lock);
5316 5317 5318 5319 5320 5321 5322 5323 5324
	}
	if (tx_bytes || tx_packets || tx_dropped) {
		stats->tx_bytes   = tx_bytes;
		stats->tx_packets = tx_packets;
		stats->tx_dropped = tx_dropped;
	}
}
EXPORT_SYMBOL(dev_txq_stats_fold);

5325 5326 5327 5328 5329 5330 5331 5332 5333 5334 5335 5336 5337 5338 5339 5340 5341 5342 5343 5344 5345
/* Convert net_device_stats to rtnl_link_stats64.  They have the same
 * fields in the same order, with only the type differing.
 */
static void netdev_stats_to_stats64(struct rtnl_link_stats64 *stats64,
				    const struct net_device_stats *netdev_stats)
{
#if BITS_PER_LONG == 64
        BUILD_BUG_ON(sizeof(*stats64) != sizeof(*netdev_stats));
        memcpy(stats64, netdev_stats, sizeof(*stats64));
#else
	size_t i, n = sizeof(*stats64) / sizeof(u64);
	const unsigned long *src = (const unsigned long *)netdev_stats;
	u64 *dst = (u64 *)stats64;

	BUILD_BUG_ON(sizeof(*netdev_stats) / sizeof(unsigned long) !=
		     sizeof(*stats64) / sizeof(u64));
	for (i = 0; i < n; i++)
		dst[i] = src[i];
#endif
}

5346 5347 5348
/**
 *	dev_get_stats	- get network device statistics
 *	@dev: device to get statistics from
5349
 *	@storage: place to store stats
5350
 *
5351 5352 5353 5354
 *	Get network statistics from device. Return @storage.
 *	The device driver may provide its own method by setting
 *	dev->netdev_ops->get_stats64 or dev->netdev_ops->get_stats;
 *	otherwise the internal statistics structure is used.
5355
 */
5356 5357
struct rtnl_link_stats64 *dev_get_stats(struct net_device *dev,
					struct rtnl_link_stats64 *storage)
5358
{
5359 5360
	const struct net_device_ops *ops = dev->netdev_ops;

5361 5362 5363 5364 5365
	if (ops->ndo_get_stats64) {
		memset(storage, 0, sizeof(*storage));
		return ops->ndo_get_stats64(dev, storage);
	}
	if (ops->ndo_get_stats) {
5366
		netdev_stats_to_stats64(storage, ops->ndo_get_stats(dev));
5367 5368
		return storage;
	}
5369 5370
	netdev_stats_to_stats64(storage, &dev->stats);
	dev_txq_stats_fold(dev, storage);
5371
	return storage;
R
Rusty Russell 已提交
5372
}
5373
EXPORT_SYMBOL(dev_get_stats);
R
Rusty Russell 已提交
5374

5375
static void netdev_init_one_queue(struct net_device *dev,
5376 5377
				  struct netdev_queue *queue,
				  void *_unused)
5378 5379 5380 5381
{
	queue->dev = dev;
}

5382 5383
static void netdev_init_queues(struct net_device *dev)
{
5384 5385
	netdev_init_one_queue(dev, &dev->rx_queue, NULL);
	netdev_for_each_tx_queue(dev, netdev_init_one_queue, NULL);
5386
	spin_lock_init(&dev->tx_global_lock);
5387 5388
}

L
Linus Torvalds 已提交
5389
/**
5390
 *	alloc_netdev_mq - allocate network device
L
Linus Torvalds 已提交
5391 5392 5393
 *	@sizeof_priv:	size of private data to allocate space for
 *	@name:		device name format string
 *	@setup:		callback to initialize device
5394
 *	@queue_count:	the number of subqueues to allocate
L
Linus Torvalds 已提交
5395 5396
 *
 *	Allocates a struct net_device with private data area for driver use
5397 5398
 *	and performs basic initialization.  Also allocates subquue structs
 *	for each queue on the device at the end of the netdevice.
L
Linus Torvalds 已提交
5399
 */
5400 5401
struct net_device *alloc_netdev_mq(int sizeof_priv, const char *name,
		void (*setup)(struct net_device *), unsigned int queue_count)
L
Linus Torvalds 已提交
5402
{
5403
	struct netdev_queue *tx;
L
Linus Torvalds 已提交
5404
	struct net_device *dev;
5405
	size_t alloc_size;
5406
	struct net_device *p;
E
Eric Dumazet 已提交
5407 5408
#ifdef CONFIG_RPS
	struct netdev_rx_queue *rx;
T
Tom Herbert 已提交
5409
	int i;
E
Eric Dumazet 已提交
5410
#endif
L
Linus Torvalds 已提交
5411

5412 5413
	BUG_ON(strlen(name) >= sizeof(dev->name));

5414
	alloc_size = sizeof(struct net_device);
5415 5416
	if (sizeof_priv) {
		/* ensure 32-byte alignment of private area */
5417
		alloc_size = ALIGN(alloc_size, NETDEV_ALIGN);
5418 5419 5420
		alloc_size += sizeof_priv;
	}
	/* ensure 32-byte alignment of whole construct */
5421
	alloc_size += NETDEV_ALIGN - 1;
L
Linus Torvalds 已提交
5422

5423
	p = kzalloc(alloc_size, GFP_KERNEL);
L
Linus Torvalds 已提交
5424
	if (!p) {
5425
		printk(KERN_ERR "alloc_netdev: Unable to allocate device.\n");
L
Linus Torvalds 已提交
5426 5427 5428
		return NULL;
	}

5429
	tx = kcalloc(queue_count, sizeof(struct netdev_queue), GFP_KERNEL);
5430 5431 5432
	if (!tx) {
		printk(KERN_ERR "alloc_netdev: Unable to allocate "
		       "tx qdiscs.\n");
5433
		goto free_p;
5434 5435
	}

E
Eric Dumazet 已提交
5436
#ifdef CONFIG_RPS
T
Tom Herbert 已提交
5437 5438 5439 5440 5441 5442 5443 5444 5445 5446 5447 5448 5449 5450 5451
	rx = kcalloc(queue_count, sizeof(struct netdev_rx_queue), GFP_KERNEL);
	if (!rx) {
		printk(KERN_ERR "alloc_netdev: Unable to allocate "
		       "rx queues.\n");
		goto free_tx;
	}

	atomic_set(&rx->count, queue_count);

	/*
	 * Set a pointer to first element in the array which holds the
	 * reference count.
	 */
	for (i = 0; i < queue_count; i++)
		rx[i].first = rx;
E
Eric Dumazet 已提交
5452
#endif
T
Tom Herbert 已提交
5453

5454
	dev = PTR_ALIGN(p, NETDEV_ALIGN);
L
Linus Torvalds 已提交
5455
	dev->padded = (char *)dev - (char *)p;
5456 5457

	if (dev_addr_init(dev))
T
Tom Herbert 已提交
5458
		goto free_rx;
5459

5460
	dev_mc_init(dev);
5461
	dev_uc_init(dev);
J
Jiri Pirko 已提交
5462

5463
	dev_net_set(dev, &init_net);
L
Linus Torvalds 已提交
5464

5465 5466
	dev->_tx = tx;
	dev->num_tx_queues = queue_count;
5467
	dev->real_num_tx_queues = queue_count;
5468

E
Eric Dumazet 已提交
5469
#ifdef CONFIG_RPS
T
Tom Herbert 已提交
5470 5471
	dev->_rx = rx;
	dev->num_rx_queues = queue_count;
E
Eric Dumazet 已提交
5472
#endif
T
Tom Herbert 已提交
5473

5474
	dev->gso_max_size = GSO_MAX_SIZE;
L
Linus Torvalds 已提交
5475

5476 5477
	netdev_init_queues(dev);

5478 5479
	INIT_LIST_HEAD(&dev->ethtool_ntuple_list.list);
	dev->ethtool_ntuple_list.count = 0;
5480
	INIT_LIST_HEAD(&dev->napi_list);
5481
	INIT_LIST_HEAD(&dev->unreg_list);
5482
	INIT_LIST_HEAD(&dev->link_watch_list);
5483
	dev->priv_flags = IFF_XMIT_DST_RELEASE;
L
Linus Torvalds 已提交
5484 5485 5486
	setup(dev);
	strcpy(dev->name, name);
	return dev;
5487

T
Tom Herbert 已提交
5488
free_rx:
E
Eric Dumazet 已提交
5489
#ifdef CONFIG_RPS
T
Tom Herbert 已提交
5490
	kfree(rx);
5491
free_tx:
E
Eric Dumazet 已提交
5492
#endif
5493 5494 5495 5496
	kfree(tx);
free_p:
	kfree(p);
	return NULL;
L
Linus Torvalds 已提交
5497
}
5498
EXPORT_SYMBOL(alloc_netdev_mq);
L
Linus Torvalds 已提交
5499 5500 5501 5502 5503

/**
 *	free_netdev - free network device
 *	@dev: device
 *
5504 5505
 *	This function does the last stage of destroying an allocated device
 * 	interface. The reference to the device object is released.
L
Linus Torvalds 已提交
5506 5507 5508 5509
 *	If this is the last reference then it will be freed.
 */
void free_netdev(struct net_device *dev)
{
5510 5511
	struct napi_struct *p, *n;

5512 5513
	release_net(dev_net(dev));

5514 5515
	kfree(dev->_tx);

5516 5517 5518
	/* Flush device addresses */
	dev_addr_flush(dev);

5519 5520 5521
	/* Clear ethtool n-tuple list */
	ethtool_ntuple_flush(dev);

5522 5523 5524
	list_for_each_entry_safe(p, n, &dev->napi_list, dev_list)
		netif_napi_del(p);

S
Stephen Hemminger 已提交
5525
	/*  Compatibility with error handling in drivers */
L
Linus Torvalds 已提交
5526 5527 5528 5529 5530 5531 5532 5533
	if (dev->reg_state == NETREG_UNINITIALIZED) {
		kfree((char *)dev - dev->padded);
		return;
	}

	BUG_ON(dev->reg_state != NETREG_UNREGISTERED);
	dev->reg_state = NETREG_RELEASED;

5534 5535
	/* will free via device release */
	put_device(&dev->dev);
L
Linus Torvalds 已提交
5536
}
E
Eric Dumazet 已提交
5537
EXPORT_SYMBOL(free_netdev);
5538

5539 5540 5541 5542 5543 5544
/**
 *	synchronize_net -  Synchronize with packet receive processing
 *
 *	Wait for packets currently being received to be done.
 *	Does not block later packets from starting.
 */
5545
void synchronize_net(void)
L
Linus Torvalds 已提交
5546 5547
{
	might_sleep();
5548
	synchronize_rcu();
L
Linus Torvalds 已提交
5549
}
E
Eric Dumazet 已提交
5550
EXPORT_SYMBOL(synchronize_net);
L
Linus Torvalds 已提交
5551 5552

/**
5553
 *	unregister_netdevice_queue - remove device from the kernel
L
Linus Torvalds 已提交
5554
 *	@dev: device
5555
 *	@head: list
5556
 *
L
Linus Torvalds 已提交
5557
 *	This function shuts down a device interface and removes it
5558
 *	from the kernel tables.
5559
 *	If head not NULL, device is queued to be unregistered later.
L
Linus Torvalds 已提交
5560 5561 5562 5563 5564
 *
 *	Callers must hold the rtnl semaphore.  You may want
 *	unregister_netdev() instead of this.
 */

5565
void unregister_netdevice_queue(struct net_device *dev, struct list_head *head)
L
Linus Torvalds 已提交
5566
{
5567 5568
	ASSERT_RTNL();

5569
	if (head) {
5570
		list_move_tail(&dev->unreg_list, head);
5571 5572 5573 5574 5575
	} else {
		rollback_registered(dev);
		/* Finish processing unregister after unlock */
		net_set_todo(dev);
	}
L
Linus Torvalds 已提交
5576
}
5577
EXPORT_SYMBOL(unregister_netdevice_queue);
L
Linus Torvalds 已提交
5578

5579 5580 5581 5582 5583 5584 5585 5586 5587 5588 5589 5590 5591 5592
/**
 *	unregister_netdevice_many - unregister many devices
 *	@head: list of devices
 */
void unregister_netdevice_many(struct list_head *head)
{
	struct net_device *dev;

	if (!list_empty(head)) {
		rollback_registered_many(head);
		list_for_each_entry(dev, head, unreg_list)
			net_set_todo(dev);
	}
}
5593
EXPORT_SYMBOL(unregister_netdevice_many);
5594

L
Linus Torvalds 已提交
5595 5596 5597 5598 5599
/**
 *	unregister_netdev - remove device from the kernel
 *	@dev: device
 *
 *	This function shuts down a device interface and removes it
5600
 *	from the kernel tables.
L
Linus Torvalds 已提交
5601 5602 5603 5604 5605 5606 5607 5608 5609 5610 5611 5612 5613
 *
 *	This is just a wrapper for unregister_netdevice that takes
 *	the rtnl semaphore.  In general you want to use this and not
 *	unregister_netdevice.
 */
void unregister_netdev(struct net_device *dev)
{
	rtnl_lock();
	unregister_netdevice(dev);
	rtnl_unlock();
}
EXPORT_SYMBOL(unregister_netdev);

5614 5615 5616 5617 5618 5619 5620 5621 5622 5623 5624 5625 5626 5627 5628 5629 5630 5631 5632 5633 5634 5635 5636 5637 5638 5639 5640 5641 5642 5643 5644 5645
/**
 *	dev_change_net_namespace - move device to different nethost namespace
 *	@dev: device
 *	@net: network namespace
 *	@pat: If not NULL name pattern to try if the current device name
 *	      is already taken in the destination network namespace.
 *
 *	This function shuts down a device interface and moves it
 *	to a new network namespace. On success 0 is returned, on
 *	a failure a netagive errno code is returned.
 *
 *	Callers must hold the rtnl semaphore.
 */

int dev_change_net_namespace(struct net_device *dev, struct net *net, const char *pat)
{
	int err;

	ASSERT_RTNL();

	/* Don't allow namespace local devices to be moved. */
	err = -EINVAL;
	if (dev->features & NETIF_F_NETNS_LOCAL)
		goto out;

	/* Ensure the device has been registrered */
	err = -EINVAL;
	if (dev->reg_state != NETREG_REGISTERED)
		goto out;

	/* Get out if there is nothing todo */
	err = 0;
5646
	if (net_eq(dev_net(dev), net))
5647 5648 5649 5650 5651 5652
		goto out;

	/* Pick the destination device name, and ensure
	 * we can use it in the destination network namespace.
	 */
	err = -EEXIST;
5653
	if (__dev_get_by_name(net, dev->name)) {
5654 5655 5656
		/* We get here if we can't use the current device name */
		if (!pat)
			goto out;
5657
		if (dev_get_valid_name(dev, pat, 1))
5658 5659 5660 5661 5662 5663 5664 5665
			goto out;
	}

	/*
	 * And now a mini version of register_netdevice unregister_netdevice.
	 */

	/* If device is running close it first. */
5666
	dev_close(dev);
5667 5668 5669 5670 5671 5672 5673 5674 5675 5676 5677 5678 5679 5680

	/* And unlink it from device chain */
	err = -ENODEV;
	unlist_netdevice(dev);

	synchronize_net();

	/* Shutdown queueing discipline. */
	dev_shutdown(dev);

	/* Notify protocols, that we are about to destroy
	   this device. They should clean all the things.
	*/
	call_netdevice_notifiers(NETDEV_UNREGISTER, dev);
5681
	call_netdevice_notifiers(NETDEV_UNREGISTER_BATCH, dev);
5682 5683 5684 5685

	/*
	 *	Flush the unicast and multicast chains
	 */
5686
	dev_uc_flush(dev);
5687
	dev_mc_flush(dev);
5688 5689

	/* Actually switch the network namespace */
5690
	dev_net_set(dev, net);
5691 5692 5693 5694 5695 5696 5697 5698 5699

	/* If there is an ifindex conflict assign a new one */
	if (__dev_get_by_index(net, dev->ifindex)) {
		int iflink = (dev->iflink == dev->ifindex);
		dev->ifindex = dev_new_index(net);
		if (iflink)
			dev->iflink = dev->ifindex;
	}

5700
	/* Fixup kobjects */
5701
	err = device_rename(&dev->dev, dev->name);
5702
	WARN_ON(err);
5703 5704 5705 5706 5707 5708 5709

	/* Add the device back in the hashes */
	list_netdevice(dev);

	/* Notify protocols, that a new device appeared. */
	call_netdevice_notifiers(NETDEV_REGISTER, dev);

5710 5711 5712 5713 5714 5715
	/*
	 *	Prevent userspace races by waiting until the network
	 *	device is fully setup before sending notifications.
	 */
	rtmsg_ifinfo(RTM_NEWLINK, dev, ~0U);

5716 5717 5718 5719 5720
	synchronize_net();
	err = 0;
out:
	return err;
}
5721
EXPORT_SYMBOL_GPL(dev_change_net_namespace);
5722

L
Linus Torvalds 已提交
5723 5724 5725 5726 5727 5728 5729 5730 5731
static int dev_cpu_callback(struct notifier_block *nfb,
			    unsigned long action,
			    void *ocpu)
{
	struct sk_buff **list_skb;
	struct sk_buff *skb;
	unsigned int cpu, oldcpu = (unsigned long)ocpu;
	struct softnet_data *sd, *oldsd;

5732
	if (action != CPU_DEAD && action != CPU_DEAD_FROZEN)
L
Linus Torvalds 已提交
5733 5734 5735 5736 5737 5738 5739 5740 5741 5742 5743 5744 5745 5746 5747 5748
		return NOTIFY_OK;

	local_irq_disable();
	cpu = smp_processor_id();
	sd = &per_cpu(softnet_data, cpu);
	oldsd = &per_cpu(softnet_data, oldcpu);

	/* Find end of our completion_queue. */
	list_skb = &sd->completion_queue;
	while (*list_skb)
		list_skb = &(*list_skb)->next;
	/* Append completion queue from offline CPU. */
	*list_skb = oldsd->completion_queue;
	oldsd->completion_queue = NULL;

	/* Append output queue from offline CPU. */
5749 5750 5751 5752 5753 5754
	if (oldsd->output_queue) {
		*sd->output_queue_tailp = oldsd->output_queue;
		sd->output_queue_tailp = oldsd->output_queue_tailp;
		oldsd->output_queue = NULL;
		oldsd->output_queue_tailp = &oldsd->output_queue;
	}
L
Linus Torvalds 已提交
5755 5756 5757 5758 5759

	raise_softirq_irqoff(NET_TX_SOFTIRQ);
	local_irq_enable();

	/* Process offline CPU's input_pkt_queue */
5760
	while ((skb = __skb_dequeue(&oldsd->process_queue))) {
L
Linus Torvalds 已提交
5761
		netif_rx(skb);
5762
		input_queue_head_incr(oldsd);
T
Tom Herbert 已提交
5763
	}
5764
	while ((skb = __skb_dequeue(&oldsd->input_pkt_queue))) {
5765
		netif_rx(skb);
5766 5767
		input_queue_head_incr(oldsd);
	}
L
Linus Torvalds 已提交
5768 5769 5770 5771 5772

	return NOTIFY_OK;
}


5773
/**
5774 5775 5776 5777
 *	netdev_increment_features - increment feature set by one
 *	@all: current feature set
 *	@one: new feature set
 *	@mask: mask feature set
5778 5779
 *
 *	Computes a new feature set after adding a device with feature set
5780 5781
 *	@one to the master device with current feature set @all.  Will not
 *	enable anything that is off in @mask. Returns the new feature set.
5782
 */
5783 5784 5785 5786
unsigned long netdev_increment_features(unsigned long all, unsigned long one,
					unsigned long mask)
{
	/* If device needs checksumming, downgrade to it. */
E
Eric Dumazet 已提交
5787
	if (all & NETIF_F_NO_CSUM && !(one & NETIF_F_NO_CSUM))
5788 5789 5790 5791 5792 5793 5794 5795
		all ^= NETIF_F_NO_CSUM | (one & NETIF_F_ALL_CSUM);
	else if (mask & NETIF_F_ALL_CSUM) {
		/* If one device supports v4/v6 checksumming, set for all. */
		if (one & (NETIF_F_IP_CSUM | NETIF_F_IPV6_CSUM) &&
		    !(all & NETIF_F_GEN_CSUM)) {
			all &= ~NETIF_F_ALL_CSUM;
			all |= one & (NETIF_F_IP_CSUM | NETIF_F_IPV6_CSUM);
		}
5796

5797 5798 5799 5800 5801 5802
		/* If one device supports hw checksumming, set for all. */
		if (one & NETIF_F_GEN_CSUM && !(all & NETIF_F_GEN_CSUM)) {
			all &= ~NETIF_F_ALL_CSUM;
			all |= NETIF_F_HW_CSUM;
		}
	}
5803

5804
	one |= NETIF_F_ALL_CSUM;
5805

5806
	one |= all & NETIF_F_ONE_FOR_ALL;
5807
	all &= one | NETIF_F_LLTX | NETIF_F_GSO | NETIF_F_UFO;
5808
	all |= one & mask & NETIF_F_ONE_FOR_ALL;
5809 5810 5811

	return all;
}
5812
EXPORT_SYMBOL(netdev_increment_features);
5813

5814 5815 5816 5817 5818 5819 5820 5821 5822 5823 5824 5825 5826
static struct hlist_head *netdev_create_hash(void)
{
	int i;
	struct hlist_head *hash;

	hash = kmalloc(sizeof(*hash) * NETDEV_HASHENTRIES, GFP_KERNEL);
	if (hash != NULL)
		for (i = 0; i < NETDEV_HASHENTRIES; i++)
			INIT_HLIST_HEAD(&hash[i]);

	return hash;
}

5827
/* Initialize per network namespace state */
5828
static int __net_init netdev_init(struct net *net)
5829 5830 5831
{
	INIT_LIST_HEAD(&net->dev_base_head);

5832 5833 5834
	net->dev_name_head = netdev_create_hash();
	if (net->dev_name_head == NULL)
		goto err_name;
5835

5836 5837 5838
	net->dev_index_head = netdev_create_hash();
	if (net->dev_index_head == NULL)
		goto err_idx;
5839 5840

	return 0;
5841 5842 5843 5844 5845

err_idx:
	kfree(net->dev_name_head);
err_name:
	return -ENOMEM;
5846 5847
}

5848 5849 5850 5851 5852 5853 5854 5855
/**
 *	netdev_drivername - network driver for the device
 *	@dev: network device
 *	@buffer: buffer for resulting name
 *	@len: size of buffer
 *
 *	Determine network driver for device.
 */
5856
char *netdev_drivername(const struct net_device *dev, char *buffer, int len)
5857
{
5858 5859
	const struct device_driver *driver;
	const struct device *parent;
5860 5861 5862 5863 5864 5865 5866 5867 5868 5869 5870 5871 5872 5873 5874 5875

	if (len <= 0 || !buffer)
		return buffer;
	buffer[0] = 0;

	parent = dev->dev.parent;

	if (!parent)
		return buffer;

	driver = parent->driver;
	if (driver && driver->name)
		strlcpy(buffer, driver->name, len);
	return buffer;
}

5876 5877 5878 5879 5880 5881 5882 5883 5884 5885 5886 5887 5888 5889 5890 5891 5892 5893 5894 5895 5896 5897 5898 5899 5900 5901 5902 5903 5904 5905 5906 5907 5908 5909 5910 5911 5912 5913 5914 5915 5916 5917 5918 5919 5920 5921 5922 5923 5924 5925 5926 5927 5928 5929 5930 5931 5932 5933 5934 5935 5936 5937
static int __netdev_printk(const char *level, const struct net_device *dev,
			   struct va_format *vaf)
{
	int r;

	if (dev && dev->dev.parent)
		r = dev_printk(level, dev->dev.parent, "%s: %pV",
			       netdev_name(dev), vaf);
	else if (dev)
		r = printk("%s%s: %pV", level, netdev_name(dev), vaf);
	else
		r = printk("%s(NULL net_device): %pV", level, vaf);

	return r;
}

int netdev_printk(const char *level, const struct net_device *dev,
		  const char *format, ...)
{
	struct va_format vaf;
	va_list args;
	int r;

	va_start(args, format);

	vaf.fmt = format;
	vaf.va = &args;

	r = __netdev_printk(level, dev, &vaf);
	va_end(args);

	return r;
}
EXPORT_SYMBOL(netdev_printk);

#define define_netdev_printk_level(func, level)			\
int func(const struct net_device *dev, const char *fmt, ...)	\
{								\
	int r;							\
	struct va_format vaf;					\
	va_list args;						\
								\
	va_start(args, fmt);					\
								\
	vaf.fmt = fmt;						\
	vaf.va = &args;						\
								\
	r = __netdev_printk(level, dev, &vaf);			\
	va_end(args);						\
								\
	return r;						\
}								\
EXPORT_SYMBOL(func);

define_netdev_printk_level(netdev_emerg, KERN_EMERG);
define_netdev_printk_level(netdev_alert, KERN_ALERT);
define_netdev_printk_level(netdev_crit, KERN_CRIT);
define_netdev_printk_level(netdev_err, KERN_ERR);
define_netdev_printk_level(netdev_warn, KERN_WARNING);
define_netdev_printk_level(netdev_notice, KERN_NOTICE);
define_netdev_printk_level(netdev_info, KERN_INFO);

5938
static void __net_exit netdev_exit(struct net *net)
5939 5940 5941 5942 5943
{
	kfree(net->dev_name_head);
	kfree(net->dev_index_head);
}

5944
static struct pernet_operations __net_initdata netdev_net_ops = {
5945 5946 5947 5948
	.init = netdev_init,
	.exit = netdev_exit,
};

5949
static void __net_exit default_device_exit(struct net *net)
5950
{
5951
	struct net_device *dev, *aux;
5952
	/*
5953
	 * Push all migratable network devices back to the
5954 5955 5956
	 * initial network namespace
	 */
	rtnl_lock();
5957
	for_each_netdev_safe(net, dev, aux) {
5958
		int err;
5959
		char fb_name[IFNAMSIZ];
5960 5961 5962 5963 5964

		/* Ignore unmoveable devices (i.e. loopback) */
		if (dev->features & NETIF_F_NETNS_LOCAL)
			continue;

5965 5966 5967
		/* Leave virtual devices for the generic cleanup */
		if (dev->rtnl_link_ops)
			continue;
5968

5969
		/* Push remaing network devices to init_net */
5970 5971
		snprintf(fb_name, IFNAMSIZ, "dev%d", dev->ifindex);
		err = dev_change_net_namespace(dev, &init_net, fb_name);
5972
		if (err) {
5973
			printk(KERN_EMERG "%s: failed to move %s to init_net: %d\n",
5974
				__func__, dev->name, err);
5975
			BUG();
5976 5977 5978 5979 5980
		}
	}
	rtnl_unlock();
}

5981 5982 5983 5984 5985 5986 5987 5988 5989 5990 5991 5992 5993 5994 5995 5996 5997 5998 5999 6000 6001 6002 6003 6004
static void __net_exit default_device_exit_batch(struct list_head *net_list)
{
	/* At exit all network devices most be removed from a network
	 * namespace.  Do this in the reverse order of registeration.
	 * Do this across as many network namespaces as possible to
	 * improve batching efficiency.
	 */
	struct net_device *dev;
	struct net *net;
	LIST_HEAD(dev_kill_list);

	rtnl_lock();
	list_for_each_entry(net, net_list, exit_list) {
		for_each_netdev_reverse(net, dev) {
			if (dev->rtnl_link_ops)
				dev->rtnl_link_ops->dellink(dev, &dev_kill_list);
			else
				unregister_netdevice_queue(dev, &dev_kill_list);
		}
	}
	unregister_netdevice_many(&dev_kill_list);
	rtnl_unlock();
}

6005
static struct pernet_operations __net_initdata default_device_ops = {
6006
	.exit = default_device_exit,
6007
	.exit_batch = default_device_exit_batch,
6008 6009
};

L
Linus Torvalds 已提交
6010 6011 6012 6013 6014 6015 6016 6017 6018 6019 6020 6021 6022 6023 6024 6025 6026 6027 6028 6029
/*
 *	Initialize the DEV module. At boot time this walks the device list and
 *	unhooks any devices that fail to initialise (normally hardware not
 *	present) and leaves us with a valid list of present and active devices.
 *
 */

/*
 *       This is called single threaded during boot, so no need
 *       to take the rtnl semaphore.
 */
static int __init net_dev_init(void)
{
	int i, rc = -ENOMEM;

	BUG_ON(!dev_boot_phase);

	if (dev_proc_init())
		goto out;

6030
	if (netdev_kobject_init())
L
Linus Torvalds 已提交
6031 6032 6033
		goto out;

	INIT_LIST_HEAD(&ptype_all);
6034
	for (i = 0; i < PTYPE_HASH_SIZE; i++)
L
Linus Torvalds 已提交
6035 6036
		INIT_LIST_HEAD(&ptype_base[i]);

6037 6038
	if (register_pernet_subsys(&netdev_net_ops))
		goto out;
L
Linus Torvalds 已提交
6039 6040 6041 6042 6043

	/*
	 *	Initialise the packet receive queues.
	 */

6044
	for_each_possible_cpu(i) {
E
Eric Dumazet 已提交
6045
		struct softnet_data *sd = &per_cpu(softnet_data, i);
L
Linus Torvalds 已提交
6046

C
Changli Gao 已提交
6047
		memset(sd, 0, sizeof(*sd));
E
Eric Dumazet 已提交
6048
		skb_queue_head_init(&sd->input_pkt_queue);
6049
		skb_queue_head_init(&sd->process_queue);
E
Eric Dumazet 已提交
6050 6051
		sd->completion_queue = NULL;
		INIT_LIST_HEAD(&sd->poll_list);
6052 6053
		sd->output_queue = NULL;
		sd->output_queue_tailp = &sd->output_queue;
E
Eric Dumazet 已提交
6054
#ifdef CONFIG_RPS
E
Eric Dumazet 已提交
6055 6056 6057 6058
		sd->csd.func = rps_trigger_softirq;
		sd->csd.info = sd;
		sd->csd.flags = 0;
		sd->cpu = i;
6059
#endif
T
Tom Herbert 已提交
6060

E
Eric Dumazet 已提交
6061 6062 6063 6064
		sd->backlog.poll = process_backlog;
		sd->backlog.weight = weight_p;
		sd->backlog.gro_list = NULL;
		sd->backlog.gro_count = 0;
L
Linus Torvalds 已提交
6065 6066 6067 6068
	}

	dev_boot_phase = 0;

6069 6070 6071 6072 6073 6074 6075 6076 6077 6078 6079 6080 6081 6082 6083
	/* The loopback device is special if any other network devices
	 * is present in a network namespace the loopback device must
	 * be present. Since we now dynamically allocate and free the
	 * loopback device ensure this invariant is maintained by
	 * keeping the loopback device as the first device on the
	 * list of network devices.  Ensuring the loopback devices
	 * is the first device that appears and the last network device
	 * that disappears.
	 */
	if (register_pernet_device(&loopback_net_ops))
		goto out;

	if (register_pernet_device(&default_device_ops))
		goto out;

6084 6085
	open_softirq(NET_TX_SOFTIRQ, net_tx_action);
	open_softirq(NET_RX_SOFTIRQ, net_rx_action);
L
Linus Torvalds 已提交
6086 6087 6088 6089 6090 6091 6092 6093 6094 6095 6096

	hotcpu_notifier(dev_cpu_callback, 0);
	dst_init();
	dev_mcast_init();
	rc = 0;
out:
	return rc;
}

subsys_initcall(net_dev_init);

6097 6098
static int __init initialize_hashrnd(void)
{
T
Tom Herbert 已提交
6099
	get_random_bytes(&hashrnd, sizeof(hashrnd));
6100 6101 6102 6103 6104
	return 0;
}

late_initcall_sync(initialize_hashrnd);