addr.c 11.1 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.
 *
S
Sean Hefty 已提交
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
 *
S
Sean Hefty 已提交
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
 *
S
Sean Hefty 已提交
17 18 19
 *      - Redistributions of source code must retain the above
 *        copyright notice, this list of conditions and the following
 *        disclaimer.
20
 *
S
Sean Hefty 已提交
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
 *
S
Sean Hefty 已提交
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 40 41 42
#include <linux/workqueue.h>
#include <net/arp.h>
#include <net/neighbour.h>
#include <net/route.h>
43
#include <net/netevent.h>
44 45
#include <net/addrconf.h>
#include <net/ip6_route.h>
46 47 48 49 50 51 52 53
#include <rdma/ib_addr.h>

MODULE_AUTHOR("Sean Hefty");
MODULE_DESCRIPTION("IB Address Translation");
MODULE_LICENSE("Dual BSD/GPL");

struct addr_req {
	struct list_head list;
54 55
	struct sockaddr_storage src_addr;
	struct sockaddr_storage dst_addr;
56
	struct rdma_dev_addr *addr;
57
	struct rdma_addr_client *client;
58 59 60 61 62 63 64
	void *context;
	void (*callback)(int status, struct sockaddr *src_addr,
			 struct rdma_dev_addr *addr, void *context);
	unsigned long timeout;
	int status;
};

65
static void process_req(struct work_struct *work);
66 67 68

static DEFINE_MUTEX(lock);
static LIST_HEAD(req_list);
69
static DECLARE_DELAYED_WORK(work, process_req);
70 71
static struct workqueue_struct *addr_wq;

72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91
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);

T
Tom Tucker 已提交
92 93
int rdma_copy_addr(struct rdma_dev_addr *dev_addr, struct net_device *dev,
		     const unsigned char *dst_dev_addr)
94
{
95
	dev_addr->dev_type = dev->type;
96 97 98 99
	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);
100
	dev_addr->bound_dev_if = dev->ifindex;
101 102
	return 0;
}
T
Tom Tucker 已提交
103
EXPORT_SYMBOL(rdma_copy_addr);
104 105 106 107

int rdma_translate_ip(struct sockaddr *addr, struct rdma_dev_addr *dev_addr)
{
	struct net_device *dev;
108
	int ret = -EADDRNOTAVAIL;
109

110 111 112 113 114 115 116 117 118
	if (dev_addr->bound_dev_if) {
		dev = dev_get_by_index(&init_net, dev_addr->bound_dev_if);
		if (!dev)
			return -ENODEV;
		ret = rdma_copy_addr(dev_addr, dev, NULL);
		dev_put(dev);
		return ret;
	}

119 120 121 122 123 124 125
	switch (addr->sa_family) {
	case AF_INET:
		dev = ip_dev_find(&init_net,
			((struct sockaddr_in *) addr)->sin_addr.s_addr);

		if (!dev)
			return ret;
126

127 128 129
		ret = rdma_copy_addr(dev_addr, dev, NULL);
		dev_put(dev);
		break;
130 131

#if defined(CONFIG_IPV6) || defined(CONFIG_IPV6_MODULE)
132
	case AF_INET6:
133 134
		rcu_read_lock();
		for_each_netdev_rcu(&init_net, dev) {
135 136 137 138 139 140 141
			if (ipv6_chk_addr(&init_net,
					  &((struct sockaddr_in6 *) addr)->sin6_addr,
					  dev, 1)) {
				ret = rdma_copy_addr(dev_addr, dev, NULL);
				break;
			}
		}
142
		rcu_read_unlock();
143
		break;
144
#endif
145
	}
146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168
	return ret;
}
EXPORT_SYMBOL(rdma_translate_ip);

static void set_timeout(unsigned long time)
{
	unsigned long delay;

	cancel_delayed_work(&work);

	delay = time - jiffies;
	if ((long)delay <= 0)
		delay = 1;

	queue_delayed_work(addr_wq, &work, delay);
}

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) {
169
		if (time_after_eq(req->timeout, temp_req->timeout))
170 171 172 173 174 175 176 177 178 179
			break;
	}

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

	if (req_list.next == &req->list)
		set_timeout(req->timeout);
	mutex_unlock(&lock);
}

180 181 182
static int addr4_resolve(struct sockaddr_in *src_in,
			 struct sockaddr_in *dst_in,
			 struct rdma_dev_addr *addr)
183
{
184 185
	__be32 src_ip = src_in->sin_addr.s_addr;
	__be32 dst_ip = dst_in->sin_addr.s_addr;
186 187 188 189
	struct rtable *rt;
	struct neighbour *neigh;
	int ret;

190
	rt = ip_route_output(&init_net, dst_ip, src_ip, 0, addr->bound_dev_if);
191 192
	if (IS_ERR(rt)) {
		ret = PTR_ERR(rt);
193
		goto out;
194
	}
195 196 197
	src_in->sin_family = AF_INET;
	src_in->sin_addr.s_addr = rt->rt_src;

198
	if (rt->dst.dev->flags & IFF_LOOPBACK) {
199 200 201 202 203 204
		ret = rdma_translate_ip((struct sockaddr *) dst_in, addr);
		if (!ret)
			memcpy(addr->dst_dev_addr, addr->src_dev_addr, MAX_ADDR_LEN);
		goto put;
	}

205
	/* If the device does ARP internally, return 'done' */
206
	if (rt->dst.dev->flags & IFF_NOARP) {
207
		ret = rdma_copy_addr(addr, rt->dst.dev, NULL);
208 209 210
		goto put;
	}

211
	neigh = neigh_lookup(&arp_tbl, &rt->rt_gateway, rt->dst.dev);
212
	if (!neigh || !(neigh->nud_state & NUD_VALID)) {
213
		neigh_event_send(rt->dst.neighbour, NULL);
214
		ret = -ENODATA;
215 216
		if (neigh)
			goto release;
217 218 219
		goto put;
	}

T
Tom Tucker 已提交
220
	ret = rdma_copy_addr(addr, neigh->dev, neigh->ha);
221 222 223 224 225 226 227 228
release:
	neigh_release(neigh);
put:
	ip_rt_put(rt);
out:
	return ret;
}

229
#if defined(CONFIG_IPV6) || defined(CONFIG_IPV6_MODULE)
230 231 232
static int addr6_resolve(struct sockaddr_in6 *src_in,
			 struct sockaddr_in6 *dst_in,
			 struct rdma_dev_addr *addr)
233
{
234
	struct flowi6 fl6;
235 236
	struct neighbour *neigh;
	struct dst_entry *dst;
237
	int ret;
238

239 240 241 242
	memset(&fl6, 0, sizeof fl6);
	ipv6_addr_copy(&fl6.daddr, &dst_in->sin6_addr);
	ipv6_addr_copy(&fl6.saddr, &src_in->sin6_addr);
	fl6.flowi6_oif = addr->bound_dev_if;
243

244
	dst = ip6_route_output(&init_net, NULL, &fl6);
245 246 247
	if ((ret = dst->error))
		goto put;

248
	if (ipv6_addr_any(&fl6.saddr)) {
249
		ret = ipv6_dev_get_saddr(&init_net, ip6_dst_idev(dst)->dev,
250
					 &fl6.daddr, 0, &fl6.saddr);
251 252
		if (ret)
			goto put;
253

254
		src_in->sin6_family = AF_INET6;
255
		ipv6_addr_copy(&src_in->sin6_addr, &fl6.saddr);
256 257 258 259 260 261 262 263 264 265
	}

	if (dst->dev->flags & IFF_LOOPBACK) {
		ret = rdma_translate_ip((struct sockaddr *) dst_in, addr);
		if (!ret)
			memcpy(addr->dst_dev_addr, addr->src_dev_addr, MAX_ADDR_LEN);
		goto put;
	}

	/* If the device does ARP internally, return 'done' */
266 267
	if (dst->dev->flags & IFF_NOARP) {
		ret = rdma_copy_addr(addr, dst->dev, NULL);
268
		goto put;
269 270
	}

271 272 273 274 275 276 277 278 279
	neigh = dst->neighbour;
	if (!neigh || !(neigh->nud_state & NUD_VALID)) {
		neigh_event_send(dst->neighbour, NULL);
		ret = -ENODATA;
		goto put;
	}

	ret = rdma_copy_addr(addr, dst->dev, neigh->ha);
put:
280 281 282
	dst_release(dst);
	return ret;
}
283
#else
284 285 286
static int addr6_resolve(struct sockaddr_in6 *src_in,
			 struct sockaddr_in6 *dst_in,
			 struct rdma_dev_addr *addr)
287 288 289 290
{
	return -EADDRNOTAVAIL;
}
#endif
291

292 293 294
static int addr_resolve(struct sockaddr *src_in,
			struct sockaddr *dst_in,
			struct rdma_dev_addr *addr)
295 296
{
	if (src_in->sa_family == AF_INET) {
297
		return addr4_resolve((struct sockaddr_in *) src_in,
298 299
			(struct sockaddr_in *) dst_in, addr);
	} else
300
		return addr6_resolve((struct sockaddr_in6 *) src_in,
301 302 303
			(struct sockaddr_in6 *) dst_in, addr);
}

304
static void process_req(struct work_struct *work)
305 306
{
	struct addr_req *req, *temp_req;
307
	struct sockaddr *src_in, *dst_in;
308 309 310 311 312 313
	struct list_head done_list;

	INIT_LIST_HEAD(&done_list);

	mutex_lock(&lock);
	list_for_each_entry_safe(req, temp_req, &req_list, list) {
314
		if (req->status == -ENODATA) {
315 316
			src_in = (struct sockaddr *) &req->src_addr;
			dst_in = (struct sockaddr *) &req->dst_addr;
317
			req->status = addr_resolve(src_in, dst_in, req->addr);
318 319 320 321
			if (req->status && time_after_eq(jiffies, req->timeout))
				req->status = -ETIMEDOUT;
			else if (req->status == -ENODATA)
				continue;
322
		}
323
		list_move_tail(&req->list, &done_list);
324 325 326 327 328 329 330 331 332 333
	}

	if (!list_empty(&req_list)) {
		req = list_entry(req_list.next, struct addr_req, list);
		set_timeout(req->timeout);
	}
	mutex_unlock(&lock);

	list_for_each_entry_safe(req, temp_req, &done_list, list) {
		list_del(&req->list);
334 335
		req->callback(req->status, (struct sockaddr *) &req->src_addr,
			req->addr, req->context);
336
		put_client(req->client);
337 338 339 340
		kfree(req);
	}
}

341 342
int rdma_resolve_ip(struct rdma_addr_client *client,
		    struct sockaddr *src_addr, struct sockaddr *dst_addr,
343 344 345 346 347
		    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)
{
348
	struct sockaddr *src_in, *dst_in;
349 350 351
	struct addr_req *req;
	int ret = 0;

352
	req = kzalloc(sizeof *req, GFP_KERNEL);
353 354 355
	if (!req)
		return -ENOMEM;

356 357 358 359 360 361 362 363 364 365 366 367 368 369 370
	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;
		}

		memcpy(src_in, src_addr, ip_addr_size(src_addr));
	} else {
		src_in->sa_family = dst_addr->sa_family;
	}

	memcpy(dst_in, dst_addr, ip_addr_size(dst_addr));
371 372 373
	req->addr = addr;
	req->callback = callback;
	req->context = context;
374 375
	req->client = client;
	atomic_inc(&client->refcount);
376

377
	req->status = addr_resolve(src_in, dst_in, addr);
378 379 380 381 382 383 384 385 386 387 388
	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;
389
		atomic_dec(&client->refcount);
390
		goto err;
391 392
	}
	return ret;
393 394 395
err:
	kfree(req);
	return ret;
396 397 398 399 400 401 402 403 404 405 406 407
}
EXPORT_SYMBOL(rdma_resolve_ip);

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;
408
			list_move(&req->list, &req_list);
409 410 411 412 413 414 415 416
			set_timeout(req->timeout);
			break;
		}
	}
	mutex_unlock(&lock);
}
EXPORT_SYMBOL(rdma_addr_cancel);

R
Roland Dreier 已提交
417
static int netevent_callback(struct notifier_block *self, unsigned long event,
418
	void *ctx)
419
{
R
Roland Dreier 已提交
420
	if (event == NETEVENT_NEIGH_UPDATE) {
421
		struct neighbour *neigh = ctx;
422

423
		if (neigh->nud_state & NUD_VALID) {
424 425 426
			set_timeout(jiffies);
		}
	}
427 428 429
	return 0;
}

430 431
static struct notifier_block nb = {
	.notifier_call = netevent_callback
432 433
};

434
static int __init addr_init(void)
435
{
436
	addr_wq = create_singlethread_workqueue("ib_addr");
437 438 439
	if (!addr_wq)
		return -ENOMEM;

440
	register_netevent_notifier(&nb);
441 442 443
	return 0;
}

444
static void __exit addr_cleanup(void)
445
{
446
	unregister_netevent_notifier(&nb);
447 448 449 450 451
	destroy_workqueue(addr_wq);
}

module_init(addr_init);
module_exit(addr_cleanup);
新手
引导
客服 返回
顶部