datapath.c 49.9 KB
Newer Older
1
/*
2
 * Copyright (c) 2007-2012 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 47 48 49 50 51
 *
 * 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>
#include <linux/openvswitch.h>
#include <linux/rculist.h>
#include <linux/dmi.h>
#include <linux/workqueue.h>
#include <net/genetlink.h>
52 53
#include <net/net_namespace.h>
#include <net/netns/generic.h>
54 55 56 57 58

#include "datapath.h"
#include "flow.h"
#include "vport-internal_dev.h"

59 60 61 62 63 64 65 66 67 68 69 70 71 72 73
/**
 * struct ovs_net - Per net-namespace data for ovs.
 * @dps: List of datapaths to enable dumping them all out.
 * Protected by genl_mutex.
 */
struct ovs_net {
	struct list_head dps;
};

static int ovs_net_id __read_mostly;

#define REHASH_FLOW_INTERVAL (10 * 60 * HZ)
static void rehash_flow_table(struct work_struct *work);
static DECLARE_DELAYED_WORK(rehash_flow_wq, rehash_flow_table);

74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91
/**
 * DOC: Locking:
 *
 * Writes to device state (add/remove datapath, port, set operations on vports,
 * etc.) are protected by RTNL.
 *
 * Writes to other state (flow table modifications, set miscellaneous datapath
 * parameters, etc.) are protected by genl_mutex.  The RTNL lock nests inside
 * genl_mutex.
 *
 * 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.
 */

static struct vport *new_vport(const struct vport_parms *);
92
static int queue_gso_packets(struct net *, int dp_ifindex, struct sk_buff *,
93
			     const struct dp_upcall_info *);
94 95
static int queue_userspace_packet(struct net *, int dp_ifindex,
				  struct sk_buff *,
96 97 98
				  const struct dp_upcall_info *);

/* Must be called with rcu_read_lock, genl_mutex, or RTNL lock. */
99
static struct datapath *get_dp(struct net *net, int dp_ifindex)
100 101 102 103 104
{
	struct datapath *dp = NULL;
	struct net_device *dev;

	rcu_read_lock();
105
	dev = dev_get_by_index_rcu(net, dp_ifindex);
106 107 108 109 110 111 112 113 114 115 116 117 118
	if (dev) {
		struct vport *vport = ovs_internal_dev_get_vport(dev);
		if (vport)
			dp = vport->dp;
	}
	rcu_read_unlock();

	return dp;
}

/* Must be called with rcu_read_lock or RTNL lock. */
const char *ovs_dp_name(const struct datapath *dp)
{
119
	struct vport *vport = ovs_vport_rtnl_rcu(dp, OVSP_LOCAL);
120 121 122 123 124 125 126 127 128 129
	return vport->ops->get_name(vport);
}

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

	rcu_read_lock();

130
	local = ovs_vport_rcu(dp, OVSP_LOCAL);
131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146
	if (local)
		ifindex = local->ops->get_ifindex(local);
	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);

	ovs_flow_tbl_destroy((__force struct flow_table *)dp->table);
	free_percpu(dp->stats_percpu);
147
	release_net(ovs_dp_get_net(dp));
148
	kfree(dp->ports);
149 150 151
	kfree(dp);
}

152 153 154 155 156 157 158 159 160 161 162 163
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);
164
	hlist_for_each_entry_rcu(vport, head, dp_hash_node) {
165 166 167 168 169 170
		if (vport->port_no == port_no)
			return vport;
	}
	return NULL;
}

171 172 173 174 175 176 177 178
/* Called with RTNL lock and genl_lock. */
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;
179
		struct hlist_head *head = vport_hash_bucket(dp, vport->port_no);
180

181
		hlist_add_head_rcu(&vport->dp_hash_node, head);
182 183 184 185 186 187 188 189 190 191 192
	}

	return vport;
}

/* Called with RTNL lock. */
void ovs_dp_detach_port(struct vport *p)
{
	ASSERT_RTNL();

	/* First drop references to device. */
193
	hlist_del_rcu(&p->dp_hash_node);
194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209

	/* 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;
	int error;
	int key_len;

210
	stats = this_cpu_ptr(dp->stats_percpu);
211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226

	/* Extract flow from 'skb' into 'key'. */
	error = ovs_flow_extract(skb, p->port_no, &key, &key_len);
	if (unlikely(error)) {
		kfree_skb(skb);
		return;
	}

	/* Look up flow. */
	flow = ovs_flow_tbl_lookup(rcu_dereference(dp->table), &key, key_len);
	if (unlikely(!flow)) {
		struct dp_upcall_info upcall;

		upcall.cmd = OVS_PACKET_CMD_MISS;
		upcall.key = &key;
		upcall.userdata = NULL;
227
		upcall.portid = p->upcall_portid;
228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251
		ovs_dp_upcall(dp, skb, &upcall);
		consume_skb(skb);
		stats_counter = &stats->n_missed;
		goto out;
	}

	OVS_CB(skb)->flow = flow;

	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)++;
	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,
252 253
	.maxattr = OVS_PACKET_ATTR_MAX,
	.netnsok = true
254 255 256
};

int ovs_dp_upcall(struct datapath *dp, struct sk_buff *skb,
257
		  const struct dp_upcall_info *upcall_info)
258 259 260 261 262
{
	struct dp_stats_percpu *stats;
	int dp_ifindex;
	int err;

263
	if (upcall_info->portid == 0) {
264 265 266 267 268 269 270 271 272 273 274
		err = -ENOTCONN;
		goto err;
	}

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

	if (!skb_is_gso(skb))
275
		err = queue_userspace_packet(ovs_dp_get_net(dp), dp_ifindex, skb, upcall_info);
276
	else
277
		err = queue_gso_packets(ovs_dp_get_net(dp), dp_ifindex, skb, upcall_info);
278 279 280 281 282 283
	if (err)
		goto err;

	return 0;

err:
284
	stats = this_cpu_ptr(dp->stats_percpu);
285 286 287 288 289 290 291 292

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

	return err;
}

293 294
static int queue_gso_packets(struct net *net, int dp_ifindex,
			     struct sk_buff *skb,
295 296
			     const struct dp_upcall_info *upcall_info)
{
297
	unsigned short gso_type = skb_shinfo(skb)->gso_type;
298 299 300 301 302
	struct dp_upcall_info later_info;
	struct sw_flow_key later_key;
	struct sk_buff *segs, *nskb;
	int err;

303
	segs = __skb_gso_segment(skb, NETIF_F_SG | NETIF_F_HW_CSUM, false);
304 305
	if (IS_ERR(segs))
		return PTR_ERR(segs);
306 307 308 309

	/* Queue all of the segments. */
	skb = segs;
	do {
310
		err = queue_userspace_packet(net, dp_ifindex, skb, upcall_info);
311 312 313
		if (err)
			break;

314
		if (skb == segs && gso_type & SKB_GSO_UDP) {
315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339
			/* 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;
}

340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368
static size_t key_attr_size(void)
{
	return    nla_total_size(4)   /* OVS_KEY_ATTR_PRIORITY */
		+ 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;
}

369 370
static int queue_userspace_packet(struct net *net, int dp_ifindex,
				  struct sk_buff *skb,
371 372 373 374 375 376 377 378 379 380 381 382 383 384
				  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;

		nskb = __vlan_put_tag(nskb, vlan_tx_tag_get(nskb));
385
		if (!nskb)
386 387 388 389 390 391 392 393 394 395 396
			return -ENOMEM;

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

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

397
	user_skb = genlmsg_new(upcall_msg_size(skb, upcall_info->userdata), GFP_ATOMIC);
398 399 400 401 402 403 404 405 406 407 408 409 410 411
	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);
	ovs_flow_to_nlattrs(upcall_info->key, user_skb);
	nla_nest_end(user_skb, nla);

	if (upcall_info->userdata)
412 413 414
		__nla_put(user_skb, OVS_PACKET_ATTR_USERDATA,
			  nla_len(upcall_info->userdata),
			  nla_data(upcall_info->userdata));
415 416 417 418 419

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

	skb_copy_and_csum_dev(skb, nla_data(nla));

420
	genlmsg_end(user_skb, upcall);
421
	err = genlmsg_unicast(net, user_skb, upcall_info->portid);
422 423 424 425 426 427 428

out:
	kfree_skb(nskb);
	return err;
}

/* Called with genl_mutex. */
429
static int flush_flows(struct datapath *dp)
430 431 432 433 434 435 436 437 438 439 440 441 442 443 444 445 446 447 448 449 450 451 452 453 454 455 456 457 458 459 460 461 462 463 464 465 466 467 468 469 470 471 472 473 474 475
{
	struct flow_table *old_table;
	struct flow_table *new_table;

	old_table = genl_dereference(dp->table);
	new_table = ovs_flow_tbl_alloc(TBL_MIN_BUCKETS);
	if (!new_table)
		return -ENOMEM;

	rcu_assign_pointer(dp->table, new_table);

	ovs_flow_tbl_deferred_destroy(old_table);
	return 0;
}

static int validate_actions(const struct nlattr *attr,
				const struct sw_flow_key *key, int depth);

static int validate_sample(const struct nlattr *attr,
				const struct sw_flow_key *key, int depth)
{
	const struct nlattr *attrs[OVS_SAMPLE_ATTR_MAX + 1];
	const struct nlattr *probability, *actions;
	const struct nlattr *a;
	int rem;

	memset(attrs, 0, sizeof(attrs));
	nla_for_each_nested(a, attr, rem) {
		int type = nla_type(a);
		if (!type || type > OVS_SAMPLE_ATTR_MAX || attrs[type])
			return -EINVAL;
		attrs[type] = a;
	}
	if (rem)
		return -EINVAL;

	probability = attrs[OVS_SAMPLE_ATTR_PROBABILITY];
	if (!probability || nla_len(probability) != sizeof(u32))
		return -EINVAL;

	actions = attrs[OVS_SAMPLE_ATTR_ACTIONS];
	if (!actions || (nla_len(actions) && nla_len(actions) < NLA_HDRLEN))
		return -EINVAL;
	return validate_actions(actions, key, depth + 1);
}

476 477 478
static int validate_tp_port(const struct sw_flow_key *flow_key)
{
	if (flow_key->eth.type == htons(ETH_P_IP)) {
479
		if (flow_key->ipv4.tp.src || flow_key->ipv4.tp.dst)
480 481
			return 0;
	} else if (flow_key->eth.type == htons(ETH_P_IPV6)) {
482
		if (flow_key->ipv6.tp.src || flow_key->ipv6.tp.dst)
483 484 485 486 487 488
			return 0;
	}

	return -EINVAL;
}

489 490 491 492 493 494 495 496 497 498 499 500 501 502 503 504
static int validate_set(const struct nlattr *a,
			const struct sw_flow_key *flow_key)
{
	const struct nlattr *ovs_key = nla_data(a);
	int key_type = nla_type(ovs_key);

	/* There can be only one key in a action */
	if (nla_total_size(nla_len(ovs_key)) != nla_len(a))
		return -EINVAL;

	if (key_type > OVS_KEY_ATTR_MAX ||
	    nla_len(ovs_key) != ovs_key_lens[key_type])
		return -EINVAL;

	switch (key_type) {
	const struct ovs_key_ipv4 *ipv4_key;
505
	const struct ovs_key_ipv6 *ipv6_key;
506 507

	case OVS_KEY_ATTR_PRIORITY:
508
	case OVS_KEY_ATTR_SKB_MARK:
509 510 511 512 513 514 515
	case OVS_KEY_ATTR_ETHERNET:
		break;

	case OVS_KEY_ATTR_IPV4:
		if (flow_key->eth.type != htons(ETH_P_IP))
			return -EINVAL;

516
		if (!flow_key->ip.proto)
517 518 519 520 521 522 523 524 525 526 527
			return -EINVAL;

		ipv4_key = nla_data(ovs_key);
		if (ipv4_key->ipv4_proto != flow_key->ip.proto)
			return -EINVAL;

		if (ipv4_key->ipv4_frag != flow_key->ip.frag)
			return -EINVAL;

		break;

528 529 530 531 532 533 534 535 536 537 538 539 540 541 542 543 544 545 546
	case OVS_KEY_ATTR_IPV6:
		if (flow_key->eth.type != htons(ETH_P_IPV6))
			return -EINVAL;

		if (!flow_key->ip.proto)
			return -EINVAL;

		ipv6_key = nla_data(ovs_key);
		if (ipv6_key->ipv6_proto != flow_key->ip.proto)
			return -EINVAL;

		if (ipv6_key->ipv6_frag != flow_key->ip.frag)
			return -EINVAL;

		if (ntohl(ipv6_key->ipv6_label) & 0xFFF00000)
			return -EINVAL;

		break;

547 548 549 550
	case OVS_KEY_ATTR_TCP:
		if (flow_key->ip.proto != IPPROTO_TCP)
			return -EINVAL;

551
		return validate_tp_port(flow_key);
552 553 554 555 556

	case OVS_KEY_ATTR_UDP:
		if (flow_key->ip.proto != IPPROTO_UDP)
			return -EINVAL;

557
		return validate_tp_port(flow_key);
558 559 560 561 562 563 564 565 566 567 568 569

	default:
		return -EINVAL;
	}

	return 0;
}

static int validate_userspace(const struct nlattr *attr)
{
	static const struct nla_policy userspace_policy[OVS_USERSPACE_ATTR_MAX + 1] =	{
		[OVS_USERSPACE_ATTR_PID] = {.type = NLA_U32 },
570
		[OVS_USERSPACE_ATTR_USERDATA] = {.type = NLA_UNSPEC },
571 572 573 574 575 576 577 578 579 580 581 582 583 584 585 586 587 588 589 590 591 592 593 594 595 596 597 598 599 600 601 602 603 604 605 606 607 608 609 610 611 612 613 614 615 616 617 618 619 620 621 622 623 624 625 626 627 628 629 630 631 632 633 634 635 636 637 638 639 640 641 642 643 644 645 646 647 648 649 650 651 652 653 654 655 656 657 658 659 660 661 662 663 664 665 666 667 668 669 670 671 672 673 674 675 676 677 678 679 680 681 682 683 684 685 686
	};
	struct nlattr *a[OVS_USERSPACE_ATTR_MAX + 1];
	int error;

	error = nla_parse_nested(a, OVS_USERSPACE_ATTR_MAX,
				 attr, userspace_policy);
	if (error)
		return error;

	if (!a[OVS_USERSPACE_ATTR_PID] ||
	    !nla_get_u32(a[OVS_USERSPACE_ATTR_PID]))
		return -EINVAL;

	return 0;
}

static int validate_actions(const struct nlattr *attr,
				const struct sw_flow_key *key,  int depth)
{
	const struct nlattr *a;
	int rem, err;

	if (depth >= SAMPLE_ACTION_DEPTH)
		return -EOVERFLOW;

	nla_for_each_nested(a, attr, rem) {
		/* Expected argument lengths, (u32)-1 for variable length. */
		static const u32 action_lens[OVS_ACTION_ATTR_MAX + 1] = {
			[OVS_ACTION_ATTR_OUTPUT] = sizeof(u32),
			[OVS_ACTION_ATTR_USERSPACE] = (u32)-1,
			[OVS_ACTION_ATTR_PUSH_VLAN] = sizeof(struct ovs_action_push_vlan),
			[OVS_ACTION_ATTR_POP_VLAN] = 0,
			[OVS_ACTION_ATTR_SET] = (u32)-1,
			[OVS_ACTION_ATTR_SAMPLE] = (u32)-1
		};
		const struct ovs_action_push_vlan *vlan;
		int type = nla_type(a);

		if (type > OVS_ACTION_ATTR_MAX ||
		    (action_lens[type] != nla_len(a) &&
		     action_lens[type] != (u32)-1))
			return -EINVAL;

		switch (type) {
		case OVS_ACTION_ATTR_UNSPEC:
			return -EINVAL;

		case OVS_ACTION_ATTR_USERSPACE:
			err = validate_userspace(a);
			if (err)
				return err;
			break;

		case OVS_ACTION_ATTR_OUTPUT:
			if (nla_get_u32(a) >= DP_MAX_PORTS)
				return -EINVAL;
			break;


		case OVS_ACTION_ATTR_POP_VLAN:
			break;

		case OVS_ACTION_ATTR_PUSH_VLAN:
			vlan = nla_data(a);
			if (vlan->vlan_tpid != htons(ETH_P_8021Q))
				return -EINVAL;
			if (!(vlan->vlan_tci & htons(VLAN_TAG_PRESENT)))
				return -EINVAL;
			break;

		case OVS_ACTION_ATTR_SET:
			err = validate_set(a, key);
			if (err)
				return err;
			break;

		case OVS_ACTION_ATTR_SAMPLE:
			err = validate_sample(a, key, depth);
			if (err)
				return err;
			break;

		default:
			return -EINVAL;
		}
	}

	if (rem > 0)
		return -EINVAL;

	return 0;
}

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;
	int key_len;

	err = -EINVAL;
	if (!a[OVS_PACKET_ATTR_PACKET] || !a[OVS_PACKET_ATTR_KEY] ||
687
	    !a[OVS_PACKET_ATTR_ACTIONS])
688 689 690 691 692 693 694 695 696
		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);

697
	nla_memcpy(__skb_put(packet, len), a[OVS_PACKET_ATTR_PACKET], len);
698 699 700 701 702 703 704

	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
705
	if (ntohs(eth->h_proto) >= ETH_P_802_3_MIN)
706 707 708 709 710 711 712 713 714 715 716 717 718 719 720
		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;

	err = ovs_flow_extract(packet, -1, &flow->key, &key_len);
	if (err)
		goto err_flow_free;

	err = ovs_flow_metadata_from_nlattrs(&flow->key.phy.priority,
721
					     &flow->key.phy.skb_mark,
722 723 724 725 726 727 728 729 730 731 732 733 734 735 736 737 738 739 740
					     &flow->key.phy.in_port,
					     a[OVS_PACKET_ATTR_KEY]);
	if (err)
		goto err_flow_free;

	err = validate_actions(a[OVS_PACKET_ATTR_ACTIONS], &flow->key, 0);
	if (err)
		goto err_flow_free;

	flow->hash = ovs_flow_hash(&flow->key, key_len);

	acts = ovs_flow_actions_alloc(a[OVS_PACKET_ATTR_ACTIONS]);
	err = PTR_ERR(acts);
	if (IS_ERR(acts))
		goto err_flow_free;
	rcu_assign_pointer(flow->sf_acts, acts);

	OVS_CB(packet)->flow = flow;
	packet->priority = flow->key.phy.priority;
741
	packet->mark = flow->key.phy.skb_mark;
742 743

	rcu_read_lock();
744
	dp = get_dp(sock_net(skb->sk), ovs_header->dp_ifindex);
745 746 747 748 749 750 751 752 753 754 755 756 757 758 759 760 761 762 763 764 765 766 767
	err = -ENODEV;
	if (!dp)
		goto err_unlock;

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

	ovs_flow_free(flow);
	return err;

err_unlock:
	rcu_read_unlock();
err_flow_free:
	ovs_flow_free(flow);
err_kfree_skb:
	kfree_skb(packet);
err:
	return err;
}

static const struct nla_policy packet_policy[OVS_PACKET_ATTR_MAX + 1] = {
768
	[OVS_PACKET_ATTR_PACKET] = { .len = ETH_HLEN },
769 770 771 772 773 774 775 776 777 778 779 780 781 782 783 784 785 786 787 788 789 790 791 792 793 794 795 796 797 798 799 800 801 802 803 804 805 806 807 808 809 810 811 812 813 814 815 816 817
	[OVS_PACKET_ATTR_KEY] = { .type = NLA_NESTED },
	[OVS_PACKET_ATTR_ACTIONS] = { .type = NLA_NESTED },
};

static struct genl_ops dp_packet_genl_ops[] = {
	{ .cmd = OVS_PACKET_CMD_EXECUTE,
	  .flags = GENL_ADMIN_PERM, /* Requires CAP_NET_ADMIN privilege. */
	  .policy = packet_policy,
	  .doit = ovs_packet_cmd_execute
	}
};

static void get_dp_stats(struct datapath *dp, struct ovs_dp_stats *stats)
{
	int i;
	struct flow_table *table = genl_dereference(dp->table);

	stats->n_flows = ovs_flow_tbl_count(table);

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

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,
818 819
	.maxattr = OVS_FLOW_ATTR_MAX,
	.netnsok = true
820 821 822 823 824 825
};

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

826 827 828 829 830 831 832 833 834 835
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 */
		+ 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 */
}

836 837
/* Called with genl_lock. */
static int ovs_flow_cmd_fill_info(struct sw_flow *flow, struct datapath *dp,
838
				  struct sk_buff *skb, u32 portid,
839 840 841 842 843 844 845 846 847 848 849 850 851 852
				  u32 seq, u32 flags, u8 cmd)
{
	const int skb_orig_len = skb->len;
	const struct sw_flow_actions *sf_acts;
	struct ovs_flow_stats stats;
	struct ovs_header *ovs_header;
	struct nlattr *nla;
	unsigned long used;
	u8 tcp_flags;
	int err;

	sf_acts = rcu_dereference_protected(flow->sf_acts,
					    lockdep_genl_is_held());

853
	ovs_header = genlmsg_put(skb, portid, seq, &dp_flow_genl_family, flags, cmd);
854 855 856 857 858 859 860 861 862 863 864 865 866 867 868 869 870 871 872 873
	if (!ovs_header)
		return -EMSGSIZE;

	ovs_header->dp_ifindex = get_dpifindex(dp);

	nla = nla_nest_start(skb, OVS_FLOW_ATTR_KEY);
	if (!nla)
		goto nla_put_failure;
	err = ovs_flow_to_nlattrs(&flow->key, skb);
	if (err)
		goto error;
	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;
	tcp_flags = flow->tcp_flags;
	spin_unlock_bh(&flow->lock);

874 875 876
	if (used &&
	    nla_put_u64(skb, OVS_FLOW_ATTR_USED, ovs_flow_used_time(used)))
		goto nla_put_failure;
877

878 879 880 881
	if (stats.n_packets &&
	    nla_put(skb, OVS_FLOW_ATTR_STATS,
		    sizeof(struct ovs_flow_stats), &stats))
		goto nla_put_failure;
882

883 884 885
	if (tcp_flags &&
	    nla_put_u8(skb, OVS_FLOW_ATTR_TCP_FLAGS, tcp_flags))
		goto nla_put_failure;
886 887 888 889 890 891 892 893 894 895 896 897 898 899 900 901 902 903 904 905 906 907 908 909 910 911 912 913 914 915 916 917

	/* 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.
	 */
	err = nla_put(skb, OVS_FLOW_ATTR_ACTIONS, sf_acts->actions_len,
		      sf_acts->actions);
	if (err < 0 && skb_orig_len)
		goto error;

	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;

	sf_acts = rcu_dereference_protected(flow->sf_acts,
					    lockdep_genl_is_held());

918
	return genlmsg_new(ovs_flow_cmd_msg_size(sf_acts), GFP_KERNEL);
919 920 921 922
}

static struct sk_buff *ovs_flow_cmd_build_info(struct sw_flow *flow,
					       struct datapath *dp,
923
					       u32 portid, u32 seq, u8 cmd)
924 925 926 927 928 929 930 931
{
	struct sk_buff *skb;
	int retval;

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

932
	retval = ovs_flow_cmd_fill_info(flow, dp, skb, portid, seq, 0, cmd);
933 934 935 936 937 938 939 940 941 942 943 944 945 946 947 948 949 950 951 952 953 954 955 956 957 958 959 960 961 962 963 964 965 966
	BUG_ON(retval < 0);
	return skb;
}

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;
	struct sw_flow_key key;
	struct sw_flow *flow;
	struct sk_buff *reply;
	struct datapath *dp;
	struct flow_table *table;
	int error;
	int key_len;

	/* Extract key. */
	error = -EINVAL;
	if (!a[OVS_FLOW_ATTR_KEY])
		goto error;
	error = ovs_flow_from_nlattrs(&key, &key_len, a[OVS_FLOW_ATTR_KEY]);
	if (error)
		goto error;

	/* Validate actions. */
	if (a[OVS_FLOW_ATTR_ACTIONS]) {
		error = validate_actions(a[OVS_FLOW_ATTR_ACTIONS], &key,  0);
		if (error)
			goto error;
	} else if (info->genlhdr->cmd == OVS_FLOW_CMD_NEW) {
		error = -EINVAL;
		goto error;
	}

967
	dp = get_dp(sock_net(skb->sk), ovs_header->dp_ifindex);
968 969 970 971 972 973 974 975 976 977 978 979 980 981 982 983 984 985 986 987 988 989 990 991 992 993 994 995 996 997 998 999 1000 1001 1002 1003 1004 1005 1006 1007 1008 1009 1010 1011 1012 1013
	error = -ENODEV;
	if (!dp)
		goto error;

	table = genl_dereference(dp->table);
	flow = ovs_flow_tbl_lookup(table, &key, key_len);
	if (!flow) {
		struct sw_flow_actions *acts;

		/* Bail out if we're not allowed to create a new flow. */
		error = -ENOENT;
		if (info->genlhdr->cmd == OVS_FLOW_CMD_SET)
			goto error;

		/* Expand table, if necessary, to make room. */
		if (ovs_flow_tbl_need_to_expand(table)) {
			struct flow_table *new_table;

			new_table = ovs_flow_tbl_expand(table);
			if (!IS_ERR(new_table)) {
				rcu_assign_pointer(dp->table, new_table);
				ovs_flow_tbl_deferred_destroy(table);
				table = genl_dereference(dp->table);
			}
		}

		/* Allocate flow. */
		flow = ovs_flow_alloc();
		if (IS_ERR(flow)) {
			error = PTR_ERR(flow);
			goto error;
		}
		flow->key = key;
		clear_stats(flow);

		/* Obtain actions. */
		acts = ovs_flow_actions_alloc(a[OVS_FLOW_ATTR_ACTIONS]);
		error = PTR_ERR(acts);
		if (IS_ERR(acts))
			goto error_free_flow;
		rcu_assign_pointer(flow->sf_acts, acts);

		/* Put flow in bucket. */
		flow->hash = ovs_flow_hash(&key, key_len);
		ovs_flow_tbl_insert(table, flow);

1014
		reply = ovs_flow_cmd_build_info(flow, dp, info->snd_portid,
1015 1016 1017 1018 1019 1020 1021 1022 1023 1024 1025 1026 1027 1028 1029 1030 1031 1032 1033 1034 1035 1036 1037 1038 1039 1040 1041 1042 1043 1044 1045 1046 1047 1048 1049 1050 1051
						info->snd_seq,
						OVS_FLOW_CMD_NEW);
	} else {
		/* We found a matching flow. */
		struct sw_flow_actions *old_acts;
		struct nlattr *acts_attrs;

		/* 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))
			goto error;

		/* Update actions. */
		old_acts = rcu_dereference_protected(flow->sf_acts,
						     lockdep_genl_is_held());
		acts_attrs = a[OVS_FLOW_ATTR_ACTIONS];
		if (acts_attrs &&
		   (old_acts->actions_len != nla_len(acts_attrs) ||
		   memcmp(old_acts->actions, nla_data(acts_attrs),
			  old_acts->actions_len))) {
			struct sw_flow_actions *new_acts;

			new_acts = ovs_flow_actions_alloc(acts_attrs);
			error = PTR_ERR(new_acts);
			if (IS_ERR(new_acts))
				goto error;

			rcu_assign_pointer(flow->sf_acts, new_acts);
			ovs_flow_deferred_free_acts(old_acts);
		}

1052
		reply = ovs_flow_cmd_build_info(flow, dp, info->snd_portid,
1053 1054 1055 1056 1057 1058 1059 1060 1061 1062 1063
					       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);
		}
	}

	if (!IS_ERR(reply))
1064
		genl_notify(reply, genl_info_net(info), info->snd_portid,
1065 1066 1067
			   ovs_dp_flow_multicast_group.id, info->nlhdr,
			   GFP_KERNEL);
	else
1068
		netlink_set_err(sock_net(skb->sk)->genl_sock, 0,
1069 1070 1071 1072 1073 1074 1075 1076 1077 1078 1079 1080 1081 1082 1083 1084 1085 1086 1087 1088 1089 1090 1091 1092 1093 1094 1095
				ovs_dp_flow_multicast_group.id, PTR_ERR(reply));
	return 0;

error_free_flow:
	ovs_flow_free(flow);
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;
	struct flow_table *table;
	int err;
	int key_len;

	if (!a[OVS_FLOW_ATTR_KEY])
		return -EINVAL;
	err = ovs_flow_from_nlattrs(&key, &key_len, a[OVS_FLOW_ATTR_KEY]);
	if (err)
		return err;

1096
	dp = get_dp(sock_net(skb->sk), ovs_header->dp_ifindex);
1097 1098 1099 1100 1101 1102 1103 1104
	if (!dp)
		return -ENODEV;

	table = genl_dereference(dp->table);
	flow = ovs_flow_tbl_lookup(table, &key, key_len);
	if (!flow)
		return -ENOENT;

1105
	reply = ovs_flow_cmd_build_info(flow, dp, info->snd_portid,
1106 1107 1108 1109 1110 1111 1112 1113 1114 1115 1116 1117 1118 1119 1120 1121 1122 1123 1124
					info->snd_seq, OVS_FLOW_CMD_NEW);
	if (IS_ERR(reply))
		return PTR_ERR(reply);

	return genlmsg_reply(reply, info);
}

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;
	struct flow_table *table;
	int err;
	int key_len;

1125 1126 1127 1128
	dp = get_dp(sock_net(skb->sk), ovs_header->dp_ifindex);
	if (!dp)
		return -ENODEV;

1129
	if (!a[OVS_FLOW_ATTR_KEY])
1130 1131
		return flush_flows(dp);

1132 1133 1134 1135 1136 1137 1138 1139 1140 1141 1142 1143 1144 1145 1146
	err = ovs_flow_from_nlattrs(&key, &key_len, a[OVS_FLOW_ATTR_KEY]);
	if (err)
		return err;

	table = genl_dereference(dp->table);
	flow = ovs_flow_tbl_lookup(table, &key, key_len);
	if (!flow)
		return -ENOENT;

	reply = ovs_flow_cmd_alloc_info(flow);
	if (!reply)
		return -ENOMEM;

	ovs_flow_tbl_remove(table, flow);

1147
	err = ovs_flow_cmd_fill_info(flow, dp, reply, info->snd_portid,
1148 1149 1150 1151 1152
				     info->snd_seq, 0, OVS_FLOW_CMD_DEL);
	BUG_ON(err < 0);

	ovs_flow_deferred_free(flow);

1153
	genl_notify(reply, genl_info_net(info), info->snd_portid,
1154 1155 1156 1157 1158 1159 1160 1161 1162 1163
		    ovs_dp_flow_multicast_group.id, info->nlhdr, GFP_KERNEL);
	return 0;
}

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));
	struct datapath *dp;
	struct flow_table *table;

1164
	dp = get_dp(sock_net(skb->sk), ovs_header->dp_ifindex);
1165 1166 1167 1168 1169 1170 1171 1172 1173 1174 1175 1176 1177 1178 1179 1180
	if (!dp)
		return -ENODEV;

	table = genl_dereference(dp->table);

	for (;;) {
		struct sw_flow *flow;
		u32 bucket, obj;

		bucket = cb->args[0];
		obj = cb->args[1];
		flow = ovs_flow_tbl_next(table, &bucket, &obj);
		if (!flow)
			break;

		if (ovs_flow_cmd_fill_info(flow, dp, skb,
1181
					   NETLINK_CB(cb->skb).portid,
1182 1183 1184 1185 1186 1187 1188 1189 1190 1191 1192 1193 1194 1195 1196 1197 1198 1199 1200 1201 1202 1203 1204 1205 1206 1207 1208 1209 1210 1211 1212 1213 1214 1215 1216 1217 1218 1219 1220 1221 1222 1223 1224 1225
					   cb->nlh->nlmsg_seq, NLM_F_MULTI,
					   OVS_FLOW_CMD_NEW) < 0)
			break;

		cb->args[0] = bucket;
		cb->args[1] = obj;
	}
	return skb->len;
}

static struct genl_ops dp_flow_genl_ops[] = {
	{ .cmd = OVS_FLOW_CMD_NEW,
	  .flags = GENL_ADMIN_PERM, /* Requires CAP_NET_ADMIN privilege. */
	  .policy = flow_policy,
	  .doit = ovs_flow_cmd_new_or_set
	},
	{ .cmd = OVS_FLOW_CMD_DEL,
	  .flags = GENL_ADMIN_PERM, /* Requires CAP_NET_ADMIN privilege. */
	  .policy = flow_policy,
	  .doit = ovs_flow_cmd_del
	},
	{ .cmd = OVS_FLOW_CMD_GET,
	  .flags = 0,		    /* OK for unprivileged users. */
	  .policy = flow_policy,
	  .doit = ovs_flow_cmd_get,
	  .dumpit = ovs_flow_cmd_dump
	},
	{ .cmd = OVS_FLOW_CMD_SET,
	  .flags = GENL_ADMIN_PERM, /* Requires CAP_NET_ADMIN privilege. */
	  .policy = flow_policy,
	  .doit = ovs_flow_cmd_new_or_set,
	},
};

static const struct nla_policy datapath_policy[OVS_DP_ATTR_MAX + 1] = {
	[OVS_DP_ATTR_NAME] = { .type = NLA_NUL_STRING, .len = IFNAMSIZ - 1 },
	[OVS_DP_ATTR_UPCALL_PID] = { .type = NLA_U32 },
};

static struct genl_family dp_datapath_genl_family = {
	.id = GENL_ID_GENERATE,
	.hdrsize = sizeof(struct ovs_header),
	.name = OVS_DATAPATH_FAMILY,
	.version = OVS_DATAPATH_VERSION,
1226 1227
	.maxattr = OVS_DP_ATTR_MAX,
	.netnsok = true
1228 1229 1230 1231 1232 1233
};

static struct genl_multicast_group ovs_dp_datapath_multicast_group = {
	.name = OVS_DATAPATH_MCGROUP
};

1234 1235 1236 1237 1238 1239 1240 1241 1242 1243
static size_t ovs_dp_cmd_msg_size(void)
{
	size_t msgsize = NLMSG_ALIGN(sizeof(struct ovs_header));

	msgsize += nla_total_size(IFNAMSIZ);
	msgsize += nla_total_size(sizeof(struct ovs_dp_stats));

	return msgsize;
}

1244
static int ovs_dp_cmd_fill_info(struct datapath *dp, struct sk_buff *skb,
1245
				u32 portid, u32 seq, u32 flags, u8 cmd)
1246 1247 1248 1249 1250
{
	struct ovs_header *ovs_header;
	struct ovs_dp_stats dp_stats;
	int err;

1251
	ovs_header = genlmsg_put(skb, portid, seq, &dp_datapath_genl_family,
1252 1253 1254 1255 1256 1257 1258 1259 1260 1261 1262 1263 1264
				   flags, cmd);
	if (!ovs_header)
		goto error;

	ovs_header->dp_ifindex = get_dpifindex(dp);

	rcu_read_lock();
	err = nla_put_string(skb, OVS_DP_ATTR_NAME, ovs_dp_name(dp));
	rcu_read_unlock();
	if (err)
		goto nla_put_failure;

	get_dp_stats(dp, &dp_stats);
1265 1266
	if (nla_put(skb, OVS_DP_ATTR_STATS, sizeof(struct ovs_dp_stats), &dp_stats))
		goto nla_put_failure;
1267 1268 1269 1270 1271 1272 1273 1274 1275

	return genlmsg_end(skb, ovs_header);

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

1276
static struct sk_buff *ovs_dp_cmd_build_info(struct datapath *dp, u32 portid,
1277 1278 1279 1280 1281
					     u32 seq, u8 cmd)
{
	struct sk_buff *skb;
	int retval;

1282
	skb = genlmsg_new(ovs_dp_cmd_msg_size(), GFP_KERNEL);
1283 1284 1285
	if (!skb)
		return ERR_PTR(-ENOMEM);

1286
	retval = ovs_dp_cmd_fill_info(dp, skb, portid, seq, 0, cmd);
1287 1288 1289 1290 1291 1292 1293 1294
	if (retval < 0) {
		kfree_skb(skb);
		return ERR_PTR(retval);
	}
	return skb;
}

/* Called with genl_mutex and optionally with RTNL lock also. */
1295 1296
static struct datapath *lookup_datapath(struct net *net,
					struct ovs_header *ovs_header,
1297 1298 1299 1300 1301
					struct nlattr *a[OVS_DP_ATTR_MAX + 1])
{
	struct datapath *dp;

	if (!a[OVS_DP_ATTR_NAME])
1302
		dp = get_dp(net, ovs_header->dp_ifindex);
1303 1304 1305 1306
	else {
		struct vport *vport;

		rcu_read_lock();
1307
		vport = ovs_vport_locate(net, nla_data(a[OVS_DP_ATTR_NAME]));
1308 1309 1310 1311 1312 1313 1314 1315 1316 1317 1318 1319 1320
		dp = vport && vport->port_no == OVSP_LOCAL ? vport->dp : NULL;
		rcu_read_unlock();
	}
	return dp ? dp : ERR_PTR(-ENODEV);
}

static int ovs_dp_cmd_new(struct sk_buff *skb, struct genl_info *info)
{
	struct nlattr **a = info->attrs;
	struct vport_parms parms;
	struct sk_buff *reply;
	struct datapath *dp;
	struct vport *vport;
1321
	struct ovs_net *ovs_net;
1322
	int err, i;
1323 1324 1325 1326 1327 1328 1329 1330 1331 1332

	err = -EINVAL;
	if (!a[OVS_DP_ATTR_NAME] || !a[OVS_DP_ATTR_UPCALL_PID])
		goto err;

	rtnl_lock();

	err = -ENOMEM;
	dp = kzalloc(sizeof(*dp), GFP_KERNEL);
	if (dp == NULL)
1333 1334 1335
		goto err_unlock_rtnl;

	ovs_dp_set_net(dp, hold_net(sock_net(skb->sk)));
1336 1337 1338 1339 1340 1341 1342 1343 1344 1345 1346 1347 1348

	/* Allocate table. */
	err = -ENOMEM;
	rcu_assign_pointer(dp->table, ovs_flow_tbl_alloc(TBL_MIN_BUCKETS));
	if (!dp->table)
		goto err_free_dp;

	dp->stats_percpu = alloc_percpu(struct dp_stats_percpu);
	if (!dp->stats_percpu) {
		err = -ENOMEM;
		goto err_destroy_table;
	}

1349 1350 1351 1352 1353 1354 1355 1356 1357 1358
	dp->ports = kmalloc(DP_VPORT_HASH_BUCKETS * sizeof(struct hlist_head),
			GFP_KERNEL);
	if (!dp->ports) {
		err = -ENOMEM;
		goto err_destroy_percpu;
	}

	for (i = 0; i < DP_VPORT_HASH_BUCKETS; i++)
		INIT_HLIST_HEAD(&dp->ports[i]);

1359 1360 1361 1362 1363 1364
	/* Set up our datapath device. */
	parms.name = nla_data(a[OVS_DP_ATTR_NAME]);
	parms.type = OVS_VPORT_TYPE_INTERNAL;
	parms.options = NULL;
	parms.dp = dp;
	parms.port_no = OVSP_LOCAL;
1365
	parms.upcall_portid = nla_get_u32(a[OVS_DP_ATTR_UPCALL_PID]);
1366 1367 1368 1369 1370 1371 1372

	vport = new_vport(&parms);
	if (IS_ERR(vport)) {
		err = PTR_ERR(vport);
		if (err == -EBUSY)
			err = -EEXIST;

1373
		goto err_destroy_ports_array;
1374 1375
	}

1376
	reply = ovs_dp_cmd_build_info(dp, info->snd_portid,
1377 1378 1379 1380 1381
				      info->snd_seq, OVS_DP_CMD_NEW);
	err = PTR_ERR(reply);
	if (IS_ERR(reply))
		goto err_destroy_local_port;

1382 1383
	ovs_net = net_generic(ovs_dp_get_net(dp), ovs_net_id);
	list_add_tail(&dp->list_node, &ovs_net->dps);
1384 1385
	rtnl_unlock();

1386
	genl_notify(reply, genl_info_net(info), info->snd_portid,
1387 1388 1389 1390 1391
		    ovs_dp_datapath_multicast_group.id, info->nlhdr,
		    GFP_KERNEL);
	return 0;

err_destroy_local_port:
1392 1393 1394
	ovs_dp_detach_port(ovs_vport_rtnl(dp, OVSP_LOCAL));
err_destroy_ports_array:
	kfree(dp->ports);
1395 1396 1397 1398 1399
err_destroy_percpu:
	free_percpu(dp->stats_percpu);
err_destroy_table:
	ovs_flow_tbl_destroy(genl_dereference(dp->table));
err_free_dp:
1400
	release_net(ovs_dp_get_net(dp));
1401 1402 1403 1404 1405 1406 1407
	kfree(dp);
err_unlock_rtnl:
	rtnl_unlock();
err:
	return err;
}

1408 1409
/* Called with genl_mutex. */
static void __dp_destroy(struct datapath *dp)
1410
{
1411
	int i;
1412 1413

	rtnl_lock();
1414 1415 1416

	for (i = 0; i < DP_VPORT_HASH_BUCKETS; i++) {
		struct vport *vport;
1417
		struct hlist_node *n;
1418

1419
		hlist_for_each_entry_safe(vport, n, &dp->ports[i], dp_hash_node)
1420 1421 1422
			if (vport->port_no != OVSP_LOCAL)
				ovs_dp_detach_port(vport);
	}
1423 1424

	list_del(&dp->list_node);
1425
	ovs_dp_detach_port(ovs_vport_rtnl(dp, OVSP_LOCAL));
1426 1427 1428 1429 1430 1431 1432 1433 1434

	/* rtnl_unlock() will wait until all the references to devices that
	 * are pending unregistration have been dropped.  We do it here to
	 * ensure that any internal devices (which contain DP pointers) are
	 * fully destroyed before freeing the datapath.
	 */
	rtnl_unlock();

	call_rcu(&dp->rcu, destroy_dp_rcu);
1435 1436 1437 1438 1439 1440 1441 1442 1443 1444 1445 1446 1447
}

static int ovs_dp_cmd_del(struct sk_buff *skb, struct genl_info *info)
{
	struct sk_buff *reply;
	struct datapath *dp;
	int err;

	dp = lookup_datapath(sock_net(skb->sk), info->userhdr, info->attrs);
	err = PTR_ERR(dp);
	if (IS_ERR(dp))
		return err;

1448
	reply = ovs_dp_cmd_build_info(dp, info->snd_portid,
1449 1450 1451 1452 1453 1454
				      info->snd_seq, OVS_DP_CMD_DEL);
	err = PTR_ERR(reply);
	if (IS_ERR(reply))
		return err;

	__dp_destroy(dp);
1455

1456
	genl_notify(reply, genl_info_net(info), info->snd_portid,
1457 1458 1459 1460 1461 1462 1463 1464 1465 1466 1467 1468
		    ovs_dp_datapath_multicast_group.id, info->nlhdr,
		    GFP_KERNEL);

	return 0;
}

static int ovs_dp_cmd_set(struct sk_buff *skb, struct genl_info *info)
{
	struct sk_buff *reply;
	struct datapath *dp;
	int err;

1469
	dp = lookup_datapath(sock_net(skb->sk), info->userhdr, info->attrs);
1470 1471 1472
	if (IS_ERR(dp))
		return PTR_ERR(dp);

1473
	reply = ovs_dp_cmd_build_info(dp, info->snd_portid,
1474 1475 1476
				      info->snd_seq, OVS_DP_CMD_NEW);
	if (IS_ERR(reply)) {
		err = PTR_ERR(reply);
1477
		netlink_set_err(sock_net(skb->sk)->genl_sock, 0,
1478 1479 1480 1481
				ovs_dp_datapath_multicast_group.id, err);
		return 0;
	}

1482
	genl_notify(reply, genl_info_net(info), info->snd_portid,
1483 1484 1485 1486 1487 1488 1489 1490 1491 1492 1493
		    ovs_dp_datapath_multicast_group.id, info->nlhdr,
		    GFP_KERNEL);

	return 0;
}

static int ovs_dp_cmd_get(struct sk_buff *skb, struct genl_info *info)
{
	struct sk_buff *reply;
	struct datapath *dp;

1494
	dp = lookup_datapath(sock_net(skb->sk), info->userhdr, info->attrs);
1495 1496 1497
	if (IS_ERR(dp))
		return PTR_ERR(dp);

1498
	reply = ovs_dp_cmd_build_info(dp, info->snd_portid,
1499 1500 1501 1502 1503 1504 1505 1506 1507
				      info->snd_seq, OVS_DP_CMD_NEW);
	if (IS_ERR(reply))
		return PTR_ERR(reply);

	return genlmsg_reply(reply, info);
}

static int ovs_dp_cmd_dump(struct sk_buff *skb, struct netlink_callback *cb)
{
1508
	struct ovs_net *ovs_net = net_generic(sock_net(skb->sk), ovs_net_id);
1509 1510 1511 1512
	struct datapath *dp;
	int skip = cb->args[0];
	int i = 0;

1513
	list_for_each_entry(dp, &ovs_net->dps, list_node) {
1514
		if (i >= skip &&