[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/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:
		read_lock(&dev_base_lock);
		for_each_netdev(&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;
			}
		}
		read_unlock(&dev_base_lock);
		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 void addr_send_arp(struct sockaddr *dst_in)
{
	struct rtable *rt;
	struct flowi fl;

	memset(&fl, 0, sizeof fl);

	switch (dst_in->sa_family) {
	case AF_INET:
		fl.nl_u.ip4_u.daddr =
			((struct sockaddr_in *) dst_in)->sin_addr.s_addr;

		if (ip_route_output_key(&init_net, &rt, &fl))
			return;

		neigh_event_send(rt->u.dst.neighbour, NULL);
		ip_rt_put(rt);
		break;

#if defined(CONFIG_IPV6) || defined(CONFIG_IPV6_MODULE)
	case AF_INET6:
	{
		struct dst_entry *dst;

		fl.nl_u.ip6_u.daddr =
			((struct sockaddr_in6 *) dst_in)->sin6_addr;

		dst = ip6_route_output(&init_net, NULL, &fl);
		if (!dst)
			return;

		neigh_event_send(dst->neighbour, NULL);
		dst_release(dst);
		break;
	}
#endif
	}
}

static int addr4_resolve_remote(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;

	/* If the device does ARP internally, return 'done' */
	if (rt->idev->dev->flags & IFF_NOARP) {
		rdma_copy_addr(addr, rt->idev->dev, NULL);
		goto put;
	}

	neigh = neigh_lookup(&arp_tbl, &rt->rt_gateway, rt->idev->dev);
	if (!neigh) {
		ret = -ENODATA;
		goto put;
	}

	if (!(neigh->nud_state & NUD_VALID)) {
		ret = -ENODATA;
		goto release;
	}

	if (!src_ip) {
		src_in->sin_family = dst_in->sin_family;
		src_in->sin_addr.s_addr = rt->rt_src;
	}

	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_remote(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 = -ENODATA;

	memset(&fl, 0, sizeof fl);
	fl.nl_u.ip6_u.daddr = dst_in->sin6_addr;
	fl.nl_u.ip6_u.saddr = src_in->sin6_addr;
	fl.oif = addr->bound_dev_if;

	dst = ip6_route_output(&init_net, NULL, &fl);
	if (!dst)
		return ret;

	if (dst->dev->flags & IFF_NOARP) {
		ret = rdma_copy_addr(addr, dst->dev, NULL);
	} else {
		neigh = dst->neighbour;
		if (neigh && (neigh->nud_state & NUD_VALID))
			ret = rdma_copy_addr(addr, neigh->dev, neigh->ha);
	}

	dst_release(dst);
	return ret;
}
#else
static int addr6_resolve_remote(struct sockaddr_in6 *src_in,
			       struct sockaddr_in6 *dst_in,
			       struct rdma_dev_addr *addr)
{
	return -EADDRNOTAVAIL;
}
#endif

static int addr_resolve_remote(struct sockaddr *src_in,
				struct sockaddr *dst_in,
				struct rdma_dev_addr *addr)
{
	if (src_in->sa_family == AF_INET) {
		return addr4_resolve_remote((struct sockaddr_in *) src_in,
			(struct sockaddr_in *) dst_in, addr);
	} else
		return addr6_resolve_remote((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_remote(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);
	}
}

static int addr_resolve_local(struct sockaddr *src_in,
			      struct sockaddr *dst_in,
			      struct rdma_dev_addr *addr)
{
	struct net_device *dev;
	int ret;

	switch (dst_in->sa_family) {
	case AF_INET:
	{
		__be32 src_ip = ((struct sockaddr_in *) src_in)->sin_addr.s_addr;
		__be32 dst_ip = ((struct sockaddr_in *) dst_in)->sin_addr.s_addr;

		dev = ip_dev_find(&init_net, dst_ip);
		if (!dev)
			return -EADDRNOTAVAIL;

		if (ipv4_is_zeronet(src_ip)) {
			src_in->sa_family = dst_in->sa_family;
			((struct sockaddr_in *) src_in)->sin_addr.s_addr = dst_ip;
			ret = rdma_copy_addr(addr, dev, dev->dev_addr);
		} else if (ipv4_is_loopback(src_ip)) {
			ret = rdma_translate_ip(dst_in, addr);
			if (!ret)
				memcpy(addr->dst_dev_addr, dev->dev_addr, MAX_ADDR_LEN);
		} else {
			ret = rdma_translate_ip(src_in, addr);
			if (!ret)
				memcpy(addr->dst_dev_addr, dev->dev_addr, MAX_ADDR_LEN);
		}
		dev_put(dev);
		break;
	}

#if defined(CONFIG_IPV6) || defined(CONFIG_IPV6_MODULE)
	case AF_INET6:
	{
		struct in6_addr *a;

		read_lock(&dev_base_lock);
		for_each_netdev(&init_net, dev)
			if (ipv6_chk_addr(&init_net,
					  &((struct sockaddr_in6 *) dst_in)->sin6_addr,
					  dev, 1))
				break;

		if (!dev) {
			read_unlock(&dev_base_lock);
			return -EADDRNOTAVAIL;
		}

		a = &((struct sockaddr_in6 *) src_in)->sin6_addr;

		if (ipv6_addr_any(a)) {
			src_in->sa_family = dst_in->sa_family;
			((struct sockaddr_in6 *) src_in)->sin6_addr =
				((struct sockaddr_in6 *) dst_in)->sin6_addr;
			ret = rdma_copy_addr(addr, dev, dev->dev_addr);
		} else if (ipv6_addr_loopback(a)) {
			ret = rdma_translate_ip(dst_in, addr);
			if (!ret)
				memcpy(addr->dst_dev_addr, dev->dev_addr, MAX_ADDR_LEN);
		} else  {
			ret = rdma_translate_ip(src_in, addr);
			if (!ret)
				memcpy(addr->dst_dev_addr, dev->dev_addr, MAX_ADDR_LEN);
		}
		read_unlock(&dev_base_lock);
		break;
	}
#endif

	default:
		ret = -EADDRNOTAVAIL;
		break;
	}

	return ret;
}

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_local(src_in, dst_in, addr);
	if (req->status == -EADDRNOTAVAIL)
		req->status = addr_resolve_remote(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);
		addr_send_arp(dst_in);
		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
7025fcd3 SH	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	*
a9474917 SH	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:
7025fcd3	12	*
a9474917 SH	13	* Redistribution and use in source and binary forms, with or
	14	* without modification, are permitted provided that the following
	15	* conditions are met:
7025fcd3	16	*
a9474917 SH	17	* - Redistributions of source code must retain the above
	18	* copyright notice, this list of conditions and the following
	19	* disclaimer.
7025fcd3	20	*
a9474917 SH	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.
7025fcd3	25	*
a9474917 SH	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.
7025fcd3 SH	34	*/
	35
	36	#include <linux/mutex.h>
	37	#include <linux/inetdevice.h>
	38	#include <linux/workqueue.h>
7025fcd3 SH	39	#include <net/arp.h>
	40	#include <net/neighbour.h>
	41	#include <net/route.h>
e795d092	42	#include <net/netevent.h>
38617c64 AS	43	#include <net/addrconf.h>
38617c64 AS	44	#include <net/ip6_route.h>
7025fcd3 SH	45	#include <rdma/ib_addr.h>
	46
	47	MODULE_AUTHOR("Sean Hefty");
	48	MODULE_DESCRIPTION("IB Address Translation");
	49	MODULE_LICENSE("Dual BSD/GPL");
	50
	51	struct addr_req {
	52	struct list_head list;
38617c64 AS	53	struct sockaddr_storage src_addr;
38617c64 AS	54	struct sockaddr_storage dst_addr;
7025fcd3	55	struct rdma_dev_addr *addr;
7a118df3	56	struct rdma_addr_client *client;
7025fcd3 SH	57	void *context;
	58	void (callback)(int status, struct sockaddr src_addr,
	59	struct rdma_dev_addr addr, void context);
	60	unsigned long timeout;
	61	int status;
	62	};
	63
c4028958	64	static void process_req(struct work_struct *work);
7025fcd3 SH	65
	66	static DEFINE_MUTEX(lock);
	67	static LIST_HEAD(req_list);
c4028958	68	static DECLARE_DELAYED_WORK(work, process_req);
7025fcd3 SH	69	static struct workqueue_struct *addr_wq;
7025fcd3 SH	70
7a118df3 SH	71	void rdma_addr_register_client(struct rdma_addr_client *client)
	72	{
	73	atomic_set(&client->refcount, 1);
	74	init_completion(&client->comp);
	75	}
	76	EXPORT_SYMBOL(rdma_addr_register_client);
	77
	78	static inline void put_client(struct rdma_addr_client *client)
	79	{
	80	if (atomic_dec_and_test(&client->refcount))
	81	complete(&client->comp);
	82	}
	83
	84	void rdma_addr_unregister_client(struct rdma_addr_client *client)
	85	{
	86	put_client(client);
	87	wait_for_completion(&client->comp);
	88	}
	89	EXPORT_SYMBOL(rdma_addr_unregister_client);
	90
07ebafba TT	91	int rdma_copy_addr(struct rdma_dev_addr dev_addr, struct net_device dev,
07ebafba TT	92	const unsigned char *dst_dev_addr)
7025fcd3	93	{
c4315d85	94	dev_addr->dev_type = dev->type;
7025fcd3 SH	95	memcpy(dev_addr->src_dev_addr, dev->dev_addr, MAX_ADDR_LEN);
	96	memcpy(dev_addr->broadcast, dev->broadcast, MAX_ADDR_LEN);
	97	if (dst_dev_addr)
	98	memcpy(dev_addr->dst_dev_addr, dst_dev_addr, MAX_ADDR_LEN);
6266ed6e	99	dev_addr->bound_dev_if = dev->ifindex;
7025fcd3 SH	100	return 0;
7025fcd3 SH	101	}
07ebafba	102	EXPORT_SYMBOL(rdma_copy_addr);
7025fcd3 SH	103
	104	int rdma_translate_ip(struct sockaddr addr, struct rdma_dev_addr dev_addr)
	105	{
	106	struct net_device *dev;
38617c64	107	int ret = -EADDRNOTAVAIL;
7025fcd3	108
6266ed6e SH	109	if (dev_addr->bound_dev_if) {
	110	dev = dev_get_by_index(&init_net, dev_addr->bound_dev_if);
	111	if (!dev)
	112	return -ENODEV;
	113	ret = rdma_copy_addr(dev_addr, dev, NULL);
	114	dev_put(dev);
	115	return ret;
	116	}
	117
38617c64 AS	118	switch (addr->sa_family) {
	119	case AF_INET:
	120	dev = ip_dev_find(&init_net,
	121	((struct sockaddr_in *) addr)->sin_addr.s_addr);
	122
	123	if (!dev)
	124	return ret;
7025fcd3	125
38617c64 AS	126	ret = rdma_copy_addr(dev_addr, dev, NULL);
	127	dev_put(dev);
	128	break;
2c4ab624 RD	129
2c4ab624 RD	130	#if defined(CONFIG_IPV6) \|\| defined(CONFIG_IPV6_MODULE)
38617c64	131	case AF_INET6:
0f9ea5d2	132	read_lock(&dev_base_lock);
38617c64 AS	133	for_each_netdev(&init_net, dev) {
	134	if (ipv6_chk_addr(&init_net,
	135	&((struct sockaddr_in6 *) addr)->sin6_addr,
	136	dev, 1)) {
	137	ret = rdma_copy_addr(dev_addr, dev, NULL);
	138	break;
	139	}
	140	}
0f9ea5d2	141	read_unlock(&dev_base_lock);
38617c64	142	break;
2c4ab624	143	#endif
38617c64	144	}
7025fcd3 SH	145	return ret;
	146	}
	147	EXPORT_SYMBOL(rdma_translate_ip);
	148
	149	static void set_timeout(unsigned long time)
	150	{
	151	unsigned long delay;
	152
	153	cancel_delayed_work(&work);
	154
	155	delay = time - jiffies;
	156	if ((long)delay <= 0)
	157	delay = 1;
	158
	159	queue_delayed_work(addr_wq, &work, delay);
	160	}
	161
	162	static void queue_req(struct addr_req *req)
	163	{
	164	struct addr_req *temp_req;
	165
	166	mutex_lock(&lock);
	167	list_for_each_entry_reverse(temp_req, &req_list, list) {
f115db48	168	if (time_after_eq(req->timeout, temp_req->timeout))
7025fcd3 SH	169	break;
	170	}
	171
	172	list_add(&req->list, &temp_req->list);
	173
	174	if (req_list.next == &req->list)
	175	set_timeout(req->timeout);
	176	mutex_unlock(&lock);
	177	}
	178
38617c64	179	static void addr_send_arp(struct sockaddr *dst_in)
7025fcd3 SH	180	{
	181	struct rtable *rt;
	182	struct flowi fl;
7025fcd3 SH	183
7025fcd3 SH	184	memset(&fl, 0, sizeof fl);
2c4ab624 RD	185
	186	switch (dst_in->sa_family) {
	187	case AF_INET:
38617c64 AS	188	fl.nl_u.ip4_u.daddr =
38617c64 AS	189	((struct sockaddr_in *) dst_in)->sin_addr.s_addr;
7025fcd3	190
38617c64 AS	191	if (ip_route_output_key(&init_net, &rt, &fl))
	192	return;
	193
	194	neigh_event_send(rt->u.dst.neighbour, NULL);
	195	ip_rt_put(rt);
2c4ab624 RD	196	break;
	197
	198	#if defined(CONFIG_IPV6) \|\| defined(CONFIG_IPV6_MODULE)
	199	case AF_INET6:
	200	{
	201	struct dst_entry *dst;
38617c64	202
38617c64 AS	203	fl.nl_u.ip6_u.daddr =
	204	((struct sockaddr_in6 *) dst_in)->sin6_addr;
	205
	206	dst = ip6_route_output(&init_net, NULL, &fl);
	207	if (!dst)
	208	return;
	209
	210	neigh_event_send(dst->neighbour, NULL);
	211	dst_release(dst);
2c4ab624 RD	212	break;
	213	}
	214	#endif
38617c64	215	}
7025fcd3 SH	216	}
7025fcd3 SH	217
38617c64	218	static int addr4_resolve_remote(struct sockaddr_in *src_in,
7025fcd3 SH	219	struct sockaddr_in *dst_in,
	220	struct rdma_dev_addr *addr)
	221	{
1b90c137 AV	222	__be32 src_ip = src_in->sin_addr.s_addr;
1b90c137 AV	223	__be32 dst_ip = dst_in->sin_addr.s_addr;
7025fcd3 SH	224	struct flowi fl;
	225	struct rtable *rt;
	226	struct neighbour *neigh;
	227	int ret;
	228
	229	memset(&fl, 0, sizeof fl);
	230	fl.nl_u.ip4_u.daddr = dst_ip;
	231	fl.nl_u.ip4_u.saddr = src_ip;
6266ed6e SH	232	fl.oif = addr->bound_dev_if;
6266ed6e SH	233
f206351a	234	ret = ip_route_output_key(&init_net, &rt, &fl);
7025fcd3 SH	235	if (ret)
	236	goto out;
	237
	238	/* If the device does ARP internally, return 'done' */
	239	if (rt->idev->dev->flags & IFF_NOARP) {
07ebafba	240	rdma_copy_addr(addr, rt->idev->dev, NULL);
7025fcd3 SH	241	goto put;
	242	}
	243
	244	neigh = neigh_lookup(&arp_tbl, &rt->rt_gateway, rt->idev->dev);
	245	if (!neigh) {
	246	ret = -ENODATA;
	247	goto put;
	248	}
	249
	250	if (!(neigh->nud_state & NUD_VALID)) {
	251	ret = -ENODATA;
	252	goto release;
	253	}
	254
	255	if (!src_ip) {
	256	src_in->sin_family = dst_in->sin_family;
	257	src_in->sin_addr.s_addr = rt->rt_src;
	258	}
	259
07ebafba	260	ret = rdma_copy_addr(addr, neigh->dev, neigh->ha);
7025fcd3 SH	261	release:
	262	neigh_release(neigh);
	263	put:
	264	ip_rt_put(rt);
	265	out:
	266	return ret;
	267	}
	268
2c4ab624	269	#if defined(CONFIG_IPV6) \|\| defined(CONFIG_IPV6_MODULE)
38617c64 AS	270	static int addr6_resolve_remote(struct sockaddr_in6 *src_in,
	271	struct sockaddr_in6 *dst_in,
	272	struct rdma_dev_addr *addr)
	273	{
	274	struct flowi fl;
	275	struct neighbour *neigh;
	276	struct dst_entry *dst;
	277	int ret = -ENODATA;
	278
	279	memset(&fl, 0, sizeof fl);
	280	fl.nl_u.ip6_u.daddr = dst_in->sin6_addr;
	281	fl.nl_u.ip6_u.saddr = src_in->sin6_addr;
6266ed6e	282	fl.oif = addr->bound_dev_if;
38617c64 AS	283
	284	dst = ip6_route_output(&init_net, NULL, &fl);
	285	if (!dst)
	286	return ret;
	287
	288	if (dst->dev->flags & IFF_NOARP) {
	289	ret = rdma_copy_addr(addr, dst->dev, NULL);
	290	} else {
	291	neigh = dst->neighbour;
	292	if (neigh && (neigh->nud_state & NUD_VALID))
	293	ret = rdma_copy_addr(addr, neigh->dev, neigh->ha);
	294	}
	295
	296	dst_release(dst);
	297	return ret;
	298	}
2c4ab624 RD	299	#else
	300	static int addr6_resolve_remote(struct sockaddr_in6 *src_in,
	301	struct sockaddr_in6 *dst_in,
	302	struct rdma_dev_addr *addr)
	303	{
	304	return -EADDRNOTAVAIL;
	305	}
	306	#endif
38617c64 AS	307
	308	static int addr_resolve_remote(struct sockaddr *src_in,
	309	struct sockaddr *dst_in,
	310	struct rdma_dev_addr *addr)
	311	{
	312	if (src_in->sa_family == AF_INET) {
	313	return addr4_resolve_remote((struct sockaddr_in *) src_in,
	314	(struct sockaddr_in *) dst_in, addr);
	315	} else
	316	return addr6_resolve_remote((struct sockaddr_in6 *) src_in,
	317	(struct sockaddr_in6 *) dst_in, addr);
	318	}
	319
c4028958	320	static void process_req(struct work_struct *work)
7025fcd3 SH	321	{
7025fcd3 SH	322	struct addr_req req, temp_req;
38617c64	323	struct sockaddr src_in, dst_in;
7025fcd3 SH	324	struct list_head done_list;
	325
	326	INIT_LIST_HEAD(&done_list);
	327
	328	mutex_lock(&lock);
	329	list_for_each_entry_safe(req, temp_req, &req_list, list) {
c78bb844	330	if (req->status == -ENODATA) {
38617c64 AS	331	src_in = (struct sockaddr *) &req->src_addr;
38617c64 AS	332	dst_in = (struct sockaddr *) &req->dst_addr;
7025fcd3 SH	333	req->status = addr_resolve_remote(src_in, dst_in,
7025fcd3 SH	334	req->addr);
c78bb844 KK	335	if (req->status && time_after_eq(jiffies, req->timeout))
	336	req->status = -ETIMEDOUT;
	337	else if (req->status == -ENODATA)
	338	continue;
7025fcd3	339	}
04699a1f	340	list_move_tail(&req->list, &done_list);
7025fcd3 SH	341	}
	342
	343	if (!list_empty(&req_list)) {
	344	req = list_entry(req_list.next, struct addr_req, list);
	345	set_timeout(req->timeout);
	346	}
	347	mutex_unlock(&lock);
	348
	349	list_for_each_entry_safe(req, temp_req, &done_list, list) {
	350	list_del(&req->list);
38617c64 AS	351	req->callback(req->status, (struct sockaddr *) &req->src_addr,
38617c64 AS	352	req->addr, req->context);
7a118df3	353	put_client(req->client);
7025fcd3 SH	354	kfree(req);
	355	}
	356	}
	357
38617c64 AS	358	static int addr_resolve_local(struct sockaddr *src_in,
38617c64 AS	359	struct sockaddr *dst_in,
7025fcd3 SH	360	struct rdma_dev_addr *addr)
	361	{
	362	struct net_device *dev;
7025fcd3 SH	363	int ret;
7025fcd3 SH	364
2c4ab624 RD	365	switch (dst_in->sa_family) {
	366	case AF_INET:
	367	{
38617c64 AS	368	__be32 src_ip = ((struct sockaddr_in *) src_in)->sin_addr.s_addr;
	369	__be32 dst_ip = ((struct sockaddr_in *) dst_in)->sin_addr.s_addr;
	370
	371	dev = ip_dev_find(&init_net, dst_ip);
	372	if (!dev)
	373	return -EADDRNOTAVAIL;
	374
	375	if (ipv4_is_zeronet(src_ip)) {
	376	src_in->sa_family = dst_in->sa_family;
	377	((struct sockaddr_in *) src_in)->sin_addr.s_addr = dst_ip;
	378	ret = rdma_copy_addr(addr, dev, dev->dev_addr);
	379	} else if (ipv4_is_loopback(src_ip)) {
	380	ret = rdma_translate_ip(dst_in, addr);
	381	if (!ret)
	382	memcpy(addr->dst_dev_addr, dev->dev_addr, MAX_ADDR_LEN);
	383	} else {
	384	ret = rdma_translate_ip(src_in, addr);
	385	if (!ret)
	386	memcpy(addr->dst_dev_addr, dev->dev_addr, MAX_ADDR_LEN);
	387	}
	388	dev_put(dev);
2c4ab624 RD	389	break;
	390	}
	391
	392	#if defined(CONFIG_IPV6) \|\| defined(CONFIG_IPV6_MODULE)
	393	case AF_INET6:
	394	{
38617c64 AS	395	struct in6_addr *a;
38617c64 AS	396
0f9ea5d2	397	read_lock(&dev_base_lock);
38617c64 AS	398	for_each_netdev(&init_net, dev)
38617c64 AS	399	if (ipv6_chk_addr(&init_net,
85f20b39	400	&((struct sockaddr_in6 *) dst_in)->sin6_addr,
38617c64 AS	401	dev, 1))
	402	break;
	403
0f9ea5d2 ED	404	if (!dev) {
0f9ea5d2 ED	405	read_unlock(&dev_base_lock);
38617c64	406	return -EADDRNOTAVAIL;
0f9ea5d2	407	}
38617c64 AS	408
	409	a = &((struct sockaddr_in6 *) src_in)->sin6_addr;
	410
	411	if (ipv6_addr_any(a)) {
	412	src_in->sa_family = dst_in->sa_family;
	413	((struct sockaddr_in6 *) src_in)->sin6_addr =
	414	((struct sockaddr_in6 *) dst_in)->sin6_addr;
	415	ret = rdma_copy_addr(addr, dev, dev->dev_addr);
	416	} else if (ipv6_addr_loopback(a)) {
	417	ret = rdma_translate_ip(dst_in, addr);
	418	if (!ret)
	419	memcpy(addr->dst_dev_addr, dev->dev_addr, MAX_ADDR_LEN);
	420	} else {
	421	ret = rdma_translate_ip(src_in, addr);
	422	if (!ret)
	423	memcpy(addr->dst_dev_addr, dev->dev_addr, MAX_ADDR_LEN);
	424	}
0f9ea5d2	425	read_unlock(&dev_base_lock);
2c4ab624 RD	426	break;
	427	}
	428	#endif
	429
	430	default:
	431	ret = -EADDRNOTAVAIL;
	432	break;
7025fcd3 SH	433	}
7025fcd3 SH	434
7025fcd3 SH	435	return ret;
	436	}
	437
7a118df3 SH	438	int rdma_resolve_ip(struct rdma_addr_client *client,
7a118df3 SH	439	struct sockaddr src_addr, struct sockaddr dst_addr,
7025fcd3 SH	440	struct rdma_dev_addr *addr, int timeout_ms,
	441	void (callback)(int status, struct sockaddr src_addr,
	442	struct rdma_dev_addr addr, void context),
	443	void *context)
	444	{
38617c64	445	struct sockaddr src_in, dst_in;
7025fcd3 SH	446	struct addr_req *req;
	447	int ret = 0;
	448
dd00cc48	449	req = kzalloc(sizeof *req, GFP_KERNEL);
7025fcd3 SH	450	if (!req)
7025fcd3 SH	451	return -ENOMEM;
7025fcd3	452
d2e08862 SH	453	src_in = (struct sockaddr *) &req->src_addr;
	454	dst_in = (struct sockaddr *) &req->dst_addr;
	455
	456	if (src_addr) {
	457	if (src_addr->sa_family != dst_addr->sa_family) {
	458	ret = -EINVAL;
	459	goto err;
	460	}
	461
	462	memcpy(src_in, src_addr, ip_addr_size(src_addr));
	463	} else {
	464	src_in->sa_family = dst_addr->sa_family;
	465	}
	466
	467	memcpy(dst_in, dst_addr, ip_addr_size(dst_addr));
7025fcd3 SH	468	req->addr = addr;
	469	req->callback = callback;
	470	req->context = context;
7a118df3 SH	471	req->client = client;
7a118df3 SH	472	atomic_inc(&client->refcount);
7025fcd3	473
7025fcd3 SH	474	req->status = addr_resolve_local(src_in, dst_in, addr);
	475	if (req->status == -EADDRNOTAVAIL)
	476	req->status = addr_resolve_remote(src_in, dst_in, addr);
	477
	478	switch (req->status) {
	479	case 0:
	480	req->timeout = jiffies;
	481	queue_req(req);
	482	break;
	483	case -ENODATA:
	484	req->timeout = msecs_to_jiffies(timeout_ms) + jiffies;
	485	queue_req(req);
	486	addr_send_arp(dst_in);
	487	break;
	488	default:
	489	ret = req->status;
7a118df3	490	atomic_dec(&client->refcount);
d2e08862	491	goto err;
7025fcd3 SH	492	}
7025fcd3 SH	493	return ret;
d2e08862 SH	494	err:
	495	kfree(req);
	496	return ret;
7025fcd3 SH	497	}
	498	EXPORT_SYMBOL(rdma_resolve_ip);
	499
	500	void rdma_addr_cancel(struct rdma_dev_addr *addr)
	501	{
	502	struct addr_req req, temp_req;
	503
	504	mutex_lock(&lock);
	505	list_for_each_entry_safe(req, temp_req, &req_list, list) {
	506	if (req->addr == addr) {
	507	req->status = -ECANCELED;
	508	req->timeout = jiffies;
04699a1f	509	list_move(&req->list, &req_list);
7025fcd3 SH	510	set_timeout(req->timeout);
	511	break;
	512	}
	513	}
	514	mutex_unlock(&lock);
	515	}
	516	EXPORT_SYMBOL(rdma_addr_cancel);
	517
3cd96564	518	static int netevent_callback(struct notifier_block *self, unsigned long event,
e795d092	519	void *ctx)
7025fcd3	520	{
3cd96564	521	if (event == NETEVENT_NEIGH_UPDATE) {
e795d092	522	struct neighbour *neigh = ctx;
7025fcd3	523
1f126670	524	if (neigh->nud_state & NUD_VALID) {
e795d092 TT	525	set_timeout(jiffies);
	526	}
	527	}
7025fcd3 SH	528	return 0;
	529	}
	530
e795d092 TT	531	static struct notifier_block nb = {
e795d092 TT	532	.notifier_call = netevent_callback
7025fcd3 SH	533	};
7025fcd3 SH	534
716abb1f	535	static int __init addr_init(void)
7025fcd3	536	{
c7f743a6	537	addr_wq = create_singlethread_workqueue("ib_addr");
7025fcd3 SH	538	if (!addr_wq)
	539	return -ENOMEM;
	540
e795d092	541	register_netevent_notifier(&nb);
7025fcd3 SH	542	return 0;
	543	}
	544
716abb1f	545	static void __exit addr_cleanup(void)
7025fcd3	546	{
e795d092	547	unregister_netevent_notifier(&nb);
7025fcd3 SH	548	destroy_workqueue(addr_wq);
	549	}
	550
	551	module_init(addr_init);
	552	module_exit(addr_cleanup);