vxlan.c 82.4 KB
Newer Older
S
stephen hemminger 已提交
1
/*
R
Rami Rosen 已提交
2
 * VXLAN: Virtual eXtensible Local Area Network
S
stephen hemminger 已提交
3
 *
4
 * Copyright (c) 2012-2013 Vyatta Inc.
S
stephen hemminger 已提交
5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26
 *
 * This program is free software; you can redistribute it and/or modify
 * it under the terms of the GNU General Public License version 2 as
 * published by the Free Software Foundation.
 */

#define pr_fmt(fmt) KBUILD_MODNAME ": " fmt

#include <linux/kernel.h>
#include <linux/types.h>
#include <linux/module.h>
#include <linux/errno.h>
#include <linux/slab.h>
#include <linux/skbuff.h>
#include <linux/rculist.h>
#include <linux/netdevice.h>
#include <linux/in.h>
#include <linux/ip.h>
#include <linux/udp.h>
#include <linux/igmp.h>
#include <linux/etherdevice.h>
#include <linux/if_ether.h>
27
#include <linux/if_vlan.h>
S
stephen hemminger 已提交
28
#include <linux/hash.h>
Y
Yan Burman 已提交
29
#include <linux/ethtool.h>
D
David Stevens 已提交
30 31
#include <net/arp.h>
#include <net/ndisc.h>
S
stephen hemminger 已提交
32
#include <net/ip.h>
33
#include <net/ip_tunnels.h>
S
stephen hemminger 已提交
34 35
#include <net/icmp.h>
#include <net/udp.h>
T
Tom Herbert 已提交
36
#include <net/udp_tunnel.h>
S
stephen hemminger 已提交
37 38 39 40 41 42
#include <net/rtnetlink.h>
#include <net/route.h>
#include <net/dsfield.h>
#include <net/inet_ecn.h>
#include <net/net_namespace.h>
#include <net/netns/generic.h>
43
#include <net/vxlan.h>
44
#include <net/protocol.h>
45
#include <net/udp_tunnel.h>
C
Cong Wang 已提交
46 47 48 49
#if IS_ENABLED(CONFIG_IPV6)
#include <net/ipv6.h>
#include <net/addrconf.h>
#include <net/ip6_tunnel.h>
50
#include <net/ip6_checksum.h>
C
Cong Wang 已提交
51
#endif
T
Thomas Graf 已提交
52
#include <net/dst_metadata.h>
S
stephen hemminger 已提交
53 54 55

#define VXLAN_VERSION	"0.1"

56 57
#define PORT_HASH_BITS	8
#define PORT_HASH_SIZE  (1<<PORT_HASH_BITS)
S
stephen hemminger 已提交
58 59 60
#define FDB_AGE_DEFAULT 300 /* 5 min */
#define FDB_AGE_INTERVAL (10 * HZ)	/* rescan interval */

61 62
/* UDP port for VXLAN traffic.
 * The IANA assigned port is 4789, but the Linux default is 8472
S
Stephen Hemminger 已提交
63
 * for compatibility with early adopters.
64
 */
65 66
static unsigned short vxlan_port __read_mostly = 8472;
module_param_named(udp_port, vxlan_port, ushort, 0444);
S
stephen hemminger 已提交
67 68 69 70 71 72
MODULE_PARM_DESC(udp_port, "Destination UDP port");

static bool log_ecn_error = true;
module_param(log_ecn_error, bool, 0644);
MODULE_PARM_DESC(log_ecn_error, "Log packets received with corrupted ECN");

P
Pravin B Shelar 已提交
73
static int vxlan_net_id;
74
static struct rtnl_link_ops vxlan_link_ops;
75

76
static const u8 all_zeros_mac[ETH_ALEN + 2];
77

78
static int vxlan_sock_add(struct vxlan_dev *vxlan);
79

80 81 82 83
/* per-network namespace private data for this module */
struct vxlan_net {
	struct list_head  vxlan_list;
	struct hlist_head sock_list[PORT_HASH_SIZE];
84
	spinlock_t	  sock_lock;
85 86
};

S
stephen hemminger 已提交
87 88 89 90 91 92
/* Forwarding table entry */
struct vxlan_fdb {
	struct hlist_node hlist;	/* linked list of entries */
	struct rcu_head	  rcu;
	unsigned long	  updated;	/* jiffies */
	unsigned long	  used;
93
	struct list_head  remotes;
94
	u8		  eth_addr[ETH_ALEN];
S
stephen hemminger 已提交
95
	u16		  state;	/* see ndm_state */
96
	u8		  flags;	/* see ndm_flags */
S
stephen hemminger 已提交
97 98 99 100
};

/* salt for hash table */
static u32 vxlan_salt __read_mostly;
101
static struct workqueue_struct *vxlan_wq;
S
stephen hemminger 已提交
102

T
Thomas Graf 已提交
103 104
static inline bool vxlan_collect_metadata(struct vxlan_sock *vs)
{
105 106
	return vs->flags & VXLAN_F_COLLECT_METADATA ||
	       ip_tunnel_collect_metadata();
T
Thomas Graf 已提交
107 108
}

C
Cong Wang 已提交
109 110 111 112
#if IS_ENABLED(CONFIG_IPV6)
static inline
bool vxlan_addr_equal(const union vxlan_addr *a, const union vxlan_addr *b)
{
J
Jiri Benc 已提交
113 114 115 116 117 118
	if (a->sa.sa_family != b->sa.sa_family)
		return false;
	if (a->sa.sa_family == AF_INET6)
		return ipv6_addr_equal(&a->sin6.sin6_addr, &b->sin6.sin6_addr);
	else
		return a->sin.sin_addr.s_addr == b->sin.sin_addr.s_addr;
C
Cong Wang 已提交
119 120 121 122
}

static inline bool vxlan_addr_any(const union vxlan_addr *ipa)
{
J
Jiri Benc 已提交
123 124 125 126
	if (ipa->sa.sa_family == AF_INET6)
		return ipv6_addr_any(&ipa->sin6.sin6_addr);
	else
		return ipa->sin.sin_addr.s_addr == htonl(INADDR_ANY);
C
Cong Wang 已提交
127 128 129 130
}

static inline bool vxlan_addr_multicast(const union vxlan_addr *ipa)
{
J
Jiri Benc 已提交
131 132 133 134
	if (ipa->sa.sa_family == AF_INET6)
		return ipv6_addr_is_multicast(&ipa->sin6.sin6_addr);
	else
		return IN_MULTICAST(ntohl(ipa->sin.sin_addr.s_addr));
C
Cong Wang 已提交
135 136 137 138
}

static int vxlan_nla_get_addr(union vxlan_addr *ip, struct nlattr *nla)
{
J
Jiri Benc 已提交
139
	if (nla_len(nla) >= sizeof(struct in6_addr)) {
140
		ip->sin6.sin6_addr = nla_get_in6_addr(nla);
J
Jiri Benc 已提交
141 142 143
		ip->sa.sa_family = AF_INET6;
		return 0;
	} else if (nla_len(nla) >= sizeof(__be32)) {
144
		ip->sin.sin_addr.s_addr = nla_get_in_addr(nla);
J
Jiri Benc 已提交
145 146 147 148 149
		ip->sa.sa_family = AF_INET;
		return 0;
	} else {
		return -EAFNOSUPPORT;
	}
C
Cong Wang 已提交
150 151 152
}

static int vxlan_nla_put_addr(struct sk_buff *skb, int attr,
J
Jiri Benc 已提交
153
			      const union vxlan_addr *ip)
C
Cong Wang 已提交
154
{
J
Jiri Benc 已提交
155
	if (ip->sa.sa_family == AF_INET6)
156
		return nla_put_in6_addr(skb, attr, &ip->sin6.sin6_addr);
J
Jiri Benc 已提交
157
	else
158
		return nla_put_in_addr(skb, attr, ip->sin.sin_addr.s_addr);
C
Cong Wang 已提交
159 160 161 162 163 164 165
}

#else /* !CONFIG_IPV6 */

static inline
bool vxlan_addr_equal(const union vxlan_addr *a, const union vxlan_addr *b)
{
J
Jiri Benc 已提交
166
	return a->sin.sin_addr.s_addr == b->sin.sin_addr.s_addr;
C
Cong Wang 已提交
167 168 169 170
}

static inline bool vxlan_addr_any(const union vxlan_addr *ipa)
{
J
Jiri Benc 已提交
171
	return ipa->sin.sin_addr.s_addr == htonl(INADDR_ANY);
C
Cong Wang 已提交
172 173 174 175
}

static inline bool vxlan_addr_multicast(const union vxlan_addr *ipa)
{
J
Jiri Benc 已提交
176
	return IN_MULTICAST(ntohl(ipa->sin.sin_addr.s_addr));
C
Cong Wang 已提交
177 178 179 180
}

static int vxlan_nla_get_addr(union vxlan_addr *ip, struct nlattr *nla)
{
J
Jiri Benc 已提交
181 182 183
	if (nla_len(nla) >= sizeof(struct in6_addr)) {
		return -EAFNOSUPPORT;
	} else if (nla_len(nla) >= sizeof(__be32)) {
184
		ip->sin.sin_addr.s_addr = nla_get_in_addr(nla);
J
Jiri Benc 已提交
185 186 187 188 189
		ip->sa.sa_family = AF_INET;
		return 0;
	} else {
		return -EAFNOSUPPORT;
	}
C
Cong Wang 已提交
190 191 192
}

static int vxlan_nla_put_addr(struct sk_buff *skb, int attr,
J
Jiri Benc 已提交
193
			      const union vxlan_addr *ip)
C
Cong Wang 已提交
194
{
195
	return nla_put_in_addr(skb, attr, ip->sin.sin_addr.s_addr);
C
Cong Wang 已提交
196 197 198
}
#endif

199 200 201 202 203 204 205 206
/* Virtual Network hash table head */
static inline struct hlist_head *vni_head(struct vxlan_sock *vs, u32 id)
{
	return &vs->vni_list[hash_32(id, VNI_HASH_BITS)];
}

/* Socket hash table head */
static inline struct hlist_head *vs_head(struct net *net, __be16 port)
S
stephen hemminger 已提交
207 208 209
{
	struct vxlan_net *vn = net_generic(net, vxlan_net_id);

210 211 212
	return &vn->sock_list[hash_32(ntohs(port), PORT_HASH_BITS)];
}

213 214 215
/* First remote destination for a forwarding entry.
 * Guaranteed to be non-NULL because remotes are never deleted.
 */
216
static inline struct vxlan_rdst *first_remote_rcu(struct vxlan_fdb *fdb)
217
{
218 219 220 221 222 223
	return list_entry_rcu(fdb->remotes.next, struct vxlan_rdst, list);
}

static inline struct vxlan_rdst *first_remote_rtnl(struct vxlan_fdb *fdb)
{
	return list_first_entry(&fdb->remotes, struct vxlan_rdst, list);
224 225
}

226 227 228 229 230
/* Find VXLAN socket based on network namespace, address family and UDP port
 * and enabled unshareable flags.
 */
static struct vxlan_sock *vxlan_find_sock(struct net *net, sa_family_t family,
					  __be16 port, u32 flags)
231 232
{
	struct vxlan_sock *vs;
233 234

	flags &= VXLAN_F_RCV_FLAGS;
235 236

	hlist_for_each_entry_rcu(vs, vs_head(net, port), hlist) {
237
		if (inet_sk(vs->sock->sk)->inet_sport == port &&
238
		    vxlan_get_sk_family(vs) == family &&
239
		    vs->flags == flags)
240 241 242
			return vs;
	}
	return NULL;
S
stephen hemminger 已提交
243 244
}

P
Pravin B Shelar 已提交
245
static struct vxlan_dev *vxlan_vs_find_vni(struct vxlan_sock *vs, u32 id)
S
stephen hemminger 已提交
246 247 248
{
	struct vxlan_dev *vxlan;

249
	hlist_for_each_entry_rcu(vxlan, vni_head(vs, id), hlist) {
250
		if (vxlan->default_dst.remote_vni == id)
S
stephen hemminger 已提交
251 252 253 254 255 256
			return vxlan;
	}

	return NULL;
}

P
Pravin B Shelar 已提交
257
/* Look up VNI in a per net namespace table */
258
static struct vxlan_dev *vxlan_find_vni(struct net *net, u32 id,
259 260
					sa_family_t family, __be16 port,
					u32 flags)
P
Pravin B Shelar 已提交
261 262 263
{
	struct vxlan_sock *vs;

264
	vs = vxlan_find_sock(net, family, port, flags);
P
Pravin B Shelar 已提交
265 266 267 268 269 270
	if (!vs)
		return NULL;

	return vxlan_vs_find_vni(vs, id);
}

S
stephen hemminger 已提交
271 272
/* Fill in neighbour message in skbuff. */
static int vxlan_fdb_info(struct sk_buff *skb, struct vxlan_dev *vxlan,
S
Stephen Hemminger 已提交
273 274 275
			  const struct vxlan_fdb *fdb,
			  u32 portid, u32 seq, int type, unsigned int flags,
			  const struct vxlan_rdst *rdst)
S
stephen hemminger 已提交
276 277 278 279 280
{
	unsigned long now = jiffies;
	struct nda_cacheinfo ci;
	struct nlmsghdr *nlh;
	struct ndmsg *ndm;
D
David Stevens 已提交
281
	bool send_ip, send_eth;
S
stephen hemminger 已提交
282 283 284 285 286 287 288

	nlh = nlmsg_put(skb, portid, seq, type, sizeof(*ndm), flags);
	if (nlh == NULL)
		return -EMSGSIZE;

	ndm = nlmsg_data(nlh);
	memset(ndm, 0, sizeof(*ndm));
D
David Stevens 已提交
289 290 291 292 293

	send_eth = send_ip = true;

	if (type == RTM_GETNEIGH) {
		ndm->ndm_family	= AF_INET;
C
Cong Wang 已提交
294
		send_ip = !vxlan_addr_any(&rdst->remote_ip);
D
David Stevens 已提交
295 296 297
		send_eth = !is_zero_ether_addr(fdb->eth_addr);
	} else
		ndm->ndm_family	= AF_BRIDGE;
S
stephen hemminger 已提交
298 299
	ndm->ndm_state = fdb->state;
	ndm->ndm_ifindex = vxlan->dev->ifindex;
300
	ndm->ndm_flags = fdb->flags;
301
	ndm->ndm_type = RTN_UNICAST;
S
stephen hemminger 已提交
302

303
	if (!net_eq(dev_net(vxlan->dev), vxlan->net) &&
304
	    nla_put_s32(skb, NDA_LINK_NETNSID,
305
			peernet2id_alloc(dev_net(vxlan->dev), vxlan->net)))
306 307
		goto nla_put_failure;

D
David Stevens 已提交
308
	if (send_eth && nla_put(skb, NDA_LLADDR, ETH_ALEN, &fdb->eth_addr))
S
stephen hemminger 已提交
309 310
		goto nla_put_failure;

C
Cong Wang 已提交
311
	if (send_ip && vxlan_nla_put_addr(skb, NDA_DST, &rdst->remote_ip))
312 313
		goto nla_put_failure;

314
	if (rdst->remote_port && rdst->remote_port != vxlan->cfg.dst_port &&
315 316
	    nla_put_be16(skb, NDA_PORT, rdst->remote_port))
		goto nla_put_failure;
317
	if (rdst->remote_vni != vxlan->default_dst.remote_vni &&
P
Pravin B Shelar 已提交
318
	    nla_put_u32(skb, NDA_VNI, rdst->remote_vni))
319 320 321
		goto nla_put_failure;
	if (rdst->remote_ifindex &&
	    nla_put_u32(skb, NDA_IFINDEX, rdst->remote_ifindex))
S
stephen hemminger 已提交
322 323 324 325 326 327 328 329 330 331
		goto nla_put_failure;

	ci.ndm_used	 = jiffies_to_clock_t(now - fdb->used);
	ci.ndm_confirmed = 0;
	ci.ndm_updated	 = jiffies_to_clock_t(now - fdb->updated);
	ci.ndm_refcnt	 = 0;

	if (nla_put(skb, NDA_CACHEINFO, sizeof(ci), &ci))
		goto nla_put_failure;

332 333
	nlmsg_end(skb, nlh);
	return 0;
S
stephen hemminger 已提交
334 335 336 337 338 339 340 341 342 343

nla_put_failure:
	nlmsg_cancel(skb, nlh);
	return -EMSGSIZE;
}

static inline size_t vxlan_nlmsg_size(void)
{
	return NLMSG_ALIGN(sizeof(struct ndmsg))
		+ nla_total_size(ETH_ALEN) /* NDA_LLADDR */
C
Cong Wang 已提交
344
		+ nla_total_size(sizeof(struct in6_addr)) /* NDA_DST */
345
		+ nla_total_size(sizeof(__be16)) /* NDA_PORT */
346 347
		+ nla_total_size(sizeof(__be32)) /* NDA_VNI */
		+ nla_total_size(sizeof(__u32)) /* NDA_IFINDEX */
348
		+ nla_total_size(sizeof(__s32)) /* NDA_LINK_NETNSID */
S
stephen hemminger 已提交
349 350 351
		+ nla_total_size(sizeof(struct nda_cacheinfo));
}

352 353
static void vxlan_fdb_notify(struct vxlan_dev *vxlan, struct vxlan_fdb *fdb,
			     struct vxlan_rdst *rd, int type)
S
stephen hemminger 已提交
354 355 356 357 358 359 360 361 362
{
	struct net *net = dev_net(vxlan->dev);
	struct sk_buff *skb;
	int err = -ENOBUFS;

	skb = nlmsg_new(vxlan_nlmsg_size(), GFP_ATOMIC);
	if (skb == NULL)
		goto errout;

363
	err = vxlan_fdb_info(skb, vxlan, fdb, 0, 0, type, 0, rd);
S
stephen hemminger 已提交
364 365 366 367 368 369 370 371 372 373 374 375 376 377
	if (err < 0) {
		/* -EMSGSIZE implies BUG in vxlan_nlmsg_size() */
		WARN_ON(err == -EMSGSIZE);
		kfree_skb(skb);
		goto errout;
	}

	rtnl_notify(skb, net, 0, RTNLGRP_NEIGH, NULL, GFP_ATOMIC);
	return;
errout:
	if (err < 0)
		rtnl_set_sk_err(net, RTNLGRP_NEIGH, err);
}

C
Cong Wang 已提交
378
static void vxlan_ip_miss(struct net_device *dev, union vxlan_addr *ipa)
D
David Stevens 已提交
379 380
{
	struct vxlan_dev *vxlan = netdev_priv(dev);
381 382 383 384
	struct vxlan_fdb f = {
		.state = NUD_STALE,
	};
	struct vxlan_rdst remote = {
C
Cong Wang 已提交
385
		.remote_ip = *ipa, /* goes to NDA_DST */
386 387
		.remote_vni = VXLAN_N_VID,
	};
388

389
	vxlan_fdb_notify(vxlan, &f, &remote, RTM_GETNEIGH);
D
David Stevens 已提交
390 391 392 393
}

static void vxlan_fdb_miss(struct vxlan_dev *vxlan, const u8 eth_addr[ETH_ALEN])
{
394 395 396
	struct vxlan_fdb f = {
		.state = NUD_STALE,
	};
397
	struct vxlan_rdst remote = { };
D
David Stevens 已提交
398 399 400

	memcpy(f.eth_addr, eth_addr, ETH_ALEN);

401
	vxlan_fdb_notify(vxlan, &f, &remote, RTM_GETNEIGH);
D
David Stevens 已提交
402 403
}

S
stephen hemminger 已提交
404 405 406 407 408 409 410 411
/* Hash Ethernet address */
static u32 eth_hash(const unsigned char *addr)
{
	u64 value = get_unaligned((u64 *)addr);

	/* only want 6 bytes */
#ifdef __BIG_ENDIAN
	value >>= 16;
412 413
#else
	value <<= 16;
S
stephen hemminger 已提交
414 415 416 417 418 419 420 421 422 423 424 425
#endif
	return hash_64(value, FDB_HASH_BITS);
}

/* Hash chain to use given mac address */
static inline struct hlist_head *vxlan_fdb_head(struct vxlan_dev *vxlan,
						const u8 *mac)
{
	return &vxlan->fdb_head[eth_hash(mac)];
}

/* Look up Ethernet address in forwarding table */
426
static struct vxlan_fdb *__vxlan_find_mac(struct vxlan_dev *vxlan,
S
stephen hemminger 已提交
427 428 429 430 431
					const u8 *mac)
{
	struct hlist_head *head = vxlan_fdb_head(vxlan, mac);
	struct vxlan_fdb *f;

432
	hlist_for_each_entry_rcu(f, head, hlist) {
433
		if (ether_addr_equal(mac, f->eth_addr))
S
stephen hemminger 已提交
434 435 436 437 438 439
			return f;
	}

	return NULL;
}

440 441 442 443 444 445 446 447 448 449 450 451
static struct vxlan_fdb *vxlan_find_mac(struct vxlan_dev *vxlan,
					const u8 *mac)
{
	struct vxlan_fdb *f;

	f = __vxlan_find_mac(vxlan, mac);
	if (f)
		f->used = jiffies;

	return f;
}

452 453
/* caller should hold vxlan->hash_lock */
static struct vxlan_rdst *vxlan_fdb_find_rdst(struct vxlan_fdb *f,
C
Cong Wang 已提交
454
					      union vxlan_addr *ip, __be16 port,
455
					      __u32 vni, __u32 ifindex)
456
{
457
	struct vxlan_rdst *rd;
458

459
	list_for_each_entry(rd, &f->remotes, list) {
C
Cong Wang 已提交
460
		if (vxlan_addr_equal(&rd->remote_ip, ip) &&
461 462 463
		    rd->remote_port == port &&
		    rd->remote_vni == vni &&
		    rd->remote_ifindex == ifindex)
464
			return rd;
465
	}
466

467 468 469
	return NULL;
}

470 471
/* Replace destination of unicast mac */
static int vxlan_fdb_replace(struct vxlan_fdb *f,
C
Cong Wang 已提交
472
			     union vxlan_addr *ip, __be16 port, __u32 vni, __u32 ifindex)
473 474 475 476 477 478 479 480 481 482
{
	struct vxlan_rdst *rd;

	rd = vxlan_fdb_find_rdst(f, ip, port, vni, ifindex);
	if (rd)
		return 0;

	rd = list_first_entry_or_null(&f->remotes, struct vxlan_rdst, list);
	if (!rd)
		return 0;
C
Cong Wang 已提交
483
	rd->remote_ip = *ip;
484 485 486 487 488 489
	rd->remote_port = port;
	rd->remote_vni = vni;
	rd->remote_ifindex = ifindex;
	return 1;
}

490 491
/* Add/update destinations for multicast */
static int vxlan_fdb_append(struct vxlan_fdb *f,
492 493
			    union vxlan_addr *ip, __be16 port, __u32 vni,
			    __u32 ifindex, struct vxlan_rdst **rdp)
494 495 496 497 498 499 500
{
	struct vxlan_rdst *rd;

	rd = vxlan_fdb_find_rdst(f, ip, port, vni, ifindex);
	if (rd)
		return 0;

501 502 503
	rd = kmalloc(sizeof(*rd), GFP_ATOMIC);
	if (rd == NULL)
		return -ENOBUFS;
C
Cong Wang 已提交
504
	rd->remote_ip = *ip;
505 506 507
	rd->remote_port = port;
	rd->remote_vni = vni;
	rd->remote_ifindex = ifindex;
508 509 510

	list_add_tail_rcu(&rd->list, &f->remotes);

511
	*rdp = rd;
512 513 514
	return 1;
}

T
Tom Herbert 已提交
515 516 517
static struct vxlanhdr *vxlan_gro_remcsum(struct sk_buff *skb,
					  unsigned int off,
					  struct vxlanhdr *vh, size_t hdrlen,
518 519
					  u32 data, struct gro_remcsum *grc,
					  bool nopartial)
T
Tom Herbert 已提交
520
{
521
	size_t start, offset;
T
Tom Herbert 已提交
522 523

	if (skb->remcsum_offload)
524
		return vh;
T
Tom Herbert 已提交
525 526 527 528 529 530 531 532 533

	if (!NAPI_GRO_CB(skb)->csum_valid)
		return NULL;

	start = (data & VXLAN_RCO_MASK) << VXLAN_RCO_SHIFT;
	offset = start + ((data & VXLAN_RCO_UDP) ?
			  offsetof(struct udphdr, check) :
			  offsetof(struct tcphdr, check));

534 535
	vh = skb_gro_remcsum_process(skb, (void *)vh, off, hdrlen,
				     start, offset, grc, nopartial);
T
Tom Herbert 已提交
536 537 538 539 540 541

	skb->remcsum_offload = 1;

	return vh;
}

542 543 544
static struct sk_buff **vxlan_gro_receive(struct sk_buff **head,
					  struct sk_buff *skb,
					  struct udp_offload *uoff)
545 546 547
{
	struct sk_buff *p, **pp = NULL;
	struct vxlanhdr *vh, *vh2;
548
	unsigned int hlen, off_vx;
549
	int flush = 1;
T
Tom Herbert 已提交
550 551 552
	struct vxlan_sock *vs = container_of(uoff, struct vxlan_sock,
					     udp_offloads);
	u32 flags;
553 554 555
	struct gro_remcsum grc;

	skb_gro_remcsum_init(&grc);
556 557 558 559 560 561 562 563 564 565

	off_vx = skb_gro_offset(skb);
	hlen = off_vx + sizeof(*vh);
	vh   = skb_gro_header_fast(skb, off_vx);
	if (skb_gro_header_hard(skb, hlen)) {
		vh = skb_gro_header_slow(skb, hlen, off_vx);
		if (unlikely(!vh))
			goto out;
	}

T
Tom Herbert 已提交
566 567 568 569 570 571
	skb_gro_postpull_rcsum(skb, vh, sizeof(struct vxlanhdr));

	flags = ntohl(vh->vx_flags);

	if ((flags & VXLAN_HF_RCO) && (vs->flags & VXLAN_F_REMCSUM_RX)) {
		vh = vxlan_gro_remcsum(skb, off_vx, vh, sizeof(struct vxlanhdr),
572 573 574
				       ntohl(vh->vx_vni), &grc,
				       !!(vs->flags &
					  VXLAN_F_REMCSUM_NOPARTIAL));
T
Tom Herbert 已提交
575 576 577 578 579

		if (!vh)
			goto out;
	}

580 581
	skb_gro_pull(skb, sizeof(struct vxlanhdr)); /* pull vxlan header */

582 583 584 585 586 587 588
	flush = 0;

	for (p = *head; p; p = p->next) {
		if (!NAPI_GRO_CB(p)->same_flow)
			continue;

		vh2 = (struct vxlanhdr *)(p->data + off_vx);
T
Thomas Graf 已提交
589 590
		if (vh->vx_flags != vh2->vx_flags ||
		    vh->vx_vni != vh2->vx_vni) {
591 592 593 594 595
			NAPI_GRO_CB(p)->same_flow = 0;
			continue;
		}
	}

596
	pp = eth_gro_receive(head, skb);
597 598

out:
599
	skb_gro_remcsum_cleanup(skb, &grc);
600 601 602 603 604
	NAPI_GRO_CB(skb)->flush |= flush;

	return pp;
}

605 606
static int vxlan_gro_complete(struct sk_buff *skb, int nhoff,
			      struct udp_offload *uoff)
607
{
608 609
	udp_tunnel_gro_complete(skb, nhoff);

610
	return eth_gro_complete(skb, nhoff + sizeof(struct vxlanhdr));
611 612
}

613
/* Notify netdevs that UDP port started listening */
614
static void vxlan_notify_add_rx_port(struct vxlan_sock *vs)
615 616
{
	struct net_device *dev;
617
	struct sock *sk = vs->sock->sk;
618
	struct net *net = sock_net(sk);
619
	sa_family_t sa_family = vxlan_get_sk_family(vs);
J
Joseph Gasparakis 已提交
620
	__be16 port = inet_sk(sk)->inet_sport;
621 622 623
	int err;

	if (sa_family == AF_INET) {
624
		err = udp_add_offload(net, &vs->udp_offloads);
625 626 627
		if (err)
			pr_warn("vxlan: udp_add_offload failed with status %d\n", err);
	}
628 629 630 631 632 633 634 635 636 637 638

	rcu_read_lock();
	for_each_netdev_rcu(net, dev) {
		if (dev->netdev_ops->ndo_add_vxlan_port)
			dev->netdev_ops->ndo_add_vxlan_port(dev, sa_family,
							    port);
	}
	rcu_read_unlock();
}

/* Notify netdevs that UDP port is no more listening */
639
static void vxlan_notify_del_rx_port(struct vxlan_sock *vs)
640 641
{
	struct net_device *dev;
642
	struct sock *sk = vs->sock->sk;
643
	struct net *net = sock_net(sk);
644
	sa_family_t sa_family = vxlan_get_sk_family(vs);
J
Joseph Gasparakis 已提交
645
	__be16 port = inet_sk(sk)->inet_sport;
646 647 648 649 650 651 652 653

	rcu_read_lock();
	for_each_netdev_rcu(net, dev) {
		if (dev->netdev_ops->ndo_del_vxlan_port)
			dev->netdev_ops->ndo_del_vxlan_port(dev, sa_family,
							    port);
	}
	rcu_read_unlock();
654 655 656

	if (sa_family == AF_INET)
		udp_del_offload(&vs->udp_offloads);
657 658
}

S
stephen hemminger 已提交
659 660
/* Add new entry to forwarding table -- assumes lock held */
static int vxlan_fdb_create(struct vxlan_dev *vxlan,
C
Cong Wang 已提交
661
			    const u8 *mac, union vxlan_addr *ip,
662
			    __u16 state, __u16 flags,
663
			    __be16 port, __u32 vni, __u32 ifindex,
664
			    __u8 ndm_flags)
S
stephen hemminger 已提交
665
{
666
	struct vxlan_rdst *rd = NULL;
S
stephen hemminger 已提交
667 668 669
	struct vxlan_fdb *f;
	int notify = 0;

670
	f = __vxlan_find_mac(vxlan, mac);
S
stephen hemminger 已提交
671 672 673 674 675 676 677 678 679 680 681
	if (f) {
		if (flags & NLM_F_EXCL) {
			netdev_dbg(vxlan->dev,
				   "lost race to create %pM\n", mac);
			return -EEXIST;
		}
		if (f->state != state) {
			f->state = state;
			f->updated = jiffies;
			notify = 1;
		}
682 683 684 685 686
		if (f->flags != ndm_flags) {
			f->flags = ndm_flags;
			f->updated = jiffies;
			notify = 1;
		}
687 688 689 690
		if ((flags & NLM_F_REPLACE)) {
			/* Only change unicasts */
			if (!(is_multicast_ether_addr(f->eth_addr) ||
			     is_zero_ether_addr(f->eth_addr))) {
691
				notify |= vxlan_fdb_replace(f, ip, port, vni,
692 693 694 695
							   ifindex);
			} else
				return -EOPNOTSUPP;
		}
696
		if ((flags & NLM_F_APPEND) &&
697 698
		    (is_multicast_ether_addr(f->eth_addr) ||
		     is_zero_ether_addr(f->eth_addr))) {
699 700
			int rc = vxlan_fdb_append(f, ip, port, vni, ifindex,
						  &rd);
701 702 703 704 705

			if (rc < 0)
				return rc;
			notify |= rc;
		}
S
stephen hemminger 已提交
706 707 708 709
	} else {
		if (!(flags & NLM_F_CREATE))
			return -ENOENT;

710 711
		if (vxlan->cfg.addrmax &&
		    vxlan->addrcnt >= vxlan->cfg.addrmax)
S
stephen hemminger 已提交
712 713
			return -ENOSPC;

714 715 716 717 718
		/* Disallow replace to add a multicast entry */
		if ((flags & NLM_F_REPLACE) &&
		    (is_multicast_ether_addr(mac) || is_zero_ether_addr(mac)))
			return -EOPNOTSUPP;

C
Cong Wang 已提交
719
		netdev_dbg(vxlan->dev, "add %pM -> %pIS\n", mac, ip);
S
stephen hemminger 已提交
720 721 722 723 724 725
		f = kmalloc(sizeof(*f), GFP_ATOMIC);
		if (!f)
			return -ENOMEM;

		notify = 1;
		f->state = state;
726
		f->flags = ndm_flags;
S
stephen hemminger 已提交
727
		f->updated = f->used = jiffies;
728
		INIT_LIST_HEAD(&f->remotes);
S
stephen hemminger 已提交
729 730
		memcpy(f->eth_addr, mac, ETH_ALEN);

731
		vxlan_fdb_append(f, ip, port, vni, ifindex, &rd);
732

S
stephen hemminger 已提交
733 734 735 736 737
		++vxlan->addrcnt;
		hlist_add_head_rcu(&f->hlist,
				   vxlan_fdb_head(vxlan, mac));
	}

738 739 740 741 742
	if (notify) {
		if (rd == NULL)
			rd = first_remote_rtnl(f);
		vxlan_fdb_notify(vxlan, f, rd, RTM_NEWNEIGH);
	}
S
stephen hemminger 已提交
743 744 745 746

	return 0;
}

W
Wei Yongjun 已提交
747
static void vxlan_fdb_free(struct rcu_head *head)
748 749
{
	struct vxlan_fdb *f = container_of(head, struct vxlan_fdb, rcu);
750
	struct vxlan_rdst *rd, *nd;
751

752
	list_for_each_entry_safe(rd, nd, &f->remotes, list)
753 754 755 756
		kfree(rd);
	kfree(f);
}

S
stephen hemminger 已提交
757 758 759 760 761 762
static void vxlan_fdb_destroy(struct vxlan_dev *vxlan, struct vxlan_fdb *f)
{
	netdev_dbg(vxlan->dev,
		    "delete %pM\n", f->eth_addr);

	--vxlan->addrcnt;
763
	vxlan_fdb_notify(vxlan, f, first_remote_rtnl(f), RTM_DELNEIGH);
S
stephen hemminger 已提交
764 765

	hlist_del_rcu(&f->hlist);
766
	call_rcu(&f->rcu, vxlan_fdb_free);
S
stephen hemminger 已提交
767 768
}

M
Mike Rapoport 已提交
769
static int vxlan_fdb_parse(struct nlattr *tb[], struct vxlan_dev *vxlan,
C
Cong Wang 已提交
770
			   union vxlan_addr *ip, __be16 *port, u32 *vni, u32 *ifindex)
S
stephen hemminger 已提交
771
{
772
	struct net *net = dev_net(vxlan->dev);
C
Cong Wang 已提交
773
	int err;
S
stephen hemminger 已提交
774

M
Mike Rapoport 已提交
775
	if (tb[NDA_DST]) {
C
Cong Wang 已提交
776 777 778
		err = vxlan_nla_get_addr(ip, tb[NDA_DST]);
		if (err)
			return err;
M
Mike Rapoport 已提交
779
	} else {
C
Cong Wang 已提交
780 781 782 783 784 785 786 787 788 789
		union vxlan_addr *remote = &vxlan->default_dst.remote_ip;
		if (remote->sa.sa_family == AF_INET) {
			ip->sin.sin_addr.s_addr = htonl(INADDR_ANY);
			ip->sa.sa_family = AF_INET;
#if IS_ENABLED(CONFIG_IPV6)
		} else {
			ip->sin6.sin6_addr = in6addr_any;
			ip->sa.sa_family = AF_INET6;
#endif
		}
M
Mike Rapoport 已提交
790
	}
S
stephen hemminger 已提交
791

792
	if (tb[NDA_PORT]) {
793
		if (nla_len(tb[NDA_PORT]) != sizeof(__be16))
794
			return -EINVAL;
M
Mike Rapoport 已提交
795 796
		*port = nla_get_be16(tb[NDA_PORT]);
	} else {
797
		*port = vxlan->cfg.dst_port;
M
Mike Rapoport 已提交
798
	}
799 800 801 802

	if (tb[NDA_VNI]) {
		if (nla_len(tb[NDA_VNI]) != sizeof(u32))
			return -EINVAL;
M
Mike Rapoport 已提交
803 804 805 806
		*vni = nla_get_u32(tb[NDA_VNI]);
	} else {
		*vni = vxlan->default_dst.remote_vni;
	}
807 808

	if (tb[NDA_IFINDEX]) {
P
Pravin B Shelar 已提交
809
		struct net_device *tdev;
810 811 812

		if (nla_len(tb[NDA_IFINDEX]) != sizeof(u32))
			return -EINVAL;
M
Mike Rapoport 已提交
813
		*ifindex = nla_get_u32(tb[NDA_IFINDEX]);
814
		tdev = __dev_get_by_index(net, *ifindex);
P
Pravin B Shelar 已提交
815
		if (!tdev)
816
			return -EADDRNOTAVAIL;
M
Mike Rapoport 已提交
817 818 819 820 821 822 823 824 825 826
	} else {
		*ifindex = 0;
	}

	return 0;
}

/* Add static entry (via netlink) */
static int vxlan_fdb_add(struct ndmsg *ndm, struct nlattr *tb[],
			 struct net_device *dev,
827
			 const unsigned char *addr, u16 vid, u16 flags)
M
Mike Rapoport 已提交
828 829 830
{
	struct vxlan_dev *vxlan = netdev_priv(dev);
	/* struct net *net = dev_net(vxlan->dev); */
C
Cong Wang 已提交
831
	union vxlan_addr ip;
M
Mike Rapoport 已提交
832 833 834 835 836 837 838 839 840 841 842 843 844 845 846 847
	__be16 port;
	u32 vni, ifindex;
	int err;

	if (!(ndm->ndm_state & (NUD_PERMANENT|NUD_REACHABLE))) {
		pr_info("RTM_NEWNEIGH with invalid state %#x\n",
			ndm->ndm_state);
		return -EINVAL;
	}

	if (tb[NDA_DST] == NULL)
		return -EINVAL;

	err = vxlan_fdb_parse(tb, vxlan, &ip, &port, &vni, &ifindex);
	if (err)
		return err;
848

849 850 851
	if (vxlan->default_dst.remote_ip.sa.sa_family != ip.sa.sa_family)
		return -EAFNOSUPPORT;

S
stephen hemminger 已提交
852
	spin_lock_bh(&vxlan->hash_lock);
C
Cong Wang 已提交
853
	err = vxlan_fdb_create(vxlan, addr, &ip, ndm->ndm_state, flags,
854
			       port, vni, ifindex, ndm->ndm_flags);
S
stephen hemminger 已提交
855 856 857 858 859 860
	spin_unlock_bh(&vxlan->hash_lock);

	return err;
}

/* Delete entry (via netlink) */
861 862
static int vxlan_fdb_delete(struct ndmsg *ndm, struct nlattr *tb[],
			    struct net_device *dev,
863
			    const unsigned char *addr, u16 vid)
S
stephen hemminger 已提交
864 865 866
{
	struct vxlan_dev *vxlan = netdev_priv(dev);
	struct vxlan_fdb *f;
867
	struct vxlan_rdst *rd = NULL;
C
Cong Wang 已提交
868
	union vxlan_addr ip;
869 870 871 872 873 874 875 876 877
	__be16 port;
	u32 vni, ifindex;
	int err;

	err = vxlan_fdb_parse(tb, vxlan, &ip, &port, &vni, &ifindex);
	if (err)
		return err;

	err = -ENOENT;
S
stephen hemminger 已提交
878 879 880

	spin_lock_bh(&vxlan->hash_lock);
	f = vxlan_find_mac(vxlan, addr);
881 882 883
	if (!f)
		goto out;

C
Cong Wang 已提交
884 885
	if (!vxlan_addr_any(&ip)) {
		rd = vxlan_fdb_find_rdst(f, &ip, port, vni, ifindex);
886 887 888 889 890 891 892 893 894 895 896
		if (!rd)
			goto out;
	}

	err = 0;

	/* remove a destination if it's not the only one on the list,
	 * otherwise destroy the fdb entry
	 */
	if (rd && !list_is_singular(&f->remotes)) {
		list_del_rcu(&rd->list);
897
		vxlan_fdb_notify(vxlan, f, rd, RTM_DELNEIGH);
898
		kfree_rcu(rd, rcu);
899
		goto out;
S
stephen hemminger 已提交
900
	}
901 902 903 904

	vxlan_fdb_destroy(vxlan, f);

out:
S
stephen hemminger 已提交
905 906 907 908 909 910 911
	spin_unlock_bh(&vxlan->hash_lock);

	return err;
}

/* Dump forwarding table */
static int vxlan_fdb_dump(struct sk_buff *skb, struct netlink_callback *cb,
912 913
			  struct net_device *dev,
			  struct net_device *filter_dev, int idx)
S
stephen hemminger 已提交
914 915 916 917 918 919 920 921
{
	struct vxlan_dev *vxlan = netdev_priv(dev);
	unsigned int h;

	for (h = 0; h < FDB_HASH_SIZE; ++h) {
		struct vxlan_fdb *f;
		int err;

922
		hlist_for_each_entry_rcu(f, &vxlan->fdb_head[h], hlist) {
923 924
			struct vxlan_rdst *rd;

925
			list_for_each_entry_rcu(rd, &f->remotes, list) {
926 927 928
				if (idx < cb->args[0])
					goto skip;

929 930 931 932 933 934
				err = vxlan_fdb_info(skb, vxlan, f,
						     NETLINK_CB(cb->skb).portid,
						     cb->nlh->nlmsg_seq,
						     RTM_NEWNEIGH,
						     NLM_F_MULTI, rd);
				if (err < 0)
935 936
					goto out;
skip:
937 938
				++idx;
			}
S
stephen hemminger 已提交
939 940
		}
	}
941
out:
S
stephen hemminger 已提交
942 943 944 945 946
	return idx;
}

/* Watch incoming packets to learn mapping between Ethernet address
 * and Tunnel endpoint.
947
 * Return true if packet is bogus and should be dropped.
S
stephen hemminger 已提交
948
 */
949
static bool vxlan_snoop(struct net_device *dev,
C
Cong Wang 已提交
950
			union vxlan_addr *src_ip, const u8 *src_mac)
S
stephen hemminger 已提交
951 952 953 954 955 956
{
	struct vxlan_dev *vxlan = netdev_priv(dev);
	struct vxlan_fdb *f;

	f = vxlan_find_mac(vxlan, src_mac);
	if (likely(f)) {
957
		struct vxlan_rdst *rdst = first_remote_rcu(f);
958

C
Cong Wang 已提交
959
		if (likely(vxlan_addr_equal(&rdst->remote_ip, src_ip)))
960 961 962
			return false;

		/* Don't migrate static entries, drop packets */
963
		if (f->state & NUD_NOARP)
964
			return true;
S
stephen hemminger 已提交
965 966 967

		if (net_ratelimit())
			netdev_info(dev,
C
Cong Wang 已提交
968
				    "%pM migrated from %pIS to %pIS\n",
969
				    src_mac, &rdst->remote_ip.sa, &src_ip->sa);
S
stephen hemminger 已提交
970

C
Cong Wang 已提交
971
		rdst->remote_ip = *src_ip;
S
stephen hemminger 已提交
972
		f->updated = jiffies;
973
		vxlan_fdb_notify(vxlan, f, rdst, RTM_NEWNEIGH);
S
stephen hemminger 已提交
974 975 976
	} else {
		/* learned new entry */
		spin_lock(&vxlan->hash_lock);
977 978 979 980 981 982

		/* close off race between vxlan_flush and incoming packets */
		if (netif_running(dev))
			vxlan_fdb_create(vxlan, src_mac, src_ip,
					 NUD_REACHABLE,
					 NLM_F_EXCL|NLM_F_CREATE,
983
					 vxlan->cfg.dst_port,
984 985
					 vxlan->default_dst.remote_vni,
					 0, NTF_SELF);
S
stephen hemminger 已提交
986 987
		spin_unlock(&vxlan->hash_lock);
	}
988 989

	return false;
S
stephen hemminger 已提交
990 991 992
}

/* See if multicast group is already in use by other ID */
993
static bool vxlan_group_used(struct vxlan_net *vn, struct vxlan_dev *dev)
S
stephen hemminger 已提交
994
{
995
	struct vxlan_dev *vxlan;
996
	unsigned short family = dev->default_dst.remote_ip.sa.sa_family;
S
stephen hemminger 已提交
997

998 999 1000
	/* The vxlan_sock is only used by dev, leaving group has
	 * no effect on other vxlan devices.
	 */
1001 1002
	if (family == AF_INET && dev->vn4_sock &&
	    atomic_read(&dev->vn4_sock->refcnt) == 1)
1003
		return false;
1004 1005 1006 1007 1008
#if IS_ENABLED(CONFIG_IPV6)
	if (family == AF_INET6 && dev->vn6_sock &&
	    atomic_read(&dev->vn6_sock->refcnt) == 1)
		return false;
#endif
1009

1010
	list_for_each_entry(vxlan, &vn->vxlan_list, next) {
1011
		if (!netif_running(vxlan->dev) || vxlan == dev)
1012
			continue;
S
stephen hemminger 已提交
1013

1014
		if (family == AF_INET && vxlan->vn4_sock != dev->vn4_sock)
1015
			continue;
1016 1017 1018 1019
#if IS_ENABLED(CONFIG_IPV6)
		if (family == AF_INET6 && vxlan->vn6_sock != dev->vn6_sock)
			continue;
#endif
1020 1021 1022 1023 1024 1025 1026 1027 1028 1029

		if (!vxlan_addr_equal(&vxlan->default_dst.remote_ip,
				      &dev->default_dst.remote_ip))
			continue;

		if (vxlan->default_dst.remote_ifindex !=
		    dev->default_dst.remote_ifindex)
			continue;

		return true;
1030
	}
S
stephen hemminger 已提交
1031 1032 1033 1034

	return false;
}

1035
static void __vxlan_sock_release(struct vxlan_sock *vs)
1036
{
1037
	struct vxlan_net *vn;
1038

1039 1040
	if (!vs)
		return;
1041 1042
	if (!atomic_dec_and_test(&vs->refcnt))
		return;
S
stephen hemminger 已提交
1043

1044
	vn = net_generic(sock_net(vs->sock->sk), vxlan_net_id);
1045
	spin_lock(&vn->sock_lock);
1046
	hlist_del_rcu(&vs->hlist);
1047
	vxlan_notify_del_rx_port(vs);
1048 1049
	spin_unlock(&vn->sock_lock);

1050
	queue_work(vxlan_wq, &vs->del_work);
S
stephen hemminger 已提交
1051 1052
}

1053 1054 1055 1056 1057 1058 1059 1060
static void vxlan_sock_release(struct vxlan_dev *vxlan)
{
	__vxlan_sock_release(vxlan->vn4_sock);
#if IS_ENABLED(CONFIG_IPV6)
	__vxlan_sock_release(vxlan->vn6_sock);
#endif
}

1061
/* Update multicast group membership when first VNI on
1062
 * multicast address is brought up
1063
 */
1064
static int vxlan_igmp_join(struct vxlan_dev *vxlan)
S
stephen hemminger 已提交
1065
{
1066
	struct sock *sk;
C
Cong Wang 已提交
1067 1068
	union vxlan_addr *ip = &vxlan->default_dst.remote_ip;
	int ifindex = vxlan->default_dst.remote_ifindex;
1069
	int ret = -EINVAL;
S
stephen hemminger 已提交
1070

C
Cong Wang 已提交
1071 1072 1073 1074 1075 1076
	if (ip->sa.sa_family == AF_INET) {
		struct ip_mreqn mreq = {
			.imr_multiaddr.s_addr	= ip->sin.sin_addr.s_addr,
			.imr_ifindex		= ifindex,
		};

1077 1078
		sk = vxlan->vn4_sock->sock->sk;
		lock_sock(sk);
1079
		ret = ip_mc_join_group(sk, &mreq);
1080
		release_sock(sk);
C
Cong Wang 已提交
1081 1082
#if IS_ENABLED(CONFIG_IPV6)
	} else {
1083 1084
		sk = vxlan->vn6_sock->sock->sk;
		lock_sock(sk);
1085 1086
		ret = ipv6_stub->ipv6_sock_mc_join(sk, ifindex,
						   &ip->sin6.sin6_addr);
1087
		release_sock(sk);
C
Cong Wang 已提交
1088 1089
#endif
	}
S
stephen hemminger 已提交
1090

1091
	return ret;
S
stephen hemminger 已提交
1092 1093 1094
}

/* Inverse of vxlan_igmp_join when last VNI is brought down */
1095
static int vxlan_igmp_leave(struct vxlan_dev *vxlan)
S
stephen hemminger 已提交
1096
{
1097
	struct sock *sk;
C
Cong Wang 已提交
1098 1099
	union vxlan_addr *ip = &vxlan->default_dst.remote_ip;
	int ifindex = vxlan->default_dst.remote_ifindex;
1100
	int ret = -EINVAL;
S
stephen hemminger 已提交
1101

C
Cong Wang 已提交
1102 1103 1104 1105 1106 1107
	if (ip->sa.sa_family == AF_INET) {
		struct ip_mreqn mreq = {
			.imr_multiaddr.s_addr	= ip->sin.sin_addr.s_addr,
			.imr_ifindex		= ifindex,
		};

1108 1109
		sk = vxlan->vn4_sock->sock->sk;
		lock_sock(sk);
1110
		ret = ip_mc_leave_group(sk, &mreq);
1111
		release_sock(sk);
C
Cong Wang 已提交
1112 1113
#if IS_ENABLED(CONFIG_IPV6)
	} else {
1114 1115
		sk = vxlan->vn6_sock->sock->sk;
		lock_sock(sk);
1116 1117
		ret = ipv6_stub->ipv6_sock_mc_drop(sk, ifindex,
						   &ip->sin6.sin6_addr);
1118
		release_sock(sk);
C
Cong Wang 已提交
1119 1120
#endif
	}
S
stephen hemminger 已提交
1121

1122
	return ret;
S
stephen hemminger 已提交
1123 1124
}

T
Tom Herbert 已提交
1125
static struct vxlanhdr *vxlan_remcsum(struct sk_buff *skb, struct vxlanhdr *vh,
1126
				      size_t hdrlen, u32 data, bool nopartial)
T
Tom Herbert 已提交
1127 1128 1129
{
	size_t start, offset, plen;

1130 1131 1132
	if (skb->remcsum_offload)
		return vh;

T
Tom Herbert 已提交
1133 1134 1135 1136 1137 1138 1139 1140 1141 1142 1143 1144
	start = (data & VXLAN_RCO_MASK) << VXLAN_RCO_SHIFT;
	offset = start + ((data & VXLAN_RCO_UDP) ?
			  offsetof(struct udphdr, check) :
			  offsetof(struct tcphdr, check));

	plen = hdrlen + offset + sizeof(u16);

	if (!pskb_may_pull(skb, plen))
		return NULL;

	vh = (struct vxlanhdr *)(udp_hdr(skb) + 1);

1145 1146
	skb_remcsum_process(skb, (void *)vh + hdrlen, start, offset,
			    nopartial);
T
Tom Herbert 已提交
1147 1148 1149 1150

	return vh;
}

1151 1152 1153 1154 1155 1156 1157 1158 1159 1160 1161 1162
static void vxlan_rcv(struct vxlan_sock *vs, struct sk_buff *skb,
		      struct vxlan_metadata *md, u32 vni,
		      struct metadata_dst *tun_dst)
{
	struct iphdr *oip = NULL;
	struct ipv6hdr *oip6 = NULL;
	struct vxlan_dev *vxlan;
	struct pcpu_sw_netstats *stats;
	union vxlan_addr saddr;
	int err = 0;

	/* For flow based devices, map all packets to VNI 0 */
1163
	if (vs->flags & VXLAN_F_COLLECT_METADATA)
1164 1165 1166 1167 1168 1169 1170 1171 1172 1173 1174 1175 1176 1177 1178 1179
		vni = 0;

	/* Is this VNI defined? */
	vxlan = vxlan_vs_find_vni(vs, vni);
	if (!vxlan)
		goto drop;

	skb_reset_mac_header(skb);
	skb_scrub_packet(skb, !net_eq(vxlan->net, dev_net(vxlan->dev)));
	skb->protocol = eth_type_trans(skb, vxlan->dev);
	skb_postpull_rcsum(skb, eth_hdr(skb), ETH_HLEN);

	/* Ignore packet loops (and multicast echo) */
	if (ether_addr_equal(eth_hdr(skb)->h_source, vxlan->dev->dev_addr))
		goto drop;

1180 1181
	/* Get data from the outer IP header */
	if (vxlan_get_sk_family(vs) == AF_INET) {
1182 1183 1184 1185 1186 1187 1188 1189 1190 1191 1192 1193 1194 1195 1196 1197 1198 1199 1200 1201 1202 1203
		oip = ip_hdr(skb);
		saddr.sin.sin_addr.s_addr = oip->saddr;
		saddr.sa.sa_family = AF_INET;
#if IS_ENABLED(CONFIG_IPV6)
	} else {
		oip6 = ipv6_hdr(skb);
		saddr.sin6.sin6_addr = oip6->saddr;
		saddr.sa.sa_family = AF_INET6;
#endif
	}

	if (tun_dst) {
		skb_dst_set(skb, (struct dst_entry *)tun_dst);
		tun_dst = NULL;
	}

	if ((vxlan->flags & VXLAN_F_LEARN) &&
	    vxlan_snoop(skb->dev, &saddr, eth_hdr(skb)->h_source))
		goto drop;

	skb_reset_network_header(skb);
	/* In flow-based mode, GBP is carried in dst_metadata */
1204
	if (!(vs->flags & VXLAN_F_COLLECT_METADATA))
1205 1206 1207 1208 1209 1210 1211 1212 1213 1214 1215 1216 1217 1218 1219 1220 1221 1222 1223 1224 1225 1226 1227 1228 1229 1230 1231 1232 1233
		skb->mark = md->gbp;

	if (oip6)
		err = IP6_ECN_decapsulate(oip6, skb);
	if (oip)
		err = IP_ECN_decapsulate(oip, skb);

	if (unlikely(err)) {
		if (log_ecn_error) {
			if (oip6)
				net_info_ratelimited("non-ECT from %pI6\n",
						     &oip6->saddr);
			if (oip)
				net_info_ratelimited("non-ECT from %pI4 with TOS=%#x\n",
						     &oip->saddr, oip->tos);
		}
		if (err > 1) {
			++vxlan->dev->stats.rx_frame_errors;
			++vxlan->dev->stats.rx_errors;
			goto drop;
		}
	}

	stats = this_cpu_ptr(vxlan->dev->tstats);
	u64_stats_update_begin(&stats->syncp);
	stats->rx_packets++;
	stats->rx_bytes += skb->len;
	u64_stats_update_end(&stats->syncp);

T
Tom Herbert 已提交
1234
	gro_cells_receive(&vxlan->gro_cells, skb);
1235 1236 1237 1238 1239 1240 1241 1242 1243 1244

	return;
drop:
	if (tun_dst)
		dst_release((struct dst_entry *)tun_dst);

	/* Consume bad packet */
	kfree_skb(skb);
}

S
stephen hemminger 已提交
1245 1246 1247
/* Callback from net/ipv4/udp.c to receive packets */
static int vxlan_udp_encap_recv(struct sock *sk, struct sk_buff *skb)
{
T
Thomas Graf 已提交
1248
	struct metadata_dst *tun_dst = NULL;
P
Pravin B Shelar 已提交
1249
	struct vxlan_sock *vs;
S
stephen hemminger 已提交
1250
	struct vxlanhdr *vxh;
1251
	u32 flags, vni;
T
Thomas Graf 已提交
1252 1253
	struct vxlan_metadata _md;
	struct vxlan_metadata *md = &_md;
S
stephen hemminger 已提交
1254 1255

	/* Need Vxlan and inner Ethernet header to be present */
1256
	if (!pskb_may_pull(skb, VXLAN_HLEN))
S
stephen hemminger 已提交
1257 1258
		goto error;

1259
	vxh = (struct vxlanhdr *)(udp_hdr(skb) + 1);
1260 1261 1262 1263 1264 1265 1266 1267
	flags = ntohl(vxh->vx_flags);
	vni = ntohl(vxh->vx_vni);

	if (flags & VXLAN_HF_VNI) {
		flags &= ~VXLAN_HF_VNI;
	} else {
		/* VNI flag always required to be set */
		goto bad_flags;
S
stephen hemminger 已提交
1268 1269
	}

P
Pravin B Shelar 已提交
1270 1271
	if (iptunnel_pull_header(skb, VXLAN_HLEN, htons(ETH_P_TEB)))
		goto drop;
T
Tom Herbert 已提交
1272
	vxh = (struct vxlanhdr *)(udp_hdr(skb) + 1);
P
Pravin B Shelar 已提交
1273

1274
	vs = rcu_dereference_sk_user_data(sk);
P
Pravin B Shelar 已提交
1275
	if (!vs)
S
stephen hemminger 已提交
1276 1277
		goto drop;

T
Tom Herbert 已提交
1278
	if ((flags & VXLAN_HF_RCO) && (vs->flags & VXLAN_F_REMCSUM_RX)) {
1279 1280
		vxh = vxlan_remcsum(skb, vxh, sizeof(struct vxlanhdr), vni,
				    !!(vs->flags & VXLAN_F_REMCSUM_NOPARTIAL));
T
Tom Herbert 已提交
1281 1282 1283 1284
		if (!vxh)
			goto drop;

		flags &= ~VXLAN_HF_RCO;
1285
		vni &= VXLAN_VNI_MASK;
T
Tom Herbert 已提交
1286 1287
	}

T
Thomas Graf 已提交
1288
	if (vxlan_collect_metadata(vs)) {
1289 1290 1291
		tun_dst = udp_tun_rx_dst(skb, vxlan_get_sk_family(vs), TUNNEL_KEY,
					 cpu_to_be64(vni >> 8), sizeof(*md));

T
Thomas Graf 已提交
1292 1293 1294
		if (!tun_dst)
			goto drop;

1295
		md = ip_tunnel_info_opts(&tun_dst->u.tun_info);
T
Thomas Graf 已提交
1296 1297 1298 1299
	} else {
		memset(md, 0, sizeof(*md));
	}

T
Thomas Graf 已提交
1300 1301 1302 1303 1304 1305 1306
	/* For backwards compatibility, only allow reserved fields to be
	 * used by VXLAN extensions if explicitly requested.
	 */
	if ((flags & VXLAN_HF_GBP) && (vs->flags & VXLAN_F_GBP)) {
		struct vxlanhdr_gbp *gbp;

		gbp = (struct vxlanhdr_gbp *)vxh;
T
Thomas Graf 已提交
1307 1308 1309
		md->gbp = ntohs(gbp->policy_id);

		if (tun_dst)
1310
			tun_dst->u.tun_info.key.tun_flags |= TUNNEL_VXLAN_OPT;
T
Thomas Graf 已提交
1311 1312

		if (gbp->dont_learn)
T
Thomas Graf 已提交
1313
			md->gbp |= VXLAN_GBP_DONT_LEARN;
T
Thomas Graf 已提交
1314 1315

		if (gbp->policy_applied)
T
Thomas Graf 已提交
1316
			md->gbp |= VXLAN_GBP_POLICY_APPLIED;
T
Thomas Graf 已提交
1317 1318 1319 1320

		flags &= ~VXLAN_GBP_USED_BITS;
	}

1321
	if (flags || vni & ~VXLAN_VNI_MASK) {
1322 1323 1324 1325
		/* If there are any unprocessed flags remaining treat
		 * this as a malformed packet. This behavior diverges from
		 * VXLAN RFC (RFC7348) which stipulates that bits in reserved
		 * in reserved fields are to be ignored. The approach here
1326
		 * maintains compatibility with previous stack code, and also
1327 1328 1329 1330 1331 1332 1333
		 * is more robust and provides a little more security in
		 * adding extensions to VXLAN.
		 */

		goto bad_flags;
	}

1334
	vxlan_rcv(vs, skb, md, vni >> 8, tun_dst);
P
Pravin B Shelar 已提交
1335 1336 1337 1338 1339 1340 1341
	return 0;

drop:
	/* Consume bad packet */
	kfree_skb(skb);
	return 0;

1342 1343 1344 1345
bad_flags:
	netdev_dbg(skb->dev, "invalid vxlan flags=%#x vni=%#x\n",
		   ntohl(vxh->vx_flags), ntohl(vxh->vx_vni));

P
Pravin B Shelar 已提交
1346
error:
T
Thomas Graf 已提交
1347 1348 1349
	if (tun_dst)
		dst_release((struct dst_entry *)tun_dst);

P
Pravin B Shelar 已提交
1350 1351 1352 1353
	/* Return non vxlan pkt */
	return 1;
}

D
David Stevens 已提交
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 1379 1380 1381 1382 1383 1384 1385 1386 1387 1388 1389 1390 1391 1392 1393 1394 1395 1396 1397 1398 1399 1400 1401
static int arp_reduce(struct net_device *dev, struct sk_buff *skb)
{
	struct vxlan_dev *vxlan = netdev_priv(dev);
	struct arphdr *parp;
	u8 *arpptr, *sha;
	__be32 sip, tip;
	struct neighbour *n;

	if (dev->flags & IFF_NOARP)
		goto out;

	if (!pskb_may_pull(skb, arp_hdr_len(dev))) {
		dev->stats.tx_dropped++;
		goto out;
	}
	parp = arp_hdr(skb);

	if ((parp->ar_hrd != htons(ARPHRD_ETHER) &&
	     parp->ar_hrd != htons(ARPHRD_IEEE802)) ||
	    parp->ar_pro != htons(ETH_P_IP) ||
	    parp->ar_op != htons(ARPOP_REQUEST) ||
	    parp->ar_hln != dev->addr_len ||
	    parp->ar_pln != 4)
		goto out;
	arpptr = (u8 *)parp + sizeof(struct arphdr);
	sha = arpptr;
	arpptr += dev->addr_len;	/* sha */
	memcpy(&sip, arpptr, sizeof(sip));
	arpptr += sizeof(sip);
	arpptr += dev->addr_len;	/* tha */
	memcpy(&tip, arpptr, sizeof(tip));

	if (ipv4_is_loopback(tip) ||
	    ipv4_is_multicast(tip))
		goto out;

	n = neigh_lookup(&arp_tbl, &tip, dev);

	if (n) {
		struct vxlan_fdb *f;
		struct sk_buff	*reply;

		if (!(n->nud_state & NUD_CONNECTED)) {
			neigh_release(n);
			goto out;
		}

		f = vxlan_find_mac(vxlan, n->ha);
C
Cong Wang 已提交
1402
		if (f && vxlan_addr_any(&(first_remote_rcu(f)->remote_ip))) {
D
David Stevens 已提交
1403 1404 1405 1406 1407 1408 1409 1410 1411 1412
			/* bridge-local neighbor */
			neigh_release(n);
			goto out;
		}

		reply = arp_create(ARPOP_REPLY, ETH_P_ARP, sip, dev, tip, sha,
				n->ha, sha);

		neigh_release(n);

1413 1414 1415
		if (reply == NULL)
			goto out;

D
David Stevens 已提交
1416 1417 1418 1419 1420 1421 1422
		skb_reset_mac_header(reply);
		__skb_pull(reply, skb_network_offset(reply));
		reply->ip_summed = CHECKSUM_UNNECESSARY;
		reply->pkt_type = PACKET_HOST;

		if (netif_rx_ni(reply) == NET_RX_DROP)
			dev->stats.rx_dropped++;
C
Cong Wang 已提交
1423 1424 1425
	} else if (vxlan->flags & VXLAN_F_L3MISS) {
		union vxlan_addr ipa = {
			.sin.sin_addr.s_addr = tip,
1426
			.sin.sin_family = AF_INET,
C
Cong Wang 已提交
1427 1428 1429 1430
		};

		vxlan_ip_miss(dev, &ipa);
	}
D
David Stevens 已提交
1431 1432 1433 1434 1435
out:
	consume_skb(skb);
	return NETDEV_TX_OK;
}

C
Cong Wang 已提交
1436
#if IS_ENABLED(CONFIG_IPV6)
1437 1438 1439 1440 1441 1442 1443 1444 1445 1446 1447 1448 1449 1450 1451 1452 1453 1454 1455 1456 1457 1458 1459 1460 1461 1462 1463 1464 1465 1466 1467 1468 1469 1470 1471 1472 1473 1474 1475 1476 1477 1478 1479 1480 1481 1482 1483 1484 1485 1486 1487 1488 1489 1490 1491 1492 1493 1494 1495 1496 1497 1498 1499 1500 1501 1502 1503 1504 1505 1506 1507 1508 1509 1510 1511 1512 1513 1514 1515 1516 1517 1518 1519 1520 1521 1522 1523 1524
static struct sk_buff *vxlan_na_create(struct sk_buff *request,
	struct neighbour *n, bool isrouter)
{
	struct net_device *dev = request->dev;
	struct sk_buff *reply;
	struct nd_msg *ns, *na;
	struct ipv6hdr *pip6;
	u8 *daddr;
	int na_olen = 8; /* opt hdr + ETH_ALEN for target */
	int ns_olen;
	int i, len;

	if (dev == NULL)
		return NULL;

	len = LL_RESERVED_SPACE(dev) + sizeof(struct ipv6hdr) +
		sizeof(*na) + na_olen + dev->needed_tailroom;
	reply = alloc_skb(len, GFP_ATOMIC);
	if (reply == NULL)
		return NULL;

	reply->protocol = htons(ETH_P_IPV6);
	reply->dev = dev;
	skb_reserve(reply, LL_RESERVED_SPACE(request->dev));
	skb_push(reply, sizeof(struct ethhdr));
	skb_set_mac_header(reply, 0);

	ns = (struct nd_msg *)skb_transport_header(request);

	daddr = eth_hdr(request)->h_source;
	ns_olen = request->len - skb_transport_offset(request) - sizeof(*ns);
	for (i = 0; i < ns_olen-1; i += (ns->opt[i+1]<<3)) {
		if (ns->opt[i] == ND_OPT_SOURCE_LL_ADDR) {
			daddr = ns->opt + i + sizeof(struct nd_opt_hdr);
			break;
		}
	}

	/* Ethernet header */
	ether_addr_copy(eth_hdr(reply)->h_dest, daddr);
	ether_addr_copy(eth_hdr(reply)->h_source, n->ha);
	eth_hdr(reply)->h_proto = htons(ETH_P_IPV6);
	reply->protocol = htons(ETH_P_IPV6);

	skb_pull(reply, sizeof(struct ethhdr));
	skb_set_network_header(reply, 0);
	skb_put(reply, sizeof(struct ipv6hdr));

	/* IPv6 header */

	pip6 = ipv6_hdr(reply);
	memset(pip6, 0, sizeof(struct ipv6hdr));
	pip6->version = 6;
	pip6->priority = ipv6_hdr(request)->priority;
	pip6->nexthdr = IPPROTO_ICMPV6;
	pip6->hop_limit = 255;
	pip6->daddr = ipv6_hdr(request)->saddr;
	pip6->saddr = *(struct in6_addr *)n->primary_key;

	skb_pull(reply, sizeof(struct ipv6hdr));
	skb_set_transport_header(reply, 0);

	na = (struct nd_msg *)skb_put(reply, sizeof(*na) + na_olen);

	/* Neighbor Advertisement */
	memset(na, 0, sizeof(*na)+na_olen);
	na->icmph.icmp6_type = NDISC_NEIGHBOUR_ADVERTISEMENT;
	na->icmph.icmp6_router = isrouter;
	na->icmph.icmp6_override = 1;
	na->icmph.icmp6_solicited = 1;
	na->target = ns->target;
	ether_addr_copy(&na->opt[2], n->ha);
	na->opt[0] = ND_OPT_TARGET_LL_ADDR;
	na->opt[1] = na_olen >> 3;

	na->icmph.icmp6_cksum = csum_ipv6_magic(&pip6->saddr,
		&pip6->daddr, sizeof(*na)+na_olen, IPPROTO_ICMPV6,
		csum_partial(na, sizeof(*na)+na_olen, 0));

	pip6->payload_len = htons(sizeof(*na)+na_olen);

	skb_push(reply, sizeof(struct ipv6hdr));

	reply->ip_summed = CHECKSUM_UNNECESSARY;

	return reply;
}

C
Cong Wang 已提交
1525 1526 1527
static int neigh_reduce(struct net_device *dev, struct sk_buff *skb)
{
	struct vxlan_dev *vxlan = netdev_priv(dev);
1528
	struct nd_msg *msg;
C
Cong Wang 已提交
1529 1530
	const struct ipv6hdr *iphdr;
	const struct in6_addr *saddr, *daddr;
1531 1532
	struct neighbour *n;
	struct inet6_dev *in6_dev;
C
Cong Wang 已提交
1533 1534 1535 1536 1537 1538 1539 1540 1541 1542 1543 1544 1545 1546

	in6_dev = __in6_dev_get(dev);
	if (!in6_dev)
		goto out;

	iphdr = ipv6_hdr(skb);
	saddr = &iphdr->saddr;
	daddr = &iphdr->daddr;

	msg = (struct nd_msg *)skb_transport_header(skb);
	if (msg->icmph.icmp6_code != 0 ||
	    msg->icmph.icmp6_type != NDISC_NEIGHBOUR_SOLICITATION)
		goto out;

1547 1548 1549 1550 1551
	if (ipv6_addr_loopback(daddr) ||
	    ipv6_addr_is_multicast(&msg->target))
		goto out;

	n = neigh_lookup(ipv6_stub->nd_tbl, &msg->target, dev);
C
Cong Wang 已提交
1552 1553 1554

	if (n) {
		struct vxlan_fdb *f;
1555
		struct sk_buff *reply;
C
Cong Wang 已提交
1556 1557 1558 1559 1560 1561 1562 1563 1564 1565 1566 1567 1568

		if (!(n->nud_state & NUD_CONNECTED)) {
			neigh_release(n);
			goto out;
		}

		f = vxlan_find_mac(vxlan, n->ha);
		if (f && vxlan_addr_any(&(first_remote_rcu(f)->remote_ip))) {
			/* bridge-local neighbor */
			neigh_release(n);
			goto out;
		}

1569 1570 1571
		reply = vxlan_na_create(skb, n,
					!!(f ? f->flags & NTF_ROUTER : 0));

C
Cong Wang 已提交
1572
		neigh_release(n);
1573 1574 1575 1576 1577 1578 1579

		if (reply == NULL)
			goto out;

		if (netif_rx_ni(reply) == NET_RX_DROP)
			dev->stats.rx_dropped++;

C
Cong Wang 已提交
1580
	} else if (vxlan->flags & VXLAN_F_L3MISS) {
1581 1582
		union vxlan_addr ipa = {
			.sin6.sin6_addr = msg->target,
1583
			.sin6.sin6_family = AF_INET6,
1584 1585
		};

C
Cong Wang 已提交
1586 1587 1588 1589 1590 1591 1592 1593 1594
		vxlan_ip_miss(dev, &ipa);
	}

out:
	consume_skb(skb);
	return NETDEV_TX_OK;
}
#endif

D
David Stevens 已提交
1595 1596 1597 1598 1599 1600 1601 1602 1603 1604 1605
static bool route_shortcircuit(struct net_device *dev, struct sk_buff *skb)
{
	struct vxlan_dev *vxlan = netdev_priv(dev);
	struct neighbour *n;

	if (is_multicast_ether_addr(eth_hdr(skb)->h_dest))
		return false;

	n = NULL;
	switch (ntohs(eth_hdr(skb)->h_proto)) {
	case ETH_P_IP:
1606 1607 1608
	{
		struct iphdr *pip;

D
David Stevens 已提交
1609 1610 1611 1612
		if (!pskb_may_pull(skb, sizeof(struct iphdr)))
			return false;
		pip = ip_hdr(skb);
		n = neigh_lookup(&arp_tbl, &pip->daddr, dev);
C
Cong Wang 已提交
1613 1614 1615
		if (!n && (vxlan->flags & VXLAN_F_L3MISS)) {
			union vxlan_addr ipa = {
				.sin.sin_addr.s_addr = pip->daddr,
1616
				.sin.sin_family = AF_INET,
C
Cong Wang 已提交
1617 1618 1619 1620 1621 1622
			};

			vxlan_ip_miss(dev, &ipa);
			return false;
		}

D
David Stevens 已提交
1623
		break;
1624 1625 1626 1627 1628 1629 1630 1631 1632 1633 1634 1635 1636
	}
#if IS_ENABLED(CONFIG_IPV6)
	case ETH_P_IPV6:
	{
		struct ipv6hdr *pip6;

		if (!pskb_may_pull(skb, sizeof(struct ipv6hdr)))
			return false;
		pip6 = ipv6_hdr(skb);
		n = neigh_lookup(ipv6_stub->nd_tbl, &pip6->daddr, dev);
		if (!n && (vxlan->flags & VXLAN_F_L3MISS)) {
			union vxlan_addr ipa = {
				.sin6.sin6_addr = pip6->daddr,
1637
				.sin6.sin6_family = AF_INET6,
1638 1639 1640 1641 1642 1643 1644 1645 1646
			};

			vxlan_ip_miss(dev, &ipa);
			return false;
		}

		break;
	}
#endif
D
David Stevens 已提交
1647 1648 1649 1650 1651 1652 1653
	default:
		return false;
	}

	if (n) {
		bool diff;

1654
		diff = !ether_addr_equal(eth_hdr(skb)->h_dest, n->ha);
D
David Stevens 已提交
1655 1656 1657 1658 1659 1660 1661
		if (diff) {
			memcpy(eth_hdr(skb)->h_source, eth_hdr(skb)->h_dest,
				dev->addr_len);
			memcpy(eth_hdr(skb)->h_dest, n->ha, dev->addr_len);
		}
		neigh_release(n);
		return diff;
C
Cong Wang 已提交
1662 1663
	}

D
David Stevens 已提交
1664 1665 1666
	return false;
}

1667
static void vxlan_build_gbp_hdr(struct vxlanhdr *vxh, u32 vxflags,
T
Thomas Graf 已提交
1668 1669 1670 1671
				struct vxlan_metadata *md)
{
	struct vxlanhdr_gbp *gbp;

1672 1673 1674
	if (!md->gbp)
		return;

T
Thomas Graf 已提交
1675 1676 1677 1678 1679 1680 1681 1682 1683 1684 1685 1686
	gbp = (struct vxlanhdr_gbp *)vxh;
	vxh->vx_flags |= htonl(VXLAN_HF_GBP);

	if (md->gbp & VXLAN_GBP_DONT_LEARN)
		gbp->dont_learn = 1;

	if (md->gbp & VXLAN_GBP_POLICY_APPLIED)
		gbp->policy_applied = 1;

	gbp->policy_id = htons(md->gbp & VXLAN_GBP_ID_MASK);
}

C
Cong Wang 已提交
1687
#if IS_ENABLED(CONFIG_IPV6)
1688 1689
static int vxlan6_xmit_skb(struct dst_entry *dst, struct sock *sk,
			   struct sk_buff *skb,
C
Cong Wang 已提交
1690 1691
			   struct net_device *dev, struct in6_addr *saddr,
			   struct in6_addr *daddr, __u8 prio, __u8 ttl,
1692
			   __be16 src_port, __be16 dst_port, __be32 vni,
1693
			   struct vxlan_metadata *md, bool xnet, u32 vxflags)
C
Cong Wang 已提交
1694 1695 1696 1697
{
	struct vxlanhdr *vxh;
	int min_headroom;
	int err;
1698
	bool udp_sum = !(vxflags & VXLAN_F_UDP_ZERO_CSUM6_TX);
T
Tom Herbert 已提交
1699 1700 1701
	int type = udp_sum ? SKB_GSO_UDP_TUNNEL_CSUM : SKB_GSO_UDP_TUNNEL;
	u16 hdrlen = sizeof(struct vxlanhdr);

1702
	if ((vxflags & VXLAN_F_REMCSUM_TX) &&
T
Tom Herbert 已提交
1703 1704 1705 1706 1707 1708 1709 1710 1711 1712 1713
	    skb->ip_summed == CHECKSUM_PARTIAL) {
		int csum_start = skb_checksum_start_offset(skb);

		if (csum_start <= VXLAN_MAX_REMCSUM_START &&
		    !(csum_start & VXLAN_RCO_SHIFT_MASK) &&
		    (skb->csum_offset == offsetof(struct udphdr, check) ||
		     skb->csum_offset == offsetof(struct tcphdr, check))) {
			udp_sum = false;
			type |= SKB_GSO_TUNNEL_REMCSUM;
		}
	}
C
Cong Wang 已提交
1714

N
Nicolas Dichtel 已提交
1715
	skb_scrub_packet(skb, xnet);
1716

C
Cong Wang 已提交
1717 1718
	min_headroom = LL_RESERVED_SPACE(dst->dev) + dst->header_len
			+ VXLAN_HLEN + sizeof(struct ipv6hdr)
1719
			+ (skb_vlan_tag_present(skb) ? VLAN_HLEN : 0);
C
Cong Wang 已提交
1720 1721 1722

	/* Need space for new headers (invalidates iph ptr) */
	err = skb_cow_head(skb, min_headroom);
P
Pravin B Shelar 已提交
1723 1724 1725 1726
	if (unlikely(err)) {
		kfree_skb(skb);
		goto err;
	}
C
Cong Wang 已提交
1727

1728
	skb = vlan_hwaccel_push_inside(skb);
P
Pravin B Shelar 已提交
1729 1730 1731 1732
	if (WARN_ON(!skb)) {
		err = -ENOMEM;
		goto err;
	}
C
Cong Wang 已提交
1733

1734 1735 1736 1737 1738 1739
	skb = iptunnel_handle_offloads(skb, udp_sum, type);
	if (IS_ERR(skb)) {
		err = -EINVAL;
		goto err;
	}

C
Cong Wang 已提交
1740
	vxh = (struct vxlanhdr *) __skb_push(skb, sizeof(*vxh));
1741
	vxh->vx_flags = htonl(VXLAN_HF_VNI);
1742
	vxh->vx_vni = vni;
C
Cong Wang 已提交
1743

T
Tom Herbert 已提交
1744 1745 1746 1747 1748 1749 1750 1751 1752 1753 1754 1755 1756 1757 1758 1759
	if (type & SKB_GSO_TUNNEL_REMCSUM) {
		u32 data = (skb_checksum_start_offset(skb) - hdrlen) >>
			   VXLAN_RCO_SHIFT;

		if (skb->csum_offset == offsetof(struct udphdr, check))
			data |= VXLAN_RCO_UDP;

		vxh->vx_vni |= htonl(data);
		vxh->vx_flags |= htonl(VXLAN_HF_RCO);

		if (!skb_is_gso(skb)) {
			skb->ip_summed = CHECKSUM_NONE;
			skb->encapsulation = 0;
		}
	}

1760 1761
	if (vxflags & VXLAN_F_GBP)
		vxlan_build_gbp_hdr(vxh, vxflags, md);
T
Thomas Graf 已提交
1762

1763 1764
	skb_set_inner_protocol(skb, htons(ETH_P_TEB));

1765
	udp_tunnel6_xmit_skb(dst, sk, skb, dev, saddr, daddr, prio,
1766
			     ttl, src_port, dst_port,
1767
			     !!(vxflags & VXLAN_F_UDP_ZERO_CSUM6_TX));
C
Cong Wang 已提交
1768
	return 0;
P
Pravin B Shelar 已提交
1769 1770 1771
err:
	dst_release(dst);
	return err;
C
Cong Wang 已提交
1772 1773 1774
}
#endif

1775 1776
static int vxlan_xmit_skb(struct rtable *rt, struct sock *sk, struct sk_buff *skb,
			  __be32 src, __be32 dst, __u8 tos, __u8 ttl, __be16 df,
1777
			  __be16 src_port, __be16 dst_port, __be32 vni,
1778
			  struct vxlan_metadata *md, bool xnet, u32 vxflags)
1779 1780
{
	struct vxlanhdr *vxh;
1781
	int min_headroom;
1782
	int err;
1783
	bool udp_sum = !!(vxflags & VXLAN_F_UDP_CSUM);
T
Tom Herbert 已提交
1784 1785 1786
	int type = udp_sum ? SKB_GSO_UDP_TUNNEL_CSUM : SKB_GSO_UDP_TUNNEL;
	u16 hdrlen = sizeof(struct vxlanhdr);

1787
	if ((vxflags & VXLAN_F_REMCSUM_TX) &&
T
Tom Herbert 已提交
1788 1789 1790 1791 1792 1793 1794 1795 1796 1797 1798
	    skb->ip_summed == CHECKSUM_PARTIAL) {
		int csum_start = skb_checksum_start_offset(skb);

		if (csum_start <= VXLAN_MAX_REMCSUM_START &&
		    !(csum_start & VXLAN_RCO_SHIFT_MASK) &&
		    (skb->csum_offset == offsetof(struct udphdr, check) ||
		     skb->csum_offset == offsetof(struct tcphdr, check))) {
			udp_sum = false;
			type |= SKB_GSO_TUNNEL_REMCSUM;
		}
	}
1799

1800
	min_headroom = LL_RESERVED_SPACE(rt->dst.dev) + rt->dst.header_len
1801
			+ VXLAN_HLEN + sizeof(struct iphdr)
1802
			+ (skb_vlan_tag_present(skb) ? VLAN_HLEN : 0);
1803 1804 1805

	/* Need space for new headers (invalidates iph ptr) */
	err = skb_cow_head(skb, min_headroom);
P
Pravin B Shelar 已提交
1806 1807
	if (unlikely(err)) {
		kfree_skb(skb);
1808
		return err;
P
Pravin B Shelar 已提交
1809
	}
1810

1811 1812 1813
	skb = vlan_hwaccel_push_inside(skb);
	if (WARN_ON(!skb))
		return -ENOMEM;
1814

1815 1816 1817 1818
	skb = iptunnel_handle_offloads(skb, udp_sum, type);
	if (IS_ERR(skb))
		return PTR_ERR(skb);

1819
	vxh = (struct vxlanhdr *) __skb_push(skb, sizeof(*vxh));
1820
	vxh->vx_flags = htonl(VXLAN_HF_VNI);
1821
	vxh->vx_vni = vni;
1822

T
Tom Herbert 已提交
1823 1824 1825 1826 1827 1828 1829 1830 1831 1832 1833 1834 1835 1836 1837 1838
	if (type & SKB_GSO_TUNNEL_REMCSUM) {
		u32 data = (skb_checksum_start_offset(skb) - hdrlen) >>
			   VXLAN_RCO_SHIFT;

		if (skb->csum_offset == offsetof(struct udphdr, check))
			data |= VXLAN_RCO_UDP;

		vxh->vx_vni |= htonl(data);
		vxh->vx_flags |= htonl(VXLAN_HF_RCO);

		if (!skb_is_gso(skb)) {
			skb->ip_summed = CHECKSUM_NONE;
			skb->encapsulation = 0;
		}
	}

1839 1840
	if (vxflags & VXLAN_F_GBP)
		vxlan_build_gbp_hdr(vxh, vxflags, md);
T
Thomas Graf 已提交
1841

1842 1843
	skb_set_inner_protocol(skb, htons(ETH_P_TEB));

1844 1845 1846 1847
	udp_tunnel_xmit_skb(rt, sk, skb, src, dst, tos, ttl, df,
			    src_port, dst_port, xnet,
			    !(vxflags & VXLAN_F_UDP_CSUM));
	return 0;
1848 1849
}

1850 1851 1852 1853 1854 1855 1856 1857 1858 1859 1860 1861 1862 1863 1864 1865 1866 1867 1868 1869 1870 1871 1872 1873 1874 1875 1876 1877
#if IS_ENABLED(CONFIG_IPV6)
static struct dst_entry *vxlan6_get_route(struct vxlan_dev *vxlan,
					  struct sk_buff *skb, int oif,
					  const struct in6_addr *daddr,
					  struct in6_addr *saddr)
{
	struct dst_entry *ndst;
	struct flowi6 fl6;
	int err;

	memset(&fl6, 0, sizeof(fl6));
	fl6.flowi6_oif = oif;
	fl6.daddr = *daddr;
	fl6.saddr = vxlan->cfg.saddr.sin6.sin6_addr;
	fl6.flowi6_mark = skb->mark;
	fl6.flowi6_proto = IPPROTO_UDP;

	err = ipv6_stub->ipv6_dst_lookup(vxlan->net,
					 vxlan->vn6_sock->sock->sk,
					 &ndst, &fl6);
	if (err < 0)
		return ERR_PTR(err);

	*saddr = fl6.saddr;
	return ndst;
}
#endif

1878 1879 1880 1881
/* Bypass encapsulation if the destination is local */
static void vxlan_encap_bypass(struct sk_buff *skb, struct vxlan_dev *src_vxlan,
			       struct vxlan_dev *dst_vxlan)
{
1882
	struct pcpu_sw_netstats *tx_stats, *rx_stats;
C
Cong Wang 已提交
1883 1884
	union vxlan_addr loopback;
	union vxlan_addr *remote_ip = &dst_vxlan->default_dst.remote_ip;
1885 1886
	struct net_device *dev = skb->dev;
	int len = skb->len;
1887

1888 1889
	tx_stats = this_cpu_ptr(src_vxlan->dev->tstats);
	rx_stats = this_cpu_ptr(dst_vxlan->dev->tstats);
1890 1891 1892 1893 1894
	skb->pkt_type = PACKET_HOST;
	skb->encapsulation = 0;
	skb->dev = dst_vxlan->dev;
	__skb_pull(skb, skb_network_offset(skb));

C
Cong Wang 已提交
1895 1896 1897 1898 1899 1900 1901 1902 1903 1904
	if (remote_ip->sa.sa_family == AF_INET) {
		loopback.sin.sin_addr.s_addr = htonl(INADDR_LOOPBACK);
		loopback.sa.sa_family =  AF_INET;
#if IS_ENABLED(CONFIG_IPV6)
	} else {
		loopback.sin6.sin6_addr = in6addr_loopback;
		loopback.sa.sa_family =  AF_INET6;
#endif
	}

1905
	if (dst_vxlan->flags & VXLAN_F_LEARN)
C
Cong Wang 已提交
1906
		vxlan_snoop(skb->dev, &loopback, eth_hdr(skb)->h_source);
1907 1908 1909

	u64_stats_update_begin(&tx_stats->syncp);
	tx_stats->tx_packets++;
1910
	tx_stats->tx_bytes += len;
1911 1912 1913 1914 1915
	u64_stats_update_end(&tx_stats->syncp);

	if (netif_rx(skb) == NET_RX_SUCCESS) {
		u64_stats_update_begin(&rx_stats->syncp);
		rx_stats->rx_packets++;
1916
		rx_stats->rx_bytes += len;
1917 1918
		u64_stats_update_end(&rx_stats->syncp);
	} else {
1919
		dev->stats.rx_dropped++;
1920 1921 1922
	}
}

1923 1924
static void vxlan_xmit_one(struct sk_buff *skb, struct net_device *dev,
			   struct vxlan_rdst *rdst, bool did_rsc)
S
stephen hemminger 已提交
1925
{
1926
	struct ip_tunnel_info *info;
S
stephen hemminger 已提交
1927
	struct vxlan_dev *vxlan = netdev_priv(dev);
1928
	struct sock *sk;
C
Cong Wang 已提交
1929
	struct rtable *rt = NULL;
S
stephen hemminger 已提交
1930 1931
	const struct iphdr *old_iph;
	struct flowi4 fl4;
C
Cong Wang 已提交
1932
	union vxlan_addr *dst;
T
Thomas Graf 已提交
1933 1934 1935
	union vxlan_addr remote_ip;
	struct vxlan_metadata _md;
	struct vxlan_metadata *md = &_md;
C
Cong Wang 已提交
1936
	__be16 src_port = 0, dst_port;
S
Stephen Hemminger 已提交
1937
	u32 vni;
S
stephen hemminger 已提交
1938 1939
	__be16 df = 0;
	__u8 tos, ttl;
1940
	int err;
T
Thomas Graf 已提交
1941
	u32 flags = vxlan->flags;
S
stephen hemminger 已提交
1942

1943
	info = skb_tunnel_info(skb);
1944

T
Thomas Graf 已提交
1945
	if (rdst) {
1946
		dst_port = rdst->remote_port ? rdst->remote_port : vxlan->cfg.dst_port;
T
Thomas Graf 已提交
1947 1948 1949 1950 1951 1952 1953 1954
		vni = rdst->remote_vni;
		dst = &rdst->remote_ip;
	} else {
		if (!info) {
			WARN_ONCE(1, "%s: Missing encapsulation instructions\n",
				  dev->name);
			goto drop;
		}
1955
		dst_port = info->key.tp_dst ? : vxlan->cfg.dst_port;
T
Thomas Graf 已提交
1956
		vni = be64_to_cpu(info->key.tun_id);
1957 1958
		remote_ip.sa.sa_family = ip_tunnel_info_af(info);
		if (remote_ip.sa.sa_family == AF_INET)
1959 1960 1961
			remote_ip.sin.sin_addr.s_addr = info->key.u.ipv4.dst;
		else
			remote_ip.sin6.sin6_addr = info->key.u.ipv6.dst;
T
Thomas Graf 已提交
1962 1963
		dst = &remote_ip;
	}
D
David Stevens 已提交
1964

C
Cong Wang 已提交
1965
	if (vxlan_addr_any(dst)) {
D
David Stevens 已提交
1966 1967
		if (did_rsc) {
			/* short-circuited back to local bridge */
1968
			vxlan_encap_bypass(skb, vxlan, vxlan);
1969
			return;
D
David Stevens 已提交
1970
		}
1971
		goto drop;
D
David Stevens 已提交
1972
	}
1973

S
stephen hemminger 已提交
1974 1975
	old_iph = ip_hdr(skb);

1976
	ttl = vxlan->cfg.ttl;
C
Cong Wang 已提交
1977
	if (!ttl && vxlan_addr_multicast(dst))
S
stephen hemminger 已提交
1978 1979
		ttl = 1;

1980
	tos = vxlan->cfg.tos;
S
stephen hemminger 已提交
1981
	if (tos == 1)
1982
		tos = ip_tunnel_get_dsfield(old_iph, skb);
S
stephen hemminger 已提交
1983

1984 1985
	src_port = udp_flow_src_port(dev_net(dev), skb, vxlan->cfg.port_min,
				     vxlan->cfg.port_max, true);
S
stephen hemminger 已提交
1986

1987 1988 1989 1990 1991
	if (info) {
		ttl = info->key.ttl;
		tos = info->key.tos;

		if (info->options_len)
1992
			md = ip_tunnel_info_opts(info);
1993 1994 1995 1996
	} else {
		md->gbp = skb->mark;
	}

C
Cong Wang 已提交
1997
	if (dst->sa.sa_family == AF_INET) {
1998 1999 2000 2001
		if (!vxlan->vn4_sock)
			goto drop;
		sk = vxlan->vn4_sock->sock->sk;

2002 2003 2004 2005 2006 2007 2008 2009 2010
		if (info) {
			if (info->key.tun_flags & TUNNEL_DONT_FRAGMENT)
				df = htons(IP_DF);

			if (info->key.tun_flags & TUNNEL_CSUM)
				flags |= VXLAN_F_UDP_CSUM;
			else
				flags &= ~VXLAN_F_UDP_CSUM;
		}
T
Thomas Graf 已提交
2011

C
Cong Wang 已提交
2012
		memset(&fl4, 0, sizeof(fl4));
T
Thomas Graf 已提交
2013
		fl4.flowi4_oif = rdst ? rdst->remote_ifindex : 0;
C
Cong Wang 已提交
2014
		fl4.flowi4_tos = RT_TOS(tos);
2015 2016
		fl4.flowi4_mark = skb->mark;
		fl4.flowi4_proto = IPPROTO_UDP;
C
Cong Wang 已提交
2017
		fl4.daddr = dst->sin.sin_addr.s_addr;
2018
		fl4.saddr = vxlan->cfg.saddr.sin.sin_addr.s_addr;
C
Cong Wang 已提交
2019

N
Nicolas Dichtel 已提交
2020
		rt = ip_route_output_key(vxlan->net, &fl4);
C
Cong Wang 已提交
2021 2022 2023 2024 2025 2026
		if (IS_ERR(rt)) {
			netdev_dbg(dev, "no route to %pI4\n",
				   &dst->sin.sin_addr.s_addr);
			dev->stats.tx_carrier_errors++;
			goto tx_error;
		}
S
stephen hemminger 已提交
2027

C
Cong Wang 已提交
2028 2029 2030 2031
		if (rt->dst.dev == dev) {
			netdev_dbg(dev, "circular route to %pI4\n",
				   &dst->sin.sin_addr.s_addr);
			dev->stats.collisions++;
2032
			goto rt_tx_error;
C
Cong Wang 已提交
2033 2034 2035 2036 2037 2038 2039 2040
		}

		/* Bypass encapsulation if the destination is local */
		if (rt->rt_flags & RTCF_LOCAL &&
		    !(rt->rt_flags & (RTCF_BROADCAST | RTCF_MULTICAST))) {
			struct vxlan_dev *dst_vxlan;

			ip_rt_put(rt);
2041
			dst_vxlan = vxlan_find_vni(vxlan->net, vni,
2042 2043
						   dst->sa.sa_family, dst_port,
						   vxlan->flags);
C
Cong Wang 已提交
2044 2045 2046 2047 2048 2049 2050 2051
			if (!dst_vxlan)
				goto tx_error;
			vxlan_encap_bypass(skb, vxlan, dst_vxlan);
			return;
		}

		tos = ip_tunnel_ecn_encap(tos, old_iph, skb);
		ttl = ttl ? : ip4_dst_hoplimit(&rt->dst);
2052
		err = vxlan_xmit_skb(rt, sk, skb, fl4.saddr,
2053
				     dst->sin.sin_addr.s_addr, tos, ttl, df,
2054
				     src_port, dst_port, htonl(vni << 8), md,
2055
				     !net_eq(vxlan->net, dev_net(vxlan->dev)),
T
Thomas Graf 已提交
2056
				     flags);
P
Pravin B Shelar 已提交
2057 2058 2059
		if (err < 0) {
			/* skb is already freed. */
			skb = NULL;
C
Cong Wang 已提交
2060
			goto rt_tx_error;
P
Pravin B Shelar 已提交
2061
		}
C
Cong Wang 已提交
2062 2063 2064
#if IS_ENABLED(CONFIG_IPV6)
	} else {
		struct dst_entry *ndst;
2065
		struct in6_addr saddr;
J
Jiri Benc 已提交
2066
		u32 rt6i_flags;
C
Cong Wang 已提交
2067

2068 2069 2070 2071
		if (!vxlan->vn6_sock)
			goto drop;
		sk = vxlan->vn6_sock->sock->sk;

2072 2073 2074 2075
		ndst = vxlan6_get_route(vxlan, skb,
					rdst ? rdst->remote_ifindex : 0,
					&dst->sin6.sin6_addr, &saddr);
		if (IS_ERR(ndst)) {
C
Cong Wang 已提交
2076 2077 2078
			netdev_dbg(dev, "no route to %pI6\n",
				   &dst->sin6.sin6_addr);
			dev->stats.tx_carrier_errors++;
2079
			goto tx_error;
C
Cong Wang 已提交
2080
		}
S
stephen hemminger 已提交
2081

C
Cong Wang 已提交
2082 2083 2084 2085 2086 2087 2088
		if (ndst->dev == dev) {
			netdev_dbg(dev, "circular route to %pI6\n",
				   &dst->sin6.sin6_addr);
			dst_release(ndst);
			dev->stats.collisions++;
			goto tx_error;
		}
2089

C
Cong Wang 已提交
2090
		/* Bypass encapsulation if the destination is local */
J
Jiri Benc 已提交
2091 2092 2093
		rt6i_flags = ((struct rt6_info *)ndst)->rt6i_flags;
		if (rt6i_flags & RTF_LOCAL &&
		    !(rt6i_flags & (RTCF_BROADCAST | RTCF_MULTICAST))) {
C
Cong Wang 已提交
2094 2095 2096
			struct vxlan_dev *dst_vxlan;

			dst_release(ndst);
2097
			dst_vxlan = vxlan_find_vni(vxlan->net, vni,
2098 2099
						   dst->sa.sa_family, dst_port,
						   vxlan->flags);
C
Cong Wang 已提交
2100 2101 2102 2103 2104
			if (!dst_vxlan)
				goto tx_error;
			vxlan_encap_bypass(skb, vxlan, dst_vxlan);
			return;
		}
2105

2106 2107 2108 2109 2110 2111 2112
		if (info) {
			if (info->key.tun_flags & TUNNEL_CSUM)
				flags &= ~VXLAN_F_UDP_ZERO_CSUM6_TX;
			else
				flags |= VXLAN_F_UDP_ZERO_CSUM6_TX;
		}

C
Cong Wang 已提交
2113
		ttl = ttl ? : ip6_dst_hoplimit(ndst);
2114
		err = vxlan6_xmit_skb(ndst, sk, skb, dev, &saddr, &dst->sin6.sin6_addr,
2115
				      0, ttl, src_port, dst_port, htonl(vni << 8), md,
2116
				      !net_eq(vxlan->net, dev_net(vxlan->dev)),
2117
				      flags);
C
Cong Wang 已提交
2118 2119
#endif
	}
2120

2121
	return;
S
stephen hemminger 已提交
2122 2123 2124 2125 2126

drop:
	dev->stats.tx_dropped++;
	goto tx_free;

2127 2128
rt_tx_error:
	ip_rt_put(rt);
S
stephen hemminger 已提交
2129 2130 2131 2132 2133 2134
tx_error:
	dev->stats.tx_errors++;
tx_free:
	dev_kfree_skb(skb);
}

2135 2136 2137 2138 2139 2140 2141 2142 2143
/* Transmit local packets over Vxlan
 *
 * Outer IP header inherits ECN and DF from inner header.
 * Outer UDP destination is the VXLAN assigned port.
 *           source port is based on hash of flow
 */
static netdev_tx_t vxlan_xmit(struct sk_buff *skb, struct net_device *dev)
{
	struct vxlan_dev *vxlan = netdev_priv(dev);
2144
	const struct ip_tunnel_info *info;
2145 2146
	struct ethhdr *eth;
	bool did_rsc = false;
2147
	struct vxlan_rdst *rdst, *fdst = NULL;
2148 2149
	struct vxlan_fdb *f;

2150
	info = skb_tunnel_info(skb);
2151

2152 2153 2154
	skb_reset_mac_header(skb);
	eth = eth_hdr(skb);

C
Cong Wang 已提交
2155 2156 2157 2158 2159
	if ((vxlan->flags & VXLAN_F_PROXY)) {
		if (ntohs(eth->h_proto) == ETH_P_ARP)
			return arp_reduce(dev, skb);
#if IS_ENABLED(CONFIG_IPV6)
		else if (ntohs(eth->h_proto) == ETH_P_IPV6 &&
2160 2161
			 pskb_may_pull(skb, sizeof(struct ipv6hdr)
				       + sizeof(struct nd_msg)) &&
C
Cong Wang 已提交
2162 2163 2164 2165 2166 2167 2168 2169
			 ipv6_hdr(skb)->nexthdr == IPPROTO_ICMPV6) {
				struct nd_msg *msg;

				msg = (struct nd_msg *)skb_transport_header(skb);
				if (msg->icmph.icmp6_code == 0 &&
				    msg->icmph.icmp6_type == NDISC_NEIGHBOUR_SOLICITATION)
					return neigh_reduce(dev, skb);
		}
L
Li RongQing 已提交
2170
		eth = eth_hdr(skb);
C
Cong Wang 已提交
2171 2172
#endif
	}
2173

2174
	if (vxlan->flags & VXLAN_F_COLLECT_METADATA &&
2175
	    info && info->mode & IP_TUNNEL_INFO_TX) {
T
Thomas Graf 已提交
2176 2177 2178 2179
		vxlan_xmit_one(skb, dev, NULL, false);
		return NETDEV_TX_OK;
	}

2180
	f = vxlan_find_mac(vxlan, eth->h_dest);
2181 2182 2183
	did_rsc = false;

	if (f && (f->flags & NTF_ROUTER) && (vxlan->flags & VXLAN_F_RSC) &&
2184 2185
	    (ntohs(eth->h_proto) == ETH_P_IP ||
	     ntohs(eth->h_proto) == ETH_P_IPV6)) {
2186 2187 2188 2189 2190
		did_rsc = route_shortcircuit(dev, skb);
		if (did_rsc)
			f = vxlan_find_mac(vxlan, eth->h_dest);
	}

2191
	if (f == NULL) {
2192 2193 2194 2195 2196 2197 2198
		f = vxlan_find_mac(vxlan, all_zeros_mac);
		if (f == NULL) {
			if ((vxlan->flags & VXLAN_F_L2MISS) &&
			    !is_multicast_ether_addr(eth->h_dest))
				vxlan_fdb_miss(vxlan, eth->h_dest);

			dev->stats.tx_dropped++;
2199
			kfree_skb(skb);
2200 2201 2202
			return NETDEV_TX_OK;
		}
	}
2203

2204 2205
	list_for_each_entry_rcu(rdst, &f->remotes, list) {
		struct sk_buff *skb1;
2206

2207 2208 2209 2210
		if (!fdst) {
			fdst = rdst;
			continue;
		}
2211 2212 2213
		skb1 = skb_clone(skb, GFP_ATOMIC);
		if (skb1)
			vxlan_xmit_one(skb1, dev, rdst, did_rsc);
2214 2215
	}

2216 2217 2218 2219
	if (fdst)
		vxlan_xmit_one(skb, dev, fdst, did_rsc);
	else
		kfree_skb(skb);
2220
	return NETDEV_TX_OK;
2221 2222
}

S
stephen hemminger 已提交
2223 2224 2225 2226 2227 2228 2229 2230 2231 2232 2233 2234
/* Walk the forwarding table and purge stale entries */
static void vxlan_cleanup(unsigned long arg)
{
	struct vxlan_dev *vxlan = (struct vxlan_dev *) arg;
	unsigned long next_timer = jiffies + FDB_AGE_INTERVAL;
	unsigned int h;

	if (!netif_running(vxlan->dev))
		return;

	for (h = 0; h < FDB_HASH_SIZE; ++h) {
		struct hlist_node *p, *n;
2235 2236

		spin_lock_bh(&vxlan->hash_lock);
S
stephen hemminger 已提交
2237 2238 2239 2240 2241
		hlist_for_each_safe(p, n, &vxlan->fdb_head[h]) {
			struct vxlan_fdb *f
				= container_of(p, struct vxlan_fdb, hlist);
			unsigned long timeout;

2242
			if (f->state & NUD_PERMANENT)
S
stephen hemminger 已提交
2243 2244
				continue;

2245
			timeout = f->used + vxlan->cfg.age_interval * HZ;
S
stephen hemminger 已提交
2246 2247 2248 2249 2250 2251 2252 2253 2254
			if (time_before_eq(timeout, jiffies)) {
				netdev_dbg(vxlan->dev,
					   "garbage collect %pM\n",
					   f->eth_addr);
				f->state = NUD_STALE;
				vxlan_fdb_destroy(vxlan, f);
			} else if (time_before(timeout, next_timer))
				next_timer = timeout;
		}
2255
		spin_unlock_bh(&vxlan->hash_lock);
S
stephen hemminger 已提交
2256 2257 2258 2259 2260
	}

	mod_timer(&vxlan->age_timer, next_timer);
}

2261 2262
static void vxlan_vs_add_dev(struct vxlan_sock *vs, struct vxlan_dev *vxlan)
{
2263
	struct vxlan_net *vn = net_generic(vxlan->net, vxlan_net_id);
2264 2265
	__u32 vni = vxlan->default_dst.remote_vni;

2266
	spin_lock(&vn->sock_lock);
2267
	hlist_add_head_rcu(&vxlan->hlist, vni_head(vs, vni));
2268
	spin_unlock(&vn->sock_lock);
2269 2270
}

S
stephen hemminger 已提交
2271 2272 2273
/* Setup stats when device is created */
static int vxlan_init(struct net_device *dev)
{
2274
	dev->tstats = netdev_alloc_pcpu_stats(struct pcpu_sw_netstats);
2275
	if (!dev->tstats)
S
stephen hemminger 已提交
2276 2277 2278 2279 2280
		return -ENOMEM;

	return 0;
}

2281
static void vxlan_fdb_delete_default(struct vxlan_dev *vxlan)
2282 2283 2284 2285 2286 2287 2288 2289 2290 2291
{
	struct vxlan_fdb *f;

	spin_lock_bh(&vxlan->hash_lock);
	f = __vxlan_find_mac(vxlan, all_zeros_mac);
	if (f)
		vxlan_fdb_destroy(vxlan, f);
	spin_unlock_bh(&vxlan->hash_lock);
}

2292 2293 2294 2295
static void vxlan_uninit(struct net_device *dev)
{
	struct vxlan_dev *vxlan = netdev_priv(dev);

2296
	vxlan_fdb_delete_default(vxlan);
2297

2298 2299 2300
	free_percpu(dev->tstats);
}

S
stephen hemminger 已提交
2301 2302 2303 2304
/* Start ageing timer and join group when device is brought up */
static int vxlan_open(struct net_device *dev)
{
	struct vxlan_dev *vxlan = netdev_priv(dev);
2305
	int ret;
2306

2307 2308 2309
	ret = vxlan_sock_add(vxlan);
	if (ret < 0)
		return ret;
S
stephen hemminger 已提交
2310

2311
	if (vxlan_addr_multicast(&vxlan->default_dst.remote_ip)) {
2312
		ret = vxlan_igmp_join(vxlan);
2313 2314
		if (ret == -EADDRINUSE)
			ret = 0;
2315
		if (ret) {
2316
			vxlan_sock_release(vxlan);
2317 2318
			return ret;
		}
S
stephen hemminger 已提交
2319 2320
	}

2321
	if (vxlan->cfg.age_interval)
S
stephen hemminger 已提交
2322 2323
		mod_timer(&vxlan->age_timer, jiffies + FDB_AGE_INTERVAL);

2324
	return ret;
S
stephen hemminger 已提交
2325 2326 2327 2328 2329
}

/* Purge the forwarding table */
static void vxlan_flush(struct vxlan_dev *vxlan)
{
2330
	unsigned int h;
S
stephen hemminger 已提交
2331 2332 2333 2334 2335 2336 2337

	spin_lock_bh(&vxlan->hash_lock);
	for (h = 0; h < FDB_HASH_SIZE; ++h) {
		struct hlist_node *p, *n;
		hlist_for_each_safe(p, n, &vxlan->fdb_head[h]) {
			struct vxlan_fdb *f
				= container_of(p, struct vxlan_fdb, hlist);
2338 2339 2340
			/* the all_zeros_mac entry is deleted at vxlan_uninit */
			if (!is_zero_ether_addr(f->eth_addr))
				vxlan_fdb_destroy(vxlan, f);
S
stephen hemminger 已提交
2341 2342 2343 2344 2345 2346 2347 2348 2349
		}
	}
	spin_unlock_bh(&vxlan->hash_lock);
}

/* Cleanup timer and forwarding table on shutdown */
static int vxlan_stop(struct net_device *dev)
{
	struct vxlan_dev *vxlan = netdev_priv(dev);
N
Nicolas Dichtel 已提交
2350
	struct vxlan_net *vn = net_generic(vxlan->net, vxlan_net_id);
2351
	int ret = 0;
S
stephen hemminger 已提交
2352

2353
	if (vxlan_addr_multicast(&vxlan->default_dst.remote_ip) &&
2354
	    !vxlan_group_used(vn, vxlan))
2355
		ret = vxlan_igmp_leave(vxlan);
S
stephen hemminger 已提交
2356 2357 2358 2359

	del_timer_sync(&vxlan->age_timer);

	vxlan_flush(vxlan);
2360
	vxlan_sock_release(vxlan);
S
stephen hemminger 已提交
2361

2362
	return ret;
S
stephen hemminger 已提交
2363 2364 2365 2366 2367 2368 2369
}

/* Stub, nothing needs to be done. */
static void vxlan_set_multicast_list(struct net_device *dev)
{
}

D
David Wragg 已提交
2370 2371 2372
static int __vxlan_change_mtu(struct net_device *dev,
			      struct net_device *lowerdev,
			      struct vxlan_rdst *dst, int new_mtu, bool strict)
2373
{
D
David Wragg 已提交
2374
	int max_mtu = IP_MAX_MTU;
2375

D
David Wragg 已提交
2376 2377
	if (lowerdev)
		max_mtu = lowerdev->mtu;
2378 2379

	if (dst->remote_ip.sa.sa_family == AF_INET6)
D
David Wragg 已提交
2380
		max_mtu -= VXLAN6_HEADROOM;
2381
	else
D
David Wragg 已提交
2382
		max_mtu -= VXLAN_HEADROOM;
2383

D
David Wragg 已提交
2384
	if (new_mtu < 68)
2385 2386
		return -EINVAL;

D
David Wragg 已提交
2387 2388 2389 2390 2391 2392 2393
	if (new_mtu > max_mtu) {
		if (strict)
			return -EINVAL;

		new_mtu = max_mtu;
	}

2394 2395 2396 2397
	dev->mtu = new_mtu;
	return 0;
}

D
David Wragg 已提交
2398 2399 2400 2401 2402 2403 2404 2405 2406
static int vxlan_change_mtu(struct net_device *dev, int new_mtu)
{
	struct vxlan_dev *vxlan = netdev_priv(dev);
	struct vxlan_rdst *dst = &vxlan->default_dst;
	struct net_device *lowerdev = __dev_get_by_index(vxlan->net,
							 dst->remote_ifindex);
	return __vxlan_change_mtu(dev, lowerdev, dst, new_mtu, true);
}

2407 2408 2409 2410 2411 2412 2413 2414 2415 2416 2417 2418 2419 2420 2421 2422 2423 2424 2425 2426 2427 2428 2429 2430 2431 2432 2433 2434 2435 2436 2437 2438 2439 2440 2441
static int egress_ipv4_tun_info(struct net_device *dev, struct sk_buff *skb,
				struct ip_tunnel_info *info,
				__be16 sport, __be16 dport)
{
	struct vxlan_dev *vxlan = netdev_priv(dev);
	struct rtable *rt;
	struct flowi4 fl4;

	memset(&fl4, 0, sizeof(fl4));
	fl4.flowi4_tos = RT_TOS(info->key.tos);
	fl4.flowi4_mark = skb->mark;
	fl4.flowi4_proto = IPPROTO_UDP;
	fl4.daddr = info->key.u.ipv4.dst;

	rt = ip_route_output_key(vxlan->net, &fl4);
	if (IS_ERR(rt))
		return PTR_ERR(rt);
	ip_rt_put(rt);

	info->key.u.ipv4.src = fl4.saddr;
	info->key.tp_src = sport;
	info->key.tp_dst = dport;
	return 0;
}

static int vxlan_fill_metadata_dst(struct net_device *dev, struct sk_buff *skb)
{
	struct vxlan_dev *vxlan = netdev_priv(dev);
	struct ip_tunnel_info *info = skb_tunnel_info(skb);
	__be16 sport, dport;

	sport = udp_flow_src_port(dev_net(dev), skb, vxlan->cfg.port_min,
				  vxlan->cfg.port_max, true);
	dport = info->key.tp_dst ? : vxlan->cfg.dst_port;

2442 2443 2444
	if (ip_tunnel_info_af(info) == AF_INET) {
		if (!vxlan->vn4_sock)
			return -EINVAL;
2445
		return egress_ipv4_tun_info(dev, skb, info, sport, dport);
2446 2447 2448 2449 2450 2451 2452 2453 2454 2455 2456 2457 2458 2459 2460 2461 2462 2463 2464 2465
	} else {
#if IS_ENABLED(CONFIG_IPV6)
		struct dst_entry *ndst;

		if (!vxlan->vn6_sock)
			return -EINVAL;
		ndst = vxlan6_get_route(vxlan, skb, 0,
					&info->key.u.ipv6.dst,
					&info->key.u.ipv6.src);
		if (IS_ERR(ndst))
			return PTR_ERR(ndst);
		dst_release(ndst);

		info->key.tp_src = sport;
		info->key.tp_dst = dport;
#else /* !CONFIG_IPV6 */
		return -EPFNOSUPPORT;
#endif
	}
	return 0;
2466 2467
}

S
stephen hemminger 已提交
2468 2469
static const struct net_device_ops vxlan_netdev_ops = {
	.ndo_init		= vxlan_init,
2470
	.ndo_uninit		= vxlan_uninit,
S
stephen hemminger 已提交
2471 2472 2473
	.ndo_open		= vxlan_open,
	.ndo_stop		= vxlan_stop,
	.ndo_start_xmit		= vxlan_xmit,
2474
	.ndo_get_stats64	= ip_tunnel_get_stats64,
S
stephen hemminger 已提交
2475
	.ndo_set_rx_mode	= vxlan_set_multicast_list,
2476
	.ndo_change_mtu		= vxlan_change_mtu,
S
stephen hemminger 已提交
2477 2478 2479 2480 2481
	.ndo_validate_addr	= eth_validate_addr,
	.ndo_set_mac_address	= eth_mac_addr,
	.ndo_fdb_add		= vxlan_fdb_add,
	.ndo_fdb_del		= vxlan_fdb_delete,
	.ndo_fdb_dump		= vxlan_fdb_dump,
2482
	.ndo_fill_metadata_dst	= vxlan_fill_metadata_dst,
S
stephen hemminger 已提交
2483 2484 2485 2486 2487 2488 2489
};

/* Info for udev, that this is a virtual tunnel endpoint */
static struct device_type vxlan_type = {
	.name = "vxlan",
};

2490
/* Calls the ndo_add_vxlan_port of the caller in order to
J
Joseph Gasparakis 已提交
2491 2492
 * supply the listening VXLAN udp ports. Callers are expected
 * to implement the ndo_add_vxlan_port.
2493 2494 2495 2496 2497 2498 2499
 */
void vxlan_get_rx_port(struct net_device *dev)
{
	struct vxlan_sock *vs;
	struct net *net = dev_net(dev);
	struct vxlan_net *vn = net_generic(net, vxlan_net_id);
	sa_family_t sa_family;
J
Joseph Gasparakis 已提交
2500 2501
	__be16 port;
	unsigned int i;
2502 2503 2504

	spin_lock(&vn->sock_lock);
	for (i = 0; i < PORT_HASH_SIZE; ++i) {
J
Joseph Gasparakis 已提交
2505 2506
		hlist_for_each_entry_rcu(vs, &vn->sock_list[i], hlist) {
			port = inet_sk(vs->sock->sk)->inet_sport;
2507
			sa_family = vxlan_get_sk_family(vs);
2508 2509 2510 2511 2512 2513 2514 2515
			dev->netdev_ops->ndo_add_vxlan_port(dev, sa_family,
							    port);
		}
	}
	spin_unlock(&vn->sock_lock);
}
EXPORT_SYMBOL_GPL(vxlan_get_rx_port);

S
stephen hemminger 已提交
2516 2517 2518 2519
/* Initialize the device structure. */
static void vxlan_setup(struct net_device *dev)
{
	struct vxlan_dev *vxlan = netdev_priv(dev);
2520
	unsigned int h;
S
stephen hemminger 已提交
2521 2522 2523 2524 2525

	eth_hw_addr_random(dev);
	ether_setup(dev);

	dev->netdev_ops = &vxlan_netdev_ops;
2526
	dev->destructor = free_netdev;
S
stephen hemminger 已提交
2527 2528 2529
	SET_NETDEV_DEVTYPE(dev, &vxlan_type);

	dev->features	|= NETIF_F_LLTX;
2530
	dev->features	|= NETIF_F_SG | NETIF_F_HW_CSUM;
2531
	dev->features   |= NETIF_F_RXCSUM;
2532
	dev->features   |= NETIF_F_GSO_SOFTWARE;
2533

2534 2535
	dev->vlan_features = dev->features;
	dev->features |= NETIF_F_HW_VLAN_CTAG_TX | NETIF_F_HW_VLAN_STAG_TX;
2536
	dev->hw_features |= NETIF_F_SG | NETIF_F_HW_CSUM | NETIF_F_RXCSUM;
2537
	dev->hw_features |= NETIF_F_GSO_SOFTWARE;
2538
	dev->hw_features |= NETIF_F_HW_VLAN_CTAG_TX | NETIF_F_HW_VLAN_STAG_TX;
2539
	netif_keep_dst(dev);
J
Jiri Benc 已提交
2540
	dev->priv_flags &= ~IFF_TX_SKB_SHARING;
2541
	dev->priv_flags |= IFF_LIVE_ADDR_CHANGE | IFF_NO_QUEUE;
S
stephen hemminger 已提交
2542

2543
	INIT_LIST_HEAD(&vxlan->next);
S
stephen hemminger 已提交
2544 2545 2546 2547 2548 2549
	spin_lock_init(&vxlan->hash_lock);

	init_timer_deferrable(&vxlan->age_timer);
	vxlan->age_timer.function = vxlan_cleanup;
	vxlan->age_timer.data = (unsigned long) vxlan;

2550
	vxlan->cfg.dst_port = htons(vxlan_port);
2551

S
stephen hemminger 已提交
2552 2553
	vxlan->dev = dev;

T
Tom Herbert 已提交
2554 2555
	gro_cells_init(&vxlan->gro_cells, dev);

S
stephen hemminger 已提交
2556 2557 2558 2559 2560 2561
	for (h = 0; h < FDB_HASH_SIZE; ++h)
		INIT_HLIST_HEAD(&vxlan->fdb_head[h]);
}

static const struct nla_policy vxlan_policy[IFLA_VXLAN_MAX + 1] = {
	[IFLA_VXLAN_ID]		= { .type = NLA_U32 },
2562
	[IFLA_VXLAN_GROUP]	= { .len = FIELD_SIZEOF(struct iphdr, daddr) },
C
Cong Wang 已提交
2563
	[IFLA_VXLAN_GROUP6]	= { .len = sizeof(struct in6_addr) },
S
stephen hemminger 已提交
2564 2565
	[IFLA_VXLAN_LINK]	= { .type = NLA_U32 },
	[IFLA_VXLAN_LOCAL]	= { .len = FIELD_SIZEOF(struct iphdr, saddr) },
C
Cong Wang 已提交
2566
	[IFLA_VXLAN_LOCAL6]	= { .len = sizeof(struct in6_addr) },
S
stephen hemminger 已提交
2567 2568 2569 2570 2571
	[IFLA_VXLAN_TOS]	= { .type = NLA_U8 },
	[IFLA_VXLAN_TTL]	= { .type = NLA_U8 },
	[IFLA_VXLAN_LEARNING]	= { .type = NLA_U8 },
	[IFLA_VXLAN_AGEING]	= { .type = NLA_U32 },
	[IFLA_VXLAN_LIMIT]	= { .type = NLA_U32 },
2572
	[IFLA_VXLAN_PORT_RANGE] = { .len  = sizeof(struct ifla_vxlan_port_range) },
D
David Stevens 已提交
2573 2574 2575 2576
	[IFLA_VXLAN_PROXY]	= { .type = NLA_U8 },
	[IFLA_VXLAN_RSC]	= { .type = NLA_U8 },
	[IFLA_VXLAN_L2MISS]	= { .type = NLA_U8 },
	[IFLA_VXLAN_L3MISS]	= { .type = NLA_U8 },
2577
	[IFLA_VXLAN_COLLECT_METADATA]	= { .type = NLA_U8 },
2578
	[IFLA_VXLAN_PORT]	= { .type = NLA_U16 },
2579 2580 2581
	[IFLA_VXLAN_UDP_CSUM]	= { .type = NLA_U8 },
	[IFLA_VXLAN_UDP_ZERO_CSUM6_TX]	= { .type = NLA_U8 },
	[IFLA_VXLAN_UDP_ZERO_CSUM6_RX]	= { .type = NLA_U8 },
T
Tom Herbert 已提交
2582 2583
	[IFLA_VXLAN_REMCSUM_TX]	= { .type = NLA_U8 },
	[IFLA_VXLAN_REMCSUM_RX]	= { .type = NLA_U8 },
T
Thomas Graf 已提交
2584
	[IFLA_VXLAN_GBP]	= { .type = NLA_FLAG, },
2585
	[IFLA_VXLAN_REMCSUM_NOPARTIAL]	= { .type = NLA_FLAG },
S
stephen hemminger 已提交
2586 2587 2588 2589 2590 2591 2592 2593 2594 2595 2596 2597 2598 2599 2600 2601 2602 2603 2604 2605 2606 2607 2608 2609 2610
};

static int vxlan_validate(struct nlattr *tb[], struct nlattr *data[])
{
	if (tb[IFLA_ADDRESS]) {
		if (nla_len(tb[IFLA_ADDRESS]) != ETH_ALEN) {
			pr_debug("invalid link address (not ethernet)\n");
			return -EINVAL;
		}

		if (!is_valid_ether_addr(nla_data(tb[IFLA_ADDRESS]))) {
			pr_debug("invalid all zero ethernet address\n");
			return -EADDRNOTAVAIL;
		}
	}

	if (!data)
		return -EINVAL;

	if (data[IFLA_VXLAN_ID]) {
		__u32 id = nla_get_u32(data[IFLA_VXLAN_ID]);
		if (id >= VXLAN_VID_MASK)
			return -ERANGE;
	}

2611 2612 2613 2614 2615 2616 2617 2618 2619 2620 2621
	if (data[IFLA_VXLAN_PORT_RANGE]) {
		const struct ifla_vxlan_port_range *p
			= nla_data(data[IFLA_VXLAN_PORT_RANGE]);

		if (ntohs(p->high) < ntohs(p->low)) {
			pr_debug("port range %u .. %u not valid\n",
				 ntohs(p->low), ntohs(p->high));
			return -EINVAL;
		}
	}

S
stephen hemminger 已提交
2622 2623 2624
	return 0;
}

Y
Yan Burman 已提交
2625 2626 2627 2628 2629 2630 2631 2632 2633 2634 2635 2636
static void vxlan_get_drvinfo(struct net_device *netdev,
			      struct ethtool_drvinfo *drvinfo)
{
	strlcpy(drvinfo->version, VXLAN_VERSION, sizeof(drvinfo->version));
	strlcpy(drvinfo->driver, "vxlan", sizeof(drvinfo->driver));
}

static const struct ethtool_ops vxlan_ethtool_ops = {
	.get_drvinfo	= vxlan_get_drvinfo,
	.get_link	= ethtool_op_get_link,
};

2637 2638 2639
static void vxlan_del_work(struct work_struct *work)
{
	struct vxlan_sock *vs = container_of(work, struct vxlan_sock, del_work);
2640
	udp_tunnel_sock_release(vs->sock);
2641 2642 2643
	kfree_rcu(vs, rcu);
}

T
Tom Herbert 已提交
2644 2645
static struct socket *vxlan_create_sock(struct net *net, bool ipv6,
					__be16 port, u32 flags)
2646
{
C
Cong Wang 已提交
2647
	struct socket *sock;
T
Tom Herbert 已提交
2648 2649
	struct udp_port_cfg udp_conf;
	int err;
C
Cong Wang 已提交
2650

T
Tom Herbert 已提交
2651
	memset(&udp_conf, 0, sizeof(udp_conf));
C
Cong Wang 已提交
2652

T
Tom Herbert 已提交
2653 2654 2655
	if (ipv6) {
		udp_conf.family = AF_INET6;
		udp_conf.use_udp6_rx_checksums =
2656
		    !(flags & VXLAN_F_UDP_ZERO_CSUM6_RX);
2657
		udp_conf.ipv6_v6only = 1;
T
Tom Herbert 已提交
2658 2659
	} else {
		udp_conf.family = AF_INET;
C
Cong Wang 已提交
2660 2661
	}

T
Tom Herbert 已提交
2662
	udp_conf.local_udp_port = port;
2663

T
Tom Herbert 已提交
2664 2665 2666 2667
	/* Open UDP socket */
	err = udp_sock_create(net, &udp_conf, &sock);
	if (err < 0)
		return ERR_PTR(err);
C
Cong Wang 已提交
2668

Z
Zhi Yong Wu 已提交
2669
	return sock;
C
Cong Wang 已提交
2670 2671 2672
}

/* Create new listen socket if needed */
2673 2674
static struct vxlan_sock *vxlan_socket_create(struct net *net, bool ipv6,
					      __be16 port, u32 flags)
C
Cong Wang 已提交
2675 2676 2677 2678 2679
{
	struct vxlan_net *vn = net_generic(net, vxlan_net_id);
	struct vxlan_sock *vs;
	struct socket *sock;
	unsigned int h;
2680
	struct udp_tunnel_sock_cfg tunnel_cfg;
C
Cong Wang 已提交
2681

2682
	vs = kzalloc(sizeof(*vs), GFP_KERNEL);
C
Cong Wang 已提交
2683 2684 2685 2686 2687 2688 2689 2690
	if (!vs)
		return ERR_PTR(-ENOMEM);

	for (h = 0; h < VNI_HASH_SIZE; ++h)
		INIT_HLIST_HEAD(&vs->vni_list[h]);

	INIT_WORK(&vs->del_work, vxlan_del_work);

T
Tom Herbert 已提交
2691
	sock = vxlan_create_sock(net, ipv6, port, flags);
Z
Zhi Yong Wu 已提交
2692
	if (IS_ERR(sock)) {
2693 2694
		pr_info("Cannot bind port %d, err=%ld\n", ntohs(port),
			PTR_ERR(sock));
2695
		kfree(vs);
2696
		return ERR_CAST(sock);
2697
	}
C
Cong Wang 已提交
2698 2699

	vs->sock = sock;
2700
	atomic_set(&vs->refcnt, 1);
2701
	vs->flags = (flags & VXLAN_F_RCV_FLAGS);
2702

2703 2704 2705 2706 2707
	/* Initialize the vxlan udp offloads structure */
	vs->udp_offloads.port = port;
	vs->udp_offloads.callbacks.gro_receive  = vxlan_gro_receive;
	vs->udp_offloads.callbacks.gro_complete = vxlan_gro_complete;

2708 2709
	spin_lock(&vn->sock_lock);
	hlist_add_head_rcu(&vs->hlist, vs_head(net, port));
2710
	vxlan_notify_add_rx_port(vs);
2711
	spin_unlock(&vn->sock_lock);
2712 2713

	/* Mark socket as an encapsulation socket. */
2714 2715 2716 2717 2718 2719
	tunnel_cfg.sk_user_data = vs;
	tunnel_cfg.encap_type = 1;
	tunnel_cfg.encap_rcv = vxlan_udp_encap_recv;
	tunnel_cfg.encap_destroy = NULL;

	setup_udp_tunnel_sock(net, sock, &tunnel_cfg);
C
Cong Wang 已提交
2720

2721 2722 2723
	return vs;
}

2724
static int __vxlan_sock_add(struct vxlan_dev *vxlan, bool ipv6)
2725
{
2726 2727
	struct vxlan_net *vn = net_generic(vxlan->net, vxlan_net_id);
	struct vxlan_sock *vs = NULL;
2728

2729
	if (!vxlan->cfg.no_share) {
2730
		spin_lock(&vn->sock_lock);
2731 2732 2733
		vs = vxlan_find_sock(vxlan->net, ipv6 ? AF_INET6 : AF_INET,
				     vxlan->cfg.dst_port, vxlan->flags);
		if (vs && !atomic_add_unless(&vs->refcnt, 1, 0)) {
2734
			spin_unlock(&vn->sock_lock);
2735
			return -EBUSY;
2736 2737 2738
		}
		spin_unlock(&vn->sock_lock);
	}
2739
	if (!vs)
2740 2741
		vs = vxlan_socket_create(vxlan->net, ipv6,
					 vxlan->cfg.dst_port, vxlan->flags);
2742 2743
	if (IS_ERR(vs))
		return PTR_ERR(vs);
2744 2745 2746 2747 2748 2749
#if IS_ENABLED(CONFIG_IPV6)
	if (ipv6)
		vxlan->vn6_sock = vs;
	else
#endif
		vxlan->vn4_sock = vs;
2750 2751
	vxlan_vs_add_dev(vs, vxlan);
	return 0;
2752 2753
}

2754 2755 2756 2757 2758 2759 2760 2761 2762 2763 2764 2765 2766 2767 2768 2769 2770 2771 2772
static int vxlan_sock_add(struct vxlan_dev *vxlan)
{
	bool ipv6 = vxlan->flags & VXLAN_F_IPV6;
	bool metadata = vxlan->flags & VXLAN_F_COLLECT_METADATA;
	int ret = 0;

	vxlan->vn4_sock = NULL;
#if IS_ENABLED(CONFIG_IPV6)
	vxlan->vn6_sock = NULL;
	if (ipv6 || metadata)
		ret = __vxlan_sock_add(vxlan, true);
#endif
	if (!ret && (!ipv6 || metadata))
		ret = __vxlan_sock_add(vxlan, false);
	if (ret < 0)
		vxlan_sock_release(vxlan);
	return ret;
}

2773 2774
static int vxlan_dev_configure(struct net *src_net, struct net_device *dev,
			       struct vxlan_config *conf)
S
stephen hemminger 已提交
2775
{
2776
	struct vxlan_net *vn = net_generic(src_net, vxlan_net_id);
2777
	struct vxlan_dev *vxlan = netdev_priv(dev), *tmp;
2778
	struct vxlan_rdst *dst = &vxlan->default_dst;
2779
	unsigned short needed_headroom = ETH_HLEN;
S
stephen hemminger 已提交
2780
	int err;
C
Cong Wang 已提交
2781
	bool use_ipv6 = false;
2782
	__be16 default_port = vxlan->cfg.dst_port;
2783
	struct net_device *lowerdev = NULL;
S
stephen hemminger 已提交
2784

2785
	vxlan->net = src_net;
N
Nicolas Dichtel 已提交
2786

2787
	dst->remote_vni = conf->vni;
S
stephen hemminger 已提交
2788

2789
	memcpy(&dst->remote_ip, &conf->remote_ip, sizeof(conf->remote_ip));
S
stephen hemminger 已提交
2790

2791 2792 2793
	/* Unless IPv6 is explicitly requested, assume IPv4 */
	if (!dst->remote_ip.sa.sa_family)
		dst->remote_ip.sa.sa_family = AF_INET;
C
Cong Wang 已提交
2794

2795
	if (dst->remote_ip.sa.sa_family == AF_INET6 ||
2796 2797 2798
	    vxlan->cfg.saddr.sa.sa_family == AF_INET6) {
		if (!IS_ENABLED(CONFIG_IPV6))
			return -EPFNOSUPPORT;
C
Cong Wang 已提交
2799
		use_ipv6 = true;
2800
		vxlan->flags |= VXLAN_F_IPV6;
2801
	}
S
stephen hemminger 已提交
2802

2803
	if (conf->remote_ifindex) {
2804
		lowerdev = __dev_get_by_index(src_net, conf->remote_ifindex);
2805
		dst->remote_ifindex = conf->remote_ifindex;
2806 2807

		if (!lowerdev) {
2808
			pr_info("ifindex %d does not exist\n", dst->remote_ifindex);
2809 2810
			return -ENODEV;
		}
S
stephen hemminger 已提交
2811

C
Cong Wang 已提交
2812 2813 2814 2815 2816 2817 2818 2819 2820 2821
#if IS_ENABLED(CONFIG_IPV6)
		if (use_ipv6) {
			struct inet6_dev *idev = __in6_dev_get(lowerdev);
			if (idev && idev->cnf.disable_ipv6) {
				pr_info("IPv6 is disabled via sysctl\n");
				return -EPERM;
			}
		}
#endif

2822
		if (!conf->mtu)
C
Cong Wang 已提交
2823
			dev->mtu = lowerdev->mtu - (use_ipv6 ? VXLAN6_HEADROOM : VXLAN_HEADROOM);
2824

2825
		needed_headroom = lowerdev->hard_header_len;
J
Jiri Benc 已提交
2826
	}
S
stephen hemminger 已提交
2827

2828 2829 2830 2831 2832 2833
	if (conf->mtu) {
		err = __vxlan_change_mtu(dev, lowerdev, dst, conf->mtu, false);
		if (err)
			return err;
	}

2834 2835 2836 2837 2838 2839
	if (use_ipv6 || conf->flags & VXLAN_F_COLLECT_METADATA)
		needed_headroom += VXLAN6_HEADROOM;
	else
		needed_headroom += VXLAN_HEADROOM;
	dev->needed_headroom = needed_headroom;

2840 2841 2842 2843 2844 2845 2846 2847
	memcpy(&vxlan->cfg, conf, sizeof(*conf));
	if (!vxlan->cfg.dst_port)
		vxlan->cfg.dst_port = default_port;
	vxlan->flags |= conf->flags;

	if (!vxlan->cfg.age_interval)
		vxlan->cfg.age_interval = FDB_AGE_DEFAULT;

2848 2849 2850 2851 2852 2853 2854
	list_for_each_entry(tmp, &vn->vxlan_list, next) {
		if (tmp->cfg.vni == conf->vni &&
		    (tmp->default_dst.remote_ip.sa.sa_family == AF_INET6 ||
		     tmp->cfg.saddr.sa.sa_family == AF_INET6) == use_ipv6 &&
		    tmp->cfg.dst_port == vxlan->cfg.dst_port &&
		    (tmp->flags & VXLAN_F_RCV_FLAGS) ==
		    (vxlan->flags & VXLAN_F_RCV_FLAGS))
2855
		return -EEXIST;
2856
	}
2857 2858 2859 2860 2861 2862 2863 2864 2865 2866 2867 2868 2869 2870 2871 2872 2873 2874 2875 2876 2877 2878 2879 2880 2881 2882 2883 2884 2885 2886 2887 2888 2889 2890 2891 2892 2893 2894 2895 2896 2897 2898 2899 2900 2901 2902 2903 2904 2905 2906 2907 2908 2909 2910 2911 2912 2913 2914 2915

	dev->ethtool_ops = &vxlan_ethtool_ops;

	/* create an fdb entry for a valid default destination */
	if (!vxlan_addr_any(&vxlan->default_dst.remote_ip)) {
		err = vxlan_fdb_create(vxlan, all_zeros_mac,
				       &vxlan->default_dst.remote_ip,
				       NUD_REACHABLE|NUD_PERMANENT,
				       NLM_F_EXCL|NLM_F_CREATE,
				       vxlan->cfg.dst_port,
				       vxlan->default_dst.remote_vni,
				       vxlan->default_dst.remote_ifindex,
				       NTF_SELF);
		if (err)
			return err;
	}

	err = register_netdevice(dev);
	if (err) {
		vxlan_fdb_delete_default(vxlan);
		return err;
	}

	list_add(&vxlan->next, &vn->vxlan_list);

	return 0;
}

struct net_device *vxlan_dev_create(struct net *net, const char *name,
				    u8 name_assign_type, struct vxlan_config *conf)
{
	struct nlattr *tb[IFLA_MAX+1];
	struct net_device *dev;
	int err;

	memset(&tb, 0, sizeof(tb));

	dev = rtnl_create_link(net, name, name_assign_type,
			       &vxlan_link_ops, tb);
	if (IS_ERR(dev))
		return dev;

	err = vxlan_dev_configure(net, dev, conf);
	if (err < 0) {
		free_netdev(dev);
		return ERR_PTR(err);
	}

	return dev;
}
EXPORT_SYMBOL_GPL(vxlan_dev_create);

static int vxlan_newlink(struct net *src_net, struct net_device *dev,
			 struct nlattr *tb[], struct nlattr *data[])
{
	struct vxlan_config conf;
	int err;

	memset(&conf, 0, sizeof(conf));
2916 2917 2918

	if (data[IFLA_VXLAN_ID])
		conf.vni = nla_get_u32(data[IFLA_VXLAN_ID]);
2919 2920 2921 2922 2923 2924 2925 2926 2927 2928 2929 2930 2931 2932 2933 2934 2935 2936 2937 2938 2939 2940 2941 2942 2943 2944

	if (data[IFLA_VXLAN_GROUP]) {
		conf.remote_ip.sin.sin_addr.s_addr = nla_get_in_addr(data[IFLA_VXLAN_GROUP]);
	} else if (data[IFLA_VXLAN_GROUP6]) {
		if (!IS_ENABLED(CONFIG_IPV6))
			return -EPFNOSUPPORT;

		conf.remote_ip.sin6.sin6_addr = nla_get_in6_addr(data[IFLA_VXLAN_GROUP6]);
		conf.remote_ip.sa.sa_family = AF_INET6;
	}

	if (data[IFLA_VXLAN_LOCAL]) {
		conf.saddr.sin.sin_addr.s_addr = nla_get_in_addr(data[IFLA_VXLAN_LOCAL]);
		conf.saddr.sa.sa_family = AF_INET;
	} else if (data[IFLA_VXLAN_LOCAL6]) {
		if (!IS_ENABLED(CONFIG_IPV6))
			return -EPFNOSUPPORT;

		/* TODO: respect scope id */
		conf.saddr.sin6.sin6_addr = nla_get_in6_addr(data[IFLA_VXLAN_LOCAL6]);
		conf.saddr.sa.sa_family = AF_INET6;
	}

	if (data[IFLA_VXLAN_LINK])
		conf.remote_ifindex = nla_get_u32(data[IFLA_VXLAN_LINK]);

S
stephen hemminger 已提交
2945
	if (data[IFLA_VXLAN_TOS])
2946
		conf.tos  = nla_get_u8(data[IFLA_VXLAN_TOS]);
S
stephen hemminger 已提交
2947

2948
	if (data[IFLA_VXLAN_TTL])
2949
		conf.ttl = nla_get_u8(data[IFLA_VXLAN_TTL]);
2950

S
stephen hemminger 已提交
2951
	if (!data[IFLA_VXLAN_LEARNING] || nla_get_u8(data[IFLA_VXLAN_LEARNING]))
2952
		conf.flags |= VXLAN_F_LEARN;
S
stephen hemminger 已提交
2953 2954

	if (data[IFLA_VXLAN_AGEING])
2955
		conf.age_interval = nla_get_u32(data[IFLA_VXLAN_AGEING]);
S
stephen hemminger 已提交
2956

D
David Stevens 已提交
2957
	if (data[IFLA_VXLAN_PROXY] && nla_get_u8(data[IFLA_VXLAN_PROXY]))
2958
		conf.flags |= VXLAN_F_PROXY;
D
David Stevens 已提交
2959 2960

	if (data[IFLA_VXLAN_RSC] && nla_get_u8(data[IFLA_VXLAN_RSC]))
2961
		conf.flags |= VXLAN_F_RSC;
D
David Stevens 已提交
2962 2963

	if (data[IFLA_VXLAN_L2MISS] && nla_get_u8(data[IFLA_VXLAN_L2MISS]))
2964
		conf.flags |= VXLAN_F_L2MISS;
D
David Stevens 已提交
2965 2966

	if (data[IFLA_VXLAN_L3MISS] && nla_get_u8(data[IFLA_VXLAN_L3MISS]))
2967
		conf.flags |= VXLAN_F_L3MISS;
D
David Stevens 已提交
2968

S
stephen hemminger 已提交
2969
	if (data[IFLA_VXLAN_LIMIT])
2970
		conf.addrmax = nla_get_u32(data[IFLA_VXLAN_LIMIT]);
S
stephen hemminger 已提交
2971

2972 2973 2974 2975
	if (data[IFLA_VXLAN_COLLECT_METADATA] &&
	    nla_get_u8(data[IFLA_VXLAN_COLLECT_METADATA]))
		conf.flags |= VXLAN_F_COLLECT_METADATA;

2976 2977 2978
	if (data[IFLA_VXLAN_PORT_RANGE]) {
		const struct ifla_vxlan_port_range *p
			= nla_data(data[IFLA_VXLAN_PORT_RANGE]);
2979 2980
		conf.port_min = ntohs(p->low);
		conf.port_max = ntohs(p->high);
2981 2982
	}

2983
	if (data[IFLA_VXLAN_PORT])
2984
		conf.dst_port = nla_get_be16(data[IFLA_VXLAN_PORT]);
2985

2986
	if (data[IFLA_VXLAN_UDP_CSUM] && nla_get_u8(data[IFLA_VXLAN_UDP_CSUM]))
2987
		conf.flags |= VXLAN_F_UDP_CSUM;
2988 2989 2990

	if (data[IFLA_VXLAN_UDP_ZERO_CSUM6_TX] &&
	    nla_get_u8(data[IFLA_VXLAN_UDP_ZERO_CSUM6_TX]))
2991
		conf.flags |= VXLAN_F_UDP_ZERO_CSUM6_TX;
2992 2993 2994

	if (data[IFLA_VXLAN_UDP_ZERO_CSUM6_RX] &&
	    nla_get_u8(data[IFLA_VXLAN_UDP_ZERO_CSUM6_RX]))
2995
		conf.flags |= VXLAN_F_UDP_ZERO_CSUM6_RX;
2996

T
Tom Herbert 已提交
2997 2998
	if (data[IFLA_VXLAN_REMCSUM_TX] &&
	    nla_get_u8(data[IFLA_VXLAN_REMCSUM_TX]))
2999
		conf.flags |= VXLAN_F_REMCSUM_TX;
T
Tom Herbert 已提交
3000 3001 3002

	if (data[IFLA_VXLAN_REMCSUM_RX] &&
	    nla_get_u8(data[IFLA_VXLAN_REMCSUM_RX]))
3003
		conf.flags |= VXLAN_F_REMCSUM_RX;
T
Tom Herbert 已提交
3004

T
Thomas Graf 已提交
3005
	if (data[IFLA_VXLAN_GBP])
3006
		conf.flags |= VXLAN_F_GBP;
T
Thomas Graf 已提交
3007

3008
	if (data[IFLA_VXLAN_REMCSUM_NOPARTIAL])
3009
		conf.flags |= VXLAN_F_REMCSUM_NOPARTIAL;
3010

3011 3012 3013 3014 3015
	err = vxlan_dev_configure(src_net, dev, &conf);
	switch (err) {
	case -ENODEV:
		pr_info("ifindex %d does not exist\n", conf.remote_ifindex);
		break;
Y
Yan Burman 已提交
3016

3017 3018 3019
	case -EPERM:
		pr_info("IPv6 is disabled via sysctl\n");
		break;
S
stephen hemminger 已提交
3020

3021 3022 3023
	case -EEXIST:
		pr_info("duplicate VNI %u\n", conf.vni);
		break;
3024 3025
	}

3026
	return err;
S
stephen hemminger 已提交
3027 3028 3029 3030 3031
}

static void vxlan_dellink(struct net_device *dev, struct list_head *head)
{
	struct vxlan_dev *vxlan = netdev_priv(dev);
N
Nicolas Dichtel 已提交
3032
	struct vxlan_net *vn = net_generic(vxlan->net, vxlan_net_id);
S
stephen hemminger 已提交
3033

3034
	spin_lock(&vn->sock_lock);
3035 3036
	if (!hlist_unhashed(&vxlan->hlist))
		hlist_del_rcu(&vxlan->hlist);
3037 3038
	spin_unlock(&vn->sock_lock);

T
Tom Herbert 已提交
3039
	gro_cells_destroy(&vxlan->gro_cells);
3040
	list_del(&vxlan->next);
S
stephen hemminger 已提交
3041 3042 3043 3044 3045 3046 3047
	unregister_netdevice_queue(dev, head);
}

static size_t vxlan_get_size(const struct net_device *dev)
{

	return nla_total_size(sizeof(__u32)) +	/* IFLA_VXLAN_ID */
C
Cong Wang 已提交
3048
		nla_total_size(sizeof(struct in6_addr)) + /* IFLA_VXLAN_GROUP{6} */
S
stephen hemminger 已提交
3049
		nla_total_size(sizeof(__u32)) +	/* IFLA_VXLAN_LINK */
C
Cong Wang 已提交
3050
		nla_total_size(sizeof(struct in6_addr)) + /* IFLA_VXLAN_LOCAL{6} */
S
stephen hemminger 已提交
3051 3052 3053
		nla_total_size(sizeof(__u8)) +	/* IFLA_VXLAN_TTL */
		nla_total_size(sizeof(__u8)) +	/* IFLA_VXLAN_TOS */
		nla_total_size(sizeof(__u8)) +	/* IFLA_VXLAN_LEARNING */
D
David Stevens 已提交
3054 3055 3056 3057
		nla_total_size(sizeof(__u8)) +	/* IFLA_VXLAN_PROXY */
		nla_total_size(sizeof(__u8)) +	/* IFLA_VXLAN_RSC */
		nla_total_size(sizeof(__u8)) +	/* IFLA_VXLAN_L2MISS */
		nla_total_size(sizeof(__u8)) +	/* IFLA_VXLAN_L3MISS */
3058
		nla_total_size(sizeof(__u8)) +	/* IFLA_VXLAN_COLLECT_METADATA */
S
stephen hemminger 已提交
3059 3060
		nla_total_size(sizeof(__u32)) +	/* IFLA_VXLAN_AGEING */
		nla_total_size(sizeof(__u32)) +	/* IFLA_VXLAN_LIMIT */
3061
		nla_total_size(sizeof(struct ifla_vxlan_port_range)) +
3062 3063 3064 3065
		nla_total_size(sizeof(__be16)) + /* IFLA_VXLAN_PORT */
		nla_total_size(sizeof(__u8)) + /* IFLA_VXLAN_UDP_CSUM */
		nla_total_size(sizeof(__u8)) + /* IFLA_VXLAN_UDP_ZERO_CSUM6_TX */
		nla_total_size(sizeof(__u8)) + /* IFLA_VXLAN_UDP_ZERO_CSUM6_RX */
T
Tom Herbert 已提交
3066 3067
		nla_total_size(sizeof(__u8)) + /* IFLA_VXLAN_REMCSUM_TX */
		nla_total_size(sizeof(__u8)) + /* IFLA_VXLAN_REMCSUM_RX */
S
stephen hemminger 已提交
3068 3069 3070 3071 3072 3073
		0;
}

static int vxlan_fill_info(struct sk_buff *skb, const struct net_device *dev)
{
	const struct vxlan_dev *vxlan = netdev_priv(dev);
3074
	const struct vxlan_rdst *dst = &vxlan->default_dst;
3075
	struct ifla_vxlan_port_range ports = {
3076 3077
		.low =  htons(vxlan->cfg.port_min),
		.high = htons(vxlan->cfg.port_max),
3078
	};
S
stephen hemminger 已提交
3079

3080
	if (nla_put_u32(skb, IFLA_VXLAN_ID, dst->remote_vni))
S
stephen hemminger 已提交
3081 3082
		goto nla_put_failure;

C
Cong Wang 已提交
3083 3084
	if (!vxlan_addr_any(&dst->remote_ip)) {
		if (dst->remote_ip.sa.sa_family == AF_INET) {
3085 3086
			if (nla_put_in_addr(skb, IFLA_VXLAN_GROUP,
					    dst->remote_ip.sin.sin_addr.s_addr))
C
Cong Wang 已提交
3087 3088 3089
				goto nla_put_failure;
#if IS_ENABLED(CONFIG_IPV6)
		} else {
3090 3091
			if (nla_put_in6_addr(skb, IFLA_VXLAN_GROUP6,
					     &dst->remote_ip.sin6.sin6_addr))
C
Cong Wang 已提交
3092 3093 3094 3095
				goto nla_put_failure;
#endif
		}
	}
S
stephen hemminger 已提交
3096

3097
	if (dst->remote_ifindex && nla_put_u32(skb, IFLA_VXLAN_LINK, dst->remote_ifindex))
S
stephen hemminger 已提交
3098 3099
		goto nla_put_failure;

3100 3101
	if (!vxlan_addr_any(&vxlan->cfg.saddr)) {
		if (vxlan->cfg.saddr.sa.sa_family == AF_INET) {
3102
			if (nla_put_in_addr(skb, IFLA_VXLAN_LOCAL,
3103
					    vxlan->cfg.saddr.sin.sin_addr.s_addr))
C
Cong Wang 已提交
3104 3105 3106
				goto nla_put_failure;
#if IS_ENABLED(CONFIG_IPV6)
		} else {
3107
			if (nla_put_in6_addr(skb, IFLA_VXLAN_LOCAL6,
3108
					     &vxlan->cfg.saddr.sin6.sin6_addr))
C
Cong Wang 已提交
3109 3110 3111 3112
				goto nla_put_failure;
#endif
		}
	}
S
stephen hemminger 已提交
3113

3114 3115
	if (nla_put_u8(skb, IFLA_VXLAN_TTL, vxlan->cfg.ttl) ||
	    nla_put_u8(skb, IFLA_VXLAN_TOS, vxlan->cfg.tos) ||
D
David Stevens 已提交
3116 3117 3118 3119 3120 3121 3122 3123 3124
	    nla_put_u8(skb, IFLA_VXLAN_LEARNING,
			!!(vxlan->flags & VXLAN_F_LEARN)) ||
	    nla_put_u8(skb, IFLA_VXLAN_PROXY,
			!!(vxlan->flags & VXLAN_F_PROXY)) ||
	    nla_put_u8(skb, IFLA_VXLAN_RSC, !!(vxlan->flags & VXLAN_F_RSC)) ||
	    nla_put_u8(skb, IFLA_VXLAN_L2MISS,
			!!(vxlan->flags & VXLAN_F_L2MISS)) ||
	    nla_put_u8(skb, IFLA_VXLAN_L3MISS,
			!!(vxlan->flags & VXLAN_F_L3MISS)) ||
3125 3126
	    nla_put_u8(skb, IFLA_VXLAN_COLLECT_METADATA,
		       !!(vxlan->flags & VXLAN_F_COLLECT_METADATA)) ||
3127 3128 3129
	    nla_put_u32(skb, IFLA_VXLAN_AGEING, vxlan->cfg.age_interval) ||
	    nla_put_u32(skb, IFLA_VXLAN_LIMIT, vxlan->cfg.addrmax) ||
	    nla_put_be16(skb, IFLA_VXLAN_PORT, vxlan->cfg.dst_port) ||
3130 3131 3132 3133 3134
	    nla_put_u8(skb, IFLA_VXLAN_UDP_CSUM,
			!!(vxlan->flags & VXLAN_F_UDP_CSUM)) ||
	    nla_put_u8(skb, IFLA_VXLAN_UDP_ZERO_CSUM6_TX,
			!!(vxlan->flags & VXLAN_F_UDP_ZERO_CSUM6_TX)) ||
	    nla_put_u8(skb, IFLA_VXLAN_UDP_ZERO_CSUM6_RX,
T
Tom Herbert 已提交
3135 3136 3137 3138 3139
			!!(vxlan->flags & VXLAN_F_UDP_ZERO_CSUM6_RX)) ||
	    nla_put_u8(skb, IFLA_VXLAN_REMCSUM_TX,
			!!(vxlan->flags & VXLAN_F_REMCSUM_TX)) ||
	    nla_put_u8(skb, IFLA_VXLAN_REMCSUM_RX,
			!!(vxlan->flags & VXLAN_F_REMCSUM_RX)))
S
stephen hemminger 已提交
3140 3141
		goto nla_put_failure;

3142 3143 3144
	if (nla_put(skb, IFLA_VXLAN_PORT_RANGE, sizeof(ports), &ports))
		goto nla_put_failure;

T
Thomas Graf 已提交
3145 3146 3147 3148
	if (vxlan->flags & VXLAN_F_GBP &&
	    nla_put_flag(skb, IFLA_VXLAN_GBP))
		goto nla_put_failure;

3149 3150 3151 3152
	if (vxlan->flags & VXLAN_F_REMCSUM_NOPARTIAL &&
	    nla_put_flag(skb, IFLA_VXLAN_REMCSUM_NOPARTIAL))
		goto nla_put_failure;

S
stephen hemminger 已提交
3153 3154 3155 3156 3157 3158
	return 0;

nla_put_failure:
	return -EMSGSIZE;
}

3159 3160 3161 3162 3163 3164 3165
static struct net *vxlan_get_link_net(const struct net_device *dev)
{
	struct vxlan_dev *vxlan = netdev_priv(dev);

	return vxlan->net;
}

S
stephen hemminger 已提交
3166 3167 3168 3169 3170 3171 3172 3173 3174 3175 3176
static struct rtnl_link_ops vxlan_link_ops __read_mostly = {
	.kind		= "vxlan",
	.maxtype	= IFLA_VXLAN_MAX,
	.policy		= vxlan_policy,
	.priv_size	= sizeof(struct vxlan_dev),
	.setup		= vxlan_setup,
	.validate	= vxlan_validate,
	.newlink	= vxlan_newlink,
	.dellink	= vxlan_dellink,
	.get_size	= vxlan_get_size,
	.fill_info	= vxlan_fill_info,
3177
	.get_link_net	= vxlan_get_link_net,
S
stephen hemminger 已提交
3178 3179
};

3180 3181 3182 3183 3184 3185 3186 3187 3188 3189 3190 3191 3192 3193 3194 3195 3196 3197 3198 3199 3200 3201 3202 3203 3204 3205
static void vxlan_handle_lowerdev_unregister(struct vxlan_net *vn,
					     struct net_device *dev)
{
	struct vxlan_dev *vxlan, *next;
	LIST_HEAD(list_kill);

	list_for_each_entry_safe(vxlan, next, &vn->vxlan_list, next) {
		struct vxlan_rdst *dst = &vxlan->default_dst;

		/* In case we created vxlan device with carrier
		 * and we loose the carrier due to module unload
		 * we also need to remove vxlan device. In other
		 * cases, it's not necessary and remote_ifindex
		 * is 0 here, so no matches.
		 */
		if (dst->remote_ifindex == dev->ifindex)
			vxlan_dellink(vxlan->dev, &list_kill);
	}

	unregister_netdevice_many(&list_kill);
}

static int vxlan_lowerdev_event(struct notifier_block *unused,
				unsigned long event, void *ptr)
{
	struct net_device *dev = netdev_notifier_info_to_dev(ptr);
3206
	struct vxlan_net *vn = net_generic(dev_net(dev), vxlan_net_id);
3207

3208
	if (event == NETDEV_UNREGISTER)
3209 3210 3211 3212 3213 3214 3215 3216 3217
		vxlan_handle_lowerdev_unregister(vn, dev);

	return NOTIFY_DONE;
}

static struct notifier_block vxlan_notifier_block __read_mostly = {
	.notifier_call = vxlan_lowerdev_event,
};

S
stephen hemminger 已提交
3218 3219 3220
static __net_init int vxlan_init_net(struct net *net)
{
	struct vxlan_net *vn = net_generic(net, vxlan_net_id);
3221
	unsigned int h;
S
stephen hemminger 已提交
3222

3223
	INIT_LIST_HEAD(&vn->vxlan_list);
3224
	spin_lock_init(&vn->sock_lock);
S
stephen hemminger 已提交
3225

3226 3227
	for (h = 0; h < PORT_HASH_SIZE; ++h)
		INIT_HLIST_HEAD(&vn->sock_list[h]);
S
stephen hemminger 已提交
3228 3229 3230 3231

	return 0;
}

N
Nicolas Dichtel 已提交
3232 3233 3234 3235 3236 3237 3238 3239 3240 3241 3242 3243 3244 3245 3246 3247
static void __net_exit vxlan_exit_net(struct net *net)
{
	struct vxlan_net *vn = net_generic(net, vxlan_net_id);
	struct vxlan_dev *vxlan, *next;
	struct net_device *dev, *aux;
	LIST_HEAD(list);

	rtnl_lock();
	for_each_netdev_safe(net, dev, aux)
		if (dev->rtnl_link_ops == &vxlan_link_ops)
			unregister_netdevice_queue(dev, &list);

	list_for_each_entry_safe(vxlan, next, &vn->vxlan_list, next) {
		/* If vxlan->dev is in the same netns, it has already been added
		 * to the list by the previous loop.
		 */
T
Tom Herbert 已提交
3248 3249
		if (!net_eq(dev_net(vxlan->dev), net)) {
			gro_cells_destroy(&vxlan->gro_cells);
3250
			unregister_netdevice_queue(vxlan->dev, &list);
T
Tom Herbert 已提交
3251
		}
N
Nicolas Dichtel 已提交
3252 3253 3254 3255 3256 3257
	}

	unregister_netdevice_many(&list);
	rtnl_unlock();
}

S
stephen hemminger 已提交
3258 3259
static struct pernet_operations vxlan_net_ops = {
	.init = vxlan_init_net,
N
Nicolas Dichtel 已提交
3260
	.exit = vxlan_exit_net,
S
stephen hemminger 已提交
3261 3262 3263 3264 3265 3266 3267 3268
	.id   = &vxlan_net_id,
	.size = sizeof(struct vxlan_net),
};

static int __init vxlan_init_module(void)
{
	int rc;

3269 3270 3271 3272
	vxlan_wq = alloc_workqueue("vxlan", 0, 0);
	if (!vxlan_wq)
		return -ENOMEM;

S
stephen hemminger 已提交
3273 3274
	get_random_bytes(&vxlan_salt, sizeof(vxlan_salt));

3275
	rc = register_pernet_subsys(&vxlan_net_ops);
S
stephen hemminger 已提交
3276 3277 3278
	if (rc)
		goto out1;

3279
	rc = register_netdevice_notifier(&vxlan_notifier_block);
S
stephen hemminger 已提交
3280 3281 3282
	if (rc)
		goto out2;

3283 3284 3285
	rc = rtnl_link_register(&vxlan_link_ops);
	if (rc)
		goto out3;
S
stephen hemminger 已提交
3286

3287 3288 3289
	return 0;
out3:
	unregister_netdevice_notifier(&vxlan_notifier_block);
S
stephen hemminger 已提交
3290
out2:
3291
	unregister_pernet_subsys(&vxlan_net_ops);
S
stephen hemminger 已提交
3292
out1:
3293
	destroy_workqueue(vxlan_wq);
S
stephen hemminger 已提交
3294 3295
	return rc;
}
3296
late_initcall(vxlan_init_module);
S
stephen hemminger 已提交
3297 3298 3299

static void __exit vxlan_cleanup_module(void)
{
3300
	rtnl_link_unregister(&vxlan_link_ops);
3301
	unregister_netdevice_notifier(&vxlan_notifier_block);
3302
	destroy_workqueue(vxlan_wq);
3303 3304
	unregister_pernet_subsys(&vxlan_net_ops);
	/* rcu_barrier() is called by netns */
S
stephen hemminger 已提交
3305 3306 3307 3308 3309
}
module_exit(vxlan_cleanup_module);

MODULE_LICENSE("GPL");
MODULE_VERSION(VXLAN_VERSION);
3310
MODULE_AUTHOR("Stephen Hemminger <stephen@networkplumber.org>");
J
Jesse Brandeburg 已提交
3311
MODULE_DESCRIPTION("Driver for VXLAN encapsulated traffic");
S
stephen hemminger 已提交
3312
MODULE_ALIAS_RTNL_LINK("vxlan");