dev.c 137.4 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 82
#include <linux/cpu.h>
#include <linux/types.h>
#include <linux/kernel.h>
#include <linux/sched.h>
A
Arjan van de Ven 已提交
83
#include <linux/mutex.h>
L
Linus Torvalds 已提交
84 85 86 87 88 89 90 91 92
#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>
93
#include <linux/ethtool.h>
L
Linus Torvalds 已提交
94 95
#include <linux/notifier.h>
#include <linux/skbuff.h>
96
#include <net/net_namespace.h>
L
Linus Torvalds 已提交
97 98 99 100 101 102
#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
103
#include <linux/if_macvlan.h>
L
Linus Torvalds 已提交
104 105 106 107 108 109 110 111 112 113
#include <net/dst.h>
#include <net/pkt_sched.h>
#include <net/checksum.h>
#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>
114
#include <net/wext.h>
L
Linus Torvalds 已提交
115 116
#include <net/iw_handler.h>
#include <asm/current.h>
S
Steve Grubb 已提交
117
#include <linux/audit.h>
118
#include <linux/dmaengine.h>
119
#include <linux/err.h>
120
#include <linux/ctype.h>
121
#include <linux/if_arp.h>
122
#include <linux/if_vlan.h>
123
#include <linux/ip.h>
124
#include <net/ip.h>
125 126
#include <linux/ipv6.h>
#include <linux/in.h>
D
David S. Miller 已提交
127 128
#include <linux/jhash.h>
#include <linux/random.h>
129
#include <trace/events/napi.h>
L
Linus Torvalds 已提交
130

131 132
#include "net-sysfs.h"

133 134 135
/* Instead of increasing this, you should create a hash table. */
#define MAX_GRO_SKBS 8

H
Herbert Xu 已提交
136 137 138
/* 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
139 140 141 142 143 144 145 146 147 148 149
/*
 *	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
150
 *             the average user (w/out VLANs) will not be adversely affected.
L
Linus Torvalds 已提交
151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166
 *             --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
 */

167 168 169
#define PTYPE_HASH_SIZE	(16)
#define PTYPE_HASH_MASK	(PTYPE_HASH_SIZE - 1)

L
Linus Torvalds 已提交
170
static DEFINE_SPINLOCK(ptype_lock);
171
static struct list_head ptype_base[PTYPE_HASH_SIZE] __read_mostly;
172
static struct list_head ptype_all __read_mostly;	/* Taps */
L
Linus Torvalds 已提交
173 174

/*
175
 * The @dev_base_head list is protected by @dev_base_lock and the rtnl
L
Linus Torvalds 已提交
176 177 178 179 180
 * semaphore.
 *
 * Pure readers hold dev_base_lock for reading.
 *
 * Writers must hold the rtnl semaphore while they loop through the
181
 * dev_base_head list, and hold dev_base_lock for writing when they do the
L
Linus Torvalds 已提交
182 183 184 185 186 187 188 189 190 191 192 193 194 195
 * 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);

196
static inline struct hlist_head *dev_name_hash(struct net *net, const char *name)
L
Linus Torvalds 已提交
197 198
{
	unsigned hash = full_name_hash(name, strnlen(name, IFNAMSIZ));
199
	return &net->dev_name_head[hash & (NETDEV_HASHENTRIES - 1)];
L
Linus Torvalds 已提交
200 201
}

202
static inline struct hlist_head *dev_index_hash(struct net *net, int ifindex)
L
Linus Torvalds 已提交
203
{
204
	return &net->dev_index_head[ifindex & (NETDEV_HASHENTRIES - 1)];
L
Linus Torvalds 已提交
205 206
}

207 208 209
/* Device list insertion */
static int list_netdevice(struct net_device *dev)
{
210
	struct net *net = dev_net(dev);
211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234

	ASSERT_RTNL();

	write_lock_bh(&dev_base_lock);
	list_add_tail(&dev->dev_list, &net->dev_base_head);
	hlist_add_head(&dev->name_hlist, dev_name_hash(net, dev->name));
	hlist_add_head(&dev->index_hlist, dev_index_hash(net, dev->ifindex));
	write_unlock_bh(&dev_base_lock);
	return 0;
}

/* Device list removal */
static void unlist_netdevice(struct net_device *dev)
{
	ASSERT_RTNL();

	/* Unlink dev from the device chain */
	write_lock_bh(&dev_base_lock);
	list_del(&dev->dev_list);
	hlist_del(&dev->name_hlist);
	hlist_del(&dev->index_hlist);
	write_unlock_bh(&dev_base_lock);
}

L
Linus Torvalds 已提交
235 236 237 238
/*
 *	Our notifier list
 */

239
static RAW_NOTIFIER_HEAD(netdev_chain);
L
Linus Torvalds 已提交
240 241 242 243 244

/*
 *	Device drivers call our routines to queue packets here. We empty the
 *	queue in the local softnet handler.
 */
245 246

DEFINE_PER_CPU(struct softnet_data, softnet_data);
E
Eric Dumazet 已提交
247
EXPORT_PER_CPU_SYMBOL(softnet_data);
L
Linus Torvalds 已提交
248

249
#ifdef CONFIG_LOCKDEP
250
/*
251
 * register_netdevice() inits txq->_xmit_lock and sets lockdep class
252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267
 * 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,
268
	 ARPHRD_IEEE80211_PRISM, ARPHRD_IEEE80211_RADIOTAP, ARPHRD_PHONET,
269
	 ARPHRD_PHONET_PIPE, ARPHRD_IEEE802154,
270
	 ARPHRD_VOID, ARPHRD_NONE};
271

272
static const char *const netdev_lock_name[] =
273 274 275 276 277 278 279 280 281 282 283 284 285
	{"_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",
286
	 "_xmit_IEEE80211_PRISM", "_xmit_IEEE80211_RADIOTAP", "_xmit_PHONET",
287
	 "_xmit_PHONET_PIPE", "_xmit_IEEE802154",
288
	 "_xmit_VOID", "_xmit_NONE"};
289 290

static struct lock_class_key netdev_xmit_lock_key[ARRAY_SIZE(netdev_lock_type)];
291
static struct lock_class_key netdev_addr_lock_key[ARRAY_SIZE(netdev_lock_type)];
292 293 294 295 296 297 298 299 300 301 302 303

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;
}

304 305
static inline void netdev_set_xmit_lockdep_class(spinlock_t *lock,
						 unsigned short dev_type)
306 307 308 309 310 311 312
{
	int i;

	i = netdev_lock_pos(dev_type);
	lockdep_set_class_and_name(lock, &netdev_xmit_lock_key[i],
				   netdev_lock_name[i]);
}
313 314 315 316 317 318 319 320 321 322

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]);
}
323
#else
324 325 326 327 328
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)
329 330 331
{
}
#endif
L
Linus Torvalds 已提交
332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362

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

		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.
 *
363
 *	This call does not sleep therefore it can not
L
Linus Torvalds 已提交
364 365 366 367 368 369 370 371 372
 *	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
373
	if (pt->type == htons(ETH_P_ALL))
L
Linus Torvalds 已提交
374
		list_add_rcu(&pt->list, &ptype_all);
S
Stephen Hemminger 已提交
375
	else {
376
		hash = ntohs(pt->type) & PTYPE_HASH_MASK;
L
Linus Torvalds 已提交
377 378 379 380
		list_add_rcu(&pt->list, &ptype_base[hash]);
	}
	spin_unlock_bh(&ptype_lock);
}
E
Eric Dumazet 已提交
381
EXPORT_SYMBOL(dev_add_pack);
L
Linus Torvalds 已提交
382 383 384 385 386 387 388 389

/**
 *	__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
390
 *	returns.
L
Linus Torvalds 已提交
391 392 393 394 395 396 397 398 399 400 401 402
 *
 *      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
403
	if (pt->type == htons(ETH_P_ALL))
L
Linus Torvalds 已提交
404
		head = &ptype_all;
S
Stephen Hemminger 已提交
405
	else
406
		head = &ptype_base[ntohs(pt->type) & PTYPE_HASH_MASK];
L
Linus Torvalds 已提交
407 408 409 410 411 412 413 414 415 416 417 418

	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
419 420
EXPORT_SYMBOL(__dev_remove_pack);

L
Linus Torvalds 已提交
421 422 423 424 425 426 427 428 429 430 431 432 433 434 435
/**
 *	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);
436

L
Linus Torvalds 已提交
437 438
	synchronize_net();
}
E
Eric Dumazet 已提交
439
EXPORT_SYMBOL(dev_remove_pack);
L
Linus Torvalds 已提交
440 441 442 443 444 445 446 447 448 449 450 451 452 453 454 455 456 457 458 459 460 461 462 463 464 465 466 467

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

		      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));
468
			strlcpy(s[i].name, name, IFNAMSIZ);
L
Linus Torvalds 已提交
469 470 471 472 473 474 475 476 477 478 479 480 481 482 483 484 485 486 487 488 489 490 491 492
			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] != ' ' &&
493
		    !strcmp(dev->name, s[i].name)) {
L
Linus Torvalds 已提交
494 495 496 497 498 499 500 501 502
			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
503
EXPORT_SYMBOL(netdev_boot_setup_check);
L
Linus Torvalds 已提交
504 505 506 507 508 509 510 511 512 513 514 515 516 517 518 519 520 521 522 523 524 525 526 527


/**
 *	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
	 */
528
	if (__dev_get_by_name(&init_net, name))
L
Linus Torvalds 已提交
529 530 531 532 533 534 535 536 537 538 539 540 541 542 543 544 545 546 547 548 549 550 551 552 553 554 555 556 557 558 559 560 561 562 563 564 565 566 567 568 569 570 571 572 573
		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
574
 *	@net: the applicable net namespace
L
Linus Torvalds 已提交
575 576 577 578 579 580 581 582 583
 *	@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.
 */

584
struct net_device *__dev_get_by_name(struct net *net, const char *name)
L
Linus Torvalds 已提交
585 586 587
{
	struct hlist_node *p;

588
	hlist_for_each(p, dev_name_hash(net, name)) {
L
Linus Torvalds 已提交
589 590 591 592 593 594 595
		struct net_device *dev
			= hlist_entry(p, struct net_device, name_hlist);
		if (!strncmp(dev->name, name, IFNAMSIZ))
			return dev;
	}
	return NULL;
}
E
Eric Dumazet 已提交
596
EXPORT_SYMBOL(__dev_get_by_name);
L
Linus Torvalds 已提交
597 598 599

/**
 *	dev_get_by_name		- find a device by its name
600
 *	@net: the applicable net namespace
L
Linus Torvalds 已提交
601 602 603 604 605 606 607 608 609
 *	@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.
 */

610
struct net_device *dev_get_by_name(struct net *net, const char *name)
L
Linus Torvalds 已提交
611 612 613 614
{
	struct net_device *dev;

	read_lock(&dev_base_lock);
615
	dev = __dev_get_by_name(net, name);
L
Linus Torvalds 已提交
616 617 618 619 620
	if (dev)
		dev_hold(dev);
	read_unlock(&dev_base_lock);
	return dev;
}
E
Eric Dumazet 已提交
621
EXPORT_SYMBOL(dev_get_by_name);
L
Linus Torvalds 已提交
622 623 624

/**
 *	__dev_get_by_index - find a device by its ifindex
625
 *	@net: the applicable net namespace
L
Linus Torvalds 已提交
626 627 628 629 630 631 632 633 634
 *	@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.
 */

635
struct net_device *__dev_get_by_index(struct net *net, int ifindex)
L
Linus Torvalds 已提交
636 637 638
{
	struct hlist_node *p;

639
	hlist_for_each(p, dev_index_hash(net, ifindex)) {
L
Linus Torvalds 已提交
640 641 642 643 644 645 646
		struct net_device *dev
			= hlist_entry(p, struct net_device, index_hlist);
		if (dev->ifindex == ifindex)
			return dev;
	}
	return NULL;
}
E
Eric Dumazet 已提交
647
EXPORT_SYMBOL(__dev_get_by_index);
L
Linus Torvalds 已提交
648 649 650 651


/**
 *	dev_get_by_index - find a device by its ifindex
652
 *	@net: the applicable net namespace
L
Linus Torvalds 已提交
653 654 655 656 657 658 659 660
 *	@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.
 */

661
struct net_device *dev_get_by_index(struct net *net, int ifindex)
L
Linus Torvalds 已提交
662 663 664 665
{
	struct net_device *dev;

	read_lock(&dev_base_lock);
666
	dev = __dev_get_by_index(net, ifindex);
L
Linus Torvalds 已提交
667 668 669 670 671
	if (dev)
		dev_hold(dev);
	read_unlock(&dev_base_lock);
	return dev;
}
E
Eric Dumazet 已提交
672
EXPORT_SYMBOL(dev_get_by_index);
L
Linus Torvalds 已提交
673 674 675

/**
 *	dev_getbyhwaddr - find a device by its hardware address
676
 *	@net: the applicable net namespace
L
Linus Torvalds 已提交
677 678 679 680 681 682 683 684 685 686 687 688
 *	@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
 */

689
struct net_device *dev_getbyhwaddr(struct net *net, unsigned short type, char *ha)
L
Linus Torvalds 已提交
690 691 692 693 694
{
	struct net_device *dev;

	ASSERT_RTNL();

695
	for_each_netdev(net, dev)
L
Linus Torvalds 已提交
696 697
		if (dev->type == type &&
		    !memcmp(dev->dev_addr, ha, dev->addr_len))
698 699 700
			return dev;

	return NULL;
L
Linus Torvalds 已提交
701
}
702 703
EXPORT_SYMBOL(dev_getbyhwaddr);

704
struct net_device *__dev_getfirstbyhwtype(struct net *net, unsigned short type)
L
Linus Torvalds 已提交
705 706 707
{
	struct net_device *dev;

708
	ASSERT_RTNL();
709
	for_each_netdev(net, dev)
710
		if (dev->type == type)
711 712 713
			return dev;

	return NULL;
714 715 716
}
EXPORT_SYMBOL(__dev_getfirstbyhwtype);

717
struct net_device *dev_getfirstbyhwtype(struct net *net, unsigned short type)
718 719 720 721
{
	struct net_device *dev;

	rtnl_lock();
722
	dev = __dev_getfirstbyhwtype(net, type);
723 724
	if (dev)
		dev_hold(dev);
L
Linus Torvalds 已提交
725 726 727 728 729 730 731
	rtnl_unlock();
	return dev;
}
EXPORT_SYMBOL(dev_getfirstbyhwtype);

/**
 *	dev_get_by_flags - find any device with given flags
732
 *	@net: the applicable net namespace
L
Linus Torvalds 已提交
733 734 735 736
 *	@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
737
 *	is not found or a pointer to the device. The device returned has
L
Linus Torvalds 已提交
738 739 740 741
 *	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
742 743
struct net_device *dev_get_by_flags(struct net *net, unsigned short if_flags,
				    unsigned short mask)
L
Linus Torvalds 已提交
744
{
745
	struct net_device *dev, *ret;
L
Linus Torvalds 已提交
746

747
	ret = NULL;
L
Linus Torvalds 已提交
748
	read_lock(&dev_base_lock);
749
	for_each_netdev(net, dev) {
L
Linus Torvalds 已提交
750 751
		if (((dev->flags ^ if_flags) & mask) == 0) {
			dev_hold(dev);
752
			ret = dev;
L
Linus Torvalds 已提交
753 754 755 756
			break;
		}
	}
	read_unlock(&dev_base_lock);
757
	return ret;
L
Linus Torvalds 已提交
758
}
E
Eric Dumazet 已提交
759
EXPORT_SYMBOL(dev_get_by_flags);
L
Linus Torvalds 已提交
760 761 762 763 764 765

/**
 *	dev_valid_name - check if name is okay for network device
 *	@name: name string
 *
 *	Network device names need to be valid file names to
766 767
 *	to allow sysfs to work.  We also disallow any kind of
 *	whitespace.
L
Linus Torvalds 已提交
768
 */
769
int dev_valid_name(const char *name)
L
Linus Torvalds 已提交
770
{
771 772
	if (*name == '\0')
		return 0;
773 774
	if (strlen(name) >= IFNAMSIZ)
		return 0;
775 776 777 778 779 780 781 782 783
	if (!strcmp(name, ".") || !strcmp(name, ".."))
		return 0;

	while (*name) {
		if (*name == '/' || isspace(*name))
			return 0;
		name++;
	}
	return 1;
L
Linus Torvalds 已提交
784
}
E
Eric Dumazet 已提交
785
EXPORT_SYMBOL(dev_valid_name);
L
Linus Torvalds 已提交
786 787

/**
788 789
 *	__dev_alloc_name - allocate a name for a device
 *	@net: network namespace to allocate the device name in
L
Linus Torvalds 已提交
790
 *	@name: name format string
791
 *	@buf:  scratch buffer and result name string
L
Linus Torvalds 已提交
792 793
 *
 *	Passed a format string - eg "lt%d" it will try and find a suitable
S
Stephen Hemminger 已提交
794 795 796 797 798 799
 *	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
800 801
 */

802
static int __dev_alloc_name(struct net *net, const char *name, char *buf)
L
Linus Torvalds 已提交
803 804 805 806
{
	int i = 0;
	const char *p;
	const int max_netdevices = 8*PAGE_SIZE;
S
Stephen Hemminger 已提交
807
	unsigned long *inuse;
L
Linus Torvalds 已提交
808 809 810 811 812 813 814 815 816 817 818 819 820
	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
821
		inuse = (unsigned long *) get_zeroed_page(GFP_ATOMIC);
L
Linus Torvalds 已提交
822 823 824
		if (!inuse)
			return -ENOMEM;

825
		for_each_netdev(net, d) {
L
Linus Torvalds 已提交
826 827 828 829 830 831
			if (!sscanf(d->name, name, &i))
				continue;
			if (i < 0 || i >= max_netdevices)
				continue;

			/*  avoid cases where sscanf is not exact inverse of printf */
832
			snprintf(buf, IFNAMSIZ, name, i);
L
Linus Torvalds 已提交
833 834 835 836 837 838 839 840
			if (!strncmp(buf, d->name, IFNAMSIZ))
				set_bit(i, inuse);
		}

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

841 842
	snprintf(buf, IFNAMSIZ, name, i);
	if (!__dev_get_by_name(net, buf))
L
Linus Torvalds 已提交
843 844 845 846 847 848 849 850 851
		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;
}

852 853 854 855 856 857 858 859 860 861 862 863 864 865 866 867 868 869 870 871
/**
 *	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;

872 873
	BUG_ON(!dev_net(dev));
	net = dev_net(dev);
874 875 876 877 878
	ret = __dev_alloc_name(net, name, buf);
	if (ret >= 0)
		strlcpy(dev->name, buf, IFNAMSIZ);
	return ret;
}
E
Eric Dumazet 已提交
879
EXPORT_SYMBOL(dev_alloc_name);
880

L
Linus Torvalds 已提交
881 882 883 884 885 886 887 888 889

/**
 *	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.
 */
890
int dev_change_name(struct net_device *dev, const char *newname)
L
Linus Torvalds 已提交
891
{
892
	char oldname[IFNAMSIZ];
L
Linus Torvalds 已提交
893
	int err = 0;
894
	int ret;
895
	struct net *net;
L
Linus Torvalds 已提交
896 897

	ASSERT_RTNL();
898
	BUG_ON(!dev_net(dev));
L
Linus Torvalds 已提交
899

900
	net = dev_net(dev);
L
Linus Torvalds 已提交
901 902 903 904 905 906
	if (dev->flags & IFF_UP)
		return -EBUSY;

	if (!dev_valid_name(newname))
		return -EINVAL;

907 908 909
	if (strncmp(newname, dev->name, IFNAMSIZ) == 0)
		return 0;

910 911
	memcpy(oldname, dev->name, IFNAMSIZ);

L
Linus Torvalds 已提交
912 913 914 915
	if (strchr(newname, '%')) {
		err = dev_alloc_name(dev, newname);
		if (err < 0)
			return err;
E
Eric Dumazet 已提交
916
	} else if (__dev_get_by_name(net, newname))
L
Linus Torvalds 已提交
917 918 919 920
		return -EEXIST;
	else
		strlcpy(dev->name, newname, IFNAMSIZ);

921
rollback:
922 923 924 925 926 927 928 929 930
	/* For now only devices in the initial network namespace
	 * are in sysfs.
	 */
	if (net == &init_net) {
		ret = device_rename(&dev->dev, dev->name);
		if (ret) {
			memcpy(dev->name, oldname, IFNAMSIZ);
			return ret;
		}
931
	}
932 933

	write_lock_bh(&dev_base_lock);
934
	hlist_del(&dev->name_hlist);
935
	hlist_add_head(&dev->name_hlist, dev_name_hash(net, dev->name));
936 937
	write_unlock_bh(&dev_base_lock);

938
	ret = call_netdevice_notifiers(NETDEV_CHANGENAME, dev);
939 940 941 942 943 944 945 946 947 948 949 950 951
	ret = notifier_to_errno(ret);

	if (ret) {
		if (err) {
			printk(KERN_ERR
			       "%s: name change rollback failed: %d.\n",
			       dev->name, ret);
		} else {
			err = ret;
			memcpy(dev->name, oldname, IFNAMSIZ);
			goto rollback;
		}
	}
L
Linus Torvalds 已提交
952 953 954 955

	return err;
}

956 957 958 959
/**
 *	dev_set_alias - change ifalias of a device
 *	@dev: device
 *	@alias: name up to IFALIASZ
960
 *	@len: limit of bytes to copy from info
961 962 963 964 965 966 967 968 969 970
 *
 *	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;

971 972 973 974 975 976 977 978
	if (!len) {
		if (dev->ifalias) {
			kfree(dev->ifalias);
			dev->ifalias = NULL;
		}
		return 0;
	}

E
Eric Dumazet 已提交
979
	dev->ifalias = krealloc(dev->ifalias, len + 1, GFP_KERNEL);
980 981 982 983 984 985 986 987
	if (!dev->ifalias)
		return -ENOMEM;

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


988
/**
S
Stephen Hemminger 已提交
989
 *	netdev_features_change - device changes features
990 991 992 993 994 995
 *	@dev: device to cause notification
 *
 *	Called to indicate a device has changed features.
 */
void netdev_features_change(struct net_device *dev)
{
996
	call_netdevice_notifiers(NETDEV_FEAT_CHANGE, dev);
997 998 999
}
EXPORT_SYMBOL(netdev_features_change);

L
Linus Torvalds 已提交
1000 1001 1002 1003 1004 1005 1006 1007 1008 1009 1010
/**
 *	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) {
1011
		call_netdevice_notifiers(NETDEV_CHANGE, dev);
L
Linus Torvalds 已提交
1012 1013 1014
		rtmsg_ifinfo(RTM_NEWLINK, dev, 0);
	}
}
E
Eric Dumazet 已提交
1015
EXPORT_SYMBOL(netdev_state_change);
L
Linus Torvalds 已提交
1016

1017
void netdev_bonding_change(struct net_device *dev, unsigned long event)
1018
{
1019
	call_netdevice_notifiers(event, dev);
1020 1021 1022
}
EXPORT_SYMBOL(netdev_bonding_change);

L
Linus Torvalds 已提交
1023 1024
/**
 *	dev_load 	- load a network module
1025
 *	@net: the applicable net namespace
L
Linus Torvalds 已提交
1026 1027 1028 1029 1030 1031 1032
 *	@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.
 */

1033
void dev_load(struct net *net, const char *name)
L
Linus Torvalds 已提交
1034
{
1035
	struct net_device *dev;
L
Linus Torvalds 已提交
1036 1037

	read_lock(&dev_base_lock);
1038
	dev = __dev_get_by_name(net, name);
L
Linus Torvalds 已提交
1039 1040
	read_unlock(&dev_base_lock);

1041
	if (!dev && capable(CAP_NET_ADMIN))
L
Linus Torvalds 已提交
1042 1043
		request_module("%s", name);
}
E
Eric Dumazet 已提交
1044
EXPORT_SYMBOL(dev_load);
L
Linus Torvalds 已提交
1045 1046 1047 1048 1049 1050 1051 1052 1053 1054 1055 1056 1057 1058 1059

/**
 *	dev_open	- prepare an interface for use.
 *	@dev:	device to open
 *
 *	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.
 */
int dev_open(struct net_device *dev)
{
1060
	const struct net_device_ops *ops = dev->netdev_ops;
1061
	int ret;
L
Linus Torvalds 已提交
1062

1063 1064
	ASSERT_RTNL();

L
Linus Torvalds 已提交
1065 1066 1067 1068 1069 1070 1071 1072 1073 1074 1075 1076 1077
	/*
	 *	Is it already up?
	 */

	if (dev->flags & IFF_UP)
		return 0;

	/*
	 *	Is it even present?
	 */
	if (!netif_device_present(dev))
		return -ENODEV;

1078 1079 1080 1081 1082
	ret = call_netdevice_notifiers(NETDEV_PRE_UP, dev);
	ret = notifier_to_errno(ret);
	if (ret)
		return ret;

L
Linus Torvalds 已提交
1083 1084 1085 1086
	/*
	 *	Call device private open method
	 */
	set_bit(__LINK_STATE_START, &dev->state);
1087

1088 1089
	if (ops->ndo_validate_addr)
		ret = ops->ndo_validate_addr(dev);
1090

1091 1092
	if (!ret && ops->ndo_open)
		ret = ops->ndo_open(dev);
L
Linus Torvalds 已提交
1093

1094
	/*
L
Linus Torvalds 已提交
1095 1096 1097
	 *	If it went open OK then:
	 */

1098 1099 1100
	if (ret)
		clear_bit(__LINK_STATE_START, &dev->state);
	else {
L
Linus Torvalds 已提交
1101 1102 1103 1104 1105
		/*
		 *	Set the flags.
		 */
		dev->flags |= IFF_UP;

1106 1107 1108
		/*
		 *	Enable NET_DMA
		 */
1109
		net_dmaengine_get();
1110

L
Linus Torvalds 已提交
1111 1112 1113
		/*
		 *	Initialize multicasting status
		 */
1114
		dev_set_rx_mode(dev);
L
Linus Torvalds 已提交
1115 1116 1117 1118 1119 1120 1121 1122 1123

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

		/*
		 *	... and announce new interface.
		 */
1124
		call_netdevice_notifiers(NETDEV_UP, dev);
L
Linus Torvalds 已提交
1125
	}
1126

L
Linus Torvalds 已提交
1127 1128
	return ret;
}
E
Eric Dumazet 已提交
1129
EXPORT_SYMBOL(dev_open);
L
Linus Torvalds 已提交
1130 1131 1132 1133 1134 1135 1136 1137 1138 1139 1140 1141

/**
 *	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)
{
1142
	const struct net_device_ops *ops = dev->netdev_ops;
1143 1144
	ASSERT_RTNL();

1145 1146
	might_sleep();

L
Linus Torvalds 已提交
1147 1148 1149 1150 1151 1152 1153
	if (!(dev->flags & IFF_UP))
		return 0;

	/*
	 *	Tell people we are going down, so that they can
	 *	prepare to death, when device is still operating.
	 */
1154
	call_netdevice_notifiers(NETDEV_GOING_DOWN, dev);
L
Linus Torvalds 已提交
1155 1156 1157 1158

	clear_bit(__LINK_STATE_START, &dev->state);

	/* Synchronize to scheduled poll. We cannot touch poll list,
1159 1160 1161 1162 1163
	 * 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
1164 1165
	smp_mb__after_clear_bit(); /* Commit netif_running(). */

1166 1167
	dev_deactivate(dev);

L
Linus Torvalds 已提交
1168 1169 1170 1171 1172 1173 1174
	/*
	 *	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.
	 */
1175 1176
	if (ops->ndo_stop)
		ops->ndo_stop(dev);
L
Linus Torvalds 已提交
1177 1178 1179 1180 1181 1182 1183 1184 1185 1186

	/*
	 *	Device is now down.
	 */

	dev->flags &= ~IFF_UP;

	/*
	 * Tell people we are down
	 */
1187
	call_netdevice_notifiers(NETDEV_DOWN, dev);
L
Linus Torvalds 已提交
1188

1189 1190 1191
	/*
	 *	Shutdown NET_DMA
	 */
1192
	net_dmaengine_put();
1193

L
Linus Torvalds 已提交
1194 1195
	return 0;
}
E
Eric Dumazet 已提交
1196
EXPORT_SYMBOL(dev_close);
L
Linus Torvalds 已提交
1197 1198


1199 1200 1201 1202 1203 1204 1205 1206 1207 1208 1209 1210 1211 1212 1213 1214 1215 1216 1217 1218 1219 1220 1221
/**
 *	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);


1222 1223
static int dev_boot_phase = 1;

L
Linus Torvalds 已提交
1224 1225 1226 1227 1228 1229 1230 1231 1232 1233 1234 1235 1236 1237 1238
/*
 *	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
1239
 *	to the new notifier to allow device to have a race free
L
Linus Torvalds 已提交
1240 1241 1242 1243 1244 1245
 *	view of the network device list.
 */

int register_netdevice_notifier(struct notifier_block *nb)
{
	struct net_device *dev;
1246
	struct net_device *last;
1247
	struct net *net;
L
Linus Torvalds 已提交
1248 1249 1250
	int err;

	rtnl_lock();
1251
	err = raw_notifier_chain_register(&netdev_chain, nb);
1252 1253
	if (err)
		goto unlock;
1254 1255 1256 1257 1258 1259 1260 1261 1262 1263 1264
	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
1265

1266 1267
			nb->notifier_call(nb, NETDEV_UP, dev);
		}
L
Linus Torvalds 已提交
1268
	}
1269 1270

unlock:
L
Linus Torvalds 已提交
1271 1272
	rtnl_unlock();
	return err;
1273 1274 1275

rollback:
	last = dev;
1276 1277 1278 1279
	for_each_net(net) {
		for_each_netdev(net, dev) {
			if (dev == last)
				break;
1280

1281 1282 1283 1284 1285
			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);
1286 1287
		}
	}
1288 1289

	raw_notifier_chain_unregister(&netdev_chain, nb);
1290
	goto unlock;
L
Linus Torvalds 已提交
1291
}
E
Eric Dumazet 已提交
1292
EXPORT_SYMBOL(register_netdevice_notifier);
L
Linus Torvalds 已提交
1293 1294 1295 1296 1297 1298 1299 1300 1301 1302 1303 1304 1305

/**
 *	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)
{
1306 1307 1308
	int err;

	rtnl_lock();
1309
	err = raw_notifier_chain_unregister(&netdev_chain, nb);
1310 1311
	rtnl_unlock();
	return err;
L
Linus Torvalds 已提交
1312
}
E
Eric Dumazet 已提交
1313
EXPORT_SYMBOL(unregister_netdevice_notifier);
L
Linus Torvalds 已提交
1314 1315 1316 1317

/**
 *	call_netdevice_notifiers - call all network notifier blocks
 *      @val: value passed unmodified to notifier function
1318
 *      @dev: net_device pointer passed unmodified to notifier function
L
Linus Torvalds 已提交
1319 1320
 *
 *	Call all network notifier blocks.  Parameters and return value
1321
 *	are as for raw_notifier_call_chain().
L
Linus Torvalds 已提交
1322 1323
 */

1324
int call_netdevice_notifiers(unsigned long val, struct net_device *dev)
L
Linus Torvalds 已提交
1325
{
1326
	return raw_notifier_call_chain(&netdev_chain, val, dev);
L
Linus Torvalds 已提交
1327 1328 1329 1330 1331 1332 1333 1334 1335
}

/* 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
1336
EXPORT_SYMBOL(net_enable_timestamp);
L
Linus Torvalds 已提交
1337 1338 1339 1340 1341

void net_disable_timestamp(void)
{
	atomic_dec(&netstamp_needed);
}
E
Eric Dumazet 已提交
1342
EXPORT_SYMBOL(net_disable_timestamp);
L
Linus Torvalds 已提交
1343

1344
static inline void net_timestamp(struct sk_buff *skb)
L
Linus Torvalds 已提交
1345 1346
{
	if (atomic_read(&netstamp_needed))
1347
		__net_timestamp(skb);
1348 1349
	else
		skb->tstamp.tv64 = 0;
L
Linus Torvalds 已提交
1350 1351 1352 1353 1354 1355 1356
}

/*
 *	Support routine. Sends outgoing frames to any network
 *	taps currently in use.
 */

1357
static void dev_queue_xmit_nit(struct sk_buff *skb, struct net_device *dev)
L
Linus Torvalds 已提交
1358 1359
{
	struct packet_type *ptype;
1360

1361 1362 1363 1364
#ifdef CONFIG_NET_CLS_ACT
	if (!(skb->tstamp.tv64 && (G_TC_FROM(skb->tc_verd) & AT_INGRESS)))
		net_timestamp(skb);
#else
1365
	net_timestamp(skb);
1366
#endif
L
Linus Torvalds 已提交
1367 1368 1369 1370 1371 1372 1373 1374 1375

	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
1376
			struct sk_buff *skb2 = skb_clone(skb, GFP_ATOMIC);
L
Linus Torvalds 已提交
1377 1378 1379 1380 1381 1382 1383
			if (!skb2)
				break;

			/* skb->nh should be correctly
			   set by sender, so that the second statement is
			   just protection against buggy protocols.
			 */
1384
			skb_reset_mac_header(skb2);
L
Linus Torvalds 已提交
1385

1386
			if (skb_network_header(skb2) < skb2->data ||
1387
			    skb2->network_header > skb2->tail) {
L
Linus Torvalds 已提交
1388 1389 1390 1391
				if (net_ratelimit())
					printk(KERN_CRIT "protocol %04x is "
					       "buggy, dev %s\n",
					       skb2->protocol, dev->name);
1392
				skb_reset_network_header(skb2);
L
Linus Torvalds 已提交
1393 1394
			}

1395
			skb2->transport_header = skb2->network_header;
L
Linus Torvalds 已提交
1396
			skb2->pkt_type = PACKET_OUTGOING;
D
David S. Miller 已提交
1397
			ptype->func(skb2, skb->dev, ptype, skb->dev);
L
Linus Torvalds 已提交
1398 1399 1400 1401 1402
		}
	}
	rcu_read_unlock();
}

1403

1404
static inline void __netif_reschedule(struct Qdisc *q)
1405
{
1406 1407
	struct softnet_data *sd;
	unsigned long flags;
1408

1409 1410 1411 1412 1413 1414 1415 1416 1417 1418 1419 1420
	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);
1421 1422 1423
}
EXPORT_SYMBOL(__netif_schedule);

1424
void dev_kfree_skb_irq(struct sk_buff *skb)
1425
{
1426 1427 1428
	if (atomic_dec_and_test(&skb->users)) {
		struct softnet_data *sd;
		unsigned long flags;
1429

1430 1431 1432 1433 1434 1435 1436
		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);
	}
1437
}
1438
EXPORT_SYMBOL(dev_kfree_skb_irq);
1439 1440 1441 1442 1443 1444 1445 1446 1447 1448 1449

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);


1450 1451 1452 1453 1454 1455
/**
 * netif_device_detach - mark device as removed
 * @dev: network device
 *
 * Mark device as removed from system and therefore no longer available.
 */
1456 1457 1458 1459
void netif_device_detach(struct net_device *dev)
{
	if (test_and_clear_bit(__LINK_STATE_PRESENT, &dev->state) &&
	    netif_running(dev)) {
1460
		netif_tx_stop_all_queues(dev);
1461 1462 1463 1464
	}
}
EXPORT_SYMBOL(netif_device_detach);

1465 1466 1467 1468 1469 1470
/**
 * netif_device_attach - mark device as attached
 * @dev: network device
 *
 * Mark device as attached from system and restart if needed.
 */
1471 1472 1473 1474
void netif_device_attach(struct net_device *dev)
{
	if (!test_and_set_bit(__LINK_STATE_PRESENT, &dev->state) &&
	    netif_running(dev)) {
1475
		netif_tx_wake_all_queues(dev);
1476
		__netdev_watchdog_up(dev);
1477 1478 1479 1480
	}
}
EXPORT_SYMBOL(netif_device_attach);

1481 1482 1483 1484 1485 1486
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) &&
1487 1488 1489
		 protocol == htons(ETH_P_IPV6)) ||
		((features & NETIF_F_FCOE_CRC) &&
		 protocol == htons(ETH_P_FCOE)));
1490 1491 1492 1493 1494 1495 1496 1497 1498 1499 1500 1501 1502 1503 1504 1505
}

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;
}
1506

L
Linus Torvalds 已提交
1507 1508 1509 1510
/*
 * Invalidate hardware checksum when packet is to be mangled, and
 * complete checksum manually on outgoing path.
 */
1511
int skb_checksum_help(struct sk_buff *skb)
L
Linus Torvalds 已提交
1512
{
1513
	__wsum csum;
1514
	int ret = 0, offset;
L
Linus Torvalds 已提交
1515

1516
	if (skb->ip_summed == CHECKSUM_COMPLETE)
1517 1518 1519 1520 1521
		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
1522 1523
	}

1524 1525 1526 1527 1528 1529 1530 1531 1532
	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
1533 1534 1535 1536 1537
		ret = pskb_expand_head(skb, 0, 0, GFP_ATOMIC);
		if (ret)
			goto out;
	}

1538
	*(__sum16 *)(skb->data + offset) = csum_fold(csum);
1539
out_set_summed:
L
Linus Torvalds 已提交
1540
	skb->ip_summed = CHECKSUM_NONE;
1541
out:
L
Linus Torvalds 已提交
1542 1543
	return ret;
}
E
Eric Dumazet 已提交
1544
EXPORT_SYMBOL(skb_checksum_help);
L
Linus Torvalds 已提交
1545

1546 1547 1548
/**
 *	skb_gso_segment - Perform segmentation on skb.
 *	@skb: buffer to segment
1549
 *	@features: features for the output path (see dev->features)
1550 1551
 *
 *	This function segments the given skb and returns a list of segments.
1552 1553 1554
 *
 *	It may return NULL if the skb requires no segmentation.  This is
 *	only possible when GSO is used for verifying header integrity.
1555
 */
1556
struct sk_buff *skb_gso_segment(struct sk_buff *skb, int features)
1557 1558 1559
{
	struct sk_buff *segs = ERR_PTR(-EPROTONOSUPPORT);
	struct packet_type *ptype;
A
Al Viro 已提交
1560
	__be16 type = skb->protocol;
1561
	int err;
1562

1563
	skb_reset_mac_header(skb);
1564
	skb->mac_len = skb->network_header - skb->mac_header;
1565 1566
	__skb_pull(skb, skb->mac_len);

1567 1568 1569 1570 1571 1572 1573 1574 1575 1576 1577 1578 1579
	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);

1580 1581 1582 1583 1584
		if (skb_header_cloned(skb) &&
		    (err = pskb_expand_head(skb, 0, 0, GFP_ATOMIC)))
			return ERR_PTR(err);
	}

1585
	rcu_read_lock();
1586 1587
	list_for_each_entry_rcu(ptype,
			&ptype_base[ntohs(type) & PTYPE_HASH_MASK], list) {
1588
		if (ptype->type == type && !ptype->dev && ptype->gso_segment) {
1589
			if (unlikely(skb->ip_summed != CHECKSUM_PARTIAL)) {
1590 1591 1592 1593
				err = ptype->gso_send_check(skb);
				segs = ERR_PTR(err);
				if (err || skb_gso_ok(skb, features))
					break;
1594 1595
				__skb_push(skb, (skb->data -
						 skb_network_header(skb)));
1596
			}
1597
			segs = ptype->gso_segment(skb, features);
1598 1599 1600 1601 1602
			break;
		}
	}
	rcu_read_unlock();

1603
	__skb_push(skb, skb->data - skb_mac_header(skb));
1604

1605 1606 1607 1608
	return segs;
}
EXPORT_SYMBOL(skb_gso_segment);

1609 1610 1611 1612 1613
/* Take action when hardware reception checksum errors are detected. */
#ifdef CONFIG_BUG
void netdev_rx_csum_fault(struct net_device *dev)
{
	if (net_ratelimit()) {
1614
		printk(KERN_ERR "%s: hw csum failure.\n",
1615
			dev ? dev->name : "<unknown>");
1616 1617 1618 1619 1620 1621
		dump_stack();
	}
}
EXPORT_SYMBOL(netdev_rx_csum_fault);
#endif

L
Linus Torvalds 已提交
1622 1623 1624 1625 1626 1627 1628
/* 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.
 */

static inline int illegal_highdma(struct net_device *dev, struct sk_buff *skb)
{
1629
#ifdef CONFIG_HIGHMEM
L
Linus Torvalds 已提交
1630 1631 1632 1633 1634 1635 1636 1637 1638
	int i;

	if (dev->features & NETIF_F_HIGHDMA)
		return 0;

	for (i = 0; i < skb_shinfo(skb)->nr_frags; i++)
		if (PageHighMem(skb_shinfo(skb)->frags[i].page))
			return 1;

1639
#endif
L
Linus Torvalds 已提交
1640 1641 1642
	return 0;
}

1643 1644 1645 1646 1647 1648 1649 1650 1651 1652 1653 1654 1655 1656 1657 1658 1659 1660 1661 1662 1663 1664 1665 1666 1667 1668 1669 1670 1671 1672 1673 1674 1675 1676
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;
1677 1678 1679 1680 1681 1682 1683 1684
	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;
1685

1686
	if (IS_ERR(segs))
1687 1688 1689 1690 1691 1692 1693 1694 1695
		return PTR_ERR(segs);

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

	return 0;
}

1696 1697
int dev_hard_start_xmit(struct sk_buff *skb, struct net_device *dev,
			struct netdev_queue *txq)
1698
{
1699
	const struct net_device_ops *ops = dev->netdev_ops;
1700
	int rc;
1701

1702
	if (likely(!skb->next)) {
S
Stephen Hemminger 已提交
1703
		if (!list_empty(&ptype_all))
1704 1705
			dev_queue_xmit_nit(skb, dev);

1706 1707 1708 1709 1710 1711
		if (netif_needs_gso(dev, skb)) {
			if (unlikely(dev_gso_segment(skb)))
				goto out_kfree_skb;
			if (skb->next)
				goto gso;
		}
1712

1713 1714 1715 1716
		/*
		 * If device doesnt need skb->dst, release it right now while
		 * its hot in this cpu cache
		 */
E
Eric Dumazet 已提交
1717 1718 1719
		if (dev->priv_flags & IFF_XMIT_DST_RELEASE)
			skb_dst_drop(skb);

1720
		rc = ops->ndo_start_xmit(skb, dev);
1721
		if (rc == NETDEV_TX_OK)
E
Eric Dumazet 已提交
1722
			txq_trans_update(txq);
1723 1724 1725 1726 1727 1728 1729 1730 1731 1732 1733 1734 1735 1736 1737
		/*
		 * 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;
1738 1739
	}

1740
gso:
1741 1742 1743 1744 1745
	do {
		struct sk_buff *nskb = skb->next;

		skb->next = nskb->next;
		nskb->next = NULL;
1746
		rc = ops->ndo_start_xmit(nskb, dev);
1747
		if (unlikely(rc != NETDEV_TX_OK)) {
1748
			nskb->next = skb->next;
1749 1750 1751
			skb->next = nskb;
			return rc;
		}
E
Eric Dumazet 已提交
1752
		txq_trans_update(txq);
1753
		if (unlikely(netif_tx_queue_stopped(txq) && skb->next))
1754
			return NETDEV_TX_BUSY;
1755
	} while (skb->next);
1756

1757 1758 1759 1760
	skb->destructor = DEV_GSO_CB(skb)->destructor;

out_kfree_skb:
	kfree_skb(skb);
1761
	return NETDEV_TX_OK;
1762 1763
}

1764
static u32 skb_tx_hashrnd;
D
David S. Miller 已提交
1765

1766
u16 skb_tx_hash(const struct net_device *dev, const struct sk_buff *skb)
1767
{
1768
	u32 hash;
D
David S. Miller 已提交
1769

1770 1771
	if (skb_rx_queue_recorded(skb)) {
		hash = skb_get_rx_queue(skb);
E
Eric Dumazet 已提交
1772
		while (unlikely(hash >= dev->real_num_tx_queues))
1773 1774 1775
			hash -= dev->real_num_tx_queues;
		return hash;
	}
1776 1777

	if (skb->sk && skb->sk->sk_hash)
1778
		hash = skb->sk->sk_hash;
1779
	else
1780
		hash = skb->protocol;
1781

1782
	hash = jhash_1word(hash, skb_tx_hashrnd);
D
David S. Miller 已提交
1783 1784

	return (u16) (((u64) hash * dev->real_num_tx_queues) >> 32);
1785
}
1786
EXPORT_SYMBOL(skb_tx_hash);
1787

1788 1789 1790
static struct netdev_queue *dev_pick_tx(struct net_device *dev,
					struct sk_buff *skb)
{
1791 1792 1793 1794 1795 1796 1797 1798 1799 1800 1801 1802 1803 1804
	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);
		} else {
			queue_index = 0;
			if (dev->real_num_tx_queues > 1)
				queue_index = skb_tx_hash(dev, skb);
1805

1806 1807 1808 1809
			if (sk && sk->sk_dst_cache)
				sk_tx_queue_set(sk, queue_index);
		}
	}
1810

1811 1812
	skb_set_queue_mapping(skb, queue_index);
	return netdev_get_tx_queue(dev, queue_index);
1813 1814
}

1815 1816 1817 1818 1819 1820 1821 1822 1823 1824 1825 1826 1827 1828 1829 1830 1831 1832 1833 1834 1835 1836 1837 1838 1839 1840 1841 1842 1843 1844 1845 1846 1847 1848
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;
}

1849 1850 1851 1852 1853 1854 1855 1856 1857 1858 1859 1860 1861 1862 1863 1864 1865 1866 1867 1868 1869 1870 1871 1872 1873
/**
 *	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
1874 1875 1876
int dev_queue_xmit(struct sk_buff *skb)
{
	struct net_device *dev = skb->dev;
1877
	struct netdev_queue *txq;
L
Linus Torvalds 已提交
1878 1879 1880
	struct Qdisc *q;
	int rc = -ENOMEM;

1881 1882 1883 1884
	/* GSO will handle the following emulations directly. */
	if (netif_needs_gso(dev, skb))
		goto gso;

1885
	if (skb_has_frags(skb) &&
L
Linus Torvalds 已提交
1886
	    !(dev->features & NETIF_F_FRAGLIST) &&
H
Herbert Xu 已提交
1887
	    __skb_linearize(skb))
L
Linus Torvalds 已提交
1888 1889 1890 1891 1892 1893 1894 1895
		goto out_kfree_skb;

	/* Fragmented skb is linearized if device does not support SG,
	 * or if at least one of fragments is in highmem and device
	 * does not support DMA from it.
	 */
	if (skb_shinfo(skb)->nr_frags &&
	    (!(dev->features & NETIF_F_SG) || illegal_highdma(dev, skb)) &&
H
Herbert Xu 已提交
1896
	    __skb_linearize(skb))
L
Linus Torvalds 已提交
1897 1898 1899 1900 1901
		goto out_kfree_skb;

	/* If packet is not checksummed and device does not support
	 * checksumming for this protocol, complete checksumming here.
	 */
1902 1903 1904
	if (skb->ip_summed == CHECKSUM_PARTIAL) {
		skb_set_transport_header(skb, skb->csum_start -
					      skb_headroom(skb));
1905 1906
		if (!dev_can_checksum(dev, skb) && skb_checksum_help(skb))
			goto out_kfree_skb;
1907
	}
L
Linus Torvalds 已提交
1908

1909
gso:
1910 1911
	/* Disable soft irqs for various locks below. Also
	 * stops preemption for RCU.
L
Linus Torvalds 已提交
1912
	 */
1913
	rcu_read_lock_bh();
L
Linus Torvalds 已提交
1914

1915
	txq = dev_pick_tx(dev, skb);
1916
	q = rcu_dereference(txq->qdisc);
1917

L
Linus Torvalds 已提交
1918
#ifdef CONFIG_NET_CLS_ACT
E
Eric Dumazet 已提交
1919
	skb->tc_verd = SET_TC_AT(skb->tc_verd, AT_EGRESS);
L
Linus Torvalds 已提交
1920 1921
#endif
	if (q->enqueue) {
1922
		rc = __dev_xmit_skb(skb, q, dev, txq);
1923
		goto out;
L
Linus Torvalds 已提交
1924 1925 1926 1927 1928
	}

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

H
Herbert Xu 已提交
1929 1930
	   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
1931 1932 1933 1934 1935 1936 1937 1938 1939 1940
	   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 */

1941
		if (txq->xmit_lock_owner != cpu) {
L
Linus Torvalds 已提交
1942

1943
			HARD_TX_LOCK(dev, txq, cpu);
L
Linus Torvalds 已提交
1944

1945
			if (!netif_tx_queue_stopped(txq)) {
1946
				rc = NET_XMIT_SUCCESS;
1947
				if (!dev_hard_start_xmit(skb, dev, txq)) {
1948
					HARD_TX_UNLOCK(dev, txq);
L
Linus Torvalds 已提交
1949 1950 1951
					goto out;
				}
			}
1952
			HARD_TX_UNLOCK(dev, txq);
L
Linus Torvalds 已提交
1953 1954 1955 1956 1957 1958 1959 1960 1961 1962 1963 1964 1965
			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;
1966
	rcu_read_unlock_bh();
L
Linus Torvalds 已提交
1967 1968 1969 1970 1971

out_kfree_skb:
	kfree_skb(skb);
	return rc;
out:
1972
	rcu_read_unlock_bh();
L
Linus Torvalds 已提交
1973 1974
	return rc;
}
E
Eric Dumazet 已提交
1975
EXPORT_SYMBOL(dev_queue_xmit);
L
Linus Torvalds 已提交
1976 1977 1978 1979 1980 1981


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

1982 1983 1984
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
1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012

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


/**
 *	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)
{
	struct softnet_data *queue;
	unsigned long flags;

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

2013
	if (!skb->tstamp.tv64)
2014
		net_timestamp(skb);
L
Linus Torvalds 已提交
2015 2016 2017 2018 2019 2020 2021 2022 2023 2024 2025 2026 2027 2028

	/*
	 * The code is rearranged so that the path is the most
	 * short when CPU is congested, but is still operating.
	 */
	local_irq_save(flags);
	queue = &__get_cpu_var(softnet_data);

	__get_cpu_var(netdev_rx_stat).total++;
	if (queue->input_pkt_queue.qlen <= netdev_max_backlog) {
		if (queue->input_pkt_queue.qlen) {
enqueue:
			__skb_queue_tail(&queue->input_pkt_queue, skb);
			local_irq_restore(flags);
2029
			return NET_RX_SUCCESS;
L
Linus Torvalds 已提交
2030 2031
		}

2032
		napi_schedule(&queue->backlog);
L
Linus Torvalds 已提交
2033 2034 2035 2036 2037 2038 2039 2040 2041
		goto enqueue;
	}

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

	kfree_skb(skb);
	return NET_RX_DROP;
}
E
Eric Dumazet 已提交
2042
EXPORT_SYMBOL(netif_rx);
L
Linus Torvalds 已提交
2043 2044 2045 2046 2047 2048 2049 2050 2051 2052 2053 2054 2055 2056 2057 2058 2059 2060 2061 2062 2063 2064 2065 2066 2067 2068 2069 2070 2071 2072 2073

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;

2074
			WARN_ON(atomic_read(&skb->users));
L
Linus Torvalds 已提交
2075 2076 2077 2078 2079
			__kfree_skb(skb);
		}
	}

	if (sd->output_queue) {
2080
		struct Qdisc *head;
L
Linus Torvalds 已提交
2081 2082 2083 2084 2085 2086 2087

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

		while (head) {
2088 2089 2090
			struct Qdisc *q = head;
			spinlock_t *root_lock;

L
Linus Torvalds 已提交
2091 2092
			head = head->next_sched;

2093
			root_lock = qdisc_lock(q);
2094
			if (spin_trylock(root_lock)) {
2095 2096 2097
				smp_mb__before_clear_bit();
				clear_bit(__QDISC_STATE_SCHED,
					  &q->state);
2098 2099
				qdisc_run(q);
				spin_unlock(root_lock);
L
Linus Torvalds 已提交
2100
			} else {
2101
				if (!test_bit(__QDISC_STATE_DEACTIVATED,
2102
					      &q->state)) {
2103
					__netif_reschedule(q);
2104 2105 2106 2107 2108
				} else {
					smp_mb__before_clear_bit();
					clear_bit(__QDISC_STATE_SCHED,
						  &q->state);
				}
L
Linus Torvalds 已提交
2109 2110 2111 2112 2113
			}
		}
	}
}

2114 2115 2116
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
2117 2118
{
	atomic_inc(&skb->users);
D
David S. Miller 已提交
2119
	return pt_prev->func(skb, skb->dev, pt_prev, orig_dev);
L
Linus Torvalds 已提交
2120 2121 2122
}

#if defined(CONFIG_BRIDGE) || defined (CONFIG_BRIDGE_MODULE)
2123 2124 2125 2126 2127

#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;
2128
EXPORT_SYMBOL_GPL(br_fdb_test_addr_hook);
2129
#endif
L
Linus Torvalds 已提交
2130

2131 2132 2133 2134 2135 2136
/*
 * 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;
2137
EXPORT_SYMBOL_GPL(br_handle_frame_hook);
2138

2139 2140 2141
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
2142 2143 2144
{
	struct net_bridge_port *port;

2145 2146 2147
	if (skb->pkt_type == PACKET_LOOPBACK ||
	    (port = rcu_dereference(skb->dev->br_port)) == NULL)
		return skb;
L
Linus Torvalds 已提交
2148 2149

	if (*pt_prev) {
2150
		*ret = deliver_skb(skb, *pt_prev, orig_dev);
L
Linus Torvalds 已提交
2151
		*pt_prev = NULL;
2152 2153
	}

2154
	return br_handle_frame_hook(port, skb);
L
Linus Torvalds 已提交
2155 2156
}
#else
2157
#define handle_bridge(skb, pt_prev, ret, orig_dev)	(skb)
L
Linus Torvalds 已提交
2158 2159
#endif

P
Patrick McHardy 已提交
2160 2161 2162 2163 2164 2165 2166 2167 2168 2169 2170 2171 2172 2173 2174 2175 2176 2177 2178 2179 2180 2181
#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
2182 2183 2184 2185 2186
#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
2187
 * NOTE: This doesnt stop any functionality; if you dont have
L
Linus Torvalds 已提交
2188 2189 2190
 * the ingress scheduler, you just cant add policies on ingress.
 *
 */
2191
static int ing_filter(struct sk_buff *skb)
L
Linus Torvalds 已提交
2192 2193
{
	struct net_device *dev = skb->dev;
2194
	u32 ttl = G_TC_RTTL(skb->tc_verd);
2195 2196 2197
	struct netdev_queue *rxq;
	int result = TC_ACT_OK;
	struct Qdisc *q;
2198

2199 2200 2201 2202 2203 2204
	if (MAX_RED_LOOP < ttl++) {
		printk(KERN_WARNING
		       "Redir loop detected Dropping packet (%d->%d)\n",
		       skb->iif, dev->ifindex);
		return TC_ACT_SHOT;
	}
L
Linus Torvalds 已提交
2205

2206 2207
	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
2208

2209 2210
	rxq = &dev->rx_queue;

2211
	q = rxq->qdisc;
2212
	if (q != &noop_qdisc) {
2213
		spin_lock(qdisc_lock(q));
2214 2215
		if (likely(!test_bit(__QDISC_STATE_DEACTIVATED, &q->state)))
			result = qdisc_enqueue_root(skb, q);
2216 2217
		spin_unlock(qdisc_lock(q));
	}
2218 2219 2220

	return result;
}
2221

2222 2223 2224 2225
static inline struct sk_buff *handle_ing(struct sk_buff *skb,
					 struct packet_type **pt_prev,
					 int *ret, struct net_device *orig_dev)
{
2226
	if (skb->dev->rx_queue.qdisc == &noop_qdisc)
2227
		goto out;
L
Linus Torvalds 已提交
2228

2229 2230 2231 2232 2233 2234
	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
2235 2236
	}

2237 2238 2239 2240 2241 2242 2243 2244 2245 2246
	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
2247 2248 2249
}
#endif

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
/*
 * 	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();
}

2277 2278 2279 2280 2281 2282 2283 2284 2285 2286 2287 2288 2289 2290 2291
/**
 *	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
 */
L
Linus Torvalds 已提交
2292 2293 2294
int netif_receive_skb(struct sk_buff *skb)
{
	struct packet_type *ptype, *pt_prev;
D
David S. Miller 已提交
2295
	struct net_device *orig_dev;
2296
	struct net_device *null_or_orig;
L
Linus Torvalds 已提交
2297
	int ret = NET_RX_DROP;
A
Al Viro 已提交
2298
	__be16 type;
L
Linus Torvalds 已提交
2299

2300 2301 2302
	if (!skb->tstamp.tv64)
		net_timestamp(skb);

E
Eric Dumazet 已提交
2303
	if (vlan_tx_tag_present(skb) && vlan_hwaccel_do_receive(skb))
2304 2305
		return NET_RX_SUCCESS;

L
Linus Torvalds 已提交
2306
	/* if we've gotten here through NAPI, check netpoll */
2307
	if (netpoll_receive_skb(skb))
L
Linus Torvalds 已提交
2308 2309
		return NET_RX_DROP;

2310 2311
	if (!skb->iif)
		skb->iif = skb->dev->ifindex;
2312

2313
	null_or_orig = NULL;
J
Joe Eykholt 已提交
2314 2315
	orig_dev = skb->dev;
	if (orig_dev->master) {
2316 2317 2318 2319
		if (skb_bond_should_drop(skb))
			null_or_orig = orig_dev; /* deliver only exact match */
		else
			skb->dev = orig_dev->master;
J
Joe Eykholt 已提交
2320
	}
2321

L
Linus Torvalds 已提交
2322 2323
	__get_cpu_var(netdev_rx_stat).total++;

2324
	skb_reset_network_header(skb);
2325
	skb_reset_transport_header(skb);
2326
	skb->mac_len = skb->network_header - skb->mac_header;
L
Linus Torvalds 已提交
2327 2328 2329 2330 2331 2332 2333 2334 2335 2336 2337 2338 2339

	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) {
2340 2341
		if (ptype->dev == null_or_orig || ptype->dev == skb->dev ||
		    ptype->dev == orig_dev) {
2342
			if (pt_prev)
D
David S. Miller 已提交
2343
				ret = deliver_skb(skb, pt_prev, orig_dev);
L
Linus Torvalds 已提交
2344 2345 2346 2347 2348
			pt_prev = ptype;
		}
	}

#ifdef CONFIG_NET_CLS_ACT
2349 2350
	skb = handle_ing(skb, &pt_prev, &ret, orig_dev);
	if (!skb)
L
Linus Torvalds 已提交
2351 2352 2353 2354
		goto out;
ncls:
#endif

2355
	skb = handle_bridge(skb, &pt_prev, &ret, orig_dev);
P
Patrick McHardy 已提交
2356 2357 2358
	if (!skb)
		goto out;
	skb = handle_macvlan(skb, &pt_prev, &ret, orig_dev);
2359
	if (!skb)
L
Linus Torvalds 已提交
2360 2361 2362
		goto out;

	type = skb->protocol;
2363 2364
	list_for_each_entry_rcu(ptype,
			&ptype_base[ntohs(type) & PTYPE_HASH_MASK], list) {
L
Linus Torvalds 已提交
2365
		if (ptype->type == type &&
2366 2367
		    (ptype->dev == null_or_orig || ptype->dev == skb->dev ||
		     ptype->dev == orig_dev)) {
2368
			if (pt_prev)
D
David S. Miller 已提交
2369
				ret = deliver_skb(skb, pt_prev, orig_dev);
L
Linus Torvalds 已提交
2370 2371 2372 2373 2374
			pt_prev = ptype;
		}
	}

	if (pt_prev) {
D
David S. Miller 已提交
2375
		ret = pt_prev->func(skb, skb->dev, pt_prev, orig_dev);
L
Linus Torvalds 已提交
2376 2377 2378 2379 2380 2381 2382 2383 2384 2385 2386 2387
	} 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;
}
E
Eric Dumazet 已提交
2388
EXPORT_SYMBOL(netif_receive_skb);
L
Linus Torvalds 已提交
2389

2390 2391 2392 2393 2394 2395 2396 2397 2398 2399 2400 2401 2402 2403
/* Network device is going away, flush any packets still pending  */
static void flush_backlog(void *arg)
{
	struct net_device *dev = arg;
	struct softnet_data *queue = &__get_cpu_var(softnet_data);
	struct sk_buff *skb, *tmp;

	skb_queue_walk_safe(&queue->input_pkt_queue, skb, tmp)
		if (skb->dev == dev) {
			__skb_unlink(skb, &queue->input_pkt_queue);
			kfree_skb(skb);
		}
}

2404 2405 2406 2407 2408 2409 2410
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;

2411 2412
	if (NAPI_GRO_CB(skb)->count == 1) {
		skb_shinfo(skb)->gso_size = 0;
2413
		goto out;
2414
	}
2415 2416 2417 2418 2419 2420 2421 2422 2423 2424 2425 2426 2427 2428 2429 2430 2431 2432 2433 2434 2435 2436 2437 2438 2439 2440 2441 2442 2443 2444 2445

	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);
}

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

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

2446
	napi->gro_count = 0;
2447 2448 2449 2450
	napi->gro_list = NULL;
}
EXPORT_SYMBOL(napi_gro_flush);

2451
int dev_gro_receive(struct napi_struct *napi, struct sk_buff *skb)
2452 2453 2454 2455 2456
{
	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
2457
	int same_flow;
2458
	int mac_len;
2459
	int ret;
2460 2461 2462 2463

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

2464
	if (skb_is_gso(skb) || skb_has_frags(skb))
2465 2466
		goto normal;

2467 2468 2469 2470 2471
	rcu_read_lock();
	list_for_each_entry_rcu(ptype, head, list) {
		if (ptype->type != type || ptype->dev || !ptype->gro_receive)
			continue;

2472
		skb_set_network_header(skb, skb_gro_offset(skb));
2473 2474 2475 2476
		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
2477
		NAPI_GRO_CB(skb)->free = 0;
2478 2479 2480 2481 2482 2483 2484 2485 2486

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

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

H
Herbert Xu 已提交
2487
	same_flow = NAPI_GRO_CB(skb)->same_flow;
2488
	ret = NAPI_GRO_CB(skb)->free ? GRO_MERGED_FREE : GRO_MERGED;
H
Herbert Xu 已提交
2489

2490 2491 2492 2493 2494 2495
	if (pp) {
		struct sk_buff *nskb = *pp;

		*pp = nskb->next;
		nskb->next = NULL;
		napi_gro_complete(nskb);
2496
		napi->gro_count--;
2497 2498
	}

H
Herbert Xu 已提交
2499
	if (same_flow)
2500 2501
		goto ok;

2502
	if (NAPI_GRO_CB(skb)->flush || napi->gro_count >= MAX_GRO_SKBS)
2503 2504
		goto normal;

2505
	napi->gro_count++;
2506
	NAPI_GRO_CB(skb)->count = 1;
2507
	skb_shinfo(skb)->gso_size = skb_gro_len(skb);
2508 2509
	skb->next = napi->gro_list;
	napi->gro_list = skb;
2510
	ret = GRO_HELD;
2511

2512
pull:
H
Herbert Xu 已提交
2513 2514 2515 2516 2517 2518 2519 2520 2521 2522 2523 2524 2525 2526 2527 2528 2529 2530 2531
	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);
		}
2532 2533
	}

2534
ok:
2535
	return ret;
2536 2537

normal:
2538 2539
	ret = GRO_NORMAL;
	goto pull;
H
Herbert Xu 已提交
2540
}
2541 2542 2543 2544 2545 2546
EXPORT_SYMBOL(dev_gro_receive);

static int __napi_gro_receive(struct napi_struct *napi, struct sk_buff *skb)
{
	struct sk_buff *p;

2547 2548 2549
	if (netpoll_rx_on(skb))
		return GRO_NORMAL;

2550
	for (p = napi->gro_list; p; p = p->next) {
2551 2552 2553
		NAPI_GRO_CB(p)->same_flow = (p->dev == skb->dev)
			&& !compare_ether_header(skb_mac_header(p),
						 skb_gro_mac_header(skb));
2554 2555 2556 2557 2558
		NAPI_GRO_CB(p)->flush = 0;
	}

	return dev_gro_receive(napi, skb);
}
H
Herbert Xu 已提交
2559

2560
int napi_skb_finish(int ret, struct sk_buff *skb)
H
Herbert Xu 已提交
2561
{
2562 2563 2564 2565
	int err = NET_RX_SUCCESS;

	switch (ret) {
	case GRO_NORMAL:
H
Herbert Xu 已提交
2566 2567
		return netif_receive_skb(skb);

2568 2569 2570 2571 2572
	case GRO_DROP:
		err = NET_RX_DROP;
		/* fall through */

	case GRO_MERGED_FREE:
H
Herbert Xu 已提交
2573 2574 2575 2576
		kfree_skb(skb);
		break;
	}

2577 2578 2579 2580
	return err;
}
EXPORT_SYMBOL(napi_skb_finish);

2581 2582 2583 2584
void skb_gro_reset_offset(struct sk_buff *skb)
{
	NAPI_GRO_CB(skb)->data_offset = 0;
	NAPI_GRO_CB(skb)->frag0 = NULL;
2585
	NAPI_GRO_CB(skb)->frag0_len = 0;
2586

2587
	if (skb->mac_header == skb->tail &&
2588
	    !PageHighMem(skb_shinfo(skb)->frags[0].page)) {
2589 2590 2591
		NAPI_GRO_CB(skb)->frag0 =
			page_address(skb_shinfo(skb)->frags[0].page) +
			skb_shinfo(skb)->frags[0].page_offset;
2592 2593
		NAPI_GRO_CB(skb)->frag0_len = skb_shinfo(skb)->frags[0].size;
	}
2594 2595 2596
}
EXPORT_SYMBOL(skb_gro_reset_offset);

2597 2598
int napi_gro_receive(struct napi_struct *napi, struct sk_buff *skb)
{
2599 2600
	skb_gro_reset_offset(skb);

2601
	return napi_skb_finish(__napi_gro_receive(napi, skb), skb);
2602 2603 2604
}
EXPORT_SYMBOL(napi_gro_receive);

2605 2606 2607 2608 2609 2610 2611 2612 2613
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);

2614
struct sk_buff *napi_get_frags(struct napi_struct *napi)
H
Herbert Xu 已提交
2615 2616 2617 2618
{
	struct sk_buff *skb = napi->skb;

	if (!skb) {
2619 2620 2621
		skb = netdev_alloc_skb_ip_align(napi->dev, GRO_MAX_HEAD);
		if (skb)
			napi->skb = skb;
2622
	}
2623 2624
	return skb;
}
2625
EXPORT_SYMBOL(napi_get_frags);
2626

2627
int napi_frags_finish(struct napi_struct *napi, struct sk_buff *skb, int ret)
2628
{
2629
	int err = NET_RX_SUCCESS;
2630

2631 2632
	switch (ret) {
	case GRO_NORMAL:
2633 2634 2635 2636 2637 2638 2639 2640
	case GRO_HELD:
		skb->protocol = eth_type_trans(skb, napi->dev);

		if (ret == GRO_NORMAL)
			return netif_receive_skb(skb);

		skb_gro_pull(skb, -ETH_HLEN);
		break;
H
Herbert Xu 已提交
2641

2642 2643 2644
	case GRO_DROP:
		err = NET_RX_DROP;
		/* fall through */
H
Herbert Xu 已提交
2645

2646 2647 2648 2649
	case GRO_MERGED_FREE:
		napi_reuse_skb(napi, skb);
		break;
	}
H
Herbert Xu 已提交
2650 2651 2652

	return err;
}
2653 2654
EXPORT_SYMBOL(napi_frags_finish);

2655 2656 2657 2658
struct sk_buff *napi_frags_skb(struct napi_struct *napi)
{
	struct sk_buff *skb = napi->skb;
	struct ethhdr *eth;
2659 2660
	unsigned int hlen;
	unsigned int off;
2661 2662 2663 2664 2665 2666

	napi->skb = NULL;

	skb_reset_mac_header(skb);
	skb_gro_reset_offset(skb);

2667 2668 2669 2670 2671 2672 2673 2674 2675 2676
	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;
		}
2677 2678 2679 2680 2681 2682 2683 2684 2685 2686 2687 2688 2689 2690 2691 2692
	}

	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);

int napi_gro_frags(struct napi_struct *napi)
2693
{
2694
	struct sk_buff *skb = napi_frags_skb(napi);
2695 2696 2697 2698 2699 2700

	if (!skb)
		return NET_RX_DROP;

	return napi_frags_finish(napi, skb, __napi_gro_receive(napi, skb));
}
H
Herbert Xu 已提交
2701 2702
EXPORT_SYMBOL(napi_gro_frags);

2703
static int process_backlog(struct napi_struct *napi, int quota)
L
Linus Torvalds 已提交
2704 2705 2706 2707 2708
{
	int work = 0;
	struct softnet_data *queue = &__get_cpu_var(softnet_data);
	unsigned long start_time = jiffies;

2709 2710
	napi->weight = weight_p;
	do {
L
Linus Torvalds 已提交
2711 2712 2713 2714
		struct sk_buff *skb;

		local_irq_disable();
		skb = __skb_dequeue(&queue->input_pkt_queue);
2715
		if (!skb) {
2716
			__napi_complete(napi);
2717
			local_irq_enable();
2718
			break;
2719
		}
L
Linus Torvalds 已提交
2720 2721
		local_irq_enable();

2722
		netif_receive_skb(skb);
2723
	} while (++work < quota && jiffies == start_time);
L
Linus Torvalds 已提交
2724

2725 2726
	return work;
}
L
Linus Torvalds 已提交
2727

2728 2729
/**
 * __napi_schedule - schedule for receive
2730
 * @n: entry to schedule
2731 2732 2733
 *
 * The entry's receive function will be scheduled to run
 */
H
Harvey Harrison 已提交
2734
void __napi_schedule(struct napi_struct *n)
2735 2736
{
	unsigned long flags;
L
Linus Torvalds 已提交
2737

2738 2739 2740 2741
	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
2742
}
2743 2744
EXPORT_SYMBOL(__napi_schedule);

2745 2746 2747 2748 2749 2750 2751 2752 2753 2754 2755 2756 2757 2758 2759 2760 2761 2762 2763 2764 2765 2766 2767 2768 2769 2770 2771 2772 2773 2774 2775 2776 2777
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);
2778
	napi->gro_count = 0;
2779
	napi->gro_list = NULL;
H
Herbert Xu 已提交
2780
	napi->skb = NULL;
2781 2782 2783 2784
	napi->poll = poll;
	napi->weight = weight;
	list_add(&napi->dev_list, &dev->napi_list);
	napi->dev = dev;
H
Herbert Xu 已提交
2785
#ifdef CONFIG_NETPOLL
2786 2787 2788 2789 2790 2791 2792 2793 2794 2795 2796
	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;

2797
	list_del_init(&napi->dev_list);
2798
	napi_free_frags(napi);
2799 2800 2801 2802 2803 2804 2805 2806

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

	napi->gro_list = NULL;
2807
	napi->gro_count = 0;
2808 2809 2810
}
EXPORT_SYMBOL(netif_napi_del);

L
Linus Torvalds 已提交
2811 2812 2813

static void net_rx_action(struct softirq_action *h)
{
2814
	struct list_head *list = &__get_cpu_var(softnet_data).poll_list;
2815
	unsigned long time_limit = jiffies + 2;
2816
	int budget = netdev_budget;
2817 2818
	void *have;

L
Linus Torvalds 已提交
2819 2820
	local_irq_disable();

2821 2822 2823
	while (!list_empty(list)) {
		struct napi_struct *n;
		int work, weight;
L
Linus Torvalds 已提交
2824

2825
		/* If softirq window is exhuasted then punt.
2826 2827
		 * Allow this to run for 2 jiffies since which will allow
		 * an average latency of 1.5/HZ.
2828
		 */
2829
		if (unlikely(budget <= 0 || time_after(jiffies, time_limit)))
L
Linus Torvalds 已提交
2830 2831 2832 2833
			goto softnet_break;

		local_irq_enable();

2834 2835 2836 2837 2838 2839
		/* 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.
		 */
		n = list_entry(list->next, struct napi_struct, poll_list);
L
Linus Torvalds 已提交
2840

2841 2842 2843 2844
		have = netpoll_poll_lock(n);

		weight = n->weight;

2845 2846 2847 2848 2849 2850 2851
		/* 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;
2852
		if (test_bit(NAPI_STATE_SCHED, &n->state)) {
2853
			work = n->poll(n, weight);
2854 2855
			trace_napi_poll(n);
		}
2856 2857 2858 2859 2860 2861 2862 2863 2864 2865 2866 2867

		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.
		 */
2868
		if (unlikely(work == weight)) {
2869 2870 2871 2872 2873
			if (unlikely(napi_disable_pending(n))) {
				local_irq_enable();
				napi_complete(n);
				local_irq_disable();
			} else
2874 2875
				list_move_tail(&n->poll_list, list);
		}
2876 2877

		netpoll_poll_unlock(have);
L
Linus Torvalds 已提交
2878 2879
	}
out:
2880
	local_irq_enable();
2881

2882 2883 2884 2885 2886
#ifdef CONFIG_NET_DMA
	/*
	 * There may not be any more sk_buffs coming right now, so push
	 * any pending DMA copies to hardware
	 */
2887
	dma_issue_pending_all();
2888
#endif
2889

L
Linus Torvalds 已提交
2890 2891 2892 2893 2894 2895 2896 2897
	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
2898
static gifconf_func_t *gifconf_list[NPROTO];
L
Linus Torvalds 已提交
2899 2900 2901 2902 2903 2904 2905 2906 2907 2908

/**
 *	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
2909
int register_gifconf(unsigned int family, gifconf_func_t *gifconf)
L
Linus Torvalds 已提交
2910 2911 2912 2913 2914 2915
{
	if (family >= NPROTO)
		return -EINVAL;
	gifconf_list[family] = gifconf;
	return 0;
}
E
Eric Dumazet 已提交
2916
EXPORT_SYMBOL(register_gifconf);
L
Linus Torvalds 已提交
2917 2918 2919 2920 2921 2922 2923 2924 2925 2926 2927 2928 2929


/*
 *	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
 */

2930
static int dev_ifname(struct net *net, struct ifreq __user *arg)
L
Linus Torvalds 已提交
2931 2932 2933 2934 2935 2936 2937 2938 2939 2940 2941 2942
{
	struct net_device *dev;
	struct ifreq ifr;

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

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

	read_lock(&dev_base_lock);
2943
	dev = __dev_get_by_index(net, ifr.ifr_ifindex);
L
Linus Torvalds 已提交
2944 2945 2946 2947 2948 2949 2950 2951 2952 2953 2954 2955 2956 2957 2958 2959 2960 2961 2962
	if (!dev) {
		read_unlock(&dev_base_lock);
		return -ENODEV;
	}

	strcpy(ifr.ifr_name, dev->name);
	read_unlock(&dev_base_lock);

	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'.
 */

2963
static int dev_ifconf(struct net *net, char __user *arg)
L
Linus Torvalds 已提交
2964 2965 2966 2967 2968 2969 2970 2971 2972 2973 2974 2975 2976 2977 2978 2979 2980 2981 2982 2983 2984 2985 2986
{
	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;
2987
	for_each_netdev(net, dev) {
L
Linus Torvalds 已提交
2988 2989 2990 2991 2992 2993 2994 2995 2996 2997 2998 2999 3000
		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;
			}
		}
3001
	}
L
Linus Torvalds 已提交
3002 3003 3004 3005 3006 3007 3008 3009 3010 3011 3012 3013 3014 3015 3016 3017 3018

	/*
	 *	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.
 */
3019
void *dev_seq_start(struct seq_file *seq, loff_t *pos)
3020
	__acquires(dev_base_lock)
L
Linus Torvalds 已提交
3021
{
3022
	struct net *net = seq_file_net(seq);
3023
	loff_t off;
L
Linus Torvalds 已提交
3024 3025
	struct net_device *dev;

3026 3027 3028
	read_lock(&dev_base_lock);
	if (!*pos)
		return SEQ_START_TOKEN;
L
Linus Torvalds 已提交
3029

3030
	off = 1;
3031
	for_each_netdev(net, dev)
3032 3033
		if (off++ == *pos)
			return dev;
L
Linus Torvalds 已提交
3034

3035
	return NULL;
L
Linus Torvalds 已提交
3036 3037 3038 3039
}

void *dev_seq_next(struct seq_file *seq, void *v, loff_t *pos)
{
3040
	struct net *net = seq_file_net(seq);
L
Linus Torvalds 已提交
3041
	++*pos;
3042
	return v == SEQ_START_TOKEN ?
3043
		first_net_device(net) : next_net_device((struct net_device *)v);
L
Linus Torvalds 已提交
3044 3045 3046
}

void dev_seq_stop(struct seq_file *seq, void *v)
3047
	__releases(dev_base_lock)
L
Linus Torvalds 已提交
3048 3049 3050 3051 3052 3053
{
	read_unlock(&dev_base_lock);
}

static void dev_seq_printf_stats(struct seq_file *seq, struct net_device *dev)
{
3054
	const struct net_device_stats *stats = dev_get_stats(dev);
L
Linus Torvalds 已提交
3055

3056 3057 3058 3059 3060 3061 3062 3063 3064 3065 3066 3067 3068 3069 3070 3071 3072
	seq_printf(seq, "%6s:%8lu %7lu %4lu %4lu %4lu %5lu %10lu %9lu "
		   "%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
3073 3074 3075 3076 3077 3078 3079 3080 3081 3082 3083 3084 3085 3086 3087 3088 3089 3090 3091 3092 3093 3094 3095
}

/*
 *	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;

3096
	while (*pos < nr_cpu_ids)
3097
		if (cpu_online(*pos)) {
L
Linus Torvalds 已提交
3098 3099 3100 3101 3102 3103 3104 3105 3106 3107 3108 3109 3110 3111 3112 3113 3114 3115 3116 3117 3118 3119 3120 3121 3122 3123 3124
			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;

	seq_printf(seq, "%08x %08x %08x %08x %08x %08x %08x %08x %08x\n",
3125
		   s->total, s->dropped, s->time_squeeze, 0,
3126
		   0, 0, 0, 0, /* was fastroute */
E
Eric Dumazet 已提交
3127
		   s->cpu_collision);
L
Linus Torvalds 已提交
3128 3129 3130
	return 0;
}

3131
static const struct seq_operations dev_seq_ops = {
L
Linus Torvalds 已提交
3132 3133 3134 3135 3136 3137 3138 3139
	.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)
{
3140 3141
	return seq_open_net(inode, file, &dev_seq_ops,
			    sizeof(struct seq_net_private));
L
Linus Torvalds 已提交
3142 3143
}

3144
static const struct file_operations dev_seq_fops = {
L
Linus Torvalds 已提交
3145 3146 3147 3148
	.owner	 = THIS_MODULE,
	.open    = dev_seq_open,
	.read    = seq_read,
	.llseek  = seq_lseek,
3149
	.release = seq_release_net,
L
Linus Torvalds 已提交
3150 3151
};

3152
static const struct seq_operations softnet_seq_ops = {
L
Linus Torvalds 已提交
3153 3154 3155 3156 3157 3158 3159 3160 3161 3162 3163
	.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);
}

3164
static const struct file_operations softnet_seq_fops = {
L
Linus Torvalds 已提交
3165 3166 3167 3168 3169 3170 3171
	.owner	 = THIS_MODULE,
	.open    = softnet_seq_open,
	.read    = seq_read,
	.llseek  = seq_lseek,
	.release = seq_release,
};

3172 3173 3174 3175 3176 3177 3178 3179 3180 3181 3182 3183
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;
	}

3184
	for (t = 0; t < PTYPE_HASH_SIZE; t++) {
3185 3186 3187 3188 3189 3190 3191 3192 3193 3194
		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)
3195
	__acquires(RCU)
3196 3197 3198 3199 3200 3201 3202 3203 3204 3205 3206 3207 3208 3209 3210 3211 3212 3213 3214 3215 3216 3217 3218
{
	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
3219
		hash = ntohs(pt->type) & PTYPE_HASH_MASK;
3220 3221

	while (nxt == &ptype_base[hash]) {
3222
		if (++hash >= PTYPE_HASH_SIZE)
3223 3224 3225 3226 3227 3228 3229 3230
			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)
3231
	__releases(RCU)
3232 3233 3234 3235 3236 3237 3238 3239 3240 3241
{
	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");
3242
	else if (pt->dev == NULL || dev_net(pt->dev) == seq_file_net(seq)) {
3243 3244 3245 3246 3247
		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
3248 3249
		seq_printf(seq, " %-8s %pF\n",
			   pt->dev ? pt->dev->name : "", pt->func);
3250 3251 3252 3253 3254 3255 3256 3257 3258 3259 3260 3261 3262 3263
	}

	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)
{
3264 3265
	return seq_open_net(inode, file, &ptype_seq_ops,
			sizeof(struct seq_net_private));
3266 3267 3268 3269 3270 3271 3272
}

static const struct file_operations ptype_seq_fops = {
	.owner	 = THIS_MODULE,
	.open    = ptype_seq_open,
	.read    = seq_read,
	.llseek  = seq_lseek,
3273
	.release = seq_release_net,
3274 3275 3276
};


3277
static int __net_init dev_proc_net_init(struct net *net)
L
Linus Torvalds 已提交
3278 3279 3280
{
	int rc = -ENOMEM;

3281
	if (!proc_net_fops_create(net, "dev", S_IRUGO, &dev_seq_fops))
L
Linus Torvalds 已提交
3282
		goto out;
3283
	if (!proc_net_fops_create(net, "softnet_stat", S_IRUGO, &softnet_seq_fops))
L
Linus Torvalds 已提交
3284
		goto out_dev;
3285
	if (!proc_net_fops_create(net, "ptype", S_IRUGO, &ptype_seq_fops))
3286
		goto out_softnet;
3287

3288
	if (wext_proc_init(net))
3289
		goto out_ptype;
L
Linus Torvalds 已提交
3290 3291 3292
	rc = 0;
out:
	return rc;
3293
out_ptype:
3294
	proc_net_remove(net, "ptype");
L
Linus Torvalds 已提交
3295
out_softnet:
3296
	proc_net_remove(net, "softnet_stat");
L
Linus Torvalds 已提交
3297
out_dev:
3298
	proc_net_remove(net, "dev");
L
Linus Torvalds 已提交
3299 3300
	goto out;
}
3301

3302
static void __net_exit dev_proc_net_exit(struct net *net)
3303 3304 3305 3306 3307 3308 3309 3310
{
	wext_proc_exit(net);

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

3311
static struct pernet_operations __net_initdata dev_proc_ops = {
3312 3313 3314 3315 3316 3317 3318 3319
	.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
3320 3321 3322 3323 3324 3325 3326 3327 3328 3329 3330 3331 3332 3333 3334 3335 3336 3337 3338 3339 3340 3341 3342 3343 3344 3345 3346 3347 3348
#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;
3349

L
Linus Torvalds 已提交
3350 3351 3352 3353 3354 3355 3356 3357 3358 3359 3360 3361 3362
	synchronize_net();

	if (old)
		dev_put(old);

	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
3363
EXPORT_SYMBOL(netdev_set_master);
L
Linus Torvalds 已提交
3364

3365 3366
static void dev_change_rx_flags(struct net_device *dev, int flags)
{
3367 3368 3369 3370
	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);
3371 3372
}

3373
static int __dev_set_promiscuity(struct net_device *dev, int inc)
L
Linus Torvalds 已提交
3374 3375
{
	unsigned short old_flags = dev->flags;
3376 3377
	uid_t uid;
	gid_t gid;
L
Linus Torvalds 已提交
3378

3379 3380
	ASSERT_RTNL();

3381 3382 3383 3384 3385 3386 3387 3388 3389 3390 3391 3392 3393 3394 3395 3396 3397
	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;
		}
	}
3398
	if (dev->flags != old_flags) {
L
Linus Torvalds 已提交
3399 3400
		printk(KERN_INFO "device %s %s promiscuous mode\n",
		       dev->name, (dev->flags & IFF_PROMISC) ? "entered" :
3401
							       "left");
3402 3403
		if (audit_enabled) {
			current_uid_gid(&uid, &gid);
3404 3405 3406 3407 3408 3409
			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),
3410
				uid, gid,
3411
				audit_get_sessionid(current));
3412
		}
3413

3414
		dev_change_rx_flags(dev, IFF_PROMISC);
L
Linus Torvalds 已提交
3415
	}
3416
	return 0;
L
Linus Torvalds 已提交
3417 3418
}

3419 3420 3421 3422 3423 3424 3425 3426 3427
/**
 *	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.
3428
 *	Return 0 if successful or a negative errno code on error.
3429
 */
3430
int dev_set_promiscuity(struct net_device *dev, int inc)
3431 3432
{
	unsigned short old_flags = dev->flags;
3433
	int err;
3434

3435
	err = __dev_set_promiscuity(dev, inc);
3436
	if (err < 0)
3437
		return err;
3438 3439
	if (dev->flags != old_flags)
		dev_set_rx_mode(dev);
3440
	return err;
3441
}
E
Eric Dumazet 已提交
3442
EXPORT_SYMBOL(dev_set_promiscuity);
3443

L
Linus Torvalds 已提交
3444 3445 3446 3447 3448 3449 3450 3451 3452 3453
/**
 *	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.
3454
 *	Return 0 if successful or a negative errno code on error.
L
Linus Torvalds 已提交
3455 3456
 */

3457
int dev_set_allmulti(struct net_device *dev, int inc)
L
Linus Torvalds 已提交
3458 3459 3460
{
	unsigned short old_flags = dev->flags;

3461 3462
	ASSERT_RTNL();

L
Linus Torvalds 已提交
3463
	dev->flags |= IFF_ALLMULTI;
3464 3465 3466 3467 3468 3469 3470 3471 3472 3473 3474 3475 3476 3477 3478 3479
	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;
		}
	}
3480
	if (dev->flags ^ old_flags) {
3481
		dev_change_rx_flags(dev, IFF_ALLMULTI);
3482
		dev_set_rx_mode(dev);
3483
	}
3484
	return 0;
3485
}
E
Eric Dumazet 已提交
3486
EXPORT_SYMBOL(dev_set_allmulti);
3487 3488 3489 3490

/*
 *	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
3491
 *	filtering it is put in promiscuous mode while unicast addresses
3492 3493 3494 3495
 *	are present.
 */
void __dev_set_rx_mode(struct net_device *dev)
{
3496 3497
	const struct net_device_ops *ops = dev->netdev_ops;

3498 3499 3500 3501 3502
	/* dev_open will call this function so the list will stay sane. */
	if (!(dev->flags&IFF_UP))
		return;

	if (!netif_device_present(dev))
3503
		return;
3504

3505 3506
	if (ops->ndo_set_rx_mode)
		ops->ndo_set_rx_mode(dev);
3507 3508 3509 3510
	else {
		/* Unicast addresses changes may only happen under the rtnl,
		 * therefore calling __dev_set_promiscuity here is safe.
		 */
3511
		if (dev->uc.count > 0 && !dev->uc_promisc) {
3512 3513
			__dev_set_promiscuity(dev, 1);
			dev->uc_promisc = 1;
3514
		} else if (dev->uc.count == 0 && dev->uc_promisc) {
3515 3516 3517 3518
			__dev_set_promiscuity(dev, -1);
			dev->uc_promisc = 0;
		}

3519 3520
		if (ops->ndo_set_multicast_list)
			ops->ndo_set_multicast_list(dev);
3521 3522 3523 3524 3525
	}
}

void dev_set_rx_mode(struct net_device *dev)
{
3526
	netif_addr_lock_bh(dev);
3527
	__dev_set_rx_mode(dev);
3528
	netif_addr_unlock_bh(dev);
L
Linus Torvalds 已提交
3529 3530
}

3531 3532
/* hw addresses list handling functions */

3533 3534
static int __hw_addr_add(struct netdev_hw_addr_list *list, unsigned char *addr,
			 int addr_len, unsigned char addr_type)
3535 3536 3537 3538 3539 3540 3541
{
	struct netdev_hw_addr *ha;
	int alloc_size;

	if (addr_len > MAX_ADDR_LEN)
		return -EINVAL;

3542
	list_for_each_entry(ha, &list->list, list) {
J
Jiri Pirko 已提交
3543 3544 3545 3546 3547 3548 3549 3550
		if (!memcmp(ha->addr, addr, addr_len) &&
		    ha->type == addr_type) {
			ha->refcount++;
			return 0;
		}
	}


3551 3552 3553 3554 3555 3556 3557 3558
	alloc_size = sizeof(*ha);
	if (alloc_size < L1_CACHE_BYTES)
		alloc_size = L1_CACHE_BYTES;
	ha = kmalloc(alloc_size, GFP_ATOMIC);
	if (!ha)
		return -ENOMEM;
	memcpy(ha->addr, addr, addr_len);
	ha->type = addr_type;
J
Jiri Pirko 已提交
3559 3560
	ha->refcount = 1;
	ha->synced = false;
3561 3562
	list_add_tail_rcu(&ha->list, &list->list);
	list->count++;
3563 3564 3565 3566 3567 3568 3569 3570 3571 3572 3573
	return 0;
}

static void ha_rcu_free(struct rcu_head *head)
{
	struct netdev_hw_addr *ha;

	ha = container_of(head, struct netdev_hw_addr, rcu_head);
	kfree(ha);
}

3574 3575
static int __hw_addr_del(struct netdev_hw_addr_list *list, unsigned char *addr,
			 int addr_len, unsigned char addr_type)
3576 3577 3578
{
	struct netdev_hw_addr *ha;

3579
	list_for_each_entry(ha, &list->list, list) {
J
Jiri Pirko 已提交
3580
		if (!memcmp(ha->addr, addr, addr_len) &&
3581
		    (ha->type == addr_type || !addr_type)) {
J
Jiri Pirko 已提交
3582 3583
			if (--ha->refcount)
				return 0;
3584 3585
			list_del_rcu(&ha->list);
			call_rcu(&ha->rcu_head, ha_rcu_free);
3586
			list->count--;
3587 3588 3589 3590 3591 3592
			return 0;
		}
	}
	return -ENOENT;
}

3593 3594 3595
static int __hw_addr_add_multiple(struct netdev_hw_addr_list *to_list,
				  struct netdev_hw_addr_list *from_list,
				  int addr_len,
J
Jiri Pirko 已提交
3596
				  unsigned char addr_type)
3597 3598 3599 3600 3601
{
	int err;
	struct netdev_hw_addr *ha, *ha2;
	unsigned char type;

3602
	list_for_each_entry(ha, &from_list->list, list) {
3603
		type = addr_type ? addr_type : ha->type;
3604
		err = __hw_addr_add(to_list, ha->addr, addr_len, type);
3605 3606 3607 3608 3609 3610
		if (err)
			goto unroll;
	}
	return 0;

unroll:
3611
	list_for_each_entry(ha2, &from_list->list, list) {
3612 3613 3614
		if (ha2 == ha)
			break;
		type = addr_type ? addr_type : ha2->type;
3615
		__hw_addr_del(to_list, ha2->addr, addr_len, type);
3616 3617 3618 3619
	}
	return err;
}

3620 3621 3622
static void __hw_addr_del_multiple(struct netdev_hw_addr_list *to_list,
				   struct netdev_hw_addr_list *from_list,
				   int addr_len,
J
Jiri Pirko 已提交
3623
				   unsigned char addr_type)
3624 3625 3626 3627
{
	struct netdev_hw_addr *ha;
	unsigned char type;

3628
	list_for_each_entry(ha, &from_list->list, list) {
3629
		type = addr_type ? addr_type : ha->type;
3630
		__hw_addr_del(to_list, ha->addr, addr_len, addr_type);
J
Jiri Pirko 已提交
3631 3632 3633
	}
}

3634 3635
static int __hw_addr_sync(struct netdev_hw_addr_list *to_list,
			  struct netdev_hw_addr_list *from_list,
J
Jiri Pirko 已提交
3636 3637 3638 3639 3640
			  int addr_len)
{
	int err = 0;
	struct netdev_hw_addr *ha, *tmp;

3641
	list_for_each_entry_safe(ha, tmp, &from_list->list, list) {
J
Jiri Pirko 已提交
3642
		if (!ha->synced) {
3643
			err = __hw_addr_add(to_list, ha->addr,
J
Jiri Pirko 已提交
3644 3645 3646 3647 3648 3649
					    addr_len, ha->type);
			if (err)
				break;
			ha->synced = true;
			ha->refcount++;
		} else if (ha->refcount == 1) {
3650 3651
			__hw_addr_del(to_list, ha->addr, addr_len, ha->type);
			__hw_addr_del(from_list, ha->addr, addr_len, ha->type);
J
Jiri Pirko 已提交
3652
		}
3653
	}
J
Jiri Pirko 已提交
3654
	return err;
3655 3656
}

3657 3658
static void __hw_addr_unsync(struct netdev_hw_addr_list *to_list,
			     struct netdev_hw_addr_list *from_list,
J
Jiri Pirko 已提交
3659 3660 3661 3662
			     int addr_len)
{
	struct netdev_hw_addr *ha, *tmp;

3663
	list_for_each_entry_safe(ha, tmp, &from_list->list, list) {
J
Jiri Pirko 已提交
3664
		if (ha->synced) {
3665
			__hw_addr_del(to_list, ha->addr,
J
Jiri Pirko 已提交
3666 3667
				      addr_len, ha->type);
			ha->synced = false;
3668
			__hw_addr_del(from_list, ha->addr,
J
Jiri Pirko 已提交
3669 3670 3671 3672 3673
				      addr_len, ha->type);
		}
	}
}

3674
static void __hw_addr_flush(struct netdev_hw_addr_list *list)
3675 3676 3677
{
	struct netdev_hw_addr *ha, *tmp;

3678
	list_for_each_entry_safe(ha, tmp, &list->list, list) {
3679 3680 3681
		list_del_rcu(&ha->list);
		call_rcu(&ha->rcu_head, ha_rcu_free);
	}
3682 3683 3684 3685 3686 3687 3688
	list->count = 0;
}

static void __hw_addr_init(struct netdev_hw_addr_list *list)
{
	INIT_LIST_HEAD(&list->list);
	list->count = 0;
3689 3690 3691 3692 3693 3694 3695 3696
}

/* Device addresses handling functions */

static void dev_addr_flush(struct net_device *dev)
{
	/* rtnl_mutex must be held here */

3697
	__hw_addr_flush(&dev->dev_addrs);
3698 3699 3700 3701 3702 3703 3704 3705 3706 3707 3708
	dev->dev_addr = NULL;
}

static int dev_addr_init(struct net_device *dev)
{
	unsigned char addr[MAX_ADDR_LEN];
	struct netdev_hw_addr *ha;
	int err;

	/* rtnl_mutex must be held here */

3709
	__hw_addr_init(&dev->dev_addrs);
E
Eric Dumazet 已提交
3710
	memset(addr, 0, sizeof(addr));
3711
	err = __hw_addr_add(&dev->dev_addrs, addr, sizeof(addr),
3712 3713 3714 3715 3716 3717
			    NETDEV_HW_ADDR_T_LAN);
	if (!err) {
		/*
		 * Get the first (previously created) address from the list
		 * and set dev_addr pointer to this location.
		 */
3718
		ha = list_first_entry(&dev->dev_addrs.list,
3719 3720 3721 3722 3723 3724 3725 3726 3727 3728 3729 3730 3731 3732 3733 3734 3735 3736 3737 3738 3739 3740 3741 3742
				      struct netdev_hw_addr, list);
		dev->dev_addr = ha->addr;
	}
	return err;
}

/**
 *	dev_addr_add	- Add a device address
 *	@dev: device
 *	@addr: address to add
 *	@addr_type: address type
 *
 *	Add a device address to the device or increase the reference count if
 *	it already exists.
 *
 *	The caller must hold the rtnl_mutex.
 */
int dev_addr_add(struct net_device *dev, unsigned char *addr,
		 unsigned char addr_type)
{
	int err;

	ASSERT_RTNL();

3743
	err = __hw_addr_add(&dev->dev_addrs, addr, dev->addr_len, addr_type);
3744 3745 3746 3747 3748 3749 3750 3751 3752 3753 3754 3755 3756 3757 3758 3759 3760 3761 3762 3763 3764
	if (!err)
		call_netdevice_notifiers(NETDEV_CHANGEADDR, dev);
	return err;
}
EXPORT_SYMBOL(dev_addr_add);

/**
 *	dev_addr_del	- Release a device address.
 *	@dev: device
 *	@addr: address to delete
 *	@addr_type: address type
 *
 *	Release reference to a device address and remove it from the device
 *	if the reference count drops to zero.
 *
 *	The caller must hold the rtnl_mutex.
 */
int dev_addr_del(struct net_device *dev, unsigned char *addr,
		 unsigned char addr_type)
{
	int err;
J
Jiri Pirko 已提交
3765
	struct netdev_hw_addr *ha;
3766 3767 3768

	ASSERT_RTNL();

J
Jiri Pirko 已提交
3769 3770 3771 3772
	/*
	 * We can not remove the first address from the list because
	 * dev->dev_addr points to that.
	 */
3773 3774
	ha = list_first_entry(&dev->dev_addrs.list,
			      struct netdev_hw_addr, list);
J
Jiri Pirko 已提交
3775 3776 3777
	if (ha->addr == dev->dev_addr && ha->refcount == 1)
		return -ENOENT;

3778
	err = __hw_addr_del(&dev->dev_addrs, addr, dev->addr_len,
J
Jiri Pirko 已提交
3779
			    addr_type);
3780 3781 3782 3783 3784 3785 3786 3787 3788 3789 3790 3791 3792 3793 3794 3795 3796 3797 3798 3799 3800 3801 3802 3803 3804 3805
	if (!err)
		call_netdevice_notifiers(NETDEV_CHANGEADDR, dev);
	return err;
}
EXPORT_SYMBOL(dev_addr_del);

/**
 *	dev_addr_add_multiple	- Add device addresses from another device
 *	@to_dev: device to which addresses will be added
 *	@from_dev: device from which addresses will be added
 *	@addr_type: address type - 0 means type will be used from from_dev
 *
 *	Add device addresses of the one device to another.
 **
 *	The caller must hold the rtnl_mutex.
 */
int dev_addr_add_multiple(struct net_device *to_dev,
			  struct net_device *from_dev,
			  unsigned char addr_type)
{
	int err;

	ASSERT_RTNL();

	if (from_dev->addr_len != to_dev->addr_len)
		return -EINVAL;
3806
	err = __hw_addr_add_multiple(&to_dev->dev_addrs, &from_dev->dev_addrs,
J
Jiri Pirko 已提交
3807
				     to_dev->addr_len, addr_type);
3808 3809 3810 3811 3812 3813 3814 3815 3816 3817 3818 3819 3820 3821 3822 3823 3824 3825 3826 3827 3828 3829 3830 3831
	if (!err)
		call_netdevice_notifiers(NETDEV_CHANGEADDR, to_dev);
	return err;
}
EXPORT_SYMBOL(dev_addr_add_multiple);

/**
 *	dev_addr_del_multiple	- Delete device addresses by another device
 *	@to_dev: device where the addresses will be deleted
 *	@from_dev: device by which addresses the addresses will be deleted
 *	@addr_type: address type - 0 means type will used from from_dev
 *
 *	Deletes addresses in to device by the list of addresses in from device.
 *
 *	The caller must hold the rtnl_mutex.
 */
int dev_addr_del_multiple(struct net_device *to_dev,
			  struct net_device *from_dev,
			  unsigned char addr_type)
{
	ASSERT_RTNL();

	if (from_dev->addr_len != to_dev->addr_len)
		return -EINVAL;
3832
	__hw_addr_del_multiple(&to_dev->dev_addrs, &from_dev->dev_addrs,
J
Jiri Pirko 已提交
3833
			       to_dev->addr_len, addr_type);
3834 3835 3836 3837 3838
	call_netdevice_notifiers(NETDEV_CHANGEADDR, to_dev);
	return 0;
}
EXPORT_SYMBOL(dev_addr_del_multiple);

3839
/* multicast addresses handling functions */
3840

3841 3842
int __dev_addr_delete(struct dev_addr_list **list, int *count,
		      void *addr, int alen, int glbl)
3843 3844 3845 3846 3847 3848 3849 3850 3851 3852 3853 3854 3855 3856 3857 3858 3859
{
	struct dev_addr_list *da;

	for (; (da = *list) != NULL; list = &da->next) {
		if (memcmp(da->da_addr, addr, da->da_addrlen) == 0 &&
		    alen == da->da_addrlen) {
			if (glbl) {
				int old_glbl = da->da_gusers;
				da->da_gusers = 0;
				if (old_glbl == 0)
					break;
			}
			if (--da->da_users)
				return 0;

			*list = da->next;
			kfree(da);
3860
			(*count)--;
3861 3862 3863 3864 3865 3866
			return 0;
		}
	}
	return -ENOENT;
}

3867 3868
int __dev_addr_add(struct dev_addr_list **list, int *count,
		   void *addr, int alen, int glbl)
3869 3870 3871 3872 3873 3874 3875 3876 3877 3878 3879 3880 3881 3882 3883 3884 3885
{
	struct dev_addr_list *da;

	for (da = *list; da != NULL; da = da->next) {
		if (memcmp(da->da_addr, addr, da->da_addrlen) == 0 &&
		    da->da_addrlen == alen) {
			if (glbl) {
				int old_glbl = da->da_gusers;
				da->da_gusers = 1;
				if (old_glbl)
					return 0;
			}
			da->da_users++;
			return 0;
		}
	}

3886
	da = kzalloc(sizeof(*da), GFP_ATOMIC);
3887 3888 3889 3890 3891 3892 3893 3894
	if (da == NULL)
		return -ENOMEM;
	memcpy(da->da_addr, addr, alen);
	da->da_addrlen = alen;
	da->da_users = 1;
	da->da_gusers = glbl ? 1 : 0;
	da->next = *list;
	*list = da;
3895
	(*count)++;
3896 3897 3898
	return 0;
}

3899 3900 3901
/**
 *	dev_unicast_delete	- Release secondary unicast address.
 *	@dev: device
R
Randy Dunlap 已提交
3902
 *	@addr: address to delete
3903 3904
 *
 *	Release reference to a secondary unicast address and remove it
R
Randy Dunlap 已提交
3905
 *	from the device if the reference count drops to zero.
3906 3907 3908
 *
 * 	The caller must hold the rtnl_mutex.
 */
J
Jiri Pirko 已提交
3909
int dev_unicast_delete(struct net_device *dev, void *addr)
3910 3911 3912 3913 3914
{
	int err;

	ASSERT_RTNL();

3915
	netif_addr_lock_bh(dev);
3916 3917
	err = __hw_addr_del(&dev->uc, addr, dev->addr_len,
			    NETDEV_HW_ADDR_T_UNICAST);
3918
	if (!err)
3919
		__dev_set_rx_mode(dev);
3920
	netif_addr_unlock_bh(dev);
3921 3922 3923 3924 3925 3926 3927
	return err;
}
EXPORT_SYMBOL(dev_unicast_delete);

/**
 *	dev_unicast_add		- add a secondary unicast address
 *	@dev: device
3928
 *	@addr: address to add
3929 3930 3931 3932 3933 3934
 *
 *	Add a secondary unicast address to the device or increase
 *	the reference count if it already exists.
 *
 *	The caller must hold the rtnl_mutex.
 */
J
Jiri Pirko 已提交
3935
int dev_unicast_add(struct net_device *dev, void *addr)
3936 3937 3938 3939 3940
{
	int err;

	ASSERT_RTNL();

3941
	netif_addr_lock_bh(dev);
3942 3943
	err = __hw_addr_add(&dev->uc, addr, dev->addr_len,
			    NETDEV_HW_ADDR_T_UNICAST);
3944
	if (!err)
3945
		__dev_set_rx_mode(dev);
3946
	netif_addr_unlock_bh(dev);
3947 3948 3949 3950
	return err;
}
EXPORT_SYMBOL(dev_unicast_add);

3951 3952 3953 3954 3955 3956 3957 3958 3959 3960 3961 3962 3963 3964 3965 3966 3967 3968 3969 3970 3971 3972 3973 3974 3975 3976
int __dev_addr_sync(struct dev_addr_list **to, int *to_count,
		    struct dev_addr_list **from, int *from_count)
{
	struct dev_addr_list *da, *next;
	int err = 0;

	da = *from;
	while (da != NULL) {
		next = da->next;
		if (!da->da_synced) {
			err = __dev_addr_add(to, to_count,
					     da->da_addr, da->da_addrlen, 0);
			if (err < 0)
				break;
			da->da_synced = 1;
			da->da_users++;
		} else if (da->da_users == 1) {
			__dev_addr_delete(to, to_count,
					  da->da_addr, da->da_addrlen, 0);
			__dev_addr_delete(from, from_count,
					  da->da_addr, da->da_addrlen, 0);
		}
		da = next;
	}
	return err;
}
3977
EXPORT_SYMBOL_GPL(__dev_addr_sync);
3978 3979 3980 3981 3982 3983 3984 3985 3986 3987 3988 3989 3990 3991 3992 3993 3994 3995 3996

void __dev_addr_unsync(struct dev_addr_list **to, int *to_count,
		       struct dev_addr_list **from, int *from_count)
{
	struct dev_addr_list *da, *next;

	da = *from;
	while (da != NULL) {
		next = da->next;
		if (da->da_synced) {
			__dev_addr_delete(to, to_count,
					  da->da_addr, da->da_addrlen, 0);
			da->da_synced = 0;
			__dev_addr_delete(from, from_count,
					  da->da_addr, da->da_addrlen, 0);
		}
		da = next;
	}
}
3997
EXPORT_SYMBOL_GPL(__dev_addr_unsync);
3998 3999 4000 4001 4002 4003 4004

/**
 *	dev_unicast_sync - Synchronize device's unicast list to another device
 *	@to: destination device
 *	@from: source device
 *
 *	Add newly added addresses to the destination device and release
4005 4006
 *	addresses that have no users left. The source device must be
 *	locked by netif_tx_lock_bh.
4007 4008 4009 4010 4011 4012 4013 4014
 *
 *	This function is intended to be called from the dev->set_rx_mode
 *	function of layered software devices.
 */
int dev_unicast_sync(struct net_device *to, struct net_device *from)
{
	int err = 0;

J
Jiri Pirko 已提交
4015 4016 4017
	if (to->addr_len != from->addr_len)
		return -EINVAL;

4018
	netif_addr_lock_bh(to);
4019
	err = __hw_addr_sync(&to->uc, &from->uc, to->addr_len);
4020 4021
	if (!err)
		__dev_set_rx_mode(to);
4022
	netif_addr_unlock_bh(to);
4023 4024 4025 4026 4027
	return err;
}
EXPORT_SYMBOL(dev_unicast_sync);

/**
4028
 *	dev_unicast_unsync - Remove synchronized addresses from the destination device
4029 4030 4031 4032 4033 4034 4035 4036 4037
 *	@to: destination device
 *	@from: source device
 *
 *	Remove all addresses that were added to the destination device by
 *	dev_unicast_sync(). This function is intended to be called from the
 *	dev->stop function of layered software devices.
 */
void dev_unicast_unsync(struct net_device *to, struct net_device *from)
{
J
Jiri Pirko 已提交
4038 4039
	if (to->addr_len != from->addr_len)
		return;
4040

4041 4042
	netif_addr_lock_bh(from);
	netif_addr_lock(to);
4043
	__hw_addr_unsync(&to->uc, &from->uc, to->addr_len);
J
Jiri Pirko 已提交
4044
	__dev_set_rx_mode(to);
4045 4046
	netif_addr_unlock(to);
	netif_addr_unlock_bh(from);
4047 4048 4049
}
EXPORT_SYMBOL(dev_unicast_unsync);

J
Jiri Pirko 已提交
4050 4051
static void dev_unicast_flush(struct net_device *dev)
{
4052
	netif_addr_lock_bh(dev);
4053
	__hw_addr_flush(&dev->uc);
4054
	netif_addr_unlock_bh(dev);
J
Jiri Pirko 已提交
4055 4056 4057 4058
}

static void dev_unicast_init(struct net_device *dev)
{
4059
	__hw_addr_init(&dev->uc);
J
Jiri Pirko 已提交
4060 4061 4062
}


4063 4064 4065 4066 4067 4068 4069 4070 4071 4072 4073 4074 4075 4076
static void __dev_addr_discard(struct dev_addr_list **list)
{
	struct dev_addr_list *tmp;

	while (*list != NULL) {
		tmp = *list;
		*list = tmp->next;
		if (tmp->da_users > tmp->da_gusers)
			printk("__dev_addr_discard: address leakage! "
			       "da_users=%d\n", tmp->da_users);
		kfree(tmp);
	}
}

4077
static void dev_addr_discard(struct net_device *dev)
4078
{
4079
	netif_addr_lock_bh(dev);
4080

4081 4082
	__dev_addr_discard(&dev->mc_list);
	dev->mc_count = 0;
4083

4084
	netif_addr_unlock_bh(dev);
4085 4086
}

4087 4088 4089 4090 4091 4092
/**
 *	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
4093 4094 4095 4096 4097 4098
unsigned dev_get_flags(const struct net_device *dev)
{
	unsigned flags;

	flags = (dev->flags & ~(IFF_PROMISC |
				IFF_ALLMULTI |
S
Stefan Rompf 已提交
4099 4100 4101
				IFF_RUNNING |
				IFF_LOWER_UP |
				IFF_DORMANT)) |
L
Linus Torvalds 已提交
4102 4103 4104
		(dev->gflags & (IFF_PROMISC |
				IFF_ALLMULTI));

S
Stefan Rompf 已提交
4105 4106 4107 4108 4109 4110 4111 4112
	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
4113 4114 4115

	return flags;
}
E
Eric Dumazet 已提交
4116
EXPORT_SYMBOL(dev_get_flags);
L
Linus Torvalds 已提交
4117

4118 4119 4120 4121 4122 4123 4124 4125
/**
 *	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.
 */
L
Linus Torvalds 已提交
4126 4127
int dev_change_flags(struct net_device *dev, unsigned flags)
{
4128
	int ret, changes;
L
Linus Torvalds 已提交
4129 4130
	int old_flags = dev->flags;

4131 4132
	ASSERT_RTNL();

L
Linus Torvalds 已提交
4133 4134 4135 4136 4137 4138 4139 4140 4141 4142 4143 4144 4145 4146
	/*
	 *	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.
	 */

4147 4148
	if ((old_flags ^ flags) & IFF_MULTICAST)
		dev_change_rx_flags(dev, IFF_MULTICAST);
4149

4150
	dev_set_rx_mode(dev);
L
Linus Torvalds 已提交
4151 4152 4153 4154 4155 4156 4157 4158 4159 4160 4161 4162

	/*
	 *	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  ? */
		ret = ((old_flags & IFF_UP) ? dev_close : dev_open)(dev);

		if (!ret)
4163
			dev_set_rx_mode(dev);
L
Linus Torvalds 已提交
4164 4165 4166
	}

	if (dev->flags & IFF_UP &&
E
Eric Dumazet 已提交
4167
	    ((old_flags ^ dev->flags) & ~(IFF_UP | IFF_PROMISC | IFF_ALLMULTI |
L
Linus Torvalds 已提交
4168
					  IFF_VOLATILE)))
4169
		call_netdevice_notifiers(NETDEV_CHANGE, dev);
L
Linus Torvalds 已提交
4170 4171

	if ((flags ^ dev->gflags) & IFF_PROMISC) {
E
Eric Dumazet 已提交
4172 4173
		int inc = (flags & IFF_PROMISC) ? 1 : -1;

L
Linus Torvalds 已提交
4174 4175 4176 4177 4178 4179 4180 4181 4182
		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
4183 4184
		int inc = (flags & IFF_ALLMULTI) ? 1 : -1;

L
Linus Torvalds 已提交
4185 4186 4187 4188
		dev->gflags ^= IFF_ALLMULTI;
		dev_set_allmulti(dev, inc);
	}

4189 4190 4191 4192
	/* Exclude state transition flags, already notified */
	changes = (old_flags ^ dev->flags) & ~(IFF_UP | IFF_RUNNING);
	if (changes)
		rtmsg_ifinfo(RTM_NEWLINK, dev, changes);
L
Linus Torvalds 已提交
4193 4194 4195

	return ret;
}
E
Eric Dumazet 已提交
4196
EXPORT_SYMBOL(dev_change_flags);
L
Linus Torvalds 已提交
4197

4198 4199 4200 4201 4202 4203 4204
/**
 *	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
4205 4206
int dev_set_mtu(struct net_device *dev, int new_mtu)
{
4207
	const struct net_device_ops *ops = dev->netdev_ops;
L
Linus Torvalds 已提交
4208 4209 4210 4211 4212 4213 4214 4215 4216 4217 4218 4219 4220
	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;
4221 4222
	if (ops->ndo_change_mtu)
		err = ops->ndo_change_mtu(dev, new_mtu);
L
Linus Torvalds 已提交
4223 4224
	else
		dev->mtu = new_mtu;
4225

L
Linus Torvalds 已提交
4226
	if (!err && dev->flags & IFF_UP)
4227
		call_netdevice_notifiers(NETDEV_CHANGEMTU, dev);
L
Linus Torvalds 已提交
4228 4229
	return err;
}
E
Eric Dumazet 已提交
4230
EXPORT_SYMBOL(dev_set_mtu);
L
Linus Torvalds 已提交
4231

4232 4233 4234 4235 4236 4237 4238
/**
 *	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
4239 4240
int dev_set_mac_address(struct net_device *dev, struct sockaddr *sa)
{
4241
	const struct net_device_ops *ops = dev->netdev_ops;
L
Linus Torvalds 已提交
4242 4243
	int err;

4244
	if (!ops->ndo_set_mac_address)
L
Linus Torvalds 已提交
4245 4246 4247 4248 4249
		return -EOPNOTSUPP;
	if (sa->sa_family != dev->type)
		return -EINVAL;
	if (!netif_device_present(dev))
		return -ENODEV;
4250
	err = ops->ndo_set_mac_address(dev, sa);
L
Linus Torvalds 已提交
4251
	if (!err)
4252
		call_netdevice_notifiers(NETDEV_CHANGEADDR, dev);
L
Linus Torvalds 已提交
4253 4254
	return err;
}
E
Eric Dumazet 已提交
4255
EXPORT_SYMBOL(dev_set_mac_address);
L
Linus Torvalds 已提交
4256 4257

/*
4258
 *	Perform the SIOCxIFxxx calls, inside read_lock(dev_base_lock)
L
Linus Torvalds 已提交
4259
 */
4260
static int dev_ifsioc_locked(struct net *net, struct ifreq *ifr, unsigned int cmd)
L
Linus Torvalds 已提交
4261 4262
{
	int err;
4263
	struct net_device *dev = __dev_get_by_name(net, ifr->ifr_name);
L
Linus Torvalds 已提交
4264 4265 4266 4267 4268

	if (!dev)
		return -ENODEV;

	switch (cmd) {
E
Eric Dumazet 已提交
4269 4270 4271
	case SIOCGIFFLAGS:	/* Get interface flags */
		ifr->ifr_flags = (short) dev_get_flags(dev);
		return 0;
L
Linus Torvalds 已提交
4272

E
Eric Dumazet 已提交
4273 4274 4275 4276
	case SIOCGIFMETRIC:	/* Get the metric on the interface
				   (currently unused) */
		ifr->ifr_metric = 0;
		return 0;
L
Linus Torvalds 已提交
4277

E
Eric Dumazet 已提交
4278 4279 4280
	case SIOCGIFMTU:	/* Get the MTU of a device */
		ifr->ifr_mtu = dev->mtu;
		return 0;
L
Linus Torvalds 已提交
4281

E
Eric Dumazet 已提交
4282 4283 4284 4285 4286 4287 4288 4289
	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
4290

E
Eric Dumazet 已提交
4291 4292 4293
	case SIOCGIFSLAVE:
		err = -EINVAL;
		break;
4294

E
Eric Dumazet 已提交
4295 4296 4297 4298 4299 4300 4301 4302
	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;
4303

E
Eric Dumazet 已提交
4304 4305 4306
	case SIOCGIFINDEX:
		ifr->ifr_ifindex = dev->ifindex;
		return 0;
4307

E
Eric Dumazet 已提交
4308 4309 4310
	case SIOCGIFTXQLEN:
		ifr->ifr_qlen = dev->tx_queue_len;
		return 0;
4311

E
Eric Dumazet 已提交
4312 4313 4314 4315 4316 4317 4318
	default:
		/* dev_ioctl() should ensure this case
		 * is never reached
		 */
		WARN_ON(1);
		err = -EINVAL;
		break;
4319 4320 4321 4322 4323 4324 4325 4326 4327 4328 4329 4330

	}
	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
4331
	const struct net_device_ops *ops;
4332 4333 4334 4335

	if (!dev)
		return -ENODEV;

J
Jarek Poplawski 已提交
4336 4337
	ops = dev->netdev_ops;

4338
	switch (cmd) {
E
Eric Dumazet 已提交
4339 4340
	case SIOCSIFFLAGS:	/* Set interface flags */
		return dev_change_flags(dev, ifr->ifr_flags);
4341

E
Eric Dumazet 已提交
4342 4343 4344
	case SIOCSIFMETRIC:	/* Set the metric on the interface
				   (currently unused) */
		return -EOPNOTSUPP;
4345

E
Eric Dumazet 已提交
4346 4347
	case SIOCSIFMTU:	/* Set the MTU of a device */
		return dev_set_mtu(dev, ifr->ifr_mtu);
L
Linus Torvalds 已提交
4348

E
Eric Dumazet 已提交
4349 4350
	case SIOCSIFHWADDR:
		return dev_set_mac_address(dev, &ifr->ifr_hwaddr);
L
Linus Torvalds 已提交
4351

E
Eric Dumazet 已提交
4352 4353 4354 4355 4356 4357 4358
	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
4359

E
Eric Dumazet 已提交
4360 4361
	case SIOCSIFMAP:
		if (ops->ndo_set_config) {
L
Linus Torvalds 已提交
4362 4363
			if (!netif_device_present(dev))
				return -ENODEV;
E
Eric Dumazet 已提交
4364 4365 4366
			return ops->ndo_set_config(dev, &ifr->ifr_map);
		}
		return -EOPNOTSUPP;
L
Linus Torvalds 已提交
4367

E
Eric Dumazet 已提交
4368 4369 4370 4371 4372 4373 4374 4375 4376 4377 4378 4379 4380 4381 4382 4383 4384
	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;
		return dev_mc_add(dev, ifr->ifr_hwaddr.sa_data,
				  dev->addr_len, 1);

	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;
		return dev_mc_delete(dev, ifr->ifr_hwaddr.sa_data,
				     dev->addr_len, 1);
L
Linus Torvalds 已提交
4385

E
Eric Dumazet 已提交
4386 4387 4388 4389 4390
	case SIOCSIFTXQLEN:
		if (ifr->ifr_qlen < 0)
			return -EINVAL;
		dev->tx_queue_len = ifr->ifr_qlen;
		return 0;
L
Linus Torvalds 已提交
4391

E
Eric Dumazet 已提交
4392 4393 4394
	case SIOCSIFNAME:
		ifr->ifr_newname[IFNAMSIZ-1] = '\0';
		return dev_change_name(dev, ifr->ifr_newname);
L
Linus Torvalds 已提交
4395

E
Eric Dumazet 已提交
4396 4397 4398 4399 4400 4401 4402 4403 4404 4405 4406 4407 4408 4409 4410 4411 4412 4413 4414 4415 4416 4417 4418 4419 4420 4421 4422 4423
	/*
	 *	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
4424 4425 4426 4427 4428 4429 4430 4431 4432 4433 4434 4435

	}
	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
4436
 *	@net: the applicable net namespace
L
Linus Torvalds 已提交
4437 4438 4439 4440 4441 4442 4443 4444 4445
 *	@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.
 */

4446
int dev_ioctl(struct net *net, unsigned int cmd, void __user *arg)
L
Linus Torvalds 已提交
4447 4448 4449 4450 4451 4452 4453 4454 4455 4456 4457
{
	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) {
4458
		rtnl_lock();
4459
		ret = dev_ifconf(net, (char __user *) arg);
4460
		rtnl_unlock();
L
Linus Torvalds 已提交
4461 4462 4463
		return ret;
	}
	if (cmd == SIOCGIFNAME)
4464
		return dev_ifname(net, (struct ifreq __user *)arg);
L
Linus Torvalds 已提交
4465 4466 4467 4468 4469 4470 4471 4472 4473 4474 4475 4476 4477 4478 4479

	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
4480 4481 4482 4483 4484 4485 4486 4487 4488 4489 4490 4491 4492 4493 4494 4495 4496 4497 4498 4499 4500 4501 4502 4503 4504 4505
	/*
	 *	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);
		read_lock(&dev_base_lock);
		ret = dev_ifsioc_locked(net, &ifr, cmd);
		read_unlock(&dev_base_lock);
		if (!ret) {
			if (colon)
				*colon = ':';
			if (copy_to_user(arg, &ifr,
					 sizeof(struct ifreq)))
				ret = -EFAULT;
		}
		return ret;
L
Linus Torvalds 已提交
4506

E
Eric Dumazet 已提交
4507 4508 4509 4510 4511 4512 4513 4514 4515 4516 4517 4518 4519
	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
4520

E
Eric Dumazet 已提交
4521 4522 4523 4524 4525 4526 4527 4528 4529 4530 4531 4532 4533 4534 4535 4536 4537 4538 4539 4540 4541 4542 4543
	/*
	 *	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
4544

E
Eric Dumazet 已提交
4545 4546 4547 4548 4549 4550 4551 4552 4553 4554 4555 4556 4557 4558 4559 4560 4561 4562 4563 4564 4565 4566 4567 4568 4569 4570 4571 4572 4573 4574 4575 4576 4577 4578 4579 4580 4581 4582 4583 4584 4585 4586 4587 4588 4589 4590 4591 4592 4593 4594 4595
	/*
	 *	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)) {
4596
			dev_load(net, ifr.ifr_name);
L
Linus Torvalds 已提交
4597
			rtnl_lock();
4598
			ret = dev_ifsioc(net, &ifr, cmd);
L
Linus Torvalds 已提交
4599
			rtnl_unlock();
E
Eric Dumazet 已提交
4600 4601 4602
			if (!ret && copy_to_user(arg, &ifr,
						 sizeof(struct ifreq)))
				ret = -EFAULT;
L
Linus Torvalds 已提交
4603
			return ret;
E
Eric Dumazet 已提交
4604 4605 4606 4607 4608
		}
		/* 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
4609 4610 4611 4612 4613 4614
	}
}


/**
 *	dev_new_index	-	allocate an ifindex
4615
 *	@net: the applicable net namespace
L
Linus Torvalds 已提交
4616 4617 4618 4619 4620
 *
 *	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.
 */
4621
static int dev_new_index(struct net *net)
L
Linus Torvalds 已提交
4622 4623 4624 4625 4626
{
	static int ifindex;
	for (;;) {
		if (++ifindex <= 0)
			ifindex = 1;
4627
		if (!__dev_get_by_index(net, ifindex))
L
Linus Torvalds 已提交
4628 4629 4630 4631 4632
			return ifindex;
	}
}

/* Delayed registration/unregisteration */
4633
static LIST_HEAD(net_todo_list);
L
Linus Torvalds 已提交
4634

4635
static void net_set_todo(struct net_device *dev)
L
Linus Torvalds 已提交
4636 4637 4638 4639
{
	list_add_tail(&dev->todo_list, &net_todo_list);
}

4640
static void rollback_registered_many(struct list_head *head)
4641
{
4642 4643
	struct net_device *dev;

4644 4645 4646
	BUG_ON(dev_boot_phase);
	ASSERT_RTNL();

4647 4648 4649 4650 4651 4652 4653
	list_for_each_entry(dev, head, unreg_list) {
		/* Some devices call without registering
		 * for initialization unwind.
		 */
		if (dev->reg_state == NETREG_UNINITIALIZED) {
			pr_debug("unregister_netdevice: device %s/%p never "
				 "was registered\n", dev->name, dev);
4654

4655 4656 4657
			WARN_ON(1);
			return;
		}
4658

4659
		BUG_ON(dev->reg_state != NETREG_REGISTERED);
4660

4661 4662
		/* If device is running, close it first. */
		dev_close(dev);
4663

4664 4665
		/* And unlink it from device chain. */
		unlist_netdevice(dev);
4666

4667 4668
		dev->reg_state = NETREG_UNREGISTERING;
	}
4669 4670 4671

	synchronize_net();

4672 4673 4674
	list_for_each_entry(dev, head, unreg_list) {
		/* Shutdown queueing discipline. */
		dev_shutdown(dev);
4675 4676


4677 4678 4679 4680
		/* Notify protocols, that we are about to destroy
		   this device. They should clean all the things.
		*/
		call_netdevice_notifiers(NETDEV_UNREGISTER, dev);
4681

4682 4683 4684 4685 4686
		/*
		 *	Flush the unicast and multicast chains
		 */
		dev_unicast_flush(dev);
		dev_addr_discard(dev);
4687

4688 4689
		if (dev->netdev_ops->ndo_uninit)
			dev->netdev_ops->ndo_uninit(dev);
4690

4691 4692
		/* Notifier chain MUST detach us from master device. */
		WARN_ON(dev->master);
4693

4694 4695 4696
		/* Remove entries from kobject tree */
		netdev_unregister_kobject(dev);
	}
4697 4698 4699

	synchronize_net();

4700 4701 4702 4703 4704 4705 4706 4707 4708 4709
	list_for_each_entry(dev, head, unreg_list)
		dev_put(dev);
}

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

	list_add(&dev->unreg_list, &single);
	rollback_registered_many(&single);
4710 4711
}

4712 4713 4714
static void __netdev_init_queue_locks_one(struct net_device *dev,
					  struct netdev_queue *dev_queue,
					  void *_unused)
4715 4716
{
	spin_lock_init(&dev_queue->_xmit_lock);
4717
	netdev_set_xmit_lockdep_class(&dev_queue->_xmit_lock, dev->type);
4718 4719 4720 4721 4722
	dev_queue->xmit_lock_owner = -1;
}

static void netdev_init_queue_locks(struct net_device *dev)
{
4723 4724
	netdev_for_each_tx_queue(dev, __netdev_init_queue_locks_one, NULL);
	__netdev_init_queue_locks_one(dev, &dev->rx_queue, NULL);
4725 4726
}

4727 4728 4729 4730 4731 4732 4733 4734 4735 4736 4737 4738 4739 4740 4741 4742 4743 4744 4745 4746 4747 4748 4749 4750 4751 4752 4753 4754 4755 4756 4757 4758 4759 4760 4761 4762 4763 4764 4765 4766
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);

L
Linus Torvalds 已提交
4767 4768 4769 4770 4771 4772 4773 4774 4775 4776 4777 4778 4779 4780 4781 4782 4783 4784 4785 4786 4787 4788
/**
 *	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)
{
	struct hlist_head *head;
	struct hlist_node *p;
	int ret;
4789
	struct net *net = dev_net(dev);
L
Linus Torvalds 已提交
4790 4791 4792 4793

	BUG_ON(dev_boot_phase);
	ASSERT_RTNL();

4794 4795
	might_sleep();

L
Linus Torvalds 已提交
4796 4797
	/* When net_device's are persistent, this will be fatal. */
	BUG_ON(dev->reg_state != NETREG_UNINITIALIZED);
4798
	BUG_ON(!net);
L
Linus Torvalds 已提交
4799

4800
	spin_lock_init(&dev->addr_list_lock);
4801
	netdev_set_addr_lockdep_class(dev);
4802
	netdev_init_queue_locks(dev);
L
Linus Torvalds 已提交
4803 4804 4805 4806

	dev->iflink = -1;

	/* Init, if this function is available */
4807 4808
	if (dev->netdev_ops->ndo_init) {
		ret = dev->netdev_ops->ndo_init(dev);
L
Linus Torvalds 已提交
4809 4810 4811
		if (ret) {
			if (ret > 0)
				ret = -EIO;
4812
			goto out;
L
Linus Torvalds 已提交
4813 4814
		}
	}
4815

L
Linus Torvalds 已提交
4816 4817
	if (!dev_valid_name(dev->name)) {
		ret = -EINVAL;
4818
		goto err_uninit;
L
Linus Torvalds 已提交
4819 4820
	}

4821
	dev->ifindex = dev_new_index(net);
L
Linus Torvalds 已提交
4822 4823 4824 4825
	if (dev->iflink == -1)
		dev->iflink = dev->ifindex;

	/* Check for existence of name */
4826
	head = dev_name_hash(net, dev->name);
L
Linus Torvalds 已提交
4827 4828 4829 4830 4831
	hlist_for_each(p, head) {
		struct net_device *d
			= hlist_entry(p, struct net_device, name_hlist);
		if (!strncmp(d->name, dev->name, IFNAMSIZ)) {
			ret = -EEXIST;
4832
			goto err_uninit;
L
Linus Torvalds 已提交
4833
		}
4834
	}
L
Linus Torvalds 已提交
4835

4836 4837 4838 4839 4840 4841 4842 4843 4844 4845 4846 4847 4848 4849 4850
	/* 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);
	}

4851
	dev->features = netdev_fix_features(dev->features, dev->name);
L
Linus Torvalds 已提交
4852

4853 4854 4855 4856
	/* Enable software GSO if SG is supported. */
	if (dev->features & NETIF_F_SG)
		dev->features |= NETIF_F_GSO;

4857
	netdev_initialize_kobject(dev);
4858 4859 4860 4861 4862 4863

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

4864
	ret = netdev_register_kobject(dev);
4865
	if (ret)
4866
		goto err_uninit;
4867 4868
	dev->reg_state = NETREG_REGISTERED;

L
Linus Torvalds 已提交
4869 4870 4871 4872 4873 4874 4875 4876 4877
	/*
	 *	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);
4878
	list_netdevice(dev);
L
Linus Torvalds 已提交
4879 4880

	/* Notify protocols, that a new device appeared. */
4881
	ret = call_netdevice_notifiers(NETDEV_REGISTER, dev);
4882
	ret = notifier_to_errno(ret);
4883 4884 4885 4886
	if (ret) {
		rollback_registered(dev);
		dev->reg_state = NETREG_UNREGISTERED;
	}
L
Linus Torvalds 已提交
4887 4888 4889

out:
	return ret;
4890 4891

err_uninit:
4892 4893
	if (dev->netdev_ops->ndo_uninit)
		dev->netdev_ops->ndo_uninit(dev);
4894
	goto out;
L
Linus Torvalds 已提交
4895
}
E
Eric Dumazet 已提交
4896
EXPORT_SYMBOL(register_netdevice);
L
Linus Torvalds 已提交
4897

4898 4899 4900 4901 4902 4903 4904 4905 4906 4907 4908 4909 4910 4911 4912 4913 4914 4915 4916 4917 4918 4919 4920 4921 4922 4923 4924 4925 4926 4927 4928 4929 4930 4931 4932 4933 4934 4935 4936
/**
 *	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
4937 4938 4939 4940 4941 4942 4943 4944 4945
/**
 *	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.
 *
4946
 *	This is a wrapper around register_netdevice that takes the rtnl semaphore
L
Linus Torvalds 已提交
4947 4948 4949 4950 4951 4952 4953 4954 4955 4956 4957 4958 4959 4960 4961 4962 4963 4964
 *	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;
	}
4965

L
Linus Torvalds 已提交
4966 4967 4968 4969 4970 4971 4972 4973 4974 4975 4976 4977 4978 4979 4980 4981
	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
4982
 * call dev_put.
L
Linus Torvalds 已提交
4983 4984 4985 4986 4987 4988 4989 4990
 */
static void netdev_wait_allrefs(struct net_device *dev)
{
	unsigned long rebroadcast_time, warning_time;

	rebroadcast_time = warning_time = jiffies;
	while (atomic_read(&dev->refcnt) != 0) {
		if (time_after(jiffies, rebroadcast_time + 1 * HZ)) {
4991
			rtnl_lock();
L
Linus Torvalds 已提交
4992 4993

			/* Rebroadcast unregister notification */
4994
			call_netdevice_notifiers(NETDEV_UNREGISTER, dev);
L
Linus Torvalds 已提交
4995 4996 4997 4998 4999 5000 5001 5002 5003 5004 5005 5006

			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();
			}

5007
			__rtnl_unlock();
L
Linus Torvalds 已提交
5008 5009 5010 5011 5012 5013 5014 5015 5016 5017 5018 5019 5020 5021 5022 5023 5024 5025 5026 5027 5028 5029 5030 5031 5032 5033 5034 5035 5036 5037

			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
5038
 * We are invoked by rtnl_unlock().
L
Linus Torvalds 已提交
5039
 * This allows us to deal with problems:
5040
 * 1) We can delete sysfs objects which invoke hotplug
L
Linus Torvalds 已提交
5041 5042 5043
 *    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
5044 5045 5046
 *
 * 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
5047 5048 5049
 */
void netdev_run_todo(void)
{
5050
	struct list_head list;
L
Linus Torvalds 已提交
5051 5052

	/* Snapshot list, allow later requests */
5053
	list_replace_init(&net_todo_list, &list);
H
Herbert Xu 已提交
5054 5055

	__rtnl_unlock();
5056

L
Linus Torvalds 已提交
5057 5058 5059 5060 5061
	while (!list_empty(&list)) {
		struct net_device *dev
			= list_entry(list.next, struct net_device, todo_list);
		list_del(&dev->todo_list);

5062 5063 5064 5065 5066 5067
		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
5068

5069
		dev->reg_state = NETREG_UNREGISTERED;
L
Linus Torvalds 已提交
5070

5071 5072
		on_each_cpu(flush_backlog, dev, 1);

5073
		netdev_wait_allrefs(dev);
L
Linus Torvalds 已提交
5074

5075 5076
		/* paranoia */
		BUG_ON(atomic_read(&dev->refcnt));
5077 5078 5079
		WARN_ON(dev->ip_ptr);
		WARN_ON(dev->ip6_ptr);
		WARN_ON(dev->dn_ptr);
L
Linus Torvalds 已提交
5080

5081 5082
		if (dev->destructor)
			dev->destructor(dev);
5083 5084 5085

		/* Free network device */
		kobject_put(&dev->dev.kobj);
L
Linus Torvalds 已提交
5086 5087 5088
	}
}

5089 5090 5091 5092 5093 5094 5095 5096 5097
/**
 *	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)
5098
{
5099 5100 5101 5102
	const struct net_device_ops *ops = dev->netdev_ops;

	if (ops->ndo_get_stats)
		return ops->ndo_get_stats(dev);
5103 5104 5105 5106 5107 5108 5109 5110 5111 5112 5113 5114 5115 5116 5117 5118 5119 5120 5121
	else {
		unsigned long tx_bytes = 0, tx_packets = 0, tx_dropped = 0;
		struct net_device_stats *stats = &dev->stats;
		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;
		}
		return stats;
	}
R
Rusty Russell 已提交
5122
}
5123
EXPORT_SYMBOL(dev_get_stats);
R
Rusty Russell 已提交
5124

5125
static void netdev_init_one_queue(struct net_device *dev,
5126 5127
				  struct netdev_queue *queue,
				  void *_unused)
5128 5129 5130 5131
{
	queue->dev = dev;
}

5132 5133
static void netdev_init_queues(struct net_device *dev)
{
5134 5135
	netdev_init_one_queue(dev, &dev->rx_queue, NULL);
	netdev_for_each_tx_queue(dev, netdev_init_one_queue, NULL);
5136
	spin_lock_init(&dev->tx_global_lock);
5137 5138
}

L
Linus Torvalds 已提交
5139
/**
5140
 *	alloc_netdev_mq - allocate network device
L
Linus Torvalds 已提交
5141 5142 5143
 *	@sizeof_priv:	size of private data to allocate space for
 *	@name:		device name format string
 *	@setup:		callback to initialize device
5144
 *	@queue_count:	the number of subqueues to allocate
L
Linus Torvalds 已提交
5145 5146
 *
 *	Allocates a struct net_device with private data area for driver use
5147 5148
 *	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
5149
 */
5150 5151
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
5152
{
5153
	struct netdev_queue *tx;
L
Linus Torvalds 已提交
5154
	struct net_device *dev;
5155
	size_t alloc_size;
5156
	struct net_device *p;
L
Linus Torvalds 已提交
5157

5158 5159
	BUG_ON(strlen(name) >= sizeof(dev->name));

5160
	alloc_size = sizeof(struct net_device);
5161 5162
	if (sizeof_priv) {
		/* ensure 32-byte alignment of private area */
5163
		alloc_size = ALIGN(alloc_size, NETDEV_ALIGN);
5164 5165 5166
		alloc_size += sizeof_priv;
	}
	/* ensure 32-byte alignment of whole construct */
5167
	alloc_size += NETDEV_ALIGN - 1;
L
Linus Torvalds 已提交
5168

5169
	p = kzalloc(alloc_size, GFP_KERNEL);
L
Linus Torvalds 已提交
5170
	if (!p) {
5171
		printk(KERN_ERR "alloc_netdev: Unable to allocate device.\n");
L
Linus Torvalds 已提交
5172 5173 5174
		return NULL;
	}

5175
	tx = kcalloc(queue_count, sizeof(struct netdev_queue), GFP_KERNEL);
5176 5177 5178
	if (!tx) {
		printk(KERN_ERR "alloc_netdev: Unable to allocate "
		       "tx qdiscs.\n");
5179
		goto free_p;
5180 5181
	}

5182
	dev = PTR_ALIGN(p, NETDEV_ALIGN);
L
Linus Torvalds 已提交
5183
	dev->padded = (char *)dev - (char *)p;
5184 5185 5186 5187

	if (dev_addr_init(dev))
		goto free_tx;

J
Jiri Pirko 已提交
5188 5189
	dev_unicast_init(dev);

5190
	dev_net_set(dev, &init_net);
L
Linus Torvalds 已提交
5191

5192 5193
	dev->_tx = tx;
	dev->num_tx_queues = queue_count;
5194
	dev->real_num_tx_queues = queue_count;
5195

5196
	dev->gso_max_size = GSO_MAX_SIZE;
L
Linus Torvalds 已提交
5197

5198 5199
	netdev_init_queues(dev);

5200
	INIT_LIST_HEAD(&dev->napi_list);
5201
	dev->priv_flags = IFF_XMIT_DST_RELEASE;
L
Linus Torvalds 已提交
5202 5203 5204
	setup(dev);
	strcpy(dev->name, name);
	return dev;
5205 5206 5207 5208 5209 5210 5211

free_tx:
	kfree(tx);

free_p:
	kfree(p);
	return NULL;
L
Linus Torvalds 已提交
5212
}
5213
EXPORT_SYMBOL(alloc_netdev_mq);
L
Linus Torvalds 已提交
5214 5215 5216 5217 5218

/**
 *	free_netdev - free network device
 *	@dev: device
 *
5219 5220
 *	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
5221 5222 5223 5224
 *	If this is the last reference then it will be freed.
 */
void free_netdev(struct net_device *dev)
{
5225 5226
	struct napi_struct *p, *n;

5227 5228
	release_net(dev_net(dev));

5229 5230
	kfree(dev->_tx);

5231 5232 5233
	/* Flush device addresses */
	dev_addr_flush(dev);

5234 5235 5236
	list_for_each_entry_safe(p, n, &dev->napi_list, dev_list)
		netif_napi_del(p);

S
Stephen Hemminger 已提交
5237
	/*  Compatibility with error handling in drivers */
L
Linus Torvalds 已提交
5238 5239 5240 5241 5242 5243 5244 5245
	if (dev->reg_state == NETREG_UNINITIALIZED) {
		kfree((char *)dev - dev->padded);
		return;
	}

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

5246 5247
	/* will free via device release */
	put_device(&dev->dev);
L
Linus Torvalds 已提交
5248
}
E
Eric Dumazet 已提交
5249
EXPORT_SYMBOL(free_netdev);
5250

5251 5252 5253 5254 5255 5256
/**
 *	synchronize_net -  Synchronize with packet receive processing
 *
 *	Wait for packets currently being received to be done.
 *	Does not block later packets from starting.
 */
5257
void synchronize_net(void)
L
Linus Torvalds 已提交
5258 5259
{
	might_sleep();
5260
	synchronize_rcu();
L
Linus Torvalds 已提交
5261
}
E
Eric Dumazet 已提交
5262
EXPORT_SYMBOL(synchronize_net);
L
Linus Torvalds 已提交
5263 5264

/**
5265
 *	unregister_netdevice_queue - remove device from the kernel
L
Linus Torvalds 已提交
5266
 *	@dev: device
5267 5268
 *	@head: list

L
Linus Torvalds 已提交
5269
 *	This function shuts down a device interface and removes it
5270
 *	from the kernel tables.
5271
 *	If head not NULL, device is queued to be unregistered later.
L
Linus Torvalds 已提交
5272 5273 5274 5275 5276
 *
 *	Callers must hold the rtnl semaphore.  You may want
 *	unregister_netdev() instead of this.
 */

5277
void unregister_netdevice_queue(struct net_device *dev, struct list_head *head)
L
Linus Torvalds 已提交
5278
{
5279 5280
	ASSERT_RTNL();

5281 5282 5283 5284 5285 5286 5287
	if (head) {
		list_add_tail(&dev->unreg_list, head);
	} else {
		rollback_registered(dev);
		/* Finish processing unregister after unlock */
		net_set_todo(dev);
	}
L
Linus Torvalds 已提交
5288
}
5289
EXPORT_SYMBOL(unregister_netdevice_queue);
L
Linus Torvalds 已提交
5290

5291 5292 5293 5294 5295 5296 5297 5298 5299 5300 5301 5302 5303 5304 5305
/**
 *	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);
	}
}
5306
EXPORT_SYMBOL(unregister_netdevice_many);
5307

L
Linus Torvalds 已提交
5308 5309 5310 5311 5312
/**
 *	unregister_netdev - remove device from the kernel
 *	@dev: device
 *
 *	This function shuts down a device interface and removes it
5313
 *	from the kernel tables.
L
Linus Torvalds 已提交
5314 5315 5316 5317 5318 5319 5320 5321 5322 5323 5324 5325 5326
 *
 *	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);

5327 5328 5329 5330 5331 5332 5333 5334 5335 5336 5337 5338 5339 5340 5341 5342 5343 5344 5345 5346 5347 5348 5349 5350 5351 5352 5353
/**
 *	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)
{
	char buf[IFNAMSIZ];
	const char *destname;
	int err;

	ASSERT_RTNL();

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

5354 5355 5356 5357 5358 5359 5360 5361 5362
#ifdef CONFIG_SYSFS
	/* Don't allow real devices to be moved when sysfs
	 * is enabled.
	 */
	err = -EINVAL;
	if (dev->dev.parent)
		goto out;
#endif

5363 5364 5365 5366 5367 5368 5369
	/* 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;
5370
	if (net_eq(dev_net(dev), net))
5371 5372 5373 5374 5375 5376 5377 5378 5379 5380 5381 5382 5383 5384 5385 5386 5387 5388 5389 5390 5391 5392 5393 5394 5395 5396 5397 5398
		goto out;

	/* Pick the destination device name, and ensure
	 * we can use it in the destination network namespace.
	 */
	err = -EEXIST;
	destname = dev->name;
	if (__dev_get_by_name(net, destname)) {
		/* We get here if we can't use the current device name */
		if (!pat)
			goto out;
		if (!dev_valid_name(pat))
			goto out;
		if (strchr(pat, '%')) {
			if (__dev_alloc_name(net, pat, buf) < 0)
				goto out;
			destname = buf;
		} else
			destname = pat;
		if (__dev_get_by_name(net, destname))
			goto out;
	}

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

	/* If device is running close it first. */
5399
	dev_close(dev);
5400 5401 5402 5403 5404 5405 5406 5407 5408 5409 5410 5411 5412 5413 5414 5415 5416 5417

	/* 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);

	/*
	 *	Flush the unicast and multicast chains
	 */
J
Jiri Pirko 已提交
5418
	dev_unicast_flush(dev);
5419 5420
	dev_addr_discard(dev);

5421 5422
	netdev_unregister_kobject(dev);

5423
	/* Actually switch the network namespace */
5424
	dev_net_set(dev, net);
5425 5426 5427 5428 5429 5430 5431 5432 5433 5434 5435 5436 5437

	/* Assign the new device name */
	if (destname != dev->name)
		strcpy(dev->name, destname);

	/* 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;
	}

5438
	/* Fixup kobjects */
5439
	err = netdev_register_kobject(dev);
5440
	WARN_ON(err);
5441 5442 5443 5444 5445 5446 5447 5448 5449 5450 5451 5452

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

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

	synchronize_net();
	err = 0;
out:
	return err;
}
5453
EXPORT_SYMBOL_GPL(dev_change_net_namespace);
5454

L
Linus Torvalds 已提交
5455 5456 5457 5458 5459
static int dev_cpu_callback(struct notifier_block *nfb,
			    unsigned long action,
			    void *ocpu)
{
	struct sk_buff **list_skb;
5460
	struct Qdisc **list_net;
L
Linus Torvalds 已提交
5461 5462 5463 5464
	struct sk_buff *skb;
	unsigned int cpu, oldcpu = (unsigned long)ocpu;
	struct softnet_data *sd, *oldsd;

5465
	if (action != CPU_DEAD && action != CPU_DEAD_FROZEN)
L
Linus Torvalds 已提交
5466 5467 5468 5469 5470 5471 5472 5473 5474 5475 5476 5477 5478 5479 5480 5481 5482 5483 5484 5485 5486 5487 5488 5489 5490 5491 5492 5493 5494 5495 5496 5497 5498 5499
		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 */
	while ((skb = __skb_dequeue(&oldsd->input_pkt_queue)))
		netif_rx(skb);

	return NOTIFY_OK;
}


5500
/**
5501 5502 5503 5504
 *	netdev_increment_features - increment feature set by one
 *	@all: current feature set
 *	@one: new feature set
 *	@mask: mask feature set
5505 5506
 *
 *	Computes a new feature set after adding a device with feature set
5507 5508
 *	@one to the master device with current feature set @all.  Will not
 *	enable anything that is off in @mask. Returns the new feature set.
5509
 */
5510 5511 5512 5513
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
5514
	if (all & NETIF_F_NO_CSUM && !(one & NETIF_F_NO_CSUM))
5515 5516 5517 5518 5519 5520 5521 5522
		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);
		}
5523

5524 5525 5526 5527 5528 5529
		/* 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;
		}
	}
5530

5531
	one |= NETIF_F_ALL_CSUM;
5532

5533
	one |= all & NETIF_F_ONE_FOR_ALL;
5534
	all &= one | NETIF_F_LLTX | NETIF_F_GSO | NETIF_F_UFO;
5535
	all |= one & mask & NETIF_F_ONE_FOR_ALL;
5536 5537 5538

	return all;
}
5539
EXPORT_SYMBOL(netdev_increment_features);
5540

5541 5542 5543 5544 5545 5546 5547 5548 5549 5550 5551 5552 5553
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;
}

5554
/* Initialize per network namespace state */
5555
static int __net_init netdev_init(struct net *net)
5556 5557 5558
{
	INIT_LIST_HEAD(&net->dev_base_head);

5559 5560 5561
	net->dev_name_head = netdev_create_hash();
	if (net->dev_name_head == NULL)
		goto err_name;
5562

5563 5564 5565
	net->dev_index_head = netdev_create_hash();
	if (net->dev_index_head == NULL)
		goto err_idx;
5566 5567

	return 0;
5568 5569 5570 5571 5572

err_idx:
	kfree(net->dev_name_head);
err_name:
	return -ENOMEM;
5573 5574
}

5575 5576 5577 5578 5579 5580 5581 5582
/**
 *	netdev_drivername - network driver for the device
 *	@dev: network device
 *	@buffer: buffer for resulting name
 *	@len: size of buffer
 *
 *	Determine network driver for device.
 */
5583
char *netdev_drivername(const struct net_device *dev, char *buffer, int len)
5584
{
5585 5586
	const struct device_driver *driver;
	const struct device *parent;
5587 5588 5589 5590 5591 5592 5593 5594 5595 5596 5597 5598 5599 5600 5601 5602

	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;
}

5603
static void __net_exit netdev_exit(struct net *net)
5604 5605 5606 5607 5608
{
	kfree(net->dev_name_head);
	kfree(net->dev_index_head);
}

5609
static struct pernet_operations __net_initdata netdev_net_ops = {
5610 5611 5612 5613
	.init = netdev_init,
	.exit = netdev_exit,
};

5614
static void __net_exit default_device_exit(struct net *net)
5615
{
5616
	struct net_device *dev;
5617 5618 5619 5620 5621
	/*
	 * Push all migratable of the network devices back to the
	 * initial network namespace
	 */
	rtnl_lock();
5622 5623
restart:
	for_each_netdev(net, dev) {
5624
		int err;
5625
		char fb_name[IFNAMSIZ];
5626 5627 5628 5629 5630

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

5631 5632
		/* Delete virtual devices */
		if (dev->rtnl_link_ops && dev->rtnl_link_ops->dellink) {
5633
			dev->rtnl_link_ops->dellink(dev, NULL);
5634
			goto restart;
5635 5636
		}

5637
		/* Push remaing network devices to init_net */
5638 5639
		snprintf(fb_name, IFNAMSIZ, "dev%d", dev->ifindex);
		err = dev_change_net_namespace(dev, &init_net, fb_name);
5640
		if (err) {
5641
			printk(KERN_EMERG "%s: failed to move %s to init_net: %d\n",
5642
				__func__, dev->name, err);
5643
			BUG();
5644
		}
5645
		goto restart;
5646 5647 5648 5649
	}
	rtnl_unlock();
}

5650
static struct pernet_operations __net_initdata default_device_ops = {
5651 5652 5653
	.exit = default_device_exit,
};

L
Linus Torvalds 已提交
5654 5655 5656 5657 5658 5659 5660 5661 5662 5663 5664 5665 5666 5667 5668 5669 5670 5671 5672 5673
/*
 *	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;

5674
	if (netdev_kobject_init())
L
Linus Torvalds 已提交
5675 5676 5677
		goto out;

	INIT_LIST_HEAD(&ptype_all);
5678
	for (i = 0; i < PTYPE_HASH_SIZE; i++)
L
Linus Torvalds 已提交
5679 5680
		INIT_LIST_HEAD(&ptype_base[i]);

5681 5682
	if (register_pernet_subsys(&netdev_net_ops))
		goto out;
L
Linus Torvalds 已提交
5683 5684 5685 5686 5687

	/*
	 *	Initialise the packet receive queues.
	 */

5688
	for_each_possible_cpu(i) {
L
Linus Torvalds 已提交
5689 5690 5691 5692 5693 5694
		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);
5695 5696 5697

		queue->backlog.poll = process_backlog;
		queue->backlog.weight = weight_p;
5698
		queue->backlog.gro_list = NULL;
5699
		queue->backlog.gro_count = 0;
L
Linus Torvalds 已提交
5700 5701 5702 5703
	}

	dev_boot_phase = 0;

5704 5705 5706 5707 5708 5709 5710 5711 5712 5713 5714 5715 5716 5717 5718
	/* 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;

5719 5720
	open_softirq(NET_TX_SOFTIRQ, net_tx_action);
	open_softirq(NET_RX_SOFTIRQ, net_rx_action);
L
Linus Torvalds 已提交
5721 5722 5723 5724 5725 5726 5727 5728 5729 5730 5731

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

subsys_initcall(net_dev_init);

5732 5733 5734 5735 5736 5737 5738 5739
static int __init initialize_hashrnd(void)
{
	get_random_bytes(&skb_tx_hashrnd, sizeof(skb_tx_hashrnd));
	return 0;
}

late_initcall_sync(initialize_hashrnd);