[net-next-2.6.git] / drivers / infiniband / core / addr.c

/*
 * 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.
 *
 * 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:
 *
 *     Redistribution and use in source and binary forms, with or
 *     without modification, are permitted provided that the following
 *     conditions are met:
 *
 *      - Redistributions of source code must retain the above
 *        copyright notice, this list of conditions and the following
 *        disclaimer.
 *
 *      - 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.
 *
 * 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.
 */

#include <linux/mutex.h>
#include <linux/inetdevice.h>
#include <linux/slab.h>
#include <linux/workqueue.h>
#include <net/arp.h>
#include <net/neighbour.h>
#include <net/route.h>
#include <net/netevent.h>
#include <net/addrconf.h>
#include <net/ip6_route.h>
#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;
	struct sockaddr_storage src_addr;
	struct sockaddr_storage dst_addr;
	struct rdma_dev_addr *addr;
	struct rdma_addr_client *client;
	void *context;
	void (*callback)(int status, struct sockaddr *src_addr,
			 struct rdma_dev_addr *addr, void *context);
	unsigned long timeout;
	int status;
};

static void process_req(struct work_struct *work);

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

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);

int rdma_copy_addr(struct rdma_dev_addr *dev_addr, struct net_device *dev,
		     const unsigned char *dst_dev_addr)
{
	dev_addr->dev_type = dev->type;
	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);
	dev_addr->bound_dev_if = dev->ifindex;
	return 0;
}
EXPORT_SYMBOL(rdma_copy_addr);

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

	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;
	}

	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;

		ret = rdma_copy_addr(dev_addr, dev, NULL);
		dev_put(dev);
		break;

#if defined(CONFIG_IPV6) || defined(CONFIG_IPV6_MODULE)
	case AF_INET6:
		rcu_read_lock();
		for_each_netdev_rcu(&init_net, dev) {
			if (ipv6_chk_addr(&init_net,
					  &((struct sockaddr_in6 *) addr)->sin6_addr,
					  dev, 1)) {
				ret = rdma_copy_addr(dev_addr, dev, NULL);
				break;
			}
		}
		rcu_read_unlock();
		break;
#endif
	}
	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) {
		if (time_after_eq(req->timeout, temp_req->timeout))
			break;
	}

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

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

static int addr4_resolve(struct sockaddr_in *src_in,
			 struct sockaddr_in *dst_in,
			 struct rdma_dev_addr *addr)
{
	__be32 src_ip = src_in->sin_addr.s_addr;
	__be32 dst_ip = dst_in->sin_addr.s_addr;
	struct flowi fl;
	struct rtable *rt;
	struct neighbour *neigh;
	int ret;

	memset(&fl, 0, sizeof fl);
	fl.nl_u.ip4_u.daddr = dst_ip;
	fl.nl_u.ip4_u.saddr = src_ip;
	fl.oif = addr->bound_dev_if;

	ret = ip_route_output_key(&init_net, &rt, &fl);
	if (ret)
		goto out;

	src_in->sin_family = AF_INET;
	src_in->sin_addr.s_addr = rt->rt_src;

	if (rt->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' */
	if (rt->dst.dev->flags & IFF_NOARP) {
		rdma_copy_addr(addr, rt->dst.dev, NULL);
		goto put;
	}

	neigh = neigh_lookup(&arp_tbl, &rt->rt_gateway, rt->dst.dev);
	if (!neigh || !(neigh->nud_state & NUD_VALID)) {
		neigh_event_send(rt->dst.neighbour, NULL);
		ret = -ENODATA;
		if (neigh)
			goto release;
		goto put;
	}

	ret = rdma_copy_addr(addr, neigh->dev, neigh->ha);
release:
	neigh_release(neigh);
put:
	ip_rt_put(rt);
out:
	return ret;
}

#if defined(CONFIG_IPV6) || defined(CONFIG_IPV6_MODULE)
static int addr6_resolve(struct sockaddr_in6 *src_in,
			 struct sockaddr_in6 *dst_in,
			 struct rdma_dev_addr *addr)
{
	struct flowi fl;
	struct neighbour *neigh;
	struct dst_entry *dst;
	int ret;

	memset(&fl, 0, sizeof fl);
	ipv6_addr_copy(&fl.fl6_dst, &dst_in->sin6_addr);
	ipv6_addr_copy(&fl.fl6_src, &src_in->sin6_addr);
	fl.oif = addr->bound_dev_if;

	dst = ip6_route_output(&init_net, NULL, &fl);
	if ((ret = dst->error))
		goto put;

	if (ipv6_addr_any(&fl.fl6_src)) {
		ret = ipv6_dev_get_saddr(&init_net, ip6_dst_idev(dst)->dev,
					 &fl.fl6_dst, 0, &fl.fl6_src);
		if (ret)
			goto put;

		src_in->sin6_family = AF_INET6;
		ipv6_addr_copy(&src_in->sin6_addr, &fl.fl6_src);
	}

	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' */
	if (dst->dev->flags & IFF_NOARP) {
		ret = rdma_copy_addr(addr, dst->dev, NULL);
		goto put;
	}

	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:
	dst_release(dst);
	return ret;
}
#else
static int addr6_resolve(struct sockaddr_in6 *src_in,
			 struct sockaddr_in6 *dst_in,
			 struct rdma_dev_addr *addr)
{
	return -EADDRNOTAVAIL;
}
#endif

static int addr_resolve(struct sockaddr *src_in,
			struct sockaddr *dst_in,
			struct rdma_dev_addr *addr)
{
	if (src_in->sa_family == AF_INET) {
		return addr4_resolve((struct sockaddr_in *) src_in,
			(struct sockaddr_in *) dst_in, addr);
	} else
		return addr6_resolve((struct sockaddr_in6 *) src_in,
			(struct sockaddr_in6 *) dst_in, addr);
}

static void process_req(struct work_struct *work)
{
	struct addr_req *req, *temp_req;
	struct sockaddr *src_in, *dst_in;
	struct list_head done_list;

	INIT_LIST_HEAD(&done_list);

	mutex_lock(&lock);
	list_for_each_entry_safe(req, temp_req, &req_list, list) {
		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);
			if (req->status && time_after_eq(jiffies, req->timeout))
				req->status = -ETIMEDOUT;
			else if (req->status == -ENODATA)
				continue;
		}
		list_move_tail(&req->list, &done_list);
	}

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

int rdma_resolve_ip(struct rdma_addr_client *client,
		    struct sockaddr *src_addr, struct sockaddr *dst_addr,
		    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)
{
	struct sockaddr *src_in, *dst_in;
	struct addr_req *req;
	int ret = 0;

	req = kzalloc(sizeof *req, GFP_KERNEL);
	if (!req)
		return -ENOMEM;

	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));
	req->addr = addr;
	req->callback = callback;
	req->context = context;
	req->client = client;
	atomic_inc(&client->refcount);

	req->status = addr_resolve(src_in, dst_in, addr);
	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;
		atomic_dec(&client->refcount);
		goto err;
	}
	return ret;
err:
	kfree(req);
	return ret;
}
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;
			list_move(&req->list, &req_list);
			set_timeout(req->timeout);
			break;
		}
	}
	mutex_unlock(&lock);
}
EXPORT_SYMBOL(rdma_addr_cancel);

static int netevent_callback(struct notifier_block *self, unsigned long event,
	void *ctx)
{
	if (event == NETEVENT_NEIGH_UPDATE) {
		struct neighbour *neigh = ctx;

		if (neigh->nud_state & NUD_VALID) {
			set_timeout(jiffies);
		}
	}
	return 0;
}

static struct notifier_block nb = {
	.notifier_call = netevent_callback
};

static int __init addr_init(void)
{
	addr_wq = create_singlethread_workqueue("ib_addr");
	if (!addr_wq)
		return -ENOMEM;

	register_netevent_notifier(&nb);
	return 0;
}

static void __exit addr_cleanup(void)
{
	unregister_netevent_notifier(&nb);
	destroy_workqueue(addr_wq);
}

module_init(addr_init);
module_exit(addr_cleanup);
Commit	Line	Data
	1	/*
	2	* Copyright (c) 2005 Voltaire Inc. All rights reserved.
	3	* Copyright (c) 2002-2005, Network Appliance, Inc. All rights reserved.
	4	* Copyright (c) 1999-2005, Mellanox Technologies, Inc. All rights reserved.
	5	* Copyright (c) 2005 Intel Corporation. All rights reserved.
	6	*
	7	* This software is available to you under a choice of one of two
	8	* licenses. You may choose to be licensed under the terms of the GNU
	9	* General Public License (GPL) Version 2, available from the file
	10	* COPYING in the main directory of this source tree, or the
	11	* OpenIB.org BSD license below:
	12	*
	13	* Redistribution and use in source and binary forms, with or
	14	* without modification, are permitted provided that the following
	15	* conditions are met:
	16	*
	17	* - Redistributions of source code must retain the above
	18	* copyright notice, this list of conditions and the following
	19	* disclaimer.
	20	*
	21	* - Redistributions in binary form must reproduce the above
	22	* copyright notice, this list of conditions and the following
	23	* disclaimer in the documentation and/or other materials
	24	* provided with the distribution.
	25	*
	26	* THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
	27	* EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
	28	* MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND
	29	* NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS
	30	* BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN
	31	* ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
	32	* CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
	33	* SOFTWARE.
	34	*/
	35
	36	#include <linux/mutex.h>
	37	#include <linux/inetdevice.h>
	38	#include <linux/slab.h>
	39	#include <linux/workqueue.h>
	40	#include <net/arp.h>
	41	#include <net/neighbour.h>
	42	#include <net/route.h>
	43	#include <net/netevent.h>
	44	#include <net/addrconf.h>
	45	#include <net/ip6_route.h>
	46	#include <rdma/ib_addr.h>
	47
	48	MODULE_AUTHOR("Sean Hefty");
	49	MODULE_DESCRIPTION("IB Address Translation");
	50	MODULE_LICENSE("Dual BSD/GPL");
	51
	52	struct addr_req {
	53	struct list_head list;
	54	struct sockaddr_storage src_addr;
	55	struct sockaddr_storage dst_addr;
	56	struct rdma_dev_addr *addr;
	57	struct rdma_addr_client *client;
	58	void *context;
	59	void (callback)(int status, struct sockaddr src_addr,
	60	struct rdma_dev_addr addr, void context);
	61	unsigned long timeout;
	62	int status;
	63	};
	64
	65	static void process_req(struct work_struct *work);
	66
	67	static DEFINE_MUTEX(lock);
	68	static LIST_HEAD(req_list);
	69	static DECLARE_DELAYED_WORK(work, process_req);
	70	static struct workqueue_struct *addr_wq;
	71
	72	void rdma_addr_register_client(struct rdma_addr_client *client)
	73	{
	74	atomic_set(&client->refcount, 1);
	75	init_completion(&client->comp);
	76	}
	77	EXPORT_SYMBOL(rdma_addr_register_client);
	78
	79	static inline void put_client(struct rdma_addr_client *client)
	80	{
	81	if (atomic_dec_and_test(&client->refcount))
	82	complete(&client->comp);
	83	}
	84
	85	void rdma_addr_unregister_client(struct rdma_addr_client *client)
	86	{
	87	put_client(client);
	88	wait_for_completion(&client->comp);
	89	}
	90	EXPORT_SYMBOL(rdma_addr_unregister_client);
	91
	92	int rdma_copy_addr(struct rdma_dev_addr dev_addr, struct net_device dev,
	93	const unsigned char *dst_dev_addr)
	94	{
	95	dev_addr->dev_type = dev->type;
	96	memcpy(dev_addr->src_dev_addr, dev->dev_addr, MAX_ADDR_LEN);
	97	memcpy(dev_addr->broadcast, dev->broadcast, MAX_ADDR_LEN);
	98	if (dst_dev_addr)
	99	memcpy(dev_addr->dst_dev_addr, dst_dev_addr, MAX_ADDR_LEN);
	100	dev_addr->bound_dev_if = dev->ifindex;
	101	return 0;
	102	}
	103	EXPORT_SYMBOL(rdma_copy_addr);
	104
	105	int rdma_translate_ip(struct sockaddr addr, struct rdma_dev_addr dev_addr)
	106	{
	107	struct net_device *dev;
	108	int ret = -EADDRNOTAVAIL;
	109
	110	if (dev_addr->bound_dev_if) {
	111	dev = dev_get_by_index(&init_net, dev_addr->bound_dev_if);
	112	if (!dev)
	113	return -ENODEV;
	114	ret = rdma_copy_addr(dev_addr, dev, NULL);
	115	dev_put(dev);
	116	return ret;
	117	}
	118
	119	switch (addr->sa_family) {
	120	case AF_INET:
	121	dev = ip_dev_find(&init_net,
	122	((struct sockaddr_in *) addr)->sin_addr.s_addr);
	123
	124	if (!dev)
	125	return ret;
	126
	127	ret = rdma_copy_addr(dev_addr, dev, NULL);
	128	dev_put(dev);
	129	break;
	130
	131	#if defined(CONFIG_IPV6) \|\| defined(CONFIG_IPV6_MODULE)
	132	case AF_INET6:
	133	rcu_read_lock();
	134	for_each_netdev_rcu(&init_net, dev) {
	135	if (ipv6_chk_addr(&init_net,
	136	&((struct sockaddr_in6 *) addr)->sin6_addr,
	137	dev, 1)) {
	138	ret = rdma_copy_addr(dev_addr, dev, NULL);
	139	break;
	140	}
	141	}
	142	rcu_read_unlock();
	143	break;
	144	#endif
	145	}
	146	return ret;
	147	}
	148	EXPORT_SYMBOL(rdma_translate_ip);
	149
	150	static void set_timeout(unsigned long time)
	151	{
	152	unsigned long delay;
	153
	154	cancel_delayed_work(&work);
	155
	156	delay = time - jiffies;
	157	if ((long)delay <= 0)
	158	delay = 1;
	159
	160	queue_delayed_work(addr_wq, &work, delay);
	161	}
	162
	163	static void queue_req(struct addr_req *req)
	164	{
	165	struct addr_req *temp_req;
	166
	167	mutex_lock(&lock);
	168	list_for_each_entry_reverse(temp_req, &req_list, list) {
	169	if (time_after_eq(req->timeout, temp_req->timeout))
	170	break;
	171	}
	172
	173	list_add(&req->list, &temp_req->list);
	174
	175	if (req_list.next == &req->list)
	176	set_timeout(req->timeout);
	177	mutex_unlock(&lock);
	178	}
	179
	180	static int addr4_resolve(struct sockaddr_in *src_in,
	181	struct sockaddr_in *dst_in,
	182	struct rdma_dev_addr *addr)
	183	{
	184	__be32 src_ip = src_in->sin_addr.s_addr;
	185	__be32 dst_ip = dst_in->sin_addr.s_addr;
	186	struct flowi fl;
	187	struct rtable *rt;
	188	struct neighbour *neigh;
	189	int ret;
	190
	191	memset(&fl, 0, sizeof fl);
	192	fl.nl_u.ip4_u.daddr = dst_ip;
	193	fl.nl_u.ip4_u.saddr = src_ip;
	194	fl.oif = addr->bound_dev_if;
	195
	196	ret = ip_route_output_key(&init_net, &rt, &fl);
	197	if (ret)
	198	goto out;
	199
	200	src_in->sin_family = AF_INET;
	201	src_in->sin_addr.s_addr = rt->rt_src;
	202
	203	if (rt->dst.dev->flags & IFF_LOOPBACK) {
	204	ret = rdma_translate_ip((struct sockaddr *) dst_in, addr);
	205	if (!ret)
	206	memcpy(addr->dst_dev_addr, addr->src_dev_addr, MAX_ADDR_LEN);
	207	goto put;
	208	}
	209
	210	/* If the device does ARP internally, return 'done' */
	211	if (rt->dst.dev->flags & IFF_NOARP) {
	212	rdma_copy_addr(addr, rt->dst.dev, NULL);
	213	goto put;
	214	}
	215
	216	neigh = neigh_lookup(&arp_tbl, &rt->rt_gateway, rt->dst.dev);
	217	if (!neigh \|\| !(neigh->nud_state & NUD_VALID)) {
	218	neigh_event_send(rt->dst.neighbour, NULL);
	219	ret = -ENODATA;
	220	if (neigh)
	221	goto release;
	222	goto put;
	223	}
	224
	225	ret = rdma_copy_addr(addr, neigh->dev, neigh->ha);
	226	release:
	227	neigh_release(neigh);
	228	put:
	229	ip_rt_put(rt);
	230	out:
	231	return ret;
	232	}
	233
	234	#if defined(CONFIG_IPV6) \|\| defined(CONFIG_IPV6_MODULE)
	235	static int addr6_resolve(struct sockaddr_in6 *src_in,
	236	struct sockaddr_in6 *dst_in,
	237	struct rdma_dev_addr *addr)
	238	{
	239	struct flowi fl;
	240	struct neighbour *neigh;
	241	struct dst_entry *dst;
	242	int ret;
	243
	244	memset(&fl, 0, sizeof fl);
	245	ipv6_addr_copy(&fl.fl6_dst, &dst_in->sin6_addr);
	246	ipv6_addr_copy(&fl.fl6_src, &src_in->sin6_addr);
	247	fl.oif = addr->bound_dev_if;
	248
	249	dst = ip6_route_output(&init_net, NULL, &fl);
	250	if ((ret = dst->error))
	251	goto put;
	252
	253	if (ipv6_addr_any(&fl.fl6_src)) {
	254	ret = ipv6_dev_get_saddr(&init_net, ip6_dst_idev(dst)->dev,
	255	&fl.fl6_dst, 0, &fl.fl6_src);
	256	if (ret)
	257	goto put;
	258
	259	src_in->sin6_family = AF_INET6;
	260	ipv6_addr_copy(&src_in->sin6_addr, &fl.fl6_src);
	261	}
	262
	263	if (dst->dev->flags & IFF_LOOPBACK) {
	264	ret = rdma_translate_ip((struct sockaddr *) dst_in, addr);
	265	if (!ret)
	266	memcpy(addr->dst_dev_addr, addr->src_dev_addr, MAX_ADDR_LEN);
	267	goto put;
	268	}
	269
	270	/* If the device does ARP internally, return 'done' */
	271	if (dst->dev->flags & IFF_NOARP) {
	272	ret = rdma_copy_addr(addr, dst->dev, NULL);
	273	goto put;
	274	}
	275
	276	neigh = dst->neighbour;
	277	if (!neigh \|\| !(neigh->nud_state & NUD_VALID)) {
	278	neigh_event_send(dst->neighbour, NULL);
	279	ret = -ENODATA;
	280	goto put;
	281	}
	282
	283	ret = rdma_copy_addr(addr, dst->dev, neigh->ha);
	284	put:
	285	dst_release(dst);
	286	return ret;
	287	}
	288	#else
	289	static int addr6_resolve(struct sockaddr_in6 *src_in,
	290	struct sockaddr_in6 *dst_in,
	291	struct rdma_dev_addr *addr)
	292	{
	293	return -EADDRNOTAVAIL;
	294	}
	295	#endif
	296
	297	static int addr_resolve(struct sockaddr *src_in,
	298	struct sockaddr *dst_in,
	299	struct rdma_dev_addr *addr)
	300	{
	301	if (src_in->sa_family == AF_INET) {
	302	return addr4_resolve((struct sockaddr_in *) src_in,
	303	(struct sockaddr_in *) dst_in, addr);
	304	} else
	305	return addr6_resolve((struct sockaddr_in6 *) src_in,
	306	(struct sockaddr_in6 *) dst_in, addr);
	307	}
	308
	309	static void process_req(struct work_struct *work)
	310	{
	311	struct addr_req req, temp_req;
	312	struct sockaddr src_in, dst_in;
	313	struct list_head done_list;
	314
	315	INIT_LIST_HEAD(&done_list);
	316
	317	mutex_lock(&lock);
	318	list_for_each_entry_safe(req, temp_req, &req_list, list) {
	319	if (req->status == -ENODATA) {
	320	src_in = (struct sockaddr *) &req->src_addr;
	321	dst_in = (struct sockaddr *) &req->dst_addr;
	322	req->status = addr_resolve(src_in, dst_in, req->addr);
	323	if (req->status && time_after_eq(jiffies, req->timeout))
	324	req->status = -ETIMEDOUT;
	325	else if (req->status == -ENODATA)
	326	continue;
	327	}
	328	list_move_tail(&req->list, &done_list);
	329	}
	330
	331	if (!list_empty(&req_list)) {
	332	req = list_entry(req_list.next, struct addr_req, list);
	333	set_timeout(req->timeout);
	334	}
	335	mutex_unlock(&lock);
	336
	337	list_for_each_entry_safe(req, temp_req, &done_list, list) {
	338	list_del(&req->list);
	339	req->callback(req->status, (struct sockaddr *) &req->src_addr,
	340	req->addr, req->context);
	341	put_client(req->client);
	342	kfree(req);
	343	}
	344	}
	345
	346	int rdma_resolve_ip(struct rdma_addr_client *client,
	347	struct sockaddr src_addr, struct sockaddr dst_addr,
	348	struct rdma_dev_addr *addr, int timeout_ms,
	349	void (callback)(int status, struct sockaddr src_addr,
	350	struct rdma_dev_addr addr, void context),
	351	void *context)
	352	{
	353	struct sockaddr src_in, dst_in;
	354	struct addr_req *req;
	355	int ret = 0;
	356
	357	req = kzalloc(sizeof *req, GFP_KERNEL);
	358	if (!req)
	359	return -ENOMEM;
	360
	361	src_in = (struct sockaddr *) &req->src_addr;
	362	dst_in = (struct sockaddr *) &req->dst_addr;
	363
	364	if (src_addr) {
	365	if (src_addr->sa_family != dst_addr->sa_family) {
	366	ret = -EINVAL;
	367	goto err;
	368	}
	369
	370	memcpy(src_in, src_addr, ip_addr_size(src_addr));
	371	} else {
	372	src_in->sa_family = dst_addr->sa_family;
	373	}
	374
	375	memcpy(dst_in, dst_addr, ip_addr_size(dst_addr));
	376	req->addr = addr;
	377	req->callback = callback;
	378	req->context = context;
	379	req->client = client;
	380	atomic_inc(&client->refcount);
	381
	382	req->status = addr_resolve(src_in, dst_in, addr);
	383	switch (req->status) {
	384	case 0:
	385	req->timeout = jiffies;
	386	queue_req(req);
	387	break;
	388	case -ENODATA:
	389	req->timeout = msecs_to_jiffies(timeout_ms) + jiffies;
	390	queue_req(req);
	391	break;
	392	default:
	393	ret = req->status;
	394	atomic_dec(&client->refcount);
	395	goto err;
	396	}
	397	return ret;
	398	err:
	399	kfree(req);
	400	return ret;
	401	}
	402	EXPORT_SYMBOL(rdma_resolve_ip);
	403
	404	void rdma_addr_cancel(struct rdma_dev_addr *addr)
	405	{
	406	struct addr_req req, temp_req;
	407
	408	mutex_lock(&lock);
	409	list_for_each_entry_safe(req, temp_req, &req_list, list) {
	410	if (req->addr == addr) {
	411	req->status = -ECANCELED;
	412	req->timeout = jiffies;
	413	list_move(&req->list, &req_list);
	414	set_timeout(req->timeout);
	415	break;
	416	}
	417	}
	418	mutex_unlock(&lock);
	419	}
	420	EXPORT_SYMBOL(rdma_addr_cancel);
	421
	422	static int netevent_callback(struct notifier_block *self, unsigned long event,
	423	void *ctx)
	424	{
	425	if (event == NETEVENT_NEIGH_UPDATE) {
	426	struct neighbour *neigh = ctx;
	427
	428	if (neigh->nud_state & NUD_VALID) {
	429	set_timeout(jiffies);
	430	}
	431	}
	432	return 0;
	433	}
	434
	435	static struct notifier_block nb = {
	436	.notifier_call = netevent_callback
	437	};
	438
	439	static int __init addr_init(void)
	440	{
	441	addr_wq = create_singlethread_workqueue("ib_addr");
	442	if (!addr_wq)
	443	return -ENOMEM;
	444
	445	register_netevent_notifier(&nb);
	446	return 0;
	447	}
	448
	449	static void __exit addr_cleanup(void)
	450	{
	451	unregister_netevent_notifier(&nb);
	452	destroy_workqueue(addr_wq);
	453	}
	454
	455	module_init(addr_init);
	456	module_exit(addr_cleanup);