dev.c 142.2 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
#include <net/sock.h>
#include <linux/rtnetlink.h>
#include <linux/proc_fs.h>
#include <linux/seq_file.h>
#include <linux/stat.h>
#include <linux/if_bridge.h>
P
Patrick McHardy 已提交
105
#include <linux/if_macvlan.h>
L
Linus Torvalds 已提交
106 107 108
#include <net/dst.h>
#include <net/pkt_sched.h>
#include <net/checksum.h>
109
#include <net/xfrm.h>
L
Linus Torvalds 已提交
110 111 112 113 114 115 116
#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>
117
#include <net/wext.h>
L
Linus Torvalds 已提交
118 119
#include <net/iw_handler.h>
#include <asm/current.h>
S
Steve Grubb 已提交
120
#include <linux/audit.h>
121
#include <linux/dmaengine.h>
122
#include <linux/err.h>
123
#include <linux/ctype.h>
124
#include <linux/if_arp.h>
125
#include <linux/if_vlan.h>
126
#include <linux/ip.h>
127
#include <net/ip.h>
128 129
#include <linux/ipv6.h>
#include <linux/in.h>
D
David S. Miller 已提交
130 131
#include <linux/jhash.h>
#include <linux/random.h>
132
#include <trace/events/napi.h>
133
#include <linux/pci.h>
L
Linus Torvalds 已提交
134

135 136
#include "net-sysfs.h"

137 138 139
/* Instead of increasing this, you should create a hash table. */
#define MAX_GRO_SKBS 8

H
Herbert Xu 已提交
140 141 142
/* 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
143 144 145 146 147 148 149 150 151 152 153
/*
 *	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
154
 *             the average user (w/out VLANs) will not be adversely affected.
L
Linus Torvalds 已提交
155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170
 *             --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
 */

171 172 173
#define PTYPE_HASH_SIZE	(16)
#define PTYPE_HASH_MASK	(PTYPE_HASH_SIZE - 1)

L
Linus Torvalds 已提交
174
static DEFINE_SPINLOCK(ptype_lock);
175
static struct list_head ptype_base[PTYPE_HASH_SIZE] __read_mostly;
176
static struct list_head ptype_all __read_mostly;	/* Taps */
L
Linus Torvalds 已提交
177 178

/*
179
 * The @dev_base_head list is protected by @dev_base_lock and the rtnl
L
Linus Torvalds 已提交
180 181
 * semaphore.
 *
182
 * Pure readers hold dev_base_lock for reading, or rcu_read_lock()
L
Linus Torvalds 已提交
183 184
 *
 * Writers must hold the rtnl semaphore while they loop through the
185
 * dev_base_head list, and hold dev_base_lock for writing when they do the
L
Linus Torvalds 已提交
186 187 188 189 190 191 192 193 194 195 196 197 198 199
 * 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);

200
static inline struct hlist_head *dev_name_hash(struct net *net, const char *name)
L
Linus Torvalds 已提交
201 202
{
	unsigned hash = full_name_hash(name, strnlen(name, IFNAMSIZ));
203
	return &net->dev_name_head[hash_32(hash, NETDEV_HASHBITS)];
L
Linus Torvalds 已提交
204 205
}

206
static inline struct hlist_head *dev_index_hash(struct net *net, int ifindex)
L
Linus Torvalds 已提交
207
{
208
	return &net->dev_index_head[ifindex & (NETDEV_HASHENTRIES - 1)];
L
Linus Torvalds 已提交
209 210
}

211 212 213 214 215 216 217 218 219 220 221 222 223 224
static inline void rps_lock(struct softnet_data *queue)
{
#ifdef CONFIG_RPS
	spin_lock(&queue->input_pkt_queue.lock);
#endif
}

static inline void rps_unlock(struct softnet_data *queue)
{
#ifdef CONFIG_RPS
	spin_unlock(&queue->input_pkt_queue.lock);
#endif
}

225 226 227
/* Device list insertion */
static int list_netdevice(struct net_device *dev)
{
228
	struct net *net = dev_net(dev);
229 230 231 232

	ASSERT_RTNL();

	write_lock_bh(&dev_base_lock);
233
	list_add_tail_rcu(&dev->dev_list, &net->dev_base_head);
234
	hlist_add_head_rcu(&dev->name_hlist, dev_name_hash(net, dev->name));
235 236
	hlist_add_head_rcu(&dev->index_hlist,
			   dev_index_hash(net, dev->ifindex));
237 238 239 240
	write_unlock_bh(&dev_base_lock);
	return 0;
}

241 242 243
/* Device list removal
 * caller must respect a RCU grace period before freeing/reusing dev
 */
244 245 246 247 248 249
static void unlist_netdevice(struct net_device *dev)
{
	ASSERT_RTNL();

	/* Unlink dev from the device chain */
	write_lock_bh(&dev_base_lock);
250
	list_del_rcu(&dev->dev_list);
251
	hlist_del_rcu(&dev->name_hlist);
252
	hlist_del_rcu(&dev->index_hlist);
253 254 255
	write_unlock_bh(&dev_base_lock);
}

L
Linus Torvalds 已提交
256 257 258 259
/*
 *	Our notifier list
 */

260
static RAW_NOTIFIER_HEAD(netdev_chain);
L
Linus Torvalds 已提交
261 262 263 264 265

/*
 *	Device drivers call our routines to queue packets here. We empty the
 *	queue in the local softnet handler.
 */
266 267

DEFINE_PER_CPU(struct softnet_data, softnet_data);
E
Eric Dumazet 已提交
268
EXPORT_PER_CPU_SYMBOL(softnet_data);
L
Linus Torvalds 已提交
269

270
#ifdef CONFIG_LOCKDEP
271
/*
272
 * register_netdevice() inits txq->_xmit_lock and sets lockdep class
273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288
 * 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,
289
	 ARPHRD_IEEE80211_PRISM, ARPHRD_IEEE80211_RADIOTAP, ARPHRD_PHONET,
290
	 ARPHRD_PHONET_PIPE, ARPHRD_IEEE802154,
291
	 ARPHRD_VOID, ARPHRD_NONE};
292

293
static const char *const netdev_lock_name[] =
294 295 296 297 298 299 300 301 302 303 304 305 306
	{"_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",
307
	 "_xmit_IEEE80211_PRISM", "_xmit_IEEE80211_RADIOTAP", "_xmit_PHONET",
308
	 "_xmit_PHONET_PIPE", "_xmit_IEEE802154",
309
	 "_xmit_VOID", "_xmit_NONE"};
310 311

static struct lock_class_key netdev_xmit_lock_key[ARRAY_SIZE(netdev_lock_type)];
312
static struct lock_class_key netdev_addr_lock_key[ARRAY_SIZE(netdev_lock_type)];
313 314 315 316 317 318 319 320 321 322 323 324

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;
}

325 326
static inline void netdev_set_xmit_lockdep_class(spinlock_t *lock,
						 unsigned short dev_type)
327 328 329 330 331 332 333
{
	int i;

	i = netdev_lock_pos(dev_type);
	lockdep_set_class_and_name(lock, &netdev_xmit_lock_key[i],
				   netdev_lock_name[i]);
}
334 335 336 337 338 339 340 341 342 343

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]);
}
344
#else
345 346 347 348 349
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)
350 351 352
{
}
#endif
L
Linus Torvalds 已提交
353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383

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

		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.
 *
384
 *	This call does not sleep therefore it can not
L
Linus Torvalds 已提交
385 386 387 388 389 390 391 392 393
 *	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
394
	if (pt->type == htons(ETH_P_ALL))
L
Linus Torvalds 已提交
395
		list_add_rcu(&pt->list, &ptype_all);
S
Stephen Hemminger 已提交
396
	else {
397
		hash = ntohs(pt->type) & PTYPE_HASH_MASK;
L
Linus Torvalds 已提交
398 399 400 401
		list_add_rcu(&pt->list, &ptype_base[hash]);
	}
	spin_unlock_bh(&ptype_lock);
}
E
Eric Dumazet 已提交
402
EXPORT_SYMBOL(dev_add_pack);
L
Linus Torvalds 已提交
403 404 405 406 407 408 409 410

/**
 *	__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
411
 *	returns.
L
Linus Torvalds 已提交
412 413 414 415 416 417 418 419 420 421 422 423
 *
 *      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
424
	if (pt->type == htons(ETH_P_ALL))
L
Linus Torvalds 已提交
425
		head = &ptype_all;
S
Stephen Hemminger 已提交
426
	else
427
		head = &ptype_base[ntohs(pt->type) & PTYPE_HASH_MASK];
L
Linus Torvalds 已提交
428 429 430 431 432 433 434 435 436 437 438 439

	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
440 441
EXPORT_SYMBOL(__dev_remove_pack);

L
Linus Torvalds 已提交
442 443 444 445 446 447 448 449 450 451 452 453 454 455 456
/**
 *	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);
457

L
Linus Torvalds 已提交
458 459
	synchronize_net();
}
E
Eric Dumazet 已提交
460
EXPORT_SYMBOL(dev_remove_pack);
L
Linus Torvalds 已提交
461 462 463 464 465 466 467 468 469 470 471 472 473 474 475 476 477 478 479 480 481 482 483 484 485 486 487 488

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

		      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));
489
			strlcpy(s[i].name, name, IFNAMSIZ);
L
Linus Torvalds 已提交
490 491 492 493 494 495 496 497 498 499 500 501 502 503 504 505 506 507 508 509 510 511 512 513
			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] != ' ' &&
514
		    !strcmp(dev->name, s[i].name)) {
L
Linus Torvalds 已提交
515 516 517 518 519 520 521 522 523
			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
524
EXPORT_SYMBOL(netdev_boot_setup_check);
L
Linus Torvalds 已提交
525 526 527 528 529 530 531 532 533 534 535 536 537 538 539 540 541 542 543 544 545 546 547 548


/**
 *	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
	 */
549
	if (__dev_get_by_name(&init_net, name))
L
Linus Torvalds 已提交
550 551 552 553 554 555 556 557 558 559 560 561 562 563 564 565 566 567 568 569 570 571 572 573 574 575 576 577 578 579 580 581 582 583 584 585 586 587 588 589 590 591 592 593 594
		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
595
 *	@net: the applicable net namespace
L
Linus Torvalds 已提交
596 597 598 599 600 601 602 603 604
 *	@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.
 */

605
struct net_device *__dev_get_by_name(struct net *net, const char *name)
L
Linus Torvalds 已提交
606 607
{
	struct hlist_node *p;
E
Eric Dumazet 已提交
608 609
	struct net_device *dev;
	struct hlist_head *head = dev_name_hash(net, name);
L
Linus Torvalds 已提交
610

E
Eric Dumazet 已提交
611
	hlist_for_each_entry(dev, p, head, name_hlist)
L
Linus Torvalds 已提交
612 613
		if (!strncmp(dev->name, name, IFNAMSIZ))
			return dev;
E
Eric Dumazet 已提交
614

L
Linus Torvalds 已提交
615 616
	return NULL;
}
E
Eric Dumazet 已提交
617
EXPORT_SYMBOL(__dev_get_by_name);
L
Linus Torvalds 已提交
618

619 620 621 622 623 624 625 626 627 628 629 630 631 632 633 634 635 636 637 638 639 640 641 642 643 644
/**
 *	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
645 646
/**
 *	dev_get_by_name		- find a device by its name
647
 *	@net: the applicable net namespace
L
Linus Torvalds 已提交
648 649 650 651 652 653 654 655 656
 *	@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.
 */

657
struct net_device *dev_get_by_name(struct net *net, const char *name)
L
Linus Torvalds 已提交
658 659 660
{
	struct net_device *dev;

661 662
	rcu_read_lock();
	dev = dev_get_by_name_rcu(net, name);
L
Linus Torvalds 已提交
663 664
	if (dev)
		dev_hold(dev);
665
	rcu_read_unlock();
L
Linus Torvalds 已提交
666 667
	return dev;
}
E
Eric Dumazet 已提交
668
EXPORT_SYMBOL(dev_get_by_name);
L
Linus Torvalds 已提交
669 670 671

/**
 *	__dev_get_by_index - find a device by its ifindex
672
 *	@net: the applicable net namespace
L
Linus Torvalds 已提交
673 674 675 676 677 678 679 680 681
 *	@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.
 */

682
struct net_device *__dev_get_by_index(struct net *net, int ifindex)
L
Linus Torvalds 已提交
683 684
{
	struct hlist_node *p;
E
Eric Dumazet 已提交
685 686
	struct net_device *dev;
	struct hlist_head *head = dev_index_hash(net, ifindex);
L
Linus Torvalds 已提交
687

E
Eric Dumazet 已提交
688
	hlist_for_each_entry(dev, p, head, index_hlist)
L
Linus Torvalds 已提交
689 690
		if (dev->ifindex == ifindex)
			return dev;
E
Eric Dumazet 已提交
691

L
Linus Torvalds 已提交
692 693
	return NULL;
}
E
Eric Dumazet 已提交
694
EXPORT_SYMBOL(__dev_get_by_index);
L
Linus Torvalds 已提交
695

696 697 698 699 700 701 702 703 704 705 706 707 708 709 710 711 712 713 714 715 716 717 718 719 720
/**
 *	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
721 722 723

/**
 *	dev_get_by_index - find a device by its ifindex
724
 *	@net: the applicable net namespace
L
Linus Torvalds 已提交
725 726 727 728 729 730 731 732
 *	@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.
 */

733
struct net_device *dev_get_by_index(struct net *net, int ifindex)
L
Linus Torvalds 已提交
734 735 736
{
	struct net_device *dev;

737 738
	rcu_read_lock();
	dev = dev_get_by_index_rcu(net, ifindex);
L
Linus Torvalds 已提交
739 740
	if (dev)
		dev_hold(dev);
741
	rcu_read_unlock();
L
Linus Torvalds 已提交
742 743
	return dev;
}
E
Eric Dumazet 已提交
744
EXPORT_SYMBOL(dev_get_by_index);
L
Linus Torvalds 已提交
745 746 747

/**
 *	dev_getbyhwaddr - find a device by its hardware address
748
 *	@net: the applicable net namespace
L
Linus Torvalds 已提交
749 750 751 752 753 754 755 756 757 758 759 760
 *	@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
 */

761
struct net_device *dev_getbyhwaddr(struct net *net, unsigned short type, char *ha)
L
Linus Torvalds 已提交
762 763 764 765 766
{
	struct net_device *dev;

	ASSERT_RTNL();

767
	for_each_netdev(net, dev)
L
Linus Torvalds 已提交
768 769
		if (dev->type == type &&
		    !memcmp(dev->dev_addr, ha, dev->addr_len))
770 771 772
			return dev;

	return NULL;
L
Linus Torvalds 已提交
773
}
774 775
EXPORT_SYMBOL(dev_getbyhwaddr);

776
struct net_device *__dev_getfirstbyhwtype(struct net *net, unsigned short type)
L
Linus Torvalds 已提交
777 778 779
{
	struct net_device *dev;

780
	ASSERT_RTNL();
781
	for_each_netdev(net, dev)
782
		if (dev->type == type)
783 784 785
			return dev;

	return NULL;
786 787 788
}
EXPORT_SYMBOL(__dev_getfirstbyhwtype);

789
struct net_device *dev_getfirstbyhwtype(struct net *net, unsigned short type)
790
{
791
	struct net_device *dev, *ret = NULL;
792

793 794 795 796 797 798 799 800 801
	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
802 803 804 805 806
}
EXPORT_SYMBOL(dev_getfirstbyhwtype);

/**
 *	dev_get_by_flags - find any device with given flags
807
 *	@net: the applicable net namespace
L
Linus Torvalds 已提交
808 809 810 811
 *	@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
812
 *	is not found or a pointer to the device. The device returned has
L
Linus Torvalds 已提交
813 814 815 816
 *	had a reference added and the pointer is safe until the user calls
 *	dev_put to indicate they have finished with it.
 */

E
Eric Dumazet 已提交
817 818
struct net_device *dev_get_by_flags(struct net *net, unsigned short if_flags,
				    unsigned short mask)
L
Linus Torvalds 已提交
819
{
820
	struct net_device *dev, *ret;
L
Linus Torvalds 已提交
821

822
	ret = NULL;
823 824
	rcu_read_lock();
	for_each_netdev_rcu(net, dev) {
L
Linus Torvalds 已提交
825 826
		if (((dev->flags ^ if_flags) & mask) == 0) {
			dev_hold(dev);
827
			ret = dev;
L
Linus Torvalds 已提交
828 829 830
			break;
		}
	}
831
	rcu_read_unlock();
832
	return ret;
L
Linus Torvalds 已提交
833
}
E
Eric Dumazet 已提交
834
EXPORT_SYMBOL(dev_get_by_flags);
L
Linus Torvalds 已提交
835 836 837 838 839 840

/**
 *	dev_valid_name - check if name is okay for network device
 *	@name: name string
 *
 *	Network device names need to be valid file names to
841 842
 *	to allow sysfs to work.  We also disallow any kind of
 *	whitespace.
L
Linus Torvalds 已提交
843
 */
844
int dev_valid_name(const char *name)
L
Linus Torvalds 已提交
845
{
846 847
	if (*name == '\0')
		return 0;
848 849
	if (strlen(name) >= IFNAMSIZ)
		return 0;
850 851 852 853 854 855 856 857 858
	if (!strcmp(name, ".") || !strcmp(name, ".."))
		return 0;

	while (*name) {
		if (*name == '/' || isspace(*name))
			return 0;
		name++;
	}
	return 1;
L
Linus Torvalds 已提交
859
}
E
Eric Dumazet 已提交
860
EXPORT_SYMBOL(dev_valid_name);
L
Linus Torvalds 已提交
861 862

/**
863 864
 *	__dev_alloc_name - allocate a name for a device
 *	@net: network namespace to allocate the device name in
L
Linus Torvalds 已提交
865
 *	@name: name format string
866
 *	@buf:  scratch buffer and result name string
L
Linus Torvalds 已提交
867 868
 *
 *	Passed a format string - eg "lt%d" it will try and find a suitable
S
Stephen Hemminger 已提交
869 870 871 872 873 874
 *	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
875 876
 */

877
static int __dev_alloc_name(struct net *net, const char *name, char *buf)
L
Linus Torvalds 已提交
878 879 880 881
{
	int i = 0;
	const char *p;
	const int max_netdevices = 8*PAGE_SIZE;
S
Stephen Hemminger 已提交
882
	unsigned long *inuse;
L
Linus Torvalds 已提交
883 884 885 886 887 888 889 890 891 892 893 894 895
	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
896
		inuse = (unsigned long *) get_zeroed_page(GFP_ATOMIC);
L
Linus Torvalds 已提交
897 898 899
		if (!inuse)
			return -ENOMEM;

900
		for_each_netdev(net, d) {
L
Linus Torvalds 已提交
901 902 903 904 905 906
			if (!sscanf(d->name, name, &i))
				continue;
			if (i < 0 || i >= max_netdevices)
				continue;

			/*  avoid cases where sscanf is not exact inverse of printf */
907
			snprintf(buf, IFNAMSIZ, name, i);
L
Linus Torvalds 已提交
908 909 910 911 912 913 914 915
			if (!strncmp(buf, d->name, IFNAMSIZ))
				set_bit(i, inuse);
		}

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

916 917
	if (buf != name)
		snprintf(buf, IFNAMSIZ, name, i);
918
	if (!__dev_get_by_name(net, buf))
L
Linus Torvalds 已提交
919 920 921 922 923 924 925 926 927
		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;
}

928 929 930 931 932 933 934 935 936 937 938 939 940 941 942 943 944 945 946 947
/**
 *	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;

948 949
	BUG_ON(!dev_net(dev));
	net = dev_net(dev);
950 951 952 953 954
	ret = __dev_alloc_name(net, name, buf);
	if (ret >= 0)
		strlcpy(dev->name, buf, IFNAMSIZ);
	return ret;
}
E
Eric Dumazet 已提交
955
EXPORT_SYMBOL(dev_alloc_name);
956

957 958 959 960 961 962 963 964 965 966 967 968 969 970 971
static int dev_get_valid_name(struct net *net, const char *name, char *buf,
			      bool fmt)
{
	if (!dev_valid_name(name))
		return -EINVAL;

	if (fmt && strchr(name, '%'))
		return __dev_alloc_name(net, name, buf);
	else if (__dev_get_by_name(net, name))
		return -EEXIST;
	else if (buf != name)
		strlcpy(buf, name, IFNAMSIZ);

	return 0;
}
L
Linus Torvalds 已提交
972 973 974 975 976 977 978 979 980

/**
 *	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.
 */
981
int dev_change_name(struct net_device *dev, const char *newname)
L
Linus Torvalds 已提交
982
{
983
	char oldname[IFNAMSIZ];
L
Linus Torvalds 已提交
984
	int err = 0;
985
	int ret;
986
	struct net *net;
L
Linus Torvalds 已提交
987 988

	ASSERT_RTNL();
989
	BUG_ON(!dev_net(dev));
L
Linus Torvalds 已提交
990

991
	net = dev_net(dev);
L
Linus Torvalds 已提交
992 993 994
	if (dev->flags & IFF_UP)
		return -EBUSY;

995 996 997
	if (strncmp(newname, dev->name, IFNAMSIZ) == 0)
		return 0;

998 999
	memcpy(oldname, dev->name, IFNAMSIZ);

1000 1001 1002
	err = dev_get_valid_name(net, newname, dev->name, 1);
	if (err < 0)
		return err;
L
Linus Torvalds 已提交
1003

1004
rollback:
1005 1006 1007
	/* For now only devices in the initial network namespace
	 * are in sysfs.
	 */
O
Octavian Purdila 已提交
1008
	if (net_eq(net, &init_net)) {
1009 1010 1011 1012 1013
		ret = device_rename(&dev->dev, dev->name);
		if (ret) {
			memcpy(dev->name, oldname, IFNAMSIZ);
			return ret;
		}
1014
	}
1015 1016

	write_lock_bh(&dev_base_lock);
1017
	hlist_del(&dev->name_hlist);
1018 1019 1020 1021 1022 1023
	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));
1024 1025
	write_unlock_bh(&dev_base_lock);

1026
	ret = call_netdevice_notifiers(NETDEV_CHANGENAME, dev);
1027 1028 1029
	ret = notifier_to_errno(ret);

	if (ret) {
1030 1031
		/* err >= 0 after dev_alloc_name() or stores the first errno */
		if (err >= 0) {
1032 1033 1034
			err = ret;
			memcpy(dev->name, oldname, IFNAMSIZ);
			goto rollback;
1035 1036 1037 1038
		} else {
			printk(KERN_ERR
			       "%s: name change rollback failed: %d.\n",
			       dev->name, ret);
1039 1040
		}
	}
L
Linus Torvalds 已提交
1041 1042 1043 1044

	return err;
}

1045 1046 1047 1048
/**
 *	dev_set_alias - change ifalias of a device
 *	@dev: device
 *	@alias: name up to IFALIASZ
1049
 *	@len: limit of bytes to copy from info
1050 1051 1052 1053 1054 1055 1056 1057 1058 1059
 *
 *	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;

1060 1061 1062 1063 1064 1065 1066 1067
	if (!len) {
		if (dev->ifalias) {
			kfree(dev->ifalias);
			dev->ifalias = NULL;
		}
		return 0;
	}

E
Eric Dumazet 已提交
1068
	dev->ifalias = krealloc(dev->ifalias, len + 1, GFP_KERNEL);
1069 1070 1071 1072 1073 1074 1075 1076
	if (!dev->ifalias)
		return -ENOMEM;

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


1077
/**
S
Stephen Hemminger 已提交
1078
 *	netdev_features_change - device changes features
1079 1080 1081 1082 1083 1084
 *	@dev: device to cause notification
 *
 *	Called to indicate a device has changed features.
 */
void netdev_features_change(struct net_device *dev)
{
1085
	call_netdevice_notifiers(NETDEV_FEAT_CHANGE, dev);
1086 1087 1088
}
EXPORT_SYMBOL(netdev_features_change);

L
Linus Torvalds 已提交
1089 1090 1091 1092 1093 1094 1095 1096 1097 1098 1099
/**
 *	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) {
1100
		call_netdevice_notifiers(NETDEV_CHANGE, dev);
L
Linus Torvalds 已提交
1101 1102 1103
		rtmsg_ifinfo(RTM_NEWLINK, dev, 0);
	}
}
E
Eric Dumazet 已提交
1104
EXPORT_SYMBOL(netdev_state_change);
L
Linus Torvalds 已提交
1105

1106
int netdev_bonding_change(struct net_device *dev, unsigned long event)
1107
{
1108
	return call_netdevice_notifiers(event, dev);
1109 1110 1111
}
EXPORT_SYMBOL(netdev_bonding_change);

L
Linus Torvalds 已提交
1112 1113
/**
 *	dev_load 	- load a network module
1114
 *	@net: the applicable net namespace
L
Linus Torvalds 已提交
1115 1116 1117 1118 1119 1120 1121
 *	@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.
 */

1122
void dev_load(struct net *net, const char *name)
L
Linus Torvalds 已提交
1123
{
1124
	struct net_device *dev;
L
Linus Torvalds 已提交
1125

1126 1127 1128
	rcu_read_lock();
	dev = dev_get_by_name_rcu(net, name);
	rcu_read_unlock();
L
Linus Torvalds 已提交
1129

1130
	if (!dev && capable(CAP_NET_ADMIN))
L
Linus Torvalds 已提交
1131 1132
		request_module("%s", name);
}
E
Eric Dumazet 已提交
1133
EXPORT_SYMBOL(dev_load);
L
Linus Torvalds 已提交
1134

1135
static int __dev_open(struct net_device *dev)
L
Linus Torvalds 已提交
1136
{
1137
	const struct net_device_ops *ops = dev->netdev_ops;
1138
	int ret;
L
Linus Torvalds 已提交
1139

1140 1141
	ASSERT_RTNL();

L
Linus Torvalds 已提交
1142 1143 1144 1145 1146 1147
	/*
	 *	Is it even present?
	 */
	if (!netif_device_present(dev))
		return -ENODEV;

1148 1149 1150 1151 1152
	ret = call_netdevice_notifiers(NETDEV_PRE_UP, dev);
	ret = notifier_to_errno(ret);
	if (ret)
		return ret;

L
Linus Torvalds 已提交
1153 1154 1155 1156
	/*
	 *	Call device private open method
	 */
	set_bit(__LINK_STATE_START, &dev->state);
1157

1158 1159
	if (ops->ndo_validate_addr)
		ret = ops->ndo_validate_addr(dev);
1160

1161 1162
	if (!ret && ops->ndo_open)
		ret = ops->ndo_open(dev);
L
Linus Torvalds 已提交
1163

1164
	/*
L
Linus Torvalds 已提交
1165 1166 1167
	 *	If it went open OK then:
	 */

1168 1169 1170
	if (ret)
		clear_bit(__LINK_STATE_START, &dev->state);
	else {
L
Linus Torvalds 已提交
1171 1172 1173 1174 1175
		/*
		 *	Set the flags.
		 */
		dev->flags |= IFF_UP;

1176 1177 1178
		/*
		 *	Enable NET_DMA
		 */
1179
		net_dmaengine_get();
1180

L
Linus Torvalds 已提交
1181 1182 1183
		/*
		 *	Initialize multicasting status
		 */
1184
		dev_set_rx_mode(dev);
L
Linus Torvalds 已提交
1185 1186 1187 1188 1189 1190

		/*
		 *	Wakeup transmit queue engine
		 */
		dev_activate(dev);
	}
1191

L
Linus Torvalds 已提交
1192 1193 1194 1195
	return ret;
}

/**
1196 1197
 *	dev_open	- prepare an interface for use.
 *	@dev:	device to open
L
Linus Torvalds 已提交
1198
 *
1199 1200 1201 1202 1203 1204 1205
 *	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
1206
 */
1207 1208 1209 1210 1211 1212 1213 1214 1215 1216 1217 1218 1219 1220 1221 1222 1223 1224 1225 1226 1227 1228 1229 1230 1231 1232 1233 1234
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
1235
{
1236
	const struct net_device_ops *ops = dev->netdev_ops;
1237

1238
	ASSERT_RTNL();
1239 1240
	might_sleep();

L
Linus Torvalds 已提交
1241 1242 1243 1244
	/*
	 *	Tell people we are going down, so that they can
	 *	prepare to death, when device is still operating.
	 */
1245
	call_netdevice_notifiers(NETDEV_GOING_DOWN, dev);
L
Linus Torvalds 已提交
1246 1247 1248 1249

	clear_bit(__LINK_STATE_START, &dev->state);

	/* Synchronize to scheduled poll. We cannot touch poll list,
1250 1251 1252 1253 1254
	 * 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
1255 1256
	smp_mb__after_clear_bit(); /* Commit netif_running(). */

1257 1258
	dev_deactivate(dev);

L
Linus Torvalds 已提交
1259 1260 1261 1262 1263 1264 1265
	/*
	 *	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.
	 */
1266 1267
	if (ops->ndo_stop)
		ops->ndo_stop(dev);
L
Linus Torvalds 已提交
1268 1269 1270 1271 1272 1273 1274 1275

	/*
	 *	Device is now down.
	 */

	dev->flags &= ~IFF_UP;

	/*
1276
	 *	Shutdown NET_DMA
L
Linus Torvalds 已提交
1277
	 */
1278 1279 1280 1281 1282 1283 1284 1285 1286 1287 1288 1289 1290 1291 1292 1293 1294 1295 1296 1297
	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
1298

1299
	/*
1300
	 * Tell people we are down
1301
	 */
1302 1303
	rtmsg_ifinfo(RTM_NEWLINK, dev, IFF_UP|IFF_RUNNING);
	call_netdevice_notifiers(NETDEV_DOWN, dev);
1304

L
Linus Torvalds 已提交
1305 1306
	return 0;
}
E
Eric Dumazet 已提交
1307
EXPORT_SYMBOL(dev_close);
L
Linus Torvalds 已提交
1308 1309


1310 1311 1312 1313 1314 1315 1316 1317 1318 1319 1320 1321 1322 1323 1324 1325 1326 1327 1328 1329 1330 1331 1332
/**
 *	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);


1333 1334
static int dev_boot_phase = 1;

L
Linus Torvalds 已提交
1335 1336 1337 1338 1339 1340 1341 1342 1343 1344 1345 1346 1347 1348 1349
/*
 *	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
1350
 *	to the new notifier to allow device to have a race free
L
Linus Torvalds 已提交
1351 1352 1353 1354 1355 1356
 *	view of the network device list.
 */

int register_netdevice_notifier(struct notifier_block *nb)
{
	struct net_device *dev;
1357
	struct net_device *last;
1358
	struct net *net;
L
Linus Torvalds 已提交
1359 1360 1361
	int err;

	rtnl_lock();
1362
	err = raw_notifier_chain_register(&netdev_chain, nb);
1363 1364
	if (err)
		goto unlock;
1365 1366 1367 1368 1369 1370 1371 1372 1373 1374 1375
	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
1376

1377 1378
			nb->notifier_call(nb, NETDEV_UP, dev);
		}
L
Linus Torvalds 已提交
1379
	}
1380 1381

unlock:
L
Linus Torvalds 已提交
1382 1383
	rtnl_unlock();
	return err;
1384 1385 1386

rollback:
	last = dev;
1387 1388 1389 1390
	for_each_net(net) {
		for_each_netdev(net, dev) {
			if (dev == last)
				break;
1391

1392 1393 1394 1395 1396
			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);
1397
			nb->notifier_call(nb, NETDEV_UNREGISTER_BATCH, dev);
1398 1399
		}
	}
1400 1401

	raw_notifier_chain_unregister(&netdev_chain, nb);
1402
	goto unlock;
L
Linus Torvalds 已提交
1403
}
E
Eric Dumazet 已提交
1404
EXPORT_SYMBOL(register_netdevice_notifier);
L
Linus Torvalds 已提交
1405 1406 1407 1408 1409 1410 1411 1412 1413 1414 1415 1416 1417

/**
 *	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)
{
1418 1419 1420
	int err;

	rtnl_lock();
1421
	err = raw_notifier_chain_unregister(&netdev_chain, nb);
1422 1423
	rtnl_unlock();
	return err;
L
Linus Torvalds 已提交
1424
}
E
Eric Dumazet 已提交
1425
EXPORT_SYMBOL(unregister_netdevice_notifier);
L
Linus Torvalds 已提交
1426 1427 1428 1429

/**
 *	call_netdevice_notifiers - call all network notifier blocks
 *      @val: value passed unmodified to notifier function
1430
 *      @dev: net_device pointer passed unmodified to notifier function
L
Linus Torvalds 已提交
1431 1432
 *
 *	Call all network notifier blocks.  Parameters and return value
1433
 *	are as for raw_notifier_call_chain().
L
Linus Torvalds 已提交
1434 1435
 */

1436
int call_netdevice_notifiers(unsigned long val, struct net_device *dev)
L
Linus Torvalds 已提交
1437
{
1438
	return raw_notifier_call_chain(&netdev_chain, val, dev);
L
Linus Torvalds 已提交
1439 1440 1441 1442 1443 1444 1445 1446 1447
}

/* 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
1448
EXPORT_SYMBOL(net_enable_timestamp);
L
Linus Torvalds 已提交
1449 1450 1451 1452 1453

void net_disable_timestamp(void)
{
	atomic_dec(&netstamp_needed);
}
E
Eric Dumazet 已提交
1454
EXPORT_SYMBOL(net_disable_timestamp);
L
Linus Torvalds 已提交
1455

1456
static inline void net_timestamp(struct sk_buff *skb)
L
Linus Torvalds 已提交
1457 1458
{
	if (atomic_read(&netstamp_needed))
1459
		__net_timestamp(skb);
1460 1461
	else
		skb->tstamp.tv64 = 0;
L
Linus Torvalds 已提交
1462 1463
}

1464 1465 1466 1467 1468 1469 1470 1471 1472 1473 1474 1475 1476 1477 1478 1479 1480 1481 1482 1483 1484 1485 1486 1487 1488 1489 1490 1491
/**
 * dev_forward_skb - loopback an skb to another netif
 *
 * @dev: destination network device
 * @skb: buffer to forward
 *
 * return values:
 *	NET_RX_SUCCESS	(no congestion)
 *	NET_RX_DROP     (packet was dropped)
 *
 * 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);

	if (!(dev->flags & IFF_UP))
		return NET_RX_DROP;

	if (skb->len > (dev->mtu + dev->hard_header_len))
		return NET_RX_DROP;

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 1511 1512
#ifdef CONFIG_NET_CLS_ACT
	if (!(skb->tstamp.tv64 && (G_TC_FROM(skb->tc_verd) & AT_INGRESS)))
		net_timestamp(skb);
#else
1513
	net_timestamp(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 1539
				if (net_ratelimit())
					printk(KERN_CRIT "protocol %04x is "
					       "buggy, dev %s\n",
					       skb2->protocol, dev->name);
1540
				skb_reset_network_header(skb2);
L
Linus Torvalds 已提交
1541 1542
			}

1543
			skb2->transport_header = skb2->network_header;
L
Linus Torvalds 已提交
1544
			skb2->pkt_type = PACKET_OUTGOING;
D
David S. Miller 已提交
1545
			ptype->func(skb2, skb->dev, ptype, skb->dev);
L
Linus Torvalds 已提交
1546 1547 1548 1549 1550
		}
	}
	rcu_read_unlock();
}

1551

1552
static inline void __netif_reschedule(struct Qdisc *q)
1553
{
1554 1555
	struct softnet_data *sd;
	unsigned long flags;
1556

1557 1558 1559 1560 1561 1562 1563 1564 1565 1566 1567 1568
	local_irq_save(flags);
	sd = &__get_cpu_var(softnet_data);
	q->next_sched = sd->output_queue;
	sd->output_queue = q;
	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);
1569 1570 1571
}
EXPORT_SYMBOL(__netif_schedule);

1572
void dev_kfree_skb_irq(struct sk_buff *skb)
1573
{
1574 1575 1576
	if (atomic_dec_and_test(&skb->users)) {
		struct softnet_data *sd;
		unsigned long flags;
1577

1578 1579 1580 1581 1582 1583 1584
		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);
	}
1585
}
1586
EXPORT_SYMBOL(dev_kfree_skb_irq);
1587 1588 1589 1590 1591 1592 1593 1594 1595 1596 1597

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);


1598 1599 1600 1601 1602 1603
/**
 * netif_device_detach - mark device as removed
 * @dev: network device
 *
 * Mark device as removed from system and therefore no longer available.
 */
1604 1605 1606 1607
void netif_device_detach(struct net_device *dev)
{
	if (test_and_clear_bit(__LINK_STATE_PRESENT, &dev->state) &&
	    netif_running(dev)) {
1608
		netif_tx_stop_all_queues(dev);
1609 1610 1611 1612
	}
}
EXPORT_SYMBOL(netif_device_detach);

1613 1614 1615 1616 1617 1618
/**
 * netif_device_attach - mark device as attached
 * @dev: network device
 *
 * Mark device as attached from system and restart if needed.
 */
1619 1620 1621 1622
void netif_device_attach(struct net_device *dev)
{
	if (!test_and_set_bit(__LINK_STATE_PRESENT, &dev->state) &&
	    netif_running(dev)) {
1623
		netif_tx_wake_all_queues(dev);
1624
		__netdev_watchdog_up(dev);
1625 1626 1627 1628
	}
}
EXPORT_SYMBOL(netif_device_attach);

1629 1630 1631 1632 1633 1634
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) &&
1635 1636 1637
		 protocol == htons(ETH_P_IPV6)) ||
		((features & NETIF_F_FCOE_CRC) &&
		 protocol == htons(ETH_P_FCOE)));
1638 1639 1640 1641 1642 1643 1644 1645 1646 1647 1648 1649 1650 1651 1652 1653
}

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;
}
1654

1655 1656 1657 1658 1659 1660 1661 1662 1663 1664 1665 1666 1667 1668 1669 1670 1671 1672 1673 1674 1675 1676 1677 1678 1679 1680 1681 1682 1683 1684
/**
 * 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
1685 1686 1687 1688
/*
 * Invalidate hardware checksum when packet is to be mangled, and
 * complete checksum manually on outgoing path.
 */
1689
int skb_checksum_help(struct sk_buff *skb)
L
Linus Torvalds 已提交
1690
{
1691
	__wsum csum;
1692
	int ret = 0, offset;
L
Linus Torvalds 已提交
1693

1694
	if (skb->ip_summed == CHECKSUM_COMPLETE)
1695 1696 1697 1698 1699
		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
1700 1701
	}

1702 1703 1704 1705 1706 1707 1708 1709 1710
	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
1711 1712 1713 1714 1715
		ret = pskb_expand_head(skb, 0, 0, GFP_ATOMIC);
		if (ret)
			goto out;
	}

1716
	*(__sum16 *)(skb->data + offset) = csum_fold(csum);
1717
out_set_summed:
L
Linus Torvalds 已提交
1718
	skb->ip_summed = CHECKSUM_NONE;
1719
out:
L
Linus Torvalds 已提交
1720 1721
	return ret;
}
E
Eric Dumazet 已提交
1722
EXPORT_SYMBOL(skb_checksum_help);
L
Linus Torvalds 已提交
1723

1724 1725 1726
/**
 *	skb_gso_segment - Perform segmentation on skb.
 *	@skb: buffer to segment
1727
 *	@features: features for the output path (see dev->features)
1728 1729
 *
 *	This function segments the given skb and returns a list of segments.
1730 1731 1732
 *
 *	It may return NULL if the skb requires no segmentation.  This is
 *	only possible when GSO is used for verifying header integrity.
1733
 */
1734
struct sk_buff *skb_gso_segment(struct sk_buff *skb, int features)
1735 1736 1737
{
	struct sk_buff *segs = ERR_PTR(-EPROTONOSUPPORT);
	struct packet_type *ptype;
A
Al Viro 已提交
1738
	__be16 type = skb->protocol;
1739
	int err;
1740

1741
	skb_reset_mac_header(skb);
1742
	skb->mac_len = skb->network_header - skb->mac_header;
1743 1744
	__skb_pull(skb, skb->mac_len);

1745 1746 1747 1748 1749 1750 1751 1752 1753 1754 1755 1756 1757
	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);

1758 1759 1760 1761 1762
		if (skb_header_cloned(skb) &&
		    (err = pskb_expand_head(skb, 0, 0, GFP_ATOMIC)))
			return ERR_PTR(err);
	}

1763
	rcu_read_lock();
1764 1765
	list_for_each_entry_rcu(ptype,
			&ptype_base[ntohs(type) & PTYPE_HASH_MASK], list) {
1766
		if (ptype->type == type && !ptype->dev && ptype->gso_segment) {
1767
			if (unlikely(skb->ip_summed != CHECKSUM_PARTIAL)) {
1768 1769 1770 1771
				err = ptype->gso_send_check(skb);
				segs = ERR_PTR(err);
				if (err || skb_gso_ok(skb, features))
					break;
1772 1773
				__skb_push(skb, (skb->data -
						 skb_network_header(skb)));
1774
			}
1775
			segs = ptype->gso_segment(skb, features);
1776 1777 1778 1779 1780
			break;
		}
	}
	rcu_read_unlock();

1781
	__skb_push(skb, skb->data - skb_mac_header(skb));
1782

1783 1784 1785 1786
	return segs;
}
EXPORT_SYMBOL(skb_gso_segment);

1787 1788 1789 1790 1791
/* Take action when hardware reception checksum errors are detected. */
#ifdef CONFIG_BUG
void netdev_rx_csum_fault(struct net_device *dev)
{
	if (net_ratelimit()) {
1792
		printk(KERN_ERR "%s: hw csum failure.\n",
1793
			dev ? dev->name : "<unknown>");
1794 1795 1796 1797 1798 1799
		dump_stack();
	}
}
EXPORT_SYMBOL(netdev_rx_csum_fault);
#endif

L
Linus Torvalds 已提交
1800 1801 1802 1803 1804
/* 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
1805
static int illegal_highdma(struct net_device *dev, struct sk_buff *skb)
L
Linus Torvalds 已提交
1806
{
1807
#ifdef CONFIG_HIGHMEM
L
Linus Torvalds 已提交
1808
	int i;
1809 1810 1811 1812 1813
	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
1814

1815 1816
	if (PCI_DMA_BUS_IS_PHYS) {
		struct device *pdev = dev->dev.parent;
L
Linus Torvalds 已提交
1817

E
Eric Dumazet 已提交
1818 1819
		if (!pdev)
			return 0;
1820 1821 1822 1823 1824 1825
		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;
		}
	}
1826
#endif
L
Linus Torvalds 已提交
1827 1828 1829
	return 0;
}

1830 1831 1832 1833 1834 1835 1836 1837 1838 1839 1840 1841 1842 1843 1844 1845 1846 1847 1848 1849 1850 1851 1852 1853 1854 1855 1856 1857 1858 1859 1860 1861 1862 1863
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;
1864 1865 1866 1867 1868 1869 1870 1871
	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;
1872

1873
	if (IS_ERR(segs))
1874 1875 1876 1877 1878 1879 1880 1881 1882
		return PTR_ERR(segs);

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

	return 0;
}

1883 1884
int dev_hard_start_xmit(struct sk_buff *skb, struct net_device *dev,
			struct netdev_queue *txq)
1885
{
1886
	const struct net_device_ops *ops = dev->netdev_ops;
1887
	int rc = NETDEV_TX_OK;
1888

1889
	if (likely(!skb->next)) {
S
Stephen Hemminger 已提交
1890
		if (!list_empty(&ptype_all))
1891 1892
			dev_queue_xmit_nit(skb, dev);

1893 1894 1895 1896 1897 1898
		if (netif_needs_gso(dev, skb)) {
			if (unlikely(dev_gso_segment(skb)))
				goto out_kfree_skb;
			if (skb->next)
				goto gso;
		}
1899

1900 1901 1902 1903
		/*
		 * If device doesnt need skb->dst, release it right now while
		 * its hot in this cpu cache
		 */
E
Eric Dumazet 已提交
1904 1905 1906
		if (dev->priv_flags & IFF_XMIT_DST_RELEASE)
			skb_dst_drop(skb);

1907
		rc = ops->ndo_start_xmit(skb, dev);
1908
		if (rc == NETDEV_TX_OK)
E
Eric Dumazet 已提交
1909
			txq_trans_update(txq);
1910 1911 1912 1913 1914 1915 1916 1917 1918 1919 1920 1921 1922 1923 1924
		/*
		 * TODO: if skb_orphan() was called by
		 * dev->hard_start_xmit() (for example, the unmodified
		 * igb driver does that; bnx2 doesn't), then
		 * skb_tx_software_timestamp() will be unable to send
		 * back the time stamp.
		 *
		 * How can this be prevented? Always create another
		 * reference to the socket before calling
		 * dev->hard_start_xmit()? Prevent that skb_orphan()
		 * does anything in dev->hard_start_xmit() by clearing
		 * the skb destructor before the call and restoring it
		 * afterwards, then doing the skb_orphan() ourselves?
		 */
		return rc;
1925 1926
	}

1927
gso:
1928 1929 1930 1931 1932
	do {
		struct sk_buff *nskb = skb->next;

		skb->next = nskb->next;
		nskb->next = NULL;
1933 1934 1935 1936 1937 1938 1939 1940

		/*
		 * 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);

1941
		rc = ops->ndo_start_xmit(nskb, dev);
1942
		if (unlikely(rc != NETDEV_TX_OK)) {
1943 1944
			if (rc & ~NETDEV_TX_MASK)
				goto out_kfree_gso_skb;
1945
			nskb->next = skb->next;
1946 1947 1948
			skb->next = nskb;
			return rc;
		}
E
Eric Dumazet 已提交
1949
		txq_trans_update(txq);
1950
		if (unlikely(netif_tx_queue_stopped(txq) && skb->next))
1951
			return NETDEV_TX_BUSY;
1952
	} while (skb->next);
1953

1954 1955 1956
out_kfree_gso_skb:
	if (likely(skb->next == NULL))
		skb->destructor = DEV_GSO_CB(skb)->destructor;
1957 1958
out_kfree_skb:
	kfree_skb(skb);
1959
	return rc;
1960 1961
}

T
Tom Herbert 已提交
1962
static u32 hashrnd __read_mostly;
D
David S. Miller 已提交
1963

1964
u16 skb_tx_hash(const struct net_device *dev, const struct sk_buff *skb)
1965
{
1966
	u32 hash;
D
David S. Miller 已提交
1967

1968 1969
	if (skb_rx_queue_recorded(skb)) {
		hash = skb_get_rx_queue(skb);
E
Eric Dumazet 已提交
1970
		while (unlikely(hash >= dev->real_num_tx_queues))
1971 1972 1973
			hash -= dev->real_num_tx_queues;
		return hash;
	}
1974 1975

	if (skb->sk && skb->sk->sk_hash)
1976
		hash = skb->sk->sk_hash;
1977
	else
1978
		hash = skb->protocol;
1979

T
Tom Herbert 已提交
1980
	hash = jhash_1word(hash, hashrnd);
D
David S. Miller 已提交
1981 1982

	return (u16) (((u64) hash * dev->real_num_tx_queues) >> 32);
1983
}
1984
EXPORT_SYMBOL(skb_tx_hash);
1985

1986 1987 1988 1989
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
1990 1991 1992
			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);
1993 1994 1995 1996 1997 1998
		}
		return 0;
	}
	return queue_index;
}

1999 2000 2001
static struct netdev_queue *dev_pick_tx(struct net_device *dev,
					struct sk_buff *skb)
{
2002 2003 2004 2005 2006 2007 2008 2009 2010 2011
	u16 queue_index;
	struct sock *sk = skb->sk;

	if (sk_tx_queue_recorded(sk)) {
		queue_index = sk_tx_queue_get(sk);
	} else {
		const struct net_device_ops *ops = dev->netdev_ops;

		if (ops->ndo_select_queue) {
			queue_index = ops->ndo_select_queue(dev, skb);
2012
			queue_index = dev_cap_txqueue(dev, queue_index);
2013 2014 2015 2016
		} else {
			queue_index = 0;
			if (dev->real_num_tx_queues > 1)
				queue_index = skb_tx_hash(dev, skb);
2017

E
Eric Dumazet 已提交
2018
			if (sk && rcu_dereference_check(sk->sk_dst_cache, 1))
2019 2020 2021
				sk_tx_queue_set(sk, queue_index);
		}
	}
2022

2023 2024
	skb_set_queue_mapping(skb, queue_index);
	return netdev_get_tx_queue(dev, queue_index);
2025 2026
}

2027 2028 2029 2030 2031 2032 2033 2034 2035 2036 2037 2038 2039 2040 2041 2042 2043 2044 2045 2046 2047 2048 2049 2050 2051 2052 2053 2054 2055 2056 2057 2058 2059 2060
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);
	int rc;

	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) &&
		   !test_and_set_bit(__QDISC_STATE_RUNNING, &q->state)) {
		/*
		 * 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.
		 */
		__qdisc_update_bstats(q, skb->len);
		if (sch_direct_xmit(skb, q, dev, txq, root_lock))
			__qdisc_run(q);
		else
			clear_bit(__QDISC_STATE_RUNNING, &q->state);

		rc = NET_XMIT_SUCCESS;
	} else {
		rc = qdisc_enqueue_root(skb, q);
		qdisc_run(q);
	}
	spin_unlock(root_lock);

	return rc;
}

2061 2062 2063 2064 2065 2066 2067 2068 2069 2070 2071 2072 2073 2074 2075
/*
 * 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_has_frags(skb) && !(dev->features & NETIF_F_FRAGLIST)) ||
	       (skb_shinfo(skb)->nr_frags && (!(dev->features & NETIF_F_SG) ||
					      illegal_highdma(dev, skb)));
}

2076 2077 2078 2079 2080 2081 2082 2083 2084 2085 2086 2087 2088 2089 2090 2091 2092 2093 2094 2095 2096 2097 2098 2099 2100
/**
 *	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
2101 2102 2103
int dev_queue_xmit(struct sk_buff *skb)
{
	struct net_device *dev = skb->dev;
2104
	struct netdev_queue *txq;
L
Linus Torvalds 已提交
2105 2106 2107
	struct Qdisc *q;
	int rc = -ENOMEM;

2108 2109 2110 2111
	/* GSO will handle the following emulations directly. */
	if (netif_needs_gso(dev, skb))
		goto gso;

2112 2113
	/* Convert a paged skb to linear, if required */
	if (skb_needs_linearize(skb, dev) && __skb_linearize(skb))
L
Linus Torvalds 已提交
2114 2115 2116 2117 2118
		goto out_kfree_skb;

	/* If packet is not checksummed and device does not support
	 * checksumming for this protocol, complete checksumming here.
	 */
2119 2120 2121
	if (skb->ip_summed == CHECKSUM_PARTIAL) {
		skb_set_transport_header(skb, skb->csum_start -
					      skb_headroom(skb));
2122 2123
		if (!dev_can_checksum(dev, skb) && skb_checksum_help(skb))
			goto out_kfree_skb;
2124
	}
L
Linus Torvalds 已提交
2125

2126
gso:
2127 2128
	/* Disable soft irqs for various locks below. Also
	 * stops preemption for RCU.
L
Linus Torvalds 已提交
2129
	 */
2130
	rcu_read_lock_bh();
L
Linus Torvalds 已提交
2131

2132
	txq = dev_pick_tx(dev, skb);
2133
	q = rcu_dereference_bh(txq->qdisc);
2134

L
Linus Torvalds 已提交
2135
#ifdef CONFIG_NET_CLS_ACT
E
Eric Dumazet 已提交
2136
	skb->tc_verd = SET_TC_AT(skb->tc_verd, AT_EGRESS);
L
Linus Torvalds 已提交
2137 2138
#endif
	if (q->enqueue) {
2139
		rc = __dev_xmit_skb(skb, q, dev, txq);
2140
		goto out;
L
Linus Torvalds 已提交
2141 2142 2143 2144 2145
	}

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

H
Herbert Xu 已提交
2146 2147
	   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
2148 2149 2150 2151 2152 2153 2154 2155 2156 2157
	   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 */

2158
		if (txq->xmit_lock_owner != cpu) {
L
Linus Torvalds 已提交
2159

2160
			HARD_TX_LOCK(dev, txq, cpu);
L
Linus Torvalds 已提交
2161

2162
			if (!netif_tx_queue_stopped(txq)) {
2163 2164
				rc = dev_hard_start_xmit(skb, dev, txq);
				if (dev_xmit_complete(rc)) {
2165
					HARD_TX_UNLOCK(dev, txq);
L
Linus Torvalds 已提交
2166 2167 2168
					goto out;
				}
			}
2169
			HARD_TX_UNLOCK(dev, txq);
L
Linus Torvalds 已提交
2170 2171 2172 2173 2174 2175 2176 2177 2178 2179 2180 2181 2182
			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;
2183
	rcu_read_unlock_bh();
L
Linus Torvalds 已提交
2184 2185 2186 2187 2188

out_kfree_skb:
	kfree_skb(skb);
	return rc;
out:
2189
	rcu_read_unlock_bh();
L
Linus Torvalds 已提交
2190 2191
	return rc;
}
E
Eric Dumazet 已提交
2192
EXPORT_SYMBOL(dev_queue_xmit);
L
Linus Torvalds 已提交
2193 2194 2195 2196 2197 2198


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

2199 2200 2201
int netdev_max_backlog __read_mostly = 1000;
int netdev_budget __read_mostly = 300;
int weight_p __read_mostly = 64;            /* old backlog weight */
L
Linus Torvalds 已提交
2202 2203 2204

DEFINE_PER_CPU(struct netif_rx_stats, netdev_rx_stat) = { 0, };

E
Eric Dumazet 已提交
2205
#ifdef CONFIG_RPS
T
Tom Herbert 已提交
2206 2207

/* One global table that all flow-based protocols share. */
2208
struct rps_sock_flow_table *rps_sock_flow_table __read_mostly;
T
Tom Herbert 已提交
2209 2210
EXPORT_SYMBOL(rps_sock_flow_table);

T
Tom Herbert 已提交
2211 2212 2213
/*
 * 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.
2214
 * rcu_read_lock must be held on entry.
T
Tom Herbert 已提交
2215
 */
T
Tom Herbert 已提交
2216 2217
static int get_rps_cpu(struct net_device *dev, struct sk_buff *skb,
		       struct rps_dev_flow **rflowp)
T
Tom Herbert 已提交
2218 2219 2220 2221 2222
{
	struct ipv6hdr *ip6;
	struct iphdr *ip;
	struct netdev_rx_queue *rxqueue;
	struct rps_map *map;
T
Tom Herbert 已提交
2223 2224
	struct rps_dev_flow_table *flow_table;
	struct rps_sock_flow_table *sock_flow_table;
T
Tom Herbert 已提交
2225 2226
	int cpu = -1;
	u8 ip_proto;
T
Tom Herbert 已提交
2227
	u16 tcpu;
T
Tom Herbert 已提交
2228 2229 2230 2231 2232 2233
	u32 addr1, addr2, ports, ihl;

	if (skb_rx_queue_recorded(skb)) {
		u16 index = skb_get_rx_queue(skb);
		if (unlikely(index >= dev->num_rx_queues)) {
			if (net_ratelimit()) {
E
Eric Dumazet 已提交
2234 2235 2236
				pr_warning("%s received packet on queue "
					"%u, but number of RX queues is %u\n",
					dev->name, index, dev->num_rx_queues);
T
Tom Herbert 已提交
2237 2238 2239 2240 2241 2242 2243
			}
			goto done;
		}
		rxqueue = dev->_rx + index;
	} else
		rxqueue = dev->_rx;

T
Tom Herbert 已提交
2244
	if (!rxqueue->rps_map && !rxqueue->rps_flow_table)
T
Tom Herbert 已提交
2245 2246 2247 2248 2249 2250 2251 2252 2253 2254 2255 2256 2257 2258 2259 2260 2261 2262 2263 2264 2265 2266 2267 2268 2269 2270 2271 2272 2273 2274 2275 2276 2277 2278 2279 2280 2281 2282 2283 2284 2285 2286 2287 2288 2289 2290 2291 2292 2293 2294 2295
		goto done;

	if (skb->rxhash)
		goto got_hash; /* Skip hash computation on packet header */

	switch (skb->protocol) {
	case __constant_htons(ETH_P_IP):
		if (!pskb_may_pull(skb, sizeof(*ip)))
			goto done;

		ip = (struct iphdr *) skb->data;
		ip_proto = ip->protocol;
		addr1 = ip->saddr;
		addr2 = ip->daddr;
		ihl = ip->ihl;
		break;
	case __constant_htons(ETH_P_IPV6):
		if (!pskb_may_pull(skb, sizeof(*ip6)))
			goto done;

		ip6 = (struct ipv6hdr *) skb->data;
		ip_proto = ip6->nexthdr;
		addr1 = ip6->saddr.s6_addr32[3];
		addr2 = ip6->daddr.s6_addr32[3];
		ihl = (40 >> 2);
		break;
	default:
		goto done;
	}
	ports = 0;
	switch (ip_proto) {
	case IPPROTO_TCP:
	case IPPROTO_UDP:
	case IPPROTO_DCCP:
	case IPPROTO_ESP:
	case IPPROTO_AH:
	case IPPROTO_SCTP:
	case IPPROTO_UDPLITE:
		if (pskb_may_pull(skb, (ihl * 4) + 4))
			ports = *((u32 *) (skb->data + (ihl * 4)));
		break;

	default:
		break;
	}

	skb->rxhash = jhash_3words(addr1, addr2, ports, hashrnd);
	if (!skb->rxhash)
		skb->rxhash = 1;

got_hash:
T
Tom Herbert 已提交
2296 2297 2298 2299 2300 2301 2302 2303 2304 2305 2306 2307 2308 2309 2310 2311 2312 2313 2314 2315 2316 2317 2318 2319 2320 2321 2322 2323 2324 2325 2326 2327 2328 2329 2330 2331 2332 2333 2334
	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
2335 2336
	map = rcu_dereference(rxqueue->rps_map);
	if (map) {
T
Tom Herbert 已提交
2337
		tcpu = map->cpus[((u64) skb->rxhash * map->len) >> 32];
T
Tom Herbert 已提交
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

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

done:
	return cpu;
}

/*
 * This structure holds the per-CPU mask of CPUs for which IPIs are scheduled
 * to be sent to kick remote softirq processing.  There are two masks since
 * the sending of IPIs must be done with interrupts enabled.  The select field
 * indicates the current mask that enqueue_backlog uses to schedule IPIs.
 * select is flipped before net_rps_action is called while still under lock,
 * net_rps_action then uses the non-selected mask to send the IPIs and clears
 * it without conflicting with enqueue_backlog operation.
 */
struct rps_remote_softirq_cpus {
	cpumask_t mask[2];
	int select;
};
static DEFINE_PER_CPU(struct rps_remote_softirq_cpus, rps_remote_softirq_cpus);

/* Called from hardirq (IPI) context */
static void trigger_softirq(void *data)
{
	struct softnet_data *queue = data;
	__napi_schedule(&queue->backlog);
	__get_cpu_var(netdev_rx_stat).received_rps++;
}
T
Tom Herbert 已提交
2371
#endif /* CONFIG_RPS */
T
Tom Herbert 已提交
2372 2373 2374 2375 2376

/*
 * 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
2377 2378
static int enqueue_to_backlog(struct sk_buff *skb, int cpu,
			      unsigned int *qtail)
T
Tom Herbert 已提交
2379 2380 2381 2382 2383 2384 2385 2386 2387
{
	struct softnet_data *queue;
	unsigned long flags;

	queue = &per_cpu(softnet_data, cpu);

	local_irq_save(flags);
	__get_cpu_var(netdev_rx_stat).total++;

2388
	rps_lock(queue);
T
Tom Herbert 已提交
2389 2390 2391 2392
	if (queue->input_pkt_queue.qlen <= netdev_max_backlog) {
		if (queue->input_pkt_queue.qlen) {
enqueue:
			__skb_queue_tail(&queue->input_pkt_queue, skb);
T
Tom Herbert 已提交
2393 2394 2395 2396
#ifdef CONFIG_RPS
			*qtail = queue->input_queue_head +
			    queue->input_pkt_queue.qlen;
#endif
2397 2398
			rps_unlock(queue);
			local_irq_restore(flags);
T
Tom Herbert 已提交
2399 2400 2401 2402 2403
			return NET_RX_SUCCESS;
		}

		/* Schedule NAPI for backlog device */
		if (napi_schedule_prep(&queue->backlog)) {
E
Eric Dumazet 已提交
2404
#ifdef CONFIG_RPS
T
Tom Herbert 已提交
2405 2406 2407 2408 2409 2410
			if (cpu != smp_processor_id()) {
				struct rps_remote_softirq_cpus *rcpus =
				    &__get_cpu_var(rps_remote_softirq_cpus);

				cpu_set(cpu, rcpus->mask[rcpus->select]);
				__raise_softirq_irqoff(NET_RX_SOFTIRQ);
T
Tom Herbert 已提交
2411 2412
				goto enqueue;
			}
2413
#endif
T
Tom Herbert 已提交
2414
			__napi_schedule(&queue->backlog);
T
Tom Herbert 已提交
2415 2416 2417 2418
		}
		goto enqueue;
	}

2419
	rps_unlock(queue);
T
Tom Herbert 已提交
2420 2421 2422 2423 2424 2425 2426

	__get_cpu_var(netdev_rx_stat).dropped++;
	local_irq_restore(flags);

	kfree_skb(skb);
	return NET_RX_DROP;
}
L
Linus Torvalds 已提交
2427 2428 2429 2430 2431 2432 2433 2434 2435 2436 2437 2438 2439 2440 2441 2442 2443 2444

/**
 *	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)
{
2445
	int ret;
L
Linus Torvalds 已提交
2446 2447 2448 2449 2450

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

2451
	if (!skb->tstamp.tv64)
2452
		net_timestamp(skb);
L
Linus Torvalds 已提交
2453

E
Eric Dumazet 已提交
2454
#ifdef CONFIG_RPS
2455
	{
T
Tom Herbert 已提交
2456
		struct rps_dev_flow voidflow, *rflow = &voidflow;
2457 2458 2459
		int cpu;

		rcu_read_lock();
T
Tom Herbert 已提交
2460 2461

		cpu = get_rps_cpu(skb->dev, skb, &rflow);
2462 2463
		if (cpu < 0)
			cpu = smp_processor_id();
T
Tom Herbert 已提交
2464 2465 2466

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

2467 2468
		rcu_read_unlock();
	}
2469
#else
T
Tom Herbert 已提交
2470 2471 2472 2473 2474
	{
		unsigned int qtail;
		ret = enqueue_to_backlog(skb, get_cpu(), &qtail);
		put_cpu();
	}
2475
#endif
2476
	return ret;
L
Linus Torvalds 已提交
2477
}
E
Eric Dumazet 已提交
2478
EXPORT_SYMBOL(netif_rx);
L
Linus Torvalds 已提交
2479 2480 2481 2482 2483 2484 2485 2486 2487 2488 2489 2490 2491 2492 2493 2494 2495 2496 2497 2498 2499 2500 2501 2502 2503 2504 2505 2506 2507 2508 2509

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;

2510
			WARN_ON(atomic_read(&skb->users));
L
Linus Torvalds 已提交
2511 2512 2513 2514 2515
			__kfree_skb(skb);
		}
	}

	if (sd->output_queue) {
2516
		struct Qdisc *head;
L
Linus Torvalds 已提交
2517 2518 2519 2520 2521 2522 2523

		local_irq_disable();
		head = sd->output_queue;
		sd->output_queue = NULL;
		local_irq_enable();

		while (head) {
2524 2525 2526
			struct Qdisc *q = head;
			spinlock_t *root_lock;

L
Linus Torvalds 已提交
2527 2528
			head = head->next_sched;

2529
			root_lock = qdisc_lock(q);
2530
			if (spin_trylock(root_lock)) {
2531 2532 2533
				smp_mb__before_clear_bit();
				clear_bit(__QDISC_STATE_SCHED,
					  &q->state);
2534 2535
				qdisc_run(q);
				spin_unlock(root_lock);
L
Linus Torvalds 已提交
2536
			} else {
2537
				if (!test_bit(__QDISC_STATE_DEACTIVATED,
2538
					      &q->state)) {
2539
					__netif_reschedule(q);
2540 2541 2542 2543 2544
				} else {
					smp_mb__before_clear_bit();
					clear_bit(__QDISC_STATE_SCHED,
						  &q->state);
				}
L
Linus Torvalds 已提交
2545 2546 2547 2548 2549
			}
		}
	}
}

2550 2551 2552
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
2553 2554
{
	atomic_inc(&skb->users);
D
David S. Miller 已提交
2555
	return pt_prev->func(skb, skb->dev, pt_prev, orig_dev);
L
Linus Torvalds 已提交
2556 2557 2558
}

#if defined(CONFIG_BRIDGE) || defined (CONFIG_BRIDGE_MODULE)
2559 2560 2561 2562 2563

#if defined(CONFIG_ATM_LANE) || defined(CONFIG_ATM_LANE_MODULE)
/* This hook is defined here for ATM LANE */
int (*br_fdb_test_addr_hook)(struct net_device *dev,
			     unsigned char *addr) __read_mostly;
2564
EXPORT_SYMBOL_GPL(br_fdb_test_addr_hook);
2565
#endif
L
Linus Torvalds 已提交
2566

2567 2568 2569 2570 2571 2572
/*
 * If bridge module is loaded call bridging hook.
 *  returns NULL if packet was consumed.
 */
struct sk_buff *(*br_handle_frame_hook)(struct net_bridge_port *p,
					struct sk_buff *skb) __read_mostly;
2573
EXPORT_SYMBOL_GPL(br_handle_frame_hook);
2574

2575 2576 2577
static inline struct sk_buff *handle_bridge(struct sk_buff *skb,
					    struct packet_type **pt_prev, int *ret,
					    struct net_device *orig_dev)
L
Linus Torvalds 已提交
2578 2579 2580
{
	struct net_bridge_port *port;

2581 2582 2583
	if (skb->pkt_type == PACKET_LOOPBACK ||
	    (port = rcu_dereference(skb->dev->br_port)) == NULL)
		return skb;
L
Linus Torvalds 已提交
2584 2585

	if (*pt_prev) {
2586
		*ret = deliver_skb(skb, *pt_prev, orig_dev);
L
Linus Torvalds 已提交
2587
		*pt_prev = NULL;
2588 2589
	}

2590
	return br_handle_frame_hook(port, skb);
L
Linus Torvalds 已提交
2591 2592
}
#else
2593
#define handle_bridge(skb, pt_prev, ret, orig_dev)	(skb)
L
Linus Torvalds 已提交
2594 2595
#endif

P
Patrick McHardy 已提交
2596 2597 2598 2599 2600 2601 2602 2603 2604 2605 2606 2607 2608 2609 2610 2611 2612 2613 2614 2615 2616 2617
#if defined(CONFIG_MACVLAN) || defined(CONFIG_MACVLAN_MODULE)
struct sk_buff *(*macvlan_handle_frame_hook)(struct sk_buff *skb) __read_mostly;
EXPORT_SYMBOL_GPL(macvlan_handle_frame_hook);

static inline struct sk_buff *handle_macvlan(struct sk_buff *skb,
					     struct packet_type **pt_prev,
					     int *ret,
					     struct net_device *orig_dev)
{
	if (skb->dev->macvlan_port == NULL)
		return skb;

	if (*pt_prev) {
		*ret = deliver_skb(skb, *pt_prev, orig_dev);
		*pt_prev = NULL;
	}
	return macvlan_handle_frame_hook(skb);
}
#else
#define handle_macvlan(skb, pt_prev, ret, orig_dev)	(skb)
#endif

L
Linus Torvalds 已提交
2618 2619 2620 2621 2622
#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
2623
 * NOTE: This doesnt stop any functionality; if you dont have
L
Linus Torvalds 已提交
2624 2625 2626
 * the ingress scheduler, you just cant add policies on ingress.
 *
 */
2627
static int ing_filter(struct sk_buff *skb)
L
Linus Torvalds 已提交
2628 2629
{
	struct net_device *dev = skb->dev;
2630
	u32 ttl = G_TC_RTTL(skb->tc_verd);
2631 2632 2633
	struct netdev_queue *rxq;
	int result = TC_ACT_OK;
	struct Qdisc *q;
2634

2635 2636 2637
	if (MAX_RED_LOOP < ttl++) {
		printk(KERN_WARNING
		       "Redir loop detected Dropping packet (%d->%d)\n",
2638
		       skb->skb_iif, dev->ifindex);
2639 2640
		return TC_ACT_SHOT;
	}
L
Linus Torvalds 已提交
2641

2642 2643
	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
2644

2645 2646
	rxq = &dev->rx_queue;

2647
	q = rxq->qdisc;
2648
	if (q != &noop_qdisc) {
2649
		spin_lock(qdisc_lock(q));
2650 2651
		if (likely(!test_bit(__QDISC_STATE_DEACTIVATED, &q->state)))
			result = qdisc_enqueue_root(skb, q);
2652 2653
		spin_unlock(qdisc_lock(q));
	}
2654 2655 2656

	return result;
}
2657

2658 2659 2660 2661
static inline struct sk_buff *handle_ing(struct sk_buff *skb,
					 struct packet_type **pt_prev,
					 int *ret, struct net_device *orig_dev)
{
2662
	if (skb->dev->rx_queue.qdisc == &noop_qdisc)
2663
		goto out;
L
Linus Torvalds 已提交
2664

2665 2666 2667 2668 2669 2670
	if (*pt_prev) {
		*ret = deliver_skb(skb, *pt_prev, orig_dev);
		*pt_prev = NULL;
	} else {
		/* Huh? Why does turning on AF_PACKET affect this? */
		skb->tc_verd = SET_TC_OK2MUNGE(skb->tc_verd);
L
Linus Torvalds 已提交
2671 2672
	}

2673 2674 2675 2676 2677 2678 2679 2680 2681 2682
	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
2683 2684 2685
}
#endif

2686 2687 2688 2689 2690 2691 2692 2693 2694 2695 2696 2697 2698 2699 2700 2701 2702 2703 2704 2705 2706 2707 2708 2709 2710 2711 2712
/*
 * 	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();
}

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 2740 2741 2742 2743 2744 2745 2746 2747 2748 2749 2750 2751 2752 2753 2754 2755 2756 2757 2758 2759 2760 2761
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;

	if ((master->priv_flags & IFF_MASTER_ALB) && master->br_port) {
		/* 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);

2762
static int __netif_receive_skb(struct sk_buff *skb)
L
Linus Torvalds 已提交
2763 2764
{
	struct packet_type *ptype, *pt_prev;
D
David S. Miller 已提交
2765
	struct net_device *orig_dev;
2766
	struct net_device *master;
2767
	struct net_device *null_or_orig;
2768
	struct net_device *null_or_bond;
L
Linus Torvalds 已提交
2769
	int ret = NET_RX_DROP;
A
Al Viro 已提交
2770
	__be16 type;
L
Linus Torvalds 已提交
2771

2772 2773 2774
	if (!skb->tstamp.tv64)
		net_timestamp(skb);

E
Eric Dumazet 已提交
2775
	if (vlan_tx_tag_present(skb) && vlan_hwaccel_do_receive(skb))
2776 2777
		return NET_RX_SUCCESS;

L
Linus Torvalds 已提交
2778
	/* if we've gotten here through NAPI, check netpoll */
2779
	if (netpoll_receive_skb(skb))
L
Linus Torvalds 已提交
2780 2781
		return NET_RX_DROP;

2782 2783
	if (!skb->skb_iif)
		skb->skb_iif = skb->dev->ifindex;
2784

2785
	null_or_orig = NULL;
J
Joe Eykholt 已提交
2786
	orig_dev = skb->dev;
2787 2788 2789
	master = ACCESS_ONCE(orig_dev->master);
	if (master) {
		if (skb_bond_should_drop(skb, master))
2790 2791
			null_or_orig = orig_dev; /* deliver only exact match */
		else
2792
			skb->dev = master;
J
Joe Eykholt 已提交
2793
	}
2794

L
Linus Torvalds 已提交
2795 2796
	__get_cpu_var(netdev_rx_stat).total++;

2797
	skb_reset_network_header(skb);
2798
	skb_reset_transport_header(skb);
2799
	skb->mac_len = skb->network_header - skb->mac_header;
L
Linus Torvalds 已提交
2800 2801 2802 2803 2804 2805 2806 2807 2808 2809 2810 2811 2812

	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) {
2813 2814
		if (ptype->dev == null_or_orig || ptype->dev == skb->dev ||
		    ptype->dev == orig_dev) {
2815
			if (pt_prev)
D
David S. Miller 已提交
2816
				ret = deliver_skb(skb, pt_prev, orig_dev);
L
Linus Torvalds 已提交
2817 2818 2819 2820 2821
			pt_prev = ptype;
		}
	}

#ifdef CONFIG_NET_CLS_ACT
2822 2823
	skb = handle_ing(skb, &pt_prev, &ret, orig_dev);
	if (!skb)
L
Linus Torvalds 已提交
2824 2825 2826 2827
		goto out;
ncls:
#endif

2828
	skb = handle_bridge(skb, &pt_prev, &ret, orig_dev);
P
Patrick McHardy 已提交
2829 2830 2831
	if (!skb)
		goto out;
	skb = handle_macvlan(skb, &pt_prev, &ret, orig_dev);
2832
	if (!skb)
L
Linus Torvalds 已提交
2833 2834
		goto out;

2835 2836 2837 2838 2839 2840
	/*
	 * 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.
	 */
2841
	null_or_bond = NULL;
2842 2843
	if ((skb->dev->priv_flags & IFF_802_1Q_VLAN) &&
	    (vlan_dev_real_dev(skb->dev)->priv_flags & IFF_BONDING)) {
2844
		null_or_bond = vlan_dev_real_dev(skb->dev);
2845 2846
	}

L
Linus Torvalds 已提交
2847
	type = skb->protocol;
2848 2849
	list_for_each_entry_rcu(ptype,
			&ptype_base[ntohs(type) & PTYPE_HASH_MASK], list) {
2850
		if (ptype->type == type && (ptype->dev == null_or_orig ||
2851 2852
		     ptype->dev == skb->dev || ptype->dev == orig_dev ||
		     ptype->dev == null_or_bond)) {
2853
			if (pt_prev)
D
David S. Miller 已提交
2854
				ret = deliver_skb(skb, pt_prev, orig_dev);
L
Linus Torvalds 已提交
2855 2856 2857 2858 2859
			pt_prev = ptype;
		}
	}

	if (pt_prev) {
D
David S. Miller 已提交
2860
		ret = pt_prev->func(skb, skb->dev, pt_prev, orig_dev);
L
Linus Torvalds 已提交
2861 2862 2863 2864 2865 2866 2867 2868 2869 2870 2871 2872
	} 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
2873 2874 2875 2876 2877 2878 2879 2880 2881 2882 2883 2884 2885 2886 2887 2888 2889 2890

/**
 *	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
2891
#ifdef CONFIG_RPS
T
Tom Herbert 已提交
2892 2893 2894 2895
	struct rps_dev_flow voidflow, *rflow = &voidflow;
	int cpu, ret;

	rcu_read_lock();
T
Tom Herbert 已提交
2896

T
Tom Herbert 已提交
2897
	cpu = get_rps_cpu(skb->dev, skb, &rflow);
T
Tom Herbert 已提交
2898

T
Tom Herbert 已提交
2899 2900 2901 2902 2903 2904 2905 2906 2907
	if (cpu >= 0) {
		ret = enqueue_to_backlog(skb, cpu, &rflow->last_qtail);
		rcu_read_unlock();
	} else {
		rcu_read_unlock();
		ret = __netif_receive_skb(skb);
	}

	return ret;
2908 2909 2910
#else
	return __netif_receive_skb(skb);
#endif
T
Tom Herbert 已提交
2911
}
E
Eric Dumazet 已提交
2912
EXPORT_SYMBOL(netif_receive_skb);
L
Linus Torvalds 已提交
2913

2914
/* Network device is going away, flush any packets still pending  */
2915
static void flush_backlog(void *arg)
2916
{
2917 2918
	struct net_device *dev = arg;
	struct softnet_data *queue = &__get_cpu_var(softnet_data);
2919 2920
	struct sk_buff *skb, *tmp;

2921
	rps_lock(queue);
2922 2923 2924 2925
	skb_queue_walk_safe(&queue->input_pkt_queue, skb, tmp)
		if (skb->dev == dev) {
			__skb_unlink(skb, &queue->input_pkt_queue);
			kfree_skb(skb);
T
Tom Herbert 已提交
2926
			incr_input_queue_head(queue);
2927
		}
2928
	rps_unlock(queue);
2929 2930
}

2931 2932 2933 2934 2935 2936 2937
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;

2938 2939
	if (NAPI_GRO_CB(skb)->count == 1) {
		skb_shinfo(skb)->gso_size = 0;
2940
		goto out;
2941
	}
2942 2943 2944 2945 2946 2947 2948 2949 2950 2951 2952 2953 2954 2955 2956 2957 2958 2959 2960 2961 2962

	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);
}

2963
static void napi_gro_flush(struct napi_struct *napi)
2964 2965 2966 2967 2968 2969 2970 2971 2972
{
	struct sk_buff *skb, *next;

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

2973
	napi->gro_count = 0;
2974 2975 2976
	napi->gro_list = NULL;
}

2977
enum gro_result dev_gro_receive(struct napi_struct *napi, struct sk_buff *skb)
2978 2979 2980 2981 2982
{
	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
2983
	int same_flow;
2984
	int mac_len;
2985
	enum gro_result ret;
2986 2987 2988 2989

	if (!(skb->dev->features & NETIF_F_GRO))
		goto normal;

2990
	if (skb_is_gso(skb) || skb_has_frags(skb))
2991 2992
		goto normal;

2993 2994 2995 2996 2997
	rcu_read_lock();
	list_for_each_entry_rcu(ptype, head, list) {
		if (ptype->type != type || ptype->dev || !ptype->gro_receive)
			continue;

2998
		skb_set_network_header(skb, skb_gro_offset(skb));
2999 3000 3001 3002
		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
3003
		NAPI_GRO_CB(skb)->free = 0;
3004 3005 3006 3007 3008 3009 3010 3011 3012

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

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

H
Herbert Xu 已提交
3013
	same_flow = NAPI_GRO_CB(skb)->same_flow;
3014
	ret = NAPI_GRO_CB(skb)->free ? GRO_MERGED_FREE : GRO_MERGED;
H
Herbert Xu 已提交
3015

3016 3017 3018 3019 3020 3021
	if (pp) {
		struct sk_buff *nskb = *pp;

		*pp = nskb->next;
		nskb->next = NULL;
		napi_gro_complete(nskb);
3022
		napi->gro_count--;
3023 3024
	}

H
Herbert Xu 已提交
3025
	if (same_flow)
3026 3027
		goto ok;

3028
	if (NAPI_GRO_CB(skb)->flush || napi->gro_count >= MAX_GRO_SKBS)
3029 3030
		goto normal;

3031
	napi->gro_count++;
3032
	NAPI_GRO_CB(skb)->count = 1;
3033
	skb_shinfo(skb)->gso_size = skb_gro_len(skb);
3034 3035
	skb->next = napi->gro_list;
	napi->gro_list = skb;
3036
	ret = GRO_HELD;
3037

3038
pull:
H
Herbert Xu 已提交
3039 3040 3041 3042 3043 3044 3045 3046 3047 3048 3049 3050 3051 3052 3053 3054 3055 3056 3057
	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);
		}
3058 3059
	}

3060
ok:
3061
	return ret;
3062 3063

normal:
3064 3065
	ret = GRO_NORMAL;
	goto pull;
H
Herbert Xu 已提交
3066
}
3067 3068
EXPORT_SYMBOL(dev_gro_receive);

3069 3070
static gro_result_t
__napi_gro_receive(struct napi_struct *napi, struct sk_buff *skb)
3071 3072 3073
{
	struct sk_buff *p;

3074 3075 3076
	if (netpoll_rx_on(skb))
		return GRO_NORMAL;

3077
	for (p = napi->gro_list; p; p = p->next) {
3078 3079 3080 3081
		NAPI_GRO_CB(p)->same_flow =
			(p->dev == skb->dev) &&
			!compare_ether_header(skb_mac_header(p),
					      skb_gro_mac_header(skb));
3082 3083 3084 3085 3086
		NAPI_GRO_CB(p)->flush = 0;
	}

	return dev_gro_receive(napi, skb);
}
H
Herbert Xu 已提交
3087

3088
gro_result_t napi_skb_finish(gro_result_t ret, struct sk_buff *skb)
H
Herbert Xu 已提交
3089
{
3090 3091
	switch (ret) {
	case GRO_NORMAL:
3092 3093 3094
		if (netif_receive_skb(skb))
			ret = GRO_DROP;
		break;
H
Herbert Xu 已提交
3095

3096 3097
	case GRO_DROP:
	case GRO_MERGED_FREE:
H
Herbert Xu 已提交
3098 3099
		kfree_skb(skb);
		break;
3100 3101 3102 3103

	case GRO_HELD:
	case GRO_MERGED:
		break;
H
Herbert Xu 已提交
3104 3105
	}

3106
	return ret;
3107 3108 3109
}
EXPORT_SYMBOL(napi_skb_finish);

3110 3111 3112 3113
void skb_gro_reset_offset(struct sk_buff *skb)
{
	NAPI_GRO_CB(skb)->data_offset = 0;
	NAPI_GRO_CB(skb)->frag0 = NULL;
3114
	NAPI_GRO_CB(skb)->frag0_len = 0;
3115

3116
	if (skb->mac_header == skb->tail &&
3117
	    !PageHighMem(skb_shinfo(skb)->frags[0].page)) {
3118 3119 3120
		NAPI_GRO_CB(skb)->frag0 =
			page_address(skb_shinfo(skb)->frags[0].page) +
			skb_shinfo(skb)->frags[0].page_offset;
3121 3122
		NAPI_GRO_CB(skb)->frag0_len = skb_shinfo(skb)->frags[0].size;
	}
3123 3124 3125
}
EXPORT_SYMBOL(skb_gro_reset_offset);

3126
gro_result_t napi_gro_receive(struct napi_struct *napi, struct sk_buff *skb)
3127
{
3128 3129
	skb_gro_reset_offset(skb);

3130
	return napi_skb_finish(__napi_gro_receive(napi, skb), skb);
3131 3132 3133
}
EXPORT_SYMBOL(napi_gro_receive);

3134 3135 3136 3137 3138 3139 3140 3141 3142
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);

3143
struct sk_buff *napi_get_frags(struct napi_struct *napi)
H
Herbert Xu 已提交
3144 3145 3146 3147
{
	struct sk_buff *skb = napi->skb;

	if (!skb) {
3148 3149 3150
		skb = netdev_alloc_skb_ip_align(napi->dev, GRO_MAX_HEAD);
		if (skb)
			napi->skb = skb;
3151
	}
3152 3153
	return skb;
}
3154
EXPORT_SYMBOL(napi_get_frags);
3155

3156 3157
gro_result_t napi_frags_finish(struct napi_struct *napi, struct sk_buff *skb,
			       gro_result_t ret)
3158
{
3159 3160
	switch (ret) {
	case GRO_NORMAL:
3161
	case GRO_HELD:
3162
		skb->protocol = eth_type_trans(skb, skb->dev);
3163

3164 3165 3166 3167
		if (ret == GRO_HELD)
			skb_gro_pull(skb, -ETH_HLEN);
		else if (netif_receive_skb(skb))
			ret = GRO_DROP;
3168
		break;
H
Herbert Xu 已提交
3169

3170 3171 3172 3173
	case GRO_DROP:
	case GRO_MERGED_FREE:
		napi_reuse_skb(napi, skb);
		break;
3174 3175 3176

	case GRO_MERGED:
		break;
3177
	}
H
Herbert Xu 已提交
3178

3179
	return ret;
H
Herbert Xu 已提交
3180
}
3181 3182
EXPORT_SYMBOL(napi_frags_finish);

3183 3184 3185 3186
struct sk_buff *napi_frags_skb(struct napi_struct *napi)
{
	struct sk_buff *skb = napi->skb;
	struct ethhdr *eth;
3187 3188
	unsigned int hlen;
	unsigned int off;
3189 3190 3191 3192 3193 3194

	napi->skb = NULL;

	skb_reset_mac_header(skb);
	skb_gro_reset_offset(skb);

3195 3196 3197 3198 3199 3200 3201 3202 3203 3204
	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;
		}
3205 3206 3207 3208 3209 3210 3211 3212 3213 3214 3215 3216 3217 3218 3219
	}

	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);

3220
gro_result_t napi_gro_frags(struct napi_struct *napi)
3221
{
3222
	struct sk_buff *skb = napi_frags_skb(napi);
3223 3224

	if (!skb)
3225
		return GRO_DROP;
3226 3227 3228

	return napi_frags_finish(napi, skb, __napi_gro_receive(napi, skb));
}
H
Herbert Xu 已提交
3229 3230
EXPORT_SYMBOL(napi_gro_frags);

3231
static int process_backlog(struct napi_struct *napi, int quota)
L
Linus Torvalds 已提交
3232 3233 3234 3235 3236
{
	int work = 0;
	struct softnet_data *queue = &__get_cpu_var(softnet_data);
	unsigned long start_time = jiffies;

3237 3238
	napi->weight = weight_p;
	do {
L
Linus Torvalds 已提交
3239 3240
		struct sk_buff *skb;

3241 3242
		local_irq_disable();
		rps_lock(queue);
L
Linus Torvalds 已提交
3243
		skb = __skb_dequeue(&queue->input_pkt_queue);
3244
		if (!skb) {
3245
			__napi_complete(napi);
T
Tom Herbert 已提交
3246
			rps_unlock(queue);
3247
			local_irq_enable();
3248
			break;
3249
		}
T
Tom Herbert 已提交
3250
		incr_input_queue_head(queue);
3251 3252
		rps_unlock(queue);
		local_irq_enable();
L
Linus Torvalds 已提交
3253

T
Tom Herbert 已提交
3254
		__netif_receive_skb(skb);
3255
	} while (++work < quota && jiffies == start_time);
L
Linus Torvalds 已提交
3256

3257 3258
	return work;
}
L
Linus Torvalds 已提交
3259

3260 3261
/**
 * __napi_schedule - schedule for receive
3262
 * @n: entry to schedule
3263 3264 3265
 *
 * The entry's receive function will be scheduled to run
 */
H
Harvey Harrison 已提交
3266
void __napi_schedule(struct napi_struct *n)
3267 3268
{
	unsigned long flags;
L
Linus Torvalds 已提交
3269

3270 3271 3272 3273
	local_irq_save(flags);
	list_add_tail(&n->poll_list, &__get_cpu_var(softnet_data).poll_list);
	__raise_softirq_irqoff(NET_RX_SOFTIRQ);
	local_irq_restore(flags);
L
Linus Torvalds 已提交
3274
}
3275 3276
EXPORT_SYMBOL(__napi_schedule);

3277 3278 3279 3280 3281 3282 3283 3284 3285 3286 3287 3288 3289 3290 3291 3292 3293 3294 3295 3296 3297 3298 3299 3300 3301 3302 3303 3304 3305 3306 3307 3308 3309
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);
3310
	napi->gro_count = 0;
3311
	napi->gro_list = NULL;
H
Herbert Xu 已提交
3312
	napi->skb = NULL;
3313 3314 3315 3316
	napi->poll = poll;
	napi->weight = weight;
	list_add(&napi->dev_list, &dev->napi_list);
	napi->dev = dev;
H
Herbert Xu 已提交
3317
#ifdef CONFIG_NETPOLL
3318 3319 3320 3321 3322 3323 3324 3325 3326 3327 3328
	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;

3329
	list_del_init(&napi->dev_list);
3330
	napi_free_frags(napi);
3331 3332 3333 3334 3335 3336 3337 3338

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

	napi->gro_list = NULL;
3339
	napi->gro_count = 0;
3340 3341 3342
}
EXPORT_SYMBOL(netif_napi_del);

E
Eric Dumazet 已提交
3343
#ifdef CONFIG_RPS
T
Tom Herbert 已提交
3344 3345 3346 3347 3348 3349 3350 3351 3352 3353 3354 3355 3356 3357 3358 3359
/*
 * net_rps_action sends any pending IPI's for rps.  This is only called from
 * softirq and interrupts must be enabled.
 */
static void net_rps_action(cpumask_t *mask)
{
	int cpu;

	/* Send pending IPI's to kick RPS processing on remote cpus. */
	for_each_cpu_mask_nr(cpu, *mask) {
		struct softnet_data *queue = &per_cpu(softnet_data, cpu);
		if (cpu_online(cpu))
			__smp_call_function_single(cpu, &queue->csd, 0);
	}
	cpus_clear(*mask);
}
3360
#endif
L
Linus Torvalds 已提交
3361 3362 3363

static void net_rx_action(struct softirq_action *h)
{
3364
	struct list_head *list = &__get_cpu_var(softnet_data).poll_list;
3365
	unsigned long time_limit = jiffies + 2;
3366
	int budget = netdev_budget;
3367
	void *have;
E
Eric Dumazet 已提交
3368
#ifdef CONFIG_RPS
T
Tom Herbert 已提交
3369 3370
	int select;
	struct rps_remote_softirq_cpus *rcpus;
3371
#endif
3372

L
Linus Torvalds 已提交
3373 3374
	local_irq_disable();

3375 3376 3377
	while (!list_empty(list)) {
		struct napi_struct *n;
		int work, weight;
L
Linus Torvalds 已提交
3378

3379
		/* If softirq window is exhuasted then punt.
3380 3381
		 * Allow this to run for 2 jiffies since which will allow
		 * an average latency of 1.5/HZ.
3382
		 */
3383
		if (unlikely(budget <= 0 || time_after(jiffies, time_limit)))
L
Linus Torvalds 已提交
3384 3385 3386 3387
			goto softnet_break;

		local_irq_enable();

3388 3389 3390 3391 3392
		/* 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.
		 */
3393
		n = list_first_entry(list, struct napi_struct, poll_list);
L
Linus Torvalds 已提交
3394

3395 3396 3397 3398
		have = netpoll_poll_lock(n);

		weight = n->weight;

3399 3400 3401 3402 3403 3404 3405
		/* 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;
3406
		if (test_bit(NAPI_STATE_SCHED, &n->state)) {
3407
			work = n->poll(n, weight);
3408 3409
			trace_napi_poll(n);
		}
3410 3411 3412 3413 3414 3415 3416 3417 3418 3419 3420 3421

		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.
		 */
3422
		if (unlikely(work == weight)) {
3423 3424 3425 3426 3427
			if (unlikely(napi_disable_pending(n))) {
				local_irq_enable();
				napi_complete(n);
				local_irq_disable();
			} else
3428 3429
				list_move_tail(&n->poll_list, list);
		}
3430 3431

		netpoll_poll_unlock(have);
L
Linus Torvalds 已提交
3432 3433
	}
out:
E
Eric Dumazet 已提交
3434
#ifdef CONFIG_RPS
T
Tom Herbert 已提交
3435 3436 3437 3438
	rcpus = &__get_cpu_var(rps_remote_softirq_cpus);
	select = rcpus->select;
	rcpus->select ^= 1;

3439
	local_irq_enable();
3440

T
Tom Herbert 已提交
3441
	net_rps_action(&rcpus->mask[select]);
3442 3443 3444
#else
	local_irq_enable();
#endif
T
Tom Herbert 已提交
3445

3446 3447 3448 3449 3450
#ifdef CONFIG_NET_DMA
	/*
	 * There may not be any more sk_buffs coming right now, so push
	 * any pending DMA copies to hardware
	 */
3451
	dma_issue_pending_all();
3452
#endif
3453

L
Linus Torvalds 已提交
3454 3455 3456 3457 3458 3459 3460 3461
	return;

softnet_break:
	__get_cpu_var(netdev_rx_stat).time_squeeze++;
	__raise_softirq_irqoff(NET_RX_SOFTIRQ);
	goto out;
}

E
Eric Dumazet 已提交
3462
static gifconf_func_t *gifconf_list[NPROTO];
L
Linus Torvalds 已提交
3463 3464 3465 3466 3467 3468 3469 3470 3471 3472

/**
 *	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
3473
int register_gifconf(unsigned int family, gifconf_func_t *gifconf)
L
Linus Torvalds 已提交
3474 3475 3476 3477 3478 3479
{
	if (family >= NPROTO)
		return -EINVAL;
	gifconf_list[family] = gifconf;
	return 0;
}
E
Eric Dumazet 已提交
3480
EXPORT_SYMBOL(register_gifconf);
L
Linus Torvalds 已提交
3481 3482 3483 3484 3485 3486 3487 3488 3489 3490 3491 3492 3493


/*
 *	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
 */

3494
static int dev_ifname(struct net *net, struct ifreq __user *arg)
L
Linus Torvalds 已提交
3495 3496 3497 3498 3499 3500 3501 3502 3503 3504 3505
{
	struct net_device *dev;
	struct ifreq ifr;

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

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

3506 3507
	rcu_read_lock();
	dev = dev_get_by_index_rcu(net, ifr.ifr_ifindex);
L
Linus Torvalds 已提交
3508
	if (!dev) {
3509
		rcu_read_unlock();
L
Linus Torvalds 已提交
3510 3511 3512 3513
		return -ENODEV;
	}

	strcpy(ifr.ifr_name, dev->name);
3514
	rcu_read_unlock();
L
Linus Torvalds 已提交
3515 3516 3517 3518 3519 3520 3521 3522 3523 3524 3525 3526

	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'.
 */

3527
static int dev_ifconf(struct net *net, char __user *arg)
L
Linus Torvalds 已提交
3528 3529 3530 3531 3532 3533 3534 3535 3536 3537 3538 3539 3540 3541 3542 3543 3544 3545 3546 3547 3548 3549 3550
{
	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;
3551
	for_each_netdev(net, dev) {
L
Linus Torvalds 已提交
3552 3553 3554 3555 3556 3557 3558 3559 3560 3561 3562 3563 3564
		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;
			}
		}
3565
	}
L
Linus Torvalds 已提交
3566 3567 3568 3569 3570 3571 3572 3573 3574 3575 3576 3577 3578 3579 3580 3581 3582

	/*
	 *	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.
 */
3583
void *dev_seq_start(struct seq_file *seq, loff_t *pos)
3584
	__acquires(RCU)
L
Linus Torvalds 已提交
3585
{
3586
	struct net *net = seq_file_net(seq);
3587
	loff_t off;
L
Linus Torvalds 已提交
3588 3589
	struct net_device *dev;

3590
	rcu_read_lock();
3591 3592
	if (!*pos)
		return SEQ_START_TOKEN;
L
Linus Torvalds 已提交
3593

3594
	off = 1;
3595
	for_each_netdev_rcu(net, dev)
3596 3597
		if (off++ == *pos)
			return dev;
L
Linus Torvalds 已提交
3598

3599
	return NULL;
L
Linus Torvalds 已提交
3600 3601 3602 3603
}

void *dev_seq_next(struct seq_file *seq, void *v, loff_t *pos)
{
3604 3605 3606 3607
	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
3608
	++*pos;
3609
	return rcu_dereference(dev);
L
Linus Torvalds 已提交
3610 3611 3612
}

void dev_seq_stop(struct seq_file *seq, void *v)
3613
	__releases(RCU)
L
Linus Torvalds 已提交
3614
{
3615
	rcu_read_unlock();
L
Linus Torvalds 已提交
3616 3617 3618 3619
}

static void dev_seq_printf_stats(struct seq_file *seq, struct net_device *dev)
{
3620
	const struct net_device_stats *stats = dev_get_stats(dev);
L
Linus Torvalds 已提交
3621

3622
	seq_printf(seq, "%6s: %7lu %7lu %4lu %4lu %4lu %5lu %10lu %9lu "
3623 3624 3625 3626 3627 3628 3629 3630 3631 3632 3633 3634 3635 3636 3637 3638
		   "%8lu %7lu %4lu %4lu %4lu %5lu %7lu %10lu\n",
		   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
3639 3640 3641 3642 3643 3644 3645 3646 3647 3648 3649 3650 3651 3652 3653 3654 3655 3656 3657 3658 3659 3660 3661
}

/*
 *	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;
}

static struct netif_rx_stats *softnet_get_online(loff_t *pos)
{
	struct netif_rx_stats *rc = NULL;

3662
	while (*pos < nr_cpu_ids)
3663
		if (cpu_online(*pos)) {
L
Linus Torvalds 已提交
3664 3665 3666 3667 3668 3669 3670 3671 3672 3673 3674 3675 3676 3677 3678 3679 3680 3681 3682 3683 3684 3685 3686 3687 3688 3689
			rc = &per_cpu(netdev_rx_stat, *pos);
			break;
		} else
			++*pos;
	return rc;
}

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)
{
	struct netif_rx_stats *s = v;

T
Tom Herbert 已提交
3690
	seq_printf(seq, "%08x %08x %08x %08x %08x %08x %08x %08x %08x %08x\n",
3691
		   s->total, s->dropped, s->time_squeeze, 0,
3692
		   0, 0, 0, 0, /* was fastroute */
T
Tom Herbert 已提交
3693
		   s->cpu_collision, s->received_rps);
L
Linus Torvalds 已提交
3694 3695 3696
	return 0;
}

3697
static const struct seq_operations dev_seq_ops = {
L
Linus Torvalds 已提交
3698 3699 3700 3701 3702 3703 3704 3705
	.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)
{
3706 3707
	return seq_open_net(inode, file, &dev_seq_ops,
			    sizeof(struct seq_net_private));
L
Linus Torvalds 已提交
3708 3709
}

3710
static const struct file_operations dev_seq_fops = {
L
Linus Torvalds 已提交
3711 3712 3713 3714
	.owner	 = THIS_MODULE,
	.open    = dev_seq_open,
	.read    = seq_read,
	.llseek  = seq_lseek,
3715
	.release = seq_release_net,
L
Linus Torvalds 已提交
3716 3717
};

3718
static const struct seq_operations softnet_seq_ops = {
L
Linus Torvalds 已提交
3719 3720 3721 3722 3723 3724 3725 3726 3727 3728 3729
	.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);
}

3730
static const struct file_operations softnet_seq_fops = {
L
Linus Torvalds 已提交
3731 3732 3733 3734 3735 3736 3737
	.owner	 = THIS_MODULE,
	.open    = softnet_seq_open,
	.read    = seq_read,
	.llseek  = seq_lseek,
	.release = seq_release,
};

3738 3739 3740 3741 3742 3743 3744 3745 3746 3747 3748 3749
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;
	}

3750
	for (t = 0; t < PTYPE_HASH_SIZE; t++) {
3751 3752 3753 3754 3755 3756 3757 3758 3759 3760
		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)
3761
	__acquires(RCU)
3762 3763 3764 3765 3766 3767 3768 3769 3770 3771 3772 3773 3774 3775 3776 3777 3778 3779 3780 3781 3782 3783 3784
{
	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
3785
		hash = ntohs(pt->type) & PTYPE_HASH_MASK;
3786 3787

	while (nxt == &ptype_base[hash]) {
3788
		if (++hash >= PTYPE_HASH_SIZE)
3789 3790 3791 3792 3793 3794 3795 3796
			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)
3797
	__releases(RCU)
3798 3799 3800 3801 3802 3803 3804 3805 3806 3807
{
	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");
3808
	else if (pt->dev == NULL || dev_net(pt->dev) == seq_file_net(seq)) {
3809 3810 3811 3812 3813
		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
3814 3815
		seq_printf(seq, " %-8s %pF\n",
			   pt->dev ? pt->dev->name : "", pt->func);
3816 3817 3818 3819 3820 3821 3822 3823 3824 3825 3826 3827 3828 3829
	}

	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)
{
3830 3831
	return seq_open_net(inode, file, &ptype_seq_ops,
			sizeof(struct seq_net_private));
3832 3833 3834 3835 3836 3837 3838
}

static const struct file_operations ptype_seq_fops = {
	.owner	 = THIS_MODULE,
	.open    = ptype_seq_open,
	.read    = seq_read,
	.llseek  = seq_lseek,
3839
	.release = seq_release_net,
3840 3841 3842
};


3843
static int __net_init dev_proc_net_init(struct net *net)
L
Linus Torvalds 已提交
3844 3845 3846
{
	int rc = -ENOMEM;

3847
	if (!proc_net_fops_create(net, "dev", S_IRUGO, &dev_seq_fops))
L
Linus Torvalds 已提交
3848
		goto out;
3849
	if (!proc_net_fops_create(net, "softnet_stat", S_IRUGO, &softnet_seq_fops))
L
Linus Torvalds 已提交
3850
		goto out_dev;
3851
	if (!proc_net_fops_create(net, "ptype", S_IRUGO, &ptype_seq_fops))
3852
		goto out_softnet;
3853

3854
	if (wext_proc_init(net))
3855
		goto out_ptype;
L
Linus Torvalds 已提交
3856 3857 3858
	rc = 0;
out:
	return rc;
3859
out_ptype:
3860
	proc_net_remove(net, "ptype");
L
Linus Torvalds 已提交
3861
out_softnet:
3862
	proc_net_remove(net, "softnet_stat");
L
Linus Torvalds 已提交
3863
out_dev:
3864
	proc_net_remove(net, "dev");
L
Linus Torvalds 已提交
3865 3866
	goto out;
}
3867

3868
static void __net_exit dev_proc_net_exit(struct net *net)
3869 3870 3871 3872 3873 3874 3875 3876
{
	wext_proc_exit(net);

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

3877
static struct pernet_operations __net_initdata dev_proc_ops = {
3878 3879 3880 3881 3882 3883 3884 3885
	.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
3886 3887 3888 3889 3890 3891 3892 3893 3894 3895 3896 3897 3898 3899 3900 3901 3902 3903 3904 3905 3906 3907 3908 3909 3910 3911 3912 3913 3914
#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;
3915

E
Eric Dumazet 已提交
3916 3917
	if (old) {
		synchronize_net();
L
Linus Torvalds 已提交
3918
		dev_put(old);
E
Eric Dumazet 已提交
3919
	}
L
Linus Torvalds 已提交
3920 3921 3922 3923 3924 3925 3926 3927
	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
3928
EXPORT_SYMBOL(netdev_set_master);
L
Linus Torvalds 已提交
3929

3930 3931
static void dev_change_rx_flags(struct net_device *dev, int flags)
{
3932 3933 3934 3935
	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);
3936 3937
}

3938
static int __dev_set_promiscuity(struct net_device *dev, int inc)
L
Linus Torvalds 已提交
3939 3940
{
	unsigned short old_flags = dev->flags;
3941 3942
	uid_t uid;
	gid_t gid;
L
Linus Torvalds 已提交
3943

3944 3945
	ASSERT_RTNL();

3946 3947 3948 3949 3950 3951 3952 3953 3954 3955 3956 3957 3958 3959 3960 3961 3962
	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;
		}
	}
3963
	if (dev->flags != old_flags) {
L
Linus Torvalds 已提交
3964 3965
		printk(KERN_INFO "device %s %s promiscuous mode\n",
		       dev->name, (dev->flags & IFF_PROMISC) ? "entered" :
3966
							       "left");
3967 3968
		if (audit_enabled) {
			current_uid_gid(&uid, &gid);
3969 3970 3971 3972 3973 3974
			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),
3975
				uid, gid,
3976
				audit_get_sessionid(current));
3977
		}
3978

3979
		dev_change_rx_flags(dev, IFF_PROMISC);
L
Linus Torvalds 已提交
3980
	}
3981
	return 0;
L
Linus Torvalds 已提交
3982 3983
}

3984 3985 3986 3987 3988 3989 3990 3991 3992
/**
 *	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.
3993
 *	Return 0 if successful or a negative errno code on error.
3994
 */
3995
int dev_set_promiscuity(struct net_device *dev, int inc)
3996 3997
{
	unsigned short old_flags = dev->flags;
3998
	int err;
3999

4000
	err = __dev_set_promiscuity(dev, inc);
4001
	if (err < 0)
4002
		return err;
4003 4004
	if (dev->flags != old_flags)
		dev_set_rx_mode(dev);
4005
	return err;
4006
}
E
Eric Dumazet 已提交
4007
EXPORT_SYMBOL(dev_set_promiscuity);
4008

L
Linus Torvalds 已提交
4009 4010 4011 4012 4013 4014 4015 4016 4017 4018
/**
 *	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.
4019
 *	Return 0 if successful or a negative errno code on error.
L
Linus Torvalds 已提交
4020 4021
 */

4022
int dev_set_allmulti(struct net_device *dev, int inc)
L
Linus Torvalds 已提交
4023 4024 4025
{
	unsigned short old_flags = dev->flags;

4026 4027
	ASSERT_RTNL();

L
Linus Torvalds 已提交
4028
	dev->flags |= IFF_ALLMULTI;
4029 4030 4031 4032 4033 4034 4035 4036 4037 4038 4039 4040 4041 4042 4043 4044
	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;
		}
	}
4045
	if (dev->flags ^ old_flags) {
4046
		dev_change_rx_flags(dev, IFF_ALLMULTI);
4047
		dev_set_rx_mode(dev);
4048
	}
4049
	return 0;
4050
}
E
Eric Dumazet 已提交
4051
EXPORT_SYMBOL(dev_set_allmulti);
4052 4053 4054 4055

/*
 *	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
4056
 *	filtering it is put in promiscuous mode while unicast addresses
4057 4058 4059 4060
 *	are present.
 */
void __dev_set_rx_mode(struct net_device *dev)
{
4061 4062
	const struct net_device_ops *ops = dev->netdev_ops;

4063 4064 4065 4066 4067
	/* dev_open will call this function so the list will stay sane. */
	if (!(dev->flags&IFF_UP))
		return;

	if (!netif_device_present(dev))
4068
		return;
4069

4070 4071
	if (ops->ndo_set_rx_mode)
		ops->ndo_set_rx_mode(dev);
4072 4073 4074 4075
	else {
		/* Unicast addresses changes may only happen under the rtnl,
		 * therefore calling __dev_set_promiscuity here is safe.
		 */
4076
		if (!netdev_uc_empty(dev) && !dev->uc_promisc) {
4077 4078
			__dev_set_promiscuity(dev, 1);
			dev->uc_promisc = 1;
4079
		} else if (netdev_uc_empty(dev) && dev->uc_promisc) {
4080 4081 4082 4083
			__dev_set_promiscuity(dev, -1);
			dev->uc_promisc = 0;
		}

4084 4085
		if (ops->ndo_set_multicast_list)
			ops->ndo_set_multicast_list(dev);
4086 4087 4088 4089 4090
	}
}

void dev_set_rx_mode(struct net_device *dev)
{
4091
	netif_addr_lock_bh(dev);
4092
	__dev_set_rx_mode(dev);
4093
	netif_addr_unlock_bh(dev);
L
Linus Torvalds 已提交
4094 4095
}

4096 4097 4098 4099 4100 4101
/**
 *	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
4102 4103 4104 4105 4106 4107
unsigned dev_get_flags(const struct net_device *dev)
{
	unsigned flags;

	flags = (dev->flags & ~(IFF_PROMISC |
				IFF_ALLMULTI |
S
Stefan Rompf 已提交
4108 4109 4110
				IFF_RUNNING |
				IFF_LOWER_UP |
				IFF_DORMANT)) |
L
Linus Torvalds 已提交
4111 4112 4113
		(dev->gflags & (IFF_PROMISC |
				IFF_ALLMULTI));

S
Stefan Rompf 已提交
4114 4115 4116 4117 4118 4119 4120 4121
	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
4122 4123 4124

	return flags;
}
E
Eric Dumazet 已提交
4125
EXPORT_SYMBOL(dev_get_flags);
L
Linus Torvalds 已提交
4126

4127
int __dev_change_flags(struct net_device *dev, unsigned int flags)
L
Linus Torvalds 已提交
4128 4129
{
	int old_flags = dev->flags;
4130
	int ret;
L
Linus Torvalds 已提交
4131

4132 4133
	ASSERT_RTNL();

L
Linus Torvalds 已提交
4134 4135 4136 4137 4138 4139 4140 4141 4142 4143 4144 4145 4146 4147
	/*
	 *	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.
	 */

4148 4149
	if ((old_flags ^ flags) & IFF_MULTICAST)
		dev_change_rx_flags(dev, IFF_MULTICAST);
4150

4151
	dev_set_rx_mode(dev);
L
Linus Torvalds 已提交
4152 4153 4154 4155 4156 4157 4158 4159 4160

	/*
	 *	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  ? */
4161
		ret = ((old_flags & IFF_UP) ? __dev_close : __dev_open)(dev);
L
Linus Torvalds 已提交
4162 4163

		if (!ret)
4164
			dev_set_rx_mode(dev);
L
Linus Torvalds 已提交
4165 4166 4167
	}

	if ((flags ^ dev->gflags) & IFF_PROMISC) {
E
Eric Dumazet 已提交
4168 4169
		int inc = (flags & IFF_PROMISC) ? 1 : -1;

L
Linus Torvalds 已提交
4170 4171 4172 4173 4174 4175 4176 4177 4178
		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
4179 4180
		int inc = (flags & IFF_ALLMULTI) ? 1 : -1;

L
Linus Torvalds 已提交
4181 4182 4183 4184
		dev->gflags ^= IFF_ALLMULTI;
		dev_set_allmulti(dev, inc);
	}

4185 4186 4187 4188 4189 4190 4191 4192 4193 4194 4195 4196 4197 4198 4199 4200 4201 4202 4203 4204 4205 4206 4207 4208 4209 4210 4211 4212 4213 4214 4215 4216 4217 4218 4219 4220 4221
	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;
4222 4223
	if (changes)
		rtmsg_ifinfo(RTM_NEWLINK, dev, changes);
L
Linus Torvalds 已提交
4224

4225
	__dev_notify_flags(dev, old_flags);
L
Linus Torvalds 已提交
4226 4227
	return ret;
}
E
Eric Dumazet 已提交
4228
EXPORT_SYMBOL(dev_change_flags);
L
Linus Torvalds 已提交
4229

4230 4231 4232 4233 4234 4235 4236
/**
 *	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
4237 4238
int dev_set_mtu(struct net_device *dev, int new_mtu)
{
4239
	const struct net_device_ops *ops = dev->netdev_ops;
L
Linus Torvalds 已提交
4240 4241 4242 4243 4244 4245 4246 4247 4248 4249 4250 4251 4252
	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;
4253 4254
	if (ops->ndo_change_mtu)
		err = ops->ndo_change_mtu(dev, new_mtu);
L
Linus Torvalds 已提交
4255 4256
	else
		dev->mtu = new_mtu;
4257

L
Linus Torvalds 已提交
4258
	if (!err && dev->flags & IFF_UP)
4259
		call_netdevice_notifiers(NETDEV_CHANGEMTU, dev);
L
Linus Torvalds 已提交
4260 4261
	return err;
}
E
Eric Dumazet 已提交
4262
EXPORT_SYMBOL(dev_set_mtu);
L
Linus Torvalds 已提交
4263

4264 4265 4266 4267 4268 4269 4270
/**
 *	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
4271 4272
int dev_set_mac_address(struct net_device *dev, struct sockaddr *sa)
{
4273
	const struct net_device_ops *ops = dev->netdev_ops;
L
Linus Torvalds 已提交
4274 4275
	int err;

4276
	if (!ops->ndo_set_mac_address)
L
Linus Torvalds 已提交
4277 4278 4279 4280 4281
		return -EOPNOTSUPP;
	if (sa->sa_family != dev->type)
		return -EINVAL;
	if (!netif_device_present(dev))
		return -ENODEV;
4282
	err = ops->ndo_set_mac_address(dev, sa);
L
Linus Torvalds 已提交
4283
	if (!err)
4284
		call_netdevice_notifiers(NETDEV_CHANGEADDR, dev);
L
Linus Torvalds 已提交
4285 4286
	return err;
}
E
Eric Dumazet 已提交
4287
EXPORT_SYMBOL(dev_set_mac_address);
L
Linus Torvalds 已提交
4288 4289

/*
E
Eric Dumazet 已提交
4290
 *	Perform the SIOCxIFxxx calls, inside rcu_read_lock()
L
Linus Torvalds 已提交
4291
 */
4292
static int dev_ifsioc_locked(struct net *net, struct ifreq *ifr, unsigned int cmd)
L
Linus Torvalds 已提交
4293 4294
{
	int err;
E
Eric Dumazet 已提交
4295
	struct net_device *dev = dev_get_by_name_rcu(net, ifr->ifr_name);
L
Linus Torvalds 已提交
4296 4297 4298 4299 4300

	if (!dev)
		return -ENODEV;

	switch (cmd) {
E
Eric Dumazet 已提交
4301 4302 4303
	case SIOCGIFFLAGS:	/* Get interface flags */
		ifr->ifr_flags = (short) dev_get_flags(dev);
		return 0;
L
Linus Torvalds 已提交
4304

E
Eric Dumazet 已提交
4305 4306 4307 4308
	case SIOCGIFMETRIC:	/* Get the metric on the interface
				   (currently unused) */
		ifr->ifr_metric = 0;
		return 0;
L
Linus Torvalds 已提交
4309

E
Eric Dumazet 已提交
4310 4311 4312
	case SIOCGIFMTU:	/* Get the MTU of a device */
		ifr->ifr_mtu = dev->mtu;
		return 0;
L
Linus Torvalds 已提交
4313

E
Eric Dumazet 已提交
4314 4315 4316 4317 4318 4319 4320 4321
	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
4322

E
Eric Dumazet 已提交
4323 4324 4325
	case SIOCGIFSLAVE:
		err = -EINVAL;
		break;
4326

E
Eric Dumazet 已提交
4327 4328 4329 4330 4331 4332 4333 4334
	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;
4335

E
Eric Dumazet 已提交
4336 4337 4338
	case SIOCGIFINDEX:
		ifr->ifr_ifindex = dev->ifindex;
		return 0;
4339

E
Eric Dumazet 已提交
4340 4341 4342
	case SIOCGIFTXQLEN:
		ifr->ifr_qlen = dev->tx_queue_len;
		return 0;
4343

E
Eric Dumazet 已提交
4344 4345 4346 4347 4348 4349 4350
	default:
		/* dev_ioctl() should ensure this case
		 * is never reached
		 */
		WARN_ON(1);
		err = -EINVAL;
		break;
4351 4352 4353 4354 4355 4356 4357 4358 4359 4360 4361 4362

	}
	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
4363
	const struct net_device_ops *ops;
4364 4365 4366 4367

	if (!dev)
		return -ENODEV;

J
Jarek Poplawski 已提交
4368 4369
	ops = dev->netdev_ops;

4370
	switch (cmd) {
E
Eric Dumazet 已提交
4371 4372
	case SIOCSIFFLAGS:	/* Set interface flags */
		return dev_change_flags(dev, ifr->ifr_flags);
4373

E
Eric Dumazet 已提交
4374 4375 4376
	case SIOCSIFMETRIC:	/* Set the metric on the interface
				   (currently unused) */
		return -EOPNOTSUPP;
4377

E
Eric Dumazet 已提交
4378 4379
	case SIOCSIFMTU:	/* Set the MTU of a device */
		return dev_set_mtu(dev, ifr->ifr_mtu);
L
Linus Torvalds 已提交
4380

E
Eric Dumazet 已提交
4381 4382
	case SIOCSIFHWADDR:
		return dev_set_mac_address(dev, &ifr->ifr_hwaddr);
L
Linus Torvalds 已提交
4383

E
Eric Dumazet 已提交
4384 4385 4386 4387 4388 4389 4390
	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
4391

E
Eric Dumazet 已提交
4392 4393
	case SIOCSIFMAP:
		if (ops->ndo_set_config) {
L
Linus Torvalds 已提交
4394 4395
			if (!netif_device_present(dev))
				return -ENODEV;
E
Eric Dumazet 已提交
4396 4397 4398
			return ops->ndo_set_config(dev, &ifr->ifr_map);
		}
		return -EOPNOTSUPP;
L
Linus Torvalds 已提交
4399

E
Eric Dumazet 已提交
4400 4401 4402 4403 4404 4405
	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;
4406
		return dev_mc_add_global(dev, ifr->ifr_hwaddr.sa_data);
E
Eric Dumazet 已提交
4407 4408 4409 4410 4411 4412 4413

	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;
4414
		return dev_mc_del_global(dev, ifr->ifr_hwaddr.sa_data);
L
Linus Torvalds 已提交
4415

E
Eric Dumazet 已提交
4416 4417 4418 4419 4420
	case SIOCSIFTXQLEN:
		if (ifr->ifr_qlen < 0)
			return -EINVAL;
		dev->tx_queue_len = ifr->ifr_qlen;
		return 0;
L
Linus Torvalds 已提交
4421

E
Eric Dumazet 已提交
4422 4423 4424
	case SIOCSIFNAME:
		ifr->ifr_newname[IFNAMSIZ-1] = '\0';
		return dev_change_name(dev, ifr->ifr_newname);
L
Linus Torvalds 已提交
4425

E
Eric Dumazet 已提交
4426 4427 4428 4429 4430 4431 4432 4433 4434 4435 4436 4437 4438 4439 4440 4441 4442 4443 4444 4445 4446 4447 4448 4449 4450 4451 4452 4453
	/*
	 *	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
4454 4455 4456 4457 4458 4459 4460 4461 4462 4463 4464 4465

	}
	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
4466
 *	@net: the applicable net namespace
L
Linus Torvalds 已提交
4467 4468 4469 4470 4471 4472 4473 4474 4475
 *	@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.
 */

4476
int dev_ioctl(struct net *net, unsigned int cmd, void __user *arg)
L
Linus Torvalds 已提交
4477 4478 4479 4480 4481 4482 4483 4484 4485 4486 4487
{
	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) {
4488
		rtnl_lock();
4489
		ret = dev_ifconf(net, (char __user *) arg);
4490
		rtnl_unlock();
L
Linus Torvalds 已提交
4491 4492 4493
		return ret;
	}
	if (cmd == SIOCGIFNAME)
4494
		return dev_ifname(net, (struct ifreq __user *)arg);
L
Linus Torvalds 已提交
4495 4496 4497 4498 4499 4500 4501 4502 4503 4504 4505 4506 4507 4508 4509

	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
4510 4511 4512 4513 4514 4515 4516 4517 4518 4519 4520 4521 4522 4523 4524
	/*
	 *	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
4525
		rcu_read_lock();
E
Eric Dumazet 已提交
4526
		ret = dev_ifsioc_locked(net, &ifr, cmd);
E
Eric Dumazet 已提交
4527
		rcu_read_unlock();
E
Eric Dumazet 已提交
4528 4529 4530 4531 4532 4533 4534 4535
		if (!ret) {
			if (colon)
				*colon = ':';
			if (copy_to_user(arg, &ifr,
					 sizeof(struct ifreq)))
				ret = -EFAULT;
		}
		return ret;
L
Linus Torvalds 已提交
4536

E
Eric Dumazet 已提交
4537 4538 4539 4540 4541 4542 4543 4544 4545 4546 4547 4548 4549
	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
4550

E
Eric Dumazet 已提交
4551 4552 4553 4554 4555 4556 4557 4558 4559 4560 4561 4562 4563 4564 4565 4566 4567 4568 4569 4570 4571 4572 4573
	/*
	 *	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
4574

E
Eric Dumazet 已提交
4575 4576 4577 4578 4579 4580 4581 4582 4583 4584 4585 4586 4587 4588 4589 4590 4591 4592 4593 4594 4595 4596 4597 4598 4599 4600 4601 4602 4603 4604 4605 4606 4607 4608 4609 4610 4611 4612 4613 4614 4615 4616 4617 4618 4619 4620 4621 4622 4623 4624 4625
	/*
	 *	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)) {
4626
			dev_load(net, ifr.ifr_name);
L
Linus Torvalds 已提交
4627
			rtnl_lock();
4628
			ret = dev_ifsioc(net, &ifr, cmd);
L
Linus Torvalds 已提交
4629
			rtnl_unlock();
E
Eric Dumazet 已提交
4630 4631 4632
			if (!ret && copy_to_user(arg, &ifr,
						 sizeof(struct ifreq)))
				ret = -EFAULT;
L
Linus Torvalds 已提交
4633
			return ret;
E
Eric Dumazet 已提交
4634 4635 4636 4637 4638
		}
		/* 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
4639 4640 4641 4642 4643 4644
	}
}


/**
 *	dev_new_index	-	allocate an ifindex
4645
 *	@net: the applicable net namespace
L
Linus Torvalds 已提交
4646 4647 4648 4649 4650
 *
 *	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.
 */
4651
static int dev_new_index(struct net *net)
L
Linus Torvalds 已提交
4652 4653 4654 4655 4656
{
	static int ifindex;
	for (;;) {
		if (++ifindex <= 0)
			ifindex = 1;
4657
		if (!__dev_get_by_index(net, ifindex))
L
Linus Torvalds 已提交
4658 4659 4660 4661 4662
			return ifindex;
	}
}

/* Delayed registration/unregisteration */
4663
static LIST_HEAD(net_todo_list);
L
Linus Torvalds 已提交
4664

4665
static void net_set_todo(struct net_device *dev)
L
Linus Torvalds 已提交
4666 4667 4668 4669
{
	list_add_tail(&dev->todo_list, &net_todo_list);
}

4670
static void rollback_registered_many(struct list_head *head)
4671
{
4672
	struct net_device *dev, *tmp;
4673

4674 4675 4676
	BUG_ON(dev_boot_phase);
	ASSERT_RTNL();

4677
	list_for_each_entry_safe(dev, tmp, head, unreg_list) {
4678
		/* Some devices call without registering
4679 4680
		 * for initialization unwind. Remove those
		 * devices and proceed with the remaining.
4681 4682 4683 4684
		 */
		if (dev->reg_state == NETREG_UNINITIALIZED) {
			pr_debug("unregister_netdevice: device %s/%p never "
				 "was registered\n", dev->name, dev);
4685

4686
			WARN_ON(1);
4687 4688
			list_del(&dev->unreg_list);
			continue;
4689
		}
4690

4691
		BUG_ON(dev->reg_state != NETREG_REGISTERED);
4692

4693 4694
		/* If device is running, close it first. */
		dev_close(dev);
4695

4696 4697
		/* And unlink it from device chain. */
		unlist_netdevice(dev);
4698

4699 4700
		dev->reg_state = NETREG_UNREGISTERING;
	}
4701 4702 4703

	synchronize_net();

4704 4705 4706
	list_for_each_entry(dev, head, unreg_list) {
		/* Shutdown queueing discipline. */
		dev_shutdown(dev);
4707 4708


4709 4710 4711 4712
		/* Notify protocols, that we are about to destroy
		   this device. They should clean all the things.
		*/
		call_netdevice_notifiers(NETDEV_UNREGISTER, dev);
4713

4714 4715 4716 4717
		if (!dev->rtnl_link_ops ||
		    dev->rtnl_link_state == RTNL_LINK_INITIALIZED)
			rtmsg_ifinfo(RTM_DELLINK, dev, ~0U);

4718 4719 4720
		/*
		 *	Flush the unicast and multicast chains
		 */
4721
		dev_uc_flush(dev);
4722
		dev_mc_flush(dev);
4723

4724 4725
		if (dev->netdev_ops->ndo_uninit)
			dev->netdev_ops->ndo_uninit(dev);
4726

4727 4728
		/* Notifier chain MUST detach us from master device. */
		WARN_ON(dev->master);
4729

4730 4731 4732
		/* Remove entries from kobject tree */
		netdev_unregister_kobject(dev);
	}
4733

4734
	/* Process any work delayed until the end of the batch */
4735
	dev = list_first_entry(head, struct net_device, unreg_list);
4736
	call_netdevice_notifiers(NETDEV_UNREGISTER_BATCH, dev);
4737

4738
	synchronize_net();
4739

4740
	list_for_each_entry(dev, head, unreg_list)
4741 4742 4743 4744 4745 4746 4747 4748 4749
		dev_put(dev);
}

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

	list_add(&dev->unreg_list, &single);
	rollback_registered_many(&single);
4750 4751
}

4752 4753 4754
static void __netdev_init_queue_locks_one(struct net_device *dev,
					  struct netdev_queue *dev_queue,
					  void *_unused)
4755 4756
{
	spin_lock_init(&dev_queue->_xmit_lock);
4757
	netdev_set_xmit_lockdep_class(&dev_queue->_xmit_lock, dev->type);
4758 4759 4760 4761 4762
	dev_queue->xmit_lock_owner = -1;
}

static void netdev_init_queue_locks(struct net_device *dev)
{
4763 4764
	netdev_for_each_tx_queue(dev, __netdev_init_queue_locks_one, NULL);
	__netdev_init_queue_locks_one(dev, &dev->rx_queue, NULL);
4765 4766
}

4767 4768 4769 4770 4771 4772 4773 4774 4775 4776 4777 4778 4779 4780 4781 4782 4783 4784 4785 4786 4787 4788 4789 4790 4791 4792 4793 4794 4795 4796 4797 4798 4799 4800 4801 4802 4803 4804 4805 4806
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);

4807 4808 4809 4810 4811 4812 4813 4814 4815 4816 4817 4818 4819 4820 4821 4822 4823 4824 4825 4826 4827 4828 4829 4830 4831 4832 4833
/**
 *	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
4834 4835 4836 4837 4838 4839 4840 4841 4842 4843 4844 4845 4846 4847 4848 4849 4850 4851 4852 4853
/**
 *	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;
4854
	struct net *net = dev_net(dev);
L
Linus Torvalds 已提交
4855 4856 4857 4858

	BUG_ON(dev_boot_phase);
	ASSERT_RTNL();

4859 4860
	might_sleep();

L
Linus Torvalds 已提交
4861 4862
	/* When net_device's are persistent, this will be fatal. */
	BUG_ON(dev->reg_state != NETREG_UNINITIALIZED);
4863
	BUG_ON(!net);
L
Linus Torvalds 已提交
4864

4865
	spin_lock_init(&dev->addr_list_lock);
4866
	netdev_set_addr_lockdep_class(dev);
4867
	netdev_init_queue_locks(dev);
L
Linus Torvalds 已提交
4868 4869 4870

	dev->iflink = -1;

E
Eric Dumazet 已提交
4871
#ifdef CONFIG_RPS
T
Tom Herbert 已提交
4872 4873 4874 4875 4876 4877 4878 4879 4880 4881 4882 4883 4884 4885 4886 4887
	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
4888
#endif
L
Linus Torvalds 已提交
4889
	/* Init, if this function is available */
4890 4891
	if (dev->netdev_ops->ndo_init) {
		ret = dev->netdev_ops->ndo_init(dev);
L
Linus Torvalds 已提交
4892 4893 4894
		if (ret) {
			if (ret > 0)
				ret = -EIO;
4895
			goto out;
L
Linus Torvalds 已提交
4896 4897
		}
	}
4898

4899 4900
	ret = dev_get_valid_name(net, dev->name, dev->name, 0);
	if (ret)
4901
		goto err_uninit;
L
Linus Torvalds 已提交
4902

4903
	dev->ifindex = dev_new_index(net);
L
Linus Torvalds 已提交
4904 4905 4906
	if (dev->iflink == -1)
		dev->iflink = dev->ifindex;

4907 4908 4909 4910 4911 4912 4913 4914 4915 4916 4917 4918 4919 4920 4921
	/* 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);
	}

4922
	dev->features = netdev_fix_features(dev->features, dev->name);
L
Linus Torvalds 已提交
4923

4924 4925 4926 4927
	/* Enable software GSO if SG is supported. */
	if (dev->features & NETIF_F_SG)
		dev->features |= NETIF_F_GSO;

4928
	netdev_initialize_kobject(dev);
4929 4930 4931 4932 4933 4934

	ret = call_netdevice_notifiers(NETDEV_POST_INIT, dev);
	ret = notifier_to_errno(ret);
	if (ret)
		goto err_uninit;

4935
	ret = netdev_register_kobject(dev);
4936
	if (ret)
4937
		goto err_uninit;
4938 4939
	dev->reg_state = NETREG_REGISTERED;

L
Linus Torvalds 已提交
4940 4941 4942 4943 4944 4945 4946 4947 4948
	/*
	 *	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);
4949
	list_netdevice(dev);
L
Linus Torvalds 已提交
4950 4951

	/* Notify protocols, that a new device appeared. */
4952
	ret = call_netdevice_notifiers(NETDEV_REGISTER, dev);
4953
	ret = notifier_to_errno(ret);
4954 4955 4956 4957
	if (ret) {
		rollback_registered(dev);
		dev->reg_state = NETREG_UNREGISTERED;
	}
4958 4959 4960 4961
	/*
	 *	Prevent userspace races by waiting until the network
	 *	device is fully setup before sending notifications.
	 */
4962 4963 4964
	if (!dev->rtnl_link_ops ||
	    dev->rtnl_link_state == RTNL_LINK_INITIALIZED)
		rtmsg_ifinfo(RTM_NEWLINK, dev, ~0U);
L
Linus Torvalds 已提交
4965 4966 4967

out:
	return ret;
4968 4969

err_uninit:
4970 4971
	if (dev->netdev_ops->ndo_uninit)
		dev->netdev_ops->ndo_uninit(dev);
4972
	goto out;
L
Linus Torvalds 已提交
4973
}
E
Eric Dumazet 已提交
4974
EXPORT_SYMBOL(register_netdevice);
L
Linus Torvalds 已提交
4975

4976 4977 4978 4979 4980 4981 4982 4983 4984 4985 4986 4987 4988 4989 4990 4991 4992 4993 4994 4995 4996 4997 4998 4999 5000 5001 5002 5003 5004 5005 5006 5007 5008 5009 5010 5011 5012 5013 5014
/**
 *	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
5015 5016 5017 5018 5019 5020 5021 5022 5023
/**
 *	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.
 *
5024
 *	This is a wrapper around register_netdevice that takes the rtnl semaphore
L
Linus Torvalds 已提交
5025 5026 5027 5028 5029 5030 5031 5032 5033 5034 5035 5036 5037 5038 5039 5040 5041 5042
 *	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;
	}
5043

L
Linus Torvalds 已提交
5044 5045 5046 5047 5048 5049 5050 5051 5052 5053 5054 5055 5056 5057 5058 5059
	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
5060
 * call dev_put.
L
Linus Torvalds 已提交
5061 5062 5063 5064 5065
 */
static void netdev_wait_allrefs(struct net_device *dev)
{
	unsigned long rebroadcast_time, warning_time;

5066 5067
	linkwatch_forget_dev(dev);

L
Linus Torvalds 已提交
5068 5069 5070
	rebroadcast_time = warning_time = jiffies;
	while (atomic_read(&dev->refcnt) != 0) {
		if (time_after(jiffies, rebroadcast_time + 1 * HZ)) {
5071
			rtnl_lock();
L
Linus Torvalds 已提交
5072 5073

			/* Rebroadcast unregister notification */
5074
			call_netdevice_notifiers(NETDEV_UNREGISTER, dev);
5075
			/* don't resend NETDEV_UNREGISTER_BATCH, _BATCH users
5076
			 * should have already handle it the first time */
L
Linus Torvalds 已提交
5077 5078 5079 5080 5081 5082 5083 5084 5085 5086 5087 5088

			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();
			}

5089
			__rtnl_unlock();
L
Linus Torvalds 已提交
5090 5091 5092 5093 5094 5095 5096 5097 5098 5099 5100 5101 5102 5103 5104 5105 5106 5107 5108 5109 5110 5111 5112 5113 5114 5115 5116 5117 5118 5119

			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
5120
 * We are invoked by rtnl_unlock().
L
Linus Torvalds 已提交
5121
 * This allows us to deal with problems:
5122
 * 1) We can delete sysfs objects which invoke hotplug
L
Linus Torvalds 已提交
5123 5124 5125
 *    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
5126 5127 5128
 *
 * 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
5129 5130 5131
 */
void netdev_run_todo(void)
{
5132
	struct list_head list;
L
Linus Torvalds 已提交
5133 5134

	/* Snapshot list, allow later requests */
5135
	list_replace_init(&net_todo_list, &list);
H
Herbert Xu 已提交
5136 5137

	__rtnl_unlock();
5138

L
Linus Torvalds 已提交
5139 5140
	while (!list_empty(&list)) {
		struct net_device *dev
5141
			= list_first_entry(&list, struct net_device, todo_list);
L
Linus Torvalds 已提交
5142 5143
		list_del(&dev->todo_list);

5144 5145 5146 5147 5148 5149
		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
5150

5151
		dev->reg_state = NETREG_UNREGISTERED;
L
Linus Torvalds 已提交
5152

5153
		on_each_cpu(flush_backlog, dev, 1);
5154

5155
		netdev_wait_allrefs(dev);
L
Linus Torvalds 已提交
5156

5157 5158
		/* paranoia */
		BUG_ON(atomic_read(&dev->refcnt));
5159 5160 5161
		WARN_ON(dev->ip_ptr);
		WARN_ON(dev->ip6_ptr);
		WARN_ON(dev->dn_ptr);
L
Linus Torvalds 已提交
5162

5163 5164
		if (dev->destructor)
			dev->destructor(dev);
5165 5166 5167

		/* Free network device */
		kobject_put(&dev->dev.kobj);
L
Linus Torvalds 已提交
5168 5169 5170
	}
}

5171 5172 5173 5174 5175 5176 5177 5178 5179 5180 5181 5182 5183 5184 5185 5186 5187 5188 5189 5190 5191 5192 5193 5194 5195 5196
/**
 *	dev_txq_stats_fold - fold tx_queues stats
 *	@dev: device to get statistics from
 *	@stats: struct net_device_stats to hold results
 */
void dev_txq_stats_fold(const struct net_device *dev,
			struct net_device_stats *stats)
{
	unsigned long tx_bytes = 0, tx_packets = 0, tx_dropped = 0;
	unsigned int i;
	struct netdev_queue *txq;

	for (i = 0; i < dev->num_tx_queues; i++) {
		txq = netdev_get_tx_queue(dev, i);
		tx_bytes   += txq->tx_bytes;
		tx_packets += txq->tx_packets;
		tx_dropped += txq->tx_dropped;
	}
	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);

5197 5198 5199 5200 5201 5202 5203 5204 5205
/**
 *	dev_get_stats	- get network device statistics
 *	@dev: device to get statistics from
 *
 *	Get network statistics from device. The device driver may provide
 *	its own method by setting dev->netdev_ops->get_stats; otherwise
 *	the internal statistics structure is used.
 */
const struct net_device_stats *dev_get_stats(struct net_device *dev)
5206
{
5207 5208 5209 5210
	const struct net_device_ops *ops = dev->netdev_ops;

	if (ops->ndo_get_stats)
		return ops->ndo_get_stats(dev);
5211 5212 5213

	dev_txq_stats_fold(dev, &dev->stats);
	return &dev->stats;
R
Rusty Russell 已提交
5214
}
5215
EXPORT_SYMBOL(dev_get_stats);
R
Rusty Russell 已提交
5216

5217
static void netdev_init_one_queue(struct net_device *dev,
5218 5219
				  struct netdev_queue *queue,
				  void *_unused)
5220 5221 5222 5223
{
	queue->dev = dev;
}

5224 5225
static void netdev_init_queues(struct net_device *dev)
{
5226 5227
	netdev_init_one_queue(dev, &dev->rx_queue, NULL);
	netdev_for_each_tx_queue(dev, netdev_init_one_queue, NULL);
5228
	spin_lock_init(&dev->tx_global_lock);
5229 5230
}

L
Linus Torvalds 已提交
5231
/**
5232
 *	alloc_netdev_mq - allocate network device
L
Linus Torvalds 已提交
5233 5234 5235
 *	@sizeof_priv:	size of private data to allocate space for
 *	@name:		device name format string
 *	@setup:		callback to initialize device
5236
 *	@queue_count:	the number of subqueues to allocate
L
Linus Torvalds 已提交
5237 5238
 *
 *	Allocates a struct net_device with private data area for driver use
5239 5240
 *	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
5241
 */
5242 5243
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
5244
{
5245
	struct netdev_queue *tx;
L
Linus Torvalds 已提交
5246
	struct net_device *dev;
5247
	size_t alloc_size;
5248
	struct net_device *p;
E
Eric Dumazet 已提交
5249 5250
#ifdef CONFIG_RPS
	struct netdev_rx_queue *rx;
T
Tom Herbert 已提交
5251
	int i;
E
Eric Dumazet 已提交
5252
#endif
L
Linus Torvalds 已提交
5253

5254 5255
	BUG_ON(strlen(name) >= sizeof(dev->name));

5256
	alloc_size = sizeof(struct net_device);
5257 5258
	if (sizeof_priv) {
		/* ensure 32-byte alignment of private area */
5259
		alloc_size = ALIGN(alloc_size, NETDEV_ALIGN);
5260 5261 5262
		alloc_size += sizeof_priv;
	}
	/* ensure 32-byte alignment of whole construct */
5263
	alloc_size += NETDEV_ALIGN - 1;
L
Linus Torvalds 已提交
5264

5265
	p = kzalloc(alloc_size, GFP_KERNEL);
L
Linus Torvalds 已提交
5266
	if (!p) {
5267
		printk(KERN_ERR "alloc_netdev: Unable to allocate device.\n");
L
Linus Torvalds 已提交
5268 5269 5270
		return NULL;
	}

5271
	tx = kcalloc(queue_count, sizeof(struct netdev_queue), GFP_KERNEL);
5272 5273 5274
	if (!tx) {
		printk(KERN_ERR "alloc_netdev: Unable to allocate "
		       "tx qdiscs.\n");
5275
		goto free_p;
5276 5277
	}

E
Eric Dumazet 已提交
5278
#ifdef CONFIG_RPS
T
Tom Herbert 已提交
5279 5280 5281 5282 5283 5284 5285 5286 5287 5288 5289 5290 5291 5292 5293
	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
5294
#endif
T
Tom Herbert 已提交
5295

5296
	dev = PTR_ALIGN(p, NETDEV_ALIGN);
L
Linus Torvalds 已提交
5297
	dev->padded = (char *)dev - (char *)p;
5298 5299

	if (dev_addr_init(dev))
T
Tom Herbert 已提交
5300
		goto free_rx;
5301

5302
	dev_mc_init(dev);
5303
	dev_uc_init(dev);
J
Jiri Pirko 已提交
5304

5305
	dev_net_set(dev, &init_net);
L
Linus Torvalds 已提交
5306

5307 5308
	dev->_tx = tx;
	dev->num_tx_queues = queue_count;
5309
	dev->real_num_tx_queues = queue_count;
5310

E
Eric Dumazet 已提交
5311
#ifdef CONFIG_RPS
T
Tom Herbert 已提交
5312 5313
	dev->_rx = rx;
	dev->num_rx_queues = queue_count;
E
Eric Dumazet 已提交
5314
#endif
T
Tom Herbert 已提交
5315

5316
	dev->gso_max_size = GSO_MAX_SIZE;
L
Linus Torvalds 已提交
5317

5318 5319
	netdev_init_queues(dev);

5320 5321
	INIT_LIST_HEAD(&dev->ethtool_ntuple_list.list);
	dev->ethtool_ntuple_list.count = 0;
5322
	INIT_LIST_HEAD(&dev->napi_list);
5323
	INIT_LIST_HEAD(&dev->unreg_list);
5324
	INIT_LIST_HEAD(&dev->link_watch_list);
5325
	dev->priv_flags = IFF_XMIT_DST_RELEASE;
L
Linus Torvalds 已提交
5326 5327 5328
	setup(dev);
	strcpy(dev->name, name);
	return dev;
5329

T
Tom Herbert 已提交
5330
free_rx:
E
Eric Dumazet 已提交
5331
#ifdef CONFIG_RPS
T
Tom Herbert 已提交
5332
	kfree(rx);
5333
free_tx:
E
Eric Dumazet 已提交
5334
#endif
5335 5336 5337 5338
	kfree(tx);
free_p:
	kfree(p);
	return NULL;
L
Linus Torvalds 已提交
5339
}
5340
EXPORT_SYMBOL(alloc_netdev_mq);
L
Linus Torvalds 已提交
5341 5342 5343 5344 5345

/**
 *	free_netdev - free network device
 *	@dev: device
 *
5346 5347
 *	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
5348 5349 5350 5351
 *	If this is the last reference then it will be freed.
 */
void free_netdev(struct net_device *dev)
{
5352 5353
	struct napi_struct *p, *n;

5354 5355
	release_net(dev_net(dev));

5356 5357
	kfree(dev->_tx);

5358 5359 5360
	/* Flush device addresses */
	dev_addr_flush(dev);

5361 5362 5363
	/* Clear ethtool n-tuple list */
	ethtool_ntuple_flush(dev);

5364 5365 5366
	list_for_each_entry_safe(p, n, &dev->napi_list, dev_list)
		netif_napi_del(p);

S
Stephen Hemminger 已提交
5367
	/*  Compatibility with error handling in drivers */
L
Linus Torvalds 已提交
5368 5369 5370 5371 5372 5373 5374 5375
	if (dev->reg_state == NETREG_UNINITIALIZED) {
		kfree((char *)dev - dev->padded);
		return;
	}

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

5376 5377
	/* will free via device release */
	put_device(&dev->dev);
L
Linus Torvalds 已提交
5378
}
E
Eric Dumazet 已提交
5379
EXPORT_SYMBOL(free_netdev);
5380

5381 5382 5383 5384 5385 5386
/**
 *	synchronize_net -  Synchronize with packet receive processing
 *
 *	Wait for packets currently being received to be done.
 *	Does not block later packets from starting.
 */
5387
void synchronize_net(void)
L
Linus Torvalds 已提交
5388 5389
{
	might_sleep();
5390
	synchronize_rcu();
L
Linus Torvalds 已提交
5391
}
E
Eric Dumazet 已提交
5392
EXPORT_SYMBOL(synchronize_net);
L
Linus Torvalds 已提交
5393 5394

/**
5395
 *	unregister_netdevice_queue - remove device from the kernel
L
Linus Torvalds 已提交
5396
 *	@dev: device
5397
 *	@head: list
5398
 *
L
Linus Torvalds 已提交
5399
 *	This function shuts down a device interface and removes it
5400
 *	from the kernel tables.
5401
 *	If head not NULL, device is queued to be unregistered later.
L
Linus Torvalds 已提交
5402 5403 5404 5405 5406
 *
 *	Callers must hold the rtnl semaphore.  You may want
 *	unregister_netdev() instead of this.
 */

5407
void unregister_netdevice_queue(struct net_device *dev, struct list_head *head)
L
Linus Torvalds 已提交
5408
{
5409 5410
	ASSERT_RTNL();

5411
	if (head) {
5412
		list_move_tail(&dev->unreg_list, head);
5413 5414 5415 5416 5417
	} else {
		rollback_registered(dev);
		/* Finish processing unregister after unlock */
		net_set_todo(dev);
	}
L
Linus Torvalds 已提交
5418
}
5419
EXPORT_SYMBOL(unregister_netdevice_queue);
L
Linus Torvalds 已提交
5420

5421 5422 5423 5424 5425 5426 5427 5428 5429 5430 5431 5432 5433 5434
/**
 *	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);
	}
}
5435
EXPORT_SYMBOL(unregister_netdevice_many);
5436

L
Linus Torvalds 已提交
5437 5438 5439 5440 5441
/**
 *	unregister_netdev - remove device from the kernel
 *	@dev: device
 *
 *	This function shuts down a device interface and removes it
5442
 *	from the kernel tables.
L
Linus Torvalds 已提交
5443 5444 5445 5446 5447 5448 5449 5450 5451 5452 5453 5454 5455
 *
 *	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);

5456 5457 5458 5459 5460 5461 5462 5463 5464 5465 5466 5467 5468 5469 5470 5471 5472 5473 5474 5475 5476 5477 5478 5479 5480
/**
 *	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;

5481 5482 5483 5484 5485 5486 5487 5488 5489
#ifdef CONFIG_SYSFS
	/* Don't allow real devices to be moved when sysfs
	 * is enabled.
	 */
	err = -EINVAL;
	if (dev->dev.parent)
		goto out;
#endif

5490 5491 5492 5493 5494 5495 5496
	/* 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;
5497
	if (net_eq(dev_net(dev), net))
5498 5499 5500 5501 5502 5503
		goto out;

	/* Pick the destination device name, and ensure
	 * we can use it in the destination network namespace.
	 */
	err = -EEXIST;
5504
	if (__dev_get_by_name(net, dev->name)) {
5505 5506 5507
		/* We get here if we can't use the current device name */
		if (!pat)
			goto out;
5508
		if (dev_get_valid_name(net, pat, dev->name, 1))
5509 5510 5511 5512 5513 5514 5515 5516
			goto out;
	}

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

	/* If device is running close it first. */
5517
	dev_close(dev);
5518 5519 5520 5521 5522 5523 5524 5525 5526 5527 5528 5529 5530 5531

	/* 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);
5532
	call_netdevice_notifiers(NETDEV_UNREGISTER_BATCH, dev);
5533 5534 5535 5536

	/*
	 *	Flush the unicast and multicast chains
	 */
5537
	dev_uc_flush(dev);
5538
	dev_mc_flush(dev);
5539

5540 5541
	netdev_unregister_kobject(dev);

5542
	/* Actually switch the network namespace */
5543
	dev_net_set(dev, net);
5544 5545 5546 5547 5548 5549 5550 5551 5552

	/* 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;
	}

5553
	/* Fixup kobjects */
5554
	err = netdev_register_kobject(dev);
5555
	WARN_ON(err);
5556 5557 5558 5559 5560 5561 5562

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

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

5563 5564 5565 5566 5567 5568
	/*
	 *	Prevent userspace races by waiting until the network
	 *	device is fully setup before sending notifications.
	 */
	rtmsg_ifinfo(RTM_NEWLINK, dev, ~0U);

5569 5570 5571 5572 5573
	synchronize_net();
	err = 0;
out:
	return err;
}
5574
EXPORT_SYMBOL_GPL(dev_change_net_namespace);
5575

L
Linus Torvalds 已提交
5576 5577 5578 5579 5580
static int dev_cpu_callback(struct notifier_block *nfb,
			    unsigned long action,
			    void *ocpu)
{
	struct sk_buff **list_skb;
5581
	struct Qdisc **list_net;
L
Linus Torvalds 已提交
5582 5583 5584 5585
	struct sk_buff *skb;
	unsigned int cpu, oldcpu = (unsigned long)ocpu;
	struct softnet_data *sd, *oldsd;

5586
	if (action != CPU_DEAD && action != CPU_DEAD_FROZEN)
L
Linus Torvalds 已提交
5587 5588 5589 5590 5591 5592 5593 5594 5595 5596 5597 5598 5599 5600 5601 5602 5603 5604 5605 5606 5607 5608 5609 5610 5611 5612 5613
		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;

	/* Find end of our output_queue. */
	list_net = &sd->output_queue;
	while (*list_net)
		list_net = &(*list_net)->next_sched;
	/* Append output queue from offline CPU. */
	*list_net = oldsd->output_queue;
	oldsd->output_queue = NULL;

	raise_softirq_irqoff(NET_TX_SOFTIRQ);
	local_irq_enable();

	/* Process offline CPU's input_pkt_queue */
T
Tom Herbert 已提交
5614
	while ((skb = __skb_dequeue(&oldsd->input_pkt_queue))) {
L
Linus Torvalds 已提交
5615
		netif_rx(skb);
T
Tom Herbert 已提交
5616 5617
		incr_input_queue_head(oldsd);
	}
L
Linus Torvalds 已提交
5618 5619 5620 5621 5622

	return NOTIFY_OK;
}


5623
/**
5624 5625 5626 5627
 *	netdev_increment_features - increment feature set by one
 *	@all: current feature set
 *	@one: new feature set
 *	@mask: mask feature set
5628 5629
 *
 *	Computes a new feature set after adding a device with feature set
5630 5631
 *	@one to the master device with current feature set @all.  Will not
 *	enable anything that is off in @mask. Returns the new feature set.
5632
 */
5633 5634 5635 5636
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
5637
	if (all & NETIF_F_NO_CSUM && !(one & NETIF_F_NO_CSUM))
5638 5639 5640 5641 5642 5643 5644 5645
		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);
		}
5646

5647 5648 5649 5650 5651 5652
		/* 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;
		}
	}
5653

5654
	one |= NETIF_F_ALL_CSUM;
5655

5656
	one |= all & NETIF_F_ONE_FOR_ALL;
5657
	all &= one | NETIF_F_LLTX | NETIF_F_GSO | NETIF_F_UFO;
5658
	all |= one & mask & NETIF_F_ONE_FOR_ALL;
5659 5660 5661

	return all;
}
5662
EXPORT_SYMBOL(netdev_increment_features);
5663

5664 5665 5666 5667 5668 5669 5670 5671 5672 5673 5674 5675 5676
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;
}

5677
/* Initialize per network namespace state */
5678
static int __net_init netdev_init(struct net *net)
5679 5680 5681
{
	INIT_LIST_HEAD(&net->dev_base_head);

5682 5683 5684
	net->dev_name_head = netdev_create_hash();
	if (net->dev_name_head == NULL)
		goto err_name;
5685

5686 5687 5688
	net->dev_index_head = netdev_create_hash();
	if (net->dev_index_head == NULL)
		goto err_idx;
5689 5690

	return 0;
5691 5692 5693 5694 5695

err_idx:
	kfree(net->dev_name_head);
err_name:
	return -ENOMEM;
5696 5697
}

5698 5699 5700 5701 5702 5703 5704 5705
/**
 *	netdev_drivername - network driver for the device
 *	@dev: network device
 *	@buffer: buffer for resulting name
 *	@len: size of buffer
 *
 *	Determine network driver for device.
 */
5706
char *netdev_drivername(const struct net_device *dev, char *buffer, int len)
5707
{
5708 5709
	const struct device_driver *driver;
	const struct device *parent;
5710 5711 5712 5713 5714 5715 5716 5717 5718 5719 5720 5721 5722 5723 5724 5725

	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;
}

5726
static void __net_exit netdev_exit(struct net *net)
5727 5728 5729 5730 5731
{
	kfree(net->dev_name_head);
	kfree(net->dev_index_head);
}

5732
static struct pernet_operations __net_initdata netdev_net_ops = {
5733 5734 5735 5736
	.init = netdev_init,
	.exit = netdev_exit,
};

5737
static void __net_exit default_device_exit(struct net *net)
5738
{
5739
	struct net_device *dev, *aux;
5740
	/*
5741
	 * Push all migratable network devices back to the
5742 5743 5744
	 * initial network namespace
	 */
	rtnl_lock();
5745
	for_each_netdev_safe(net, dev, aux) {
5746
		int err;
5747
		char fb_name[IFNAMSIZ];
5748 5749 5750 5751 5752

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

5753 5754 5755
		/* Leave virtual devices for the generic cleanup */
		if (dev->rtnl_link_ops)
			continue;
5756

5757
		/* Push remaing network devices to init_net */
5758 5759
		snprintf(fb_name, IFNAMSIZ, "dev%d", dev->ifindex);
		err = dev_change_net_namespace(dev, &init_net, fb_name);
5760
		if (err) {
5761
			printk(KERN_EMERG "%s: failed to move %s to init_net: %d\n",
5762
				__func__, dev->name, err);
5763
			BUG();
5764 5765 5766 5767 5768
		}
	}
	rtnl_unlock();
}

5769 5770 5771 5772 5773 5774 5775 5776 5777 5778 5779 5780 5781 5782 5783 5784 5785 5786 5787 5788 5789 5790 5791 5792
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();
}

5793
static struct pernet_operations __net_initdata default_device_ops = {
5794
	.exit = default_device_exit,
5795
	.exit_batch = default_device_exit_batch,
5796 5797
};

L
Linus Torvalds 已提交
5798 5799 5800 5801 5802 5803 5804 5805 5806 5807 5808 5809 5810 5811 5812 5813 5814 5815 5816 5817
/*
 *	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;

5818
	if (netdev_kobject_init())
L
Linus Torvalds 已提交
5819 5820 5821
		goto out;

	INIT_LIST_HEAD(&ptype_all);
5822
	for (i = 0; i < PTYPE_HASH_SIZE; i++)
L
Linus Torvalds 已提交
5823 5824
		INIT_LIST_HEAD(&ptype_base[i]);

5825 5826
	if (register_pernet_subsys(&netdev_net_ops))
		goto out;
L
Linus Torvalds 已提交
5827 5828 5829 5830 5831

	/*
	 *	Initialise the packet receive queues.
	 */

5832
	for_each_possible_cpu(i) {
L
Linus Torvalds 已提交
5833 5834 5835 5836 5837 5838
		struct softnet_data *queue;

		queue = &per_cpu(softnet_data, i);
		skb_queue_head_init(&queue->input_pkt_queue);
		queue->completion_queue = NULL;
		INIT_LIST_HEAD(&queue->poll_list);
5839

E
Eric Dumazet 已提交
5840
#ifdef CONFIG_RPS
T
Tom Herbert 已提交
5841 5842 5843
		queue->csd.func = trigger_softirq;
		queue->csd.info = queue;
		queue->csd.flags = 0;
5844
#endif
T
Tom Herbert 已提交
5845

5846 5847
		queue->backlog.poll = process_backlog;
		queue->backlog.weight = weight_p;
5848
		queue->backlog.gro_list = NULL;
5849
		queue->backlog.gro_count = 0;
L
Linus Torvalds 已提交
5850 5851 5852 5853
	}

	dev_boot_phase = 0;

5854 5855 5856 5857 5858 5859 5860 5861 5862 5863 5864 5865 5866 5867 5868
	/* 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;

5869 5870
	open_softirq(NET_TX_SOFTIRQ, net_tx_action);
	open_softirq(NET_RX_SOFTIRQ, net_rx_action);
L
Linus Torvalds 已提交
5871 5872 5873 5874 5875 5876 5877 5878 5879 5880 5881

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

subsys_initcall(net_dev_init);

5882 5883
static int __init initialize_hashrnd(void)
{
T
Tom Herbert 已提交
5884
	get_random_bytes(&hashrnd, sizeof(hashrnd));
5885 5886 5887 5888 5889
	return 0;
}

late_initcall_sync(initialize_hashrnd);