geneve.c 46.6 KB
Newer Older
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16
/*
 * GENEVE: Generic Network Virtualization Encapsulation
 *
 * Copyright (c) 2015 Red Hat, Inc.
 *
 * 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/module.h>
#include <linux/etherdevice.h>
#include <linux/hash.h>
17
#include <net/dst_metadata.h>
18
#include <net/gro_cells.h>
19 20
#include <net/rtnetlink.h>
#include <net/geneve.h>
21
#include <net/protocol.h>
22 23 24 25 26 27 28 29 30 31 32 33 34 35 36

#define GENEVE_NETDEV_VER	"0.6"

#define GENEVE_UDP_PORT		6081

#define GENEVE_N_VID		(1u << 24)
#define GENEVE_VID_MASK		(GENEVE_N_VID - 1)

#define VNI_HASH_BITS		10
#define VNI_HASH_SIZE		(1<<VNI_HASH_BITS)

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");

37 38
#define GENEVE_VER 0
#define GENEVE_BASE_HLEN (sizeof(struct udphdr) + sizeof(struct genevehdr))
39 40
#define GENEVE_IPV4_HLEN (ETH_HLEN + sizeof(struct iphdr) + GENEVE_BASE_HLEN)
#define GENEVE_IPV6_HLEN (ETH_HLEN + sizeof(struct ipv6hdr) + GENEVE_BASE_HLEN)
41

42 43
/* per-network namespace private data for this module */
struct geneve_net {
44 45
	struct list_head	geneve_list;
	struct list_head	sock_list;
46 47
};

48
static unsigned int geneve_net_id;
49

Jiri Benc's avatar
Jiri Benc committed
50 51 52 53 54
struct geneve_dev_node {
	struct hlist_node hlist;
	struct geneve_dev *geneve;
};

55 56
/* Pseudo network device */
struct geneve_dev {
Jiri Benc's avatar
Jiri Benc committed
57 58 59 60
	struct geneve_dev_node hlist4;	/* vni hash table for IPv4 socket */
#if IS_ENABLED(CONFIG_IPV6)
	struct geneve_dev_node hlist6;	/* vni hash table for IPv6 socket */
#endif
61 62
	struct net	   *net;	/* netns for packet i/o */
	struct net_device  *dev;	/* netdev for geneve tunnel */
63
	struct ip_tunnel_info info;
64
	struct geneve_sock __rcu *sock4;	/* IPv4 socket used for geneve tunnel */
65
#if IS_ENABLED(CONFIG_IPV6)
66
	struct geneve_sock __rcu *sock6;	/* IPv6 socket used for geneve tunnel */
67
#endif
68
	struct list_head   next;	/* geneve's per namespace list */
69
	struct gro_cells   gro_cells;
70 71
	bool		   collect_md;
	bool		   use_udp6_rx_checksums;
72
	bool		   ttl_inherit;
73 74
};

75 76 77 78 79 80
struct geneve_sock {
	bool			collect_md;
	struct list_head	list;
	struct socket		*sock;
	struct rcu_head		rcu;
	int			refcnt;
81
	struct hlist_head	vni_list[VNI_HASH_SIZE];
82
};
83 84 85 86 87 88 89 90 91

static inline __u32 geneve_net_vni_hash(u8 vni[3])
{
	__u32 vnid;

	vnid = (vni[0] << 16) | (vni[1] << 8) | vni[2];
	return hash_32(vnid, VNI_HASH_BITS);
}

92 93 94 95 96 97 98 99 100 101 102
static __be64 vni_to_tunnel_id(const __u8 *vni)
{
#ifdef __BIG_ENDIAN
	return (vni[0] << 16) | (vni[1] << 8) | vni[2];
#else
	return (__force __be64)(((__force u64)vni[0] << 40) |
				((__force u64)vni[1] << 48) |
				((__force u64)vni[2] << 56));
#endif
}

103 104 105 106 107 108 109 110 111 112 113 114 115 116
/* Convert 64 bit tunnel ID to 24 bit VNI. */
static void tunnel_id_to_vni(__be64 tun_id, __u8 *vni)
{
#ifdef __BIG_ENDIAN
	vni[0] = (__force __u8)(tun_id >> 16);
	vni[1] = (__force __u8)(tun_id >> 8);
	vni[2] = (__force __u8)tun_id;
#else
	vni[0] = (__force __u8)((__force u64)tun_id >> 40);
	vni[1] = (__force __u8)((__force u64)tun_id >> 48);
	vni[2] = (__force __u8)((__force u64)tun_id >> 56);
#endif
}

117 118 119 120 121
static bool eq_tun_id_and_vni(u8 *tun_id, u8 *vni)
{
	return !memcmp(vni, &tun_id[5], 3);
}

122 123 124 125 126
static sa_family_t geneve_get_sk_family(struct geneve_sock *gs)
{
	return gs->sock->sk->sk_family;
}

127
static struct geneve_dev *geneve_lookup(struct geneve_sock *gs,
128
					__be32 addr, u8 vni[])
129 130
{
	struct hlist_head *vni_list_head;
Jiri Benc's avatar
Jiri Benc committed
131
	struct geneve_dev_node *node;
132 133 134
	__u32 hash;

	/* Find the device for this VNI */
135
	hash = geneve_net_vni_hash(vni);
136
	vni_list_head = &gs->vni_list[hash];
Jiri Benc's avatar
Jiri Benc committed
137 138 139 140
	hlist_for_each_entry_rcu(node, vni_list_head, hlist) {
		if (eq_tun_id_and_vni((u8 *)&node->geneve->info.key.tun_id, vni) &&
		    addr == node->geneve->info.key.u.ipv4.dst)
			return node->geneve;
141 142 143 144 145 146 147 148 149
	}
	return NULL;
}

#if IS_ENABLED(CONFIG_IPV6)
static struct geneve_dev *geneve6_lookup(struct geneve_sock *gs,
					 struct in6_addr addr6, u8 vni[])
{
	struct hlist_head *vni_list_head;
Jiri Benc's avatar
Jiri Benc committed
150
	struct geneve_dev_node *node;
151 152 153 154 155
	__u32 hash;

	/* Find the device for this VNI */
	hash = geneve_net_vni_hash(vni);
	vni_list_head = &gs->vni_list[hash];
Jiri Benc's avatar
Jiri Benc committed
156 157 158 159
	hlist_for_each_entry_rcu(node, vni_list_head, hlist) {
		if (eq_tun_id_and_vni((u8 *)&node->geneve->info.key.tun_id, vni) &&
		    ipv6_addr_equal(&addr6, &node->geneve->info.key.u.ipv6.dst))
			return node->geneve;
160
	}
161 162
	return NULL;
}
163
#endif
164

165 166 167 168 169
static inline struct genevehdr *geneve_hdr(const struct sk_buff *skb)
{
	return (struct genevehdr *)(udp_hdr(skb) + 1);
}

170 171
static struct geneve_dev *geneve_lookup_skb(struct geneve_sock *gs,
					    struct sk_buff *skb)
172
{
173
	static u8 zero_vni[3];
174
	u8 *vni;
175

176
	if (geneve_get_sk_family(gs) == AF_INET) {
177
		struct iphdr *iph;
178
		__be32 addr;
179

180
		iph = ip_hdr(skb); /* outer IP header... */
181

182 183 184 185
		if (gs->collect_md) {
			vni = zero_vni;
			addr = 0;
		} else {
186
			vni = geneve_hdr(skb)->vni;
187 188 189
			addr = iph->saddr;
		}

190
		return geneve_lookup(gs, addr, vni);
191
#if IS_ENABLED(CONFIG_IPV6)
192
	} else if (geneve_get_sk_family(gs) == AF_INET6) {
193
		static struct in6_addr zero_addr6;
194 195 196
		struct ipv6hdr *ip6h;
		struct in6_addr addr6;

197
		ip6h = ipv6_hdr(skb); /* outer IPv6 header... */
198

199 200 201 202
		if (gs->collect_md) {
			vni = zero_vni;
			addr6 = zero_addr6;
		} else {
203
			vni = geneve_hdr(skb)->vni;
204 205 206
			addr6 = ip6h->saddr;
		}

207
		return geneve6_lookup(gs, addr6, vni);
208 209
#endif
	}
210 211 212 213 214 215 216 217 218 219
	return NULL;
}

/* geneve receive/decap routine */
static void geneve_rx(struct geneve_dev *geneve, struct geneve_sock *gs,
		      struct sk_buff *skb)
{
	struct genevehdr *gnvh = geneve_hdr(skb);
	struct metadata_dst *tun_dst = NULL;
	struct pcpu_sw_netstats *stats;
220
	unsigned int len;
221 222
	int err = 0;
	void *oiph;
223

224
	if (ip_tunnel_collect_metadata() || gs->collect_md) {
225 226 227 228 229 230
		__be16 flags;

		flags = TUNNEL_KEY | TUNNEL_GENEVE_OPT |
			(gnvh->oam ? TUNNEL_OAM : 0) |
			(gnvh->critical ? TUNNEL_CRIT_OPT : 0);

231
		tun_dst = udp_tun_rx_dst(skb, geneve_get_sk_family(gs), flags,
232 233
					 vni_to_tunnel_id(gnvh->vni),
					 gnvh->opt_len * 4);
234 235
		if (!tun_dst) {
			geneve->dev->stats.rx_dropped++;
236
			goto drop;
237
		}
238
		/* Update tunnel dst according to Geneve options. */
239
		ip_tunnel_info_opts_set(&tun_dst->u.tun_info,
240 241
					gnvh->options, gnvh->opt_len * 4,
					TUNNEL_GENEVE_OPT);
242 243 244 245
	} else {
		/* Drop packets w/ critical options,
		 * since we don't support any...
		 */
246 247 248
		if (gnvh->critical) {
			geneve->dev->stats.rx_frame_errors++;
			geneve->dev->stats.rx_errors++;
249
			goto drop;
250
		}
251
	}
252 253 254 255 256

	skb_reset_mac_header(skb);
	skb->protocol = eth_type_trans(skb, geneve->dev);
	skb_postpull_rcsum(skb, eth_hdr(skb), ETH_HLEN);

257 258 259
	if (tun_dst)
		skb_dst_set(skb, &tun_dst->dst);

260
	/* Ignore packet loops (and multicast echo) */
261 262
	if (ether_addr_equal(eth_hdr(skb)->h_source, geneve->dev->dev_addr)) {
		geneve->dev->stats.rx_errors++;
263
		goto drop;
264
	}
265

266
	oiph = skb_network_header(skb);
267 268
	skb_reset_network_header(skb);

269 270
	if (geneve_get_sk_family(gs) == AF_INET)
		err = IP_ECN_decapsulate(oiph, skb);
271
#if IS_ENABLED(CONFIG_IPV6)
272 273
	else
		err = IP6_ECN_decapsulate(oiph, skb);
274
#endif
275 276

	if (unlikely(err)) {
277
		if (log_ecn_error) {
278
			if (geneve_get_sk_family(gs) == AF_INET)
279 280
				net_info_ratelimited("non-ECT from %pI4 "
						     "with TOS=%#x\n",
281 282
						     &((struct iphdr *)oiph)->saddr,
						     ((struct iphdr *)oiph)->tos);
283
#if IS_ENABLED(CONFIG_IPV6)
284
			else
285
				net_info_ratelimited("non-ECT from %pI6\n",
286
						     &((struct ipv6hdr *)oiph)->saddr);
287 288
#endif
		}
289 290 291 292 293 294 295
		if (err > 1) {
			++geneve->dev->stats.rx_frame_errors;
			++geneve->dev->stats.rx_errors;
			goto drop;
		}
	}

296 297 298 299 300 301 302 303 304
	len = skb->len;
	err = gro_cells_receive(&geneve->gro_cells, skb);
	if (likely(err == NET_RX_SUCCESS)) {
		stats = this_cpu_ptr(geneve->dev->tstats);
		u64_stats_update_begin(&stats->syncp);
		stats->rx_packets++;
		stats->rx_bytes += len;
		u64_stats_update_end(&stats->syncp);
	}
305 306 307 308 309 310 311 312 313
	return;
drop:
	/* Consume bad packet */
	kfree_skb(skb);
}

/* Setup stats when device is created */
static int geneve_init(struct net_device *dev)
{
314 315 316
	struct geneve_dev *geneve = netdev_priv(dev);
	int err;

317 318 319
	dev->tstats = netdev_alloc_pcpu_stats(struct pcpu_sw_netstats);
	if (!dev->tstats)
		return -ENOMEM;
320 321 322 323 324 325 326

	err = gro_cells_init(&geneve->gro_cells, dev);
	if (err) {
		free_percpu(dev->tstats);
		return err;
	}

327
	err = dst_cache_init(&geneve->info.dst_cache, GFP_KERNEL);
328 329 330 331 332
	if (err) {
		free_percpu(dev->tstats);
		gro_cells_destroy(&geneve->gro_cells);
		return err;
	}
333 334 335 336 337
	return 0;
}

static void geneve_uninit(struct net_device *dev)
{
338 339
	struct geneve_dev *geneve = netdev_priv(dev);

340
	dst_cache_destroy(&geneve->info.dst_cache);
341
	gro_cells_destroy(&geneve->gro_cells);
342 343 344
	free_percpu(dev->tstats);
}

345 346 347 348
/* Callback from net/ipv4/udp.c to receive packets */
static int geneve_udp_encap_recv(struct sock *sk, struct sk_buff *skb)
{
	struct genevehdr *geneveh;
349
	struct geneve_dev *geneve;
350 351 352
	struct geneve_sock *gs;
	int opts_len;

353
	/* Need UDP and Geneve header to be present */
354
	if (unlikely(!pskb_may_pull(skb, GENEVE_BASE_HLEN)))
355
		goto drop;
356 357 358 359

	/* Return packets with reserved bits set */
	geneveh = geneve_hdr(skb);
	if (unlikely(geneveh->ver != GENEVE_VER))
360
		goto drop;
361 362

	if (unlikely(geneveh->proto_type != htons(ETH_P_TEB)))
363
		goto drop;
364

365 366 367 368 369 370 371 372
	gs = rcu_dereference_sk_user_data(sk);
	if (!gs)
		goto drop;

	geneve = geneve_lookup_skb(gs, skb);
	if (!geneve)
		goto drop;

373 374
	opts_len = geneveh->opt_len * 4;
	if (iptunnel_pull_header(skb, GENEVE_BASE_HLEN + opts_len,
375
				 htons(ETH_P_TEB),
376 377
				 !net_eq(geneve->net, dev_net(geneve->dev)))) {
		geneve->dev->stats.rx_dropped++;
378
		goto drop;
379
	}
380

381
	geneve_rx(geneve, gs, skb);
382 383 384 385 386 387 388 389 390
	return 0;

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

static struct socket *geneve_create_sock(struct net *net, bool ipv6,
391
					 __be16 port, bool ipv6_rx_csum)
392 393 394 395 396 397 398 399 400
{
	struct socket *sock;
	struct udp_port_cfg udp_conf;
	int err;

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

	if (ipv6) {
		udp_conf.family = AF_INET6;
401
		udp_conf.ipv6_v6only = 1;
402
		udp_conf.use_udp6_rx_checksums = ipv6_rx_csum;
403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422
	} else {
		udp_conf.family = AF_INET;
		udp_conf.local_ip.s_addr = htonl(INADDR_ANY);
	}

	udp_conf.local_udp_port = port;

	/* Open UDP socket */
	err = udp_sock_create(net, &udp_conf, &sock);
	if (err < 0)
		return ERR_PTR(err);

	return sock;
}

static int geneve_hlen(struct genevehdr *gh)
{
	return sizeof(*gh) + gh->opt_len * 4;
}

423 424 425
static struct sk_buff *geneve_gro_receive(struct sock *sk,
					  struct list_head *head,
					  struct sk_buff *skb)
426
{
427 428
	struct sk_buff *pp = NULL;
	struct sk_buff *p;
429 430 431 432 433 434 435 436 437 438 439 440 441 442 443 444 445 446 447 448 449 450 451 452 453 454
	struct genevehdr *gh, *gh2;
	unsigned int hlen, gh_len, off_gnv;
	const struct packet_offload *ptype;
	__be16 type;
	int flush = 1;

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

	if (gh->ver != GENEVE_VER || gh->oam)
		goto out;
	gh_len = geneve_hlen(gh);

	hlen = off_gnv + gh_len;
	if (skb_gro_header_hard(skb, hlen)) {
		gh = skb_gro_header_slow(skb, hlen, off_gnv);
		if (unlikely(!gh))
			goto out;
	}

455
	list_for_each_entry(p, head, list) {
456 457 458 459 460 461 462 463 464 465 466 467 468 469 470
		if (!NAPI_GRO_CB(p)->same_flow)
			continue;

		gh2 = (struct genevehdr *)(p->data + off_gnv);
		if (gh->opt_len != gh2->opt_len ||
		    memcmp(gh, gh2, gh_len)) {
			NAPI_GRO_CB(p)->same_flow = 0;
			continue;
		}
	}

	type = gh->proto_type;

	rcu_read_lock();
	ptype = gro_find_receive_by_type(type);
471
	if (!ptype)
472 473 474 475
		goto out_unlock;

	skb_gro_pull(skb, gh_len);
	skb_gro_postpull_rcsum(skb, gh, gh_len);
476
	pp = call_gro_receive(ptype->callbacks.gro_receive, head, skb);
477
	flush = 0;
478 479 480 481

out_unlock:
	rcu_read_unlock();
out:
482
	skb_gro_flush_final(skb, pp, flush);
483 484 485 486

	return pp;
}

487 488
static int geneve_gro_complete(struct sock *sk, struct sk_buff *skb,
			       int nhoff)
489 490 491 492 493 494 495 496 497 498 499 500 501 502 503 504 505
{
	struct genevehdr *gh;
	struct packet_offload *ptype;
	__be16 type;
	int gh_len;
	int err = -ENOSYS;

	gh = (struct genevehdr *)(skb->data + nhoff);
	gh_len = geneve_hlen(gh);
	type = gh->proto_type;

	rcu_read_lock();
	ptype = gro_find_complete_by_type(type);
	if (ptype)
		err = ptype->callbacks.gro_complete(skb, nhoff + gh_len);

	rcu_read_unlock();
506 507 508

	skb_set_inner_mac_header(skb, nhoff + gh_len);

509 510 511 512 513
	return err;
}

/* Create new listen socket if needed */
static struct geneve_sock *geneve_socket_create(struct net *net, __be16 port,
514
						bool ipv6, bool ipv6_rx_csum)
515 516 517 518 519
{
	struct geneve_net *gn = net_generic(net, geneve_net_id);
	struct geneve_sock *gs;
	struct socket *sock;
	struct udp_tunnel_sock_cfg tunnel_cfg;
520
	int h;
521 522 523 524 525

	gs = kzalloc(sizeof(*gs), GFP_KERNEL);
	if (!gs)
		return ERR_PTR(-ENOMEM);

526
	sock = geneve_create_sock(net, ipv6, port, ipv6_rx_csum);
527 528 529 530 531 532 533
	if (IS_ERR(sock)) {
		kfree(gs);
		return ERR_CAST(sock);
	}

	gs->sock = sock;
	gs->refcnt = 1;
534 535
	for (h = 0; h < VNI_HASH_SIZE; ++h)
		INIT_HLIST_HEAD(&gs->vni_list[h]);
536 537

	/* Initialize the geneve udp offloads structure */
538
	udp_tunnel_notify_add_rx_port(gs->sock, UDP_TUNNEL_TYPE_GENEVE);
539 540

	/* Mark socket as an encapsulation socket */
541
	memset(&tunnel_cfg, 0, sizeof(tunnel_cfg));
542 543
	tunnel_cfg.sk_user_data = gs;
	tunnel_cfg.encap_type = 1;
544 545
	tunnel_cfg.gro_receive = geneve_gro_receive;
	tunnel_cfg.gro_complete = geneve_gro_complete;
546 547 548 549 550 551 552
	tunnel_cfg.encap_rcv = geneve_udp_encap_recv;
	tunnel_cfg.encap_destroy = NULL;
	setup_udp_tunnel_sock(net, sock, &tunnel_cfg);
	list_add(&gs->list, &gn->sock_list);
	return gs;
}

553
static void __geneve_sock_release(struct geneve_sock *gs)
554
{
555
	if (!gs || --gs->refcnt)
556 557 558
		return;

	list_del(&gs->list);
559
	udp_tunnel_notify_del_rx_port(gs->sock, UDP_TUNNEL_TYPE_GENEVE);
560 561 562 563
	udp_tunnel_sock_release(gs->sock);
	kfree_rcu(gs, rcu);
}

564 565
static void geneve_sock_release(struct geneve_dev *geneve)
{
566
	struct geneve_sock *gs4 = rtnl_dereference(geneve->sock4);
567
#if IS_ENABLED(CONFIG_IPV6)
568 569 570 571 572 573 574 575 576 577 578
	struct geneve_sock *gs6 = rtnl_dereference(geneve->sock6);

	rcu_assign_pointer(geneve->sock6, NULL);
#endif

	rcu_assign_pointer(geneve->sock4, NULL);
	synchronize_net();

	__geneve_sock_release(gs4);
#if IS_ENABLED(CONFIG_IPV6)
	__geneve_sock_release(gs6);
579 580 581
#endif
}

582
static struct geneve_sock *geneve_find_sock(struct geneve_net *gn,
583
					    sa_family_t family,
584 585 586 587 588 589
					    __be16 dst_port)
{
	struct geneve_sock *gs;

	list_for_each_entry(gs, &gn->sock_list, list) {
		if (inet_sk(gs->sock->sk)->inet_sport == dst_port &&
590
		    geneve_get_sk_family(gs) == family) {
591 592 593 594 595 596
			return gs;
		}
	}
	return NULL;
}

597
static int geneve_sock_add(struct geneve_dev *geneve, bool ipv6)
598 599
{
	struct net *net = geneve->net;
600
	struct geneve_net *gn = net_generic(net, geneve_net_id);
Jiri Benc's avatar
Jiri Benc committed
601
	struct geneve_dev_node *node;
602
	struct geneve_sock *gs;
603
	__u8 vni[3];
604
	__u32 hash;
605

606
	gs = geneve_find_sock(gn, ipv6 ? AF_INET6 : AF_INET, geneve->info.key.tp_dst);
607 608 609 610 611
	if (gs) {
		gs->refcnt++;
		goto out;
	}

612 613
	gs = geneve_socket_create(net, geneve->info.key.tp_dst, ipv6,
				  geneve->use_udp6_rx_checksums);
614 615 616
	if (IS_ERR(gs))
		return PTR_ERR(gs);

617 618
out:
	gs->collect_md = geneve->collect_md;
619
#if IS_ENABLED(CONFIG_IPV6)
Jiri Benc's avatar
Jiri Benc committed
620
	if (ipv6) {
621
		rcu_assign_pointer(geneve->sock6, gs);
Jiri Benc's avatar
Jiri Benc committed
622 623
		node = &geneve->hlist6;
	} else
624
#endif
Jiri Benc's avatar
Jiri Benc committed
625
	{
626
		rcu_assign_pointer(geneve->sock4, gs);
Jiri Benc's avatar
Jiri Benc committed
627 628 629
		node = &geneve->hlist4;
	}
	node->geneve = geneve;
630

631 632
	tunnel_id_to_vni(geneve->info.key.tun_id, vni);
	hash = geneve_net_vni_hash(vni);
Jiri Benc's avatar
Jiri Benc committed
633
	hlist_add_head_rcu(&node->hlist, &gs->vni_list[hash]);
634 635 636
	return 0;
}

637 638 639
static int geneve_open(struct net_device *dev)
{
	struct geneve_dev *geneve = netdev_priv(dev);
640
	bool ipv6 = !!(geneve->info.mode & IP_TUNNEL_INFO_IPV6);
641 642 643 644 645 646 647 648 649 650 651 652 653 654 655
	bool metadata = geneve->collect_md;
	int ret = 0;

#if IS_ENABLED(CONFIG_IPV6)
	if (ipv6 || metadata)
		ret = geneve_sock_add(geneve, true);
#endif
	if (!ret && (!ipv6 || metadata))
		ret = geneve_sock_add(geneve, false);
	if (ret < 0)
		geneve_sock_release(geneve);

	return ret;
}

656 657 658 659
static int geneve_stop(struct net_device *dev)
{
	struct geneve_dev *geneve = netdev_priv(dev);

Jiri Benc's avatar
Jiri Benc committed
660 661 662 663
	hlist_del_init_rcu(&geneve->hlist4.hlist);
#if IS_ENABLED(CONFIG_IPV6)
	hlist_del_init_rcu(&geneve->hlist6.hlist);
#endif
664
	geneve_sock_release(geneve);
665 666 667
	return 0;
}

668
static void geneve_build_header(struct genevehdr *geneveh,
669
				const struct ip_tunnel_info *info)
670 671
{
	geneveh->ver = GENEVE_VER;
672 673 674
	geneveh->opt_len = info->options_len / 4;
	geneveh->oam = !!(info->key.tun_flags & TUNNEL_OAM);
	geneveh->critical = !!(info->key.tun_flags & TUNNEL_CRIT_OPT);
675
	geneveh->rsvd1 = 0;
676
	tunnel_id_to_vni(info->key.tun_id, geneveh->vni);
677 678 679
	geneveh->proto_type = htons(ETH_P_TEB);
	geneveh->rsvd2 = 0;

680 681
	if (info->key.tun_flags & TUNNEL_GENEVE_OPT)
		ip_tunnel_info_opts_get(geneveh->options, info);
682 683
}

684 685 686
static int geneve_build_skb(struct dst_entry *dst, struct sk_buff *skb,
			    const struct ip_tunnel_info *info,
			    bool xnet, int ip_hdr_len)
687
{
688
	bool udp_sum = !!(info->key.tun_flags & TUNNEL_CSUM);
689 690 691 692
	struct genevehdr *gnvh;
	int min_headroom;
	int err;

693
	skb_reset_mac_header(skb);
694 695
	skb_scrub_packet(skb, xnet);

696 697
	min_headroom = LL_RESERVED_SPACE(dst->dev) + dst->header_len +
		       GENEVE_BASE_HLEN + info->options_len + ip_hdr_len;
698
	err = skb_cow_head(skb, min_headroom);
699
	if (unlikely(err))
700 701
		goto free_dst;

702
	err = udp_tunnel_handle_offloads(skb, udp_sum);
703
	if (err)
704 705
		goto free_dst;

706
	gnvh = __skb_push(skb, sizeof(*gnvh) + info->options_len);
707
	geneve_build_header(gnvh, info);
708 709 710 711 712 713 714 715 716 717
	skb_set_inner_protocol(skb, htons(ETH_P_TEB));
	return 0;

free_dst:
	dst_release(dst);
	return err;
}

static struct rtable *geneve_get_v4_rt(struct sk_buff *skb,
				       struct net_device *dev,
718
				       struct geneve_sock *gs4,
719
				       struct flowi4 *fl4,
720
				       const struct ip_tunnel_info *info)
721
{
722
	bool use_cache = ip_tunnel_dst_cache_usable(skb, info);
723
	struct geneve_dev *geneve = netdev_priv(dev);
724
	struct dst_cache *dst_cache;
725 726 727
	struct rtable *rt = NULL;
	__u8 tos;

728
	if (!gs4)
729 730
		return ERR_PTR(-EIO);

731 732 733
	memset(fl4, 0, sizeof(*fl4));
	fl4->flowi4_mark = skb->mark;
	fl4->flowi4_proto = IPPROTO_UDP;
734 735
	fl4->daddr = info->key.u.ipv4.dst;
	fl4->saddr = info->key.u.ipv4.src;
736

737 738 739 740
	tos = info->key.tos;
	if ((tos == 1) && !geneve->collect_md) {
		tos = ip_tunnel_get_dsfield(ip_hdr(skb), skb);
		use_cache = false;
741
	}
742
	fl4->flowi4_tos = RT_TOS(tos);
743

744
	dst_cache = (struct dst_cache *)&info->dst_cache;
745 746 747 748
	if (use_cache) {
		rt = dst_cache_get_ip4(dst_cache, &fl4->saddr);
		if (rt)
			return rt;
749 750 751 752
	}
	rt = ip_route_output_key(geneve->net, fl4);
	if (IS_ERR(rt)) {
		netdev_dbg(dev, "no route to %pI4\n", &fl4->daddr);
753
		return ERR_PTR(-ENETUNREACH);
754 755 756 757
	}
	if (rt->dst.dev == dev) { /* is this necessary? */
		netdev_dbg(dev, "circular route to %pI4\n", &fl4->daddr);
		ip_rt_put(rt);
758
		return ERR_PTR(-ELOOP);
759
	}
760 761
	if (use_cache)
		dst_cache_set_ip4(dst_cache, &rt->dst, fl4->saddr);
762 763 764
	return rt;
}

765 766 767
#if IS_ENABLED(CONFIG_IPV6)
static struct dst_entry *geneve_get_v6_dst(struct sk_buff *skb,
					   struct net_device *dev,
768
					   struct geneve_sock *gs6,
769
					   struct flowi6 *fl6,
770
					   const struct ip_tunnel_info *info)
771
{
772
	bool use_cache = ip_tunnel_dst_cache_usable(skb, info);
773 774
	struct geneve_dev *geneve = netdev_priv(dev);
	struct dst_entry *dst = NULL;
775
	struct dst_cache *dst_cache;
776
	__u8 prio;
777

778 779 780
	if (!gs6)
		return ERR_PTR(-EIO);

781 782 783
	memset(fl6, 0, sizeof(*fl6));
	fl6->flowi6_mark = skb->mark;
	fl6->flowi6_proto = IPPROTO_UDP;
784 785 786 787 788 789
	fl6->daddr = info->key.u.ipv6.dst;
	fl6->saddr = info->key.u.ipv6.src;
	prio = info->key.tos;
	if ((prio == 1) && !geneve->collect_md) {
		prio = ip_tunnel_get_dsfield(ip_hdr(skb), skb);
		use_cache = false;
790 791
	}

792 793
	fl6->flowlabel = ip6_make_flowinfo(RT_TOS(prio),
					   info->key.label);
794
	dst_cache = (struct dst_cache *)&info->dst_cache;
795 796 797 798
	if (use_cache) {
		dst = dst_cache_get_ip6(dst_cache, &fl6->saddr);
		if (dst)
			return dst;
799 800 801 802 803 804 805 806 807 808 809
	}
	if (ipv6_stub->ipv6_dst_lookup(geneve->net, gs6->sock->sk, &dst, fl6)) {
		netdev_dbg(dev, "no route to %pI6\n", &fl6->daddr);
		return ERR_PTR(-ENETUNREACH);
	}
	if (dst->dev == dev) { /* is this necessary? */
		netdev_dbg(dev, "circular route to %pI6\n", &fl6->daddr);
		dst_release(dst);
		return ERR_PTR(-ELOOP);
	}

810 811
	if (use_cache)
		dst_cache_set_ip6(dst_cache, dst, &fl6->saddr);
812 813 814 815
	return dst;
}
#endif

816
static int geneve_xmit_skb(struct sk_buff *skb, struct net_device *dev,
817 818
			   struct geneve_dev *geneve,
			   const struct ip_tunnel_info *info)
819
{
820 821 822 823
	bool xnet = !net_eq(geneve->net, dev_net(geneve->dev));
	struct geneve_sock *gs4 = rcu_dereference(geneve->sock4);
	const struct ip_tunnel_key *key = &info->key;
	struct rtable *rt;
824
	struct flowi4 fl4;
825
	__u8 tos, ttl;
826
	__be16 sport;
827
	__be16 df;
828
	int err;
829

830
	rt = geneve_get_v4_rt(skb, dev, gs4, &fl4, info);
831 832
	if (IS_ERR(rt))
		return PTR_ERR(rt);
833

834 835
	skb_tunnel_check_pmtu(skb, &rt->dst,
			      GENEVE_IPV4_HLEN + info->options_len);
836

837
	sport = udp_flow_src_port(geneve->net, skb, 1, USHRT_MAX, true);
838 839
	if (geneve->collect_md) {
		tos = ip_tunnel_ecn_encap(key->tos, ip_hdr(skb), skb);
840
		ttl = key->ttl;
841
	} else {
842
		tos = ip_tunnel_ecn_encap(fl4.flowi4_tos, ip_hdr(skb), skb);
843 844 845 846 847
		if (geneve->ttl_inherit)
			ttl = ip_tunnel_get_ttl(ip_hdr(skb), skb);
		else
			ttl = key->ttl;
		ttl = ttl ? : ip4_dst_hoplimit(&rt->dst);
848
	}
849
	df = key->tun_flags & TUNNEL_DONT_FRAGMENT ? htons(IP_DF) : 0;
850

851
	err = geneve_build_skb(&rt->dst, skb, info, xnet, sizeof(struct iphdr));
852 853
	if (unlikely(err))
		return err;
854

855 856 857 858 859
	udp_tunnel_xmit_skb(rt, gs4->sock->sk, skb, fl4.saddr, fl4.daddr,
			    tos, ttl, df, sport, geneve->info.key.tp_dst,
			    !net_eq(geneve->net, dev_net(geneve->dev)),
			    !(info->key.tun_flags & TUNNEL_CSUM));
	return 0;
860 861
}

862
#if IS_ENABLED(CONFIG_IPV6)
863
static int geneve6_xmit_skb(struct sk_buff *skb, struct net_device *dev,
864 865
			    struct geneve_dev *geneve,
			    const struct ip_tunnel_info *info)
866
{
867 868 869
	bool xnet = !net_eq(geneve->net, dev_net(geneve->dev));
	struct geneve_sock *gs6 = rcu_dereference(geneve->sock6);
	const struct ip_tunnel_key *key = &info->key;
870 871
	struct dst_entry *dst = NULL;
	struct flowi6 fl6;
872
	__u8 prio, ttl;
873
	__be16 sport;
874
	int err;
875

876
	dst = geneve_get_v6_dst(skb, dev, gs6, &fl6, info);
877 878
	if (IS_ERR(dst))
		return PTR_ERR(dst);
879

880
	skb_tunnel_check_pmtu(skb, dst, GENEVE_IPV6_HLEN + info->options_len);
881

882
	sport = udp_flow_src_port(geneve->net, skb, 1, USHRT_MAX, true);
883 884 885 886 887 888
	if (geneve->collect_md) {
		prio = ip_tunnel_ecn_encap(key->tos, ip_hdr(skb), skb);
		ttl = key->ttl;
	} else {
		prio = ip_tunnel_ecn_encap(ip6_tclass(fl6.flowlabel),
					   ip_hdr(skb), skb);
889 890 891 892 893
		if (geneve->ttl_inherit)
			ttl = ip_tunnel_get_ttl(ip_hdr(skb), skb);
		else
			ttl = key->ttl;
		ttl = ttl ? : ip6_dst_hoplimit(dst);
894
	}
895
	err = geneve_build_skb(dst, skb, info, xnet, sizeof(struct ipv6hdr));
896 897
	if (unlikely(err))
		return err;
898

899 900 901 902 903 904 905
	udp_tunnel6_xmit_skb(dst, gs6->sock->sk, skb, dev,
			     &fl6.saddr, &fl6.daddr, prio, ttl,
			     info->key.label, sport, geneve->info.key.tp_dst,
			     !(info->key.tun_flags & TUNNEL_CSUM));
	return 0;
}
#endif
906

907 908 909 910 911
static netdev_tx_t geneve_xmit(struct sk_buff *skb, struct net_device *dev)
{
	struct geneve_dev *geneve = netdev_priv(dev);
	struct ip_tunnel_info *info = NULL;
	int err;
912

913 914 915 916 917
	if (geneve->collect_md) {
		info = skb_tunnel_info(skb);
		if (unlikely(!info || !(info->mode & IP_TUNNEL_INFO_TX))) {
			err = -EINVAL;
			netdev_dbg(dev, "no tunnel metadata\n");
918
			goto tx_error;
919
		}
920
	} else {
921
		info = &geneve->info;
922
	}
923

924
	rcu_read_lock();
925 926 927 928 929 930
#if IS_ENABLED(CONFIG_IPV6)
	if (info->mode & IP_TUNNEL_INFO_IPV6)
		err = geneve6_xmit_skb(skb, dev, geneve, info);
	else
#endif
		err = geneve_xmit_skb(skb, dev, geneve, info);
931
	rcu_read_unlock();
932

933 934
	if (likely(!err))
		return NETDEV_TX_OK;
935 936
tx_error:
	dev_kfree_skb(skb);
937

938 939 940 941
	if (err == -ELOOP)
		dev->stats.collisions++;
	else if (err == -ENETUNREACH)
		dev->stats.tx_carrier_errors++;
942 943

	dev->stats.tx_errors++;
944 945 946
	return NETDEV_TX_OK;
}

947
static int geneve_change_mtu(struct net_device *dev, int new_mtu)
948
{
949 950
	if (new_mtu > dev->max_mtu)
		new_mtu = dev->max_mtu;
951 952
	else if (new_mtu < dev->min_mtu)
		new_mtu = dev->min_mtu;
David Wragg's avatar
David Wragg committed
953

954 955 956 957
	dev->mtu = new_mtu;
	return 0;
}

958 959 960 961 962
static int geneve_fill_metadata_dst(struct net_device *dev, struct sk_buff *skb)
{
	struct ip_tunnel_info *info = skb_tunnel_info(skb);
	struct geneve_dev *geneve = netdev_priv(dev);

963
	if (ip_tunnel_info_af(info) == AF_INET) {
964 965
		struct rtable *rt;
		struct flowi4 fl4;
966
		struct geneve_sock *gs4 = rcu_dereference(geneve->sock4);
967

968
		rt = geneve_get_v4_rt(skb, dev, gs4, &fl4, info);
969 970
		if (IS_ERR(rt))
			return PTR_ERR(rt);
971

972 973 974 975
		ip_rt_put(rt);
		info->key.u.ipv4.src = fl4.saddr;
#if IS_ENABLED(CONFIG_IPV6)
	} else if (ip_tunnel_info_af(info) == AF_INET6) {
976 977
		struct dst_entry *dst;
		struct flowi6 fl6;
978
		struct geneve_sock *gs6 = rcu_dereference(geneve->sock6);
979

980
		dst = geneve_get_v6_dst(skb, dev, gs6, &fl6, info);
981 982 983 984 985 986 987 988 989
		if (IS_ERR(dst))
			return PTR_ERR(dst);

		dst_release(dst);
		info->key.u.ipv6.src = fl6.saddr;
#endif
	} else {
		return -EINVAL;
	}
990 991 992

	info->key.tp_src = udp_flow_src_port(geneve->net, skb,
					     1, USHRT_MAX, true);
993
	info->key.tp_dst = geneve->info.key.tp_dst;
994 995 996
	return 0;
}

997 998 999 1000 1001 1002 1003
static const struct net_device_ops geneve_netdev_ops = {
	.ndo_init		= geneve_init,
	.ndo_uninit		= geneve_uninit,
	.ndo_open		= geneve_open,
	.ndo_stop		= geneve_stop,
	.ndo_start_xmit		= geneve_xmit,
	.ndo_get_stats64	= ip_tunnel_get_stats64,
1004
	.ndo_change_mtu		= geneve_change_mtu,
1005 1006
	.ndo_validate_addr	= eth_validate_addr,
	.ndo_set_mac_address	= eth_mac_addr,
1007
	.ndo_fill_metadata_dst	= geneve_fill_metadata_dst,
1008 1009 1010 1011 1012 1013 1014 1015 1016 1017 1018 1019 1020 1021 1022 1023 1024 1025 1026
};

static void geneve_get_drvinfo(struct net_device *dev,
			       struct ethtool_drvinfo *drvinfo)
{
	strlcpy(drvinfo->version, GENEVE_NETDEV_VER, sizeof(drvinfo->version));
	strlcpy(drvinfo->driver, "geneve", sizeof(drvinfo->driver));
}

static const struct ethtool_ops geneve_ethtool_ops = {
	.get_drvinfo	= geneve_get_drvinfo,
	.get_link	= ethtool_op_get_link,
};

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

1027
/* Calls the ndo_udp_tunnel_add of the caller in order to
1028
 * supply the listening GENEVE udp ports. Callers are expected
1029
 * to implement the ndo_udp_tunnel_add.
1030
 */
1031
static void geneve_offload_rx_ports(struct net_device *dev, bool push)
1032 1033 1034 1035
{
	struct net *net = dev_net(dev);
	struct geneve_net *gn = net_generic(net, geneve_net_id);
	struct geneve_sock *gs;
1036

1037
	rcu_read_lock();
1038 1039 1040 1041 1042 1043 1044 1045 1046
	list_for_each_entry_rcu(gs, &gn->sock_list, list) {
		if (push) {
			udp_tunnel_push_rx_port(dev, gs->sock,
						UDP_TUNNEL_TYPE_GENEVE);
		} else {
			udp_tunnel_drop_rx_port(dev, gs->sock,
						UDP_TUNNEL_TYPE_GENEVE);
		}
	}
1047 1048 1049
	rcu_read_unlock();
}

1050 1051 1052 1053 1054 1055 1056
/* Initialize the device structure. */
static void geneve_setup(struct net_device *dev)
{
	ether_setup(dev);

	dev->netdev_ops = &geneve_netdev_ops;
	dev->ethtool_ops = &geneve_ethtool_ops;
1057
	dev->needs_free_netdev = true;
1058 1059 1060 1061 1062 1063 1064 1065 1066 1067 1068

	SET_NETDEV_DEVTYPE(dev, &geneve_type);

	dev->features    |= NETIF_F_LLTX;
	dev->features    |= NETIF_F_SG | NETIF_F_HW_CSUM;
	dev->features    |= NETIF_F_RXCSUM;
	dev->features    |= NETIF_F_GSO_SOFTWARE;

	dev->hw_features |= NETIF_F_SG | NETIF_F_HW_CSUM | NETIF_F_RXCSUM;
	dev->hw_features |= NETIF_F_GSO_SOFTWARE;

1069 1070 1071 1072 1073 1074 1075 1076
	/* MTU range: 68 - (something less than 65535) */
	dev->min_mtu = ETH_MIN_MTU;
	/* The max_mtu calculation does not take account of GENEVE
	 * options, to avoid excluding potentially valid
	 * configurations. This will be further reduced by IPvX hdr size.
	 */
	dev->max_mtu = IP_MAX_MTU - GENEVE_BASE_HLEN - dev->hard_header_len;

1077
	netif_keep_dst(dev);
1078
	dev->priv_flags &= ~IFF_TX_SKB_SHARING;
1079
	dev->priv_flags |= IFF_LIVE_ADDR_CHANGE | IFF_NO_QUEUE;
1080
	eth_hw_addr_random(dev);
1081 1082 1083 1084 1085
}

static const struct nla_policy geneve_policy[IFLA_GENEVE_MAX + 1] = {
	[IFLA_GENEVE_ID]		= { .type = NLA_U32 },
	[IFLA_GENEVE_REMOTE]		= { .len = FIELD_SIZEOF(struct iphdr, daddr) },
1086
	[IFLA_GENEVE_REMOTE6]		= { .len = sizeof(struct in6_addr) },
1087
	[IFLA_GENEVE_TTL]		= { .type = NLA_U8 },
1088
	[IFLA_GENEVE_TOS]		= { .type = NLA_U8 },
1089
	[IFLA_GENEVE_LABEL]		= { .type = NLA_U32 },
1090
	[IFLA_GENEVE_PORT]		= { .type = NLA_U16 },
1091
	[IFLA_GENEVE_COLLECT_METADATA]	= { .type = NLA_FLAG },
1092 1093 1094
	[IFLA_GENEVE_UDP_CSUM]		= { .type = NLA_U8 },
	[IFLA_GENEVE_UDP_ZERO_CSUM6_TX]	= { .type = NLA_U8 },
	[IFLA_GENEVE_UDP_ZERO_CSUM6_RX]	= { .type = NLA_U8 },
1095
	[IFLA_GENEVE_TTL_INHERIT]	= { .type = NLA_U8 },
1096 1097
};

1098 1099
static int geneve_validate(struct nlattr *tb[], struct nlattr *data[],
			   struct netlink_ext_ack *extack)
1100 1101
{
	if (tb[IFLA_ADDRESS]) {
1102 1103 1104
		if (nla_len(tb[IFLA_ADDRESS]) != ETH_ALEN) {
			NL_SET_ERR_MSG_ATTR(extack, tb[IFLA_ADDRESS],
					    "Provided link layer address is not Ethernet");
1105
			return -EINVAL;
1106
		}
1107

1108 1109 1110
		if (!is_valid_ether_addr(nla_data(tb[IFLA_ADDRESS]))) {
			NL_SET_ERR_MSG_ATTR(extack, tb[IFLA_ADDRESS],
					    "Provided Ethernet address is not unicast");
1111
			return -EADDRNOTAVAIL;
1112
		}
1113 1114
	}

1115 1116 1117
	if (!data) {
		NL_SET_ERR_MSG(extack,
			       "Not enough attributes provided to perform the operation");
1118
		return -EINVAL;
1119
	}
1120 1121 1122 1123

	if (data[IFLA_GENEVE_ID]) {
		__u32 vni =  nla_get_u32(data[IFLA_GENEVE_ID]);

1124 1125 1126
		if (vni >= GENEVE_N_VID) {
			NL_SET_ERR_MSG_ATTR(extack, data[IFLA_GENEVE_ID],
					    "Geneve ID must be lower than 16777216");
1127
			return -ERANGE;
1128
		}
1129 1130 1131 1132 1133
	}

	return 0;
}

1134
static struct geneve_dev *geneve_find_dev(struct geneve_net *gn,
1135
					  const struct ip_tunnel_info *info,
1136 1137 1138
					  bool *tun_on_same_port,
					  bool *tun_collect_md)
{
1139
	struct geneve_dev *geneve, *t = NULL;
1140 1141 1142 1143

	*tun_on_same_port = false;
	*tun_collect_md = false;
	list_for_each_entry(geneve, &gn->geneve_list, next) {
1144
		if (info->key.tp_dst == geneve->info.key.tp_dst) {
1145 1146 1147
			*tun_collect_md = geneve->collect_md;
			*tun_on_same_port = true;
		}
1148 1149 1150
		if (info->key.tun_id == geneve->info.key.tun_id &&
		    info->key.tp_dst == geneve->info.key.tp_dst &&
		    !memcmp(&info->key.u, &geneve->info.key.u, sizeof(info->key.u)))
1151 1152 1153 1154 1155
			t = geneve;
	}
	return t;
}

1156 1157
static bool is_tnl_info_zero(const struct ip_tunnel_info *info)
{
1158 1159 1160
	return !(info->key.tun_id || info->key.tun_flags || info->key.tos ||
		 info->key.ttl || info->key.label || info->key.tp_src ||
		 memchr_inv(&info->key.u, 0, sizeof(info->key.u)));
1161 1162
}

1163 1164 1165 1166 1167 1168 1169 1170 1171
static bool geneve_dst_addr_equal(struct ip_tunnel_info *a,
				  struct ip_tunnel_info *b)
{
	if (ip_tunnel_info_af(a) == AF_INET)
		return a->key.u.ipv4.dst == b->key.u.ipv4.dst;
	else
		return ipv6_addr_equal(&a->key.u.ipv6.dst, &b->key.u.ipv6.dst);
}

1172
static int geneve_configure(struct net *net, struct net_device *dev,
1173
			    struct netlink_ext_ack *extack,
1174
			    const struct ip_tunnel_info *info,
1175 1176
			    bool metadata, bool ipv6_rx_csum,
			    bool ttl_inherit)
1177 1178
{
	struct geneve_net *gn = net_generic(net, geneve_net_id);
1179 1180
	struct geneve_dev *t, *geneve = netdev_priv(dev);
	bool tun_collect_md, tun_on_same_port;
1181
	int err, encap_len;
1182

1183 1184 1185
	if (metadata && !is_tnl_info_zero(info)) {
		NL_SET_ERR_MSG(extack,
			       "Device is externally controlled, so attributes (VNI, Port, and so on) must not be specified");
1186
		return -EINVAL;
1187
	}
1188 1189 1190 1191

	geneve->net = net;
	geneve->dev = dev;

1192
	t = geneve_find_dev(gn, info, &tun_on_same_port, &tun_collect_md);
1193 1194 1195
	if (t)
		return -EBUSY;

1196 1197
	/* make enough headroom for basic scenario */
	encap_len = GENEVE_BASE_HLEN + ETH_HLEN;
1198
	if (!metadata && ip_tunnel_info_af(info) == AF_INET) {
1199
		encap_len += sizeof(struct iphdr);
1200 1201
		dev->max_mtu -= sizeof(struct iphdr);
	} else {
1202
		encap_len += sizeof(struct ipv6hdr);
1203 1204
		dev->max_mtu -= sizeof(struct ipv6hdr);
	}
1205 1206
	dev->needed_headroom = encap_len + ETH_HLEN;

1207
	if (metadata) {
1208 1209 1210
		if (tun_on_same_port) {
			NL_SET_ERR_MSG(extack,
				       "There can be only one externally controlled device on a destination port");
1211
			return -EPERM;
1212
		}
1213
	} else {
1214 1215 1216
		if (tun_collect_md) {
			NL_SET_ERR_MSG(extack,
				       "There already exists an externally controlled device on this destination port");
1217
			return -EPERM;
1218
		}
1219 1220
	}

1221 1222 1223 1224
	dst_cache_reset(&geneve->info.dst_cache);
	geneve->info = *info;
	geneve->collect_md = metadata;
	geneve->use_udp6_rx_checksums = ipv6_rx_csum;
1225
	geneve->ttl_inherit = ttl_inherit;
1226

1227 1228 1229 1230
	err = register_netdevice(dev);
	if (err)
		return err;

1231 1232 1233 1234
	list_add(&geneve->next, &gn->geneve_list);
	return 0;
}

1235 1236 1237 1238 1239 1240
static void init_tnl_info(struct ip_tunnel_info *info, __u16 dst_port)
{
	memset(info, 0, sizeof(*info));
	info->key.tp_dst = htons(dst_port);
}

1241 1242 1243
static int geneve_nl2info(struct nlattr *tb[], struct nlattr *data[],
			  struct netlink_ext_ack *extack,
			  struct ip_tunnel_info *info, bool *metadata,
1244 1245
			  bool *use_udp6_rx_checksums, bool *ttl_inherit,
			  bool changelink)
1246
{
Girish Moodalbail's avatar