bond_main.c 141.2 KB
Newer Older
L
Linus Torvalds 已提交
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33
/*
 * originally based on the dummy device.
 *
 * Copyright 1999, Thomas Davis, tadavis@lbl.gov.
 * Licensed under the GPL. Based on dummy.c, and eql.c devices.
 *
 * bonding.c: an Ethernet Bonding driver
 *
 * This is useful to talk to a Cisco EtherChannel compatible equipment:
 *	Cisco 5500
 *	Sun Trunking (Solaris)
 *	Alteon AceDirector Trunks
 *	Linux Bonding
 *	and probably many L2 switches ...
 *
 * How it works:
 *    ifconfig bond0 ipaddress netmask up
 *      will setup a network device, with an ip address.  No mac address
 *	will be assigned at this time.  The hw mac address will come from
 *	the first slave bonded to the channel.  All slaves will then use
 *	this hw mac address.
 *
 *    ifconfig bond0 down
 *         will release all slaves, marking them as down.
 *
 *    ifenslave bond0 eth0
 *	will attach eth0 to bond0 as a slave.  eth0 hw mac address will either
 *	a: be used as initial mac address
 *	b: if a hw mac address already is there, eth0's hw mac address
 *	   will then be set from bond0.
 *
 */

J
Joe Perches 已提交
34 35
#define pr_fmt(fmt) KBUILD_MODNAME ": " fmt

L
Linus Torvalds 已提交
36 37 38 39 40 41 42 43
#include <linux/kernel.h>
#include <linux/module.h>
#include <linux/types.h>
#include <linux/fcntl.h>
#include <linux/interrupt.h>
#include <linux/ptrace.h>
#include <linux/ioport.h>
#include <linux/in.h>
44
#include <net/ip.h>
L
Linus Torvalds 已提交
45
#include <linux/ip.h>
46 47
#include <linux/tcp.h>
#include <linux/udp.h>
L
Linus Torvalds 已提交
48 49 50 51 52 53 54 55
#include <linux/slab.h>
#include <linux/string.h>
#include <linux/init.h>
#include <linux/timer.h>
#include <linux/socket.h>
#include <linux/ctype.h>
#include <linux/inet.h>
#include <linux/bitops.h>
S
Stephen Hemminger 已提交
56
#include <linux/io.h>
L
Linus Torvalds 已提交
57 58
#include <asm/system.h>
#include <asm/dma.h>
S
Stephen Hemminger 已提交
59
#include <linux/uaccess.h>
L
Linus Torvalds 已提交
60 61
#include <linux/errno.h>
#include <linux/netdevice.h>
62
#include <linux/netpoll.h>
L
Linus Torvalds 已提交
63
#include <linux/inetdevice.h>
64
#include <linux/igmp.h>
L
Linus Torvalds 已提交
65 66 67 68 69 70 71 72 73 74 75 76 77
#include <linux/etherdevice.h>
#include <linux/skbuff.h>
#include <net/sock.h>
#include <linux/rtnetlink.h>
#include <linux/proc_fs.h>
#include <linux/seq_file.h>
#include <linux/smp.h>
#include <linux/if_ether.h>
#include <net/arp.h>
#include <linux/mii.h>
#include <linux/ethtool.h>
#include <linux/if_vlan.h>
#include <linux/if_bonding.h>
D
David Sterba 已提交
78
#include <linux/jiffies.h>
79
#include <linux/preempt.h>
J
Jay Vosburgh 已提交
80
#include <net/route.h>
81
#include <net/net_namespace.h>
82
#include <net/netns/generic.h>
L
Linus Torvalds 已提交
83 84 85 86 87 88 89 90 91 92 93
#include "bonding.h"
#include "bond_3ad.h"
#include "bond_alb.h"

/*---------------------------- Module parameters ----------------------------*/

/* monitor all links that often (in milliseconds). <=0 disables monitoring */
#define BOND_LINK_MON_INTERV	0
#define BOND_LINK_ARP_INTERV	0

static int max_bonds	= BOND_DEFAULT_MAX_BONDS;
94
static int tx_queues	= BOND_DEFAULT_TX_QUEUES;
95
static int num_grat_arp = 1;
96
static int num_unsol_na = 1;
L
Linus Torvalds 已提交
97
static int miimon	= BOND_LINK_MON_INTERV;
S
Stephen Hemminger 已提交
98 99
static int updelay;
static int downdelay;
L
Linus Torvalds 已提交
100
static int use_carrier	= 1;
S
Stephen Hemminger 已提交
101 102
static char *mode;
static char *primary;
103
static char *primary_reselect;
S
Stephen Hemminger 已提交
104 105 106
static char *lacp_rate;
static char *ad_select;
static char *xmit_hash_policy;
L
Linus Torvalds 已提交
107
static int arp_interval = BOND_LINK_ARP_INTERV;
S
Stephen Hemminger 已提交
108 109 110
static char *arp_ip_target[BOND_MAX_ARP_TARGETS];
static char *arp_validate;
static char *fail_over_mac;
111
static int all_slaves_active = 0;
112
static struct bond_params bonding_defaults;
113
static int resend_igmp = BOND_DEFAULT_RESEND_IGMP;
L
Linus Torvalds 已提交
114 115 116

module_param(max_bonds, int, 0);
MODULE_PARM_DESC(max_bonds, "Max number of bonded devices");
117 118
module_param(tx_queues, int, 0);
MODULE_PARM_DESC(tx_queues, "Max number of transmit queues (default = 16)");
119 120
module_param(num_grat_arp, int, 0644);
MODULE_PARM_DESC(num_grat_arp, "Number of gratuitous ARP packets to send on failover event");
121 122
module_param(num_unsol_na, int, 0644);
MODULE_PARM_DESC(num_unsol_na, "Number of unsolicited IPv6 Neighbor Advertisements packets to send on failover event");
L
Linus Torvalds 已提交
123 124 125 126 127
module_param(miimon, int, 0);
MODULE_PARM_DESC(miimon, "Link check interval in milliseconds");
module_param(updelay, int, 0);
MODULE_PARM_DESC(updelay, "Delay before considering link up, in milliseconds");
module_param(downdelay, int, 0);
128 129
MODULE_PARM_DESC(downdelay, "Delay before considering link down, "
			    "in milliseconds");
L
Linus Torvalds 已提交
130
module_param(use_carrier, int, 0);
131 132
MODULE_PARM_DESC(use_carrier, "Use netif_carrier_ok (vs MII ioctls) in miimon; "
			      "0 for off, 1 for on (default)");
L
Linus Torvalds 已提交
133
module_param(mode, charp, 0);
134 135 136 137
MODULE_PARM_DESC(mode, "Mode of operation : 0 for balance-rr, "
		       "1 for active-backup, 2 for balance-xor, "
		       "3 for broadcast, 4 for 802.3ad, 5 for balance-tlb, "
		       "6 for balance-alb");
L
Linus Torvalds 已提交
138 139
module_param(primary, charp, 0);
MODULE_PARM_DESC(primary, "Primary network device to use");
140 141 142 143 144 145 146 147
module_param(primary_reselect, charp, 0);
MODULE_PARM_DESC(primary_reselect, "Reselect primary slave "
				   "once it comes up; "
				   "0 for always (default), "
				   "1 for only if speed of primary is "
				   "better, "
				   "2 for only on active slave "
				   "failure");
L
Linus Torvalds 已提交
148
module_param(lacp_rate, charp, 0);
149 150
MODULE_PARM_DESC(lacp_rate, "LACPDU tx rate to request from 802.3ad partner "
			    "(slow/fast)");
151 152
module_param(ad_select, charp, 0);
MODULE_PARM_DESC(ad_select, "803.ad aggregation selection logic: stable (0, default), bandwidth (1), count (2)");
153
module_param(xmit_hash_policy, charp, 0);
154 155
MODULE_PARM_DESC(xmit_hash_policy, "XOR hashing method: 0 for layer 2 (default)"
				   ", 1 for layer 3+4");
L
Linus Torvalds 已提交
156 157 158 159
module_param(arp_interval, int, 0);
MODULE_PARM_DESC(arp_interval, "arp interval in milliseconds");
module_param_array(arp_ip_target, charp, NULL, 0);
MODULE_PARM_DESC(arp_ip_target, "arp targets in n.n.n.n form");
160 161
module_param(arp_validate, charp, 0);
MODULE_PARM_DESC(arp_validate, "validate src/dst of ARP probes: none (default), active, backup or all");
162 163
module_param(fail_over_mac, charp, 0);
MODULE_PARM_DESC(fail_over_mac, "For active-backup, do not set all slaves to the same MAC.  none (default), active or follow");
164 165 166 167
module_param(all_slaves_active, int, 0);
MODULE_PARM_DESC(all_slaves_active, "Keep all frames received on an interface"
				     "by setting active flag for all slaves.  "
				     "0 for never (default), 1 for always.");
168 169
module_param(resend_igmp, int, 0);
MODULE_PARM_DESC(resend_igmp, "Number of IGMP membership reports to send on link failure");
L
Linus Torvalds 已提交
170 171 172

/*----------------------------- Global variables ----------------------------*/

173 174 175 176
#ifdef CONFIG_NET_POLL_CONTROLLER
cpumask_var_t netpoll_block_tx;
#endif

177
static const char * const version =
L
Linus Torvalds 已提交
178 179
	DRV_DESCRIPTION ": v" DRV_VERSION " (" DRV_RELDATE ")\n";

180
int bond_net_id __read_mostly;
L
Linus Torvalds 已提交
181

S
Stephen Hemminger 已提交
182 183
static __be32 arp_target[BOND_MAX_ARP_TARGETS];
static int arp_ip_count;
L
Linus Torvalds 已提交
184
static int bond_mode	= BOND_MODE_ROUNDROBIN;
S
Stephen Hemminger 已提交
185 186
static int xmit_hashtype = BOND_XMIT_POLICY_LAYER2;
static int lacp_fast;
E
Eric Dumazet 已提交
187
#ifdef CONFIG_NET_POLL_CONTROLLER
188
static int disable_netpoll = 1;
E
Eric Dumazet 已提交
189
#endif
L
Linus Torvalds 已提交
190

191
const struct bond_parm_tbl bond_lacp_tbl[] = {
L
Linus Torvalds 已提交
192 193 194 195 196
{	"slow",		AD_LACP_SLOW},
{	"fast",		AD_LACP_FAST},
{	NULL,		-1},
};

197
const struct bond_parm_tbl bond_mode_tbl[] = {
L
Linus Torvalds 已提交
198 199 200 201 202 203 204 205 206 207
{	"balance-rr",		BOND_MODE_ROUNDROBIN},
{	"active-backup",	BOND_MODE_ACTIVEBACKUP},
{	"balance-xor",		BOND_MODE_XOR},
{	"broadcast",		BOND_MODE_BROADCAST},
{	"802.3ad",		BOND_MODE_8023AD},
{	"balance-tlb",		BOND_MODE_TLB},
{	"balance-alb",		BOND_MODE_ALB},
{	NULL,			-1},
};

208
const struct bond_parm_tbl xmit_hashtype_tbl[] = {
209 210
{	"layer2",		BOND_XMIT_POLICY_LAYER2},
{	"layer3+4",		BOND_XMIT_POLICY_LAYER34},
211
{	"layer2+3",		BOND_XMIT_POLICY_LAYER23},
212 213 214
{	NULL,			-1},
};

215
const struct bond_parm_tbl arp_validate_tbl[] = {
216 217 218 219 220 221 222
{	"none",			BOND_ARP_VALIDATE_NONE},
{	"active",		BOND_ARP_VALIDATE_ACTIVE},
{	"backup",		BOND_ARP_VALIDATE_BACKUP},
{	"all",			BOND_ARP_VALIDATE_ALL},
{	NULL,			-1},
};

223
const struct bond_parm_tbl fail_over_mac_tbl[] = {
224 225 226 227 228 229
{	"none",			BOND_FOM_NONE},
{	"active",		BOND_FOM_ACTIVE},
{	"follow",		BOND_FOM_FOLLOW},
{	NULL,			-1},
};

230 231 232 233 234 235 236
const struct bond_parm_tbl pri_reselect_tbl[] = {
{	"always",		BOND_PRI_RESELECT_ALWAYS},
{	"better",		BOND_PRI_RESELECT_BETTER},
{	"failure",		BOND_PRI_RESELECT_FAILURE},
{	NULL,			-1},
};

237 238 239 240 241 242 243
struct bond_parm_tbl ad_select_tbl[] = {
{	"stable",	BOND_AD_STABLE},
{	"bandwidth",	BOND_AD_BANDWIDTH},
{	"count",	BOND_AD_COUNT},
{	NULL,		-1},
};

L
Linus Torvalds 已提交
244 245
/*-------------------------- Forward declarations ---------------------------*/

J
Jay Vosburgh 已提交
246
static void bond_send_gratuitous_arp(struct bonding *bond);
247
static int bond_init(struct net_device *bond_dev);
248
static void bond_uninit(struct net_device *bond_dev);
L
Linus Torvalds 已提交
249 250 251

/*---------------------------- General routines -----------------------------*/

252
static const char *bond_mode_name(int mode)
L
Linus Torvalds 已提交
253
{
254 255 256 257 258
	static const char *names[] = {
		[BOND_MODE_ROUNDROBIN] = "load balancing (round-robin)",
		[BOND_MODE_ACTIVEBACKUP] = "fault-tolerance (active-backup)",
		[BOND_MODE_XOR] = "load balancing (xor)",
		[BOND_MODE_BROADCAST] = "fault-tolerance (broadcast)",
S
Stephen Hemminger 已提交
259
		[BOND_MODE_8023AD] = "IEEE 802.3ad Dynamic link aggregation",
260 261 262 263 264
		[BOND_MODE_TLB] = "transmit load balancing",
		[BOND_MODE_ALB] = "adaptive load balancing",
	};

	if (mode < 0 || mode > BOND_MODE_ALB)
L
Linus Torvalds 已提交
265
		return "unknown";
266 267

	return names[mode];
L
Linus Torvalds 已提交
268 269 270 271 272 273 274 275 276 277 278 279 280 281 282
}

/*---------------------------------- VLAN -----------------------------------*/

/**
 * bond_add_vlan - add a new vlan id on bond
 * @bond: bond that got the notification
 * @vlan_id: the vlan id to add
 *
 * Returns -ENOMEM if allocation failed.
 */
static int bond_add_vlan(struct bonding *bond, unsigned short vlan_id)
{
	struct vlan_entry *vlan;

283
	pr_debug("bond: %s, vlan id %d\n",
J
Joe Perches 已提交
284
		 (bond ? bond->dev->name : "None"), vlan_id);
L
Linus Torvalds 已提交
285

286
	vlan = kzalloc(sizeof(struct vlan_entry), GFP_KERNEL);
S
Stephen Hemminger 已提交
287
	if (!vlan)
L
Linus Torvalds 已提交
288 289 290 291 292 293 294 295 296 297 298
		return -ENOMEM;

	INIT_LIST_HEAD(&vlan->vlan_list);
	vlan->vlan_id = vlan_id;

	write_lock_bh(&bond->lock);

	list_add_tail(&vlan->vlan_list, &bond->vlan_list);

	write_unlock_bh(&bond->lock);

299
	pr_debug("added VLAN ID %d on bond %s\n", vlan_id, bond->dev->name);
L
Linus Torvalds 已提交
300 301 302 303 304 305 306 307 308 309 310 311 312

	return 0;
}

/**
 * bond_del_vlan - delete a vlan id from bond
 * @bond: bond that got the notification
 * @vlan_id: the vlan id to delete
 *
 * returns -ENODEV if @vlan_id was not found in @bond.
 */
static int bond_del_vlan(struct bonding *bond, unsigned short vlan_id)
{
313
	struct vlan_entry *vlan;
L
Linus Torvalds 已提交
314 315
	int res = -ENODEV;

316
	pr_debug("bond: %s, vlan id %d\n", bond->dev->name, vlan_id);
L
Linus Torvalds 已提交
317

318
	block_netpoll_tx();
L
Linus Torvalds 已提交
319 320
	write_lock_bh(&bond->lock);

321
	list_for_each_entry(vlan, &bond->vlan_list, vlan_list) {
L
Linus Torvalds 已提交
322 323 324
		if (vlan->vlan_id == vlan_id) {
			list_del(&vlan->vlan_list);

325
			if (bond_is_lb(bond))
L
Linus Torvalds 已提交
326 327
				bond_alb_clear_vlan(bond, vlan_id);

J
Joe Perches 已提交
328 329
			pr_debug("removed VLAN ID %d from bond %s\n",
				 vlan_id, bond->dev->name);
L
Linus Torvalds 已提交
330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347

			kfree(vlan);

			if (list_empty(&bond->vlan_list) &&
			    (bond->slave_cnt == 0)) {
				/* Last VLAN removed and no slaves, so
				 * restore block on adding VLANs. This will
				 * be removed once new slaves that are not
				 * VLAN challenged will be added.
				 */
				bond->dev->features |= NETIF_F_VLAN_CHALLENGED;
			}

			res = 0;
			goto out;
		}
	}

J
Joe Perches 已提交
348 349
	pr_debug("couldn't find VLAN ID %d in bond %s\n",
		 vlan_id, bond->dev->name);
L
Linus Torvalds 已提交
350 351 352

out:
	write_unlock_bh(&bond->lock);
353
	unblock_netpoll_tx();
L
Linus Torvalds 已提交
354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372
	return res;
}

/**
 * bond_has_challenged_slaves
 * @bond: the bond we're working on
 *
 * Searches the slave list. Returns 1 if a vlan challenged slave
 * was found, 0 otherwise.
 *
 * Assumes bond->lock is held.
 */
static int bond_has_challenged_slaves(struct bonding *bond)
{
	struct slave *slave;
	int i;

	bond_for_each_slave(bond, slave, i) {
		if (slave->dev->features & NETIF_F_VLAN_CHALLENGED) {
373
			pr_debug("found VLAN challenged slave - %s\n",
J
Joe Perches 已提交
374
				 slave->dev->name);
L
Linus Torvalds 已提交
375 376 377 378
			return 1;
		}
	}

379
	pr_debug("no VLAN challenged slaves found\n");
L
Linus Torvalds 已提交
380 381 382 383 384 385 386 387 388 389
	return 0;
}

/**
 * bond_next_vlan - safely skip to the next item in the vlans list.
 * @bond: the bond we're working on
 * @curr: item we're advancing from
 *
 * Returns %NULL if list is empty, bond->next_vlan if @curr is %NULL,
 * or @curr->next otherwise (even if it is @curr itself again).
S
Stephen Hemminger 已提交
390
 *
L
Linus Torvalds 已提交
391 392 393 394 395 396
 * Caller must hold bond->lock
 */
struct vlan_entry *bond_next_vlan(struct bonding *bond, struct vlan_entry *curr)
{
	struct vlan_entry *next, *last;

S
Stephen Hemminger 已提交
397
	if (list_empty(&bond->vlan_list))
L
Linus Torvalds 已提交
398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419
		return NULL;

	if (!curr) {
		next = list_entry(bond->vlan_list.next,
				  struct vlan_entry, vlan_list);
	} else {
		last = list_entry(bond->vlan_list.prev,
				  struct vlan_entry, vlan_list);
		if (last == curr) {
			next = list_entry(bond->vlan_list.next,
					  struct vlan_entry, vlan_list);
		} else {
			next = list_entry(curr->vlan_list.next,
					  struct vlan_entry, vlan_list);
		}
	}

	return next;
}

/**
 * bond_dev_queue_xmit - Prepare skb for xmit.
S
Stephen Hemminger 已提交
420
 *
L
Linus Torvalds 已提交
421 422 423
 * @bond: bond device that got this skb for tx.
 * @skb: hw accel VLAN tagged skb to transmit
 * @slave_dev: slave that is supposed to xmit this skbuff
S
Stephen Hemminger 已提交
424
 *
L
Linus Torvalds 已提交
425 426 427 428 429 430 431
 * When the bond gets an skb to transmit that is
 * already hardware accelerated VLAN tagged, and it
 * needs to relay this skb to a slave that is not
 * hw accel capable, the skb needs to be "unaccelerated",
 * i.e. strip the hwaccel tag and re-insert it as part
 * of the payload.
 */
S
Stephen Hemminger 已提交
432 433
int bond_dev_queue_xmit(struct bonding *bond, struct sk_buff *skb,
			struct net_device *slave_dev)
L
Linus Torvalds 已提交
434
{
435
	unsigned short uninitialized_var(vlan_id);
L
Linus Torvalds 已提交
436

437
	/* Test vlan_list not vlgrp to catch and handle 802.1p tags */
L
Linus Torvalds 已提交
438 439 440 441 442 443 444 445 446 447 448 449 450 451 452 453 454
	if (!list_empty(&bond->vlan_list) &&
	    !(slave_dev->features & NETIF_F_HW_VLAN_TX) &&
	    vlan_get_tag(skb, &vlan_id) == 0) {
		skb->dev = slave_dev;
		skb = vlan_put_tag(skb, vlan_id);
		if (!skb) {
			/* vlan_put_tag() frees the skb in case of error,
			 * so return success here so the calling functions
			 * won't attempt to free is again.
			 */
			return 0;
		}
	} else {
		skb->dev = slave_dev;
	}

	skb->priority = 1;
455 456 457 458 459
#ifdef CONFIG_NET_POLL_CONTROLLER
	if (unlikely(bond->dev->priv_flags & IFF_IN_NETPOLL)) {
		struct netpoll *np = bond->dev->npinfo->netpoll;
		slave_dev->npinfo = bond->dev->npinfo;
		slave_dev->priv_flags |= IFF_IN_NETPOLL;
460
		netpoll_send_skb_on_dev(np, skb, slave_dev);
461 462 463 464
		slave_dev->priv_flags &= ~IFF_IN_NETPOLL;
	} else
#endif
		dev_queue_xmit(skb);
L
Linus Torvalds 已提交
465 466 467 468 469 470 471 472 473 474 475 476

	return 0;
}

/*
 * In the following 3 functions, bond_vlan_rx_register(), bond_vlan_rx_add_vid
 * and bond_vlan_rx_kill_vid, We don't protect the slave list iteration with a
 * lock because:
 * a. This operation is performed in IOCTL context,
 * b. The operation is protected by the RTNL semaphore in the 8021q code,
 * c. Holding a lock with BH disabled while directly calling a base driver
 *    entry point is generally a BAD idea.
S
Stephen Hemminger 已提交
477
 *
L
Linus Torvalds 已提交
478 479 480 481 482 483 484 485 486 487 488 489 490 491
 * The design of synchronization/protection for this operation in the 8021q
 * module is good for one or more VLAN devices over a single physical device
 * and cannot be extended for a teaming solution like bonding, so there is a
 * potential race condition here where a net device from the vlan group might
 * be referenced (either by a base driver or the 8021q code) while it is being
 * removed from the system. However, it turns out we're not making matters
 * worse, and if it works for regular VLAN usage it will work here too.
*/

/**
 * bond_vlan_rx_register - Propagates registration to slaves
 * @bond_dev: bonding net device that got called
 * @grp: vlan group being registered
 */
S
Stephen Hemminger 已提交
492 493
static void bond_vlan_rx_register(struct net_device *bond_dev,
				  struct vlan_group *grp)
L
Linus Torvalds 已提交
494
{
495
	struct bonding *bond = netdev_priv(bond_dev);
L
Linus Torvalds 已提交
496 497 498
	struct slave *slave;
	int i;

499
	write_lock(&bond->lock);
L
Linus Torvalds 已提交
500
	bond->vlgrp = grp;
501
	write_unlock(&bond->lock);
L
Linus Torvalds 已提交
502 503 504

	bond_for_each_slave(bond, slave, i) {
		struct net_device *slave_dev = slave->dev;
505
		const struct net_device_ops *slave_ops = slave_dev->netdev_ops;
L
Linus Torvalds 已提交
506 507

		if ((slave_dev->features & NETIF_F_HW_VLAN_RX) &&
508 509
		    slave_ops->ndo_vlan_rx_register) {
			slave_ops->ndo_vlan_rx_register(slave_dev, grp);
L
Linus Torvalds 已提交
510 511 512 513 514 515 516 517 518 519 520
		}
	}
}

/**
 * bond_vlan_rx_add_vid - Propagates adding an id to slaves
 * @bond_dev: bonding net device that got called
 * @vid: vlan id being added
 */
static void bond_vlan_rx_add_vid(struct net_device *bond_dev, uint16_t vid)
{
521
	struct bonding *bond = netdev_priv(bond_dev);
L
Linus Torvalds 已提交
522 523 524 525 526
	struct slave *slave;
	int i, res;

	bond_for_each_slave(bond, slave, i) {
		struct net_device *slave_dev = slave->dev;
527
		const struct net_device_ops *slave_ops = slave_dev->netdev_ops;
L
Linus Torvalds 已提交
528 529

		if ((slave_dev->features & NETIF_F_HW_VLAN_FILTER) &&
530 531
		    slave_ops->ndo_vlan_rx_add_vid) {
			slave_ops->ndo_vlan_rx_add_vid(slave_dev, vid);
L
Linus Torvalds 已提交
532 533 534 535 536
		}
	}

	res = bond_add_vlan(bond, vid);
	if (res) {
J
Joe Perches 已提交
537
		pr_err("%s: Error: Failed to add vlan id %d\n",
L
Linus Torvalds 已提交
538 539 540 541 542 543 544 545 546 547 548
		       bond_dev->name, vid);
	}
}

/**
 * bond_vlan_rx_kill_vid - Propagates deleting an id to slaves
 * @bond_dev: bonding net device that got called
 * @vid: vlan id being removed
 */
static void bond_vlan_rx_kill_vid(struct net_device *bond_dev, uint16_t vid)
{
549
	struct bonding *bond = netdev_priv(bond_dev);
L
Linus Torvalds 已提交
550 551 552 553 554 555
	struct slave *slave;
	struct net_device *vlan_dev;
	int i, res;

	bond_for_each_slave(bond, slave, i) {
		struct net_device *slave_dev = slave->dev;
556
		const struct net_device_ops *slave_ops = slave_dev->netdev_ops;
L
Linus Torvalds 已提交
557 558

		if ((slave_dev->features & NETIF_F_HW_VLAN_FILTER) &&
559
		    slave_ops->ndo_vlan_rx_kill_vid) {
L
Linus Torvalds 已提交
560 561 562
			/* Save and then restore vlan_dev in the grp array,
			 * since the slave's driver might clear it.
			 */
D
Dan Aloni 已提交
563
			vlan_dev = vlan_group_get_device(bond->vlgrp, vid);
564
			slave_ops->ndo_vlan_rx_kill_vid(slave_dev, vid);
D
Dan Aloni 已提交
565
			vlan_group_set_device(bond->vlgrp, vid, vlan_dev);
L
Linus Torvalds 已提交
566 567 568 569 570
		}
	}

	res = bond_del_vlan(bond, vid);
	if (res) {
J
Joe Perches 已提交
571
		pr_err("%s: Error: Failed to remove vlan id %d\n",
L
Linus Torvalds 已提交
572 573 574 575 576 577 578
		       bond_dev->name, vid);
	}
}

static void bond_add_vlans_on_slave(struct bonding *bond, struct net_device *slave_dev)
{
	struct vlan_entry *vlan;
579
	const struct net_device_ops *slave_ops = slave_dev->netdev_ops;
L
Linus Torvalds 已提交
580

581
	if (!bond->vlgrp)
582
		return;
L
Linus Torvalds 已提交
583 584

	if ((slave_dev->features & NETIF_F_HW_VLAN_RX) &&
585 586
	    slave_ops->ndo_vlan_rx_register)
		slave_ops->ndo_vlan_rx_register(slave_dev, bond->vlgrp);
L
Linus Torvalds 已提交
587 588

	if (!(slave_dev->features & NETIF_F_HW_VLAN_FILTER) ||
589
	    !(slave_ops->ndo_vlan_rx_add_vid))
590
		return;
L
Linus Torvalds 已提交
591

592 593
	list_for_each_entry(vlan, &bond->vlan_list, vlan_list)
		slave_ops->ndo_vlan_rx_add_vid(slave_dev, vlan->vlan_id);
L
Linus Torvalds 已提交
594 595
}

S
Stephen Hemminger 已提交
596 597
static void bond_del_vlans_from_slave(struct bonding *bond,
				      struct net_device *slave_dev)
L
Linus Torvalds 已提交
598
{
599
	const struct net_device_ops *slave_ops = slave_dev->netdev_ops;
L
Linus Torvalds 已提交
600 601 602
	struct vlan_entry *vlan;
	struct net_device *vlan_dev;

603
	if (!bond->vlgrp)
604
		return;
L
Linus Torvalds 已提交
605 606

	if (!(slave_dev->features & NETIF_F_HW_VLAN_FILTER) ||
607
	    !(slave_ops->ndo_vlan_rx_kill_vid))
L
Linus Torvalds 已提交
608 609 610
		goto unreg;

	list_for_each_entry(vlan, &bond->vlan_list, vlan_list) {
611 612
		if (!vlan->vlan_id)
			continue;
L
Linus Torvalds 已提交
613 614 615
		/* Save and then restore vlan_dev in the grp array,
		 * since the slave's driver might clear it.
		 */
D
Dan Aloni 已提交
616
		vlan_dev = vlan_group_get_device(bond->vlgrp, vlan->vlan_id);
617
		slave_ops->ndo_vlan_rx_kill_vid(slave_dev, vlan->vlan_id);
D
Dan Aloni 已提交
618
		vlan_group_set_device(bond->vlgrp, vlan->vlan_id, vlan_dev);
L
Linus Torvalds 已提交
619 620 621 622
	}

unreg:
	if ((slave_dev->features & NETIF_F_HW_VLAN_RX) &&
623 624
	    slave_ops->ndo_vlan_rx_register)
		slave_ops->ndo_vlan_rx_register(slave_dev, NULL);
L
Linus Torvalds 已提交
625 626 627 628
}

/*------------------------------- Link status -------------------------------*/

629 630 631 632 633 634 635 636 637 638 639 640 641 642 643 644 645 646 647 648 649 650 651 652 653 654 655 656 657 658 659 660 661 662 663 664
/*
 * Set the carrier state for the master according to the state of its
 * slaves.  If any slaves are up, the master is up.  In 802.3ad mode,
 * do special 802.3ad magic.
 *
 * Returns zero if carrier state does not change, nonzero if it does.
 */
static int bond_set_carrier(struct bonding *bond)
{
	struct slave *slave;
	int i;

	if (bond->slave_cnt == 0)
		goto down;

	if (bond->params.mode == BOND_MODE_8023AD)
		return bond_3ad_set_carrier(bond);

	bond_for_each_slave(bond, slave, i) {
		if (slave->link == BOND_LINK_UP) {
			if (!netif_carrier_ok(bond->dev)) {
				netif_carrier_on(bond->dev);
				return 1;
			}
			return 0;
		}
	}

down:
	if (netif_carrier_ok(bond->dev)) {
		netif_carrier_off(bond->dev);
		return 1;
	}
	return 0;
}

L
Linus Torvalds 已提交
665 666 667 668 669 670 671 672 673 674
/*
 * Get link speed and duplex from the slave's base driver
 * using ethtool. If for some reason the call fails or the
 * values are invalid, fake speed and duplex to 100/Full
 * and return error.
 */
static int bond_update_speed_duplex(struct slave *slave)
{
	struct net_device *slave_dev = slave->dev;
	struct ethtool_cmd etool;
675
	int res;
L
Linus Torvalds 已提交
676 677 678 679 680

	/* Fake speed and duplex */
	slave->speed = SPEED_100;
	slave->duplex = DUPLEX_FULL;

681 682
	if (!slave_dev->ethtool_ops || !slave_dev->ethtool_ops->get_settings)
		return -1;
L
Linus Torvalds 已提交
683

684 685
	res = slave_dev->ethtool_ops->get_settings(slave_dev, &etool);
	if (res < 0)
L
Linus Torvalds 已提交
686 687 688 689 690 691
		return -1;

	switch (etool.speed) {
	case SPEED_10:
	case SPEED_100:
	case SPEED_1000:
692
	case SPEED_10000:
L
Linus Torvalds 已提交
693 694 695 696 697 698 699 700 701 702 703 704 705 706 707 708 709 710 711 712 713 714 715
		break;
	default:
		return -1;
	}

	switch (etool.duplex) {
	case DUPLEX_FULL:
	case DUPLEX_HALF:
		break;
	default:
		return -1;
	}

	slave->speed = etool.speed;
	slave->duplex = etool.duplex;

	return 0;
}

/*
 * if <dev> supports MII link status reporting, check its link status.
 *
 * We either do MII/ETHTOOL ioctls, or check netif_carrier_ok(),
S
Stephen Hemminger 已提交
716
 * depending upon the setting of the use_carrier parameter.
L
Linus Torvalds 已提交
717 718 719 720 721 722 723 724 725 726 727
 *
 * Return either BMSR_LSTATUS, meaning that the link is up (or we
 * can't tell and just pretend it is), or 0, meaning that the link is
 * down.
 *
 * If reporting is non-zero, instead of faking link up, return -1 if
 * both ETHTOOL and MII ioctls fail (meaning the device does not
 * support them).  If use_carrier is set, return whatever it says.
 * It'd be nice if there was a good way to tell if a driver supports
 * netif_carrier, but there really isn't.
 */
S
Stephen Hemminger 已提交
728 729
static int bond_check_dev_link(struct bonding *bond,
			       struct net_device *slave_dev, int reporting)
L
Linus Torvalds 已提交
730
{
731
	const struct net_device_ops *slave_ops = slave_dev->netdev_ops;
732
	int (*ioctl)(struct net_device *, struct ifreq *, int);
L
Linus Torvalds 已提交
733 734 735
	struct ifreq ifr;
	struct mii_ioctl_data *mii;

736 737 738
	if (!reporting && !netif_running(slave_dev))
		return 0;

739
	if (bond->params.use_carrier)
L
Linus Torvalds 已提交
740 741
		return netif_carrier_ok(slave_dev) ? BMSR_LSTATUS : 0;

742 743 744 745 746 747 748 749 750 751 752
	/* Try to get link status using Ethtool first. */
	if (slave_dev->ethtool_ops) {
		if (slave_dev->ethtool_ops->get_link) {
			u32 link;

			link = slave_dev->ethtool_ops->get_link(slave_dev);

			return link ? BMSR_LSTATUS : 0;
		}
	}

S
Stephen Hemminger 已提交
753
	/* Ethtool can't be used, fallback to MII ioctls. */
754
	ioctl = slave_ops->ndo_do_ioctl;
L
Linus Torvalds 已提交
755 756 757 758 759 760 761 762 763 764 765 766 767 768 769 770 771 772
	if (ioctl) {
		/* TODO: set pointer to correct ioctl on a per team member */
		/*       bases to make this more efficient. that is, once  */
		/*       we determine the correct ioctl, we will always    */
		/*       call it and not the others for that team          */
		/*       member.                                           */

		/*
		 * We cannot assume that SIOCGMIIPHY will also read a
		 * register; not all network drivers (e.g., e100)
		 * support that.
		 */

		/* Yes, the mii is overlaid on the ifreq.ifr_ifru */
		strncpy(ifr.ifr_name, slave_dev->name, IFNAMSIZ);
		mii = if_mii(&ifr);
		if (IOCTL(slave_dev, &ifr, SIOCGMIIPHY) == 0) {
			mii->reg_num = MII_BMSR;
S
Stephen Hemminger 已提交
773 774
			if (IOCTL(slave_dev, &ifr, SIOCGMIIREG) == 0)
				return mii->val_out & BMSR_LSTATUS;
L
Linus Torvalds 已提交
775 776 777 778 779
		}
	}

	/*
	 * If reporting, report that either there's no dev->do_ioctl,
780
	 * or both SIOCGMIIREG and get_link failed (meaning that we
L
Linus Torvalds 已提交
781 782 783
	 * cannot report link status).  If not reporting, pretend
	 * we're ok.
	 */
S
Stephen Hemminger 已提交
784
	return reporting ? -1 : BMSR_LSTATUS;
L
Linus Torvalds 已提交
785 786 787 788 789 790 791
}

/*----------------------------- Multicast list ------------------------------*/

/*
 * Push the promiscuity flag down to appropriate slaves
 */
792
static int bond_set_promiscuity(struct bonding *bond, int inc)
L
Linus Torvalds 已提交
793
{
794
	int err = 0;
L
Linus Torvalds 已提交
795 796 797
	if (USES_PRIMARY(bond->params.mode)) {
		/* write lock already acquired */
		if (bond->curr_active_slave) {
798 799
			err = dev_set_promiscuity(bond->curr_active_slave->dev,
						  inc);
L
Linus Torvalds 已提交
800 801 802 803 804
		}
	} else {
		struct slave *slave;
		int i;
		bond_for_each_slave(bond, slave, i) {
805 806 807
			err = dev_set_promiscuity(slave->dev, inc);
			if (err)
				return err;
L
Linus Torvalds 已提交
808 809
		}
	}
810
	return err;
L
Linus Torvalds 已提交
811 812 813 814 815
}

/*
 * Push the allmulti flag down to all slaves
 */
816
static int bond_set_allmulti(struct bonding *bond, int inc)
L
Linus Torvalds 已提交
817
{
818
	int err = 0;
L
Linus Torvalds 已提交
819 820 821
	if (USES_PRIMARY(bond->params.mode)) {
		/* write lock already acquired */
		if (bond->curr_active_slave) {
822 823
			err = dev_set_allmulti(bond->curr_active_slave->dev,
					       inc);
L
Linus Torvalds 已提交
824 825 826 827 828
		}
	} else {
		struct slave *slave;
		int i;
		bond_for_each_slave(bond, slave, i) {
829 830 831
			err = dev_set_allmulti(slave->dev, inc);
			if (err)
				return err;
L
Linus Torvalds 已提交
832 833
		}
	}
834
	return err;
L
Linus Torvalds 已提交
835 836 837 838 839 840
}

/*
 * Add a Multicast address to slaves
 * according to mode
 */
841
static void bond_mc_add(struct bonding *bond, void *addr)
L
Linus Torvalds 已提交
842 843 844
{
	if (USES_PRIMARY(bond->params.mode)) {
		/* write lock already acquired */
S
Stephen Hemminger 已提交
845
		if (bond->curr_active_slave)
846
			dev_mc_add(bond->curr_active_slave->dev, addr);
L
Linus Torvalds 已提交
847 848 849
	} else {
		struct slave *slave;
		int i;
S
Stephen Hemminger 已提交
850 851

		bond_for_each_slave(bond, slave, i)
852
			dev_mc_add(slave->dev, addr);
L
Linus Torvalds 已提交
853 854 855 856 857 858 859
	}
}

/*
 * Remove a multicast address from slave
 * according to mode
 */
860
static void bond_mc_del(struct bonding *bond, void *addr)
L
Linus Torvalds 已提交
861 862 863
{
	if (USES_PRIMARY(bond->params.mode)) {
		/* write lock already acquired */
S
Stephen Hemminger 已提交
864
		if (bond->curr_active_slave)
865
			dev_mc_del(bond->curr_active_slave->dev, addr);
L
Linus Torvalds 已提交
866 867 868 869
	} else {
		struct slave *slave;
		int i;
		bond_for_each_slave(bond, slave, i) {
870
			dev_mc_del(slave->dev, addr);
L
Linus Torvalds 已提交
871 872 873 874
		}
	}
}

875

876
static void __bond_resend_igmp_join_requests(struct net_device *dev)
877 878 879 880 881
{
	struct in_device *in_dev;
	struct ip_mc_list *im;

	rcu_read_lock();
882
	in_dev = __in_dev_get_rcu(dev);
883
	if (in_dev) {
S
Stephen Hemminger 已提交
884
		for (im = in_dev->mc_list; im; im = im->next)
885 886 887 888 889 890
			ip_mc_rejoin_group(im);
	}

	rcu_read_unlock();
}

891 892 893 894 895 896 897 898 899 900 901 902 903 904 905 906 907 908 909 910 911 912 913 914 915
/*
 * Retrieve the list of registered multicast addresses for the bonding
 * device and retransmit an IGMP JOIN request to the current active
 * slave.
 */
static void bond_resend_igmp_join_requests(struct bonding *bond)
{
	struct net_device *vlan_dev;
	struct vlan_entry *vlan;

	read_lock(&bond->lock);

	/* rejoin all groups on bond device */
	__bond_resend_igmp_join_requests(bond->dev);

	/* rejoin all groups on vlan devices */
	if (bond->vlgrp) {
		list_for_each_entry(vlan, &bond->vlan_list, vlan_list) {
			vlan_dev = vlan_group_get_device(bond->vlgrp,
							 vlan->vlan_id);
			if (vlan_dev)
				__bond_resend_igmp_join_requests(vlan_dev);
		}
	}

916 917 918
	if (--bond->igmp_retrans > 0)
		queue_delayed_work(bond->wq, &bond->mcast_work, HZ/5);

919 920 921 922 923 924 925 926 927 928
	read_unlock(&bond->lock);
}

void bond_resend_igmp_join_requests_delayed(struct work_struct *work)
{
	struct bonding *bond = container_of(work, struct bonding,
							mcast_work.work);
	bond_resend_igmp_join_requests(bond);
}

L
Linus Torvalds 已提交
929 930 931
/*
 * flush all members of flush->mc_list from device dev->mc_list
 */
S
Stephen Hemminger 已提交
932 933
static void bond_mc_list_flush(struct net_device *bond_dev,
			       struct net_device *slave_dev)
L
Linus Torvalds 已提交
934
{
935
	struct bonding *bond = netdev_priv(bond_dev);
936
	struct netdev_hw_addr *ha;
L
Linus Torvalds 已提交
937

938 939
	netdev_for_each_mc_addr(ha, bond_dev)
		dev_mc_del(slave_dev, ha->addr);
L
Linus Torvalds 已提交
940 941 942 943 944

	if (bond->params.mode == BOND_MODE_8023AD) {
		/* del lacpdu mc addr from mc list */
		u8 lacpdu_multicast[ETH_ALEN] = MULTICAST_LACPDU_ADDR;

945
		dev_mc_del(slave_dev, lacpdu_multicast);
L
Linus Torvalds 已提交
946 947 948 949 950 951 952 953 954 955
	}
}

/*--------------------------- Active slave change ---------------------------*/

/*
 * Update the mc list and multicast-related flags for the new and
 * old active slaves (if any) according to the multicast mode, and
 * promiscuous flags unconditionally.
 */
S
Stephen Hemminger 已提交
956 957
static void bond_mc_swap(struct bonding *bond, struct slave *new_active,
			 struct slave *old_active)
L
Linus Torvalds 已提交
958
{
959
	struct netdev_hw_addr *ha;
L
Linus Torvalds 已提交
960

S
Stephen Hemminger 已提交
961
	if (!USES_PRIMARY(bond->params.mode))
L
Linus Torvalds 已提交
962 963 964 965 966 967
		/* nothing to do -  mc list is already up-to-date on
		 * all slaves
		 */
		return;

	if (old_active) {
S
Stephen Hemminger 已提交
968
		if (bond->dev->flags & IFF_PROMISC)
L
Linus Torvalds 已提交
969 970
			dev_set_promiscuity(old_active->dev, -1);

S
Stephen Hemminger 已提交
971
		if (bond->dev->flags & IFF_ALLMULTI)
L
Linus Torvalds 已提交
972 973
			dev_set_allmulti(old_active->dev, -1);

974 975
		netdev_for_each_mc_addr(ha, bond->dev)
			dev_mc_del(old_active->dev, ha->addr);
L
Linus Torvalds 已提交
976 977 978
	}

	if (new_active) {
979
		/* FIXME: Signal errors upstream. */
S
Stephen Hemminger 已提交
980
		if (bond->dev->flags & IFF_PROMISC)
L
Linus Torvalds 已提交
981 982
			dev_set_promiscuity(new_active->dev, 1);

S
Stephen Hemminger 已提交
983
		if (bond->dev->flags & IFF_ALLMULTI)
L
Linus Torvalds 已提交
984 985
			dev_set_allmulti(new_active->dev, 1);

986 987
		netdev_for_each_mc_addr(ha, bond->dev)
			dev_mc_add(new_active->dev, ha->addr);
L
Linus Torvalds 已提交
988 989 990
	}
}

991 992 993 994 995 996 997 998 999 1000
/*
 * bond_do_fail_over_mac
 *
 * Perform special MAC address swapping for fail_over_mac settings
 *
 * Called with RTNL, bond->lock for read, curr_slave_lock for write_bh.
 */
static void bond_do_fail_over_mac(struct bonding *bond,
				  struct slave *new_active,
				  struct slave *old_active)
1001 1002 1003 1004
	__releases(&bond->curr_slave_lock)
	__releases(&bond->lock)
	__acquires(&bond->lock)
	__acquires(&bond->curr_slave_lock)
1005 1006 1007 1008 1009 1010 1011 1012 1013 1014 1015 1016 1017 1018 1019 1020 1021 1022 1023 1024 1025 1026 1027 1028 1029 1030 1031 1032 1033 1034 1035 1036 1037 1038 1039
{
	u8 tmp_mac[ETH_ALEN];
	struct sockaddr saddr;
	int rv;

	switch (bond->params.fail_over_mac) {
	case BOND_FOM_ACTIVE:
		if (new_active)
			memcpy(bond->dev->dev_addr,  new_active->dev->dev_addr,
			       new_active->dev->addr_len);
		break;
	case BOND_FOM_FOLLOW:
		/*
		 * if new_active && old_active, swap them
		 * if just old_active, do nothing (going to no active slave)
		 * if just new_active, set new_active to bond's MAC
		 */
		if (!new_active)
			return;

		write_unlock_bh(&bond->curr_slave_lock);
		read_unlock(&bond->lock);

		if (old_active) {
			memcpy(tmp_mac, new_active->dev->dev_addr, ETH_ALEN);
			memcpy(saddr.sa_data, old_active->dev->dev_addr,
			       ETH_ALEN);
			saddr.sa_family = new_active->dev->type;
		} else {
			memcpy(saddr.sa_data, bond->dev->dev_addr, ETH_ALEN);
			saddr.sa_family = bond->dev->type;
		}

		rv = dev_set_mac_address(new_active->dev, &saddr);
		if (rv) {
J
Joe Perches 已提交
1040
			pr_err("%s: Error %d setting MAC of slave %s\n",
1041 1042 1043 1044 1045 1046 1047 1048 1049 1050 1051 1052
			       bond->dev->name, -rv, new_active->dev->name);
			goto out;
		}

		if (!old_active)
			goto out;

		memcpy(saddr.sa_data, tmp_mac, ETH_ALEN);
		saddr.sa_family = old_active->dev->type;

		rv = dev_set_mac_address(old_active->dev, &saddr);
		if (rv)
J
Joe Perches 已提交
1053
			pr_err("%s: Error %d setting MAC of slave %s\n",
1054 1055 1056 1057 1058 1059
			       bond->dev->name, -rv, new_active->dev->name);
out:
		read_lock(&bond->lock);
		write_lock_bh(&bond->curr_slave_lock);
		break;
	default:
J
Joe Perches 已提交
1060
		pr_err("%s: bond_do_fail_over_mac impossible: bad policy %d\n",
1061 1062 1063 1064 1065 1066
		       bond->dev->name, bond->params.fail_over_mac);
		break;
	}

}

1067 1068 1069 1070 1071 1072 1073 1074 1075 1076 1077 1078 1079 1080 1081 1082 1083 1084 1085
static bool bond_should_change_active(struct bonding *bond)
{
	struct slave *prim = bond->primary_slave;
	struct slave *curr = bond->curr_active_slave;

	if (!prim || !curr || curr->link != BOND_LINK_UP)
		return true;
	if (bond->force_primary) {
		bond->force_primary = false;
		return true;
	}
	if (bond->params.primary_reselect == BOND_PRI_RESELECT_BETTER &&
	    (prim->speed < curr->speed ||
	     (prim->speed == curr->speed && prim->duplex <= curr->duplex)))
		return false;
	if (bond->params.primary_reselect == BOND_PRI_RESELECT_FAILURE)
		return false;
	return true;
}
1086

L
Linus Torvalds 已提交
1087 1088 1089 1090 1091 1092 1093 1094 1095 1096 1097 1098 1099
/**
 * find_best_interface - select the best available slave to be the active one
 * @bond: our bonding struct
 *
 * Warning: Caller must hold curr_slave_lock for writing.
 */
static struct slave *bond_find_best_slave(struct bonding *bond)
{
	struct slave *new_active, *old_active;
	struct slave *bestslave = NULL;
	int mintime = bond->params.updelay;
	int i;

1100
	new_active = bond->curr_active_slave;
L
Linus Torvalds 已提交
1101 1102

	if (!new_active) { /* there were no active slaves left */
S
Stephen Hemminger 已提交
1103
		if (bond->slave_cnt > 0)   /* found one slave */
L
Linus Torvalds 已提交
1104
			new_active = bond->first_slave;
S
Stephen Hemminger 已提交
1105
		else
L
Linus Torvalds 已提交
1106 1107 1108 1109
			return NULL; /* still no slave, return NULL */
	}

	if ((bond->primary_slave) &&
1110 1111
	    bond->primary_slave->link == BOND_LINK_UP &&
	    bond_should_change_active(bond)) {
L
Linus Torvalds 已提交
1112 1113 1114 1115 1116 1117 1118
		new_active = bond->primary_slave;
	}

	/* remember where to stop iterating over the slaves */
	old_active = new_active;

	bond_for_each_slave_from(bond, new_active, i, old_active) {
1119 1120 1121 1122 1123 1124 1125 1126
		if (new_active->link == BOND_LINK_UP) {
			return new_active;
		} else if (new_active->link == BOND_LINK_BACK &&
			   IS_UP(new_active->dev)) {
			/* link up, but waiting for stabilization */
			if (new_active->delay < mintime) {
				mintime = new_active->delay;
				bestslave = new_active;
L
Linus Torvalds 已提交
1127 1128 1129 1130 1131 1132 1133 1134 1135 1136 1137 1138 1139 1140 1141 1142 1143 1144 1145 1146
			}
		}
	}

	return bestslave;
}

/**
 * change_active_interface - change the active slave into the specified one
 * @bond: our bonding struct
 * @new: the new slave to make the active one
 *
 * Set the new slave to the bond's settings and unset them on the old
 * curr_active_slave.
 * Setting include flags, mc-list, promiscuity, allmulti, etc.
 *
 * If @new's link state is %BOND_LINK_BACK we'll set it to %BOND_LINK_UP,
 * because it is apparently the best available slave we have, even though its
 * updelay hasn't timed out yet.
 *
1147 1148
 * If new_active is not NULL, caller must hold bond->lock for read and
 * curr_slave_lock for write_bh.
L
Linus Torvalds 已提交
1149
 */
1150
void bond_change_active_slave(struct bonding *bond, struct slave *new_active)
L
Linus Torvalds 已提交
1151 1152 1153
{
	struct slave *old_active = bond->curr_active_slave;

S
Stephen Hemminger 已提交
1154
	if (old_active == new_active)
L
Linus Torvalds 已提交
1155 1156 1157
		return;

	if (new_active) {
1158 1159
		new_active->jiffies = jiffies;

L
Linus Torvalds 已提交
1160 1161
		if (new_active->link == BOND_LINK_BACK) {
			if (USES_PRIMARY(bond->params.mode)) {
J
Joe Perches 已提交
1162 1163 1164
				pr_info("%s: making interface %s the new active one %d ms earlier.\n",
					bond->dev->name, new_active->dev->name,
					(bond->params.updelay - new_active->delay) * bond->params.miimon);
L
Linus Torvalds 已提交
1165 1166 1167 1168 1169
			}

			new_active->delay = 0;
			new_active->link = BOND_LINK_UP;

S
Stephen Hemminger 已提交
1170
			if (bond->params.mode == BOND_MODE_8023AD)
L
Linus Torvalds 已提交
1171 1172
				bond_3ad_handle_link_change(new_active, BOND_LINK_UP);

1173
			if (bond_is_lb(bond))
L
Linus Torvalds 已提交
1174 1175 1176
				bond_alb_handle_link_change(bond, new_active, BOND_LINK_UP);
		} else {
			if (USES_PRIMARY(bond->params.mode)) {
J
Joe Perches 已提交
1177 1178
				pr_info("%s: making interface %s the new active one.\n",
					bond->dev->name, new_active->dev->name);
L
Linus Torvalds 已提交
1179 1180 1181 1182
			}
		}
	}

S
Stephen Hemminger 已提交
1183
	if (USES_PRIMARY(bond->params.mode))
L
Linus Torvalds 已提交
1184 1185
		bond_mc_swap(bond, new_active, old_active);

1186
	if (bond_is_lb(bond)) {
L
Linus Torvalds 已提交
1187
		bond_alb_handle_active_change(bond, new_active);
1188 1189 1190 1191
		if (old_active)
			bond_set_slave_inactive_flags(old_active);
		if (new_active)
			bond_set_slave_active_flags(new_active);
L
Linus Torvalds 已提交
1192 1193 1194
	} else {
		bond->curr_active_slave = new_active;
	}
J
Jay Vosburgh 已提交
1195 1196

	if (bond->params.mode == BOND_MODE_ACTIVEBACKUP) {
S
Stephen Hemminger 已提交
1197
		if (old_active)
J
Jay Vosburgh 已提交
1198 1199 1200 1201
			bond_set_slave_inactive_flags(old_active);

		if (new_active) {
			bond_set_slave_active_flags(new_active);
1202

1203 1204 1205
			if (bond->params.fail_over_mac)
				bond_do_fail_over_mac(bond, new_active,
						      old_active);
1206

1207
			bond->send_grat_arp = bond->params.num_grat_arp;
1208
			bond_send_gratuitous_arp(bond);
1209

1210 1211 1212
			bond->send_unsol_na = bond->params.num_unsol_na;
			bond_send_unsolicited_na(bond);

1213 1214 1215
			write_unlock_bh(&bond->curr_slave_lock);
			read_unlock(&bond->lock);

1216
			netdev_bonding_change(bond->dev, NETDEV_BONDING_FAILOVER);
1217 1218 1219

			read_lock(&bond->lock);
			write_lock_bh(&bond->curr_slave_lock);
1220
		}
J
Jay Vosburgh 已提交
1221
	}
1222

1223 1224 1225 1226
	/* resend IGMP joins since active slave has changed or
	 * all were sent on curr_active_slave */
	if ((USES_PRIMARY(bond->params.mode) && new_active) ||
	    bond->params.mode == BOND_MODE_ROUNDROBIN) {
1227
		bond->igmp_retrans = bond->params.resend_igmp;
1228
		queue_delayed_work(bond->wq, &bond->mcast_work, 0);
1229
	}
L
Linus Torvalds 已提交
1230 1231 1232 1233 1234 1235
}

/**
 * bond_select_active_slave - select a new active slave, if needed
 * @bond: our bonding struct
 *
S
Stephen Hemminger 已提交
1236
 * This functions should be called when one of the following occurs:
L
Linus Torvalds 已提交
1237 1238 1239 1240
 * - The old curr_active_slave has been released or lost its link.
 * - The primary_slave has got its link back.
 * - A slave has got its link back and there's no old curr_active_slave.
 *
1241
 * Caller must hold bond->lock for read and curr_slave_lock for write_bh.
L
Linus Torvalds 已提交
1242
 */
1243
void bond_select_active_slave(struct bonding *bond)
L
Linus Torvalds 已提交
1244 1245
{
	struct slave *best_slave;
1246
	int rv;
L
Linus Torvalds 已提交
1247 1248 1249 1250

	best_slave = bond_find_best_slave(bond);
	if (best_slave != bond->curr_active_slave) {
		bond_change_active_slave(bond, best_slave);
1251 1252 1253 1254 1255
		rv = bond_set_carrier(bond);
		if (!rv)
			return;

		if (netif_carrier_ok(bond->dev)) {
J
Joe Perches 已提交
1256 1257
			pr_info("%s: first active interface up!\n",
				bond->dev->name);
1258
		} else {
J
Joe Perches 已提交
1259 1260
			pr_info("%s: now running without any active interface !\n",
				bond->dev->name);
1261
		}
L
Linus Torvalds 已提交
1262 1263 1264 1265 1266 1267 1268 1269 1270 1271 1272 1273 1274 1275 1276 1277 1278 1279 1280 1281 1282 1283 1284 1285 1286 1287 1288 1289 1290 1291 1292 1293 1294 1295 1296 1297 1298 1299
	}
}

/*--------------------------- slave list handling ---------------------------*/

/*
 * This function attaches the slave to the end of list.
 *
 * bond->lock held for writing by caller.
 */
static void bond_attach_slave(struct bonding *bond, struct slave *new_slave)
{
	if (bond->first_slave == NULL) { /* attaching the first slave */
		new_slave->next = new_slave;
		new_slave->prev = new_slave;
		bond->first_slave = new_slave;
	} else {
		new_slave->next = bond->first_slave;
		new_slave->prev = bond->first_slave->prev;
		new_slave->next->prev = new_slave;
		new_slave->prev->next = new_slave;
	}

	bond->slave_cnt++;
}

/*
 * This function detaches the slave from the list.
 * WARNING: no check is made to verify if the slave effectively
 * belongs to <bond>.
 * Nothing is freed on return, structures are just unchained.
 * If any slave pointer in bond was pointing to <slave>,
 * it should be changed by the calling function.
 *
 * bond->lock held for writing by caller.
 */
static void bond_detach_slave(struct bonding *bond, struct slave *slave)
{
S
Stephen Hemminger 已提交
1300
	if (slave->next)
L
Linus Torvalds 已提交
1301 1302
		slave->next->prev = slave->prev;

S
Stephen Hemminger 已提交
1303
	if (slave->prev)
L
Linus Torvalds 已提交
1304 1305 1306 1307 1308 1309 1310 1311 1312 1313 1314 1315 1316 1317 1318
		slave->prev->next = slave->next;

	if (bond->first_slave == slave) { /* slave is the first slave */
		if (bond->slave_cnt > 1) { /* there are more slave */
			bond->first_slave = slave->next;
		} else {
			bond->first_slave = NULL; /* slave was the last one */
		}
	}

	slave->next = NULL;
	slave->prev = NULL;
	bond->slave_cnt--;
}

1319 1320 1321 1322 1323 1324 1325 1326 1327 1328 1329 1330 1331 1332 1333 1334 1335 1336 1337 1338 1339
#ifdef CONFIG_NET_POLL_CONTROLLER
/*
 * You must hold read lock on bond->lock before calling this.
 */
static bool slaves_support_netpoll(struct net_device *bond_dev)
{
	struct bonding *bond = netdev_priv(bond_dev);
	struct slave *slave;
	int i = 0;
	bool ret = true;

	bond_for_each_slave(bond, slave, i) {
		if ((slave->dev->priv_flags & IFF_DISABLE_NETPOLL) ||
		    !slave->dev->netdev_ops->ndo_poll_controller)
			ret = false;
	}
	return i != 0 && ret;
}

static void bond_poll_controller(struct net_device *bond_dev)
{
1340 1341 1342 1343 1344 1345 1346 1347
	struct bonding *bond = netdev_priv(bond_dev);
	struct slave *slave;
	int i;

	bond_for_each_slave(bond, slave, i) {
		if (slave->dev && IS_UP(slave->dev))
			netpoll_poll_dev(slave->dev);
	}
1348 1349 1350 1351 1352 1353 1354 1355 1356 1357 1358 1359 1360 1361 1362 1363 1364 1365 1366 1367 1368 1369 1370 1371 1372 1373 1374 1375 1376 1377 1378
}

static void bond_netpoll_cleanup(struct net_device *bond_dev)
{
	struct bonding *bond = netdev_priv(bond_dev);
	struct slave *slave;
	const struct net_device_ops *ops;
	int i;

	read_lock(&bond->lock);
	bond_dev->npinfo = NULL;
	bond_for_each_slave(bond, slave, i) {
		if (slave->dev) {
			ops = slave->dev->netdev_ops;
			if (ops->ndo_netpoll_cleanup)
				ops->ndo_netpoll_cleanup(slave->dev);
			else
				slave->dev->npinfo = NULL;
		}
	}
	read_unlock(&bond->lock);
}

#else

static void bond_netpoll_cleanup(struct net_device *bond_dev)
{
}

#endif

L
Linus Torvalds 已提交
1379 1380
/*---------------------------------- IOCTL ----------------------------------*/

1381 1382
static int bond_sethwaddr(struct net_device *bond_dev,
			  struct net_device *slave_dev)
L
Linus Torvalds 已提交
1383
{
1384 1385 1386
	pr_debug("bond_dev=%p\n", bond_dev);
	pr_debug("slave_dev=%p\n", slave_dev);
	pr_debug("slave_dev->addr_len=%d\n", slave_dev->addr_len);
L
Linus Torvalds 已提交
1387 1388 1389 1390
	memcpy(bond_dev->dev_addr, slave_dev->dev_addr, slave_dev->addr_len);
	return 0;
}

1391 1392 1393
#define BOND_VLAN_FEATURES \
	(NETIF_F_VLAN_CHALLENGED | NETIF_F_HW_VLAN_RX | NETIF_F_HW_VLAN_TX | \
	 NETIF_F_HW_VLAN_FILTER)
1394

S
Stephen Hemminger 已提交
1395
/*
1396
 * Compute the common dev->feature set available to all slaves.  Some
1397 1398
 * feature bits are managed elsewhere, so preserve those feature bits
 * on the master device.
1399 1400 1401 1402 1403
 */
static int bond_compute_features(struct bonding *bond)
{
	struct slave *slave;
	struct net_device *bond_dev = bond->dev;
1404
	unsigned long features = bond_dev->features;
1405
	unsigned long vlan_features = 0;
1406 1407
	unsigned short max_hard_header_len = max((u16)ETH_HLEN,
						bond_dev->hard_header_len);
1408
	int i;
1409

1410
	features &= ~(NETIF_F_ALL_CSUM | BOND_VLAN_FEATURES);
1411 1412 1413 1414 1415 1416
	features |=  NETIF_F_GSO_MASK | NETIF_F_NO_CSUM;

	if (!bond->first_slave)
		goto done;

	features &= ~NETIF_F_ONE_FOR_ALL;
1417

1418
	vlan_features = bond->first_slave->dev->vlan_features;
1419
	bond_for_each_slave(bond, slave, i) {
1420 1421 1422
		features = netdev_increment_features(features,
						     slave->dev->features,
						     NETIF_F_ONE_FOR_ALL);
1423 1424 1425
		vlan_features = netdev_increment_features(vlan_features,
							slave->dev->vlan_features,
							NETIF_F_ONE_FOR_ALL);
1426 1427 1428
		if (slave->dev->hard_header_len > max_hard_header_len)
			max_hard_header_len = slave->dev->hard_header_len;
	}
1429

1430
done:
1431
	features |= (bond_dev->features & BOND_VLAN_FEATURES);
1432
	bond_dev->features = netdev_fix_features(features, NULL);
1433
	bond_dev->vlan_features = netdev_fix_features(vlan_features, NULL);
1434
	bond_dev->hard_header_len = max_hard_header_len;
1435 1436 1437 1438

	return 0;
}

1439 1440 1441
static void bond_setup_by_slave(struct net_device *bond_dev,
				struct net_device *slave_dev)
{
1442
	struct bonding *bond = netdev_priv(bond_dev);
1443

1444
	bond_dev->header_ops	    = slave_dev->header_ops;
1445 1446 1447 1448 1449 1450 1451

	bond_dev->type		    = slave_dev->type;
	bond_dev->hard_header_len   = slave_dev->hard_header_len;
	bond_dev->addr_len	    = slave_dev->addr_len;

	memcpy(bond_dev->broadcast, slave_dev->broadcast,
		slave_dev->addr_len);
1452
	bond->setup_by_slave = 1;
1453 1454
}

L
Linus Torvalds 已提交
1455
/* enslave device <slave> to bond device <master> */
1456
int bond_enslave(struct net_device *bond_dev, struct net_device *slave_dev)
L
Linus Torvalds 已提交
1457
{
1458
	struct bonding *bond = netdev_priv(bond_dev);
1459
	const struct net_device_ops *slave_ops = slave_dev->netdev_ops;
L
Linus Torvalds 已提交
1460
	struct slave *new_slave = NULL;
1461
	struct netdev_hw_addr *ha;
L
Linus Torvalds 已提交
1462 1463 1464 1465 1466
	struct sockaddr addr;
	int link_reporting;
	int old_features = bond_dev->features;
	int res = 0;

1467
	if (!bond->params.use_carrier && slave_dev->ethtool_ops == NULL &&
1468
		slave_ops->ndo_do_ioctl == NULL) {
J
Joe Perches 已提交
1469 1470
		pr_warning("%s: Warning: no link monitoring support for %s\n",
			   bond_dev->name, slave_dev->name);
L
Linus Torvalds 已提交
1471 1472 1473 1474
	}

	/* bond must be initialized by bond_open() before enslaving */
	if (!(bond_dev->flags & IFF_UP)) {
J
Joe Perches 已提交
1475 1476
		pr_warning("%s: master_dev is not up in bond_enslave\n",
			   bond_dev->name);
L
Linus Torvalds 已提交
1477 1478 1479 1480
	}

	/* already enslaved */
	if (slave_dev->flags & IFF_SLAVE) {
1481
		pr_debug("Error, Device was already enslaved\n");
L
Linus Torvalds 已提交
1482 1483 1484 1485 1486 1487
		return -EBUSY;
	}

	/* vlan challenged mutual exclusion */
	/* no need to lock since we're protected by rtnl_lock */
	if (slave_dev->features & NETIF_F_VLAN_CHALLENGED) {
1488
		pr_debug("%s: NETIF_F_VLAN_CHALLENGED\n", slave_dev->name);
1489
		if (bond->vlgrp) {
J
Joe Perches 已提交
1490 1491
			pr_err("%s: Error: cannot enslave VLAN challenged slave %s on VLAN enabled bond %s\n",
			       bond_dev->name, slave_dev->name, bond_dev->name);
L
Linus Torvalds 已提交
1492 1493
			return -EPERM;
		} else {
J
Joe Perches 已提交
1494 1495 1496
			pr_warning("%s: Warning: enslaved VLAN challenged slave %s. Adding VLANs will be blocked as long as %s is part of bond %s\n",
				   bond_dev->name, slave_dev->name,
				   slave_dev->name, bond_dev->name);
L
Linus Torvalds 已提交
1497 1498 1499
			bond_dev->features |= NETIF_F_VLAN_CHALLENGED;
		}
	} else {
1500
		pr_debug("%s: ! NETIF_F_VLAN_CHALLENGED\n", slave_dev->name);
L
Linus Torvalds 已提交
1501 1502 1503 1504 1505 1506 1507 1508
		if (bond->slave_cnt == 0) {
			/* First slave, and it is not VLAN challenged,
			 * so remove the block of adding VLANs over the bond.
			 */
			bond_dev->features &= ~NETIF_F_VLAN_CHALLENGED;
		}
	}

1509 1510
	/*
	 * Old ifenslave binaries are no longer supported.  These can
S
Stephen Hemminger 已提交
1511
	 * be identified with moderate accuracy by the state of the slave:
1512 1513 1514 1515
	 * the current ifenslave will set the interface down prior to
	 * enslaving it; the old ifenslave will not.
	 */
	if ((slave_dev->flags & IFF_UP)) {
J
Joe Perches 已提交
1516
		pr_err("%s is up. This may be due to an out of date ifenslave.\n",
1517 1518 1519 1520
		       slave_dev->name);
		res = -EPERM;
		goto err_undo_flags;
	}
L
Linus Torvalds 已提交
1521

1522 1523 1524 1525 1526 1527 1528 1529
	/* set bonding device ether type by slave - bonding netdevices are
	 * created with ether_setup, so when the slave type is not ARPHRD_ETHER
	 * there is a need to override some of the type dependent attribs/funcs.
	 *
	 * bond ether type mutual exclusion - don't allow slaves of dissimilar
	 * ether type (eg ARPHRD_ETHER and ARPHRD_INFINIBAND) share the same bond
	 */
	if (bond->slave_cnt == 0) {
1530 1531
		if (bond_dev->type != slave_dev->type) {
			pr_debug("%s: change device type from %d to %d\n",
J
Joe Perches 已提交
1532 1533
				 bond_dev->name,
				 bond_dev->type, slave_dev->type);
1534

1535 1536 1537 1538 1539 1540 1541 1542 1543
			res = netdev_bonding_change(bond_dev,
						    NETDEV_PRE_TYPE_CHANGE);
			res = notifier_to_errno(res);
			if (res) {
				pr_err("%s: refused to change device type\n",
				       bond_dev->name);
				res = -EBUSY;
				goto err_undo_flags;
			}
1544

1545
			/* Flush unicast and multicast addresses */
1546
			dev_uc_flush(bond_dev);
1547
			dev_mc_flush(bond_dev);
1548

1549 1550 1551 1552
			if (slave_dev->type != ARPHRD_ETHER)
				bond_setup_by_slave(bond_dev, slave_dev);
			else
				ether_setup(bond_dev);
1553

1554 1555
			netdev_bonding_change(bond_dev,
					      NETDEV_POST_TYPE_CHANGE);
1556
		}
1557
	} else if (bond_dev->type != slave_dev->type) {
J
Joe Perches 已提交
1558 1559 1560 1561 1562
		pr_err("%s ether type (%d) is different from other slaves (%d), can not enslave it.\n",
		       slave_dev->name,
		       slave_dev->type, bond_dev->type);
		res = -EINVAL;
		goto err_undo_flags;
1563 1564
	}

1565
	if (slave_ops->ndo_set_mac_address == NULL) {
1566
		if (bond->slave_cnt == 0) {
J
Joe Perches 已提交
1567 1568
			pr_warning("%s: Warning: The first slave device specified does not support setting the MAC address. Setting fail_over_mac to active.",
				   bond_dev->name);
1569 1570
			bond->params.fail_over_mac = BOND_FOM_ACTIVE;
		} else if (bond->params.fail_over_mac != BOND_FOM_ACTIVE) {
J
Joe Perches 已提交
1571 1572
			pr_err("%s: Error: The slave device specified does not support setting the MAC address, but fail_over_mac is not set to active.\n",
			       bond_dev->name);
1573 1574 1575
			res = -EOPNOTSUPP;
			goto err_undo_flags;
		}
L
Linus Torvalds 已提交
1576 1577
	}

1578 1579 1580 1581 1582 1583 1584
	/* If this is the first slave, then we need to set the master's hardware
	 * address to be the same as the slave's. */
	if (bond->slave_cnt == 0)
		memcpy(bond->dev->dev_addr, slave_dev->dev_addr,
		       slave_dev->addr_len);


1585
	new_slave = kzalloc(sizeof(struct slave), GFP_KERNEL);
L
Linus Torvalds 已提交
1586 1587 1588 1589 1590
	if (!new_slave) {
		res = -ENOMEM;
		goto err_undo_flags;
	}

1591 1592 1593 1594 1595 1596
	/*
	 * Set the new_slave's queue_id to be zero.  Queue ID mapping
	 * is set via sysfs or module option if desired.
	 */
	new_slave->queue_id = 0;

1597 1598 1599 1600 1601 1602 1603 1604
	/* Save slave's original mtu and then set it to match the bond */
	new_slave->original_mtu = slave_dev->mtu;
	res = dev_set_mtu(slave_dev, bond->dev->mtu);
	if (res) {
		pr_debug("Error %d calling dev_set_mtu\n", res);
		goto err_free;
	}

1605 1606 1607 1608 1609 1610
	/*
	 * Save slave's original ("permanent") mac address for modes
	 * that need it, and for restoring it upon release, and then
	 * set it to the master's address
	 */
	memcpy(new_slave->perm_hwaddr, slave_dev->dev_addr, ETH_ALEN);
L
Linus Torvalds 已提交
1611

1612
	if (!bond->params.fail_over_mac) {
1613 1614 1615 1616 1617 1618 1619 1620
		/*
		 * Set slave to master's mac address.  The application already
		 * set the master's mac address to that of the first slave
		 */
		memcpy(addr.sa_data, bond_dev->dev_addr, bond_dev->addr_len);
		addr.sa_family = slave_dev->type;
		res = dev_set_mac_address(slave_dev, &addr);
		if (res) {
1621
			pr_debug("Error %d calling set_mac_address\n", res);
1622
			goto err_restore_mtu;
1623
		}
1624
	}
L
Linus Torvalds 已提交
1625

1626 1627
	res = netdev_set_master(slave_dev, bond_dev);
	if (res) {
1628
		pr_debug("Error %d calling netdev_set_master\n", res);
1629
		goto err_restore_mac;
1630
	}
1631 1632 1633
	/* open the slave since the application closed it */
	res = dev_open(slave_dev);
	if (res) {
S
Stephen Hemminger 已提交
1634
		pr_debug("Opening slave %s failed\n", slave_dev->name);
1635
		goto err_unset_master;
L
Linus Torvalds 已提交
1636 1637 1638
	}

	new_slave->dev = slave_dev;
1639
	slave_dev->priv_flags |= IFF_BONDING;
L
Linus Torvalds 已提交
1640

1641
	if (bond_is_lb(bond)) {
L
Linus Torvalds 已提交
1642 1643 1644 1645
		/* bond_alb_init_slave() must be called before all other stages since
		 * it might fail and we do not want to have to undo everything
		 */
		res = bond_alb_init_slave(bond, new_slave);
S
Stephen Hemminger 已提交
1646
		if (res)
1647
			goto err_close;
L
Linus Torvalds 已提交
1648 1649 1650 1651 1652 1653 1654 1655 1656 1657
	}

	/* If the mode USES_PRIMARY, then the new slave gets the
	 * master's promisc (and mc) settings only if it becomes the
	 * curr_active_slave, and that is taken care of later when calling
	 * bond_change_active()
	 */
	if (!USES_PRIMARY(bond->params.mode)) {
		/* set promiscuity level to new slave */
		if (bond_dev->flags & IFF_PROMISC) {
1658 1659 1660
			res = dev_set_promiscuity(slave_dev, 1);
			if (res)
				goto err_close;
L
Linus Torvalds 已提交
1661 1662 1663 1664
		}

		/* set allmulti level to new slave */
		if (bond_dev->flags & IFF_ALLMULTI) {
1665 1666 1667
			res = dev_set_allmulti(slave_dev, 1);
			if (res)
				goto err_close;
L
Linus Torvalds 已提交
1668 1669
		}

1670
		netif_addr_lock_bh(bond_dev);
L
Linus Torvalds 已提交
1671
		/* upload master's mc_list to new slave */
1672 1673
		netdev_for_each_mc_addr(ha, bond_dev)
			dev_mc_add(slave_dev, ha->addr);
1674
		netif_addr_unlock_bh(bond_dev);
L
Linus Torvalds 已提交
1675 1676 1677 1678 1679 1680
	}

	if (bond->params.mode == BOND_MODE_8023AD) {
		/* add lacpdu mc addr to mc list */
		u8 lacpdu_multicast[ETH_ALEN] = MULTICAST_LACPDU_ADDR;

1681
		dev_mc_add(slave_dev, lacpdu_multicast);
L
Linus Torvalds 已提交
1682 1683 1684 1685 1686 1687 1688 1689 1690 1691 1692
	}

	bond_add_vlans_on_slave(bond, slave_dev);

	write_lock_bh(&bond->lock);

	bond_attach_slave(bond, new_slave);

	new_slave->delay = 0;
	new_slave->link_failure_count = 0;

1693 1694
	bond_compute_features(bond);

1695 1696 1697 1698
	write_unlock_bh(&bond->lock);

	read_lock(&bond->lock);

1699 1700
	new_slave->last_arp_rx = jiffies;

L
Linus Torvalds 已提交
1701 1702 1703 1704 1705 1706 1707 1708 1709 1710 1711 1712 1713
	if (bond->params.miimon && !bond->params.use_carrier) {
		link_reporting = bond_check_dev_link(bond, slave_dev, 1);

		if ((link_reporting == -1) && !bond->params.arp_interval) {
			/*
			 * miimon is set but a bonded network driver
			 * does not support ETHTOOL/MII and
			 * arp_interval is not set.  Note: if
			 * use_carrier is enabled, we will never go
			 * here (because netif_carrier is always
			 * supported); thus, we don't need to change
			 * the messages for netif_carrier.
			 */
J
Joe Perches 已提交
1714
			pr_warning("%s: Warning: MII and ETHTOOL support not available for interface %s, and arp_interval/arp_ip_target module parameters not specified, thus bonding will not detect link failures! see bonding.txt for details.\n",
1715
			       bond_dev->name, slave_dev->name);
L
Linus Torvalds 已提交
1716 1717
		} else if (link_reporting == -1) {
			/* unable get link status using mii/ethtool */
J
Joe Perches 已提交
1718 1719
			pr_warning("%s: Warning: can't get link status from interface %s; the network driver associated with this interface does not support MII or ETHTOOL link status reporting, thus miimon has no effect on this interface.\n",
				   bond_dev->name, slave_dev->name);
L
Linus Torvalds 已提交
1720 1721 1722 1723 1724 1725 1726
		}
	}

	/* check for initial state */
	if (!bond->params.miimon ||
	    (bond_check_dev_link(bond, slave_dev, 0) == BMSR_LSTATUS)) {
		if (bond->params.updelay) {
J
Joe Perches 已提交
1727
			pr_debug("Initial state of slave_dev is BOND_LINK_BACK\n");
L
Linus Torvalds 已提交
1728 1729 1730
			new_slave->link  = BOND_LINK_BACK;
			new_slave->delay = bond->params.updelay;
		} else {
J
Joe Perches 已提交
1731
			pr_debug("Initial state of slave_dev is BOND_LINK_UP\n");
L
Linus Torvalds 已提交
1732 1733 1734 1735
			new_slave->link  = BOND_LINK_UP;
		}
		new_slave->jiffies = jiffies;
	} else {
J
Joe Perches 已提交
1736
		pr_debug("Initial state of slave_dev is BOND_LINK_DOWN\n");
L
Linus Torvalds 已提交
1737 1738 1739 1740 1741
		new_slave->link  = BOND_LINK_DOWN;
	}

	if (bond_update_speed_duplex(new_slave) &&
	    (new_slave->link != BOND_LINK_DOWN)) {
J
Joe Perches 已提交
1742 1743
		pr_warning("%s: Warning: failed to get speed and duplex from %s, assumed to be 100Mb/sec and Full.\n",
			   bond_dev->name, new_slave->dev->name);
L
Linus Torvalds 已提交
1744 1745

		if (bond->params.mode == BOND_MODE_8023AD) {
J
Joe Perches 已提交
1746 1747
			pr_warning("%s: Warning: Operation of 802.3ad mode requires ETHTOOL support in base driver for proper aggregator selection.\n",
				   bond_dev->name);
L
Linus Torvalds 已提交
1748 1749 1750 1751 1752
		}
	}

	if (USES_PRIMARY(bond->params.mode) && bond->params.primary[0]) {
		/* if there is a primary slave, remember it */
1753
		if (strcmp(bond->params.primary, new_slave->dev->name) == 0) {
L
Linus Torvalds 已提交
1754
			bond->primary_slave = new_slave;
1755 1756
			bond->force_primary = true;
		}
L
Linus Torvalds 已提交
1757 1758
	}

1759 1760
	write_lock_bh(&bond->curr_slave_lock);

L
Linus Torvalds 已提交
1761 1762
	switch (bond->params.mode) {
	case BOND_MODE_ACTIVEBACKUP:
1763 1764
		bond_set_slave_inactive_flags(new_slave);
		bond_select_active_slave(bond);
L
Linus Torvalds 已提交
1765 1766 1767 1768 1769 1770 1771 1772 1773 1774 1775 1776 1777 1778 1779 1780 1781 1782 1783 1784 1785 1786 1787 1788 1789
		break;
	case BOND_MODE_8023AD:
		/* in 802.3ad mode, the internal mechanism
		 * will activate the slaves in the selected
		 * aggregator
		 */
		bond_set_slave_inactive_flags(new_slave);
		/* if this is the first slave */
		if (bond->slave_cnt == 1) {
			SLAVE_AD_INFO(new_slave).id = 1;
			/* Initialize AD with the number of times that the AD timer is called in 1 second
			 * can be called only after the mac address of the bond is set
			 */
			bond_3ad_initialize(bond, 1000/AD_TIMER_INTERVAL,
					    bond->params.lacp_fast);
		} else {
			SLAVE_AD_INFO(new_slave).id =
				SLAVE_AD_INFO(new_slave->prev).id + 1;
		}

		bond_3ad_bind_slave(new_slave);
		break;
	case BOND_MODE_TLB:
	case BOND_MODE_ALB:
		new_slave->state = BOND_STATE_ACTIVE;
1790
		bond_set_slave_inactive_flags(new_slave);
1791
		bond_select_active_slave(bond);
L
Linus Torvalds 已提交
1792 1793
		break;
	default:
1794
		pr_debug("This slave is always active in trunk mode\n");
L
Linus Torvalds 已提交
1795 1796 1797 1798 1799 1800 1801 1802

		/* always active in trunk mode */
		new_slave->state = BOND_STATE_ACTIVE;

		/* In trunking mode there is little meaning to curr_active_slave
		 * anyway (it holds no special properties of the bond device),
		 * so we can change it without calling change_active_interface()
		 */
S
Stephen Hemminger 已提交
1803
		if (!bond->curr_active_slave)
L
Linus Torvalds 已提交
1804
			bond->curr_active_slave = new_slave;
S
Stephen Hemminger 已提交
1805

L
Linus Torvalds 已提交
1806 1807 1808
		break;
	} /* switch(bond_mode) */

1809 1810
	write_unlock_bh(&bond->curr_slave_lock);

1811 1812
	bond_set_carrier(bond);

1813
#ifdef CONFIG_NET_POLL_CONTROLLER
1814
	if (disable_netpoll) {
1815
		bond_dev->priv_flags |= IFF_DISABLE_NETPOLL;
1816 1817 1818 1819 1820 1821 1822 1823 1824 1825 1826
	} else {
		if (slaves_support_netpoll(bond_dev)) {
			bond_dev->priv_flags &= ~IFF_DISABLE_NETPOLL;
			if (bond_dev->npinfo)
				slave_dev->npinfo = bond_dev->npinfo;
		} else if (!(bond_dev->priv_flags & IFF_DISABLE_NETPOLL)) {
			bond_dev->priv_flags |= IFF_DISABLE_NETPOLL;
			pr_info("New slave device %s does not support netpoll\n",
				slave_dev->name);
			pr_info("Disabling netpoll support for %s\n", bond_dev->name);
		}
1827 1828
	}
#endif
1829
	read_unlock(&bond->lock);
L
Linus Torvalds 已提交
1830

1831 1832
	res = bond_create_slave_symlinks(bond_dev, slave_dev);
	if (res)
1833
		goto err_close;
1834

J
Joe Perches 已提交
1835 1836 1837 1838
	pr_info("%s: enslaving %s as a%s interface with a%s link.\n",
		bond_dev->name, slave_dev->name,
		new_slave->state == BOND_STATE_ACTIVE ? "n active" : " backup",
		new_slave->link != BOND_LINK_DOWN ? "n up" : " down");
L
Linus Torvalds 已提交
1839 1840 1841 1842 1843 1844 1845 1846

	/* enslave is successful */
	return 0;

/* Undo stages on error */
err_close:
	dev_close(slave_dev);

1847 1848 1849
err_unset_master:
	netdev_set_master(slave_dev, NULL);

L
Linus Torvalds 已提交
1850
err_restore_mac:
1851
	if (!bond->params.fail_over_mac) {
1852 1853 1854 1855
		/* XXX TODO - fom follow mode needs to change master's
		 * MAC if this slave's MAC is in use by the bond, or at
		 * least print a warning.
		 */
1856 1857 1858 1859
		memcpy(addr.sa_data, new_slave->perm_hwaddr, ETH_ALEN);
		addr.sa_family = slave_dev->type;
		dev_set_mac_address(slave_dev, &addr);
	}
L
Linus Torvalds 已提交
1860

1861 1862 1863
err_restore_mtu:
	dev_set_mtu(slave_dev, new_slave->original_mtu);

L
Linus Torvalds 已提交
1864 1865 1866 1867 1868
err_free:
	kfree(new_slave);

err_undo_flags:
	bond_dev->features = old_features;
S
Stephen Hemminger 已提交
1869

L
Linus Torvalds 已提交
1870 1871 1872 1873 1874 1875 1876 1877 1878 1879 1880 1881 1882 1883
	return res;
}

/*
 * Try to release the slave device <slave> from the bond device <master>
 * It is legal to access curr_active_slave without a lock because all the function
 * is write-locked.
 *
 * The rules for slave state should be:
 *   for Active/Backup:
 *     Active stays on all backups go down
 *   for Bonded connections:
 *     The first up interface should be left on and all others downed.
 */
1884
int bond_release(struct net_device *bond_dev, struct net_device *slave_dev)
L
Linus Torvalds 已提交
1885
{
1886
	struct bonding *bond = netdev_priv(bond_dev);
L
Linus Torvalds 已提交
1887 1888 1889 1890 1891 1892
	struct slave *slave, *oldcurrent;
	struct sockaddr addr;

	/* slave is not a slave or master is not master of this slave */
	if (!(slave_dev->flags & IFF_SLAVE) ||
	    (slave_dev->master != bond_dev)) {
J
Joe Perches 已提交
1893
		pr_err("%s: Error: cannot release %s.\n",
L
Linus Torvalds 已提交
1894 1895 1896 1897
		       bond_dev->name, slave_dev->name);
		return -EINVAL;
	}

1898
	block_netpoll_tx();
1899
	netdev_bonding_change(bond_dev, NETDEV_BONDING_DESLAVE);
L
Linus Torvalds 已提交
1900 1901 1902 1903 1904
	write_lock_bh(&bond->lock);

	slave = bond_get_slave_by_dev(bond, slave_dev);
	if (!slave) {
		/* not a slave of this bond */
J
Joe Perches 已提交
1905 1906
		pr_info("%s: %s not enslaved\n",
			bond_dev->name, slave_dev->name);
1907
		write_unlock_bh(&bond->lock);
1908
		unblock_netpoll_tx();
L
Linus Torvalds 已提交
1909 1910 1911
		return -EINVAL;
	}

1912
	if (!bond->params.fail_over_mac) {
1913 1914
		if (!compare_ether_addr(bond_dev->dev_addr, slave->perm_hwaddr) &&
		    bond->slave_cnt > 1)
J
Joe Perches 已提交
1915 1916 1917 1918
			pr_warning("%s: Warning: the permanent HWaddr of %s - %pM - is still in use by %s. Set the HWaddr of %s to a different address to avoid conflicts.\n",
				   bond_dev->name, slave_dev->name,
				   slave->perm_hwaddr,
				   bond_dev->name, slave_dev->name);
L
Linus Torvalds 已提交
1919 1920 1921 1922 1923 1924 1925 1926 1927 1928
	}

	/* Inform AD package of unbinding of slave. */
	if (bond->params.mode == BOND_MODE_8023AD) {
		/* must be called before the slave is
		 * detached from the list
		 */
		bond_3ad_unbind_slave(slave);
	}

J
Joe Perches 已提交
1929 1930 1931 1932
	pr_info("%s: releasing %s interface %s\n",
		bond_dev->name,
		(slave->state == BOND_STATE_ACTIVE) ? "active" : "backup",
		slave_dev->name);
L
Linus Torvalds 已提交
1933 1934 1935 1936 1937 1938 1939 1940

	oldcurrent = bond->curr_active_slave;

	bond->current_arp_slave = NULL;

	/* release the slave from its bond */
	bond_detach_slave(bond, slave);

1941 1942
	bond_compute_features(bond);

S
Stephen Hemminger 已提交
1943
	if (bond->primary_slave == slave)
L
Linus Torvalds 已提交
1944 1945
		bond->primary_slave = NULL;

S
Stephen Hemminger 已提交
1946
	if (oldcurrent == slave)
L
Linus Torvalds 已提交
1947 1948
		bond_change_active_slave(bond, NULL);

1949
	if (bond_is_lb(bond)) {
L
Linus Torvalds 已提交
1950 1951 1952 1953 1954
		/* Must be called only after the slave has been
		 * detached from the list and the curr_active_slave
		 * has been cleared (if our_slave == old_current),
		 * but before a new active slave is selected.
		 */
1955
		write_unlock_bh(&bond->lock);
L
Linus Torvalds 已提交
1956
		bond_alb_deinit_slave(bond, slave);
1957
		write_lock_bh(&bond->lock);
L
Linus Torvalds 已提交
1958 1959
	}

1960 1961 1962 1963 1964 1965 1966 1967 1968 1969
	if (oldcurrent == slave) {
		/*
		 * Note that we hold RTNL over this sequence, so there
		 * is no concern that another slave add/remove event
		 * will interfere.
		 */
		write_unlock_bh(&bond->lock);
		read_lock(&bond->lock);
		write_lock_bh(&bond->curr_slave_lock);

L
Linus Torvalds 已提交
1970 1971
		bond_select_active_slave(bond);

1972 1973 1974 1975 1976
		write_unlock_bh(&bond->curr_slave_lock);
		read_unlock(&bond->lock);
		write_lock_bh(&bond->lock);
	}

L
Linus Torvalds 已提交
1977
	if (bond->slave_cnt == 0) {
1978 1979
		bond_set_carrier(bond);

L
Linus Torvalds 已提交
1980 1981 1982 1983 1984 1985
		/* if the last slave was removed, zero the mac address
		 * of the master so it will be set by the application
		 * to the mac address of the first slave
		 */
		memset(bond_dev->dev_addr, 0, bond_dev->addr_len);

1986
		if (!bond->vlgrp) {
L
Linus Torvalds 已提交
1987 1988
			bond_dev->features |= NETIF_F_VLAN_CHALLENGED;
		} else {
J
Joe Perches 已提交
1989 1990 1991 1992
			pr_warning("%s: Warning: clearing HW address of %s while it still has VLANs.\n",
				   bond_dev->name, bond_dev->name);
			pr_warning("%s: When re-adding slaves, make sure the bond's HW address matches its VLANs'.\n",
				   bond_dev->name);
L
Linus Torvalds 已提交
1993 1994 1995
		}
	} else if ((bond_dev->features & NETIF_F_VLAN_CHALLENGED) &&
		   !bond_has_challenged_slaves(bond)) {
J
Joe Perches 已提交
1996 1997
		pr_info("%s: last VLAN challenged slave %s left bond %s. VLAN blocking is removed\n",
			bond_dev->name, slave_dev->name, bond_dev->name);
L
Linus Torvalds 已提交
1998 1999 2000 2001
		bond_dev->features &= ~NETIF_F_VLAN_CHALLENGED;
	}

	write_unlock_bh(&bond->lock);
2002
	unblock_netpoll_tx();
L
Linus Torvalds 已提交
2003

2004 2005 2006
	/* must do this from outside any spinlocks */
	bond_destroy_slave_symlinks(bond_dev, slave_dev);

L
Linus Torvalds 已提交
2007 2008 2009 2010 2011 2012 2013 2014
	bond_del_vlans_from_slave(bond, slave_dev);

	/* If the mode USES_PRIMARY, then we should only remove its
	 * promisc and mc settings if it was the curr_active_slave, but that was
	 * already taken care of above when we detached the slave
	 */
	if (!USES_PRIMARY(bond->params.mode)) {
		/* unset promiscuity level from slave */
S
Stephen Hemminger 已提交
2015
		if (bond_dev->flags & IFF_PROMISC)
L
Linus Torvalds 已提交
2016 2017 2018
			dev_set_promiscuity(slave_dev, -1);

		/* unset allmulti level from slave */
S
Stephen Hemminger 已提交
2019
		if (bond_dev->flags & IFF_ALLMULTI)
L
Linus Torvalds 已提交
2020 2021 2022
			dev_set_allmulti(slave_dev, -1);

		/* flush master's mc_list from slave */
2023
		netif_addr_lock_bh(bond_dev);
L
Linus Torvalds 已提交
2024
		bond_mc_list_flush(bond_dev, slave_dev);
2025
		netif_addr_unlock_bh(bond_dev);
L
Linus Torvalds 已提交
2026 2027 2028 2029
	}

	netdev_set_master(slave_dev, NULL);

2030 2031
#ifdef CONFIG_NET_POLL_CONTROLLER
	read_lock_bh(&bond->lock);
2032 2033 2034 2035 2036

	 /* Make sure netpoll over stays disabled until fixed. */
	if (!disable_netpoll)
		if (slaves_support_netpoll(bond_dev))
				bond_dev->priv_flags &= ~IFF_DISABLE_NETPOLL;
2037 2038 2039 2040 2041 2042 2043
	read_unlock_bh(&bond->lock);
	if (slave_dev->netdev_ops->ndo_netpoll_cleanup)
		slave_dev->netdev_ops->ndo_netpoll_cleanup(slave_dev);
	else
		slave_dev->npinfo = NULL;
#endif

L
Linus Torvalds 已提交
2044 2045 2046
	/* close slave before restoring its mac address */
	dev_close(slave_dev);

2047
	if (bond->params.fail_over_mac != BOND_FOM_ACTIVE) {
2048 2049 2050 2051 2052
		/* restore original ("permanent") mac address */
		memcpy(addr.sa_data, slave->perm_hwaddr, ETH_ALEN);
		addr.sa_family = slave_dev->type;
		dev_set_mac_address(slave_dev, &addr);
	}
L
Linus Torvalds 已提交
2053

2054 2055
	dev_set_mtu(slave_dev, slave->original_mtu);

2056
	slave_dev->priv_flags &= ~(IFF_MASTER_8023AD | IFF_MASTER_ALB |
2057 2058
				   IFF_SLAVE_INACTIVE | IFF_BONDING |
				   IFF_SLAVE_NEEDARP);
L
Linus Torvalds 已提交
2059 2060 2061 2062 2063 2064

	kfree(slave);

	return 0;  /* deletion OK */
}

2065
/*
S
Stephen Hemminger 已提交
2066
* First release a slave and than destroy the bond if no more slaves are left.
2067 2068
* Must be under rtnl_lock when this function is called.
*/
S
Stephen Hemminger 已提交
2069 2070
int  bond_release_and_destroy(struct net_device *bond_dev,
			      struct net_device *slave_dev)
2071
{
2072
	struct bonding *bond = netdev_priv(bond_dev);
2073 2074 2075 2076
	int ret;

	ret = bond_release(bond_dev, slave_dev);
	if ((ret == 0) && (bond->slave_cnt == 0)) {
J
Joe Perches 已提交
2077 2078
		pr_info("%s: destroying bond %s.\n",
			bond_dev->name, bond_dev->name);
S
Stephen Hemminger 已提交
2079
		unregister_netdevice(bond_dev);
2080 2081 2082 2083
	}
	return ret;
}

L
Linus Torvalds 已提交
2084 2085 2086 2087 2088
/*
 * This function releases all slaves.
 */
static int bond_release_all(struct net_device *bond_dev)
{
2089
	struct bonding *bond = netdev_priv(bond_dev);
L
Linus Torvalds 已提交
2090 2091 2092 2093
	struct slave *slave;
	struct net_device *slave_dev;
	struct sockaddr addr;

2094
	block_netpoll_tx();
L
Linus Torvalds 已提交
2095 2096
	write_lock_bh(&bond->lock);

2097 2098
	netif_carrier_off(bond_dev);

S
Stephen Hemminger 已提交
2099
	if (bond->slave_cnt == 0)
L
Linus Torvalds 已提交
2100 2101 2102 2103 2104 2105 2106 2107 2108 2109
		goto out;

	bond->current_arp_slave = NULL;
	bond->primary_slave = NULL;
	bond_change_active_slave(bond, NULL);

	while ((slave = bond->first_slave) != NULL) {
		/* Inform AD package of unbinding of slave
		 * before slave is detached from the list.
		 */
S
Stephen Hemminger 已提交
2110
		if (bond->params.mode == BOND_MODE_8023AD)
L
Linus Torvalds 已提交
2111 2112 2113 2114 2115
			bond_3ad_unbind_slave(slave);

		slave_dev = slave->dev;
		bond_detach_slave(bond, slave);

2116 2117 2118 2119 2120 2121
		/* now that the slave is detached, unlock and perform
		 * all the undo steps that should not be called from
		 * within a lock.
		 */
		write_unlock_bh(&bond->lock);

2122
		if (bond_is_lb(bond)) {
L
Linus Torvalds 已提交
2123 2124 2125 2126 2127 2128
			/* must be called only after the slave
			 * has been detached from the list
			 */
			bond_alb_deinit_slave(bond, slave);
		}

2129 2130
		bond_compute_features(bond);

2131
		bond_destroy_slave_symlinks(bond_dev, slave_dev);
L
Linus Torvalds 已提交
2132 2133 2134 2135 2136 2137 2138 2139
		bond_del_vlans_from_slave(bond, slave_dev);

		/* If the mode USES_PRIMARY, then we should only remove its
		 * promisc and mc settings if it was the curr_active_slave, but that was
		 * already taken care of above when we detached the slave
		 */
		if (!USES_PRIMARY(bond->params.mode)) {
			/* unset promiscuity level from slave */
S
Stephen Hemminger 已提交
2140
			if (bond_dev->flags & IFF_PROMISC)
L
Linus Torvalds 已提交
2141 2142 2143
				dev_set_promiscuity(slave_dev, -1);

			/* unset allmulti level from slave */
S
Stephen Hemminger 已提交
2144
			if (bond_dev->flags & IFF_ALLMULTI)
L
Linus Torvalds 已提交
2145 2146 2147
				dev_set_allmulti(slave_dev, -1);

			/* flush master's mc_list from slave */
2148
			netif_addr_lock_bh(bond_dev);
L
Linus Torvalds 已提交
2149
			bond_mc_list_flush(bond_dev, slave_dev);
2150
			netif_addr_unlock_bh(bond_dev);
L
Linus Torvalds 已提交
2151 2152 2153 2154 2155 2156 2157
		}

		netdev_set_master(slave_dev, NULL);

		/* close slave before restoring its mac address */
		dev_close(slave_dev);

2158
		if (!bond->params.fail_over_mac) {
2159 2160 2161 2162 2163
			/* restore original ("permanent") mac address*/
			memcpy(addr.sa_data, slave->perm_hwaddr, ETH_ALEN);
			addr.sa_family = slave_dev->type;
			dev_set_mac_address(slave_dev, &addr);
		}
L
Linus Torvalds 已提交
2164

2165 2166
		slave_dev->priv_flags &= ~(IFF_MASTER_8023AD | IFF_MASTER_ALB |
					   IFF_SLAVE_INACTIVE);
L
Linus Torvalds 已提交
2167 2168 2169 2170 2171 2172 2173 2174 2175 2176 2177 2178 2179

		kfree(slave);

		/* re-acquire the lock before getting the next slave */
		write_lock_bh(&bond->lock);
	}

	/* zero the mac address of the master so it will be
	 * set by the application to the mac address of the
	 * first slave
	 */
	memset(bond_dev->dev_addr, 0, bond_dev->addr_len);

2180
	if (!bond->vlgrp) {
L
Linus Torvalds 已提交
2181
		bond_dev->features |= NETIF_F_VLAN_CHALLENGED;
2182
	} else {
J
Joe Perches 已提交
2183 2184 2185 2186
		pr_warning("%s: Warning: clearing HW address of %s while it still has VLANs.\n",
			   bond_dev->name, bond_dev->name);
		pr_warning("%s: When re-adding slaves, make sure the bond's HW address matches its VLANs'.\n",
			   bond_dev->name);
L
Linus Torvalds 已提交
2187 2188
	}

J
Joe Perches 已提交
2189
	pr_info("%s: released all slaves\n", bond_dev->name);
L
Linus Torvalds 已提交
2190 2191 2192

out:
	write_unlock_bh(&bond->lock);
2193
	unblock_netpoll_tx();
L
Linus Torvalds 已提交
2194 2195 2196 2197 2198 2199 2200 2201 2202 2203 2204 2205

	return 0;
}

/*
 * This function changes the active slave to slave <slave_dev>.
 * It returns -EINVAL in the following cases.
 *  - <slave_dev> is not found in the list.
 *  - There is not active slave now.
 *  - <slave_dev> is already active.
 *  - The link state of <slave_dev> is not BOND_LINK_UP.
 *  - <slave_dev> is not running.
S
Stephen Hemminger 已提交
2206 2207
 * In these cases, this function does nothing.
 * In the other cases, current_slave pointer is changed and 0 is returned.
L
Linus Torvalds 已提交
2208 2209 2210
 */
static int bond_ioctl_change_active(struct net_device *bond_dev, struct net_device *slave_dev)
{
2211
	struct bonding *bond = netdev_priv(bond_dev);
L
Linus Torvalds 已提交
2212 2213 2214 2215
	struct slave *old_active = NULL;
	struct slave *new_active = NULL;
	int res = 0;

S
Stephen Hemminger 已提交
2216
	if (!USES_PRIMARY(bond->params.mode))
L
Linus Torvalds 已提交
2217 2218 2219
		return -EINVAL;

	/* Verify that master_dev is indeed the master of slave_dev */
S
Stephen Hemminger 已提交
2220
	if (!(slave_dev->flags & IFF_SLAVE) || (slave_dev->master != bond_dev))
L
Linus Torvalds 已提交
2221 2222
		return -EINVAL;

2223
	read_lock(&bond->lock);
L
Linus Torvalds 已提交
2224

2225
	read_lock(&bond->curr_slave_lock);
L
Linus Torvalds 已提交
2226
	old_active = bond->curr_active_slave;
2227 2228
	read_unlock(&bond->curr_slave_lock);

L
Linus Torvalds 已提交
2229 2230 2231 2232 2233 2234
	new_active = bond_get_slave_by_dev(bond, slave_dev);

	/*
	 * Changing to the current active: do nothing; return success.
	 */
	if (new_active && (new_active == old_active)) {
2235
		read_unlock(&bond->lock);
L
Linus Torvalds 已提交
2236 2237 2238 2239 2240 2241 2242
		return 0;
	}

	if ((new_active) &&
	    (old_active) &&
	    (new_active->link == BOND_LINK_UP) &&
	    IS_UP(new_active->dev)) {
2243
		block_netpoll_tx();
2244
		write_lock_bh(&bond->curr_slave_lock);
L
Linus Torvalds 已提交
2245
		bond_change_active_slave(bond, new_active);
2246
		write_unlock_bh(&bond->curr_slave_lock);
2247
		unblock_netpoll_tx();
S
Stephen Hemminger 已提交
2248
	} else
L
Linus Torvalds 已提交
2249 2250
		res = -EINVAL;

2251
	read_unlock(&bond->lock);
L
Linus Torvalds 已提交
2252 2253 2254 2255 2256 2257

	return res;
}

static int bond_info_query(struct net_device *bond_dev, struct ifbond *info)
{
2258
	struct bonding *bond = netdev_priv(bond_dev);
L
Linus Torvalds 已提交
2259 2260 2261 2262

	info->bond_mode = bond->params.mode;
	info->miimon = bond->params.miimon;

2263
	read_lock(&bond->lock);
L
Linus Torvalds 已提交
2264
	info->num_slaves = bond->slave_cnt;
2265
	read_unlock(&bond->lock);
L
Linus Torvalds 已提交
2266 2267 2268 2269 2270 2271

	return 0;
}

static int bond_slave_info_query(struct net_device *bond_dev, struct ifslave *info)
{
2272
	struct bonding *bond = netdev_priv(bond_dev);
L
Linus Torvalds 已提交
2273
	struct slave *slave;
2274
	int i, res = -ENODEV;
L
Linus Torvalds 已提交
2275

2276
	read_lock(&bond->lock);
L
Linus Torvalds 已提交
2277 2278 2279

	bond_for_each_slave(bond, slave, i) {
		if (i == (int)info->slave_id) {
2280 2281 2282 2283 2284
			res = 0;
			strcpy(info->slave_name, slave->dev->name);
			info->link = slave->link;
			info->state = slave->state;
			info->link_failure_count = slave->link_failure_count;
L
Linus Torvalds 已提交
2285 2286 2287 2288
			break;
		}
	}

2289
	read_unlock(&bond->lock);
L
Linus Torvalds 已提交
2290

2291
	return res;
L
Linus Torvalds 已提交
2292 2293 2294 2295 2296
}

/*-------------------------------- Monitoring -------------------------------*/


J
Jay Vosburgh 已提交
2297 2298 2299 2300
static int bond_miimon_inspect(struct bonding *bond)
{
	struct slave *slave;
	int i, link_state, commit = 0;
2301 2302 2303
	bool ignore_updelay;

	ignore_updelay = !bond->curr_active_slave ? true : false;
L
Linus Torvalds 已提交
2304 2305

	bond_for_each_slave(bond, slave, i) {
J
Jay Vosburgh 已提交
2306
		slave->new_link = BOND_LINK_NOCHANGE;
L
Linus Torvalds 已提交
2307

J
Jay Vosburgh 已提交
2308
		link_state = bond_check_dev_link(bond, slave->dev, 0);
L
Linus Torvalds 已提交
2309 2310

		switch (slave->link) {
J
Jay Vosburgh 已提交
2311 2312 2313
		case BOND_LINK_UP:
			if (link_state)
				continue;
L
Linus Torvalds 已提交
2314

J
Jay Vosburgh 已提交
2315 2316 2317
			slave->link = BOND_LINK_FAIL;
			slave->delay = bond->params.downdelay;
			if (slave->delay) {
J
Joe Perches 已提交
2318 2319 2320 2321 2322 2323 2324 2325
				pr_info("%s: link status down for %sinterface %s, disabling it in %d ms.\n",
					bond->dev->name,
					(bond->params.mode ==
					 BOND_MODE_ACTIVEBACKUP) ?
					((slave->state == BOND_STATE_ACTIVE) ?
					 "active " : "backup ") : "",
					slave->dev->name,
					bond->params.downdelay * bond->params.miimon);
L
Linus Torvalds 已提交
2326
			}
J
Jay Vosburgh 已提交
2327 2328 2329 2330 2331 2332 2333
			/*FALLTHRU*/
		case BOND_LINK_FAIL:
			if (link_state) {
				/*
				 * recovered before downdelay expired
				 */
				slave->link = BOND_LINK_UP;
L
Linus Torvalds 已提交
2334
				slave->jiffies = jiffies;
J
Joe Perches 已提交
2335 2336 2337 2338 2339
				pr_info("%s: link status up again after %d ms for interface %s.\n",
					bond->dev->name,
					(bond->params.downdelay - slave->delay) *
					bond->params.miimon,
					slave->dev->name);
J
Jay Vosburgh 已提交
2340
				continue;
L
Linus Torvalds 已提交
2341
			}
J
Jay Vosburgh 已提交
2342 2343 2344 2345 2346

			if (slave->delay <= 0) {
				slave->new_link = BOND_LINK_DOWN;
				commit++;
				continue;
L
Linus Torvalds 已提交
2347 2348
			}

J
Jay Vosburgh 已提交
2349 2350 2351 2352 2353 2354 2355 2356 2357 2358 2359
			slave->delay--;
			break;

		case BOND_LINK_DOWN:
			if (!link_state)
				continue;

			slave->link = BOND_LINK_BACK;
			slave->delay = bond->params.updelay;

			if (slave->delay) {
J
Joe Perches 已提交
2360 2361 2362 2363 2364
				pr_info("%s: link status up for interface %s, enabling it in %d ms.\n",
					bond->dev->name, slave->dev->name,
					ignore_updelay ? 0 :
					bond->params.updelay *
					bond->params.miimon);
J
Jay Vosburgh 已提交
2365 2366 2367 2368 2369
			}
			/*FALLTHRU*/
		case BOND_LINK_BACK:
			if (!link_state) {
				slave->link = BOND_LINK_DOWN;
J
Joe Perches 已提交
2370 2371 2372 2373 2374
				pr_info("%s: link status down again after %d ms for interface %s.\n",
					bond->dev->name,
					(bond->params.updelay - slave->delay) *
					bond->params.miimon,
					slave->dev->name);
J
Jay Vosburgh 已提交
2375 2376 2377 2378

				continue;
			}

2379 2380 2381
			if (ignore_updelay)
				slave->delay = 0;

J
Jay Vosburgh 已提交
2382 2383 2384
			if (slave->delay <= 0) {
				slave->new_link = BOND_LINK_UP;
				commit++;
2385
				ignore_updelay = false;
J
Jay Vosburgh 已提交
2386
				continue;
L
Linus Torvalds 已提交
2387
			}
J
Jay Vosburgh 已提交
2388 2389

			slave->delay--;
L
Linus Torvalds 已提交
2390
			break;
J
Jay Vosburgh 已提交
2391 2392
		}
	}
L
Linus Torvalds 已提交
2393

J
Jay Vosburgh 已提交
2394 2395
	return commit;
}
L
Linus Torvalds 已提交
2396

J
Jay Vosburgh 已提交
2397 2398 2399 2400 2401 2402 2403 2404 2405
static void bond_miimon_commit(struct bonding *bond)
{
	struct slave *slave;
	int i;

	bond_for_each_slave(bond, slave, i) {
		switch (slave->new_link) {
		case BOND_LINK_NOCHANGE:
			continue;
L
Linus Torvalds 已提交
2406

J
Jay Vosburgh 已提交
2407 2408 2409 2410 2411 2412 2413 2414 2415 2416 2417 2418 2419
		case BOND_LINK_UP:
			slave->link = BOND_LINK_UP;
			slave->jiffies = jiffies;

			if (bond->params.mode == BOND_MODE_8023AD) {
				/* prevent it from being the active one */
				slave->state = BOND_STATE_BACKUP;
			} else if (bond->params.mode != BOND_MODE_ACTIVEBACKUP) {
				/* make it immediately active */
				slave->state = BOND_STATE_ACTIVE;
			} else if (slave != bond->primary_slave) {
				/* prevent it from being the active one */
				slave->state = BOND_STATE_BACKUP;
L
Linus Torvalds 已提交
2420 2421
			}

2422 2423
			bond_update_speed_duplex(slave);

2424 2425 2426
			pr_info("%s: link status definitely up for interface %s, %d Mbps %s duplex.\n",
				bond->dev->name, slave->dev->name,
				slave->speed, slave->duplex ? "full" : "half");
L
Linus Torvalds 已提交
2427

J
Jay Vosburgh 已提交
2428 2429 2430
			/* notify ad that the link status has changed */
			if (bond->params.mode == BOND_MODE_8023AD)
				bond_3ad_handle_link_change(slave, BOND_LINK_UP);
2431

2432
			if (bond_is_lb(bond))
J
Jay Vosburgh 已提交
2433 2434
				bond_alb_handle_link_change(bond, slave,
							    BOND_LINK_UP);
L
Linus Torvalds 已提交
2435

J
Jay Vosburgh 已提交
2436 2437 2438
			if (!bond->curr_active_slave ||
			    (slave == bond->primary_slave))
				goto do_failover;
L
Linus Torvalds 已提交
2439

J
Jay Vosburgh 已提交
2440
			continue;
2441

J
Jay Vosburgh 已提交
2442
		case BOND_LINK_DOWN:
J
Jay Vosburgh 已提交
2443 2444 2445
			if (slave->link_failure_count < UINT_MAX)
				slave->link_failure_count++;

J
Jay Vosburgh 已提交
2446
			slave->link = BOND_LINK_DOWN;
L
Linus Torvalds 已提交
2447

J
Jay Vosburgh 已提交
2448 2449 2450 2451
			if (bond->params.mode == BOND_MODE_ACTIVEBACKUP ||
			    bond->params.mode == BOND_MODE_8023AD)
				bond_set_slave_inactive_flags(slave);

J
Joe Perches 已提交
2452 2453
			pr_info("%s: link status definitely down for interface %s, disabling it\n",
				bond->dev->name, slave->dev->name);
J
Jay Vosburgh 已提交
2454 2455 2456 2457 2458

			if (bond->params.mode == BOND_MODE_8023AD)
				bond_3ad_handle_link_change(slave,
							    BOND_LINK_DOWN);

2459
			if (bond_is_lb(bond))
J
Jay Vosburgh 已提交
2460 2461 2462 2463 2464 2465 2466 2467 2468
				bond_alb_handle_link_change(bond, slave,
							    BOND_LINK_DOWN);

			if (slave == bond->curr_active_slave)
				goto do_failover;

			continue;

		default:
J
Joe Perches 已提交
2469
			pr_err("%s: invalid new link %d on slave %s\n",
J
Jay Vosburgh 已提交
2470 2471 2472 2473 2474 2475 2476 2477 2478
			       bond->dev->name, slave->new_link,
			       slave->dev->name);
			slave->new_link = BOND_LINK_NOCHANGE;

			continue;
		}

do_failover:
		ASSERT_RTNL();
2479
		block_netpoll_tx();
J
Jay Vosburgh 已提交
2480 2481 2482
		write_lock_bh(&bond->curr_slave_lock);
		bond_select_active_slave(bond);
		write_unlock_bh(&bond->curr_slave_lock);
2483
		unblock_netpoll_tx();
J
Jay Vosburgh 已提交
2484 2485 2486
	}

	bond_set_carrier(bond);
L
Linus Torvalds 已提交
2487 2488
}

2489 2490 2491 2492
/*
 * bond_mii_monitor
 *
 * Really a wrapper that splits the mii monitor into two phases: an
J
Jay Vosburgh 已提交
2493 2494 2495
 * inspection, then (if inspection indicates something needs to be done)
 * an acquisition of appropriate locks followed by a commit phase to
 * implement whatever link state changes are indicated.
2496 2497 2498 2499 2500 2501 2502
 */
void bond_mii_monitor(struct work_struct *work)
{
	struct bonding *bond = container_of(work, struct bonding,
					    mii_work.work);

	read_lock(&bond->lock);
J
Jay Vosburgh 已提交
2503 2504 2505 2506 2507
	if (bond->kill_timers)
		goto out;

	if (bond->slave_cnt == 0)
		goto re_arm;
2508 2509 2510 2511 2512 2513 2514

	if (bond->send_grat_arp) {
		read_lock(&bond->curr_slave_lock);
		bond_send_gratuitous_arp(bond);
		read_unlock(&bond->curr_slave_lock);
	}

2515 2516 2517 2518 2519 2520
	if (bond->send_unsol_na) {
		read_lock(&bond->curr_slave_lock);
		bond_send_unsolicited_na(bond);
		read_unlock(&bond->curr_slave_lock);
	}

J
Jay Vosburgh 已提交
2521
	if (bond_miimon_inspect(bond)) {
2522 2523 2524
		read_unlock(&bond->lock);
		rtnl_lock();
		read_lock(&bond->lock);
J
Jay Vosburgh 已提交
2525 2526 2527

		bond_miimon_commit(bond);

2528 2529 2530
		read_unlock(&bond->lock);
		rtnl_unlock();	/* might sleep, hold no other locks */
		read_lock(&bond->lock);
2531 2532
	}

J
Jay Vosburgh 已提交
2533 2534 2535 2536 2537
re_arm:
	if (bond->params.miimon)
		queue_delayed_work(bond->wq, &bond->mii_work,
				   msecs_to_jiffies(bond->params.miimon));
out:
2538 2539
	read_unlock(&bond->lock);
}
J
Jay Vosburgh 已提交
2540

2541
static __be32 bond_glean_dev_ip(struct net_device *dev)
J
Jay Vosburgh 已提交
2542 2543 2544
{
	struct in_device *idev;
	struct in_ifaddr *ifa;
A
Al Viro 已提交
2545
	__be32 addr = 0;
J
Jay Vosburgh 已提交
2546 2547 2548 2549 2550

	if (!dev)
		return 0;

	rcu_read_lock();
2551
	idev = __in_dev_get_rcu(dev);
J
Jay Vosburgh 已提交
2552 2553 2554 2555 2556 2557 2558 2559 2560 2561 2562 2563 2564
	if (!idev)
		goto out;

	ifa = idev->ifa_list;
	if (!ifa)
		goto out;

	addr = ifa->ifa_local;
out:
	rcu_read_unlock();
	return addr;
}

2565
static int bond_has_this_ip(struct bonding *bond, __be32 ip)
2566
{
2567
	struct vlan_entry *vlan;
2568 2569 2570 2571

	if (ip == bond->master_ip)
		return 1;

2572
	list_for_each_entry(vlan, &bond->vlan_list, vlan_list) {
2573 2574 2575 2576 2577 2578 2579
		if (ip == vlan->vlan_ip)
			return 1;
	}

	return 0;
}

J
Jay Vosburgh 已提交
2580 2581 2582 2583 2584
/*
 * We go to the (large) trouble of VLAN tagging ARP frames because
 * switches in VLAN mode (especially if ports are configured as
 * "native" to a VLAN) might not pass non-tagged frames.
 */
2585
static void bond_arp_send(struct net_device *slave_dev, int arp_op, __be32 dest_ip, __be32 src_ip, unsigned short vlan_id)
J
Jay Vosburgh 已提交
2586 2587 2588
{
	struct sk_buff *skb;

2589
	pr_debug("arp %d on slave %s: dst %x src %x vid %d\n", arp_op,
J
Joe Perches 已提交
2590
		 slave_dev->name, dest_ip, src_ip, vlan_id);
S
Stephen Hemminger 已提交
2591

J
Jay Vosburgh 已提交
2592 2593 2594 2595
	skb = arp_create(arp_op, ETH_P_ARP, dest_ip, slave_dev, src_ip,
			 NULL, slave_dev->dev_addr, NULL);

	if (!skb) {
J
Joe Perches 已提交
2596
		pr_err("ARP packet allocation failed\n");
J
Jay Vosburgh 已提交
2597 2598 2599 2600 2601
		return;
	}
	if (vlan_id) {
		skb = vlan_put_tag(skb, vlan_id);
		if (!skb) {
J
Joe Perches 已提交
2602
			pr_err("failed to insert VLAN tag\n");
J
Jay Vosburgh 已提交
2603 2604 2605 2606 2607 2608 2609
			return;
		}
	}
	arp_xmit(skb);
}


L
Linus Torvalds 已提交
2610 2611
static void bond_arp_send_all(struct bonding *bond, struct slave *slave)
{
J
Jay Vosburgh 已提交
2612
	int i, vlan_id, rv;
2613
	__be32 *targets = bond->params.arp_targets;
2614
	struct vlan_entry *vlan;
J
Jay Vosburgh 已提交
2615 2616 2617
	struct net_device *vlan_dev;
	struct flowi fl;
	struct rtable *rt;
L
Linus Torvalds 已提交
2618

2619 2620
	for (i = 0; (i < BOND_MAX_ARP_TARGETS); i++) {
		if (!targets[i])
2621
			break;
2622
		pr_debug("basa: target %x\n", targets[i]);
2623
		if (!bond->vlgrp) {
2624
			pr_debug("basa: empty vlan: arp_send\n");
J
Jay Vosburgh 已提交
2625 2626 2627 2628 2629 2630 2631 2632 2633 2634 2635 2636 2637 2638
			bond_arp_send(slave->dev, ARPOP_REQUEST, targets[i],
				      bond->master_ip, 0);
			continue;
		}

		/*
		 * If VLANs are configured, we do a route lookup to
		 * determine which VLAN interface would be used, so we
		 * can tag the ARP with the proper VLAN tag.
		 */
		memset(&fl, 0, sizeof(fl));
		fl.fl4_dst = targets[i];
		fl.fl4_tos = RTO_ONLINK;

2639
		rv = ip_route_output_key(dev_net(bond->dev), &rt, &fl);
J
Jay Vosburgh 已提交
2640 2641
		if (rv) {
			if (net_ratelimit()) {
J
Joe Perches 已提交
2642 2643
				pr_warning("%s: no route to arp_ip_target %pI4\n",
					   bond->dev->name, &fl.fl4_dst);
J
Jay Vosburgh 已提交
2644 2645 2646 2647 2648 2649 2650
			}
			continue;
		}

		/*
		 * This target is not on a VLAN
		 */
2651
		if (rt->dst.dev == bond->dev) {
2652
			ip_rt_put(rt);
2653
			pr_debug("basa: rtdev == bond->dev: arp_send\n");
J
Jay Vosburgh 已提交
2654 2655 2656 2657 2658 2659
			bond_arp_send(slave->dev, ARPOP_REQUEST, targets[i],
				      bond->master_ip, 0);
			continue;
		}

		vlan_id = 0;
2660
		list_for_each_entry(vlan, &bond->vlan_list, vlan_list) {
D
Dan Aloni 已提交
2661
			vlan_dev = vlan_group_get_device(bond->vlgrp, vlan->vlan_id);
2662
			if (vlan_dev == rt->dst.dev) {
J
Jay Vosburgh 已提交
2663
				vlan_id = vlan->vlan_id;
2664
				pr_debug("basa: vlan match on %s %d\n",
J
Jay Vosburgh 已提交
2665 2666 2667 2668 2669 2670
				       vlan_dev->name, vlan_id);
				break;
			}
		}

		if (vlan_id) {
2671
			ip_rt_put(rt);
J
Jay Vosburgh 已提交
2672 2673 2674 2675 2676 2677
			bond_arp_send(slave->dev, ARPOP_REQUEST, targets[i],
				      vlan->vlan_ip, vlan_id);
			continue;
		}

		if (net_ratelimit()) {
J
Joe Perches 已提交
2678 2679
			pr_warning("%s: no path to arp_ip_target %pI4 via rt.dev %s\n",
				   bond->dev->name, &fl.fl4_dst,
2680
				   rt->dst.dev ? rt->dst.dev->name : "NULL");
J
Jay Vosburgh 已提交
2681
		}
2682
		ip_rt_put(rt);
J
Jay Vosburgh 已提交
2683 2684 2685 2686 2687 2688
	}
}

/*
 * Kick out a gratuitous ARP for an IP on the bonding master plus one
 * for each VLAN above us.
2689 2690
 *
 * Caller must hold curr_slave_lock for read or better
J
Jay Vosburgh 已提交
2691 2692 2693 2694 2695 2696 2697
 */
static void bond_send_gratuitous_arp(struct bonding *bond)
{
	struct slave *slave = bond->curr_active_slave;
	struct vlan_entry *vlan;
	struct net_device *vlan_dev;

J
Joe Perches 已提交
2698 2699
	pr_debug("bond_send_grat_arp: bond %s slave %s\n",
		 bond->dev->name, slave ? slave->dev->name : "NULL");
2700 2701 2702

	if (!slave || !bond->send_grat_arp ||
	    test_bit(__LINK_STATE_LINKWATCH_PENDING, &slave->dev->state))
J
Jay Vosburgh 已提交
2703 2704
		return;

2705 2706
	bond->send_grat_arp--;

J
Jay Vosburgh 已提交
2707 2708
	if (bond->master_ip) {
		bond_arp_send(slave->dev, ARPOP_REPLY, bond->master_ip,
2709
				bond->master_ip, 0);
J
Jay Vosburgh 已提交
2710 2711
	}

2712 2713 2714
	if (!bond->vlgrp)
		return;

J
Jay Vosburgh 已提交
2715
	list_for_each_entry(vlan, &bond->vlan_list, vlan_list) {
D
Dan Aloni 已提交
2716
		vlan_dev = vlan_group_get_device(bond->vlgrp, vlan->vlan_id);
J
Jay Vosburgh 已提交
2717 2718 2719 2720
		if (vlan->vlan_ip) {
			bond_arp_send(slave->dev, ARPOP_REPLY, vlan->vlan_ip,
				      vlan->vlan_ip, vlan->vlan_id);
		}
L
Linus Torvalds 已提交
2721 2722 2723
	}
}

2724
static void bond_validate_arp(struct bonding *bond, struct slave *slave, __be32 sip, __be32 tip)
2725 2726
{
	int i;
2727
	__be32 *targets = bond->params.arp_targets;
2728 2729

	for (i = 0; (i < BOND_MAX_ARP_TARGETS) && targets[i]; i++) {
2730
		pr_debug("bva: sip %pI4 tip %pI4 t[%d] %pI4 bhti(tip) %d\n",
J
Joe Perches 已提交
2731 2732
			 &sip, &tip, i, &targets[i],
			 bond_has_this_ip(bond, tip));
2733 2734 2735 2736 2737 2738 2739 2740 2741 2742 2743 2744 2745 2746
		if (sip == targets[i]) {
			if (bond_has_this_ip(bond, tip))
				slave->last_arp_rx = jiffies;
			return;
		}
	}
}

static int bond_arp_rcv(struct sk_buff *skb, struct net_device *dev, struct packet_type *pt, struct net_device *orig_dev)
{
	struct arphdr *arp;
	struct slave *slave;
	struct bonding *bond;
	unsigned char *arp_ptr;
2747
	__be32 sip, tip;
2748

2749 2750 2751 2752 2753 2754 2755 2756 2757 2758 2759
	if (dev->priv_flags & IFF_802_1Q_VLAN) {
		/*
		 * When using VLANS and bonding, dev and oriv_dev may be
		 * incorrect if the physical interface supports VLAN
		 * acceleration.  With this change ARP validation now
		 * works for hosts only reachable on the VLAN interface.
		 */
		dev = vlan_dev_real_dev(dev);
		orig_dev = dev_get_by_index_rcu(dev_net(skb->dev),skb->skb_iif);
	}

2760 2761 2762
	if (!(dev->priv_flags & IFF_BONDING) || !(dev->flags & IFF_MASTER))
		goto out;

2763
	bond = netdev_priv(dev);
2764 2765
	read_lock(&bond->lock);

2766
	pr_debug("bond_arp_rcv: bond %s skb->dev %s orig_dev %s\n",
J
Joe Perches 已提交
2767 2768
		 bond->dev->name, skb->dev ? skb->dev->name : "NULL",
		 orig_dev ? orig_dev->name : "NULL");
2769 2770 2771 2772 2773

	slave = bond_get_slave_by_dev(bond, orig_dev);
	if (!slave || !slave_do_arp_validate(bond, slave))
		goto out_unlock;

2774
	if (!pskb_may_pull(skb, arp_hdr_len(dev)))
2775 2776
		goto out_unlock;

2777
	arp = arp_hdr(skb);
2778 2779 2780 2781 2782 2783 2784 2785 2786 2787 2788 2789 2790 2791
	if (arp->ar_hln != dev->addr_len ||
	    skb->pkt_type == PACKET_OTHERHOST ||
	    skb->pkt_type == PACKET_LOOPBACK ||
	    arp->ar_hrd != htons(ARPHRD_ETHER) ||
	    arp->ar_pro != htons(ETH_P_IP) ||
	    arp->ar_pln != 4)
		goto out_unlock;

	arp_ptr = (unsigned char *)(arp + 1);
	arp_ptr += dev->addr_len;
	memcpy(&sip, arp_ptr, 4);
	arp_ptr += 4 + dev->addr_len;
	memcpy(&tip, arp_ptr, 4);

2792
	pr_debug("bond_arp_rcv: %s %s/%d av %d sv %d sip %pI4 tip %pI4\n",
J
Joe Perches 已提交
2793 2794 2795
		 bond->dev->name, slave->dev->name, slave->state,
		 bond->params.arp_validate, slave_do_arp_validate(bond, slave),
		 &sip, &tip);
2796 2797 2798 2799 2800 2801 2802 2803 2804 2805 2806 2807 2808 2809 2810 2811 2812 2813 2814 2815 2816

	/*
	 * Backup slaves won't see the ARP reply, but do come through
	 * here for each ARP probe (so we swap the sip/tip to validate
	 * the probe).  In a "redundant switch, common router" type of
	 * configuration, the ARP probe will (hopefully) travel from
	 * the active, through one switch, the router, then the other
	 * switch before reaching the backup.
	 */
	if (slave->state == BOND_STATE_ACTIVE)
		bond_validate_arp(bond, slave, sip, tip);
	else
		bond_validate_arp(bond, slave, tip, sip);

out_unlock:
	read_unlock(&bond->lock);
out:
	dev_kfree_skb(skb);
	return NET_RX_SUCCESS;
}

L
Linus Torvalds 已提交
2817 2818 2819 2820 2821 2822 2823
/*
 * this function is called regularly to monitor each slave's link
 * ensuring that traffic is being sent and received when arp monitoring
 * is used in load-balancing mode. if the adapter has been dormant, then an
 * arp is transmitted to generate traffic. see activebackup_arp_monitor for
 * arp monitoring in active backup mode.
 */
2824
void bond_loadbalance_arp_mon(struct work_struct *work)
L
Linus Torvalds 已提交
2825
{
2826 2827
	struct bonding *bond = container_of(work, struct bonding,
					    arp_work.work);
L
Linus Torvalds 已提交
2828 2829 2830 2831 2832 2833 2834
	struct slave *slave, *oldcurrent;
	int do_failover = 0;
	int delta_in_ticks;
	int i;

	read_lock(&bond->lock);

2835
	delta_in_ticks = msecs_to_jiffies(bond->params.arp_interval);
L
Linus Torvalds 已提交
2836

S
Stephen Hemminger 已提交
2837
	if (bond->kill_timers)
L
Linus Torvalds 已提交
2838 2839
		goto out;

S
Stephen Hemminger 已提交
2840
	if (bond->slave_cnt == 0)
L
Linus Torvalds 已提交
2841 2842 2843 2844 2845 2846 2847 2848 2849 2850 2851 2852 2853 2854 2855
		goto re_arm;

	read_lock(&bond->curr_slave_lock);
	oldcurrent = bond->curr_active_slave;
	read_unlock(&bond->curr_slave_lock);

	/* see if any of the previous devices are up now (i.e. they have
	 * xmt and rcv traffic). the curr_active_slave does not come into
	 * the picture unless it is null. also, slave->jiffies is not needed
	 * here because we send an arp on each slave and give a slave as
	 * long as it needs to get the tx/rx within the delta.
	 * TODO: what about up/down delay in arp mode? it wasn't here before
	 *       so it can wait
	 */
	bond_for_each_slave(bond, slave, i) {
2856 2857
		unsigned long trans_start = dev_trans_start(slave->dev);

L
Linus Torvalds 已提交
2858
		if (slave->link != BOND_LINK_UP) {
2859 2860 2861 2862 2863 2864
			if (time_in_range(jiffies,
				trans_start - delta_in_ticks,
				trans_start + delta_in_ticks) &&
			    time_in_range(jiffies,
				slave->dev->last_rx - delta_in_ticks,
				slave->dev->last_rx + delta_in_ticks)) {
L
Linus Torvalds 已提交
2865 2866 2867 2868 2869 2870 2871 2872 2873 2874

				slave->link  = BOND_LINK_UP;
				slave->state = BOND_STATE_ACTIVE;

				/* primary_slave has no meaning in round-robin
				 * mode. the window of a slave being up and
				 * curr_active_slave being null after enslaving
				 * is closed.
				 */
				if (!oldcurrent) {
J
Joe Perches 已提交
2875 2876 2877
					pr_info("%s: link status definitely up for interface %s, ",
						bond->dev->name,
						slave->dev->name);
L
Linus Torvalds 已提交
2878 2879
					do_failover = 1;
				} else {
J
Joe Perches 已提交
2880 2881 2882
					pr_info("%s: interface %s is now up\n",
						bond->dev->name,
						slave->dev->name);
L
Linus Torvalds 已提交
2883 2884 2885 2886 2887 2888 2889 2890 2891
				}
			}
		} else {
			/* slave->link == BOND_LINK_UP */

			/* not all switches will respond to an arp request
			 * when the source ip is 0, so don't take the link down
			 * if we don't know our ip yet
			 */
2892 2893 2894 2895 2896 2897
			if (!time_in_range(jiffies,
				trans_start - delta_in_ticks,
				trans_start + 2 * delta_in_ticks) ||
			    !time_in_range(jiffies,
				slave->dev->last_rx - delta_in_ticks,
				slave->dev->last_rx + 2 * delta_in_ticks)) {
L
Linus Torvalds 已提交
2898 2899 2900 2901

				slave->link  = BOND_LINK_DOWN;
				slave->state = BOND_STATE_BACKUP;

S
Stephen Hemminger 已提交
2902
				if (slave->link_failure_count < UINT_MAX)
L
Linus Torvalds 已提交
2903 2904
					slave->link_failure_count++;

J
Joe Perches 已提交
2905 2906 2907
				pr_info("%s: interface %s is now down.\n",
					bond->dev->name,
					slave->dev->name);
L
Linus Torvalds 已提交
2908

S
Stephen Hemminger 已提交
2909
				if (slave == oldcurrent)
L
Linus Torvalds 已提交
2910 2911 2912 2913 2914 2915 2916 2917 2918 2919 2920
					do_failover = 1;
			}
		}

		/* note: if switch is in round-robin mode, all links
		 * must tx arp to ensure all links rx an arp - otherwise
		 * links may oscillate or not come up at all; if switch is
		 * in something like xor mode, there is nothing we can
		 * do - all replies will be rx'ed on same link causing slaves
		 * to be unstable during low/no traffic periods
		 */
S
Stephen Hemminger 已提交
2921
		if (IS_UP(slave->dev))
L
Linus Torvalds 已提交
2922 2923 2924 2925
			bond_arp_send_all(bond, slave);
	}

	if (do_failover) {
2926
		block_netpoll_tx();
2927
		write_lock_bh(&bond->curr_slave_lock);
L
Linus Torvalds 已提交
2928 2929 2930

		bond_select_active_slave(bond);

2931
		write_unlock_bh(&bond->curr_slave_lock);
2932
		unblock_netpoll_tx();
L
Linus Torvalds 已提交
2933 2934 2935
	}

re_arm:
2936 2937
	if (bond->params.arp_interval)
		queue_delayed_work(bond->wq, &bond->arp_work, delta_in_ticks);
L
Linus Torvalds 已提交
2938 2939 2940 2941 2942
out:
	read_unlock(&bond->lock);
}

/*
2943 2944 2945 2946 2947 2948
 * Called to inspect slaves for active-backup mode ARP monitor link state
 * changes.  Sets new_link in slaves to specify what action should take
 * place for the slave.  Returns 0 if no changes are found, >0 if changes
 * to link states must be committed.
 *
 * Called with bond->lock held for read.
L
Linus Torvalds 已提交
2949
 */
2950
static int bond_ab_arp_inspect(struct bonding *bond, int delta_in_ticks)
L
Linus Torvalds 已提交
2951 2952
{
	struct slave *slave;
2953
	int i, commit = 0;
2954
	unsigned long trans_start;
L
Linus Torvalds 已提交
2955

2956 2957
	bond_for_each_slave(bond, slave, i) {
		slave->new_link = BOND_LINK_NOCHANGE;
L
Linus Torvalds 已提交
2958

2959
		if (slave->link != BOND_LINK_UP) {
2960 2961 2962 2963
			if (time_in_range(jiffies,
				slave_last_rx(bond, slave) - delta_in_ticks,
				slave_last_rx(bond, slave) + delta_in_ticks)) {

2964 2965 2966
				slave->new_link = BOND_LINK_UP;
				commit++;
			}
L
Linus Torvalds 已提交
2967

2968 2969
			continue;
		}
L
Linus Torvalds 已提交
2970

2971 2972 2973 2974 2975
		/*
		 * Give slaves 2*delta after being enslaved or made
		 * active.  This avoids bouncing, as the last receive
		 * times need a full ARP monitor cycle to be updated.
		 */
2976 2977 2978
		if (time_in_range(jiffies,
				  slave->jiffies - delta_in_ticks,
				  slave->jiffies + 2 * delta_in_ticks))
2979 2980 2981 2982 2983 2984 2985 2986 2987 2988 2989 2990 2991 2992 2993 2994 2995
			continue;

		/*
		 * Backup slave is down if:
		 * - No current_arp_slave AND
		 * - more than 3*delta since last receive AND
		 * - the bond has an IP address
		 *
		 * Note: a non-null current_arp_slave indicates
		 * the curr_active_slave went down and we are
		 * searching for a new one; under this condition
		 * we only take the curr_active_slave down - this
		 * gives each slave a chance to tx/rx traffic
		 * before being taken out
		 */
		if (slave->state == BOND_STATE_BACKUP &&
		    !bond->current_arp_slave &&
2996 2997 2998 2999
		    !time_in_range(jiffies,
			slave_last_rx(bond, slave) - delta_in_ticks,
			slave_last_rx(bond, slave) + 3 * delta_in_ticks)) {

3000 3001 3002 3003 3004 3005 3006 3007 3008 3009
			slave->new_link = BOND_LINK_DOWN;
			commit++;
		}

		/*
		 * Active slave is down if:
		 * - more than 2*delta since transmitting OR
		 * - (more than 2*delta since receive AND
		 *    the bond has an IP address)
		 */
3010
		trans_start = dev_trans_start(slave->dev);
3011
		if ((slave->state == BOND_STATE_ACTIVE) &&
3012 3013 3014 3015 3016 3017 3018
		    (!time_in_range(jiffies,
			trans_start - delta_in_ticks,
			trans_start + 2 * delta_in_ticks) ||
		     !time_in_range(jiffies,
			slave_last_rx(bond, slave) - delta_in_ticks,
			slave_last_rx(bond, slave) + 2 * delta_in_ticks))) {

3019 3020 3021
			slave->new_link = BOND_LINK_DOWN;
			commit++;
		}
L
Linus Torvalds 已提交
3022 3023
	}

3024 3025
	return commit;
}
L
Linus Torvalds 已提交
3026

3027 3028 3029 3030 3031 3032 3033 3034 3035 3036
/*
 * Called to commit link state changes noted by inspection step of
 * active-backup mode ARP monitor.
 *
 * Called with RTNL and bond->lock for read.
 */
static void bond_ab_arp_commit(struct bonding *bond, int delta_in_ticks)
{
	struct slave *slave;
	int i;
3037
	unsigned long trans_start;
L
Linus Torvalds 已提交
3038

3039 3040 3041 3042
	bond_for_each_slave(bond, slave, i) {
		switch (slave->new_link) {
		case BOND_LINK_NOCHANGE:
			continue;
3043

3044
		case BOND_LINK_UP:
3045
			trans_start = dev_trans_start(slave->dev);
3046
			if ((!bond->curr_active_slave &&
3047 3048 3049
			     time_in_range(jiffies,
					   trans_start - delta_in_ticks,
					   trans_start + delta_in_ticks)) ||
3050
			    bond->curr_active_slave != slave) {
3051 3052 3053
				slave->link = BOND_LINK_UP;
				bond->current_arp_slave = NULL;

J
Joe Perches 已提交
3054
				pr_info("%s: link status definitely up for interface %s.\n",
3055
					bond->dev->name, slave->dev->name);
3056

3057 3058 3059
				if (!bond->curr_active_slave ||
				    (slave == bond->primary_slave))
					goto do_failover;
L
Linus Torvalds 已提交
3060

3061
			}
L
Linus Torvalds 已提交
3062

3063
			continue;
L
Linus Torvalds 已提交
3064

3065 3066 3067 3068 3069
		case BOND_LINK_DOWN:
			if (slave->link_failure_count < UINT_MAX)
				slave->link_failure_count++;

			slave->link = BOND_LINK_DOWN;
3070
			bond_set_slave_inactive_flags(slave);
3071

J
Joe Perches 已提交
3072
			pr_info("%s: link status definitely down for interface %s, disabling it\n",
3073
				bond->dev->name, slave->dev->name);
3074

3075
			if (slave == bond->curr_active_slave) {
3076
				bond->current_arp_slave = NULL;
3077
				goto do_failover;
L
Linus Torvalds 已提交
3078
			}
3079 3080

			continue;
3081 3082

		default:
J
Joe Perches 已提交
3083
			pr_err("%s: impossible: new_link %d on slave %s\n",
3084 3085
			       bond->dev->name, slave->new_link,
			       slave->dev->name);
3086
			continue;
L
Linus Torvalds 已提交
3087 3088
		}

3089 3090
do_failover:
		ASSERT_RTNL();
3091
		block_netpoll_tx();
3092
		write_lock_bh(&bond->curr_slave_lock);
3093
		bond_select_active_slave(bond);
3094
		write_unlock_bh(&bond->curr_slave_lock);
3095
		unblock_netpoll_tx();
3096
	}
L
Linus Torvalds 已提交
3097

3098 3099
	bond_set_carrier(bond);
}
L
Linus Torvalds 已提交
3100

3101 3102 3103 3104 3105 3106 3107 3108 3109
/*
 * Send ARP probes for active-backup mode ARP monitor.
 *
 * Called with bond->lock held for read.
 */
static void bond_ab_arp_probe(struct bonding *bond)
{
	struct slave *slave;
	int i;
L
Linus Torvalds 已提交
3110

3111
	read_lock(&bond->curr_slave_lock);
L
Linus Torvalds 已提交
3112

3113
	if (bond->current_arp_slave && bond->curr_active_slave)
J
Joe Perches 已提交
3114 3115 3116
		pr_info("PROBE: c_arp %s && cas %s BAD\n",
			bond->current_arp_slave->dev->name,
			bond->curr_active_slave->dev->name);
L
Linus Torvalds 已提交
3117

3118 3119 3120 3121 3122
	if (bond->curr_active_slave) {
		bond_arp_send_all(bond, bond->curr_active_slave);
		read_unlock(&bond->curr_slave_lock);
		return;
	}
L
Linus Torvalds 已提交
3123

3124
	read_unlock(&bond->curr_slave_lock);
3125

3126 3127 3128 3129
	/* if we don't have a curr_active_slave, search for the next available
	 * backup slave from the current_arp_slave and make it the candidate
	 * for becoming the curr_active_slave
	 */
L
Linus Torvalds 已提交
3130

3131 3132 3133 3134 3135
	if (!bond->current_arp_slave) {
		bond->current_arp_slave = bond->first_slave;
		if (!bond->current_arp_slave)
			return;
	}
L
Linus Torvalds 已提交
3136

3137
	bond_set_slave_inactive_flags(bond->current_arp_slave);
3138

3139 3140 3141 3142 3143 3144
	/* search for next candidate */
	bond_for_each_slave_from(bond, slave, i, bond->current_arp_slave->next) {
		if (IS_UP(slave->dev)) {
			slave->link = BOND_LINK_BACK;
			bond_set_slave_active_flags(slave);
			bond_arp_send_all(bond, slave);
L
Linus Torvalds 已提交
3145
			slave->jiffies = jiffies;
3146 3147
			bond->current_arp_slave = slave;
			break;
L
Linus Torvalds 已提交
3148 3149
		}

3150 3151 3152 3153 3154 3155
		/* if the link state is up at this point, we
		 * mark it down - this can happen if we have
		 * simultaneous link failures and
		 * reselect_active_interface doesn't make this
		 * one the current slave so it is still marked
		 * up when it is actually down
L
Linus Torvalds 已提交
3156
		 */
3157 3158 3159 3160
		if (slave->link == BOND_LINK_UP) {
			slave->link = BOND_LINK_DOWN;
			if (slave->link_failure_count < UINT_MAX)
				slave->link_failure_count++;
L
Linus Torvalds 已提交
3161

3162 3163
			bond_set_slave_inactive_flags(slave);

J
Joe Perches 已提交
3164 3165
			pr_info("%s: backup interface %s is now down.\n",
				bond->dev->name, slave->dev->name);
L
Linus Torvalds 已提交
3166
		}
3167 3168
	}
}
L
Linus Torvalds 已提交
3169

3170 3171 3172 3173 3174
void bond_activebackup_arp_mon(struct work_struct *work)
{
	struct bonding *bond = container_of(work, struct bonding,
					    arp_work.work);
	int delta_in_ticks;
L
Linus Torvalds 已提交
3175

3176
	read_lock(&bond->lock);
L
Linus Torvalds 已提交
3177

3178 3179
	if (bond->kill_timers)
		goto out;
L
Linus Torvalds 已提交
3180

3181
	delta_in_ticks = msecs_to_jiffies(bond->params.arp_interval);
L
Linus Torvalds 已提交
3182

3183 3184 3185
	if (bond->slave_cnt == 0)
		goto re_arm;

3186 3187 3188 3189 3190 3191
	if (bond->send_grat_arp) {
		read_lock(&bond->curr_slave_lock);
		bond_send_gratuitous_arp(bond);
		read_unlock(&bond->curr_slave_lock);
	}

3192 3193 3194 3195 3196 3197
	if (bond->send_unsol_na) {
		read_lock(&bond->curr_slave_lock);
		bond_send_unsolicited_na(bond);
		read_unlock(&bond->curr_slave_lock);
	}

3198 3199 3200 3201 3202 3203 3204 3205 3206 3207
	if (bond_ab_arp_inspect(bond, delta_in_ticks)) {
		read_unlock(&bond->lock);
		rtnl_lock();
		read_lock(&bond->lock);

		bond_ab_arp_commit(bond, delta_in_ticks);

		read_unlock(&bond->lock);
		rtnl_unlock();
		read_lock(&bond->lock);
L
Linus Torvalds 已提交
3208 3209
	}

3210 3211
	bond_ab_arp_probe(bond);

L
Linus Torvalds 已提交
3212
re_arm:
S
Stephen Hemminger 已提交
3213
	if (bond->params.arp_interval)
3214
		queue_delayed_work(bond->wq, &bond->arp_work, delta_in_ticks);
L
Linus Torvalds 已提交
3215 3216 3217 3218 3219 3220 3221 3222 3223
out:
	read_unlock(&bond->lock);
}

/*------------------------------ proc/seq_file-------------------------------*/

#ifdef CONFIG_PROC_FS

static void *bond_info_seq_start(struct seq_file *seq, loff_t *pos)
3224 3225
	__acquires(&dev_base_lock)
	__acquires(&bond->lock)
L
Linus Torvalds 已提交
3226 3227 3228 3229 3230 3231 3232 3233
{
	struct bonding *bond = seq->private;
	loff_t off = 0;
	struct slave *slave;
	int i;

	/* make sure the bond won't be taken away */
	read_lock(&dev_base_lock);
3234
	read_lock(&bond->lock);
L
Linus Torvalds 已提交
3235

S
Stephen Hemminger 已提交
3236
	if (*pos == 0)
L
Linus Torvalds 已提交
3237 3238 3239
		return SEQ_START_TOKEN;

	bond_for_each_slave(bond, slave, i) {
S
Stephen Hemminger 已提交
3240
		if (++off == *pos)
L
Linus Torvalds 已提交
3241 3242 3243 3244 3245 3246 3247 3248 3249 3250 3251 3252
			return slave;
	}

	return NULL;
}

static void *bond_info_seq_next(struct seq_file *seq, void *v, loff_t *pos)
{
	struct bonding *bond = seq->private;
	struct slave *slave = v;

	++*pos;
S
Stephen Hemminger 已提交
3253
	if (v == SEQ_START_TOKEN)
L
Linus Torvalds 已提交
3254 3255 3256 3257 3258 3259 3260 3261
		return bond->first_slave;

	slave = slave->next;

	return (slave == bond->first_slave) ? NULL : slave;
}

static void bond_info_seq_stop(struct seq_file *seq, void *v)
3262 3263
	__releases(&bond->lock)
	__releases(&dev_base_lock)
L
Linus Torvalds 已提交
3264 3265 3266
{
	struct bonding *bond = seq->private;

3267
	read_unlock(&bond->lock);
L
Linus Torvalds 已提交
3268 3269 3270 3271 3272 3273 3274
	read_unlock(&dev_base_lock);
}

static void bond_info_show_master(struct seq_file *seq)
{
	struct bonding *bond = seq->private;
	struct slave *curr;
3275
	int i;
L
Linus Torvalds 已提交
3276 3277 3278 3279 3280

	read_lock(&bond->curr_slave_lock);
	curr = bond->curr_active_slave;
	read_unlock(&bond->curr_slave_lock);

3281
	seq_printf(seq, "Bonding Mode: %s",
L
Linus Torvalds 已提交
3282 3283
		   bond_mode_name(bond->params.mode));

3284 3285
	if (bond->params.mode == BOND_MODE_ACTIVEBACKUP &&
	    bond->params.fail_over_mac)
3286 3287
		seq_printf(seq, " (fail_over_mac %s)",
		   fail_over_mac_tbl[bond->params.fail_over_mac].modename);
3288 3289 3290

	seq_printf(seq, "\n");

3291 3292 3293 3294 3295 3296 3297
	if (bond->params.mode == BOND_MODE_XOR ||
		bond->params.mode == BOND_MODE_8023AD) {
		seq_printf(seq, "Transmit Hash Policy: %s (%d)\n",
			xmit_hashtype_tbl[bond->params.xmit_policy].modename,
			bond->params.xmit_policy);
	}

L
Linus Torvalds 已提交
3298
	if (USES_PRIMARY(bond->params.mode)) {
3299
		seq_printf(seq, "Primary Slave: %s",
3300 3301
			   (bond->primary_slave) ?
			   bond->primary_slave->dev->name : "None");
3302 3303 3304
		if (bond->primary_slave)
			seq_printf(seq, " (primary_reselect %s)",
		   pri_reselect_tbl[bond->params.primary_reselect].modename);
L
Linus Torvalds 已提交
3305

3306
		seq_printf(seq, "\nCurrently Active Slave: %s\n",
L
Linus Torvalds 已提交
3307 3308 3309
			   (curr) ? curr->dev->name : "None");
	}

3310 3311
	seq_printf(seq, "MII Status: %s\n", netif_carrier_ok(bond->dev) ?
		   "up" : "down");
L
Linus Torvalds 已提交
3312 3313 3314 3315 3316 3317
	seq_printf(seq, "MII Polling Interval (ms): %d\n", bond->params.miimon);
	seq_printf(seq, "Up Delay (ms): %d\n",
		   bond->params.updelay * bond->params.miimon);
	seq_printf(seq, "Down Delay (ms): %d\n",
		   bond->params.downdelay * bond->params.miimon);

3318 3319

	/* ARP information */
S
Stephen Hemminger 已提交
3320 3321
	if (bond->params.arp_interval > 0) {
		int printed = 0;
3322 3323 3324 3325 3326
		seq_printf(seq, "ARP Polling Interval (ms): %d\n",
				bond->params.arp_interval);

		seq_printf(seq, "ARP IP target/s (n.n.n.n form):");

S
Stephen Hemminger 已提交
3327
		for (i = 0; (i < BOND_MAX_ARP_TARGETS); i++) {
3328
			if (!bond->params.arp_targets[i])
3329
				break;
3330 3331
			if (printed)
				seq_printf(seq, ",");
3332
			seq_printf(seq, " %pI4", &bond->params.arp_targets[i]);
3333 3334 3335 3336 3337
			printed = 1;
		}
		seq_printf(seq, "\n");
	}

L
Linus Torvalds 已提交
3338 3339 3340 3341 3342 3343
	if (bond->params.mode == BOND_MODE_8023AD) {
		struct ad_info ad_info;

		seq_puts(seq, "\n802.3ad info\n");
		seq_printf(seq, "LACP rate: %s\n",
			   (bond->params.lacp_fast) ? "fast" : "slow");
3344 3345
		seq_printf(seq, "Aggregator selection policy (ad_select): %s\n",
			   ad_select_tbl[bond->params.ad_select].modename);
L
Linus Torvalds 已提交
3346 3347 3348 3349 3350 3351 3352 3353 3354 3355 3356 3357 3358 3359 3360

		if (bond_3ad_get_active_agg_info(bond, &ad_info)) {
			seq_printf(seq, "bond %s has no active aggregator\n",
				   bond->dev->name);
		} else {
			seq_printf(seq, "Active Aggregator Info:\n");

			seq_printf(seq, "\tAggregator ID: %d\n",
				   ad_info.aggregator_id);
			seq_printf(seq, "\tNumber of ports: %d\n",
				   ad_info.ports);
			seq_printf(seq, "\tActor Key: %d\n",
				   ad_info.actor_key);
			seq_printf(seq, "\tPartner Key: %d\n",
				   ad_info.partner_key);
J
Johannes Berg 已提交
3361 3362
			seq_printf(seq, "\tPartner Mac Address: %pM\n",
				   ad_info.partner_system);
L
Linus Torvalds 已提交
3363 3364 3365 3366
		}
	}
}

S
Stephen Hemminger 已提交
3367 3368
static void bond_info_show_slave(struct seq_file *seq,
				 const struct slave *slave)
L
Linus Torvalds 已提交
3369 3370 3371 3372 3373 3374
{
	struct bonding *bond = seq->private;

	seq_printf(seq, "\nSlave Interface: %s\n", slave->dev->name);
	seq_printf(seq, "MII Status: %s\n",
		   (slave->link == BOND_LINK_UP) ?  "up" : "down");
3375 3376
	seq_printf(seq, "Speed: %d Mbps\n", slave->speed);
	seq_printf(seq, "Duplex: %s\n", slave->duplex ? "full" : "half");
3377
	seq_printf(seq, "Link Failure Count: %u\n",
L
Linus Torvalds 已提交
3378 3379
		   slave->link_failure_count);

J
Johannes Berg 已提交
3380
	seq_printf(seq, "Permanent HW addr: %pM\n", slave->perm_hwaddr);
L
Linus Torvalds 已提交
3381 3382 3383 3384 3385

	if (bond->params.mode == BOND_MODE_8023AD) {
		const struct aggregator *agg
			= SLAVE_AD_INFO(slave).port.aggregator;

S
Stephen Hemminger 已提交
3386
		if (agg)
L
Linus Torvalds 已提交
3387 3388
			seq_printf(seq, "Aggregator ID: %d\n",
				   agg->aggregator_identifier);
S
Stephen Hemminger 已提交
3389
		else
L
Linus Torvalds 已提交
3390 3391
			seq_puts(seq, "Aggregator ID: N/A\n");
	}
3392
	seq_printf(seq, "Slave queue ID: %d\n", slave->queue_id);
L
Linus Torvalds 已提交
3393 3394 3395 3396 3397 3398 3399
}

static int bond_info_seq_show(struct seq_file *seq, void *v)
{
	if (v == SEQ_START_TOKEN) {
		seq_printf(seq, "%s\n", version);
		bond_info_show_master(seq);
S
Stephen Hemminger 已提交
3400
	} else
L
Linus Torvalds 已提交
3401 3402 3403 3404 3405
		bond_info_show_slave(seq, v);

	return 0;
}

J
Jan Engelhardt 已提交
3406
static const struct seq_operations bond_info_seq_ops = {
L
Linus Torvalds 已提交
3407 3408 3409 3410 3411 3412 3413 3414 3415 3416 3417 3418 3419 3420 3421 3422 3423 3424 3425 3426 3427 3428 3429
	.start = bond_info_seq_start,
	.next  = bond_info_seq_next,
	.stop  = bond_info_seq_stop,
	.show  = bond_info_seq_show,
};

static int bond_info_open(struct inode *inode, struct file *file)
{
	struct seq_file *seq;
	struct proc_dir_entry *proc;
	int res;

	res = seq_open(file, &bond_info_seq_ops);
	if (!res) {
		/* recover the pointer buried in proc_dir_entry data */
		seq = file->private_data;
		proc = PDE(inode);
		seq->private = proc->data;
	}

	return res;
}

3430
static const struct file_operations bond_info_fops = {
L
Linus Torvalds 已提交
3431 3432 3433 3434 3435 3436 3437
	.owner   = THIS_MODULE,
	.open    = bond_info_open,
	.read    = seq_read,
	.llseek  = seq_lseek,
	.release = seq_release,
};

3438
static void bond_create_proc_entry(struct bonding *bond)
L
Linus Torvalds 已提交
3439 3440
{
	struct net_device *bond_dev = bond->dev;
3441
	struct bond_net *bn = net_generic(dev_net(bond_dev), bond_net_id);
L
Linus Torvalds 已提交
3442

3443
	if (bn->proc_dir) {
3444
		bond->proc_entry = proc_create_data(bond_dev->name,
3445
						    S_IRUGO, bn->proc_dir,
3446
						    &bond_info_fops, bond);
S
Stephen Hemminger 已提交
3447
		if (bond->proc_entry == NULL)
J
Joe Perches 已提交
3448 3449
			pr_warning("Warning: Cannot create /proc/net/%s/%s\n",
				   DRV_NAME, bond_dev->name);
S
Stephen Hemminger 已提交
3450
		else
L
Linus Torvalds 已提交
3451 3452 3453 3454 3455 3456
			memcpy(bond->proc_file_name, bond_dev->name, IFNAMSIZ);
	}
}

static void bond_remove_proc_entry(struct bonding *bond)
{
3457 3458 3459 3460 3461
	struct net_device *bond_dev = bond->dev;
	struct bond_net *bn = net_generic(dev_net(bond_dev), bond_net_id);

	if (bn->proc_dir && bond->proc_entry) {
		remove_proc_entry(bond->proc_file_name, bn->proc_dir);
L
Linus Torvalds 已提交
3462 3463 3464 3465 3466 3467 3468 3469
		memset(bond->proc_file_name, 0, IFNAMSIZ);
		bond->proc_entry = NULL;
	}
}

/* Create the bonding directory under /proc/net, if doesn't exist yet.
 * Caller must hold rtnl_lock.
 */
3470
static void __net_init bond_create_proc_dir(struct bond_net *bn)
L
Linus Torvalds 已提交
3471
{
3472 3473 3474
	if (!bn->proc_dir) {
		bn->proc_dir = proc_mkdir(DRV_NAME, bn->net->proc_net);
		if (!bn->proc_dir)
J
Joe Perches 已提交
3475 3476
			pr_warning("Warning: cannot create /proc/net/%s\n",
				   DRV_NAME);
L
Linus Torvalds 已提交
3477 3478 3479 3480 3481 3482
	}
}

/* Destroy the bonding directory under /proc/net, if empty.
 * Caller must hold rtnl_lock.
 */
3483
static void __net_exit bond_destroy_proc_dir(struct bond_net *bn)
L
Linus Torvalds 已提交
3484
{
3485 3486 3487
	if (bn->proc_dir) {
		remove_proc_entry(DRV_NAME, bn->net->proc_net);
		bn->proc_dir = NULL;
L
Linus Torvalds 已提交
3488 3489
	}
}
3490 3491 3492

#else /* !CONFIG_PROC_FS */

3493
static void bond_create_proc_entry(struct bonding *bond)
3494 3495 3496 3497 3498 3499 3500
{
}

static void bond_remove_proc_entry(struct bonding *bond)
{
}

3501
static inline void bond_create_proc_dir(struct bond_net *bn)
3502 3503 3504
{
}

3505
static inline void bond_destroy_proc_dir(struct bond_net *bn)
3506 3507 3508
{
}

L
Linus Torvalds 已提交
3509 3510
#endif /* CONFIG_PROC_FS */

3511

L
Linus Torvalds 已提交
3512 3513 3514 3515 3516 3517 3518 3519 3520
/*-------------------------- netdev event handling --------------------------*/

/*
 * Change device name
 */
static int bond_event_changename(struct bonding *bond)
{
	bond_remove_proc_entry(bond);
	bond_create_proc_entry(bond);
3521

L
Linus Torvalds 已提交
3522 3523 3524
	return NOTIFY_DONE;
}

S
Stephen Hemminger 已提交
3525 3526
static int bond_master_netdev_event(unsigned long event,
				    struct net_device *bond_dev)
L
Linus Torvalds 已提交
3527
{
3528
	struct bonding *event_bond = netdev_priv(bond_dev);
L
Linus Torvalds 已提交
3529 3530 3531 3532 3533 3534 3535 3536 3537 3538 3539

	switch (event) {
	case NETDEV_CHANGENAME:
		return bond_event_changename(event_bond);
	default:
		break;
	}

	return NOTIFY_DONE;
}

S
Stephen Hemminger 已提交
3540 3541
static int bond_slave_netdev_event(unsigned long event,
				   struct net_device *slave_dev)
L
Linus Torvalds 已提交
3542 3543
{
	struct net_device *bond_dev = slave_dev->master;
3544
	struct bonding *bond = netdev_priv(bond_dev);
L
Linus Torvalds 已提交
3545 3546 3547 3548

	switch (event) {
	case NETDEV_UNREGISTER:
		if (bond_dev) {
3549 3550 3551 3552
			if (bond->setup_by_slave)
				bond_release_and_destroy(bond_dev, slave_dev);
			else
				bond_release(bond_dev, slave_dev);
L
Linus Torvalds 已提交
3553 3554 3555
		}
		break;
	case NETDEV_CHANGE:
3556 3557 3558 3559 3560 3561 3562 3563 3564 3565 3566 3567 3568 3569 3570 3571 3572 3573 3574 3575
		if (bond->params.mode == BOND_MODE_8023AD || bond_is_lb(bond)) {
			struct slave *slave;

			slave = bond_get_slave_by_dev(bond, slave_dev);
			if (slave) {
				u16 old_speed = slave->speed;
				u16 old_duplex = slave->duplex;

				bond_update_speed_duplex(slave);

				if (bond_is_lb(bond))
					break;

				if (old_speed != slave->speed)
					bond_3ad_adapter_speed_changed(slave);
				if (old_duplex != slave->duplex)
					bond_3ad_adapter_duplex_changed(slave);
			}
		}

L
Linus Torvalds 已提交
3576 3577 3578 3579 3580 3581 3582 3583 3584 3585 3586 3587 3588 3589 3590 3591 3592 3593 3594 3595 3596 3597 3598 3599 3600
		break;
	case NETDEV_DOWN:
		/*
		 * ... Or is it this?
		 */
		break;
	case NETDEV_CHANGEMTU:
		/*
		 * TODO: Should slaves be allowed to
		 * independently alter their MTU?  For
		 * an active-backup bond, slaves need
		 * not be the same type of device, so
		 * MTUs may vary.  For other modes,
		 * slaves arguably should have the
		 * same MTUs. To do this, we'd need to
		 * take over the slave's change_mtu
		 * function for the duration of their
		 * servitude.
		 */
		break;
	case NETDEV_CHANGENAME:
		/*
		 * TODO: handle changing the primary's name
		 */
		break;
3601 3602 3603
	case NETDEV_FEAT_CHANGE:
		bond_compute_features(bond);
		break;
L
Linus Torvalds 已提交
3604 3605 3606 3607 3608 3609 3610 3611 3612 3613 3614
	default:
		break;
	}

	return NOTIFY_DONE;
}

/*
 * bond_netdev_event: handle netdev notifier chain events.
 *
 * This function receives events for the netdev chain.  The caller (an
3615
 * ioctl handler calling blocking_notifier_call_chain) holds the necessary
L
Linus Torvalds 已提交
3616 3617 3618
 * locks for us to safely manipulate the slave devices (RTNL lock,
 * dev_probe_lock).
 */
S
Stephen Hemminger 已提交
3619 3620
static int bond_netdev_event(struct notifier_block *this,
			     unsigned long event, void *ptr)
L
Linus Torvalds 已提交
3621 3622 3623
{
	struct net_device *event_dev = (struct net_device *)ptr;

3624
	pr_debug("event_dev: %s, event: %lx\n",
J
Joe Perches 已提交
3625 3626
		 event_dev ? event_dev->name : "None",
		 event);
L
Linus Torvalds 已提交
3627

3628 3629 3630
	if (!(event_dev->priv_flags & IFF_BONDING))
		return NOTIFY_DONE;

L
Linus Torvalds 已提交
3631
	if (event_dev->flags & IFF_MASTER) {
3632
		pr_debug("IFF_MASTER\n");
L
Linus Torvalds 已提交
3633 3634 3635 3636
		return bond_master_netdev_event(event, event_dev);
	}

	if (event_dev->flags & IFF_SLAVE) {
3637
		pr_debug("IFF_SLAVE\n");
L
Linus Torvalds 已提交
3638 3639 3640 3641 3642 3643
		return bond_slave_netdev_event(event, event_dev);
	}

	return NOTIFY_DONE;
}

J
Jay Vosburgh 已提交
3644 3645 3646 3647 3648 3649 3650 3651 3652 3653 3654 3655
/*
 * bond_inetaddr_event: handle inetaddr notifier chain events.
 *
 * We keep track of device IPs primarily to use as source addresses in
 * ARP monitor probes (rather than spewing out broadcasts all the time).
 *
 * We track one IP for the main device (if it has one), plus one per VLAN.
 */
static int bond_inetaddr_event(struct notifier_block *this, unsigned long event, void *ptr)
{
	struct in_ifaddr *ifa = ptr;
	struct net_device *vlan_dev, *event_dev = ifa->ifa_dev->dev;
3656
	struct bond_net *bn = net_generic(dev_net(event_dev), bond_net_id);
3657 3658
	struct bonding *bond;
	struct vlan_entry *vlan;
J
Jay Vosburgh 已提交
3659

3660
	list_for_each_entry(bond, &bn->dev_list, bond_list) {
J
Jay Vosburgh 已提交
3661 3662 3663 3664 3665 3666 3667 3668 3669 3670 3671 3672 3673
		if (bond->dev == event_dev) {
			switch (event) {
			case NETDEV_UP:
				bond->master_ip = ifa->ifa_local;
				return NOTIFY_OK;
			case NETDEV_DOWN:
				bond->master_ip = bond_glean_dev_ip(bond->dev);
				return NOTIFY_OK;
			default:
				return NOTIFY_DONE;
			}
		}

3674
		list_for_each_entry(vlan, &bond->vlan_list, vlan_list) {
3675 3676
			if (!bond->vlgrp)
				continue;
D
Dan Aloni 已提交
3677
			vlan_dev = vlan_group_get_device(bond->vlgrp, vlan->vlan_id);
J
Jay Vosburgh 已提交
3678 3679 3680 3681 3682 3683 3684 3685 3686 3687 3688 3689 3690 3691 3692 3693 3694 3695
			if (vlan_dev == event_dev) {
				switch (event) {
				case NETDEV_UP:
					vlan->vlan_ip = ifa->ifa_local;
					return NOTIFY_OK;
				case NETDEV_DOWN:
					vlan->vlan_ip =
						bond_glean_dev_ip(vlan_dev);
					return NOTIFY_OK;
				default:
					return NOTIFY_DONE;
				}
			}
		}
	}
	return NOTIFY_DONE;
}

L
Linus Torvalds 已提交
3696 3697 3698 3699
static struct notifier_block bond_netdev_notifier = {
	.notifier_call = bond_netdev_event,
};

J
Jay Vosburgh 已提交
3700 3701 3702 3703
static struct notifier_block bond_inetaddr_notifier = {
	.notifier_call = bond_inetaddr_event,
};

L
Linus Torvalds 已提交
3704 3705 3706 3707 3708 3709 3710 3711 3712 3713 3714 3715 3716 3717 3718 3719 3720 3721 3722 3723 3724
/*-------------------------- Packet type handling ---------------------------*/

/* register to receive lacpdus on a bond */
static void bond_register_lacpdu(struct bonding *bond)
{
	struct packet_type *pk_type = &(BOND_AD_INFO(bond).ad_pkt_type);

	/* initialize packet type */
	pk_type->type = PKT_TYPE_LACPDU;
	pk_type->dev = bond->dev;
	pk_type->func = bond_3ad_lacpdu_recv;

	dev_add_pack(pk_type);
}

/* unregister to receive lacpdus on a bond */
static void bond_unregister_lacpdu(struct bonding *bond)
{
	dev_remove_pack(&(BOND_AD_INFO(bond).ad_pkt_type));
}

3725 3726 3727 3728
void bond_register_arp(struct bonding *bond)
{
	struct packet_type *pt = &bond->arp_mon_pt;

J
Jay Vosburgh 已提交
3729 3730 3731
	if (pt->type)
		return;

3732
	pt->type = htons(ETH_P_ARP);
J
Jay Vosburgh 已提交
3733
	pt->dev = bond->dev;
3734 3735 3736 3737 3738 3739
	pt->func = bond_arp_rcv;
	dev_add_pack(pt);
}

void bond_unregister_arp(struct bonding *bond)
{
J
Jay Vosburgh 已提交
3740 3741 3742 3743
	struct packet_type *pt = &bond->arp_mon_pt;

	dev_remove_pack(pt);
	pt->type = 0;
3744 3745
}

3746 3747
/*---------------------------- Hashing Policies -----------------------------*/

3748 3749 3750 3751
/*
 * Hash for the output device based upon layer 2 and layer 3 data. If
 * the packet is not IP mimic bond_xmit_hash_policy_l2()
 */
3752
static int bond_xmit_hash_policy_l23(struct sk_buff *skb, int count)
3753 3754 3755 3756
{
	struct ethhdr *data = (struct ethhdr *)skb->data;
	struct iphdr *iph = ip_hdr(skb);

3757
	if (skb->protocol == htons(ETH_P_IP)) {
3758
		return ((ntohl(iph->saddr ^ iph->daddr) & 0xffff) ^
3759
			(data->h_dest[5] ^ data->h_source[5])) % count;
3760 3761
	}

3762
	return (data->h_dest[5] ^ data->h_source[5]) % count;
3763 3764
}

3765
/*
3766
 * Hash for the output device based upon layer 3 and layer 4 data. If
3767 3768 3769
 * the packet is a frag or not TCP or UDP, just use layer 3 data.  If it is
 * altogether not IP, mimic bond_xmit_hash_policy_l2()
 */
3770
static int bond_xmit_hash_policy_l34(struct sk_buff *skb, int count)
3771 3772
{
	struct ethhdr *data = (struct ethhdr *)skb->data;
3773
	struct iphdr *iph = ip_hdr(skb);
3774
	__be16 *layer4hdr = (__be16 *)((u32 *)iph + iph->ihl);
3775 3776
	int layer4_xor = 0;

3777 3778
	if (skb->protocol == htons(ETH_P_IP)) {
		if (!(iph->frag_off & htons(IP_MF|IP_OFFSET)) &&
3779 3780
		    (iph->protocol == IPPROTO_TCP ||
		     iph->protocol == IPPROTO_UDP)) {
3781
			layer4_xor = ntohs((*layer4hdr ^ *(layer4hdr + 1)));
3782 3783 3784 3785 3786 3787
		}
		return (layer4_xor ^
			((ntohl(iph->saddr ^ iph->daddr)) & 0xffff)) % count;

	}

3788
	return (data->h_dest[5] ^ data->h_source[5]) % count;
3789 3790 3791 3792 3793
}

/*
 * Hash for the output device based upon layer 2 data
 */
3794
static int bond_xmit_hash_policy_l2(struct sk_buff *skb, int count)
3795 3796 3797
{
	struct ethhdr *data = (struct ethhdr *)skb->data;

3798
	return (data->h_dest[5] ^ data->h_source[5]) % count;
3799 3800
}

L
Linus Torvalds 已提交
3801 3802 3803 3804
/*-------------------------- Device entry points ----------------------------*/

static int bond_open(struct net_device *bond_dev)
{
3805
	struct bonding *bond = netdev_priv(bond_dev);
L
Linus Torvalds 已提交
3806 3807 3808

	bond->kill_timers = 0;

3809 3810
	INIT_DELAYED_WORK(&bond->mcast_work, bond_resend_igmp_join_requests_delayed);

3811
	if (bond_is_lb(bond)) {
L
Linus Torvalds 已提交
3812 3813 3814 3815 3816
		/* bond_alb_initialize must be called before the timer
		 * is started.
		 */
		if (bond_alb_initialize(bond, (bond->params.mode == BOND_MODE_ALB))) {
			/* something went wrong - fail the open operation */
3817
			return -ENOMEM;
L
Linus Torvalds 已提交
3818 3819
		}

3820 3821
		INIT_DELAYED_WORK(&bond->alb_work, bond_alb_monitor);
		queue_delayed_work(bond->wq, &bond->alb_work, 0);
L
Linus Torvalds 已提交
3822 3823 3824
	}

	if (bond->params.miimon) {  /* link check interval, in milliseconds. */
3825 3826
		INIT_DELAYED_WORK(&bond->mii_work, bond_mii_monitor);
		queue_delayed_work(bond->wq, &bond->mii_work, 0);
L
Linus Torvalds 已提交
3827 3828 3829
	}

	if (bond->params.arp_interval) {  /* arp interval, in milliseconds. */
3830 3831 3832 3833 3834 3835 3836 3837
		if (bond->params.mode == BOND_MODE_ACTIVEBACKUP)
			INIT_DELAYED_WORK(&bond->arp_work,
					  bond_activebackup_arp_mon);
		else
			INIT_DELAYED_WORK(&bond->arp_work,
					  bond_loadbalance_arp_mon);

		queue_delayed_work(bond->wq, &bond->arp_work, 0);
3838 3839
		if (bond->params.arp_validate)
			bond_register_arp(bond);
L
Linus Torvalds 已提交
3840 3841 3842
	}

	if (bond->params.mode == BOND_MODE_8023AD) {
3843
		INIT_DELAYED_WORK(&bond->ad_work, bond_3ad_state_machine_handler);
3844
		queue_delayed_work(bond->wq, &bond->ad_work, 0);
L
Linus Torvalds 已提交
3845 3846
		/* register to receive LACPDUs */
		bond_register_lacpdu(bond);
3847
		bond_3ad_initiate_agg_selection(bond, 1);
L
Linus Torvalds 已提交
3848 3849 3850 3851 3852 3853 3854
	}

	return 0;
}

static int bond_close(struct net_device *bond_dev)
{
3855
	struct bonding *bond = netdev_priv(bond_dev);
L
Linus Torvalds 已提交
3856 3857 3858 3859 3860 3861

	if (bond->params.mode == BOND_MODE_8023AD) {
		/* Unregister the receive of LACPDUs */
		bond_unregister_lacpdu(bond);
	}

3862 3863 3864
	if (bond->params.arp_validate)
		bond_unregister_arp(bond);

L
Linus Torvalds 已提交
3865 3866
	write_lock_bh(&bond->lock);

3867
	bond->send_grat_arp = 0;
3868
	bond->send_unsol_na = 0;
L
Linus Torvalds 已提交
3869 3870 3871 3872 3873 3874 3875

	/* signal timers not to re-arm */
	bond->kill_timers = 1;

	write_unlock_bh(&bond->lock);

	if (bond->params.miimon) {  /* link check interval, in milliseconds. */
3876
		cancel_delayed_work(&bond->mii_work);
L
Linus Torvalds 已提交
3877 3878 3879
	}

	if (bond->params.arp_interval) {  /* arp interval, in milliseconds. */
3880
		cancel_delayed_work(&bond->arp_work);
L
Linus Torvalds 已提交
3881 3882 3883 3884
	}

	switch (bond->params.mode) {
	case BOND_MODE_8023AD:
3885
		cancel_delayed_work(&bond->ad_work);
L
Linus Torvalds 已提交
3886 3887 3888
		break;
	case BOND_MODE_TLB:
	case BOND_MODE_ALB:
3889
		cancel_delayed_work(&bond->alb_work);
L
Linus Torvalds 已提交
3890 3891 3892 3893 3894
		break;
	default:
		break;
	}

3895 3896
	if (delayed_work_pending(&bond->mcast_work))
		cancel_delayed_work(&bond->mcast_work);
L
Linus Torvalds 已提交
3897

3898
	if (bond_is_lb(bond)) {
L
Linus Torvalds 已提交
3899 3900 3901 3902 3903 3904 3905 3906 3907
		/* Must be called only after all
		 * slaves have been released
		 */
		bond_alb_deinitialize(bond);
	}

	return 0;
}

3908 3909
static struct rtnl_link_stats64 *bond_get_stats(struct net_device *bond_dev,
						struct rtnl_link_stats64 *stats)
L
Linus Torvalds 已提交
3910
{
3911
	struct bonding *bond = netdev_priv(bond_dev);
3912
	struct rtnl_link_stats64 temp;
L
Linus Torvalds 已提交
3913 3914 3915
	struct slave *slave;
	int i;

3916
	memset(stats, 0, sizeof(*stats));
L
Linus Torvalds 已提交
3917 3918 3919 3920

	read_lock_bh(&bond->lock);

	bond_for_each_slave(bond, slave, i) {
3921
		const struct rtnl_link_stats64 *sstats =
3922 3923 3924 3925 3926 3927 3928 3929 3930 3931 3932 3933 3934 3935 3936 3937 3938 3939 3940 3941 3942 3943 3944 3945 3946 3947 3948
			dev_get_stats(slave->dev, &temp);

		stats->rx_packets += sstats->rx_packets;
		stats->rx_bytes += sstats->rx_bytes;
		stats->rx_errors += sstats->rx_errors;
		stats->rx_dropped += sstats->rx_dropped;

		stats->tx_packets += sstats->tx_packets;
		stats->tx_bytes += sstats->tx_bytes;
		stats->tx_errors += sstats->tx_errors;
		stats->tx_dropped += sstats->tx_dropped;

		stats->multicast += sstats->multicast;
		stats->collisions += sstats->collisions;

		stats->rx_length_errors += sstats->rx_length_errors;
		stats->rx_over_errors += sstats->rx_over_errors;
		stats->rx_crc_errors += sstats->rx_crc_errors;
		stats->rx_frame_errors += sstats->rx_frame_errors;
		stats->rx_fifo_errors += sstats->rx_fifo_errors;
		stats->rx_missed_errors += sstats->rx_missed_errors;

		stats->tx_aborted_errors += sstats->tx_aborted_errors;
		stats->tx_carrier_errors += sstats->tx_carrier_errors;
		stats->tx_fifo_errors += sstats->tx_fifo_errors;
		stats->tx_heartbeat_errors += sstats->tx_heartbeat_errors;
		stats->tx_window_errors += sstats->tx_window_errors;
3949 3950
	}

L
Linus Torvalds 已提交
3951 3952 3953 3954 3955 3956 3957 3958 3959 3960 3961 3962 3963 3964 3965
	read_unlock_bh(&bond->lock);

	return stats;
}

static int bond_do_ioctl(struct net_device *bond_dev, struct ifreq *ifr, int cmd)
{
	struct net_device *slave_dev = NULL;
	struct ifbond k_binfo;
	struct ifbond __user *u_binfo = NULL;
	struct ifslave k_sinfo;
	struct ifslave __user *u_sinfo = NULL;
	struct mii_ioctl_data *mii = NULL;
	int res = 0;

J
Joe Perches 已提交
3966
	pr_debug("bond_ioctl: master=%s, cmd=%d\n", bond_dev->name, cmd);
L
Linus Torvalds 已提交
3967 3968 3969 3970

	switch (cmd) {
	case SIOCGMIIPHY:
		mii = if_mii(ifr);
S
Stephen Hemminger 已提交
3971
		if (!mii)
L
Linus Torvalds 已提交
3972
			return -EINVAL;
S
Stephen Hemminger 已提交
3973

L
Linus Torvalds 已提交
3974 3975 3976 3977 3978 3979 3980 3981
		mii->phy_id = 0;
		/* Fall Through */
	case SIOCGMIIREG:
		/*
		 * We do this again just in case we were called by SIOCGMIIREG
		 * instead of SIOCGMIIPHY.
		 */
		mii = if_mii(ifr);
S
Stephen Hemminger 已提交
3982
		if (!mii)
L
Linus Torvalds 已提交
3983
			return -EINVAL;
S
Stephen Hemminger 已提交
3984

L
Linus Torvalds 已提交
3985 3986

		if (mii->reg_num == 1) {
3987
			struct bonding *bond = netdev_priv(bond_dev);
L
Linus Torvalds 已提交
3988
			mii->val_out = 0;
3989
			read_lock(&bond->lock);
L
Linus Torvalds 已提交
3990
			read_lock(&bond->curr_slave_lock);
S
Stephen Hemminger 已提交
3991
			if (netif_carrier_ok(bond->dev))
L
Linus Torvalds 已提交
3992
				mii->val_out = BMSR_LSTATUS;
S
Stephen Hemminger 已提交
3993

L
Linus Torvalds 已提交
3994
			read_unlock(&bond->curr_slave_lock);
3995
			read_unlock(&bond->lock);
L
Linus Torvalds 已提交
3996 3997 3998 3999 4000 4001 4002
		}

		return 0;
	case BOND_INFO_QUERY_OLD:
	case SIOCBONDINFOQUERY:
		u_binfo = (struct ifbond __user *)ifr->ifr_data;

S
Stephen Hemminger 已提交
4003
		if (copy_from_user(&k_binfo, u_binfo, sizeof(ifbond)))
L
Linus Torvalds 已提交
4004 4005 4006
			return -EFAULT;

		res = bond_info_query(bond_dev, &k_binfo);
S
Stephen Hemminger 已提交
4007 4008 4009
		if (res == 0 &&
		    copy_to_user(u_binfo, &k_binfo, sizeof(ifbond)))
			return -EFAULT;
L
Linus Torvalds 已提交
4010 4011 4012 4013 4014 4015

		return res;
	case BOND_SLAVE_INFO_QUERY_OLD:
	case SIOCBONDSLAVEINFOQUERY:
		u_sinfo = (struct ifslave __user *)ifr->ifr_data;

S
Stephen Hemminger 已提交
4016
		if (copy_from_user(&k_sinfo, u_sinfo, sizeof(ifslave)))
L
Linus Torvalds 已提交
4017 4018 4019
			return -EFAULT;

		res = bond_slave_info_query(bond_dev, &k_sinfo);
S
Stephen Hemminger 已提交
4020 4021 4022
		if (res == 0 &&
		    copy_to_user(u_sinfo, &k_sinfo, sizeof(ifslave)))
			return -EFAULT;
L
Linus Torvalds 已提交
4023 4024 4025 4026 4027 4028 4029

		return res;
	default:
		/* Go on */
		break;
	}

S
Stephen Hemminger 已提交
4030
	if (!capable(CAP_NET_ADMIN))
L
Linus Torvalds 已提交
4031 4032
		return -EPERM;

4033
	slave_dev = dev_get_by_name(dev_net(bond_dev), ifr->ifr_slave);
L
Linus Torvalds 已提交
4034

J
Joe Perches 已提交
4035
	pr_debug("slave_dev=%p:\n", slave_dev);
L
Linus Torvalds 已提交
4036

S
Stephen Hemminger 已提交
4037
	if (!slave_dev)
L
Linus Torvalds 已提交
4038
		res = -ENODEV;
S
Stephen Hemminger 已提交
4039
	else {
J
Joe Perches 已提交
4040
		pr_debug("slave_dev->name=%s:\n", slave_dev->name);
L
Linus Torvalds 已提交
4041 4042 4043 4044 4045 4046 4047 4048 4049 4050 4051 4052 4053 4054 4055 4056 4057 4058 4059 4060 4061 4062 4063 4064 4065 4066 4067
		switch (cmd) {
		case BOND_ENSLAVE_OLD:
		case SIOCBONDENSLAVE:
			res = bond_enslave(bond_dev, slave_dev);
			break;
		case BOND_RELEASE_OLD:
		case SIOCBONDRELEASE:
			res = bond_release(bond_dev, slave_dev);
			break;
		case BOND_SETHWADDR_OLD:
		case SIOCBONDSETHWADDR:
			res = bond_sethwaddr(bond_dev, slave_dev);
			break;
		case BOND_CHANGE_ACTIVE_OLD:
		case SIOCBONDCHANGEACTIVE:
			res = bond_ioctl_change_active(bond_dev, slave_dev);
			break;
		default:
			res = -EOPNOTSUPP;
		}

		dev_put(slave_dev);
	}

	return res;
}

4068 4069 4070 4071 4072 4073 4074 4075 4076 4077 4078 4079 4080
static bool bond_addr_in_mc_list(unsigned char *addr,
				 struct netdev_hw_addr_list *list,
				 int addrlen)
{
	struct netdev_hw_addr *ha;

	netdev_hw_addr_list_for_each(ha, list)
		if (!memcmp(ha->addr, addr, addrlen))
			return true;

	return false;
}

L
Linus Torvalds 已提交
4081 4082
static void bond_set_multicast_list(struct net_device *bond_dev)
{
4083
	struct bonding *bond = netdev_priv(bond_dev);
4084 4085
	struct netdev_hw_addr *ha;
	bool found;
L
Linus Torvalds 已提交
4086 4087 4088 4089

	/*
	 * Do promisc before checking multicast_mode
	 */
S
Stephen Hemminger 已提交
4090
	if ((bond_dev->flags & IFF_PROMISC) && !(bond->flags & IFF_PROMISC))
4091 4092 4093 4094
		/*
		 * FIXME: Need to handle the error when one of the multi-slaves
		 * encounters error.
		 */
L
Linus Torvalds 已提交
4095 4096
		bond_set_promiscuity(bond, 1);

S
Stephen Hemminger 已提交
4097 4098

	if (!(bond_dev->flags & IFF_PROMISC) && (bond->flags & IFF_PROMISC))
L
Linus Torvalds 已提交
4099
		bond_set_promiscuity(bond, -1);
S
Stephen Hemminger 已提交
4100

L
Linus Torvalds 已提交
4101 4102

	/* set allmulti flag to slaves */
S
Stephen Hemminger 已提交
4103
	if ((bond_dev->flags & IFF_ALLMULTI) && !(bond->flags & IFF_ALLMULTI))
4104 4105 4106 4107
		/*
		 * FIXME: Need to handle the error when one of the multi-slaves
		 * encounters error.
		 */
L
Linus Torvalds 已提交
4108 4109
		bond_set_allmulti(bond, 1);

S
Stephen Hemminger 已提交
4110 4111

	if (!(bond_dev->flags & IFF_ALLMULTI) && (bond->flags & IFF_ALLMULTI))
L
Linus Torvalds 已提交
4112
		bond_set_allmulti(bond, -1);
S
Stephen Hemminger 已提交
4113

L
Linus Torvalds 已提交
4114

4115 4116
	read_lock(&bond->lock);

L
Linus Torvalds 已提交
4117 4118 4119
	bond->flags = bond_dev->flags;

	/* looking for addresses to add to slaves' mc list */
4120 4121 4122 4123 4124
	netdev_for_each_mc_addr(ha, bond_dev) {
		found = bond_addr_in_mc_list(ha->addr, &bond->mc_list,
					     bond_dev->addr_len);
		if (!found)
			bond_mc_add(bond, ha->addr);
L
Linus Torvalds 已提交
4125 4126 4127
	}

	/* looking for addresses to delete from slaves' list */
4128 4129 4130 4131 4132
	netdev_hw_addr_list_for_each(ha, &bond->mc_list) {
		found = bond_addr_in_mc_list(ha->addr, &bond_dev->mc,
					     bond_dev->addr_len);
		if (!found)
			bond_mc_del(bond, ha->addr);
L
Linus Torvalds 已提交
4133 4134 4135
	}

	/* save master's multicast list */
4136 4137 4138
	__hw_addr_flush(&bond->mc_list);
	__hw_addr_add_multiple(&bond->mc_list, &bond_dev->mc,
			       bond_dev->addr_len, NETDEV_HW_ADDR_T_MULTICAST);
L
Linus Torvalds 已提交
4139

4140
	read_unlock(&bond->lock);
L
Linus Torvalds 已提交
4141 4142
}

4143 4144 4145 4146 4147 4148 4149 4150 4151
static int bond_neigh_setup(struct net_device *dev, struct neigh_parms *parms)
{
	struct bonding *bond = netdev_priv(dev);
	struct slave *slave = bond->first_slave;

	if (slave) {
		const struct net_device_ops *slave_ops
			= slave->dev->netdev_ops;
		if (slave_ops->ndo_neigh_setup)
4152
			return slave_ops->ndo_neigh_setup(slave->dev, parms);
4153 4154 4155 4156
	}
	return 0;
}

L
Linus Torvalds 已提交
4157 4158 4159 4160 4161
/*
 * Change the MTU of all of a master's slaves to match the master
 */
static int bond_change_mtu(struct net_device *bond_dev, int new_mtu)
{
4162
	struct bonding *bond = netdev_priv(bond_dev);
L
Linus Torvalds 已提交
4163 4164 4165 4166
	struct slave *slave, *stop_at;
	int res = 0;
	int i;

4167
	pr_debug("bond=%p, name=%s, new_mtu=%d\n", bond,
J
Joe Perches 已提交
4168
		 (bond_dev ? bond_dev->name : "None"), new_mtu);
L
Linus Torvalds 已提交
4169 4170 4171 4172 4173 4174 4175 4176 4177 4178 4179 4180 4181 4182 4183 4184 4185

	/* Can't hold bond->lock with bh disabled here since
	 * some base drivers panic. On the other hand we can't
	 * hold bond->lock without bh disabled because we'll
	 * deadlock. The only solution is to rely on the fact
	 * that we're under rtnl_lock here, and the slaves
	 * list won't change. This doesn't solve the problem
	 * of setting the slave's MTU while it is
	 * transmitting, but the assumption is that the base
	 * driver can handle that.
	 *
	 * TODO: figure out a way to safely iterate the slaves
	 * list, but without holding a lock around the actual
	 * call to the base driver.
	 */

	bond_for_each_slave(bond, slave, i) {
J
Joe Perches 已提交
4186 4187 4188 4189
		pr_debug("s %p s->p %p c_m %p\n",
			 slave,
			 slave->prev,
			 slave->dev->netdev_ops->ndo_change_mtu);
4190

L
Linus Torvalds 已提交
4191 4192 4193 4194 4195 4196 4197 4198 4199 4200 4201
		res = dev_set_mtu(slave->dev, new_mtu);

		if (res) {
			/* If we failed to set the slave's mtu to the new value
			 * we must abort the operation even in ACTIVE_BACKUP
			 * mode, because if we allow the backup slaves to have
			 * different mtu values than the active slave we'll
			 * need to change their mtu when doing a failover. That
			 * means changing their mtu from timer context, which
			 * is probably not a good idea.
			 */
4202
			pr_debug("err %d %s\n", res, slave->dev->name);
L
Linus Torvalds 已提交
4203 4204 4205 4206 4207 4208 4209 4210 4211 4212 4213 4214 4215 4216 4217 4218
			goto unwind;
		}
	}

	bond_dev->mtu = new_mtu;

	return 0;

unwind:
	/* unwind from head to the slave that failed */
	stop_at = slave;
	bond_for_each_slave_from_to(bond, slave, i, bond->first_slave, stop_at) {
		int tmp_res;

		tmp_res = dev_set_mtu(slave->dev, bond_dev->mtu);
		if (tmp_res) {
J
Joe Perches 已提交
4219 4220
			pr_debug("unwind err %d dev %s\n",
				 tmp_res, slave->dev->name);
L
Linus Torvalds 已提交
4221 4222 4223 4224 4225 4226 4227 4228 4229 4230 4231 4232 4233 4234 4235
		}
	}

	return res;
}

/*
 * Change HW address
 *
 * Note that many devices must be down to change the HW address, and
 * downing the master releases all slaves.  We can make bonds full of
 * bonding devices to test this, however.
 */
static int bond_set_mac_address(struct net_device *bond_dev, void *addr)
{
4236
	struct bonding *bond = netdev_priv(bond_dev);
L
Linus Torvalds 已提交
4237 4238 4239 4240 4241
	struct sockaddr *sa = addr, tmp_sa;
	struct slave *slave, *stop_at;
	int res = 0;
	int i;

4242 4243 4244 4245
	if (bond->params.mode == BOND_MODE_ALB)
		return bond_alb_set_mac_address(bond_dev, addr);


J
Joe Perches 已提交
4246 4247
	pr_debug("bond=%p, name=%s\n",
		 bond, bond_dev ? bond_dev->name : "None");
L
Linus Torvalds 已提交
4248

4249
	/*
4250 4251
	 * If fail_over_mac is set to active, do nothing and return
	 * success.  Returning an error causes ifenslave to fail.
4252
	 */
4253
	if (bond->params.fail_over_mac == BOND_FOM_ACTIVE)
4254
		return 0;
4255

S
Stephen Hemminger 已提交
4256
	if (!is_valid_ether_addr(sa->sa_data))
L
Linus Torvalds 已提交
4257 4258 4259 4260 4261 4262 4263 4264 4265 4266 4267 4268 4269 4270 4271 4272 4273 4274
		return -EADDRNOTAVAIL;

	/* Can't hold bond->lock with bh disabled here since
	 * some base drivers panic. On the other hand we can't
	 * hold bond->lock without bh disabled because we'll
	 * deadlock. The only solution is to rely on the fact
	 * that we're under rtnl_lock here, and the slaves
	 * list won't change. This doesn't solve the problem
	 * of setting the slave's hw address while it is
	 * transmitting, but the assumption is that the base
	 * driver can handle that.
	 *
	 * TODO: figure out a way to safely iterate the slaves
	 * list, but without holding a lock around the actual
	 * call to the base driver.
	 */

	bond_for_each_slave(bond, slave, i) {
4275
		const struct net_device_ops *slave_ops = slave->dev->netdev_ops;
4276
		pr_debug("slave %p %s\n", slave, slave->dev->name);
L
Linus Torvalds 已提交
4277

4278
		if (slave_ops->ndo_set_mac_address == NULL) {
L
Linus Torvalds 已提交
4279
			res = -EOPNOTSUPP;
4280
			pr_debug("EOPNOTSUPP %s\n", slave->dev->name);
L
Linus Torvalds 已提交
4281 4282 4283 4284 4285 4286 4287 4288 4289 4290 4291
			goto unwind;
		}

		res = dev_set_mac_address(slave->dev, addr);
		if (res) {
			/* TODO: consider downing the slave
			 * and retry ?
			 * User should expect communications
			 * breakage anyway until ARP finish
			 * updating, so...
			 */
4292
			pr_debug("err %d %s\n", res, slave->dev->name);
L
Linus Torvalds 已提交
4293 4294 4295 4296 4297 4298 4299 4300 4301 4302 4303 4304 4305 4306 4307 4308 4309 4310 4311
			goto unwind;
		}
	}

	/* success */
	memcpy(bond_dev->dev_addr, sa->sa_data, bond_dev->addr_len);
	return 0;

unwind:
	memcpy(tmp_sa.sa_data, bond_dev->dev_addr, bond_dev->addr_len);
	tmp_sa.sa_family = bond_dev->type;

	/* unwind from head to the slave that failed */
	stop_at = slave;
	bond_for_each_slave_from_to(bond, slave, i, bond->first_slave, stop_at) {
		int tmp_res;

		tmp_res = dev_set_mac_address(slave->dev, &tmp_sa);
		if (tmp_res) {
J
Joe Perches 已提交
4312 4313
			pr_debug("unwind err %d dev %s\n",
				 tmp_res, slave->dev->name);
L
Linus Torvalds 已提交
4314 4315 4316 4317 4318 4319 4320 4321
		}
	}

	return res;
}

static int bond_xmit_roundrobin(struct sk_buff *skb, struct net_device *bond_dev)
{
4322
	struct bonding *bond = netdev_priv(bond_dev);
L
Linus Torvalds 已提交
4323
	struct slave *slave, *start_at;
4324
	int i, slave_no, res = 1;
4325
	struct iphdr *iph = ip_hdr(skb);
L
Linus Torvalds 已提交
4326 4327 4328

	read_lock(&bond->lock);

S
Stephen Hemminger 已提交
4329
	if (!BOND_IS_OK(bond))
L
Linus Torvalds 已提交
4330
		goto out;
4331
	/*
4332 4333 4334 4335 4336
	 * Start with the curr_active_slave that joined the bond as the
	 * default for sending IGMP traffic.  For failover purposes one
	 * needs to maintain some consistency for the interface that will
	 * send the join/membership reports.  The curr_active_slave found
	 * will send all of this type of traffic.
4337
	 */
4338
	if ((iph->protocol == IPPROTO_IGMP) &&
4339
	    (skb->protocol == htons(ETH_P_IP))) {
L
Linus Torvalds 已提交
4340

4341 4342 4343 4344 4345 4346 4347 4348 4349 4350 4351 4352 4353 4354 4355 4356 4357 4358 4359
		read_lock(&bond->curr_slave_lock);
		slave = bond->curr_active_slave;
		read_unlock(&bond->curr_slave_lock);

		if (!slave)
			goto out;
	} else {
		/*
		 * Concurrent TX may collide on rr_tx_counter; we accept
		 * that as being rare enough not to justify using an
		 * atomic op here.
		 */
		slave_no = bond->rr_tx_counter++ % bond->slave_cnt;

		bond_for_each_slave(bond, slave, i) {
			slave_no--;
			if (slave_no < 0)
				break;
		}
L
Linus Torvalds 已提交
4360 4361
	}

4362
	start_at = slave;
L
Linus Torvalds 已提交
4363 4364 4365 4366 4367 4368 4369 4370 4371 4372 4373 4374 4375 4376 4377
	bond_for_each_slave_from(bond, slave, i, start_at) {
		if (IS_UP(slave->dev) &&
		    (slave->link == BOND_LINK_UP) &&
		    (slave->state == BOND_STATE_ACTIVE)) {
			res = bond_dev_queue_xmit(bond, skb, slave->dev);
			break;
		}
	}

out:
	if (res) {
		/* no suitable interface, frame not sent */
		dev_kfree_skb(skb);
	}
	read_unlock(&bond->lock);
4378
	return NETDEV_TX_OK;
L
Linus Torvalds 已提交
4379 4380
}

4381

L
Linus Torvalds 已提交
4382 4383 4384 4385 4386 4387
/*
 * in active-backup mode, we know that bond->curr_active_slave is always valid if
 * the bond has a usable interface.
 */
static int bond_xmit_activebackup(struct sk_buff *skb, struct net_device *bond_dev)
{
4388
	struct bonding *bond = netdev_priv(bond_dev);
L
Linus Torvalds 已提交
4389 4390 4391 4392 4393
	int res = 1;

	read_lock(&bond->lock);
	read_lock(&bond->curr_slave_lock);

S
Stephen Hemminger 已提交
4394
	if (!BOND_IS_OK(bond))
L
Linus Torvalds 已提交
4395 4396
		goto out;

4397 4398 4399 4400 4401
	if (!bond->curr_active_slave)
		goto out;

	res = bond_dev_queue_xmit(bond, skb, bond->curr_active_slave->dev);

L
Linus Torvalds 已提交
4402
out:
S
Stephen Hemminger 已提交
4403
	if (res)
L
Linus Torvalds 已提交
4404 4405
		/* no suitable interface, frame not sent */
		dev_kfree_skb(skb);
S
Stephen Hemminger 已提交
4406

L
Linus Torvalds 已提交
4407 4408
	read_unlock(&bond->curr_slave_lock);
	read_unlock(&bond->lock);
4409
	return NETDEV_TX_OK;
L
Linus Torvalds 已提交
4410 4411 4412
}

/*
4413 4414 4415
 * In bond_xmit_xor() , we determine the output device by using a pre-
 * determined xmit_hash_policy(), If the selected device is not enabled,
 * find the next active slave.
L
Linus Torvalds 已提交
4416 4417 4418
 */
static int bond_xmit_xor(struct sk_buff *skb, struct net_device *bond_dev)
{
4419
	struct bonding *bond = netdev_priv(bond_dev);
L
Linus Torvalds 已提交
4420 4421 4422 4423 4424 4425 4426
	struct slave *slave, *start_at;
	int slave_no;
	int i;
	int res = 1;

	read_lock(&bond->lock);

S
Stephen Hemminger 已提交
4427
	if (!BOND_IS_OK(bond))
L
Linus Torvalds 已提交
4428 4429
		goto out;

4430
	slave_no = bond->xmit_hash_policy(skb, bond->slave_cnt);
L
Linus Torvalds 已提交
4431 4432 4433

	bond_for_each_slave(bond, slave, i) {
		slave_no--;
S
Stephen Hemminger 已提交
4434
		if (slave_no < 0)
L
Linus Torvalds 已提交
4435 4436 4437 4438 4439 4440 4441 4442 4443 4444 4445 4446 4447 4448 4449 4450 4451 4452 4453 4454
			break;
	}

	start_at = slave;

	bond_for_each_slave_from(bond, slave, i, start_at) {
		if (IS_UP(slave->dev) &&
		    (slave->link == BOND_LINK_UP) &&
		    (slave->state == BOND_STATE_ACTIVE)) {
			res = bond_dev_queue_xmit(bond, skb, slave->dev);
			break;
		}
	}

out:
	if (res) {
		/* no suitable interface, frame not sent */
		dev_kfree_skb(skb);
	}
	read_unlock(&bond->lock);
4455
	return NETDEV_TX_OK;
L
Linus Torvalds 已提交
4456 4457 4458 4459 4460 4461 4462
}

/*
 * in broadcast mode, we send everything to all usable interfaces.
 */
static int bond_xmit_broadcast(struct sk_buff *skb, struct net_device *bond_dev)
{
4463
	struct bonding *bond = netdev_priv(bond_dev);
L
Linus Torvalds 已提交
4464 4465 4466 4467 4468 4469 4470
	struct slave *slave, *start_at;
	struct net_device *tx_dev = NULL;
	int i;
	int res = 1;

	read_lock(&bond->lock);

S
Stephen Hemminger 已提交
4471
	if (!BOND_IS_OK(bond))
L
Linus Torvalds 已提交
4472 4473 4474 4475 4476 4477
		goto out;

	read_lock(&bond->curr_slave_lock);
	start_at = bond->curr_active_slave;
	read_unlock(&bond->curr_slave_lock);

S
Stephen Hemminger 已提交
4478
	if (!start_at)
L
Linus Torvalds 已提交
4479 4480 4481 4482 4483 4484 4485 4486 4487
		goto out;

	bond_for_each_slave_from(bond, slave, i, start_at) {
		if (IS_UP(slave->dev) &&
		    (slave->link == BOND_LINK_UP) &&
		    (slave->state == BOND_STATE_ACTIVE)) {
			if (tx_dev) {
				struct sk_buff *skb2 = skb_clone(skb, GFP_ATOMIC);
				if (!skb2) {
J
Joe Perches 已提交
4488
					pr_err("%s: Error: bond_xmit_broadcast(): skb_clone() failed\n",
4489
					       bond_dev->name);
L
Linus Torvalds 已提交
4490 4491 4492 4493 4494 4495 4496 4497 4498 4499 4500 4501 4502
					continue;
				}

				res = bond_dev_queue_xmit(bond, skb2, tx_dev);
				if (res) {
					dev_kfree_skb(skb2);
					continue;
				}
			}
			tx_dev = slave->dev;
		}
	}

S
Stephen Hemminger 已提交
4503
	if (tx_dev)
L
Linus Torvalds 已提交
4504 4505 4506
		res = bond_dev_queue_xmit(bond, skb, tx_dev);

out:
S
Stephen Hemminger 已提交
4507
	if (res)
L
Linus Torvalds 已提交
4508 4509
		/* no suitable interface, frame not sent */
		dev_kfree_skb(skb);
S
Stephen Hemminger 已提交
4510

L
Linus Torvalds 已提交
4511 4512
	/* frame sent to all suitable interfaces */
	read_unlock(&bond->lock);
4513
	return NETDEV_TX_OK;
L
Linus Torvalds 已提交
4514 4515 4516 4517
}

/*------------------------- Device initialization ---------------------------*/

4518 4519 4520 4521 4522 4523 4524 4525 4526 4527 4528 4529 4530 4531 4532 4533
static void bond_set_xmit_hash_policy(struct bonding *bond)
{
	switch (bond->params.xmit_policy) {
	case BOND_XMIT_POLICY_LAYER23:
		bond->xmit_hash_policy = bond_xmit_hash_policy_l23;
		break;
	case BOND_XMIT_POLICY_LAYER34:
		bond->xmit_hash_policy = bond_xmit_hash_policy_l34;
		break;
	case BOND_XMIT_POLICY_LAYER2:
	default:
		bond->xmit_hash_policy = bond_xmit_hash_policy_l2;
		break;
	}
}

4534 4535 4536 4537 4538 4539 4540 4541 4542 4543 4544 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
/*
 * Lookup the slave that corresponds to a qid
 */
static inline int bond_slave_override(struct bonding *bond,
				      struct sk_buff *skb)
{
	int i, res = 1;
	struct slave *slave = NULL;
	struct slave *check_slave;

	read_lock(&bond->lock);

	if (!BOND_IS_OK(bond) || !skb->queue_mapping)
		goto out;

	/* Find out if any slaves have the same mapping as this skb. */
	bond_for_each_slave(bond, check_slave, i) {
		if (check_slave->queue_id == skb->queue_mapping) {
			slave = check_slave;
			break;
		}
	}

	/* If the slave isn't UP, use default transmit policy. */
	if (slave && slave->queue_id && IS_UP(slave->dev) &&
	    (slave->link == BOND_LINK_UP)) {
		res = bond_dev_queue_xmit(bond, skb, slave->dev);
	}

out:
	read_unlock(&bond->lock);
	return res;
}

static u16 bond_select_queue(struct net_device *dev, struct sk_buff *skb)
{
	/*
	 * This helper function exists to help dev_pick_tx get the correct
	 * destination queue.  Using a helper function skips the a call to
	 * skb_tx_hash and will put the skbs in the queue we expect on their
	 * way down to the bonding driver.
	 */
	return skb->queue_mapping;
}

4579
static netdev_tx_t bond_start_xmit(struct sk_buff *skb, struct net_device *dev)
4580
{
4581 4582
	struct bonding *bond = netdev_priv(dev);

4583 4584 4585 4586 4587 4588 4589
	/*
	 * If we risk deadlock from transmitting this in the
	 * netpoll path, tell netpoll to queue the frame for later tx
	 */
	if (is_netpoll_tx_blocked(dev))
		return NETDEV_TX_BUSY;

4590 4591 4592 4593
	if (TX_QUEUE_OVERRIDE(bond->params.mode)) {
		if (!bond_slave_override(bond, skb))
			return NETDEV_TX_OK;
	}
4594 4595 4596 4597 4598 4599 4600 4601 4602 4603 4604 4605 4606 4607 4608 4609 4610

	switch (bond->params.mode) {
	case BOND_MODE_ROUNDROBIN:
		return bond_xmit_roundrobin(skb, dev);
	case BOND_MODE_ACTIVEBACKUP:
		return bond_xmit_activebackup(skb, dev);
	case BOND_MODE_XOR:
		return bond_xmit_xor(skb, dev);
	case BOND_MODE_BROADCAST:
		return bond_xmit_broadcast(skb, dev);
	case BOND_MODE_8023AD:
		return bond_3ad_xmit_xor(skb, dev);
	case BOND_MODE_ALB:
	case BOND_MODE_TLB:
		return bond_alb_xmit(skb, dev);
	default:
		/* Should never happen, mode already checked */
J
Joe Perches 已提交
4611 4612
		pr_err("%s: Error: Unknown bonding mode %d\n",
		       dev->name, bond->params.mode);
4613 4614 4615 4616 4617 4618 4619
		WARN_ON_ONCE(1);
		dev_kfree_skb(skb);
		return NETDEV_TX_OK;
	}
}


L
Linus Torvalds 已提交
4620 4621 4622
/*
 * set bond mode specific net device operations
 */
4623
void bond_set_mode_ops(struct bonding *bond, int mode)
L
Linus Torvalds 已提交
4624
{
4625 4626
	struct net_device *bond_dev = bond->dev;

L
Linus Torvalds 已提交
4627 4628 4629 4630 4631 4632
	switch (mode) {
	case BOND_MODE_ROUNDROBIN:
		break;
	case BOND_MODE_ACTIVEBACKUP:
		break;
	case BOND_MODE_XOR:
4633
		bond_set_xmit_hash_policy(bond);
L
Linus Torvalds 已提交
4634 4635 4636 4637
		break;
	case BOND_MODE_BROADCAST:
		break;
	case BOND_MODE_8023AD:
4638
		bond_set_master_3ad_flags(bond);
4639
		bond_set_xmit_hash_policy(bond);
L
Linus Torvalds 已提交
4640 4641
		break;
	case BOND_MODE_ALB:
4642 4643 4644
		bond_set_master_alb_flags(bond);
		/* FALLTHRU */
	case BOND_MODE_TLB:
L
Linus Torvalds 已提交
4645 4646 4647
		break;
	default:
		/* Should never happen, mode already checked */
J
Joe Perches 已提交
4648 4649
		pr_err("%s: Error: Unknown bonding mode %d\n",
		       bond_dev->name, mode);
L
Linus Torvalds 已提交
4650 4651 4652 4653
		break;
	}
}

4654 4655 4656 4657 4658 4659 4660 4661
static void bond_ethtool_get_drvinfo(struct net_device *bond_dev,
				    struct ethtool_drvinfo *drvinfo)
{
	strncpy(drvinfo->driver, DRV_NAME, 32);
	strncpy(drvinfo->version, DRV_VERSION, 32);
	snprintf(drvinfo->fw_version, 32, "%d", BOND_ABI_VERSION);
}

4662
static const struct ethtool_ops bond_ethtool_ops = {
4663
	.get_drvinfo		= bond_ethtool_get_drvinfo,
4664 4665 4666 4667 4668 4669
	.get_link		= ethtool_op_get_link,
	.get_tx_csum		= ethtool_op_get_tx_csum,
	.get_sg			= ethtool_op_get_sg,
	.get_tso		= ethtool_op_get_tso,
	.get_ufo		= ethtool_op_get_ufo,
	.get_flags		= ethtool_op_get_flags,
4670 4671
};

4672
static const struct net_device_ops bond_netdev_ops = {
4673
	.ndo_init		= bond_init,
S
Stephen Hemminger 已提交
4674
	.ndo_uninit		= bond_uninit,
4675 4676
	.ndo_open		= bond_open,
	.ndo_stop		= bond_close,
4677
	.ndo_start_xmit		= bond_start_xmit,
4678
	.ndo_select_queue	= bond_select_queue,
4679
	.ndo_get_stats64	= bond_get_stats,
4680 4681 4682 4683
	.ndo_do_ioctl		= bond_do_ioctl,
	.ndo_set_multicast_list	= bond_set_multicast_list,
	.ndo_change_mtu		= bond_change_mtu,
	.ndo_set_mac_address 	= bond_set_mac_address,
4684
	.ndo_neigh_setup	= bond_neigh_setup,
4685 4686 4687
	.ndo_vlan_rx_register	= bond_vlan_rx_register,
	.ndo_vlan_rx_add_vid 	= bond_vlan_rx_add_vid,
	.ndo_vlan_rx_kill_vid	= bond_vlan_rx_kill_vid,
4688 4689 4690 4691
#ifdef CONFIG_NET_POLL_CONTROLLER
	.ndo_netpoll_cleanup	= bond_netpoll_cleanup,
	.ndo_poll_controller	= bond_poll_controller,
#endif
4692 4693
};

4694 4695 4696 4697 4698 4699 4700 4701
static void bond_destructor(struct net_device *bond_dev)
{
	struct bonding *bond = netdev_priv(bond_dev);
	if (bond->wq)
		destroy_workqueue(bond->wq);
	free_netdev(bond_dev);
}

4702
static void bond_setup(struct net_device *bond_dev)
L
Linus Torvalds 已提交
4703
{
4704
	struct bonding *bond = netdev_priv(bond_dev);
L
Linus Torvalds 已提交
4705 4706 4707 4708 4709

	/* initialize rwlocks */
	rwlock_init(&bond->lock);
	rwlock_init(&bond->curr_slave_lock);

4710
	bond->params = bonding_defaults;
L
Linus Torvalds 已提交
4711 4712 4713 4714 4715 4716

	/* Initialize pointers */
	bond->dev = bond_dev;
	INIT_LIST_HEAD(&bond->vlan_list);

	/* Initialize the device entry points */
4717
	ether_setup(bond_dev);
4718
	bond_dev->netdev_ops = &bond_netdev_ops;
4719
	bond_dev->ethtool_ops = &bond_ethtool_ops;
4720
	bond_set_mode_ops(bond, bond->params.mode);
L
Linus Torvalds 已提交
4721

4722
	bond_dev->destructor = bond_destructor;
L
Linus Torvalds 已提交
4723 4724 4725 4726

	/* Initialize the device options */
	bond_dev->tx_queue_len = 0;
	bond_dev->flags |= IFF_MASTER|IFF_MULTICAST;
4727
	bond_dev->priv_flags |= IFF_BONDING;
4728 4729
	bond_dev->priv_flags &= ~IFF_XMIT_DST_RELEASE;

4730 4731
	if (bond->params.arp_interval)
		bond_dev->priv_flags |= IFF_MASTER_ARPMON;
L
Linus Torvalds 已提交
4732 4733 4734 4735 4736 4737 4738 4739

	/* At first, we block adding VLANs. That's the only way to
	 * prevent problems that occur when adding VLANs over an
	 * empty bond. The block will be removed once non-challenged
	 * slaves are enslaved.
	 */
	bond_dev->features |= NETIF_F_VLAN_CHALLENGED;

H
Herbert Xu 已提交
4740
	/* don't acquire bond device's netif_tx_lock when
L
Linus Torvalds 已提交
4741 4742 4743 4744 4745 4746 4747 4748 4749 4750 4751 4752 4753
	 * transmitting */
	bond_dev->features |= NETIF_F_LLTX;

	/* By default, we declare the bond to be fully
	 * VLAN hardware accelerated capable. Special
	 * care is taken in the various xmit functions
	 * when there are slaves that are not hw accel
	 * capable
	 */
	bond_dev->features |= (NETIF_F_HW_VLAN_TX |
			       NETIF_F_HW_VLAN_RX |
			       NETIF_F_HW_VLAN_FILTER);

E
Eric Dumazet 已提交
4754 4755 4756 4757
	/* By default, we enable GRO on bonding devices.
	 * Actual support requires lowlevel drivers are GRO ready.
	 */
	bond_dev->features |= NETIF_F_GRO;
L
Linus Torvalds 已提交
4758 4759
}

J
Jay Vosburgh 已提交
4760 4761 4762 4763 4764 4765 4766 4767 4768 4769 4770 4771 4772 4773 4774 4775 4776 4777 4778
static void bond_work_cancel_all(struct bonding *bond)
{
	write_lock_bh(&bond->lock);
	bond->kill_timers = 1;
	write_unlock_bh(&bond->lock);

	if (bond->params.miimon && delayed_work_pending(&bond->mii_work))
		cancel_delayed_work(&bond->mii_work);

	if (bond->params.arp_interval && delayed_work_pending(&bond->arp_work))
		cancel_delayed_work(&bond->arp_work);

	if (bond->params.mode == BOND_MODE_ALB &&
	    delayed_work_pending(&bond->alb_work))
		cancel_delayed_work(&bond->alb_work);

	if (bond->params.mode == BOND_MODE_8023AD &&
	    delayed_work_pending(&bond->ad_work))
		cancel_delayed_work(&bond->ad_work);
4779 4780 4781

	if (delayed_work_pending(&bond->mcast_work))
		cancel_delayed_work(&bond->mcast_work);
J
Jay Vosburgh 已提交
4782 4783
}

4784 4785 4786 4787 4788
/*
* Destroy a bonding device.
* Must be under rtnl_lock when this function is called.
*/
static void bond_uninit(struct net_device *bond_dev)
J
Jay Vosburgh 已提交
4789
{
4790
	struct bonding *bond = netdev_priv(bond_dev);
4791
	struct vlan_entry *vlan, *tmp;
J
Jay Vosburgh 已提交
4792

4793 4794
	bond_netpoll_cleanup(bond_dev);

4795 4796 4797
	/* Release the bonded slaves */
	bond_release_all(bond_dev);

J
Jay Vosburgh 已提交
4798 4799 4800 4801 4802
	list_del(&bond->bond_list);

	bond_work_cancel_all(bond);

	bond_remove_proc_entry(bond);
4803

4804
	__hw_addr_flush(&bond->mc_list);
4805 4806 4807 4808 4809

	list_for_each_entry_safe(vlan, tmp, &bond->vlan_list, vlan_list) {
		list_del(&vlan->vlan_list);
		kfree(vlan);
	}
J
Jay Vosburgh 已提交
4810 4811
}

L
Linus Torvalds 已提交
4812 4813 4814 4815
/*------------------------- Module initialization ---------------------------*/

/*
 * Convert string input module parms.  Accept either the
J
Jay Vosburgh 已提交
4816 4817 4818
 * number of the mode or its string name.  A bit complicated because
 * some mode names are substrings of other names, and calls from sysfs
 * may have whitespace in the name (trailing newlines, for example).
L
Linus Torvalds 已提交
4819
 */
4820
int bond_parse_parm(const char *buf, const struct bond_parm_tbl *tbl)
L
Linus Torvalds 已提交
4821
{
4822
	int modeint = -1, i, rv;
J
Jay Vosburgh 已提交
4823
	char *p, modestr[BOND_MAX_MODENAME_LEN + 1] = { 0, };
J
Jay Vosburgh 已提交
4824

J
Jay Vosburgh 已提交
4825 4826 4827 4828 4829
	for (p = (char *)buf; *p; p++)
		if (!(isdigit(*p) || isspace(*p)))
			break;

	if (*p)
J
Jay Vosburgh 已提交
4830
		rv = sscanf(buf, "%20s", modestr);
J
Jay Vosburgh 已提交
4831
	else
4832
		rv = sscanf(buf, "%d", &modeint);
J
Jay Vosburgh 已提交
4833 4834 4835

	if (!rv)
		return -1;
L
Linus Torvalds 已提交
4836 4837

	for (i = 0; tbl[i].modename; i++) {
4838
		if (modeint == tbl[i].mode)
J
Jay Vosburgh 已提交
4839 4840
			return tbl[i].mode;
		if (strcmp(modestr, tbl[i].modename) == 0)
L
Linus Torvalds 已提交
4841 4842 4843 4844 4845 4846 4847 4848
			return tbl[i].mode;
	}

	return -1;
}

static int bond_check_params(struct bond_params *params)
{
4849
	int arp_validate_value, fail_over_mac_value, primary_reselect_value;
4850

L
Linus Torvalds 已提交
4851 4852 4853 4854 4855 4856
	/*
	 * Convert string parameters.
	 */
	if (mode) {
		bond_mode = bond_parse_parm(mode, bond_mode_tbl);
		if (bond_mode == -1) {
J
Joe Perches 已提交
4857
			pr_err("Error: Invalid bonding mode \"%s\"\n",
L
Linus Torvalds 已提交
4858 4859 4860 4861 4862
			       mode == NULL ? "NULL" : mode);
			return -EINVAL;
		}
	}

4863 4864 4865
	if (xmit_hash_policy) {
		if ((bond_mode != BOND_MODE_XOR) &&
		    (bond_mode != BOND_MODE_8023AD)) {
J
Joe Perches 已提交
4866
			pr_info("xmit_hash_policy param is irrelevant in mode %s\n",
4867 4868 4869 4870 4871
			       bond_mode_name(bond_mode));
		} else {
			xmit_hashtype = bond_parse_parm(xmit_hash_policy,
							xmit_hashtype_tbl);
			if (xmit_hashtype == -1) {
J
Joe Perches 已提交
4872
				pr_err("Error: Invalid xmit_hash_policy \"%s\"\n",
S
Stephen Hemminger 已提交
4873
				       xmit_hash_policy == NULL ? "NULL" :
4874 4875 4876 4877 4878 4879
				       xmit_hash_policy);
				return -EINVAL;
			}
		}
	}

L
Linus Torvalds 已提交
4880 4881
	if (lacp_rate) {
		if (bond_mode != BOND_MODE_8023AD) {
J
Joe Perches 已提交
4882 4883
			pr_info("lacp_rate param is irrelevant in mode %s\n",
				bond_mode_name(bond_mode));
L
Linus Torvalds 已提交
4884 4885 4886
		} else {
			lacp_fast = bond_parse_parm(lacp_rate, bond_lacp_tbl);
			if (lacp_fast == -1) {
J
Joe Perches 已提交
4887
				pr_err("Error: Invalid lacp rate \"%s\"\n",
L
Linus Torvalds 已提交
4888 4889 4890 4891 4892 4893
				       lacp_rate == NULL ? "NULL" : lacp_rate);
				return -EINVAL;
			}
		}
	}

4894 4895 4896
	if (ad_select) {
		params->ad_select = bond_parse_parm(ad_select, ad_select_tbl);
		if (params->ad_select == -1) {
J
Joe Perches 已提交
4897
			pr_err("Error: Invalid ad_select \"%s\"\n",
4898 4899 4900 4901 4902
			       ad_select == NULL ? "NULL" : ad_select);
			return -EINVAL;
		}

		if (bond_mode != BOND_MODE_8023AD) {
J
Joe Perches 已提交
4903
			pr_warning("ad_select param only affects 802.3ad mode\n");
4904 4905 4906 4907 4908
		}
	} else {
		params->ad_select = BOND_AD_STABLE;
	}

4909
	if (max_bonds < 0) {
J
Joe Perches 已提交
4910 4911
		pr_warning("Warning: max_bonds (%d) not in range %d-%d, so it was reset to BOND_DEFAULT_MAX_BONDS (%d)\n",
			   max_bonds, 0, INT_MAX, BOND_DEFAULT_MAX_BONDS);
L
Linus Torvalds 已提交
4912 4913 4914 4915
		max_bonds = BOND_DEFAULT_MAX_BONDS;
	}

	if (miimon < 0) {
J
Joe Perches 已提交
4916 4917
		pr_warning("Warning: miimon module parameter (%d), not in range 0-%d, so it was reset to %d\n",
			   miimon, INT_MAX, BOND_LINK_MON_INTERV);
L
Linus Torvalds 已提交
4918 4919 4920 4921
		miimon = BOND_LINK_MON_INTERV;
	}

	if (updelay < 0) {
J
Joe Perches 已提交
4922 4923
		pr_warning("Warning: updelay module parameter (%d), not in range 0-%d, so it was reset to 0\n",
			   updelay, INT_MAX);
L
Linus Torvalds 已提交
4924 4925 4926 4927
		updelay = 0;
	}

	if (downdelay < 0) {
J
Joe Perches 已提交
4928 4929
		pr_warning("Warning: downdelay module parameter (%d), not in range 0-%d, so it was reset to 0\n",
			   downdelay, INT_MAX);
L
Linus Torvalds 已提交
4930 4931 4932 4933
		downdelay = 0;
	}

	if ((use_carrier != 0) && (use_carrier != 1)) {
J
Joe Perches 已提交
4934 4935
		pr_warning("Warning: use_carrier module parameter (%d), not of valid value (0/1), so it was set to 1\n",
			   use_carrier);
L
Linus Torvalds 已提交
4936 4937 4938
		use_carrier = 1;
	}

4939
	if (num_grat_arp < 0 || num_grat_arp > 255) {
4940
		pr_warning("Warning: num_grat_arp (%d) not in range 0-255 so it was reset to 1\n",
J
Joe Perches 已提交
4941
			   num_grat_arp);
4942 4943 4944
		num_grat_arp = 1;
	}

4945
	if (num_unsol_na < 0 || num_unsol_na > 255) {
4946
		pr_warning("Warning: num_unsol_na (%d) not in range 0-255 so it was reset to 1\n",
J
Joe Perches 已提交
4947
			   num_unsol_na);
4948 4949 4950
		num_unsol_na = 1;
	}

L
Linus Torvalds 已提交
4951 4952 4953
	/* reset values for 802.3ad */
	if (bond_mode == BOND_MODE_8023AD) {
		if (!miimon) {
J
Joe Perches 已提交
4954
			pr_warning("Warning: miimon must be specified, otherwise bonding will not detect link failure, speed and duplex which are essential for 802.3ad operation\n");
S
Stephen Hemminger 已提交
4955
			pr_warning("Forcing miimon to 100msec\n");
L
Linus Torvalds 已提交
4956 4957 4958 4959
			miimon = 100;
		}
	}

4960 4961 4962 4963 4964 4965 4966
	if (tx_queues < 1 || tx_queues > 255) {
		pr_warning("Warning: tx_queues (%d) should be between "
			   "1 and 255, resetting to %d\n",
			   tx_queues, BOND_DEFAULT_TX_QUEUES);
		tx_queues = BOND_DEFAULT_TX_QUEUES;
	}

4967 4968 4969 4970 4971 4972 4973
	if ((all_slaves_active != 0) && (all_slaves_active != 1)) {
		pr_warning("Warning: all_slaves_active module parameter (%d), "
			   "not of valid value (0/1), so it was set to "
			   "0\n", all_slaves_active);
		all_slaves_active = 0;
	}

4974 4975 4976 4977 4978 4979 4980
	if (resend_igmp < 0 || resend_igmp > 255) {
		pr_warning("Warning: resend_igmp (%d) should be between "
			   "0 and 255, resetting to %d\n",
			   resend_igmp, BOND_DEFAULT_RESEND_IGMP);
		resend_igmp = BOND_DEFAULT_RESEND_IGMP;
	}

L
Linus Torvalds 已提交
4981 4982 4983 4984
	/* reset values for TLB/ALB */
	if ((bond_mode == BOND_MODE_TLB) ||
	    (bond_mode == BOND_MODE_ALB)) {
		if (!miimon) {
J
Joe Perches 已提交
4985
			pr_warning("Warning: miimon must be specified, otherwise bonding will not detect link failure and link speed which are essential for TLB/ALB load balancing\n");
S
Stephen Hemminger 已提交
4986
			pr_warning("Forcing miimon to 100msec\n");
L
Linus Torvalds 已提交
4987 4988 4989 4990 4991
			miimon = 100;
		}
	}

	if (bond_mode == BOND_MODE_ALB) {
J
Joe Perches 已提交
4992 4993
		pr_notice("In ALB mode you might experience client disconnections upon reconnection of a link if the bonding module updelay parameter (%d msec) is incompatible with the forwarding delay time of the switch\n",
			  updelay);
L
Linus Torvalds 已提交
4994 4995 4996 4997 4998 4999 5000
	}

	if (!miimon) {
		if (updelay || downdelay) {
			/* just warn the user the up/down delay will have
			 * no effect since miimon is zero...
			 */
J
Joe Perches 已提交
5001 5002
			pr_warning("Warning: miimon module parameter not set and updelay (%d) or downdelay (%d) module parameter is set; updelay and downdelay have no effect unless miimon is set\n",
				   updelay, downdelay);
L
Linus Torvalds 已提交
5003 5004 5005 5006
		}
	} else {
		/* don't allow arp monitoring */
		if (arp_interval) {
J
Joe Perches 已提交
5007 5008
			pr_warning("Warning: miimon (%d) and arp_interval (%d) can't be used simultaneously, disabling ARP monitoring\n",
				   miimon, arp_interval);
L
Linus Torvalds 已提交
5009 5010 5011 5012
			arp_interval = 0;
		}

		if ((updelay % miimon) != 0) {
J
Joe Perches 已提交
5013 5014 5015
			pr_warning("Warning: updelay (%d) is not a multiple of miimon (%d), updelay rounded to %d ms\n",
				   updelay, miimon,
				   (updelay / miimon) * miimon);
L
Linus Torvalds 已提交
5016 5017 5018 5019 5020
		}

		updelay /= miimon;

		if ((downdelay % miimon) != 0) {
J
Joe Perches 已提交
5021 5022 5023
			pr_warning("Warning: downdelay (%d) is not a multiple of miimon (%d), downdelay rounded to %d ms\n",
				   downdelay, miimon,
				   (downdelay / miimon) * miimon);
L
Linus Torvalds 已提交
5024 5025 5026 5027 5028 5029
		}

		downdelay /= miimon;
	}

	if (arp_interval < 0) {
J
Joe Perches 已提交
5030 5031
		pr_warning("Warning: arp_interval module parameter (%d) , not in range 0-%d, so it was reset to %d\n",
			   arp_interval, INT_MAX, BOND_LINK_ARP_INTERV);
L
Linus Torvalds 已提交
5032 5033 5034 5035 5036 5037 5038 5039 5040
		arp_interval = BOND_LINK_ARP_INTERV;
	}

	for (arp_ip_count = 0;
	     (arp_ip_count < BOND_MAX_ARP_TARGETS) && arp_ip_target[arp_ip_count];
	     arp_ip_count++) {
		/* not complete check, but should be good enough to
		   catch mistakes */
		if (!isdigit(arp_ip_target[arp_ip_count][0])) {
J
Joe Perches 已提交
5041 5042
			pr_warning("Warning: bad arp_ip_target module parameter (%s), ARP monitoring will not be performed\n",
				   arp_ip_target[arp_ip_count]);
L
Linus Torvalds 已提交
5043 5044
			arp_interval = 0;
		} else {
5045
			__be32 ip = in_aton(arp_ip_target[arp_ip_count]);
L
Linus Torvalds 已提交
5046 5047 5048 5049 5050 5051
			arp_target[arp_ip_count] = ip;
		}
	}

	if (arp_interval && !arp_ip_count) {
		/* don't allow arping if no arp_ip_target given... */
J
Joe Perches 已提交
5052 5053
		pr_warning("Warning: arp_interval module parameter (%d) specified without providing an arp_ip_target parameter, arp_interval was reset to 0\n",
			   arp_interval);
L
Linus Torvalds 已提交
5054 5055 5056
		arp_interval = 0;
	}

5057 5058
	if (arp_validate) {
		if (bond_mode != BOND_MODE_ACTIVEBACKUP) {
J
Joe Perches 已提交
5059
			pr_err("arp_validate only supported in active-backup mode\n");
5060 5061 5062
			return -EINVAL;
		}
		if (!arp_interval) {
J
Joe Perches 已提交
5063
			pr_err("arp_validate requires arp_interval\n");
5064 5065 5066 5067 5068 5069
			return -EINVAL;
		}

		arp_validate_value = bond_parse_parm(arp_validate,
						     arp_validate_tbl);
		if (arp_validate_value == -1) {
J
Joe Perches 已提交
5070
			pr_err("Error: invalid arp_validate \"%s\"\n",
5071 5072 5073 5074 5075 5076
			       arp_validate == NULL ? "NULL" : arp_validate);
			return -EINVAL;
		}
	} else
		arp_validate_value = 0;

L
Linus Torvalds 已提交
5077
	if (miimon) {
J
Joe Perches 已提交
5078
		pr_info("MII link monitoring set to %d ms\n", miimon);
L
Linus Torvalds 已提交
5079 5080 5081
	} else if (arp_interval) {
		int i;

J
Joe Perches 已提交
5082 5083 5084 5085
		pr_info("ARP monitoring set to %d ms, validate %s, with %d target(s):",
			arp_interval,
			arp_validate_tbl[arp_validate_value].modename,
			arp_ip_count);
L
Linus Torvalds 已提交
5086 5087

		for (i = 0; i < arp_ip_count; i++)
J
Jiri Pirko 已提交
5088
			pr_info(" %s", arp_ip_target[i]);
L
Linus Torvalds 已提交
5089

J
Jiri Pirko 已提交
5090
		pr_info("\n");
L
Linus Torvalds 已提交
5091

5092
	} else if (max_bonds) {
L
Linus Torvalds 已提交
5093 5094 5095
		/* miimon and arp_interval not set, we need one so things
		 * work as expected, see bonding.txt for details
		 */
J
Joe Perches 已提交
5096
		pr_warning("Warning: either miimon or arp_interval and arp_ip_target module parameters must be specified, otherwise bonding will not detect link failures! see bonding.txt for details.\n");
L
Linus Torvalds 已提交
5097 5098 5099 5100 5101 5102
	}

	if (primary && !USES_PRIMARY(bond_mode)) {
		/* currently, using a primary only makes sense
		 * in active backup, TLB or ALB modes
		 */
J
Joe Perches 已提交
5103 5104
		pr_warning("Warning: %s primary device specified but has no effect in %s mode\n",
			   primary, bond_mode_name(bond_mode));
L
Linus Torvalds 已提交
5105 5106 5107
		primary = NULL;
	}

5108 5109 5110 5111
	if (primary && primary_reselect) {
		primary_reselect_value = bond_parse_parm(primary_reselect,
							 pri_reselect_tbl);
		if (primary_reselect_value == -1) {
J
Joe Perches 已提交
5112
			pr_err("Error: Invalid primary_reselect \"%s\"\n",
5113 5114 5115 5116 5117 5118 5119 5120
			       primary_reselect ==
					NULL ? "NULL" : primary_reselect);
			return -EINVAL;
		}
	} else {
		primary_reselect_value = BOND_PRI_RESELECT_ALWAYS;
	}

5121 5122 5123 5124
	if (fail_over_mac) {
		fail_over_mac_value = bond_parse_parm(fail_over_mac,
						      fail_over_mac_tbl);
		if (fail_over_mac_value == -1) {
J
Joe Perches 已提交
5125
			pr_err("Error: invalid fail_over_mac \"%s\"\n",
5126 5127 5128 5129 5130
			       arp_validate == NULL ? "NULL" : arp_validate);
			return -EINVAL;
		}

		if (bond_mode != BOND_MODE_ACTIVEBACKUP)
J
Joe Perches 已提交
5131
			pr_warning("Warning: fail_over_mac only affects active-backup mode.\n");
5132 5133 5134
	} else {
		fail_over_mac_value = BOND_FOM_NONE;
	}
5135

L
Linus Torvalds 已提交
5136 5137
	/* fill params struct with the proper values */
	params->mode = bond_mode;
5138
	params->xmit_policy = xmit_hashtype;
L
Linus Torvalds 已提交
5139
	params->miimon = miimon;
5140
	params->num_grat_arp = num_grat_arp;
5141
	params->num_unsol_na = num_unsol_na;
L
Linus Torvalds 已提交
5142
	params->arp_interval = arp_interval;
5143
	params->arp_validate = arp_validate_value;
L
Linus Torvalds 已提交
5144 5145 5146 5147 5148
	params->updelay = updelay;
	params->downdelay = downdelay;
	params->use_carrier = use_carrier;
	params->lacp_fast = lacp_fast;
	params->primary[0] = 0;
5149
	params->primary_reselect = primary_reselect_value;
5150
	params->fail_over_mac = fail_over_mac_value;
5151
	params->tx_queues = tx_queues;
5152
	params->all_slaves_active = all_slaves_active;
5153
	params->resend_igmp = resend_igmp;
L
Linus Torvalds 已提交
5154 5155 5156 5157 5158 5159 5160 5161 5162 5163 5164

	if (primary) {
		strncpy(params->primary, primary, IFNAMSIZ);
		params->primary[IFNAMSIZ - 1] = 0;
	}

	memcpy(params->arp_targets, arp_target, sizeof(arp_target));

	return 0;
}

5165
static struct lock_class_key bonding_netdev_xmit_lock_key;
5166
static struct lock_class_key bonding_netdev_addr_lock_key;
5167

5168 5169 5170
static void bond_set_lockdep_class_one(struct net_device *dev,
				       struct netdev_queue *txq,
				       void *_unused)
5171 5172 5173 5174 5175 5176 5177
{
	lockdep_set_class(&txq->_xmit_lock,
			  &bonding_netdev_xmit_lock_key);
}

static void bond_set_lockdep_class(struct net_device *dev)
{
5178 5179
	lockdep_set_class(&dev->addr_list_lock,
			  &bonding_netdev_addr_lock_key);
5180
	netdev_for_each_tx_queue(dev, bond_set_lockdep_class_one, NULL);
5181 5182
}

5183 5184 5185 5186 5187 5188
/*
 * Called from registration process
 */
static int bond_init(struct net_device *bond_dev)
{
	struct bonding *bond = netdev_priv(bond_dev);
5189
	struct bond_net *bn = net_generic(dev_net(bond_dev), bond_net_id);
5190 5191 5192 5193 5194 5195 5196 5197 5198 5199 5200 5201

	pr_debug("Begin bond_init for %s\n", bond_dev->name);

	bond->wq = create_singlethread_workqueue(bond_dev->name);
	if (!bond->wq)
		return -ENOMEM;

	bond_set_lockdep_class(bond_dev);

	netif_carrier_off(bond_dev);

	bond_create_proc_entry(bond);
5202
	list_add_tail(&bond->bond_list, &bn->dev_list);
5203

5204
	bond_prepare_sysfs_group(bond);
5205 5206

	__hw_addr_init(&bond->mc_list);
5207 5208 5209
	return 0;
}

5210 5211 5212 5213 5214 5215 5216 5217 5218 5219 5220 5221 5222
static int bond_validate(struct nlattr *tb[], struct nlattr *data[])
{
	if (tb[IFLA_ADDRESS]) {
		if (nla_len(tb[IFLA_ADDRESS]) != ETH_ALEN)
			return -EINVAL;
		if (!is_valid_ether_addr(nla_data(tb[IFLA_ADDRESS])))
			return -EADDRNOTAVAIL;
	}
	return 0;
}

static struct rtnl_link_ops bond_link_ops __read_mostly = {
	.kind		= "bond",
5223
	.priv_size	= sizeof(struct bonding),
5224 5225 5226 5227
	.setup		= bond_setup,
	.validate	= bond_validate,
};

5228
/* Create a new bond based on the specified name and bonding parameters.
5229
 * If name is NULL, obtain a suitable "bond%d" name for us.
5230 5231 5232
 * Caller must NOT hold rtnl_lock; we need to release it here before we
 * set up our sysfs entries.
 */
5233
int bond_create(struct net *net, const char *name)
5234 5235 5236 5237 5238
{
	struct net_device *bond_dev;
	int res;

	rtnl_lock();
5239

5240 5241
	bond_dev = alloc_netdev_mq(sizeof(struct bonding), name ? name : "",
				bond_setup, tx_queues);
5242
	if (!bond_dev) {
J
Joe Perches 已提交
5243
		pr_err("%s: eek! can't alloc netdev!\n", name);
5244 5245
		rtnl_unlock();
		return -ENOMEM;
5246 5247
	}

5248
	dev_net_set(bond_dev, net);
5249 5250
	bond_dev->rtnl_link_ops = &bond_link_ops;

5251 5252 5253
	if (!name) {
		res = dev_alloc_name(bond_dev, "bond%d");
		if (res < 0)
5254
			goto out;
5255 5256 5257 5258 5259 5260 5261 5262 5263
	} else {
		/*
		 * If we're given a name to register
		 * we need to ensure that its not already
		 * registered
		 */
		res = -EEXIST;
		if (__dev_get_by_name(net, name) != NULL)
			goto out;
5264 5265
	}

5266
	res = register_netdevice(bond_dev);
5267

E
Eric W. Biederman 已提交
5268
out:
5269
	rtnl_unlock();
5270 5271
	if (res < 0)
		bond_destructor(bond_dev);
E
Eric W. Biederman 已提交
5272
	return res;
5273 5274
}

5275
static int __net_init bond_net_init(struct net *net)
5276
{
5277
	struct bond_net *bn = net_generic(net, bond_net_id);
5278 5279 5280 5281 5282

	bn->net = net;
	INIT_LIST_HEAD(&bn->dev_list);

	bond_create_proc_dir(bn);
5283 5284
	
	return 0;
5285 5286
}

5287
static void __net_exit bond_net_exit(struct net *net)
5288
{
5289
	struct bond_net *bn = net_generic(net, bond_net_id);
5290 5291 5292 5293 5294 5295 5296

	bond_destroy_proc_dir(bn);
}

static struct pernet_operations bond_net_ops = {
	.init = bond_net_init,
	.exit = bond_net_exit,
5297 5298
	.id   = &bond_net_id,
	.size = sizeof(struct bond_net),
5299 5300
};

L
Linus Torvalds 已提交
5301 5302 5303 5304 5305
static int __init bonding_init(void)
{
	int i;
	int res;

S
Stephen Hemminger 已提交
5306
	pr_info("%s", version);
L
Linus Torvalds 已提交
5307

5308
	res = bond_check_params(&bonding_defaults);
S
Stephen Hemminger 已提交
5309
	if (res)
5310
		goto out;
L
Linus Torvalds 已提交
5311

5312 5313 5314 5315 5316 5317 5318
#ifdef CONFIG_NET_POLL_CONTROLLER
	if (!alloc_cpumask_var(&netpoll_block_tx, GFP_KERNEL)) {
		res = -ENOMEM;
		goto out;
	}
#endif

5319
	res = register_pernet_subsys(&bond_net_ops);
5320 5321
	if (res)
		goto out;
5322

5323 5324
	res = rtnl_link_register(&bond_link_ops);
	if (res)
5325
		goto err_link;
5326

L
Linus Torvalds 已提交
5327
	for (i = 0; i < max_bonds; i++) {
5328
		res = bond_create(&init_net, NULL);
5329 5330
		if (res)
			goto err;
L
Linus Torvalds 已提交
5331 5332
	}

5333 5334 5335 5336
	res = bond_create_sysfs();
	if (res)
		goto err;

5337

L
Linus Torvalds 已提交
5338
	register_netdevice_notifier(&bond_netdev_notifier);
J
Jay Vosburgh 已提交
5339
	register_inetaddr_notifier(&bond_inetaddr_notifier);
5340
	bond_register_ipv6_notifier();
5341
out:
L
Linus Torvalds 已提交
5342
	return res;
5343 5344
err:
	rtnl_link_unregister(&bond_link_ops);
5345
err_link:
5346
	unregister_pernet_subsys(&bond_net_ops);
5347 5348 5349
#ifdef CONFIG_NET_POLL_CONTROLLER
	free_cpumask_var(netpoll_block_tx);
#endif
5350
	goto out;
5351

L
Linus Torvalds 已提交
5352 5353 5354 5355 5356
}

static void __exit bonding_exit(void)
{
	unregister_netdevice_notifier(&bond_netdev_notifier);
J
Jay Vosburgh 已提交
5357
	unregister_inetaddr_notifier(&bond_inetaddr_notifier);
5358
	bond_unregister_ipv6_notifier();
L
Linus Torvalds 已提交
5359

5360 5361
	bond_destroy_sysfs();

5362
	rtnl_link_unregister(&bond_link_ops);
5363
	unregister_pernet_subsys(&bond_net_ops);
5364 5365 5366 5367

#ifdef CONFIG_NET_POLL_CONTROLLER
	free_cpumask_var(netpoll_block_tx);
#endif
L
Linus Torvalds 已提交
5368 5369 5370 5371 5372 5373 5374 5375
}

module_init(bonding_init);
module_exit(bonding_exit);
MODULE_LICENSE("GPL");
MODULE_VERSION(DRV_VERSION);
MODULE_DESCRIPTION(DRV_DESCRIPTION ", v" DRV_VERSION);
MODULE_AUTHOR("Thomas Davis, tadavis@lbl.gov and many others");
5376
MODULE_ALIAS_RTNL_LINK("bond");