addr.c 21.1 KB
Newer Older
1
2
3
4
5
6
/*
 * Copyright (c) 2005 Voltaire Inc.  All rights reserved.
 * Copyright (c) 2002-2005, Network Appliance, Inc. All rights reserved.
 * Copyright (c) 1999-2005, Mellanox Technologies, Inc. All rights reserved.
 * Copyright (c) 2005 Intel Corporation.  All rights reserved.
 *
Sean Hefty's avatar
Sean Hefty committed
7
8
9
10
11
 * This software is available to you under a choice of one of two
 * licenses.  You may choose to be licensed under the terms of the GNU
 * General Public License (GPL) Version 2, available from the file
 * COPYING in the main directory of this source tree, or the
 * OpenIB.org BSD license below:
12
 *
Sean Hefty's avatar
Sean Hefty committed
13
14
15
 *     Redistribution and use in source and binary forms, with or
 *     without modification, are permitted provided that the following
 *     conditions are met:
16
 *
Sean Hefty's avatar
Sean Hefty committed
17
18
19
 *      - Redistributions of source code must retain the above
 *        copyright notice, this list of conditions and the following
 *        disclaimer.
20
 *
Sean Hefty's avatar
Sean Hefty committed
21
22
23
24
 *      - Redistributions in binary form must reproduce the above
 *        copyright notice, this list of conditions and the following
 *        disclaimer in the documentation and/or other materials
 *        provided with the distribution.
25
 *
Sean Hefty's avatar
Sean Hefty committed
26
27
28
29
30
31
32
33
 * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
 * EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
 * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND
 * NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS
 * BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN
 * ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
 * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
 * SOFTWARE.
34
35
36
37
 */

#include <linux/mutex.h>
#include <linux/inetdevice.h>
38
#include <linux/slab.h>
39
#include <linux/workqueue.h>
40
#include <linux/module.h>
41
42
43
#include <net/arp.h>
#include <net/neighbour.h>
#include <net/route.h>
44
#include <net/netevent.h>
45
46
#include <net/addrconf.h>
#include <net/ip6_route.h>
47
#include <rdma/ib_addr.h>
48
#include <rdma/ib.h>
49
50
51
52
#include <rdma/rdma_netlink.h>
#include <net/netlink.h>

#include "core_priv.h"
53
54
55

struct addr_req {
	struct list_head list;
56
57
	struct sockaddr_storage src_addr;
	struct sockaddr_storage dst_addr;
58
	struct rdma_dev_addr *addr;
59
	struct rdma_addr_client *client;
60
61
62
63
	void *context;
	void (*callback)(int status, struct sockaddr *src_addr,
			 struct rdma_dev_addr *addr, void *context);
	unsigned long timeout;
64
	struct delayed_work work;
65
	int status;
66
	u32 seq;
67
68
};

69
70
static atomic_t ib_nl_addr_request_seq = ATOMIC_INIT(0);

David Howells's avatar
David Howells committed
71
static void process_req(struct work_struct *work);
72
73
74

static DEFINE_MUTEX(lock);
static LIST_HEAD(req_list);
David Howells's avatar
David Howells committed
75
static DECLARE_DELAYED_WORK(work, process_req);
76
77
static struct workqueue_struct *addr_wq;

78
79
80
81
82
83
84
85
86
87
88
89
90
91
static const struct nla_policy ib_nl_addr_policy[LS_NLA_TYPE_MAX] = {
	[LS_NLA_TYPE_DGID] = {.type = NLA_BINARY,
		.len = sizeof(struct rdma_nla_ls_gid)},
};

static inline bool ib_nl_is_good_ip_resp(const struct nlmsghdr *nlh)
{
	struct nlattr *tb[LS_NLA_TYPE_MAX] = {};
	int ret;

	if (nlh->nlmsg_flags & RDMA_NL_LS_F_ERR)
		return false;

	ret = nla_parse(tb, LS_NLA_TYPE_MAX - 1, nlmsg_data(nlh),
92
			nlmsg_len(nlh), ib_nl_addr_policy, NULL);
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
	if (ret)
		return false;

	return true;
}

static void ib_nl_process_good_ip_rsep(const struct nlmsghdr *nlh)
{
	const struct nlattr *head, *curr;
	union ib_gid gid;
	struct addr_req *req;
	int len, rem;
	int found = 0;

	head = (const struct nlattr *)nlmsg_data(nlh);
	len = nlmsg_len(nlh);

	nla_for_each_attr(curr, head, len, rem) {
		if (curr->nla_type == LS_NLA_TYPE_DGID)
			memcpy(&gid, nla_data(curr), nla_len(curr));
	}

	mutex_lock(&lock);
	list_for_each_entry(req, &req_list, list) {
		if (nlh->nlmsg_seq != req->seq)
			continue;
		/* We set the DGID part, the rest was set earlier */
		rdma_addr_set_dgid(req->addr, &gid);
		req->status = 0;
		found = 1;
		break;
	}
	mutex_unlock(&lock);

	if (!found)
		pr_info("Couldn't find request waiting for DGID: %pI6\n",
			&gid);
}

int ib_nl_handle_ip_res_resp(struct sk_buff *skb,
133
134
			     struct nlmsghdr *nlh,
			     struct netlink_ext_ack *extack)
135
136
{
	if ((nlh->nlmsg_flags & NLM_F_REQUEST) ||
137
	    !(NETLINK_CB(skb).sk))
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
		return -EPERM;

	if (ib_nl_is_good_ip_resp(nlh))
		ib_nl_process_good_ip_rsep(nlh);

	return skb->len;
}

static int ib_nl_ip_send_msg(struct rdma_dev_addr *dev_addr,
			     const void *daddr,
			     u32 seq, u16 family)
{
	struct sk_buff *skb = NULL;
	struct nlmsghdr *nlh;
	struct rdma_ls_ip_resolve_header *header;
	void *data;
	size_t size;
	int attrtype;
	int len;

	if (family == AF_INET) {
		size = sizeof(struct in_addr);
		attrtype = RDMA_NLA_F_MANDATORY | LS_NLA_TYPE_IPV4;
	} else {
		size = sizeof(struct in6_addr);
		attrtype = RDMA_NLA_F_MANDATORY | LS_NLA_TYPE_IPV6;
	}

	len = nla_total_size(sizeof(size));
	len += NLMSG_ALIGN(sizeof(*header));

	skb = nlmsg_new(len, GFP_KERNEL);
	if (!skb)
		return -ENOMEM;

	data = ibnl_put_msg(skb, &nlh, seq, 0, RDMA_NL_LS,
			    RDMA_NL_LS_OP_IP_RESOLVE, NLM_F_REQUEST);
	if (!data) {
		nlmsg_free(skb);
		return -ENODATA;
	}

	/* Construct the family header first */
181
	header = skb_put(skb, NLMSG_ALIGN(sizeof(*header)));
182
183
184
185
186
	header->ifindex = dev_addr->bound_dev_if;
	nla_put(skb, attrtype, size, daddr);

	/* Repair the nlmsg header length */
	nlmsg_end(skb, nlh);
187
	rdma_nl_multicast(skb, RDMA_NL_GROUP_LS, GFP_KERNEL);
188
189
190
191
192
193
194

	/* Make the request retry, so when we get the response from userspace
	 * we will have something.
	 */
	return -ENODATA;
}

195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
int rdma_addr_size(struct sockaddr *addr)
{
	switch (addr->sa_family) {
	case AF_INET:
		return sizeof(struct sockaddr_in);
	case AF_INET6:
		return sizeof(struct sockaddr_in6);
	case AF_IB:
		return sizeof(struct sockaddr_ib);
	default:
		return 0;
	}
}
EXPORT_SYMBOL(rdma_addr_size);

210
211
static struct rdma_addr_client self;

212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
void rdma_addr_register_client(struct rdma_addr_client *client)
{
	atomic_set(&client->refcount, 1);
	init_completion(&client->comp);
}
EXPORT_SYMBOL(rdma_addr_register_client);

static inline void put_client(struct rdma_addr_client *client)
{
	if (atomic_dec_and_test(&client->refcount))
		complete(&client->comp);
}

void rdma_addr_unregister_client(struct rdma_addr_client *client)
{
	put_client(client);
	wait_for_completion(&client->comp);
}
EXPORT_SYMBOL(rdma_addr_unregister_client);

232
233
234
void rdma_copy_addr(struct rdma_dev_addr *dev_addr,
		    const struct net_device *dev,
		    const unsigned char *dst_dev_addr)
235
{
236
	dev_addr->dev_type = dev->type;
237
238
239
240
	memcpy(dev_addr->src_dev_addr, dev->dev_addr, MAX_ADDR_LEN);
	memcpy(dev_addr->broadcast, dev->broadcast, MAX_ADDR_LEN);
	if (dst_dev_addr)
		memcpy(dev_addr->dst_dev_addr, dst_dev_addr, MAX_ADDR_LEN);
241
	dev_addr->bound_dev_if = dev->ifindex;
242
}
Tom Tucker's avatar
Tom Tucker committed
243
EXPORT_SYMBOL(rdma_copy_addr);
244

245
246
int rdma_translate_ip(const struct sockaddr *addr,
		      struct rdma_dev_addr *dev_addr,
247
		      u16 *vlan_id)
248
249
250
{
	struct net_device *dev;

251
	if (dev_addr->bound_dev_if) {
252
		dev = dev_get_by_index(dev_addr->net, dev_addr->bound_dev_if);
253
254
		if (!dev)
			return -ENODEV;
255
		rdma_copy_addr(dev_addr, dev, NULL);
256
		dev_put(dev);
257
		return 0;
258
259
	}

260
261
	switch (addr->sa_family) {
	case AF_INET:
262
		dev = ip_dev_find(dev_addr->net,
263
			((const struct sockaddr_in *)addr)->sin_addr.s_addr);
264
265

		if (!dev)
266
			return -EADDRNOTAVAIL;
267

268
		rdma_copy_addr(dev_addr, dev, NULL);
269
		dev_addr->bound_dev_if = dev->ifindex;
270
271
		if (vlan_id)
			*vlan_id = rdma_vlan_dev_vlan_id(dev);
272
273
		dev_put(dev);
		break;
274
#if IS_ENABLED(CONFIG_IPV6)
275
	case AF_INET6:
276
		rcu_read_lock();
277
278
		for_each_netdev_rcu(dev_addr->net, dev) {
			if (ipv6_chk_addr(dev_addr->net,
279
					  &((const struct sockaddr_in6 *)addr)->sin6_addr,
280
					  dev, 1)) {
281
				rdma_copy_addr(dev_addr, dev, NULL);
282
				dev_addr->bound_dev_if = dev->ifindex;
283
284
				if (vlan_id)
					*vlan_id = rdma_vlan_dev_vlan_id(dev);
285
286
287
				break;
			}
		}
288
		rcu_read_unlock();
289
		break;
290
#endif
291
	}
292
	return 0;
293
294
295
}
EXPORT_SYMBOL(rdma_translate_ip);

296
static void set_timeout(struct delayed_work *delayed_work, unsigned long time)
297
298
299
300
{
	unsigned long delay;

	delay = time - jiffies;
301
302
	if ((long)delay < 0)
		delay = 0;
303

304
	mod_delayed_work(addr_wq, delayed_work, delay);
305
306
307
308
309
310
311
312
}

static void queue_req(struct addr_req *req)
{
	struct addr_req *temp_req;

	mutex_lock(&lock);
	list_for_each_entry_reverse(temp_req, &req_list, list) {
313
		if (time_after_eq(req->timeout, temp_req->timeout))
314
315
316
317
318
			break;
	}

	list_add(&req->list, &temp_req->list);

319
	set_timeout(&req->work, req->timeout);
320
321
322
	mutex_unlock(&lock);
}

323
324
325
static int ib_nl_fetch_ha(struct dst_entry *dst, struct rdma_dev_addr *dev_addr,
			  const void *daddr, u32 seq, u16 family)
{
326
	if (rdma_nl_chk_listeners(RDMA_NL_GROUP_LS))
327
328
329
330
331
332
333
		return -EADDRNOTAVAIL;

	/* We fill in what we can, the response will fill the rest */
	rdma_copy_addr(dev_addr, dst->dev, NULL);
	return ib_nl_ip_send_msg(dev_addr, daddr, seq, family);
}

334
335
static int dst_fetch_ha(struct dst_entry *dst, struct rdma_dev_addr *dev_addr,
			const void *daddr)
336
337
{
	struct neighbour *n;
338
	int ret = 0;
339

340
341
	n = dst_neigh_lookup(dst, daddr);

342
343
344
345
346
347
	rcu_read_lock();
	if (!n || !(n->nud_state & NUD_VALID)) {
		if (n)
			neigh_event_send(n, NULL);
		ret = -ENODATA;
	} else {
348
		rdma_copy_addr(dev_addr, dst->dev, n->ha);
349
350
351
	}
	rcu_read_unlock();

352
353
354
	if (n)
		neigh_release(n);

355
356
357
	return ret;
}

358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
static bool has_gateway(struct dst_entry *dst, sa_family_t family)
{
	struct rtable *rt;
	struct rt6_info *rt6;

	if (family == AF_INET) {
		rt = container_of(dst, struct rtable, dst);
		return rt->rt_uses_gateway;
	}

	rt6 = container_of(dst, struct rt6_info, dst);
	return rt6->rt6i_flags & RTF_GATEWAY;
}

static int fetch_ha(struct dst_entry *dst, struct rdma_dev_addr *dev_addr,
		    const struct sockaddr *dst_in, u32 seq)
{
	const struct sockaddr_in *dst_in4 =
		(const struct sockaddr_in *)dst_in;
	const struct sockaddr_in6 *dst_in6 =
		(const struct sockaddr_in6 *)dst_in;
	const void *daddr = (dst_in->sa_family == AF_INET) ?
		(const void *)&dst_in4->sin_addr.s_addr :
		(const void *)&dst_in6->sin6_addr;
	sa_family_t family = dst_in->sa_family;

	/* Gateway + ARPHRD_INFINIBAND -> IB router */
	if (has_gateway(dst, family) && dst->dev->type == ARPHRD_INFINIBAND)
		return ib_nl_fetch_ha(dst, dev_addr, daddr, seq, family);
	else
		return dst_fetch_ha(dst, dev_addr, daddr);
}

391
static int addr4_resolve(struct sockaddr_in *src_in,
392
393
394
			 const struct sockaddr_in *dst_in,
			 struct rdma_dev_addr *addr,
			 struct rtable **prt)
395
{
396
397
	__be32 src_ip = src_in->sin_addr.s_addr;
	__be32 dst_ip = dst_in->sin_addr.s_addr;
398
	struct rtable *rt;
399
	struct flowi4 fl4;
400
401
	int ret;

402
403
404
405
	memset(&fl4, 0, sizeof(fl4));
	fl4.daddr = dst_ip;
	fl4.saddr = src_ip;
	fl4.flowi4_oif = addr->bound_dev_if;
406
	rt = ip_route_output_key(addr->net, &fl4);
407
408
409
410
	ret = PTR_ERR_OR_ZERO(rt);
	if (ret)
		return ret;

411
	src_in->sin_family = AF_INET;
412
	src_in->sin_addr.s_addr = fl4.saddr;
413

414
415
416
	/* If there's a gateway and type of device not ARPHRD_INFINIBAND, we're
	 * definitely in RoCE v2 (as RoCE v1 isn't routable) set the network
	 * type accordingly.
417
	 */
418
	if (rt->rt_uses_gateway && rt->dst.dev->type != ARPHRD_INFINIBAND)
419
420
		addr->network = RDMA_NETWORK_IPV4;

421
422
	addr->hoplimit = ip4_dst_hoplimit(&rt->dst);

423
424
	*prt = rt;
	return 0;
425
426
}

427
#if IS_ENABLED(CONFIG_IPV6)
Sean Hefty's avatar
Sean Hefty committed
428
static int addr6_resolve(struct sockaddr_in6 *src_in,
429
430
431
			 const struct sockaddr_in6 *dst_in,
			 struct rdma_dev_addr *addr,
			 struct dst_entry **pdst)
432
{
433
	struct flowi6 fl6;
434
	struct dst_entry *dst;
435
	struct rt6_info *rt;
Sean Hefty's avatar
Sean Hefty committed
436
	int ret;
437

438
	memset(&fl6, 0, sizeof fl6);
Alexey Dobriyan's avatar
Alexey Dobriyan committed
439
440
	fl6.daddr = dst_in->sin6_addr;
	fl6.saddr = src_in->sin6_addr;
441
	fl6.flowi6_oif = addr->bound_dev_if;
442

443
444
	ret = ipv6_stub->ipv6_dst_lookup(addr->net, NULL, &dst, &fl6);
	if (ret < 0)
445
		return ret;
Sean Hefty's avatar
Sean Hefty committed
446

447
	rt = (struct rt6_info *)dst;
448
	if (ipv6_addr_any(&src_in->sin6_addr)) {
Sean Hefty's avatar
Sean Hefty committed
449
		src_in->sin6_family = AF_INET6;
Alexey Dobriyan's avatar
Alexey Dobriyan committed
450
		src_in->sin6_addr = fl6.saddr;
Sean Hefty's avatar
Sean Hefty committed
451
452
	}

453
454
455
	/* If there's a gateway and type of device not ARPHRD_INFINIBAND, we're
	 * definitely in RoCE v2 (as RoCE v1 isn't routable) set the network
	 * type accordingly.
456
	 */
457
458
	if (rt->rt6i_flags & RTF_GATEWAY &&
	    ip6_dst_idev(dst)->dev->type != ARPHRD_INFINIBAND)
459
460
		addr->network = RDMA_NETWORK_IPV6;

461
462
	addr->hoplimit = ip6_dst_hoplimit(dst);

463
464
	*pdst = dst;
	return 0;
465
}
466
#else
Sean Hefty's avatar
Sean Hefty committed
467
static int addr6_resolve(struct sockaddr_in6 *src_in,
468
469
470
			 const struct sockaddr_in6 *dst_in,
			 struct rdma_dev_addr *addr,
			 struct dst_entry **pdst)
471
472
473
474
{
	return -EADDRNOTAVAIL;
}
#endif
475

476
477
static int addr_resolve_neigh(struct dst_entry *dst,
			      const struct sockaddr *dst_in,
478
479
			      struct rdma_dev_addr *addr,
			      u32 seq)
480
481
482
483
484
485
486
487
488
489
490
491
492
{
	if (dst->dev->flags & IFF_LOOPBACK) {
		int ret;

		ret = rdma_translate_ip(dst_in, addr, NULL);
		if (!ret)
			memcpy(addr->dst_dev_addr, addr->src_dev_addr,
			       MAX_ADDR_LEN);

		return ret;
	}

	/* If the device doesn't do ARP internally */
493
494
	if (!(dst->dev->flags & IFF_NOARP))
		return fetch_ha(dst, addr, dst_in, seq);
495

496
497
498
	rdma_copy_addr(addr, dst->dev, NULL);

	return 0;
499
500
}

501
static int addr_resolve(struct sockaddr *src_in,
502
503
			const struct sockaddr *dst_in,
			struct rdma_dev_addr *addr,
504
505
			bool resolve_neigh,
			u32 seq)
506
{
507
508
509
510
	struct net_device *ndev;
	struct dst_entry *dst;
	int ret;

511
512
513
514
515
	if (!addr->net) {
		pr_warn_ratelimited("%s: missing namespace\n", __func__);
		return -EINVAL;
	}

516
	if (src_in->sa_family == AF_INET) {
517
518
519
520
521
522
523
524
525
526
		struct rtable *rt = NULL;
		const struct sockaddr_in *dst_in4 =
			(const struct sockaddr_in *)dst_in;

		ret = addr4_resolve((struct sockaddr_in *)src_in,
				    dst_in4, addr, &rt);
		if (ret)
			return ret;

		if (resolve_neigh)
527
			ret = addr_resolve_neigh(&rt->dst, dst_in, addr, seq);
528

529
530
531
532
533
534
		if (addr->bound_dev_if) {
			ndev = dev_get_by_index(addr->net, addr->bound_dev_if);
		} else {
			ndev = rt->dst.dev;
			dev_hold(ndev);
		}
535
536
537
538
539
540
541
542
543
544
545
546
547

		ip_rt_put(rt);
	} else {
		const struct sockaddr_in6 *dst_in6 =
			(const struct sockaddr_in6 *)dst_in;

		ret = addr6_resolve((struct sockaddr_in6 *)src_in,
				    dst_in6, addr,
				    &dst);
		if (ret)
			return ret;

		if (resolve_neigh)
548
			ret = addr_resolve_neigh(dst, dst_in, addr, seq);
549

550
551
552
553
554
555
		if (addr->bound_dev_if) {
			ndev = dev_get_by_index(addr->net, addr->bound_dev_if);
		} else {
			ndev = dst->dev;
			dev_hold(ndev);
		}
556
557
558
559

		dst_release(dst);
	}

560
561
562
563
564
565
566
567
568
569
570
	if (ndev->flags & IFF_LOOPBACK) {
		ret = rdma_translate_ip(dst_in, addr, NULL);
		/*
		 * Put the loopback device and get the translated
		 * device instead.
		 */
		dev_put(ndev);
		ndev = dev_get_by_index(addr->net, addr->bound_dev_if);
	} else {
		addr->bound_dev_if = ndev->ifindex;
	}
571
572
573
	dev_put(ndev);

	return ret;
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
static void process_one_req(struct work_struct *_work)
{
	struct addr_req *req;
	struct sockaddr *src_in, *dst_in;

	mutex_lock(&lock);
	req = container_of(_work, struct addr_req, work.work);

	if (req->status == -ENODATA) {
		src_in = (struct sockaddr *)&req->src_addr;
		dst_in = (struct sockaddr *)&req->dst_addr;
		req->status = addr_resolve(src_in, dst_in, req->addr,
					   true, req->seq);
		if (req->status && time_after_eq(jiffies, req->timeout)) {
			req->status = -ETIMEDOUT;
		} else if (req->status == -ENODATA) {
			/* requeue the work for retrying again */
			set_timeout(&req->work, req->timeout);
			mutex_unlock(&lock);
			return;
		}
	}
	list_del(&req->list);
	mutex_unlock(&lock);

	req->callback(req->status, (struct sockaddr *)&req->src_addr,
		req->addr, req->context);
	put_client(req->client);
	kfree(req);
}

David Howells's avatar
David Howells committed
607
static void process_req(struct work_struct *work)
608
609
{
	struct addr_req *req, *temp_req;
610
	struct sockaddr *src_in, *dst_in;
611
612
613
614
615
616
	struct list_head done_list;

	INIT_LIST_HEAD(&done_list);

	mutex_lock(&lock);
	list_for_each_entry_safe(req, temp_req, &req_list, list) {
617
		if (req->status == -ENODATA) {
618
619
			src_in = (struct sockaddr *) &req->src_addr;
			dst_in = (struct sockaddr *) &req->dst_addr;
620
			req->status = addr_resolve(src_in, dst_in, req->addr,
621
						   true, req->seq);
622
623
			if (req->status && time_after_eq(jiffies, req->timeout))
				req->status = -ETIMEDOUT;
624
625
			else if (req->status == -ENODATA) {
				set_timeout(&req->work, req->timeout);
626
				continue;
627
			}
628
		}
629
		list_move_tail(&req->list, &done_list);
630
631
632
633
634
635
	}

	mutex_unlock(&lock);

	list_for_each_entry_safe(req, temp_req, &done_list, list) {
		list_del(&req->list);
636
637
638
639
640
		/* It is safe to cancel other work items from this work item
		 * because at a time there can be only one work item running
		 * with this single threaded work queue.
		 */
		cancel_delayed_work(&req->work);
641
642
		req->callback(req->status, (struct sockaddr *) &req->src_addr,
			req->addr, req->context);
643
		put_client(req->client);
644
645
646
647
		kfree(req);
	}
}

648
649
int rdma_resolve_ip(struct rdma_addr_client *client,
		    struct sockaddr *src_addr, struct sockaddr *dst_addr,
650
651
652
653
654
		    struct rdma_dev_addr *addr, int timeout_ms,
		    void (*callback)(int status, struct sockaddr *src_addr,
				     struct rdma_dev_addr *addr, void *context),
		    void *context)
{
655
	struct sockaddr *src_in, *dst_in;
656
657
658
	struct addr_req *req;
	int ret = 0;

659
	req = kzalloc(sizeof *req, GFP_KERNEL);
660
661
662
	if (!req)
		return -ENOMEM;

663
664
665
666
667
668
669
670
671
	src_in = (struct sockaddr *) &req->src_addr;
	dst_in = (struct sockaddr *) &req->dst_addr;

	if (src_addr) {
		if (src_addr->sa_family != dst_addr->sa_family) {
			ret = -EINVAL;
			goto err;
		}

672
		memcpy(src_in, src_addr, rdma_addr_size(src_addr));
673
674
675
676
	} else {
		src_in->sa_family = dst_addr->sa_family;
	}

677
	memcpy(dst_in, dst_addr, rdma_addr_size(dst_addr));
678
679
680
	req->addr = addr;
	req->callback = callback;
	req->context = context;
681
682
	req->client = client;
	atomic_inc(&client->refcount);
683
	INIT_DELAYED_WORK(&req->work, process_one_req);
684
	req->seq = (u32)atomic_inc_return(&ib_nl_addr_request_seq);
685

686
	req->status = addr_resolve(src_in, dst_in, addr, true, req->seq);
687
688
689
690
691
692
693
694
695
696
697
	switch (req->status) {
	case 0:
		req->timeout = jiffies;
		queue_req(req);
		break;
	case -ENODATA:
		req->timeout = msecs_to_jiffies(timeout_ms) + jiffies;
		queue_req(req);
		break;
	default:
		ret = req->status;
698
		atomic_dec(&client->refcount);
699
		goto err;
700
701
	}
	return ret;
702
703
704
err:
	kfree(req);
	return ret;
705
706
707
}
EXPORT_SYMBOL(rdma_resolve_ip);

708
709
710
711
712
713
714
int rdma_resolve_ip_route(struct sockaddr *src_addr,
			  const struct sockaddr *dst_addr,
			  struct rdma_dev_addr *addr)
{
	struct sockaddr_storage ssrc_addr = {};
	struct sockaddr *src_in = (struct sockaddr *)&ssrc_addr;

715
716
717
	if (src_addr) {
		if (src_addr->sa_family != dst_addr->sa_family)
			return -EINVAL;
718
719

		memcpy(src_in, src_addr, rdma_addr_size(src_addr));
720
	} else {
721
		src_in->sa_family = dst_addr->sa_family;
722
	}
723

724
	return addr_resolve(src_in, dst_addr, addr, false, 0);
725
726
727
}
EXPORT_SYMBOL(rdma_resolve_ip_route);

728
729
730
731
732
733
734
735
736
void rdma_addr_cancel(struct rdma_dev_addr *addr)
{
	struct addr_req *req, *temp_req;

	mutex_lock(&lock);
	list_for_each_entry_safe(req, temp_req, &req_list, list) {
		if (req->addr == addr) {
			req->status = -ECANCELED;
			req->timeout = jiffies;
737
			list_move(&req->list, &req_list);
738
			set_timeout(&req->work, req->timeout);
739
740
741
742
743
744
745
			break;
		}
	}
	mutex_unlock(&lock);
}
EXPORT_SYMBOL(rdma_addr_cancel);

746
747
748
struct resolve_cb_context {
	struct rdma_dev_addr *addr;
	struct completion comp;
749
	int status;
750
751
752
753
754
};

static void resolve_cb(int status, struct sockaddr *src_addr,
	     struct rdma_dev_addr *addr, void *context)
{
755
756
757
758
	if (!status)
		memcpy(((struct resolve_cb_context *)context)->addr,
		       addr, sizeof(struct rdma_dev_addr));
	((struct resolve_cb_context *)context)->status = status;
759
760
761
	complete(&((struct resolve_cb_context *)context)->comp);
}

762
763
int rdma_addr_find_l2_eth_by_grh(const union ib_gid *sgid,
				 const union ib_gid *dgid,
764
765
				 u8 *dmac, u16 *vlan_id, int *if_index,
				 int *hoplimit)
766
767
768
769
770
771
772
773
774
775
776
777
778
{
	int ret = 0;
	struct rdma_dev_addr dev_addr;
	struct resolve_cb_context ctx;
	struct net_device *dev;

	union {
		struct sockaddr     _sockaddr;
		struct sockaddr_in  _sockaddr_in;
		struct sockaddr_in6 _sockaddr_in6;
	} sgid_addr, dgid_addr;


779
780
	rdma_gid2ip(&sgid_addr._sockaddr, sgid);
	rdma_gid2ip(&dgid_addr._sockaddr, dgid);
781
782

	memset(&dev_addr, 0, sizeof(dev_addr));
783
784
	if (if_index)
		dev_addr.bound_dev_if = *if_index;
785
	dev_addr.net = &init_net;
786
787
788
789
790
791
792
793
794
795

	ctx.addr = &dev_addr;
	init_completion(&ctx.comp);
	ret = rdma_resolve_ip(&self, &sgid_addr._sockaddr, &dgid_addr._sockaddr,
			&dev_addr, 1000, resolve_cb, &ctx);
	if (ret)
		return ret;

	wait_for_completion(&ctx.comp);

796
797
798
799
	ret = ctx.status;
	if (ret)
		return ret;

800
801
802
803
	memcpy(dmac, dev_addr.dst_dev_addr, ETH_ALEN);
	dev = dev_get_by_index(&init_net, dev_addr.bound_dev_if);
	if (!dev)
		return -ENODEV;
804
805
	if (if_index)
		*if_index = dev_addr.bound_dev_if;
806
807
	if (vlan_id)
		*vlan_id = rdma_vlan_dev_vlan_id(dev);
808
809
	if (hoplimit)
		*hoplimit = dev_addr.hoplimit;
810
811
812
	dev_put(dev);
	return ret;
}
813
EXPORT_SYMBOL(rdma_addr_find_l2_eth_by_grh);
814
815
816
817
818
819
820
821
822
823
824

int rdma_addr_find_smac_by_sgid(union ib_gid *sgid, u8 *smac, u16 *vlan_id)
{
	int ret = 0;
	struct rdma_dev_addr dev_addr;
	union {
		struct sockaddr     _sockaddr;
		struct sockaddr_in  _sockaddr_in;
		struct sockaddr_in6 _sockaddr_in6;
	} gid_addr;

825
	rdma_gid2ip(&gid_addr._sockaddr, sgid);
826
827

	memset(&dev_addr, 0, sizeof(dev_addr));
828
	dev_addr.net = &init_net;
829
830
831
832
833
834
835
836
837
	ret = rdma_translate_ip(&gid_addr._sockaddr, &dev_addr, vlan_id);
	if (ret)
		return ret;

	memcpy(smac, dev_addr.src_dev_addr, ETH_ALEN);
	return ret;
}
EXPORT_SYMBOL(rdma_addr_find_smac_by_sgid);

Roland Dreier's avatar
Roland Dreier committed
838
static int netevent_callback(struct notifier_block *self, unsigned long event,
839
	void *ctx)
840
{
Roland Dreier's avatar
Roland Dreier committed
841
	if (event == NETEVENT_NEIGH_UPDATE) {
842
		struct neighbour *neigh = ctx;
843

844
845
		if (neigh->nud_state & NUD_VALID)
			set_timeout(&work, jiffies);
846
	}
847
848
849
	return 0;
}

850
851
static struct notifier_block nb = {
	.notifier_call = netevent_callback
852
853
};

854
int addr_init(void)
855
{
856
	addr_wq = alloc_ordered_workqueue("ib_addr", 0);
857
858
859
	if (!addr_wq)
		return -ENOMEM;

860
	register_netevent_notifier(&nb);
861
	rdma_addr_register_client(&self);
862

863
864
865
	return 0;
}

866
void addr_cleanup(void)
867
{
868
	rdma_addr_unregister_client(&self);
869
	unregister_netevent_notifier(&nb);
870
871
	destroy_workqueue(addr_wq);
}