addr.c 21.2 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);

Tom Tucker's avatar
Tom Tucker committed
232
233
int rdma_copy_addr(struct rdma_dev_addr *dev_addr, struct net_device *dev,
		     const unsigned char *dst_dev_addr)
234
{
235
	dev_addr->dev_type = dev->type;
236
237
238
239
	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);
240
	dev_addr->bound_dev_if = dev->ifindex;
241
242
	return 0;
}
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
{
	struct net_device *dev;
250
	int ret = -EADDRNOTAVAIL;
251

252
	if (dev_addr->bound_dev_if) {
253
		dev = dev_get_by_index(dev_addr->net, dev_addr->bound_dev_if);
254
255
256
257
258
259
260
		if (!dev)
			return -ENODEV;
		ret = rdma_copy_addr(dev_addr, dev, NULL);
		dev_put(dev);
		return ret;
	}

261
262
	switch (addr->sa_family) {
	case AF_INET:
263
		dev = ip_dev_find(dev_addr->net,
264
			((const struct sockaddr_in *)addr)->sin_addr.s_addr);
265
266
267

		if (!dev)
			return ret;
268

269
		ret = rdma_copy_addr(dev_addr, dev, NULL);
270
		dev_addr->bound_dev_if = dev->ifindex;
271
272
		if (vlan_id)
			*vlan_id = rdma_vlan_dev_vlan_id(dev);
273
274
		dev_put(dev);
		break;
275
#if IS_ENABLED(CONFIG_IPV6)
276
	case AF_INET6:
277
		rcu_read_lock();
278
279
		for_each_netdev_rcu(dev_addr->net, dev) {
			if (ipv6_chk_addr(dev_addr->net,
280
					  &((const struct sockaddr_in6 *)addr)->sin6_addr,
281
282
					  dev, 1)) {
				ret = rdma_copy_addr(dev_addr, dev, NULL);
283
				dev_addr->bound_dev_if = dev->ifindex;
284
285
				if (vlan_id)
					*vlan_id = rdma_vlan_dev_vlan_id(dev);
286
287
288
				break;
			}
		}
289
		rcu_read_unlock();
290
		break;
291
#endif
292
	}
293
294
295
296
	return ret;
}
EXPORT_SYMBOL(rdma_translate_ip);

297
static void set_timeout(struct delayed_work *delayed_work, unsigned long time)
298
299
300
301
{
	unsigned long delay;

	delay = time - jiffies;
302
303
	if ((long)delay < 0)
		delay = 0;
304

305
	mod_delayed_work(addr_wq, delayed_work, delay);
306
307
308
309
310
311
312
313
}

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) {
314
		if (time_after_eq(req->timeout, temp_req->timeout))
315
316
317
318
319
			break;
	}

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

320
	set_timeout(&req->work, req->timeout);
321
322
323
	mutex_unlock(&lock);
}

324
325
326
static int ib_nl_fetch_ha(struct dst_entry *dst, struct rdma_dev_addr *dev_addr,
			  const void *daddr, u32 seq, u16 family)
{
327
	if (rdma_nl_chk_listeners(RDMA_NL_GROUP_LS))
328
329
330
331
332
333
334
		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);
}

335
336
static int dst_fetch_ha(struct dst_entry *dst, struct rdma_dev_addr *dev_addr,
			const void *daddr)
337
338
339
340
{
	struct neighbour *n;
	int ret;

341
342
	n = dst_neigh_lookup(dst, daddr);

343
344
345
346
347
348
	rcu_read_lock();
	if (!n || !(n->nud_state & NUD_VALID)) {
		if (n)
			neigh_event_send(n, NULL);
		ret = -ENODATA;
	} else {
349
		ret = rdma_copy_addr(dev_addr, dst->dev, n->ha);
350
351
352
	}
	rcu_read_unlock();

353
354
355
	if (n)
		neigh_release(n);

356
357
358
	return ret;
}

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
391
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);
}

392
static int addr4_resolve(struct sockaddr_in *src_in,
393
394
395
			 const struct sockaddr_in *dst_in,
			 struct rdma_dev_addr *addr,
			 struct rtable **prt)
396
{
397
398
	__be32 src_ip = src_in->sin_addr.s_addr;
	__be32 dst_ip = dst_in->sin_addr.s_addr;
399
	struct rtable *rt;
400
	struct flowi4 fl4;
401
402
	int ret;

403
404
405
406
	memset(&fl4, 0, sizeof(fl4));
	fl4.daddr = dst_ip;
	fl4.saddr = src_ip;
	fl4.flowi4_oif = addr->bound_dev_if;
407
	rt = ip_route_output_key(addr->net, &fl4);
408
409
410
411
	ret = PTR_ERR_OR_ZERO(rt);
	if (ret)
		return ret;

412
	src_in->sin_family = AF_INET;
413
	src_in->sin_addr.s_addr = fl4.saddr;
414

415
416
417
	/* 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.
418
	 */
419
	if (rt->rt_uses_gateway && rt->dst.dev->type != ARPHRD_INFINIBAND)
420
421
		addr->network = RDMA_NETWORK_IPV4;

422
423
	addr->hoplimit = ip4_dst_hoplimit(&rt->dst);

424
425
	*prt = rt;
	return 0;
426
427
}

428
#if IS_ENABLED(CONFIG_IPV6)
Sean Hefty's avatar
Sean Hefty committed
429
static int addr6_resolve(struct sockaddr_in6 *src_in,
430
431
432
			 const struct sockaddr_in6 *dst_in,
			 struct rdma_dev_addr *addr,
			 struct dst_entry **pdst)
433
{
434
	struct flowi6 fl6;
435
	struct dst_entry *dst;
436
	struct rt6_info *rt;
Sean Hefty's avatar
Sean Hefty committed
437
	int ret;
438

439
	memset(&fl6, 0, sizeof fl6);
Alexey Dobriyan's avatar
Alexey Dobriyan committed
440
441
	fl6.daddr = dst_in->sin6_addr;
	fl6.saddr = src_in->sin6_addr;
442
	fl6.flowi6_oif = addr->bound_dev_if;
443

444
445
	ret = ipv6_stub->ipv6_dst_lookup(addr->net, NULL, &dst, &fl6);
	if (ret < 0)
446
		return ret;
Sean Hefty's avatar
Sean Hefty committed
447

448
	rt = (struct rt6_info *)dst;
449
	if (ipv6_addr_any(&src_in->sin6_addr)) {
Sean Hefty's avatar
Sean Hefty committed
450
		src_in->sin6_family = AF_INET6;
Alexey Dobriyan's avatar
Alexey Dobriyan committed
451
		src_in->sin6_addr = fl6.saddr;
Sean Hefty's avatar
Sean Hefty committed
452
453
	}

454
455
456
	/* 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.
457
	 */
458
459
	if (rt->rt6i_flags & RTF_GATEWAY &&
	    ip6_dst_idev(dst)->dev->type != ARPHRD_INFINIBAND)
460
461
		addr->network = RDMA_NETWORK_IPV6;

462
463
	addr->hoplimit = ip6_dst_hoplimit(dst);

464
465
	*pdst = dst;
	return 0;
466
}
467
#else
Sean Hefty's avatar
Sean Hefty committed
468
static int addr6_resolve(struct sockaddr_in6 *src_in,
469
470
471
			 const struct sockaddr_in6 *dst_in,
			 struct rdma_dev_addr *addr,
			 struct dst_entry **pdst)
472
473
474
475
{
	return -EADDRNOTAVAIL;
}
#endif
476

477
478
static int addr_resolve_neigh(struct dst_entry *dst,
			      const struct sockaddr *dst_in,
479
480
			      struct rdma_dev_addr *addr,
			      u32 seq)
481
482
483
484
485
486
487
488
489
490
491
492
493
{
	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 */
494
495
	if (!(dst->dev->flags & IFF_NOARP))
		return fetch_ha(dst, addr, dst_in, seq);
496
497
498
499

	return rdma_copy_addr(addr, dst->dev, NULL);
}

500
static int addr_resolve(struct sockaddr *src_in,
501
502
			const struct sockaddr *dst_in,
			struct rdma_dev_addr *addr,
503
504
			bool resolve_neigh,
			u32 seq)
505
{
506
507
508
509
	struct net_device *ndev;
	struct dst_entry *dst;
	int ret;

510
511
512
513
514
	if (!addr->net) {
		pr_warn_ratelimited("%s: missing namespace\n", __func__);
		return -EINVAL;
	}

515
	if (src_in->sa_family == AF_INET) {
516
517
518
519
520
521
522
523
524
525
		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)
526
			ret = addr_resolve_neigh(&rt->dst, dst_in, addr, seq);
527

528
529
530
531
532
533
		if (addr->bound_dev_if) {
			ndev = dev_get_by_index(addr->net, addr->bound_dev_if);
		} else {
			ndev = rt->dst.dev;
			dev_hold(ndev);
		}
534
535
536
537
538
539
540
541
542
543
544
545
546

		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)
547
			ret = addr_resolve_neigh(dst, dst_in, addr, seq);
548

549
550
551
552
553
554
		if (addr->bound_dev_if) {
			ndev = dev_get_by_index(addr->net, addr->bound_dev_if);
		} else {
			ndev = dst->dev;
			dev_hold(ndev);
		}
555
556
557
558

		dst_release(dst);
	}

559
560
561
562
563
564
565
566
567
568
569
	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;
	}
570
571
572
	dev_put(ndev);

	return ret;
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
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
606
static void process_req(struct work_struct *work)
607
608
{
	struct addr_req *req, *temp_req;
609
	struct sockaddr *src_in, *dst_in;
610
611
612
613
614
615
	struct list_head done_list;

	INIT_LIST_HEAD(&done_list);

	mutex_lock(&lock);
	list_for_each_entry_safe(req, temp_req, &req_list, list) {
616
		if (req->status == -ENODATA) {
617
618
			src_in = (struct sockaddr *) &req->src_addr;
			dst_in = (struct sockaddr *) &req->dst_addr;
619
			req->status = addr_resolve(src_in, dst_in, req->addr,
620
						   true, req->seq);
621
622
			if (req->status && time_after_eq(jiffies, req->timeout))
				req->status = -ETIMEDOUT;
623
624
			else if (req->status == -ENODATA) {
				set_timeout(&req->work, req->timeout);
625
				continue;
626
			}
627
		}
628
		list_move_tail(&req->list, &done_list);
629
630
631
632
633
634
	}

	mutex_unlock(&lock);

	list_for_each_entry_safe(req, temp_req, &done_list, list) {
		list_del(&req->list);
635
636
637
638
639
		/* 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);
640
641
		req->callback(req->status, (struct sockaddr *) &req->src_addr,
			req->addr, req->context);
642
		put_client(req->client);
643
644
645
646
		kfree(req);
	}
}

647
648
int rdma_resolve_ip(struct rdma_addr_client *client,
		    struct sockaddr *src_addr, struct sockaddr *dst_addr,
649
650
651
652
653
		    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)
{
654
	struct sockaddr *src_in, *dst_in;
655
656
657
	struct addr_req *req;
	int ret = 0;

658
	req = kzalloc(sizeof *req, GFP_KERNEL);
659
660
661
	if (!req)
		return -ENOMEM;

662
663
664
665
666
667
668
669
670
	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;
		}

671
		memcpy(src_in, src_addr, rdma_addr_size(src_addr));
672
673
674
675
	} else {
		src_in->sa_family = dst_addr->sa_family;
	}

676
	memcpy(dst_in, dst_addr, rdma_addr_size(dst_addr));
677
678
679
	req->addr = addr;
	req->callback = callback;
	req->context = context;
680
681
	req->client = client;
	atomic_inc(&client->refcount);
682
	INIT_DELAYED_WORK(&req->work, process_one_req);
683
	req->seq = (u32)atomic_inc_return(&ib_nl_addr_request_seq);
684

685
	req->status = addr_resolve(src_in, dst_in, addr, true, req->seq);
686
687
688
689
690
691
692
693
694
695
696
	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;
697
		atomic_dec(&client->refcount);
698
		goto err;
699
700
	}
	return ret;
701
702
703
err:
	kfree(req);
	return ret;
704
705
706
}
EXPORT_SYMBOL(rdma_resolve_ip);

707
708
709
710
711
712
713
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;

714
715
716
	if (src_addr) {
		if (src_addr->sa_family != dst_addr->sa_family)
			return -EINVAL;
717
718

		memcpy(src_in, src_addr, rdma_addr_size(src_addr));
719
	} else {
720
		src_in->sa_family = dst_addr->sa_family;
721
	}
722

723
	return addr_resolve(src_in, dst_addr, addr, false, 0);
724
725
726
}
EXPORT_SYMBOL(rdma_resolve_ip_route);

727
728
729
730
731
732
733
734
735
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;
736
			list_move(&req->list, &req_list);
737
			set_timeout(&req->work, req->timeout);
738
739
740
741
742
743
744
			break;
		}
	}
	mutex_unlock(&lock);
}
EXPORT_SYMBOL(rdma_addr_cancel);

745
746
747
struct resolve_cb_context {
	struct rdma_dev_addr *addr;
	struct completion comp;
748
	int status;
749
750
751
752
753
};

static void resolve_cb(int status, struct sockaddr *src_addr,
	     struct rdma_dev_addr *addr, void *context)
{
754
755
756
757
	if (!status)
		memcpy(((struct resolve_cb_context *)context)->addr,
		       addr, sizeof(struct rdma_dev_addr));
	((struct resolve_cb_context *)context)->status = status;
758
759
760
	complete(&((struct resolve_cb_context *)context)->comp);
}

761
762
int rdma_addr_find_l2_eth_by_grh(const union ib_gid *sgid,
				 const union ib_gid *dgid,
763
764
				 u8 *dmac, u16 *vlan_id, int *if_index,
				 int *hoplimit)
765
766
767
768
769
770
771
772
773
774
775
776
777
{
	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;


778
779
	rdma_gid2ip(&sgid_addr._sockaddr, sgid);
	rdma_gid2ip(&dgid_addr._sockaddr, dgid);
780
781

	memset(&dev_addr, 0, sizeof(dev_addr));
782
783
	if (if_index)
		dev_addr.bound_dev_if = *if_index;
784
	dev_addr.net = &init_net;
785
786
787
788
789
790
791
792
793
794

	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);

795
796
797
798
	ret = ctx.status;
	if (ret)
		return ret;

799
800
801
802
	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;
803
804
	if (if_index)
		*if_index = dev_addr.bound_dev_if;
805
806
	if (vlan_id)
		*vlan_id = rdma_vlan_dev_vlan_id(dev);
807
808
	if (hoplimit)
		*hoplimit = dev_addr.hoplimit;
809
810
811
	dev_put(dev);
	return ret;
}
812
EXPORT_SYMBOL(rdma_addr_find_l2_eth_by_grh);
813
814
815
816
817
818
819
820
821
822
823

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;

824
	rdma_gid2ip(&gid_addr._sockaddr, sgid);
825
826

	memset(&dev_addr, 0, sizeof(dev_addr));
827
	dev_addr.net = &init_net;
828
829
830
831
832
833
834
835
836
	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
837
static int netevent_callback(struct notifier_block *self, unsigned long event,
838
	void *ctx)
839
{
Roland Dreier's avatar
Roland Dreier committed
840
	if (event == NETEVENT_NEIGH_UPDATE) {
841
		struct neighbour *neigh = ctx;
842

843
844
		if (neigh->nud_state & NUD_VALID)
			set_timeout(&work, jiffies);
845
	}
846
847
848
	return 0;
}

849
850
static struct notifier_block nb = {
	.notifier_call = netevent_callback
851
852
};

853
int addr_init(void)
854
{
855
	addr_wq = alloc_ordered_workqueue("ib_addr", 0);
856
857
858
	if (!addr_wq)
		return -ENOMEM;

859
	register_netevent_notifier(&nb);
860
	rdma_addr_register_client(&self);
861

862
863
864
	return 0;
}

865
void addr_cleanup(void)
866
{
867
	rdma_addr_unregister_client(&self);
868
	unregister_netevent_notifier(&nb);
869
870
	destroy_workqueue(addr_wq);
}