[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 <linux/if_arp.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)
{
	switch (dev->type) {
	case ARPHRD_INFINIBAND:
		dev_addr->dev_type = RDMA_NODE_IB_CA;
		break;
	case ARPHRD_ETHER:
		dev_addr->dev_type = RDMA_NODE_RNIC;
		break;
	default:
		return -EADDRNOTAVAIL;
	}

	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->src_dev = dev;
	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;

	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;
	case AF_INET6:
		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;
			}
		}
		break;
	default:
		break;
	}
	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;
	struct dst_entry *dst;

	memset(&fl, 0, sizeof fl);
	if (dst_in->sa_family == 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);

	} else {
		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);
	}
}

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;
	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;
}

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;

	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;
}

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;

	if (dst_in->sa_family == 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);
	} else {
		struct in6_addr *a;

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

		if (!dev)
			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);
		}
	}

	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;

	if (src_addr)
		memcpy(&req->src_addr, src_addr, ip_addr_size(src_addr));
	memcpy(&req->dst_addr, dst_addr, ip_addr_size(dst_addr));
	req->addr = addr;
	req->callback = callback;
	req->context = context;
	req->client = client;
	atomic_inc(&client->refcount);

	src_in = (struct sockaddr *) &req->src_addr;
	dst_in = (struct sockaddr *) &req->dst_addr;

	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);
		kfree(req);
		break;
	}
	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 addr_init(void)
{
	addr_wq = create_singlethread_workqueue("ib_addr");
	if (!addr_wq)
		return -ENOMEM;

	register_netevent_notifier(&nb);
	return 0;
}

static void 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>
	39	#include <linux/if_arp.h>
	40	#include <net/arp.h>
	41	#include <net/neighbour.h>
	42	#include <net/route.h>
e795d092	43	#include <net/netevent.h>
38617c64 AS	44	#include <net/addrconf.h>
38617c64 AS	45	#include <net/ip6_route.h>
7025fcd3 SH	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;
38617c64 AS	54	struct sockaddr_storage src_addr;
38617c64 AS	55	struct sockaddr_storage dst_addr;
7025fcd3	56	struct rdma_dev_addr *addr;
7a118df3	57	struct rdma_addr_client *client;
7025fcd3 SH	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
c4028958	65	static void process_req(struct work_struct *work);
7025fcd3 SH	66
	67	static DEFINE_MUTEX(lock);
	68	static LIST_HEAD(req_list);
c4028958	69	static DECLARE_DELAYED_WORK(work, process_req);
7025fcd3 SH	70	static struct workqueue_struct *addr_wq;
7025fcd3 SH	71
7a118df3 SH	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
07ebafba TT	92	int rdma_copy_addr(struct rdma_dev_addr dev_addr, struct net_device dev,
07ebafba TT	93	const unsigned char *dst_dev_addr)
7025fcd3 SH	94	{
	95	switch (dev->type) {
	96	case ARPHRD_INFINIBAND:
07ebafba TT	97	dev_addr->dev_type = RDMA_NODE_IB_CA;
	98	break;
	99	case ARPHRD_ETHER:
	100	dev_addr->dev_type = RDMA_NODE_RNIC;
7025fcd3 SH	101	break;
	102	default:
	103	return -EADDRNOTAVAIL;
	104	}
	105
	106	memcpy(dev_addr->src_dev_addr, dev->dev_addr, MAX_ADDR_LEN);
	107	memcpy(dev_addr->broadcast, dev->broadcast, MAX_ADDR_LEN);
	108	if (dst_dev_addr)
	109	memcpy(dev_addr->dst_dev_addr, dst_dev_addr, MAX_ADDR_LEN);
64c5e613	110	dev_addr->src_dev = dev;
7025fcd3 SH	111	return 0;
7025fcd3 SH	112	}
07ebafba	113	EXPORT_SYMBOL(rdma_copy_addr);
7025fcd3 SH	114
	115	int rdma_translate_ip(struct sockaddr addr, struct rdma_dev_addr dev_addr)
	116	{
	117	struct net_device *dev;
38617c64	118	int ret = -EADDRNOTAVAIL;
7025fcd3	119
38617c64 AS	120	switch (addr->sa_family) {
	121	case AF_INET:
	122	dev = ip_dev_find(&init_net,
	123	((struct sockaddr_in *) addr)->sin_addr.s_addr);
	124
	125	if (!dev)
	126	return ret;
7025fcd3	127
38617c64 AS	128	ret = rdma_copy_addr(dev_addr, dev, NULL);
	129	dev_put(dev);
	130	break;
	131	case AF_INET6:
	132	for_each_netdev(&init_net, dev) {
	133	if (ipv6_chk_addr(&init_net,
	134	&((struct sockaddr_in6 *) addr)->sin6_addr,
	135	dev, 1)) {
	136	ret = rdma_copy_addr(dev_addr, dev, NULL);
	137	break;
	138	}
	139	}
	140	break;
	141	default:
	142	break;
	143	}
7025fcd3 SH	144	return ret;
	145	}
	146	EXPORT_SYMBOL(rdma_translate_ip);
	147
	148	static void set_timeout(unsigned long time)
	149	{
	150	unsigned long delay;
	151
	152	cancel_delayed_work(&work);
	153
	154	delay = time - jiffies;
	155	if ((long)delay <= 0)
	156	delay = 1;
	157
	158	queue_delayed_work(addr_wq, &work, delay);
	159	}
	160
	161	static void queue_req(struct addr_req *req)
	162	{
	163	struct addr_req *temp_req;
	164
	165	mutex_lock(&lock);
	166	list_for_each_entry_reverse(temp_req, &req_list, list) {
f115db48	167	if (time_after_eq(req->timeout, temp_req->timeout))
7025fcd3 SH	168	break;
	169	}
	170
	171	list_add(&req->list, &temp_req->list);
	172
	173	if (req_list.next == &req->list)
	174	set_timeout(req->timeout);
	175	mutex_unlock(&lock);
	176	}
	177
38617c64	178	static void addr_send_arp(struct sockaddr *dst_in)
7025fcd3 SH	179	{
	180	struct rtable *rt;
	181	struct flowi fl;
38617c64	182	struct dst_entry *dst;
7025fcd3 SH	183
7025fcd3 SH	184	memset(&fl, 0, sizeof fl);
38617c64 AS	185	if (dst_in->sa_family == AF_INET) {
	186	fl.nl_u.ip4_u.daddr =
	187	((struct sockaddr_in *) dst_in)->sin_addr.s_addr;
7025fcd3	188
38617c64 AS	189	if (ip_route_output_key(&init_net, &rt, &fl))
	190	return;
	191
	192	neigh_event_send(rt->u.dst.neighbour, NULL);
	193	ip_rt_put(rt);
	194
	195	} else {
	196	fl.nl_u.ip6_u.daddr =
	197	((struct sockaddr_in6 *) dst_in)->sin6_addr;
	198
	199	dst = ip6_route_output(&init_net, NULL, &fl);
	200	if (!dst)
	201	return;
	202
	203	neigh_event_send(dst->neighbour, NULL);
	204	dst_release(dst);
	205	}
7025fcd3 SH	206	}
7025fcd3 SH	207
38617c64	208	static int addr4_resolve_remote(struct sockaddr_in *src_in,
7025fcd3 SH	209	struct sockaddr_in *dst_in,
	210	struct rdma_dev_addr *addr)
	211	{
1b90c137 AV	212	__be32 src_ip = src_in->sin_addr.s_addr;
1b90c137 AV	213	__be32 dst_ip = dst_in->sin_addr.s_addr;
7025fcd3 SH	214	struct flowi fl;
	215	struct rtable *rt;
	216	struct neighbour *neigh;
	217	int ret;
	218
	219	memset(&fl, 0, sizeof fl);
	220	fl.nl_u.ip4_u.daddr = dst_ip;
	221	fl.nl_u.ip4_u.saddr = src_ip;
f206351a	222	ret = ip_route_output_key(&init_net, &rt, &fl);
7025fcd3 SH	223	if (ret)
	224	goto out;
	225
	226	/* If the device does ARP internally, return 'done' */
	227	if (rt->idev->dev->flags & IFF_NOARP) {
07ebafba	228	rdma_copy_addr(addr, rt->idev->dev, NULL);
7025fcd3 SH	229	goto put;
	230	}
	231
	232	neigh = neigh_lookup(&arp_tbl, &rt->rt_gateway, rt->idev->dev);
	233	if (!neigh) {
	234	ret = -ENODATA;
	235	goto put;
	236	}
	237
	238	if (!(neigh->nud_state & NUD_VALID)) {
	239	ret = -ENODATA;
	240	goto release;
	241	}
	242
	243	if (!src_ip) {
	244	src_in->sin_family = dst_in->sin_family;
	245	src_in->sin_addr.s_addr = rt->rt_src;
	246	}
	247
07ebafba	248	ret = rdma_copy_addr(addr, neigh->dev, neigh->ha);
7025fcd3 SH	249	release:
	250	neigh_release(neigh);
	251	put:
	252	ip_rt_put(rt);
	253	out:
	254	return ret;
	255	}
	256
38617c64 AS	257	static int addr6_resolve_remote(struct sockaddr_in6 *src_in,
	258	struct sockaddr_in6 *dst_in,
	259	struct rdma_dev_addr *addr)
	260	{
	261	struct flowi fl;
	262	struct neighbour *neigh;
	263	struct dst_entry *dst;
	264	int ret = -ENODATA;
	265
	266	memset(&fl, 0, sizeof fl);
	267	fl.nl_u.ip6_u.daddr = dst_in->sin6_addr;
	268	fl.nl_u.ip6_u.saddr = src_in->sin6_addr;
	269
	270	dst = ip6_route_output(&init_net, NULL, &fl);
	271	if (!dst)
	272	return ret;
	273
	274	if (dst->dev->flags & IFF_NOARP) {
	275	ret = rdma_copy_addr(addr, dst->dev, NULL);
	276	} else {
	277	neigh = dst->neighbour;
	278	if (neigh && (neigh->nud_state & NUD_VALID))
	279	ret = rdma_copy_addr(addr, neigh->dev, neigh->ha);
	280	}
	281
	282	dst_release(dst);
	283	return ret;
	284	}
	285
	286	static int addr_resolve_remote(struct sockaddr *src_in,
	287	struct sockaddr *dst_in,
	288	struct rdma_dev_addr *addr)
	289	{
	290	if (src_in->sa_family == AF_INET) {
	291	return addr4_resolve_remote((struct sockaddr_in *) src_in,
	292	(struct sockaddr_in *) dst_in, addr);
	293	} else
	294	return addr6_resolve_remote((struct sockaddr_in6 *) src_in,
	295	(struct sockaddr_in6 *) dst_in, addr);
	296	}
	297
c4028958	298	static void process_req(struct work_struct *work)
7025fcd3 SH	299	{
7025fcd3 SH	300	struct addr_req req, temp_req;
38617c64	301	struct sockaddr src_in, dst_in;
7025fcd3 SH	302	struct list_head done_list;
	303
	304	INIT_LIST_HEAD(&done_list);
	305
	306	mutex_lock(&lock);
	307	list_for_each_entry_safe(req, temp_req, &req_list, list) {
c78bb844	308	if (req->status == -ENODATA) {
38617c64 AS	309	src_in = (struct sockaddr *) &req->src_addr;
38617c64 AS	310	dst_in = (struct sockaddr *) &req->dst_addr;
7025fcd3 SH	311	req->status = addr_resolve_remote(src_in, dst_in,
7025fcd3 SH	312	req->addr);
c78bb844 KK	313	if (req->status && time_after_eq(jiffies, req->timeout))
	314	req->status = -ETIMEDOUT;
	315	else if (req->status == -ENODATA)
	316	continue;
7025fcd3	317	}
04699a1f	318	list_move_tail(&req->list, &done_list);
7025fcd3 SH	319	}
	320
	321	if (!list_empty(&req_list)) {
	322	req = list_entry(req_list.next, struct addr_req, list);
	323	set_timeout(req->timeout);
	324	}
	325	mutex_unlock(&lock);
	326
	327	list_for_each_entry_safe(req, temp_req, &done_list, list) {
	328	list_del(&req->list);
38617c64 AS	329	req->callback(req->status, (struct sockaddr *) &req->src_addr,
38617c64 AS	330	req->addr, req->context);
7a118df3	331	put_client(req->client);
7025fcd3 SH	332	kfree(req);
	333	}
	334	}
	335
38617c64 AS	336	static int addr_resolve_local(struct sockaddr *src_in,
38617c64 AS	337	struct sockaddr *dst_in,
7025fcd3 SH	338	struct rdma_dev_addr *addr)
	339	{
	340	struct net_device *dev;
7025fcd3 SH	341	int ret;
7025fcd3 SH	342
38617c64 AS	343	if (dst_in->sa_family == AF_INET) {
	344	__be32 src_ip = ((struct sockaddr_in *) src_in)->sin_addr.s_addr;
	345	__be32 dst_ip = ((struct sockaddr_in *) dst_in)->sin_addr.s_addr;
	346
	347	dev = ip_dev_find(&init_net, dst_ip);
	348	if (!dev)
	349	return -EADDRNOTAVAIL;
	350
	351	if (ipv4_is_zeronet(src_ip)) {
	352	src_in->sa_family = dst_in->sa_family;
	353	((struct sockaddr_in *) src_in)->sin_addr.s_addr = dst_ip;
	354	ret = rdma_copy_addr(addr, dev, dev->dev_addr);
	355	} else if (ipv4_is_loopback(src_ip)) {
	356	ret = rdma_translate_ip(dst_in, addr);
	357	if (!ret)
	358	memcpy(addr->dst_dev_addr, dev->dev_addr, MAX_ADDR_LEN);
	359	} else {
	360	ret = rdma_translate_ip(src_in, addr);
	361	if (!ret)
	362	memcpy(addr->dst_dev_addr, dev->dev_addr, MAX_ADDR_LEN);
	363	}
	364	dev_put(dev);
7025fcd3	365	} else {
38617c64 AS	366	struct in6_addr *a;
	367
	368	for_each_netdev(&init_net, dev)
	369	if (ipv6_chk_addr(&init_net,
	370	&((struct sockaddr_in6 *) addr)->sin6_addr,
	371	dev, 1))
	372	break;
	373
	374	if (!dev)
	375	return -EADDRNOTAVAIL;
	376
	377	a = &((struct sockaddr_in6 *) src_in)->sin6_addr;
	378
	379	if (ipv6_addr_any(a)) {
	380	src_in->sa_family = dst_in->sa_family;
	381	((struct sockaddr_in6 *) src_in)->sin6_addr =
	382	((struct sockaddr_in6 *) dst_in)->sin6_addr;
	383	ret = rdma_copy_addr(addr, dev, dev->dev_addr);
	384	} else if (ipv6_addr_loopback(a)) {
	385	ret = rdma_translate_ip(dst_in, addr);
	386	if (!ret)
	387	memcpy(addr->dst_dev_addr, dev->dev_addr, MAX_ADDR_LEN);
	388	} else {
	389	ret = rdma_translate_ip(src_in, addr);
	390	if (!ret)
	391	memcpy(addr->dst_dev_addr, dev->dev_addr, MAX_ADDR_LEN);
	392	}
7025fcd3 SH	393	}
7025fcd3 SH	394
7025fcd3 SH	395	return ret;
	396	}
	397
7a118df3 SH	398	int rdma_resolve_ip(struct rdma_addr_client *client,
7a118df3 SH	399	struct sockaddr src_addr, struct sockaddr dst_addr,
7025fcd3 SH	400	struct rdma_dev_addr *addr, int timeout_ms,
	401	void (callback)(int status, struct sockaddr src_addr,
	402	struct rdma_dev_addr addr, void context),
	403	void *context)
	404	{
38617c64	405	struct sockaddr src_in, dst_in;
7025fcd3 SH	406	struct addr_req *req;
	407	int ret = 0;
	408
dd00cc48	409	req = kzalloc(sizeof *req, GFP_KERNEL);
7025fcd3 SH	410	if (!req)
7025fcd3 SH	411	return -ENOMEM;
7025fcd3 SH	412
	413	if (src_addr)
	414	memcpy(&req->src_addr, src_addr, ip_addr_size(src_addr));
	415	memcpy(&req->dst_addr, dst_addr, ip_addr_size(dst_addr));
	416	req->addr = addr;
	417	req->callback = callback;
	418	req->context = context;
7a118df3 SH	419	req->client = client;
7a118df3 SH	420	atomic_inc(&client->refcount);
7025fcd3	421
38617c64 AS	422	src_in = (struct sockaddr *) &req->src_addr;
38617c64 AS	423	dst_in = (struct sockaddr *) &req->dst_addr;
7025fcd3 SH	424
	425	req->status = addr_resolve_local(src_in, dst_in, addr);
	426	if (req->status == -EADDRNOTAVAIL)
	427	req->status = addr_resolve_remote(src_in, dst_in, addr);
	428
	429	switch (req->status) {
	430	case 0:
	431	req->timeout = jiffies;
	432	queue_req(req);
	433	break;
	434	case -ENODATA:
	435	req->timeout = msecs_to_jiffies(timeout_ms) + jiffies;
	436	queue_req(req);
	437	addr_send_arp(dst_in);
	438	break;
	439	default:
	440	ret = req->status;
7a118df3	441	atomic_dec(&client->refcount);
7025fcd3 SH	442	kfree(req);
	443	break;
	444	}
	445	return ret;
	446	}
	447	EXPORT_SYMBOL(rdma_resolve_ip);
	448
	449	void rdma_addr_cancel(struct rdma_dev_addr *addr)
	450	{
	451	struct addr_req req, temp_req;
	452
	453	mutex_lock(&lock);
	454	list_for_each_entry_safe(req, temp_req, &req_list, list) {
	455	if (req->addr == addr) {
	456	req->status = -ECANCELED;
	457	req->timeout = jiffies;
04699a1f	458	list_move(&req->list, &req_list);
7025fcd3 SH	459	set_timeout(req->timeout);
	460	break;
	461	}
	462	}
	463	mutex_unlock(&lock);
	464	}
	465	EXPORT_SYMBOL(rdma_addr_cancel);
	466
3cd96564	467	static int netevent_callback(struct notifier_block *self, unsigned long event,
e795d092	468	void *ctx)
7025fcd3	469	{
3cd96564	470	if (event == NETEVENT_NEIGH_UPDATE) {
e795d092	471	struct neighbour *neigh = ctx;
7025fcd3	472
1f126670	473	if (neigh->nud_state & NUD_VALID) {
e795d092 TT	474	set_timeout(jiffies);
	475	}
	476	}
7025fcd3 SH	477	return 0;
	478	}
	479
e795d092 TT	480	static struct notifier_block nb = {
e795d092 TT	481	.notifier_call = netevent_callback
7025fcd3 SH	482	};
	483
	484	static int addr_init(void)
	485	{
c7f743a6	486	addr_wq = create_singlethread_workqueue("ib_addr");
7025fcd3 SH	487	if (!addr_wq)
	488	return -ENOMEM;
	489
e795d092	490	register_netevent_notifier(&nb);
7025fcd3 SH	491	return 0;
	492	}
	493
	494	static void addr_cleanup(void)
	495	{
e795d092	496	unregister_netevent_notifier(&nb);
7025fcd3 SH	497	destroy_workqueue(addr_wq);
	498	}
	499
	500	module_init(addr_init);
	501	module_exit(addr_cleanup);