[deliverable/linux.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 <linux/module.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>
#include <rdma/ib.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;

int rdma_addr_size(struct sockaddr *addr)
{
	switch (addr->sa_family) {
	case AF_INET:
		return sizeof(struct sockaddr_in);
	case AF_INET6:
		return sizeof(struct sockaddr_in6);
	case AF_IB:
		return sizeof(struct sockaddr_ib);
	default:
		return 0;
	}
}
EXPORT_SYMBOL(rdma_addr_size);

static struct rdma_addr_client self;

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,
		      u16 *vlan_id)
{
	struct net_device *dev;
	int ret = -EADDRNOTAVAIL;

	if (dev_addr->bound_dev_if) {
		dev = dev_get_by_index(dev_addr->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(dev_addr->net,
			((struct sockaddr_in *) addr)->sin_addr.s_addr);

		if (!dev)
			return ret;

		ret = rdma_copy_addr(dev_addr, dev, NULL);
		if (vlan_id)
			*vlan_id = rdma_vlan_dev_vlan_id(dev);
		dev_put(dev);
		break;
#if IS_ENABLED(CONFIG_IPV6)
	case AF_INET6:
		rcu_read_lock();
		for_each_netdev_rcu(dev_addr->net, dev) {
			if (ipv6_chk_addr(dev_addr->net,
					  &((struct sockaddr_in6 *) addr)->sin6_addr,
					  dev, 1)) {
				ret = rdma_copy_addr(dev_addr, dev, NULL);
				if (vlan_id)
					*vlan_id = rdma_vlan_dev_vlan_id(dev);
				break;
			}
		}
		rcu_read_unlock();
		break;
#endif
	}
	return ret;
}
EXPORT_SYMBOL(rdma_translate_ip);

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

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

	mod_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 dst_fetch_ha(struct dst_entry *dst, struct rdma_dev_addr *dev_addr, void *daddr)
{
	struct neighbour *n;
	int ret;

	n = dst_neigh_lookup(dst, daddr);

	rcu_read_lock();
	if (!n || !(n->nud_state & NUD_VALID)) {
		if (n)
			neigh_event_send(n, NULL);
		ret = -ENODATA;
	} else {
		ret = rdma_copy_addr(dev_addr, dst->dev, n->ha);
	}
	rcu_read_unlock();

	if (n)
		neigh_release(n);

	return ret;
}

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 rtable *rt;
	struct flowi4 fl4;
	int ret;

	memset(&fl4, 0, sizeof(fl4));
	fl4.daddr = dst_ip;
	fl4.saddr = src_ip;
	fl4.flowi4_oif = addr->bound_dev_if;
	rt = ip_route_output_key(addr->net, &fl4);
	if (IS_ERR(rt)) {
		ret = PTR_ERR(rt);
		goto out;
	}
	src_in->sin_family = AF_INET;
	src_in->sin_addr.s_addr = fl4.saddr;

	if (rt->dst.dev->flags & IFF_LOOPBACK) {
		ret = rdma_translate_ip((struct sockaddr *)dst_in, addr, NULL);
		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) {
		ret = rdma_copy_addr(addr, rt->dst.dev, NULL);
		goto put;
	}

	ret = dst_fetch_ha(&rt->dst, addr, &fl4.daddr);
put:
	ip_rt_put(rt);
out:
	return ret;
}

#if IS_ENABLED(CONFIG_IPV6)
static int addr6_resolve(struct sockaddr_in6 *src_in,
			 struct sockaddr_in6 *dst_in,
			 struct rdma_dev_addr *addr)
{
	struct flowi6 fl6;
	struct dst_entry *dst;
	int ret;

	memset(&fl6, 0, sizeof fl6);
	fl6.daddr = dst_in->sin6_addr;
	fl6.saddr = src_in->sin6_addr;
	fl6.flowi6_oif = addr->bound_dev_if;

	dst = ip6_route_output(addr->net, NULL, &fl6);
	if ((ret = dst->error))
		goto put;

	if (ipv6_addr_any(&fl6.saddr)) {
		ret = ipv6_dev_get_saddr(addr->net, ip6_dst_idev(dst)->dev,
					 &fl6.daddr, 0, &fl6.saddr);
		if (ret)
			goto put;

		src_in->sin6_family = AF_INET6;
		src_in->sin6_addr = fl6.saddr;
	}

	if (dst->dev->flags & IFF_LOOPBACK) {
		ret = rdma_translate_ip((struct sockaddr *)dst_in, addr, NULL);
		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;
	}

	ret = dst_fetch_ha(dst, addr, &fl6.daddr);
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, rdma_addr_size(src_addr));
	} else {
		src_in->sa_family = dst_addr->sa_family;
	}

	memcpy(dst_in, dst_addr, rdma_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);

struct resolve_cb_context {
	struct rdma_dev_addr *addr;
	struct completion comp;
};

static void resolve_cb(int status, struct sockaddr *src_addr,
	     struct rdma_dev_addr *addr, void *context)
{
	memcpy(((struct resolve_cb_context *)context)->addr, addr, sizeof(struct
				rdma_dev_addr));
	complete(&((struct resolve_cb_context *)context)->comp);
}

int rdma_addr_find_dmac_by_grh(const union ib_gid *sgid, const union ib_gid *dgid,
			       u8 *dmac, u16 *vlan_id, int if_index)
{
	int ret = 0;
	struct rdma_dev_addr dev_addr;
	struct resolve_cb_context ctx;
	struct net_device *dev;

	union {
		struct sockaddr     _sockaddr;
		struct sockaddr_in  _sockaddr_in;
		struct sockaddr_in6 _sockaddr_in6;
	} sgid_addr, dgid_addr;


	rdma_gid2ip(&sgid_addr._sockaddr, sgid);
	rdma_gid2ip(&dgid_addr._sockaddr, dgid);

	memset(&dev_addr, 0, sizeof(dev_addr));
	dev_addr.bound_dev_if = if_index;
	dev_addr.net = &init_net;

	ctx.addr = &dev_addr;
	init_completion(&ctx.comp);
	ret = rdma_resolve_ip(&self, &sgid_addr._sockaddr, &dgid_addr._sockaddr,
			&dev_addr, 1000, resolve_cb, &ctx);
	if (ret)
		return ret;

	wait_for_completion(&ctx.comp);

	memcpy(dmac, dev_addr.dst_dev_addr, ETH_ALEN);
	dev = dev_get_by_index(&init_net, dev_addr.bound_dev_if);
	if (!dev)
		return -ENODEV;
	if (vlan_id)
		*vlan_id = rdma_vlan_dev_vlan_id(dev);
	dev_put(dev);
	return ret;
}
EXPORT_SYMBOL(rdma_addr_find_dmac_by_grh);

int rdma_addr_find_smac_by_sgid(union ib_gid *sgid, u8 *smac, u16 *vlan_id)
{
	int ret = 0;
	struct rdma_dev_addr dev_addr;
	union {
		struct sockaddr     _sockaddr;
		struct sockaddr_in  _sockaddr_in;
		struct sockaddr_in6 _sockaddr_in6;
	} gid_addr;

	rdma_gid2ip(&gid_addr._sockaddr, sgid);

	memset(&dev_addr, 0, sizeof(dev_addr));
	dev_addr.net = &init_net;
	ret = rdma_translate_ip(&gid_addr._sockaddr, &dev_addr, vlan_id);
	if (ret)
		return ret;

	memcpy(smac, dev_addr.src_dev_addr, ETH_ALEN);
	return ret;
}
EXPORT_SYMBOL(rdma_addr_find_smac_by_sgid);

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);
	rdma_addr_register_client(&self);
	return 0;
}

static void __exit addr_cleanup(void)
{
	rdma_addr_unregister_client(&self);
	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>
5a0e3ad6	38	#include <linux/slab.h>
7025fcd3	39	#include <linux/workqueue.h>
e4dd23d7	40	#include <linux/module.h>
7025fcd3 SH	41	#include <net/arp.h>
	42	#include <net/neighbour.h>
	43	#include <net/route.h>
e795d092	44	#include <net/netevent.h>
38617c64 AS	45	#include <net/addrconf.h>
38617c64 AS	46	#include <net/ip6_route.h>
7025fcd3	47	#include <rdma/ib_addr.h>
ef560861	48	#include <rdma/ib.h>
7025fcd3 SH	49
	50	MODULE_AUTHOR("Sean Hefty");
	51	MODULE_DESCRIPTION("IB Address Translation");
	52	MODULE_LICENSE("Dual BSD/GPL");
	53
	54	struct addr_req {
	55	struct list_head list;
38617c64 AS	56	struct sockaddr_storage src_addr;
38617c64 AS	57	struct sockaddr_storage dst_addr;
7025fcd3	58	struct rdma_dev_addr *addr;
7a118df3	59	struct rdma_addr_client *client;
7025fcd3 SH	60	void *context;
	61	void (callback)(int status, struct sockaddr src_addr,
	62	struct rdma_dev_addr addr, void context);
	63	unsigned long timeout;
	64	int status;
	65	};
	66
c4028958	67	static void process_req(struct work_struct *work);
7025fcd3 SH	68
	69	static DEFINE_MUTEX(lock);
	70	static LIST_HEAD(req_list);
c4028958	71	static DECLARE_DELAYED_WORK(work, process_req);
7025fcd3 SH	72	static struct workqueue_struct *addr_wq;
7025fcd3 SH	73
ef560861 SH	74	int rdma_addr_size(struct sockaddr *addr)
	75	{
	76	switch (addr->sa_family) {
	77	case AF_INET:
	78	return sizeof(struct sockaddr_in);
	79	case AF_INET6:
	80	return sizeof(struct sockaddr_in6);
	81	case AF_IB:
	82	return sizeof(struct sockaddr_ib);
	83	default:
	84	return 0;
	85	}
	86	}
	87	EXPORT_SYMBOL(rdma_addr_size);
	88
dd5f03be MB	89	static struct rdma_addr_client self;
dd5f03be MB	90
7a118df3 SH	91	void rdma_addr_register_client(struct rdma_addr_client *client)
	92	{
	93	atomic_set(&client->refcount, 1);
	94	init_completion(&client->comp);
	95	}
	96	EXPORT_SYMBOL(rdma_addr_register_client);
	97
	98	static inline void put_client(struct rdma_addr_client *client)
	99	{
	100	if (atomic_dec_and_test(&client->refcount))
	101	complete(&client->comp);
	102	}
	103
	104	void rdma_addr_unregister_client(struct rdma_addr_client *client)
	105	{
	106	put_client(client);
	107	wait_for_completion(&client->comp);
	108	}
	109	EXPORT_SYMBOL(rdma_addr_unregister_client);
	110
07ebafba TT	111	int rdma_copy_addr(struct rdma_dev_addr dev_addr, struct net_device dev,
07ebafba TT	112	const unsigned char *dst_dev_addr)
7025fcd3	113	{
c4315d85	114	dev_addr->dev_type = dev->type;
7025fcd3 SH	115	memcpy(dev_addr->src_dev_addr, dev->dev_addr, MAX_ADDR_LEN);
	116	memcpy(dev_addr->broadcast, dev->broadcast, MAX_ADDR_LEN);
	117	if (dst_dev_addr)
	118	memcpy(dev_addr->dst_dev_addr, dst_dev_addr, MAX_ADDR_LEN);
6266ed6e	119	dev_addr->bound_dev_if = dev->ifindex;
7025fcd3 SH	120	return 0;
7025fcd3 SH	121	}
07ebafba	122	EXPORT_SYMBOL(rdma_copy_addr);
7025fcd3	123
dd5f03be MB	124	int rdma_translate_ip(struct sockaddr addr, struct rdma_dev_addr dev_addr,
dd5f03be MB	125	u16 *vlan_id)
7025fcd3 SH	126	{
7025fcd3 SH	127	struct net_device *dev;
38617c64	128	int ret = -EADDRNOTAVAIL;
7025fcd3	129
6266ed6e	130	if (dev_addr->bound_dev_if) {
565edd1d	131	dev = dev_get_by_index(dev_addr->net, dev_addr->bound_dev_if);
6266ed6e SH	132	if (!dev)
	133	return -ENODEV;
	134	ret = rdma_copy_addr(dev_addr, dev, NULL);
	135	dev_put(dev);
	136	return ret;
	137	}
	138
38617c64 AS	139	switch (addr->sa_family) {
38617c64 AS	140	case AF_INET:
565edd1d	141	dev = ip_dev_find(dev_addr->net,
38617c64 AS	142	((struct sockaddr_in *) addr)->sin_addr.s_addr);
	143
	144	if (!dev)
	145	return ret;
7025fcd3	146
38617c64	147	ret = rdma_copy_addr(dev_addr, dev, NULL);
dd5f03be MB	148	if (vlan_id)
dd5f03be MB	149	*vlan_id = rdma_vlan_dev_vlan_id(dev);
38617c64 AS	150	dev_put(dev);
38617c64 AS	151	break;
d90f9b35	152	#if IS_ENABLED(CONFIG_IPV6)
38617c64	153	case AF_INET6:
22f4fbd9	154	rcu_read_lock();
565edd1d GS	155	for_each_netdev_rcu(dev_addr->net, dev) {
565edd1d GS	156	if (ipv6_chk_addr(dev_addr->net,
38617c64 AS	157	&((struct sockaddr_in6 *) addr)->sin6_addr,
	158	dev, 1)) {
	159	ret = rdma_copy_addr(dev_addr, dev, NULL);
dd5f03be MB	160	if (vlan_id)
dd5f03be MB	161	*vlan_id = rdma_vlan_dev_vlan_id(dev);
38617c64 AS	162	break;
	163	}
	164	}
22f4fbd9	165	rcu_read_unlock();
38617c64	166	break;
2c4ab624	167	#endif
38617c64	168	}
7025fcd3 SH	169	return ret;
	170	}
	171	EXPORT_SYMBOL(rdma_translate_ip);
	172
	173	static void set_timeout(unsigned long time)
	174	{
	175	unsigned long delay;
	176
7025fcd3	177	delay = time - jiffies;
346f98b4 OK	178	if ((long)delay < 0)
346f98b4 OK	179	delay = 0;
7025fcd3	180
41f63c53	181	mod_delayed_work(addr_wq, &work, delay);
7025fcd3 SH	182	}
	183
	184	static void queue_req(struct addr_req *req)
	185	{
	186	struct addr_req *temp_req;
	187
	188	mutex_lock(&lock);
	189	list_for_each_entry_reverse(temp_req, &req_list, list) {
f115db48	190	if (time_after_eq(req->timeout, temp_req->timeout))
7025fcd3 SH	191	break;
	192	}
	193
	194	list_add(&req->list, &temp_req->list);
	195
	196	if (req_list.next == &req->list)
	197	set_timeout(req->timeout);
	198	mutex_unlock(&lock);
	199	}
	200
02b61955	201	static int dst_fetch_ha(struct dst_entry dst, struct rdma_dev_addr dev_addr, void *daddr)
51d45974 DM	202	{
	203	struct neighbour *n;
	204	int ret;
	205
02b61955 DM	206	n = dst_neigh_lookup(dst, daddr);
02b61955 DM	207
51d45974	208	rcu_read_lock();
51d45974 DM	209	if (!n \|\| !(n->nud_state & NUD_VALID)) {
	210	if (n)
	211	neigh_event_send(n, NULL);
	212	ret = -ENODATA;
	213	} else {
02b61955	214	ret = rdma_copy_addr(dev_addr, dst->dev, n->ha);
51d45974 DM	215	}
	216	rcu_read_unlock();
	217
02b61955 DM	218	if (n)
	219	neigh_release(n);
	220
51d45974 DM	221	return ret;
	222	}
	223
923c100e SH	224	static int addr4_resolve(struct sockaddr_in *src_in,
	225	struct sockaddr_in *dst_in,
	226	struct rdma_dev_addr *addr)
7025fcd3	227	{
1b90c137 AV	228	__be32 src_ip = src_in->sin_addr.s_addr;
1b90c137 AV	229	__be32 dst_ip = dst_in->sin_addr.s_addr;
7025fcd3	230	struct rtable *rt;
5fc3590c	231	struct flowi4 fl4;
7025fcd3 SH	232	int ret;
7025fcd3 SH	233
5fc3590c DM	234	memset(&fl4, 0, sizeof(fl4));
	235	fl4.daddr = dst_ip;
	236	fl4.saddr = src_ip;
	237	fl4.flowi4_oif = addr->bound_dev_if;
565edd1d	238	rt = ip_route_output_key(addr->net, &fl4);
b23dd4fe DM	239	if (IS_ERR(rt)) {
b23dd4fe DM	240	ret = PTR_ERR(rt);
7025fcd3	241	goto out;
b23dd4fe	242	}
923c100e	243	src_in->sin_family = AF_INET;
5fc3590c	244	src_in->sin_addr.s_addr = fl4.saddr;
923c100e	245
72cdd1d9	246	if (rt->dst.dev->flags & IFF_LOOPBACK) {
dd5f03be	247	ret = rdma_translate_ip((struct sockaddr *)dst_in, addr, NULL);
923c100e SH	248	if (!ret)
	249	memcpy(addr->dst_dev_addr, addr->src_dev_addr, MAX_ADDR_LEN);
	250	goto put;
	251	}
	252
7025fcd3	253	/* If the device does ARP internally, return 'done' */
72cdd1d9	254	if (rt->dst.dev->flags & IFF_NOARP) {
1bdd6384	255	ret = rdma_copy_addr(addr, rt->dst.dev, NULL);
7025fcd3 SH	256	goto put;
	257	}
	258
02b61955	259	ret = dst_fetch_ha(&rt->dst, addr, &fl4.daddr);
7025fcd3 SH	260	put:
	261	ip_rt_put(rt);
	262	out:
	263	return ret;
	264	}
	265
d90f9b35	266	#if IS_ENABLED(CONFIG_IPV6)
d14714df SH	267	static int addr6_resolve(struct sockaddr_in6 *src_in,
	268	struct sockaddr_in6 *dst_in,
	269	struct rdma_dev_addr *addr)
38617c64	270	{
4c9483b2	271	struct flowi6 fl6;
38617c64	272	struct dst_entry *dst;
d14714df	273	int ret;
38617c64	274
4c9483b2	275	memset(&fl6, 0, sizeof fl6);
4e3fd7a0 AD	276	fl6.daddr = dst_in->sin6_addr;
4e3fd7a0 AD	277	fl6.saddr = src_in->sin6_addr;
4c9483b2	278	fl6.flowi6_oif = addr->bound_dev_if;
38617c64	279
565edd1d	280	dst = ip6_route_output(addr->net, NULL, &fl6);
d14714df SH	281	if ((ret = dst->error))
	282	goto put;
	283
4c9483b2	284	if (ipv6_addr_any(&fl6.saddr)) {
565edd1d	285	ret = ipv6_dev_get_saddr(addr->net, ip6_dst_idev(dst)->dev,
4c9483b2	286	&fl6.daddr, 0, &fl6.saddr);
d14714df SH	287	if (ret)
d14714df SH	288	goto put;
38617c64	289
d14714df	290	src_in->sin6_family = AF_INET6;
4e3fd7a0	291	src_in->sin6_addr = fl6.saddr;
d14714df SH	292	}
	293
	294	if (dst->dev->flags & IFF_LOOPBACK) {
dd5f03be	295	ret = rdma_translate_ip((struct sockaddr *)dst_in, addr, NULL);
d14714df SH	296	if (!ret)
	297	memcpy(addr->dst_dev_addr, addr->src_dev_addr, MAX_ADDR_LEN);
	298	goto put;
	299	}
	300
	301	/* If the device does ARP internally, return 'done' */
38617c64 AS	302	if (dst->dev->flags & IFF_NOARP) {
38617c64 AS	303	ret = rdma_copy_addr(addr, dst->dev, NULL);
d14714df	304	goto put;
38617c64 AS	305	}
38617c64 AS	306
02b61955	307	ret = dst_fetch_ha(dst, addr, &fl6.daddr);
d14714df	308	put:
38617c64 AS	309	dst_release(dst);
	310	return ret;
	311	}
2c4ab624	312	#else
d14714df SH	313	static int addr6_resolve(struct sockaddr_in6 *src_in,
	314	struct sockaddr_in6 *dst_in,
	315	struct rdma_dev_addr *addr)
2c4ab624 RD	316	{
	317	return -EADDRNOTAVAIL;
	318	}
	319	#endif
38617c64	320
923c100e SH	321	static int addr_resolve(struct sockaddr *src_in,
	322	struct sockaddr *dst_in,
	323	struct rdma_dev_addr *addr)
38617c64 AS	324	{
38617c64 AS	325	if (src_in->sa_family == AF_INET) {
923c100e	326	return addr4_resolve((struct sockaddr_in *) src_in,
38617c64 AS	327	(struct sockaddr_in *) dst_in, addr);
38617c64 AS	328	} else
d14714df	329	return addr6_resolve((struct sockaddr_in6 *) src_in,
38617c64 AS	330	(struct sockaddr_in6 *) dst_in, addr);
	331	}
	332
c4028958	333	static void process_req(struct work_struct *work)
7025fcd3 SH	334	{
7025fcd3 SH	335	struct addr_req req, temp_req;
38617c64	336	struct sockaddr src_in, dst_in;
7025fcd3 SH	337	struct list_head done_list;
	338
	339	INIT_LIST_HEAD(&done_list);
	340
	341	mutex_lock(&lock);
	342	list_for_each_entry_safe(req, temp_req, &req_list, list) {
c78bb844	343	if (req->status == -ENODATA) {
38617c64 AS	344	src_in = (struct sockaddr *) &req->src_addr;
38617c64 AS	345	dst_in = (struct sockaddr *) &req->dst_addr;
923c100e	346	req->status = addr_resolve(src_in, dst_in, req->addr);
c78bb844 KK	347	if (req->status && time_after_eq(jiffies, req->timeout))
	348	req->status = -ETIMEDOUT;
	349	else if (req->status == -ENODATA)
	350	continue;
7025fcd3	351	}
04699a1f	352	list_move_tail(&req->list, &done_list);
7025fcd3 SH	353	}
	354
	355	if (!list_empty(&req_list)) {
	356	req = list_entry(req_list.next, struct addr_req, list);
	357	set_timeout(req->timeout);
	358	}
	359	mutex_unlock(&lock);
	360
	361	list_for_each_entry_safe(req, temp_req, &done_list, list) {
	362	list_del(&req->list);
38617c64 AS	363	req->callback(req->status, (struct sockaddr *) &req->src_addr,
38617c64 AS	364	req->addr, req->context);
7a118df3	365	put_client(req->client);
7025fcd3 SH	366	kfree(req);
	367	}
	368	}
	369
7a118df3 SH	370	int rdma_resolve_ip(struct rdma_addr_client *client,
7a118df3 SH	371	struct sockaddr src_addr, struct sockaddr dst_addr,
7025fcd3 SH	372	struct rdma_dev_addr *addr, int timeout_ms,
	373	void (callback)(int status, struct sockaddr src_addr,
	374	struct rdma_dev_addr addr, void context),
	375	void *context)
	376	{
38617c64	377	struct sockaddr src_in, dst_in;
7025fcd3 SH	378	struct addr_req *req;
	379	int ret = 0;
	380
dd00cc48	381	req = kzalloc(sizeof *req, GFP_KERNEL);
7025fcd3 SH	382	if (!req)
7025fcd3 SH	383	return -ENOMEM;
7025fcd3	384
d2e08862 SH	385	src_in = (struct sockaddr *) &req->src_addr;
	386	dst_in = (struct sockaddr *) &req->dst_addr;
	387
	388	if (src_addr) {
	389	if (src_addr->sa_family != dst_addr->sa_family) {
	390	ret = -EINVAL;
	391	goto err;
	392	}
	393
ef560861	394	memcpy(src_in, src_addr, rdma_addr_size(src_addr));
d2e08862 SH	395	} else {
	396	src_in->sa_family = dst_addr->sa_family;
	397	}
	398
ef560861	399	memcpy(dst_in, dst_addr, rdma_addr_size(dst_addr));
7025fcd3 SH	400	req->addr = addr;
	401	req->callback = callback;
	402	req->context = context;
7a118df3 SH	403	req->client = client;
7a118df3 SH	404	atomic_inc(&client->refcount);
7025fcd3	405
d14714df	406	req->status = addr_resolve(src_in, dst_in, addr);
7025fcd3 SH	407	switch (req->status) {
	408	case 0:
	409	req->timeout = jiffies;
	410	queue_req(req);
	411	break;
	412	case -ENODATA:
	413	req->timeout = msecs_to_jiffies(timeout_ms) + jiffies;
	414	queue_req(req);
7025fcd3 SH	415	break;
	416	default:
	417	ret = req->status;
7a118df3	418	atomic_dec(&client->refcount);
d2e08862	419	goto err;
7025fcd3 SH	420	}
7025fcd3 SH	421	return ret;
d2e08862 SH	422	err:
	423	kfree(req);
	424	return ret;
7025fcd3 SH	425	}
	426	EXPORT_SYMBOL(rdma_resolve_ip);
	427
	428	void rdma_addr_cancel(struct rdma_dev_addr *addr)
	429	{
	430	struct addr_req req, temp_req;
	431
	432	mutex_lock(&lock);
	433	list_for_each_entry_safe(req, temp_req, &req_list, list) {
	434	if (req->addr == addr) {
	435	req->status = -ECANCELED;
	436	req->timeout = jiffies;
04699a1f	437	list_move(&req->list, &req_list);
7025fcd3 SH	438	set_timeout(req->timeout);
	439	break;
	440	}
	441	}
	442	mutex_unlock(&lock);
	443	}
	444	EXPORT_SYMBOL(rdma_addr_cancel);
	445
dd5f03be MB	446	struct resolve_cb_context {
	447	struct rdma_dev_addr *addr;
	448	struct completion comp;
	449	};
	450
	451	static void resolve_cb(int status, struct sockaddr *src_addr,
	452	struct rdma_dev_addr addr, void context)
	453	{
	454	memcpy(((struct resolve_cb_context *)context)->addr, addr, sizeof(struct
	455	rdma_dev_addr));
	456	complete(&((struct resolve_cb_context *)context)->comp);
	457	}
	458
73cdaaee	459	int rdma_addr_find_dmac_by_grh(const union ib_gid sgid, const union ib_gid dgid,
dbf727de	460	u8 dmac, u16 vlan_id, int if_index)
dd5f03be MB	461	{
	462	int ret = 0;
	463	struct rdma_dev_addr dev_addr;
	464	struct resolve_cb_context ctx;
	465	struct net_device *dev;
	466
	467	union {
	468	struct sockaddr _sockaddr;
	469	struct sockaddr_in _sockaddr_in;
	470	struct sockaddr_in6 _sockaddr_in6;
	471	} sgid_addr, dgid_addr;
	472
	473
471e7058 HL	474	rdma_gid2ip(&sgid_addr._sockaddr, sgid);
471e7058 HL	475	rdma_gid2ip(&dgid_addr._sockaddr, dgid);
dd5f03be MB	476
dd5f03be MB	477	memset(&dev_addr, 0, sizeof(dev_addr));
dbf727de	478	dev_addr.bound_dev_if = if_index;
565edd1d	479	dev_addr.net = &init_net;
dd5f03be MB	480
	481	ctx.addr = &dev_addr;
	482	init_completion(&ctx.comp);
	483	ret = rdma_resolve_ip(&self, &sgid_addr._sockaddr, &dgid_addr._sockaddr,
	484	&dev_addr, 1000, resolve_cb, &ctx);
	485	if (ret)
	486	return ret;
	487
	488	wait_for_completion(&ctx.comp);
	489
	490	memcpy(dmac, dev_addr.dst_dev_addr, ETH_ALEN);
	491	dev = dev_get_by_index(&init_net, dev_addr.bound_dev_if);
	492	if (!dev)
	493	return -ENODEV;
	494	if (vlan_id)
	495	*vlan_id = rdma_vlan_dev_vlan_id(dev);
	496	dev_put(dev);
	497	return ret;
	498	}
	499	EXPORT_SYMBOL(rdma_addr_find_dmac_by_grh);
	500
	501	int rdma_addr_find_smac_by_sgid(union ib_gid sgid, u8 smac, u16 *vlan_id)
	502	{
	503	int ret = 0;
	504	struct rdma_dev_addr dev_addr;
	505	union {
	506	struct sockaddr _sockaddr;
	507	struct sockaddr_in _sockaddr_in;
	508	struct sockaddr_in6 _sockaddr_in6;
	509	} gid_addr;
	510
471e7058	511	rdma_gid2ip(&gid_addr._sockaddr, sgid);
dd5f03be	512
dd5f03be	513	memset(&dev_addr, 0, sizeof(dev_addr));
565edd1d	514	dev_addr.net = &init_net;
dd5f03be MB	515	ret = rdma_translate_ip(&gid_addr._sockaddr, &dev_addr, vlan_id);
	516	if (ret)
	517	return ret;
	518
	519	memcpy(smac, dev_addr.src_dev_addr, ETH_ALEN);
	520	return ret;
	521	}
	522	EXPORT_SYMBOL(rdma_addr_find_smac_by_sgid);
	523
3cd96564	524	static int netevent_callback(struct notifier_block *self, unsigned long event,
e795d092	525	void *ctx)
7025fcd3	526	{
3cd96564	527	if (event == NETEVENT_NEIGH_UPDATE) {
e795d092	528	struct neighbour *neigh = ctx;
7025fcd3	529
1f126670	530	if (neigh->nud_state & NUD_VALID) {
e795d092 TT	531	set_timeout(jiffies);
	532	}
	533	}
7025fcd3 SH	534	return 0;
	535	}
	536
e795d092 TT	537	static struct notifier_block nb = {
e795d092 TT	538	.notifier_call = netevent_callback
7025fcd3 SH	539	};
7025fcd3 SH	540
716abb1f	541	static int __init addr_init(void)
7025fcd3	542	{
c7f743a6	543	addr_wq = create_singlethread_workqueue("ib_addr");
7025fcd3 SH	544	if (!addr_wq)
	545	return -ENOMEM;
	546
e795d092	547	register_netevent_notifier(&nb);
dd5f03be	548	rdma_addr_register_client(&self);
7025fcd3 SH	549	return 0;
	550	}
	551
716abb1f	552	static void __exit addr_cleanup(void)
7025fcd3	553	{
dd5f03be	554	rdma_addr_unregister_client(&self);
e795d092	555	unregister_netevent_notifier(&nb);
7025fcd3 SH	556	destroy_workqueue(addr_wq);
	557	}
	558
	559	module_init(addr_init);
	560	module_exit(addr_cleanup);