[mirror_ovs.git] / datapath / vport-lisp.c

/*
 * Copyright (c) 2011 Nicira, Inc.
 * Copyright (c) 2013 Cisco Systems, Inc.
 *
 * This program is free software; you can redistribute it and/or
 * modify it under the terms of version 2 of the GNU General Public
 * License as published by the Free Software Foundation.
 *
 * This program is distributed in the hope that it will be useful, but
 * WITHOUT ANY WARRANTY; without even the implied warranty of
 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
 * General Public License for more details.
 *
 * You should have received a copy of the GNU General Public License
 * along with this program; if not, write to the Free Software
 * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA
 * 02110-1301, USA
 */

#define pr_fmt(fmt) KBUILD_MODNAME ": " fmt

#include <linux/version.h>

#include <linux/in.h>
#include <linux/ip.h>
#include <linux/net.h>
#include <linux/rculist.h>
#include <linux/udp.h>

#include <net/icmp.h>
#include <net/ip.h>
#include <net/route.h>
#include <net/udp.h>
#include <net/udp_tunnel.h>
#include <net/xfrm.h>

#include "datapath.h"
#include "gso.h"
#include "vport.h"

/*
 *  LISP encapsulation header:
 *
 *  +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
 *  |N|L|E|V|I|flags|            Nonce/Map-Version                  |
 *  +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
 *  |                 Instance ID/Locator Status Bits               |
 *  +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
 *
 */

/**
 * struct lisphdr - LISP header
 * @nonce_present: Flag indicating the presence of a 24 bit nonce value.
 * @locator_status_bits_present: Flag indicating the presence of Locator Status
 *                               Bits (LSB).
 * @solicit_echo_nonce: Flag indicating the use of the echo noncing mechanism.
 * @map_version_present: Flag indicating the use of mapping versioning.
 * @instance_id_present: Flag indicating the presence of a 24 bit Instance ID.
 * @reserved_flags: 3 bits reserved for future flags.
 * @nonce: 24 bit nonce value.
 * @map_version: 24 bit mapping version.
 * @locator_status_bits: Locator Status Bits: 32 bits when instance_id_present
 *                       is not set, 8 bits when it is.
 * @instance_id: 24 bit Instance ID
 */
struct lisphdr {
#ifdef __LITTLE_ENDIAN_BITFIELD
	__u8 reserved_flags:3;
	__u8 instance_id_present:1;
	__u8 map_version_present:1;
	__u8 solicit_echo_nonce:1;
	__u8 locator_status_bits_present:1;
	__u8 nonce_present:1;
#else
	__u8 nonce_present:1;
	__u8 locator_status_bits_present:1;
	__u8 solicit_echo_nonce:1;
	__u8 map_version_present:1;
	__u8 instance_id_present:1;
	__u8 reserved_flags:3;
#endif
	union {
		__u8 nonce[3];
		__u8 map_version[3];
	} u1;
	union {
		__be32 locator_status_bits;
		struct {
			__u8 instance_id[3];
			__u8 locator_status_bits;
		} word2;
	} u2;
};

#define LISP_HLEN (sizeof(struct udphdr) + sizeof(struct lisphdr))

/**
 * struct lisp_port - Keeps track of open UDP ports
 * @dst_port: lisp UDP port no.
 * @list: list element in @lisp_ports.
 * @lisp_rcv_socket: The socket created for this port number.
 * @name: vport name.
 */
struct lisp_port {
	__be16 dst_port;
	struct list_head list;
	struct socket *lisp_rcv_socket;
	char name[IFNAMSIZ];
};

static LIST_HEAD(lisp_ports);

static inline struct lisp_port *lisp_vport(const struct vport *vport)
{
	return vport_priv(vport);
}

static struct lisp_port *lisp_find_port(struct net *net, __be16 port)
{
	struct lisp_port *lisp_port;

	list_for_each_entry_rcu(lisp_port, &lisp_ports, list) {
		if (lisp_port->dst_port == port &&
			net_eq(sock_net(lisp_port->lisp_rcv_socket->sk), net))
			return lisp_port;
	}

	return NULL;
}

static inline struct lisphdr *lisp_hdr(const struct sk_buff *skb)
{
	return (struct lisphdr *)(udp_hdr(skb) + 1);
}

/* Convert 64 bit tunnel ID to 24 bit Instance ID. */
static void tunnel_id_to_instance_id(__be64 tun_id, __u8 *iid)
{

#ifdef __BIG_ENDIAN
	iid[0] = (__force __u8)(tun_id >> 16);
	iid[1] = (__force __u8)(tun_id >> 8);
	iid[2] = (__force __u8)tun_id;
#else
	iid[0] = (__force __u8)((__force u64)tun_id >> 40);
	iid[1] = (__force __u8)((__force u64)tun_id >> 48);
	iid[2] = (__force __u8)((__force u64)tun_id >> 56);
#endif
}

/* Convert 24 bit Instance ID to 64 bit tunnel ID. */
static __be64 instance_id_to_tunnel_id(__u8 *iid)
{
#ifdef __BIG_ENDIAN
	return (iid[0] << 16) | (iid[1] << 8) | iid[2];
#else
	return (__force __be64)(((__force u64)iid[0] << 40) |
				((__force u64)iid[1] << 48) |
				((__force u64)iid[2] << 56));
#endif
}

/* Compute source UDP port for outgoing packet.
 * Currently we use the flow hash.
 */
static u16 get_src_port(struct net *net, struct sk_buff *skb)
{
	u32 hash = skb_get_hash(skb);
	unsigned int range;
	int high;
	int low;

	if (!hash) {
		if (skb->protocol == htons(ETH_P_IP)) {
			struct iphdr *iph;
			int size = (sizeof(iph->saddr) * 2) / sizeof(u32);

			iph = (struct iphdr *) skb_network_header(skb);
			hash = jhash2((const u32 *)&iph->saddr, size, 0);
		} else if (skb->protocol == htons(ETH_P_IPV6)) {
			struct ipv6hdr *ipv6hdr;

			ipv6hdr = (struct ipv6hdr *) skb_network_header(skb);
			hash = jhash2((const u32 *)&ipv6hdr->saddr,
				      (sizeof(struct in6_addr) * 2) / sizeof(u32), 0);
		} else {
			pr_warn_once("LISP inner protocol is not IP when "
				     "calculating hash.\n");
		}
	}

	inet_get_local_port_range(net, &low, &high);
	range = (high - low) + 1;
	return (((u64) hash * range) >> 32) + low;
}

static void lisp_build_header(struct sk_buff *skb)
{
	struct lisphdr *lisph;
	const struct ovs_key_ipv4_tunnel *tun_key;

	tun_key = &OVS_CB(skb)->egress_tun_info->tunnel;

	lisph = (struct lisphdr *)__skb_push(skb, sizeof(struct lisphdr));
	lisph->nonce_present = 0;	/* We don't support echo nonce algorithm */
	lisph->locator_status_bits_present = 1;	/* Set LSB */
	lisph->solicit_echo_nonce = 0;	/* No echo noncing */
	lisph->map_version_present = 0;	/* No mapping versioning, nonce instead */
	lisph->instance_id_present = 1;	/* Store the tun_id as Instance ID  */
	lisph->reserved_flags = 0;	/* Reserved flags, set to 0  */

	lisph->u1.nonce[0] = 0;
	lisph->u1.nonce[1] = 0;
	lisph->u1.nonce[2] = 0;

	tunnel_id_to_instance_id(tun_key->tun_id, &lisph->u2.word2.instance_id[0]);
	lisph->u2.word2.locator_status_bits = 1;
}

/* Called with rcu_read_lock and BH disabled. */
static int lisp_rcv(struct sock *sk, struct sk_buff *skb)
{
	struct lisp_port *lisp_port;
	struct lisphdr *lisph;
	struct iphdr *iph, *inner_iph;
	struct ovs_tunnel_info tun_info;
	__be64 key;
	struct ethhdr *ethh;
	__be16 protocol;

	lisp_port = rcu_dereference_sk_user_data(sk);
	if (unlikely(!lisp_port))
		goto error;

	if (iptunnel_pull_header(skb, LISP_HLEN, 0))
		goto error;

	lisph = lisp_hdr(skb);

	if (lisph->instance_id_present != 1)
		key = 0;
	else
		key = instance_id_to_tunnel_id(&lisph->u2.word2.instance_id[0]);

	/* Save outer tunnel values */
	iph = ip_hdr(skb);
	ovs_flow_tun_info_init(&tun_info, iph,
			       udp_hdr(skb)->source, udp_hdr(skb)->dest,
			       key, TUNNEL_KEY, NULL, 0);

	/* Drop non-IP inner packets */
	inner_iph = (struct iphdr *)(lisph + 1);
	switch (inner_iph->version) {
	case 4:
		protocol = htons(ETH_P_IP);
		break;
	case 6:
		protocol = htons(ETH_P_IPV6);
		break;
	default:
		goto error;
	}
	skb->protocol = protocol;

	/* Add Ethernet header */
	ethh = (struct ethhdr *)skb_push(skb, ETH_HLEN);
	memset(ethh, 0, ETH_HLEN);
	ethh->h_dest[0] = 0x02;
	ethh->h_source[0] = 0x02;
	ethh->h_proto = protocol;

	ovs_skb_postpush_rcsum(skb, skb->data, ETH_HLEN);

	ovs_vport_receive(vport_from_priv(lisp_port), skb, &tun_info);
	goto out;

error:
	kfree_skb(skb);
out:
	return 0;
}

static int lisp_socket_init(struct lisp_port *lisp_port, struct net *net)
{
	struct udp_port_cfg udp_conf;
	struct udp_tunnel_sock_cfg tunnel_cfg;
	int err;

	memset(&udp_conf, 0, sizeof(udp_conf));

	udp_conf.family = AF_INET;
	udp_conf.local_ip.s_addr = htonl(INADDR_ANY);
	udp_conf.local_udp_port = lisp_port->dst_port;

        err = udp_sock_create(net, &udp_conf, &lisp_port->lisp_rcv_socket);
        if (err < 0) {
		pr_warn("cannot register lisp protocol handler: %d\n", err);
                return err;
	}

	tunnel_cfg.sk_user_data = lisp_port;
	tunnel_cfg.encap_type = 1;
	tunnel_cfg.encap_rcv = lisp_rcv;
	tunnel_cfg.encap_destroy = NULL;

	setup_udp_tunnel_sock(net, lisp_port->lisp_rcv_socket, &tunnel_cfg);

	return 0;
}

static int lisp_get_options(const struct vport *vport, struct sk_buff *skb)
{
	struct lisp_port *lisp_port = lisp_vport(vport);

	if (nla_put_u16(skb, OVS_TUNNEL_ATTR_DST_PORT, ntohs(lisp_port->dst_port)))
		return -EMSGSIZE;
	return 0;
}

static void lisp_tnl_destroy(struct vport *vport)
{
	struct lisp_port *lisp_port = lisp_vport(vport);

	list_del_rcu(&lisp_port->list);
	udp_tunnel_sock_release(lisp_port->lisp_rcv_socket);
	ovs_vport_deferred_free(vport);
}

static struct vport *lisp_tnl_create(const struct vport_parms *parms)
{
	struct net *net = ovs_dp_get_net(parms->dp);
	struct nlattr *options = parms->options;
	struct lisp_port *lisp_port;
	struct vport *vport;
	struct nlattr *a;
	int err;
	u16 dst_port;

	if (!options) {
		err = -EINVAL;
		goto error;
	}

	a = nla_find_nested(options, OVS_TUNNEL_ATTR_DST_PORT);
	if (a && nla_len(a) == sizeof(u16)) {
		dst_port = nla_get_u16(a);
	} else {
		/* Require destination port from userspace. */
		err = -EINVAL;
		goto error;
	}

	/* Verify if we already have a socket created for this port */
	if (lisp_find_port(net, htons(dst_port))) {
		err = -EEXIST;
		goto error;
	}

	vport = ovs_vport_alloc(sizeof(struct lisp_port),
				&ovs_lisp_vport_ops, parms);
	if (IS_ERR(vport))
		return vport;

	lisp_port = lisp_vport(vport);
	lisp_port->dst_port = htons(dst_port);
	strncpy(lisp_port->name, parms->name, IFNAMSIZ);

	err = lisp_socket_init(lisp_port, net);
	if (err)
		goto error_free;

	list_add_tail_rcu(&lisp_port->list, &lisp_ports);
	return vport;

error_free:
	ovs_vport_free(vport);
error:
	return ERR_PTR(err);
}

static int lisp_send(struct vport *vport, struct sk_buff *skb)
{
	struct ovs_key_ipv4_tunnel *tun_key;
	struct lisp_port *lisp_port = lisp_vport(vport);
	struct net *net = ovs_dp_get_net(vport->dp);
	int network_offset = skb_network_offset(skb);
	struct rtable *rt;
	int min_headroom;
	__be32 saddr;
	__be16 src_port, dst_port;
	__be16 df;
	int sent_len;
	int err;

	if (unlikely(!OVS_CB(skb)->egress_tun_info)) {
		err = -EINVAL;
		goto error;
	}

	tun_key = &OVS_CB(skb)->egress_tun_info->tunnel;

	if (skb->protocol != htons(ETH_P_IP) &&
	    skb->protocol != htons(ETH_P_IPV6)) {
		err = 0;
		goto error;
	}

	/* Route lookup */
	saddr = tun_key->ipv4_src;
	rt = find_route(ovs_dp_get_net(vport->dp),
			&saddr, tun_key->ipv4_dst,
			IPPROTO_UDP, tun_key->ipv4_tos,
			skb->mark);
	if (IS_ERR(rt)) {
		err = PTR_ERR(rt);
		goto error;
	}

	min_headroom = LL_RESERVED_SPACE(rt_dst(rt).dev) + rt_dst(rt).header_len
			+ sizeof(struct iphdr) + LISP_HLEN;

	if (skb_headroom(skb) < min_headroom || skb_header_cloned(skb)) {
		int head_delta = SKB_DATA_ALIGN(min_headroom -
						skb_headroom(skb) +
						16);

		err = pskb_expand_head(skb, max_t(int, head_delta, 0),
					0, GFP_ATOMIC);
		if (unlikely(err))
			goto err_free_rt;
	}

	/* Reset l2 headers. */
	skb_pull(skb, network_offset);
	skb_reset_mac_header(skb);
	vlan_set_tci(skb, 0);

	skb = udp_tunnel_handle_offloads(skb, false, false);
	if (IS_ERR(skb)) {
		err = PTR_ERR(skb);
		skb = NULL;
		goto err_free_rt;
	}

	src_port = htons(get_src_port(net, skb));
	dst_port = lisp_port->dst_port;

	lisp_build_header(skb);

	skb->ignore_df = 1;

	ovs_skb_set_inner_protocol(skb, skb->protocol);

	df = tun_key->tun_flags & TUNNEL_DONT_FRAGMENT ? htons(IP_DF) : 0;
	sent_len = udp_tunnel_xmit_skb(rt, skb, saddr, tun_key->ipv4_dst,
				       tun_key->ipv4_tos, tun_key->ipv4_ttl,
				       df, src_port, dst_port, false, true);

	return sent_len > 0 ? sent_len + network_offset : sent_len;

err_free_rt:
	ip_rt_put(rt);
error:
	kfree_skb(skb);
	return err;
}

static const char *lisp_get_name(const struct vport *vport)
{
	struct lisp_port *lisp_port = lisp_vport(vport);
	return lisp_port->name;
}

static int lisp_get_egress_tun_info(struct vport *vport, struct sk_buff *skb,
				    struct ovs_tunnel_info *egress_tun_info)
{
	struct net *net = ovs_dp_get_net(vport->dp);
	struct lisp_port *lisp_port = lisp_vport(vport);

	if (skb->protocol != htons(ETH_P_IP) &&
	    skb->protocol != htons(ETH_P_IPV6)) {
		return -EINVAL;
	}

	/*
	 * Get tp_src and tp_dst, refert to lisp_build_header().
	 */
	return ovs_tunnel_get_egress_info(egress_tun_info, net,
					  OVS_CB(skb)->egress_tun_info,
					  IPPROTO_UDP, skb->mark,
					  htons(get_src_port(net, skb)),
					  lisp_port->dst_port);
}

const struct vport_ops ovs_lisp_vport_ops = {
	.type			= OVS_VPORT_TYPE_LISP,
	.create			= lisp_tnl_create,
	.destroy		= lisp_tnl_destroy,
	.get_name		= lisp_get_name,
	.get_options		= lisp_get_options,
	.send			= lisp_send,
	.get_egress_tun_info	= lisp_get_egress_tun_info,
};
Commit	Line	Data
a6ae068b LJ	1	/*
	2	* Copyright (c) 2011 Nicira, Inc.
	3	* Copyright (c) 2013 Cisco Systems, Inc.
	4	*
	5	* This program is free software; you can redistribute it and/or
	6	* modify it under the terms of version 2 of the GNU General Public
	7	* License as published by the Free Software Foundation.
	8	*
	9	* This program is distributed in the hope that it will be useful, but
	10	* WITHOUT ANY WARRANTY; without even the implied warranty of
	11	* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
	12	* General Public License for more details.
	13	*
	14	* You should have received a copy of the GNU General Public License
	15	* along with this program; if not, write to the Free Software
	16	* Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA
	17	* 02110-1301, USA
	18	*/
	19
	20	#define pr_fmt(fmt) KBUILD_MODNAME ": " fmt
	21
	22	#include <linux/version.h>
a6ae068b LJ	23
	24	#include <linux/in.h>
	25	#include <linux/ip.h>
a6ae068b	26	#include <linux/net.h>
85c9de19	27	#include <linux/rculist.h>
a6ae068b LJ	28	#include <linux/udp.h>
	29
	30	#include <net/icmp.h>
	31	#include <net/ip.h>
11aa8dff	32	#include <net/route.h>
a6ae068b	33	#include <net/udp.h>
3dd9bffd	34	#include <net/udp_tunnel.h>
11aa8dff	35	#include <net/xfrm.h>
a6ae068b LJ	36
a6ae068b LJ	37	#include "datapath.h"
3cf54a56	38	#include "gso.h"
a6ae068b LJ	39	#include "vport.h"
a6ae068b LJ	40
a6ae068b LJ	41	/*
	42	* LISP encapsulation header:
	43	*
	44	* +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
	45	* \|N\|L\|E\|V\|I\|flags\| Nonce/Map-Version \|
	46	* +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
	47	* \| Instance ID/Locator Status Bits \|
	48	* +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
	49	*
	50	*/
	51
	52	/**
	53	* struct lisphdr - LISP header
	54	* @nonce_present: Flag indicating the presence of a 24 bit nonce value.
	55	* @locator_status_bits_present: Flag indicating the presence of Locator Status
	56	* Bits (LSB).
	57	* @solicit_echo_nonce: Flag indicating the use of the echo noncing mechanism.
	58	* @map_version_present: Flag indicating the use of mapping versioning.
	59	* @instance_id_present: Flag indicating the presence of a 24 bit Instance ID.
	60	* @reserved_flags: 3 bits reserved for future flags.
	61	* @nonce: 24 bit nonce value.
	62	* @map_version: 24 bit mapping version.
	63	* @locator_status_bits: Locator Status Bits: 32 bits when instance_id_present
	64	* is not set, 8 bits when it is.
	65	* @instance_id: 24 bit Instance ID
	66	*/
	67	struct lisphdr {
	68	#ifdef __LITTLE_ENDIAN_BITFIELD
	69	__u8 reserved_flags:3;
	70	__u8 instance_id_present:1;
	71	__u8 map_version_present:1;
	72	__u8 solicit_echo_nonce:1;
	73	__u8 locator_status_bits_present:1;
	74	__u8 nonce_present:1;
	75	#else
	76	__u8 nonce_present:1;
	77	__u8 locator_status_bits_present:1;
	78	__u8 solicit_echo_nonce:1;
	79	__u8 map_version_present:1;
	80	__u8 instance_id_present:1;
	81	__u8 reserved_flags:3;
	82	#endif
	83	union {
	84	__u8 nonce[3];
	85	__u8 map_version[3];
	86	} u1;
	87	union {
	88	__be32 locator_status_bits;
	89	struct {
	90	__u8 instance_id[3];
	91	__u8 locator_status_bits;
	92	} word2;
	93	} u2;
	94	};
	95
	96	#define LISP_HLEN (sizeof(struct udphdr) + sizeof(struct lisphdr))
	97
a6ae068b LJ	98	/**
a6ae068b LJ	99	* struct lisp_port - Keeps track of open UDP ports
c405d282 PS	100	* @dst_port: lisp UDP port no.
	101	* @list: list element in @lisp_ports.
	102	* @lisp_rcv_socket: The socket created for this port number.
	103	* @name: vport name.
a6ae068b LJ	104	*/
a6ae068b LJ	105	struct lisp_port {
c405d282	106	__be16 dst_port;
a6ae068b	107	struct list_head list;
a6ae068b	108	struct socket *lisp_rcv_socket;
c405d282	109	char name[IFNAMSIZ];
a6ae068b LJ	110	};
	111
	112	static LIST_HEAD(lisp_ports);
	113
c405d282 PS	114	static inline struct lisp_port lisp_vport(const struct vport vport)
	115	{
	116	return vport_priv(vport);
	117	}
	118
85c9de19	119	static struct lisp_port lisp_find_port(struct net net, __be16 port)
a6ae068b LJ	120	{
	121	struct lisp_port *lisp_port;
	122
85c9de19	123	list_for_each_entry_rcu(lisp_port, &lisp_ports, list) {
c405d282	124	if (lisp_port->dst_port == port &&
a6ae068b LJ	125	net_eq(sock_net(lisp_port->lisp_rcv_socket->sk), net))
	126	return lisp_port;
	127	}
	128
	129	return NULL;
	130	}
	131
	132	static inline struct lisphdr lisp_hdr(const struct sk_buff skb)
	133	{
	134	return (struct lisphdr *)(udp_hdr(skb) + 1);
	135	}
	136
a6ae068b LJ	137	/* Convert 64 bit tunnel ID to 24 bit Instance ID. */
	138	static void tunnel_id_to_instance_id(__be64 tun_id, __u8 *iid)
	139	{
	140
	141	#ifdef __BIG_ENDIAN
	142	iid[0] = (__force __u8)(tun_id >> 16);
	143	iid[1] = (__force __u8)(tun_id >> 8);
	144	iid[2] = (__force __u8)tun_id;
	145	#else
	146	iid[0] = (__force __u8)((__force u64)tun_id >> 40);
	147	iid[1] = (__force __u8)((__force u64)tun_id >> 48);
	148	iid[2] = (__force __u8)((__force u64)tun_id >> 56);
	149	#endif
	150	}
	151
	152	/* Convert 24 bit Instance ID to 64 bit tunnel ID. */
	153	static __be64 instance_id_to_tunnel_id(__u8 *iid)
	154	{
	155	#ifdef __BIG_ENDIAN
	156	return (iid[0] << 16) \| (iid[1] << 8) \| iid[2];
	157	#else
	158	return (__force __be64)(((__force u64)iid[0] << 40) \|
	159	((__force u64)iid[1] << 48) \|
	160	((__force u64)iid[2] << 56));
	161	#endif
	162	}
	163
11aa8dff PS	164	/* Compute source UDP port for outgoing packet.
	165	* Currently we use the flow hash.
	166	*/
cb25142c	167	static u16 get_src_port(struct net net, struct sk_buff skb)
11aa8dff	168	{
e2f3178f	169	u32 hash = skb_get_hash(skb);
11aa8dff	170	unsigned int range;
5c89b171 PS	171	int high;
	172	int low;
	173
	174	if (!hash) {
e74d4817 PS	175	if (skb->protocol == htons(ETH_P_IP)) {
	176	struct iphdr *iph;
	177	int size = (sizeof(iph->saddr) * 2) / sizeof(u32);
	178
22c18a56	179	iph = (struct iphdr *) skb_network_header(skb);
e74d4817 PS	180	hash = jhash2((const u32 *)&iph->saddr, size, 0);
	181	} else if (skb->protocol == htons(ETH_P_IPV6)) {
	182	struct ipv6hdr *ipv6hdr;
	183
22c18a56	184	ipv6hdr = (struct ipv6hdr *) skb_network_header(skb);
e74d4817 PS	185	hash = jhash2((const u32 *)&ipv6hdr->saddr,
	186	(sizeof(struct in6_addr) * 2) / sizeof(u32), 0);
	187	} else {
31125ebd JG	188	pr_warn_once("LISP inner protocol is not IP when "
31125ebd JG	189	"calculating hash.\n");
e74d4817	190	}
5c89b171	191	}
11aa8dff	192
cb25142c	193	inet_get_local_port_range(net, &low, &high);
11aa8dff PS	194	range = (high - low) + 1;
	195	return (((u64) hash * range) >> 32) + low;
	196	}
	197
22c18a56	198	static void lisp_build_header(struct sk_buff *skb)
a6ae068b	199	{
22c18a56	200	struct lisphdr *lisph;
fb66fbd1	201	const struct ovs_key_ipv4_tunnel *tun_key;
a6ae068b	202
fb66fbd1	203	tun_key = &OVS_CB(skb)->egress_tun_info->tunnel;
a6ae068b	204
22c18a56	205	lisph = (struct lisphdr *)__skb_push(skb, sizeof(struct lisphdr));
a6ae068b LJ	206	lisph->nonce_present = 0; /* We don't support echo nonce algorithm */
	207	lisph->locator_status_bits_present = 1; /* Set LSB */
	208	lisph->solicit_echo_nonce = 0; /* No echo noncing */
	209	lisph->map_version_present = 0; /* No mapping versioning, nonce instead */
	210	lisph->instance_id_present = 1; /* Store the tun_id as Instance ID */
	211	lisph->reserved_flags = 0; /* Reserved flags, set to 0 */
	212
	213	lisph->u1.nonce[0] = 0;
	214	lisph->u1.nonce[1] = 0;
	215	lisph->u1.nonce[2] = 0;
	216
85c9de19	217	tunnel_id_to_instance_id(tun_key->tun_id, &lisph->u2.word2.instance_id[0]);
a6ae068b	218	lisph->u2.word2.locator_status_bits = 1;
a6ae068b LJ	219	}
	220
	221	/* Called with rcu_read_lock and BH disabled. */
	222	static int lisp_rcv(struct sock sk, struct sk_buff skb)
	223	{
85c9de19	224	struct lisp_port *lisp_port;
a6ae068b	225	struct lisphdr *lisph;
a6ae068b	226	struct iphdr iph, inner_iph;
f0cd669f	227	struct ovs_tunnel_info tun_info;
a6ae068b	228	__be64 key;
a6ae068b LJ	229	struct ethhdr *ethh;
	230	__be16 protocol;
	231
22c18a56	232	lisp_port = rcu_dereference_sk_user_data(sk);
85c9de19 PS	233	if (unlikely(!lisp_port))
	234	goto error;
	235
971fd4ab	236	if (iptunnel_pull_header(skb, LISP_HLEN, 0))
a6ae068b LJ	237	goto error;
	238
	239	lisph = lisp_hdr(skb);
	240
a6ae068b LJ	241	if (lisph->instance_id_present != 1)
	242	key = 0;
	243	else
	244	key = instance_id_to_tunnel_id(&lisph->u2.word2.instance_id[0]);
	245
a6ae068b	246	/* Save outer tunnel values */
85c9de19	247	iph = ip_hdr(skb);
8b7ea2d4 WZ	248	ovs_flow_tun_info_init(&tun_info, iph,
	249	udp_hdr(skb)->source, udp_hdr(skb)->dest,
	250	key, TUNNEL_KEY, NULL, 0);
a6ae068b LJ	251
	252	/* Drop non-IP inner packets */
	253	inner_iph = (struct iphdr *)(lisph + 1);
	254	switch (inner_iph->version) {
	255	case 4:
	256	protocol = htons(ETH_P_IP);
	257	break;
	258	case 6:
	259	protocol = htons(ETH_P_IPV6);
	260	break;
	261	default:
	262	goto error;
	263	}
971fd4ab	264	skb->protocol = protocol;
a6ae068b LJ	265
	266	/* Add Ethernet header */
	267	ethh = (struct ethhdr *)skb_push(skb, ETH_HLEN);
	268	memset(ethh, 0, ETH_HLEN);
	269	ethh->h_dest[0] = 0x02;
	270	ethh->h_source[0] = 0x02;
	271	ethh->h_proto = protocol;
	272
3cfede14 PS	273	ovs_skb_postpush_rcsum(skb, skb->data, ETH_HLEN);
3cfede14 PS	274
f0cd669f	275	ovs_vport_receive(vport_from_priv(lisp_port), skb, &tun_info);
a6ae068b LJ	276	goto out;
	277
	278	error:
	279	kfree_skb(skb);
	280	out:
	281	return 0;
	282	}
	283
a6ae068b LJ	284	static int lisp_socket_init(struct lisp_port lisp_port, struct net net)
a6ae068b LJ	285	{
3dd9bffd	286	struct udp_port_cfg udp_conf;
22c18a56	287	struct udp_tunnel_sock_cfg tunnel_cfg;
c405d282	288	int err;
a6ae068b	289
3dd9bffd	290	memset(&udp_conf, 0, sizeof(udp_conf));
a6ae068b	291
3dd9bffd JG	292	udp_conf.family = AF_INET;
	293	udp_conf.local_ip.s_addr = htonl(INADDR_ANY);
	294	udp_conf.local_udp_port = lisp_port->dst_port;
a6ae068b	295
3dd9bffd JG	296	err = udp_sock_create(net, &udp_conf, &lisp_port->lisp_rcv_socket);
	297	if (err < 0) {
	298	pr_warn("cannot register lisp protocol handler: %d\n", err);
	299	return err;
	300	}
a6ae068b	301
22c18a56 JG	302	tunnel_cfg.sk_user_data = lisp_port;
	303	tunnel_cfg.encap_type = 1;
	304	tunnel_cfg.encap_rcv = lisp_rcv;
	305	tunnel_cfg.encap_destroy = NULL;
a6ae068b	306
22c18a56	307	setup_udp_tunnel_sock(net, lisp_port->lisp_rcv_socket, &tunnel_cfg);
a6ae068b LJ	308
a6ae068b LJ	309	return 0;
a6ae068b LJ	310	}
a6ae068b LJ	311
c405d282	312	static int lisp_get_options(const struct vport vport, struct sk_buff skb)
a6ae068b	313	{
c405d282	314	struct lisp_port *lisp_port = lisp_vport(vport);
a6ae068b	315
c405d282 PS	316	if (nla_put_u16(skb, OVS_TUNNEL_ATTR_DST_PORT, ntohs(lisp_port->dst_port)))
	317	return -EMSGSIZE;
	318	return 0;
a6ae068b LJ	319	}
a6ae068b LJ	320
c405d282	321	static void lisp_tnl_destroy(struct vport *vport)
a6ae068b	322	{
c405d282 PS	323	struct lisp_port *lisp_port = lisp_vport(vport);
c405d282 PS	324
85c9de19	325	list_del_rcu(&lisp_port->list);
22c18a56	326	udp_tunnel_sock_release(lisp_port->lisp_rcv_socket);
c405d282	327	ovs_vport_deferred_free(vport);
85c9de19 PS	328	}
85c9de19 PS	329
c405d282	330	static struct vport lisp_tnl_create(const struct vport_parms parms)
85c9de19	331	{
c405d282 PS	332	struct net *net = ovs_dp_get_net(parms->dp);
c405d282 PS	333	struct nlattr *options = parms->options;
85c9de19	334	struct lisp_port *lisp_port;
c405d282	335	struct vport *vport;
a6ae068b LJ	336	struct nlattr *a;
	337	int err;
	338	u16 dst_port;
a6ae068b LJ	339
	340	if (!options) {
	341	err = -EINVAL;
c405d282	342	goto error;
a6ae068b LJ	343	}
	344
	345	a = nla_find_nested(options, OVS_TUNNEL_ATTR_DST_PORT);
	346	if (a && nla_len(a) == sizeof(u16)) {
	347	dst_port = nla_get_u16(a);
	348	} else {
	349	/* Require destination port from userspace. */
	350	err = -EINVAL;
c405d282	351	goto error;
a6ae068b LJ	352	}
	353
	354	/* Verify if we already have a socket created for this port */
c405d282	355	if (lisp_find_port(net, htons(dst_port))) {
85c9de19	356	err = -EEXIST;
c405d282	357	goto error;
a6ae068b LJ	358	}
a6ae068b LJ	359
c405d282 PS	360	vport = ovs_vport_alloc(sizeof(struct lisp_port),
	361	&ovs_lisp_vport_ops, parms);
	362	if (IS_ERR(vport))
	363	return vport;
a6ae068b	364
c405d282 PS	365	lisp_port = lisp_vport(vport);
	366	lisp_port->dst_port = htons(dst_port);
	367	strncpy(lisp_port->name, parms->name, IFNAMSIZ);
a6ae068b LJ	368
	369	err = lisp_socket_init(lisp_port, net);
	370	if (err)
c405d282	371	goto error_free;
a6ae068b	372
c405d282 PS	373	list_add_tail_rcu(&lisp_port->list, &lisp_ports);
c405d282 PS	374	return vport;
a6ae068b	375
c405d282 PS	376	error_free:
c405d282 PS	377	ovs_vport_free(vport);
a6ae068b	378	error:
c405d282	379	return ERR_PTR(err);
a6ae068b LJ	380	}
a6ae068b LJ	381
3cf54a56	382	static int lisp_send(struct vport vport, struct sk_buff skb)
11aa8dff	383	{
27914610	384	struct ovs_key_ipv4_tunnel *tun_key;
22c18a56 JG	385	struct lisp_port *lisp_port = lisp_vport(vport);
22c18a56 JG	386	struct net *net = ovs_dp_get_net(vport->dp);
3cf54a56	387	int network_offset = skb_network_offset(skb);
11aa8dff	388	struct rtable *rt;
3cf54a56	389	int min_headroom;
11aa8dff	390	__be32 saddr;
22c18a56	391	__be16 src_port, dst_port;
3cf54a56 PS	392	__be16 df;
3cf54a56 PS	393	int sent_len;
11aa8dff	394	int err;
3cf54a56	395
93258bd7 PS	396	if (unlikely(!OVS_CB(skb)->egress_tun_info)) {
	397	err = -EINVAL;
	398	goto error;
	399	}
3cf54a56	400
fb66fbd1	401	tun_key = &OVS_CB(skb)->egress_tun_info->tunnel;
27914610	402
3cf54a56 PS	403	if (skb->protocol != htons(ETH_P_IP) &&
3cf54a56 PS	404	skb->protocol != htons(ETH_P_IPV6)) {
93258bd7 PS	405	err = 0;
93258bd7 PS	406	goto error;
3cf54a56	407	}
11aa8dff PS	408
11aa8dff PS	409	/* Route lookup */
f0cd669f	410	saddr = tun_key->ipv4_src;
11aa8dff	411	rt = find_route(ovs_dp_get_net(vport->dp),
f0cd669f JG	412	&saddr, tun_key->ipv4_dst,
f0cd669f JG	413	IPPROTO_UDP, tun_key->ipv4_tos,
3025a772	414	skb->mark);
11aa8dff PS	415	if (IS_ERR(rt)) {
	416	err = PTR_ERR(rt);
	417	goto error;
	418	}
	419
11aa8dff	420	min_headroom = LL_RESERVED_SPACE(rt_dst(rt).dev) + rt_dst(rt).header_len
3cf54a56	421	+ sizeof(struct iphdr) + LISP_HLEN;
11aa8dff PS	422
	423	if (skb_headroom(skb) < min_headroom \|\| skb_header_cloned(skb)) {
	424	int head_delta = SKB_DATA_ALIGN(min_headroom -
	425	skb_headroom(skb) +
	426	16);
	427
	428	err = pskb_expand_head(skb, max_t(int, head_delta, 0),
	429	0, GFP_ATOMIC);
	430	if (unlikely(err))
	431	goto err_free_rt;
	432	}
	433
df7d2c59	434	/* Reset l2 headers. */
3cf54a56 PS	435	skb_pull(skb, network_offset);
3cf54a56 PS	436	skb_reset_mac_header(skb);
df7d2c59 PS	437	vlan_set_tci(skb, 0);
df7d2c59 PS	438
22c18a56	439	skb = udp_tunnel_handle_offloads(skb, false, false);
b1b3fb13	440	if (IS_ERR(skb)) {
93258bd7 PS	441	err = PTR_ERR(skb);
93258bd7 PS	442	skb = NULL;
29c71cfa	443	goto err_free_rt;
b1b3fb13	444	}
29c71cfa	445
22c18a56 JG	446	src_port = htons(get_src_port(net, skb));
	447	dst_port = lisp_port->dst_port;
	448
	449	lisp_build_header(skb);
	450
f6a0c895	451	skb->ignore_df = 1;
3cf54a56	452
e4bafe59 JG	453	ovs_skb_set_inner_protocol(skb, skb->protocol);
e4bafe59 JG	454
f0cd669f	455	df = tun_key->tun_flags & TUNNEL_DONT_FRAGMENT ? htons(IP_DF) : 0;
7bdcee3e	456	sent_len = udp_tunnel_xmit_skb(rt, skb, saddr, tun_key->ipv4_dst,
22c18a56	457	tun_key->ipv4_tos, tun_key->ipv4_ttl,
7bdcee3e	458	df, src_port, dst_port, false, true);
11aa8dff	459
3cf54a56	460	return sent_len > 0 ? sent_len + network_offset : sent_len;
11aa8dff PS	461
	462	err_free_rt:
	463	ip_rt_put(rt);
	464	error:
93258bd7	465	kfree_skb(skb);
11aa8dff PS	466	return err;
	467	}
	468
c405d282	469	static const char lisp_get_name(const struct vport vport)
a6ae068b	470	{
c405d282 PS	471	struct lisp_port *lisp_port = lisp_vport(vport);
c405d282 PS	472	return lisp_port->name;
a6ae068b LJ	473	}
a6ae068b LJ	474
8b7ea2d4 WZ	475	static int lisp_get_egress_tun_info(struct vport vport, struct sk_buff skb,
	476	struct ovs_tunnel_info *egress_tun_info)
	477	{
	478	struct net *net = ovs_dp_get_net(vport->dp);
	479	struct lisp_port *lisp_port = lisp_vport(vport);
	480
	481	if (skb->protocol != htons(ETH_P_IP) &&
	482	skb->protocol != htons(ETH_P_IPV6)) {
	483	return -EINVAL;
	484	}
	485
	486	/*
	487	* Get tp_src and tp_dst, refert to lisp_build_header().
	488	*/
	489	return ovs_tunnel_get_egress_info(egress_tun_info, net,
	490	OVS_CB(skb)->egress_tun_info,
	491	IPPROTO_UDP, skb->mark,
	492	htons(get_src_port(net, skb)),
	493	lisp_port->dst_port);
	494	}
	495
a6ae068b	496	const struct vport_ops ovs_lisp_vport_ops = {
8b7ea2d4 WZ	497	.type = OVS_VPORT_TYPE_LISP,
	498	.create = lisp_tnl_create,
	499	.destroy = lisp_tnl_destroy,
	500	.get_name = lisp_get_name,
	501	.get_options = lisp_get_options,
	502	.send = lisp_send,
	503	.get_egress_tun_info = lisp_get_egress_tun_info,
a6ae068b	504	};