datapath.c 45.7 KB
Newer Older
1
/*
2
 * Copyright (c) 2007-2013 Nicira, Inc.
3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46
 *
 * This program is free software; you can redistribute it and/or
 * modify it under the terms of version 2 of the GNU General Public
 * License as published by the Free Software Foundation.
 *
 * This program is distributed in the hope that it will be useful, but
 * WITHOUT ANY WARRANTY; without even the implied warranty of
 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
 * General Public License for more details.
 *
 * You should have received a copy of the GNU General Public License
 * along with this program; if not, write to the Free Software
 * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA
 * 02110-1301, USA
 */

#define pr_fmt(fmt) KBUILD_MODNAME ": " fmt

#include <linux/init.h>
#include <linux/module.h>
#include <linux/if_arp.h>
#include <linux/if_vlan.h>
#include <linux/in.h>
#include <linux/ip.h>
#include <linux/jhash.h>
#include <linux/delay.h>
#include <linux/time.h>
#include <linux/etherdevice.h>
#include <linux/genetlink.h>
#include <linux/kernel.h>
#include <linux/kthread.h>
#include <linux/mutex.h>
#include <linux/percpu.h>
#include <linux/rcupdate.h>
#include <linux/tcp.h>
#include <linux/udp.h>
#include <linux/ethtool.h>
#include <linux/wait.h>
#include <asm/div64.h>
#include <linux/highmem.h>
#include <linux/netfilter_bridge.h>
#include <linux/netfilter_ipv4.h>
#include <linux/inetdevice.h>
#include <linux/list.h>
47
#include <linux/lockdep.h>
48 49 50 51 52
#include <linux/openvswitch.h>
#include <linux/rculist.h>
#include <linux/dmi.h>
#include <linux/workqueue.h>
#include <net/genetlink.h>
53 54
#include <net/net_namespace.h>
#include <net/netns/generic.h>
55 56 57

#include "datapath.h"
#include "flow.h"
58
#include "flow_netlink.h"
59
#include "vport-internal_dev.h"
60
#include "vport-netdev.h"
61

62 63
int ovs_net_id __read_mostly;

64
static void ovs_notify(struct genl_family *family,
65
		       struct sk_buff *skb, struct genl_info *info)
66
{
67
	genl_notify(family, skb, genl_info_net(info), info->snd_portid,
68
		    0, info->nlhdr, GFP_KERNEL);
69 70
}

71 72 73
/**
 * DOC: Locking:
 *
74 75 76 77
 * All writes e.g. Writes to device state (add/remove datapath, port, set
 * operations on vports, etc.), Writes to other state (flow table
 * modifications, set miscellaneous datapath parameters, etc.) are protected
 * by ovs_lock.
78 79 80 81 82 83
 *
 * Reads are protected by RCU.
 *
 * There are a few special cases (mostly stats) that have their own
 * synchronization but they nest under all of above and don't interact with
 * each other.
84 85
 *
 * The RTNL lock nests inside ovs_mutex.
86 87
 */

88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109
static DEFINE_MUTEX(ovs_mutex);

void ovs_lock(void)
{
	mutex_lock(&ovs_mutex);
}

void ovs_unlock(void)
{
	mutex_unlock(&ovs_mutex);
}

#ifdef CONFIG_LOCKDEP
int lockdep_ovsl_is_held(void)
{
	if (debug_locks)
		return lockdep_is_held(&ovs_mutex);
	else
		return 1;
}
#endif

110
static struct vport *new_vport(const struct vport_parms *);
111
static int queue_gso_packets(struct net *, int dp_ifindex, struct sk_buff *,
112
			     const struct dp_upcall_info *);
113 114
static int queue_userspace_packet(struct net *, int dp_ifindex,
				  struct sk_buff *,
115 116
				  const struct dp_upcall_info *);

117
/* Must be called with rcu_read_lock or ovs_mutex. */
118
static struct datapath *get_dp(struct net *net, int dp_ifindex)
119 120 121 122 123
{
	struct datapath *dp = NULL;
	struct net_device *dev;

	rcu_read_lock();
124
	dev = dev_get_by_index_rcu(net, dp_ifindex);
125 126 127 128 129 130 131 132 133 134
	if (dev) {
		struct vport *vport = ovs_internal_dev_get_vport(dev);
		if (vport)
			dp = vport->dp;
	}
	rcu_read_unlock();

	return dp;
}

135
/* Must be called with rcu_read_lock or ovs_mutex. */
136 137
const char *ovs_dp_name(const struct datapath *dp)
{
138
	struct vport *vport = ovs_vport_ovsl_rcu(dp, OVSP_LOCAL);
139 140 141 142 143 144 145 146 147 148
	return vport->ops->get_name(vport);
}

static int get_dpifindex(struct datapath *dp)
{
	struct vport *local;
	int ifindex;

	rcu_read_lock();

149
	local = ovs_vport_rcu(dp, OVSP_LOCAL);
150
	if (local)
151
		ifindex = netdev_vport_priv(local)->dev->ifindex;
152 153 154 155 156 157 158 159 160 161 162 163
	else
		ifindex = 0;

	rcu_read_unlock();

	return ifindex;
}

static void destroy_dp_rcu(struct rcu_head *rcu)
{
	struct datapath *dp = container_of(rcu, struct datapath, rcu);

164
	ovs_flow_tbl_destroy(&dp->table);
165
	free_percpu(dp->stats_percpu);
166
	release_net(ovs_dp_get_net(dp));
167
	kfree(dp->ports);
168 169 170
	kfree(dp);
}

171 172 173 174 175 176 177 178 179 180 181 182
static struct hlist_head *vport_hash_bucket(const struct datapath *dp,
					    u16 port_no)
{
	return &dp->ports[port_no & (DP_VPORT_HASH_BUCKETS - 1)];
}

struct vport *ovs_lookup_vport(const struct datapath *dp, u16 port_no)
{
	struct vport *vport;
	struct hlist_head *head;

	head = vport_hash_bucket(dp, port_no);
183
	hlist_for_each_entry_rcu(vport, head, dp_hash_node) {
184 185 186 187 188 189
		if (vport->port_no == port_no)
			return vport;
	}
	return NULL;
}

190
/* Called with ovs_mutex. */
191 192 193 194 195 196 197
static struct vport *new_vport(const struct vport_parms *parms)
{
	struct vport *vport;

	vport = ovs_vport_add(parms);
	if (!IS_ERR(vport)) {
		struct datapath *dp = parms->dp;
198
		struct hlist_head *head = vport_hash_bucket(dp, vport->port_no);
199

200
		hlist_add_head_rcu(&vport->dp_hash_node, head);
201 202 203 204 205 206
	}
	return vport;
}

void ovs_dp_detach_port(struct vport *p)
{
207
	ASSERT_OVSL();
208 209

	/* First drop references to device. */
210
	hlist_del_rcu(&p->dp_hash_node);
211 212 213 214 215 216 217 218 219 220 221 222 223

	/* Then destroy it. */
	ovs_vport_del(p);
}

/* Must be called with rcu_read_lock. */
void ovs_dp_process_received_packet(struct vport *p, struct sk_buff *skb)
{
	struct datapath *dp = p->dp;
	struct sw_flow *flow;
	struct dp_stats_percpu *stats;
	struct sw_flow_key key;
	u64 *stats_counter;
224
	u32 n_mask_hit;
225 226
	int error;

227
	stats = this_cpu_ptr(dp->stats_percpu);
228 229

	/* Extract flow from 'skb' into 'key'. */
230
	error = ovs_flow_extract(skb, p->port_no, &key);
231 232 233 234 235 236
	if (unlikely(error)) {
		kfree_skb(skb);
		return;
	}

	/* Look up flow. */
237
	flow = ovs_flow_tbl_lookup(&dp->table, &key, &n_mask_hit);
238 239 240 241 242 243
	if (unlikely(!flow)) {
		struct dp_upcall_info upcall;

		upcall.cmd = OVS_PACKET_CMD_MISS;
		upcall.key = &key;
		upcall.userdata = NULL;
244
		upcall.portid = p->upcall_portid;
245 246 247 248 249 250 251
		ovs_dp_upcall(dp, skb, &upcall);
		consume_skb(skb);
		stats_counter = &stats->n_missed;
		goto out;
	}

	OVS_CB(skb)->flow = flow;
252
	OVS_CB(skb)->pkt_key = &key;
253 254 255 256 257 258 259 260 261

	stats_counter = &stats->n_hit;
	ovs_flow_used(OVS_CB(skb)->flow, skb);
	ovs_execute_actions(dp, skb);

out:
	/* Update datapath statistics. */
	u64_stats_update_begin(&stats->sync);
	(*stats_counter)++;
262
	stats->n_mask_hit += n_mask_hit;
263 264 265 266 267 268 269 270
	u64_stats_update_end(&stats->sync);
}

static struct genl_family dp_packet_genl_family = {
	.id = GENL_ID_GENERATE,
	.hdrsize = sizeof(struct ovs_header),
	.name = OVS_PACKET_FAMILY,
	.version = OVS_PACKET_VERSION,
271
	.maxattr = OVS_PACKET_ATTR_MAX,
272 273
	.netnsok = true,
	.parallel_ops = true,
274 275 276
};

int ovs_dp_upcall(struct datapath *dp, struct sk_buff *skb,
277
		  const struct dp_upcall_info *upcall_info)
278 279 280 281 282
{
	struct dp_stats_percpu *stats;
	int dp_ifindex;
	int err;

283
	if (upcall_info->portid == 0) {
284 285 286 287 288 289 290 291 292 293 294
		err = -ENOTCONN;
		goto err;
	}

	dp_ifindex = get_dpifindex(dp);
	if (!dp_ifindex) {
		err = -ENODEV;
		goto err;
	}

	if (!skb_is_gso(skb))
295
		err = queue_userspace_packet(ovs_dp_get_net(dp), dp_ifindex, skb, upcall_info);
296
	else
297
		err = queue_gso_packets(ovs_dp_get_net(dp), dp_ifindex, skb, upcall_info);
298 299 300 301 302 303
	if (err)
		goto err;

	return 0;

err:
304
	stats = this_cpu_ptr(dp->stats_percpu);
305 306 307 308 309 310 311 312

	u64_stats_update_begin(&stats->sync);
	stats->n_lost++;
	u64_stats_update_end(&stats->sync);

	return err;
}

313 314
static int queue_gso_packets(struct net *net, int dp_ifindex,
			     struct sk_buff *skb,
315 316
			     const struct dp_upcall_info *upcall_info)
{
317
	unsigned short gso_type = skb_shinfo(skb)->gso_type;
318 319 320 321 322
	struct dp_upcall_info later_info;
	struct sw_flow_key later_key;
	struct sk_buff *segs, *nskb;
	int err;

323
	segs = __skb_gso_segment(skb, NETIF_F_SG | NETIF_F_HW_CSUM, false);
324 325
	if (IS_ERR(segs))
		return PTR_ERR(segs);
326 327 328 329

	/* Queue all of the segments. */
	skb = segs;
	do {
330
		err = queue_userspace_packet(net, dp_ifindex, skb, upcall_info);
331 332 333
		if (err)
			break;

334
		if (skb == segs && gso_type & SKB_GSO_UDP) {
335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359
			/* The initial flow key extracted by ovs_flow_extract()
			 * in this case is for a first fragment, so we need to
			 * properly mark later fragments.
			 */
			later_key = *upcall_info->key;
			later_key.ip.frag = OVS_FRAG_TYPE_LATER;

			later_info = *upcall_info;
			later_info.key = &later_key;
			upcall_info = &later_info;
		}
	} while ((skb = skb->next));

	/* Free all of the segments. */
	skb = segs;
	do {
		nskb = skb->next;
		if (err)
			kfree_skb(skb);
		else
			consume_skb(skb);
	} while ((skb = nskb));
	return err;
}

360 361 362
static size_t key_attr_size(void)
{
	return    nla_total_size(4)   /* OVS_KEY_ATTR_PRIORITY */
363 364 365 366 367 368 369 370
		+ nla_total_size(0)   /* OVS_KEY_ATTR_TUNNEL */
		  + nla_total_size(8)   /* OVS_TUNNEL_KEY_ATTR_ID */
		  + nla_total_size(4)   /* OVS_TUNNEL_KEY_ATTR_IPV4_SRC */
		  + nla_total_size(4)   /* OVS_TUNNEL_KEY_ATTR_IPV4_DST */
		  + nla_total_size(1)   /* OVS_TUNNEL_KEY_ATTR_TOS */
		  + nla_total_size(1)   /* OVS_TUNNEL_KEY_ATTR_TTL */
		  + nla_total_size(0)   /* OVS_TUNNEL_KEY_ATTR_DONT_FRAGMENT */
		  + nla_total_size(0)   /* OVS_TUNNEL_KEY_ATTR_CSUM */
371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396
		+ nla_total_size(4)   /* OVS_KEY_ATTR_IN_PORT */
		+ nla_total_size(4)   /* OVS_KEY_ATTR_SKB_MARK */
		+ nla_total_size(12)  /* OVS_KEY_ATTR_ETHERNET */
		+ nla_total_size(2)   /* OVS_KEY_ATTR_ETHERTYPE */
		+ nla_total_size(4)   /* OVS_KEY_ATTR_8021Q */
		+ nla_total_size(0)   /* OVS_KEY_ATTR_ENCAP */
		+ nla_total_size(2)   /* OVS_KEY_ATTR_ETHERTYPE */
		+ nla_total_size(40)  /* OVS_KEY_ATTR_IPV6 */
		+ nla_total_size(2)   /* OVS_KEY_ATTR_ICMPV6 */
		+ nla_total_size(28); /* OVS_KEY_ATTR_ND */
}

static size_t upcall_msg_size(const struct sk_buff *skb,
			      const struct nlattr *userdata)
{
	size_t size = NLMSG_ALIGN(sizeof(struct ovs_header))
		+ nla_total_size(skb->len) /* OVS_PACKET_ATTR_PACKET */
		+ nla_total_size(key_attr_size()); /* OVS_PACKET_ATTR_KEY */

	/* OVS_PACKET_ATTR_USERDATA */
	if (userdata)
		size += NLA_ALIGN(userdata->nla_len);

	return size;
}

397 398
static int queue_userspace_packet(struct net *net, int dp_ifindex,
				  struct sk_buff *skb,
399 400 401 402 403 404 405 406 407 408 409 410 411
				  const struct dp_upcall_info *upcall_info)
{
	struct ovs_header *upcall;
	struct sk_buff *nskb = NULL;
	struct sk_buff *user_skb; /* to be queued to userspace */
	struct nlattr *nla;
	int err;

	if (vlan_tx_tag_present(skb)) {
		nskb = skb_clone(skb, GFP_ATOMIC);
		if (!nskb)
			return -ENOMEM;

412
		nskb = __vlan_put_tag(nskb, nskb->vlan_proto, vlan_tx_tag_get(nskb));
413
		if (!nskb)
414 415 416 417 418 419 420 421 422 423 424
			return -ENOMEM;

		nskb->vlan_tci = 0;
		skb = nskb;
	}

	if (nla_attr_size(skb->len) > USHRT_MAX) {
		err = -EFBIG;
		goto out;
	}

425
	user_skb = genlmsg_new(upcall_msg_size(skb, upcall_info->userdata), GFP_ATOMIC);
426 427 428 429 430 431 432 433 434 435
	if (!user_skb) {
		err = -ENOMEM;
		goto out;
	}

	upcall = genlmsg_put(user_skb, 0, 0, &dp_packet_genl_family,
			     0, upcall_info->cmd);
	upcall->dp_ifindex = dp_ifindex;

	nla = nla_nest_start(user_skb, OVS_PACKET_ATTR_KEY);
436
	ovs_nla_put_flow(upcall_info->key, upcall_info->key, user_skb);
437 438 439
	nla_nest_end(user_skb, nla);

	if (upcall_info->userdata)
440 441 442
		__nla_put(user_skb, OVS_PACKET_ATTR_USERDATA,
			  nla_len(upcall_info->userdata),
			  nla_data(upcall_info->userdata));
443 444 445 446 447

	nla = __nla_reserve(user_skb, OVS_PACKET_ATTR_PACKET, skb->len);

	skb_copy_and_csum_dev(skb, nla_data(nla));

448
	genlmsg_end(user_skb, upcall);
449
	err = genlmsg_unicast(net, user_skb, upcall_info->portid);
450 451 452 453 454 455 456 457 458 459 460 461 462 463 464 465 466 467 468 469 470 471 472 473 474 475 476 477

out:
	kfree_skb(nskb);
	return err;
}

static void clear_stats(struct sw_flow *flow)
{
	flow->used = 0;
	flow->tcp_flags = 0;
	flow->packet_count = 0;
	flow->byte_count = 0;
}

static int ovs_packet_cmd_execute(struct sk_buff *skb, struct genl_info *info)
{
	struct ovs_header *ovs_header = info->userhdr;
	struct nlattr **a = info->attrs;
	struct sw_flow_actions *acts;
	struct sk_buff *packet;
	struct sw_flow *flow;
	struct datapath *dp;
	struct ethhdr *eth;
	int len;
	int err;

	err = -EINVAL;
	if (!a[OVS_PACKET_ATTR_PACKET] || !a[OVS_PACKET_ATTR_KEY] ||
478
	    !a[OVS_PACKET_ATTR_ACTIONS])
479 480 481 482 483 484 485 486 487
		goto err;

	len = nla_len(a[OVS_PACKET_ATTR_PACKET]);
	packet = __dev_alloc_skb(NET_IP_ALIGN + len, GFP_KERNEL);
	err = -ENOMEM;
	if (!packet)
		goto err;
	skb_reserve(packet, NET_IP_ALIGN);

488
	nla_memcpy(__skb_put(packet, len), a[OVS_PACKET_ATTR_PACKET], len);
489 490 491 492 493 494 495

	skb_reset_mac_header(packet);
	eth = eth_hdr(packet);

	/* Normally, setting the skb 'protocol' field would be handled by a
	 * call to eth_type_trans(), but it assumes there's a sending
	 * device, which we may not have. */
Simon Horman's avatar
Simon Horman committed
496
	if (ntohs(eth->h_proto) >= ETH_P_802_3_MIN)
497 498 499 500 501 502 503 504 505 506
		packet->protocol = eth->h_proto;
	else
		packet->protocol = htons(ETH_P_802_2);

	/* Build an sw_flow for sending this packet. */
	flow = ovs_flow_alloc();
	err = PTR_ERR(flow);
	if (IS_ERR(flow))
		goto err_kfree_skb;

507
	err = ovs_flow_extract(packet, -1, &flow->key);
508 509 510
	if (err)
		goto err_flow_free;

511
	err = ovs_nla_get_flow_metadata(flow, a[OVS_PACKET_ATTR_KEY]);
512 513
	if (err)
		goto err_flow_free;
514
	acts = ovs_nla_alloc_flow_actions(nla_len(a[OVS_PACKET_ATTR_ACTIONS]));
515 516 517
	err = PTR_ERR(acts);
	if (IS_ERR(acts))
		goto err_flow_free;
518

519 520
	err = ovs_nla_copy_actions(a[OVS_PACKET_ATTR_ACTIONS],
				   &flow->key, 0, &acts);
521
	rcu_assign_pointer(flow->sf_acts, acts);
522 523
	if (err)
		goto err_flow_free;
524 525

	OVS_CB(packet)->flow = flow;
526
	OVS_CB(packet)->pkt_key = &flow->key;
527
	packet->priority = flow->key.phy.priority;
528
	packet->mark = flow->key.phy.skb_mark;
529 530

	rcu_read_lock();
531
	dp = get_dp(sock_net(skb->sk), ovs_header->dp_ifindex);
532 533 534 535 536 537 538 539 540
	err = -ENODEV;
	if (!dp)
		goto err_unlock;

	local_bh_disable();
	err = ovs_execute_actions(dp, packet);
	local_bh_enable();
	rcu_read_unlock();

541
	ovs_flow_free(flow, false);
542 543 544 545 546
	return err;

err_unlock:
	rcu_read_unlock();
err_flow_free:
547
	ovs_flow_free(flow, false);
548 549 550 551 552 553 554
err_kfree_skb:
	kfree_skb(packet);
err:
	return err;
}

static const struct nla_policy packet_policy[OVS_PACKET_ATTR_MAX + 1] = {
555
	[OVS_PACKET_ATTR_PACKET] = { .len = ETH_HLEN },
556 557 558 559
	[OVS_PACKET_ATTR_KEY] = { .type = NLA_NESTED },
	[OVS_PACKET_ATTR_ACTIONS] = { .type = NLA_NESTED },
};

560
static const struct genl_ops dp_packet_genl_ops[] = {
561 562 563 564 565 566 567
	{ .cmd = OVS_PACKET_CMD_EXECUTE,
	  .flags = GENL_ADMIN_PERM, /* Requires CAP_NET_ADMIN privilege. */
	  .policy = packet_policy,
	  .doit = ovs_packet_cmd_execute
	}
};

568 569
static void get_dp_stats(struct datapath *dp, struct ovs_dp_stats *stats,
			 struct ovs_dp_megaflow_stats *mega_stats)
570 571 572
{
	int i;

573 574
	memset(mega_stats, 0, sizeof(*mega_stats));

575
	stats->n_flows = ovs_flow_tbl_count(&dp->table);
576
	mega_stats->n_masks = ovs_flow_tbl_num_masks(&dp->table);
577 578

	stats->n_hit = stats->n_missed = stats->n_lost = 0;
579

580 581 582 583 584 585 586 587 588 589 590 591 592 593 594
	for_each_possible_cpu(i) {
		const struct dp_stats_percpu *percpu_stats;
		struct dp_stats_percpu local_stats;
		unsigned int start;

		percpu_stats = per_cpu_ptr(dp->stats_percpu, i);

		do {
			start = u64_stats_fetch_begin_bh(&percpu_stats->sync);
			local_stats = *percpu_stats;
		} while (u64_stats_fetch_retry_bh(&percpu_stats->sync, start));

		stats->n_hit += local_stats.n_hit;
		stats->n_missed += local_stats.n_missed;
		stats->n_lost += local_stats.n_lost;
595
		mega_stats->n_mask_hit += local_stats.n_mask_hit;
596 597 598 599 600 601 602 603 604 605 606 607 608 609
	}
}

static const struct nla_policy flow_policy[OVS_FLOW_ATTR_MAX + 1] = {
	[OVS_FLOW_ATTR_KEY] = { .type = NLA_NESTED },
	[OVS_FLOW_ATTR_ACTIONS] = { .type = NLA_NESTED },
	[OVS_FLOW_ATTR_CLEAR] = { .type = NLA_FLAG },
};

static struct genl_family dp_flow_genl_family = {
	.id = GENL_ID_GENERATE,
	.hdrsize = sizeof(struct ovs_header),
	.name = OVS_FLOW_FAMILY,
	.version = OVS_FLOW_VERSION,
610
	.maxattr = OVS_FLOW_ATTR_MAX,
611 612
	.netnsok = true,
	.parallel_ops = true,
613 614 615 616 617 618
};

static struct genl_multicast_group ovs_dp_flow_multicast_group = {
	.name = OVS_FLOW_MCGROUP
};

619 620 621 622
static size_t ovs_flow_cmd_msg_size(const struct sw_flow_actions *acts)
{
	return NLMSG_ALIGN(sizeof(struct ovs_header))
		+ nla_total_size(key_attr_size()) /* OVS_FLOW_ATTR_KEY */
623
		+ nla_total_size(key_attr_size()) /* OVS_FLOW_ATTR_MASK */
624 625 626 627 628 629
		+ nla_total_size(sizeof(struct ovs_flow_stats)) /* OVS_FLOW_ATTR_STATS */
		+ nla_total_size(1) /* OVS_FLOW_ATTR_TCP_FLAGS */
		+ nla_total_size(8) /* OVS_FLOW_ATTR_USED */
		+ nla_total_size(acts->actions_len); /* OVS_FLOW_ATTR_ACTIONS */
}

630
/* Called with ovs_mutex. */
631
static int ovs_flow_cmd_fill_info(struct sw_flow *flow, struct datapath *dp,
632
				  struct sk_buff *skb, u32 portid,
633 634 635
				  u32 seq, u32 flags, u8 cmd)
{
	const int skb_orig_len = skb->len;
636
	struct nlattr *start;
637 638 639 640 641 642 643
	struct ovs_flow_stats stats;
	struct ovs_header *ovs_header;
	struct nlattr *nla;
	unsigned long used;
	u8 tcp_flags;
	int err;

644
	ovs_header = genlmsg_put(skb, portid, seq, &dp_flow_genl_family, flags, cmd);
645 646 647 648 649
	if (!ovs_header)
		return -EMSGSIZE;

	ovs_header->dp_ifindex = get_dpifindex(dp);

650
	/* Fill flow key. */
651 652 653
	nla = nla_nest_start(skb, OVS_FLOW_ATTR_KEY);
	if (!nla)
		goto nla_put_failure;
654

655
	err = ovs_nla_put_flow(&flow->unmasked_key, &flow->unmasked_key, skb);
656 657 658 659 660 661 662 663
	if (err)
		goto error;
	nla_nest_end(skb, nla);

	nla = nla_nest_start(skb, OVS_FLOW_ATTR_MASK);
	if (!nla)
		goto nla_put_failure;

664
	err = ovs_nla_put_flow(&flow->key, &flow->mask->key, skb);
665 666
	if (err)
		goto error;
667

668 669 670 671 672 673
	nla_nest_end(skb, nla);

	spin_lock_bh(&flow->lock);
	used = flow->used;
	stats.n_packets = flow->packet_count;
	stats.n_bytes = flow->byte_count;
674
	tcp_flags = (u8)ntohs(flow->tcp_flags);
675 676
	spin_unlock_bh(&flow->lock);

677 678 679
	if (used &&
	    nla_put_u64(skb, OVS_FLOW_ATTR_USED, ovs_flow_used_time(used)))
		goto nla_put_failure;
680

681 682 683 684
	if (stats.n_packets &&
	    nla_put(skb, OVS_FLOW_ATTR_STATS,
		    sizeof(struct ovs_flow_stats), &stats))
		goto nla_put_failure;
685

686 687 688
	if (tcp_flags &&
	    nla_put_u8(skb, OVS_FLOW_ATTR_TCP_FLAGS, tcp_flags))
		goto nla_put_failure;
689 690 691 692 693 694 695 696 697 698 699

	/* If OVS_FLOW_ATTR_ACTIONS doesn't fit, skip dumping the actions if
	 * this is the first flow to be dumped into 'skb'.  This is unusual for
	 * Netlink but individual action lists can be longer than
	 * NLMSG_GOODSIZE and thus entirely undumpable if we didn't do this.
	 * The userspace caller can always fetch the actions separately if it
	 * really wants them.  (Most userspace callers in fact don't care.)
	 *
	 * This can only fail for dump operations because the skb is always
	 * properly sized for single flows.
	 */
700 701
	start = nla_nest_start(skb, OVS_FLOW_ATTR_ACTIONS);
	if (start) {
702 703 704 705 706
		const struct sw_flow_actions *sf_acts;

		sf_acts = rcu_dereference_check(flow->sf_acts,
						lockdep_ovsl_is_held());

707 708
		err = ovs_nla_put_actions(sf_acts->actions,
					  sf_acts->actions_len, skb);
709 710 711 712 713 714 715 716 717 718
		if (!err)
			nla_nest_end(skb, start);
		else {
			if (skb_orig_len)
				goto error;

			nla_nest_cancel(skb, start);
		}
	} else if (skb_orig_len)
		goto nla_put_failure;
719 720 721 722 723 724 725 726 727 728 729 730 731 732

	return genlmsg_end(skb, ovs_header);

nla_put_failure:
	err = -EMSGSIZE;
error:
	genlmsg_cancel(skb, ovs_header);
	return err;
}

static struct sk_buff *ovs_flow_cmd_alloc_info(struct sw_flow *flow)
{
	const struct sw_flow_actions *sf_acts;

733
	sf_acts = ovsl_dereference(flow->sf_acts);
734

735
	return genlmsg_new(ovs_flow_cmd_msg_size(sf_acts), GFP_KERNEL);
736 737 738 739
}

static struct sk_buff *ovs_flow_cmd_build_info(struct sw_flow *flow,
					       struct datapath *dp,
740
					       u32 portid, u32 seq, u8 cmd)
741 742 743 744 745 746 747 748
{
	struct sk_buff *skb;
	int retval;

	skb = ovs_flow_cmd_alloc_info(flow);
	if (!skb)
		return ERR_PTR(-ENOMEM);

749
	retval = ovs_flow_cmd_fill_info(flow, dp, skb, portid, seq, 0, cmd);
750 751 752 753
	BUG_ON(retval < 0);
	return skb;
}

754 755 756 757 758 759 760 761
static struct sw_flow *__ovs_flow_tbl_lookup(struct flow_table *tbl,
					      const struct sw_flow_key *key)
{
	u32 __always_unused n_mask_hit;

	return ovs_flow_tbl_lookup(tbl, key, &n_mask_hit);
}

762 763 764 765
static int ovs_flow_cmd_new_or_set(struct sk_buff *skb, struct genl_info *info)
{
	struct nlattr **a = info->attrs;
	struct ovs_header *ovs_header = info->userhdr;
766 767 768
	struct sw_flow_key key, masked_key;
	struct sw_flow *flow = NULL;
	struct sw_flow_mask mask;
769 770
	struct sk_buff *reply;
	struct datapath *dp;
771
	struct sw_flow_actions *acts = NULL;
772
	struct sw_flow_match match;
773 774 775 776 777 778
	int error;

	/* Extract key. */
	error = -EINVAL;
	if (!a[OVS_FLOW_ATTR_KEY])
		goto error;
779 780

	ovs_match_init(&match, &key, &mask);
781 782
	error = ovs_nla_get_match(&match,
				  a[OVS_FLOW_ATTR_KEY], a[OVS_FLOW_ATTR_MASK]);
783 784 785 786 787
	if (error)
		goto error;

	/* Validate actions. */
	if (a[OVS_FLOW_ATTR_ACTIONS]) {
788
		acts = ovs_nla_alloc_flow_actions(nla_len(a[OVS_FLOW_ATTR_ACTIONS]));
789 790
		error = PTR_ERR(acts);
		if (IS_ERR(acts))
791
			goto error;
792

793 794 795
		ovs_flow_mask_key(&masked_key, &key, &mask);
		error = ovs_nla_copy_actions(a[OVS_FLOW_ATTR_ACTIONS],
					     &masked_key, 0, &acts);
796 797
		if (error) {
			OVS_NLERR("Flow actions may not be safe on all matching packets.\n");
798
			goto err_kfree;
799
		}
800 801 802 803 804
	} else if (info->genlhdr->cmd == OVS_FLOW_CMD_NEW) {
		error = -EINVAL;
		goto error;
	}

805
	ovs_lock();
806
	dp = get_dp(sock_net(skb->sk), ovs_header->dp_ifindex);
807 808
	error = -ENODEV;
	if (!dp)
809
		goto err_unlock_ovs;
810

811
	/* Check if this is a duplicate flow */
812
	flow = __ovs_flow_tbl_lookup(&dp->table, &key);
813 814 815 816
	if (!flow) {
		/* Bail out if we're not allowed to create a new flow. */
		error = -ENOENT;
		if (info->genlhdr->cmd == OVS_FLOW_CMD_SET)
817
			goto err_unlock_ovs;
818 819 820 821 822

		/* Allocate flow. */
		flow = ovs_flow_alloc();
		if (IS_ERR(flow)) {
			error = PTR_ERR(flow);
823
			goto err_unlock_ovs;
824 825 826
		}
		clear_stats(flow);

827 828
		flow->key = masked_key;
		flow->unmasked_key = key;
829 830 831
		rcu_assign_pointer(flow->sf_acts, acts);

		/* Put flow in bucket. */
832 833 834 835 836
		error = ovs_flow_tbl_insert(&dp->table, flow, &mask);
		if (error) {
			acts = NULL;
			goto err_flow_free;
		}
837

838
		reply = ovs_flow_cmd_build_info(flow, dp, info->snd_portid,
839
						info->snd_seq, OVS_FLOW_CMD_NEW);
840 841 842 843 844 845 846 847 848 849 850 851 852
	} else {
		/* We found a matching flow. */
		struct sw_flow_actions *old_acts;

		/* Bail out if we're not allowed to modify an existing flow.
		 * We accept NLM_F_CREATE in place of the intended NLM_F_EXCL
		 * because Generic Netlink treats the latter as a dump
		 * request.  We also accept NLM_F_EXCL in case that bug ever
		 * gets fixed.
		 */
		error = -EEXIST;
		if (info->genlhdr->cmd == OVS_FLOW_CMD_NEW &&
		    info->nlhdr->nlmsg_flags & (NLM_F_CREATE | NLM_F_EXCL))
853
			goto err_unlock_ovs;
854

855 856
		/* The unmasked key has to be the same for flow updates. */
		error = -EINVAL;
857
		if (!ovs_flow_cmp_unmasked_key(flow, &match)) {
858 859 860 861
			OVS_NLERR("Flow modification message rejected, unmasked key does not match.\n");
			goto err_unlock_ovs;
		}

862
		/* Update actions. */
863
		old_acts = ovsl_dereference(flow->sf_acts);
864
		rcu_assign_pointer(flow->sf_acts, acts);
865
		ovs_nla_free_flow_actions(old_acts);
866

867
		reply = ovs_flow_cmd_build_info(flow, dp, info->snd_portid,
868 869 870 871 872 873 874 875 876
					       info->snd_seq, OVS_FLOW_CMD_NEW);

		/* Clear stats. */
		if (a[OVS_FLOW_ATTR_CLEAR]) {
			spin_lock_bh(&flow->lock);
			clear_stats(flow);
			spin_unlock_bh(&flow->lock);
		}
	}
877
	ovs_unlock();
878 879

	if (!IS_ERR(reply))
880
		ovs_notify(&dp_flow_genl_family, reply, info);
881
	else
882
		genl_set_err(&dp_flow_genl_family, sock_net(skb->sk), 0,
883
			     0, PTR_ERR(reply));
884 885
	return 0;

886 887
err_flow_free:
	ovs_flow_free(flow, false);
888 889
err_unlock_ovs:
	ovs_unlock();
890 891
err_kfree:
	kfree(acts);
892 893 894 895 896 897 898 899 900 901 902 903
error:
	return error;
}

static int ovs_flow_cmd_get(struct sk_buff *skb, struct genl_info *info)
{
	struct nlattr **a = info->attrs;
	struct ovs_header *ovs_header = info->userhdr;
	struct sw_flow_key key;
	struct sk_buff *reply;
	struct sw_flow *flow;
	struct datapath *dp;
904
	struct sw_flow_match match;
905 906
	int err;

907 908
	if (!a[OVS_FLOW_ATTR_KEY]) {
		OVS_NLERR("Flow get message rejected, Key attribute missing.\n");
909
		return -EINVAL;
910 911 912
	}

	ovs_match_init(&match, &key, NULL);
913
	err = ovs_nla_get_match(&match, a[OVS_FLOW_ATTR_KEY], NULL);
914 915 916
	if (err)
		return err;

917
	ovs_lock();
918
	dp = get_dp(sock_net(skb->sk), ovs_header->dp_ifindex);
919 920 921 922
	if (!dp) {
		err = -ENODEV;
		goto unlock;
	}
923

924
	flow = __ovs_flow_tbl_lookup(&dp->table, &key);
925
	if (!flow || !ovs_flow_cmp_unmasked_key(flow, &match)) {
926 927 928
		err = -ENOENT;
		goto unlock;
	}
929

930
	reply = ovs_flow_cmd_build_info(flow, dp, info->snd_portid,
931
					info->snd_seq, OVS_FLOW_CMD_NEW);
932 933 934 935
	if (IS_ERR(reply)) {
		err = PTR_ERR(reply);
		goto unlock;
	}
936

937
	ovs_unlock();
938
	return genlmsg_reply(reply, info);
939 940 941
unlock:
	ovs_unlock();
	return err;
942 943 944 945 946 947 948 949 950 951
}

static int ovs_flow_cmd_del(struct sk_buff *skb, struct genl_info *info)
{
	struct nlattr **a = info->attrs;
	struct ovs_header *ovs_header = info->userhdr;
	struct sw_flow_key key;
	struct sk_buff *reply;
	struct sw_flow *flow;
	struct datapath *dp;
952
	struct sw_flow_match match;
953 954
	int err;

955
	ovs_lock();
956
	dp = get_dp(sock_net(skb->sk), ovs_header->dp_ifindex);
957 958 959 960
	if (!dp) {
		err = -ENODEV;
		goto unlock;
	}
961

962
	if (!a[OVS_FLOW_ATTR_KEY]) {
963
		err = ovs_flow_tbl_flush(&dp->table);
964 965
		goto unlock;
	}
966 967

	ovs_match_init(&match, &key, NULL);
968
	err = ovs_nla_get_match(&match, a[OVS_FLOW_ATTR_KEY], NULL);
969
	if (err)
970
		goto unlock;
971

972
	flow = __ovs_flow_tbl_lookup(&dp->table, &key);
973
	if (!flow || !ovs_flow_cmp_unmasked_key(flow, &match)) {
974 975 976
		err = -ENOENT;
		goto unlock;
	}
977 978

	reply = ovs_flow_cmd_alloc_info(flow);
979 980 981 982
	if (!reply) {
		err = -ENOMEM;
		goto unlock;
	}
983

984
	ovs_flow_tbl_remove(&dp->table, flow);
985

986
	err = ovs_flow_cmd_fill_info(flow, dp, reply, info->snd_portid,
987 988 989
				     info->snd_seq, 0, OVS_FLOW_CMD_DEL);
	BUG_ON(err < 0);

990
	ovs_flow_free(flow, true);
991
	ovs_unlock();
992

993
	ovs_notify(&dp_flow_genl_family, reply, info);
994
	return 0;
995 996 997
unlock:
	ovs_unlock();
	return err;
998 999 1000 1001 1002
}

static int ovs_flow_cmd_dump(struct sk_buff *skb, struct netlink_callback *cb)
{
	struct ovs_header *ovs_header = genlmsg_data(nlmsg_data(cb->nlh));
1003
	struct table_instance *ti;
1004 1005
	struct datapath *dp;

1006
	rcu_read_lock();
1007
	dp = get_dp(sock_net(skb->sk), ovs_header->dp_ifindex);
1008
	if (!dp) {
1009
		rcu_read_unlock();
1010
		return -ENODEV;
1011
	}
1012

1013
	ti = rcu_dereference(dp->table.ti);
1014 1015 1016 1017 1018 1019
	for (;;) {
		struct sw_flow *flow;
		u32 bucket, obj;

		bucket = cb->args[0];
		obj = cb->args[1];
1020
		flow = ovs_flow_tbl_dump_next(ti, &bucket, &obj);
1021 1022 1023 1024
		if (!flow)
			break;

		if (ovs_flow_cmd_fill_info(flow, dp, skb,
1025
					   NETLINK_CB(cb->skb).portid,
1026 1027 1028 1029 1030 1031 1032
					   cb->nlh->nlmsg_seq, NLM_F_MULTI,
					   OVS_FLOW_CMD_NEW) < 0)
			break;

		cb->args[0] = bucket;
		cb->args[1] = obj;
	}
1033
	rcu_read_unlock();
1034 1035 1036
	return skb->len;
}

1037
static const struct genl_ops dp_flow_genl_ops[] = {
1038 1039 1040 1041 1042 1043 1044 1045 1046 1047 1048 1049 1050 1051 1052 1053 1054 1055 1056 1057 1058 1059 1060 1061 1062