src/plugins/dns/dns_packet.h


1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156

/*
 * Copyright (c) 2015 Cisco and/or its affiliates.
 * Licensed under the Apache License, Version 2.0 (the "License");
 * you may not use this file except in compliance with the License.
 * You may obtain a copy of the License at:
 *
 *     http://www.apache.org/licenses/LICENSE-2.0
 *
 * Unless required by applicable law or agreed to in writing, software
 * distributed under the License is distributed on an "AS IS" BASIS,
 * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
 * See the License for the specific language governing permissions and
 * limitations under the License.
 */

#ifndef included_dns_packet_h
#define included_dns_packet_h

/**
 * DNS packet header format
 */

/* *INDENT-OFF* */
typedef CLIB_PACKED (struct {
  u16 id;                       /**< transaction ID */
  u16 flags;                    /**< flags  */
  u16 qdcount;                  /**< number of questions */
  u16 anscount;                 /**< number of answers */
  u16 nscount;                  /**< number of name servers */
  u16 arcount;                  /**< number of additional records */
}) dns_header_t;
/* *INDENT-ON* */

#define DNS_RCODE_MASK (0xf)
#define DNS_RCODE_NO_ERROR 0
#define DNS_RCODE_FORMAT_ERROR 1
#define DNS_RCODE_SERVER_FAILURE 2
#define DNS_RCODE_NAME_ERROR 3
#define DNS_RCODE_NOT_IMPLEMENTED 4
#define DNS_RCODE_REFUSED 5

#define DNS_RA (1<<7)		/**< recursion available */
#define DNS_RD (1<<8)		/**< recursion desired */
#define DNS_TC (1<<9)	       /**< truncation  */
#define DNS_AA (1<<10)		/**< authoritative answer  */
#define DNS_OPCODE_MASK (0xf<<11) /**< opcode mask */
#define DNS_OPCODE_QUERY (0<<11)  /**< standard query */
#define DNS_OPCODE_IQUERY (1<<11) /**< inverse query (deprecated) */
#define DNS_OPCODE_STATUS (2<<11) /**< server status  */
#define DNS_QR (1<<15)		/**< query=0, response=1  */


/*
 * Note: in DNS-land, www.foobar.com is encoded as three "labels,"
 * each of which amount to a 1 octet length followed by up to 63
 * octets of name. Don't forget to add a "null root label" after the last
 * real one, or the poor slob trying to parse the name will have
 * no chance whatsoever.
 *
 * All RRs have the same top level format shown below:
 *
 *                                    1  1  1  1  1  1
 *      0  1  2  3  4  5  6  7  8  9  0  1  2  3  4  5
 *    +--+--+--+--+--+--+--+--+--+--+--+--+--+--+--+--+
 *    |                                               |
 *    /                                               /
 *    /                      NAME                     /
 *    |                                               |
 *    +--+--+--+--+--+--+--+--+--+--+--+--+--+--+--+--+
 *    |                      TYPE                     |
 *    +--+--+--+--+--+--+--+--+--+--+--+--+--+--+--+--+
 *    |                     CLASS                     |
 *    +--+--+--+--+--+--+--+--+--+--+--+--+--+--+--+--+
 *    |                      TTL                      |
 *    |                                               |
 *    +--+--+--+--+--+--+--+--+--+--+--+--+--+--+--+--+
 *    |                   RDLENGTH                    |
 *    +--+--+--+--+--+--+--+--+--+--+--+--+--+--+--+--|
 *    /                     RDATA                     /
 *    /                                               /
 *    +--+--+--+--+--+--+--+--+--+--+--+--+--+--+--+--+
 *
 *
 *  DNS "questions" have the following format:
 *
 *                                     1  1  1  1  1  1
 *       0  1  2  3  4  5  6  7  8  9  0  1  2  3  4  5
 *     +--+--+--+--+--+--+--+--+--+--+--+--+--+--+--+--+
 *     |                                               |
 *     /                     QNAME                     /
 *     /                                               /
 *     +--+--+--+--+--+--+--+--+--+--+--+--+--+--+--+--+
 *     |                     QTYPE                     |
 *     +--+--+--+--+--+--+--+--+--+--+--+--+--+--+--+--+
 *     |                     QCLASS                    |
 *     +--+--+--+--+--+--+--+--+--+--+--+--+--+--+--+--+
 */

/**
 * DNS "question" fixed header.
 */
/* *INDENT-OFF* */
typedef CLIB_PACKED (struct {
  u16 type;	/**< record type requested */
  u16 class;	/**< class, 1 = internet */
}) dns_query_t;
/* *INDENT-ON* */

/**
 * DNS RR fixed header.
 */
/* *INDENT-OFF* */
typedef CLIB_PACKED (struct {
  u16 type;	/**< record type */
  u16 class;	/**< class, 1 = internet */
  u32 ttl;	/**< time to live, in seconds */
  u16 rdlength;
  /**< length of r */
  u8 rdata[0];
}) dns_rr_t;
/* *INDENT-ON* */

/*
 * There are quite a number of DNS record types
 * Feel free to add as needed
 */
#define foreach_dns_type                        \
_(A, 1) 	/**< ip4 host address */        \
_(AAAA, 28)     /**< ip6 host address */        \
_(ALL, 255)     /**< all available data */      \
_(TEXT, 16)     /**< a text string */           \
_(NAMESERVER, 2) /**< a nameserver */           \
_(CNAME, 5)      /**< a CNAME (alias) */	\
_(MAIL_EXCHANGE, 15) /**< a mail exchange  */	\
_(PTR, 12)      /**< a PTR (pointer) record */	\
_(HINFO, 13)	/**< Host info */

typedef enum
{
#define _(name,value) DNS_TYPE_##name = value,
  foreach_dns_type
#undef _
} dns_type_t;

#define DNS_CLASS_IN	1	/**< The Internet */


#endif /* included_dns_packet_h */

/*
 * fd.io coding-style-patch-verification: ON
 *
 * Local Variables:
 * eval: (c-set-style "gnu")
 * End:
 */
/*
 * ipsec_tun.h : IPSEC tunnel protection
 *
 * Copyright (c) 2015 Cisco and/or its affiliates.
 * Licensed under the Apache License, Version 2.0 (the "License");
 * you may not use this file except in compliance with the License.
 * You may obtain a copy of the License at:
 *
 *     http://www.apache.org/licenses/LICENSE-2.0
 *
 * Unless required by applicable law or agreed to in writing, software
 * distributed under the License is distributed on an "AS IS" BASIS,
 * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
 * See the License for the specific language governing permissions and
 * limitations under the License.
 */

#include <vnet/ipsec/ipsec_tun.h>
#include <vnet/ipsec/ipsec_itf.h>
#include <vnet/ipsec/esp.h>
#include <vnet/udp/udp_local.h>
#include <vnet/adj/adj_delegate.h>
#include <vnet/adj/adj_midchain.h>
#include <vnet/teib/teib.h>

/* instantiate the bihash functions */
#include <vppinfra/bihash_8_16.h>
#include <vppinfra/bihash_template.c>
#include <vppinfra/bihash_24_16.h>
#include <vppinfra/bihash_template.c>

#define IPSEC_TUN_DEFAULT_HASH_NUM_BUCKETS (64 * 1024)
#define IPSEC_TUN_DEFAULT_HASH_MEMORY_SIZE 512 << 20

/**
 * The logger
 */
vlib_log_class_t ipsec_tun_protect_logger;

/**
 * Pool of tunnel protection objects
 */
ipsec_tun_protect_t *ipsec_tun_protect_pool;

/**
 * Adj delegate registered type
 */
static adj_delegate_type_t ipsec_tun_adj_delegate_type;

/**
 * Adj index to TX SA mapping
 */
index_t *ipsec_tun_protect_sa_by_adj_index;

const ip_address_t IP_ADDR_ALL_0 = IP_ADDRESS_V4_ALL_0S;

/**
 * The DB of all added per-nh tunnel protectiond
 */
typedef struct ipsec_tun_protect_itf_db_t_
{
  /** A hash table key'd on IP (4 or 6) address */
  uword *id_hash;
  /** If the interface is P2P then there is only one protect
   * object associated with the auto-adj for each NH proto */
  index_t id_itp;
} ipsec_tun_protect_itf_db_t;

typedef struct ipsec_tun_protect_db_t_
{
  /** Per-interface vector */
  ipsec_tun_protect_itf_db_t *id_itf;
} ipsec_tun_protect_db_t;

static ipsec_tun_protect_db_t itp_db;

const static ipsec_tun_protect_itf_db_t IPSEC_TUN_PROTECT_DEFAULT_DB_ENTRY = {
  .id_itp = INDEX_INVALID,
};

#define ITP_DBG(_itp, _fmt, _args...)                   \
{   		          				\
  vlib_log_debug(ipsec_tun_protect_logger,              \
                 "[%U]: " _fmt,                         \
                 format_ipsec_tun_protect,              \
                 _itp, ##_args);                        \
}

#define ITP_DBG2(_fmt, _args...)                        \
{   		          				\
  vlib_log_debug(ipsec_tun_protect_logger,              \
                 _fmt, ##_args);                        \
}

static u32 ipsec_tun_node_regs[N_AF];

void
ipsec_tun_register_nodes (ip_address_family_t af)
{
  if (0 == ipsec_tun_node_regs[af]++)
    {
      if (AF_IP4 == af)
	{
	  ipsec_register_udp_port (UDP_DST_PORT_ipsec);
	  ip4_register_protocol (IP_PROTOCOL_IPSEC_ESP,
				 ipsec4_tun_input_node.index);
	}
      else
	ip6_register_protocol (IP_PROTOCOL_IPSEC_ESP,
			       ipsec6_tun_input_node.index);
    }
}

void
ipsec_tun_unregister_nodes (ip_address_family_t af)
{
  ASSERT (0 != ipsec_tun_node_regs[af]);
  if (0 == --ipsec_tun_node_regs[af])
    {
      if (AF_IP4 == af)
	{
	  ipsec_unregister_udp_port (UDP_DST_PORT_ipsec);
	  ip4_unregister_protocol (IP_PROTOCOL_IPSEC_ESP);
	}
      else
	ip6_unregister_protocol (IP_PROTOCOL_IPSEC_ESP);
    }
}

static inline const ipsec_tun_protect_t *
ipsec_tun_protect_from_const_base (const adj_delegate_t * ad)
{
  if (ad == NULL)
    return (NULL);
  return (pool_elt_at_index (ipsec_tun_protect_pool, ad->ad_index));
}

static u32
ipsec_tun_protect_get_adj_next (vnet_link_t linkt,
				const ipsec_tun_protect_t * itp)
{
  ipsec_main_t *im;
  ipsec_sa_t *sa;
  u32 next;

  if (!(itp->itp_flags & IPSEC_PROTECT_ITF))
    {
      if (ip46_address_is_ip4 (&itp->itp_tun.src))
	linkt = VNET_LINK_IP4;
      else
	linkt = VNET_LINK_IP6;
    }

  sa = ipsec_sa_get (itp->itp_out_sa);
  im = &ipsec_main;
  next = 0;

  if ((sa->crypto_alg == IPSEC_CRYPTO_ALG_NONE &&
       sa->integ_alg == IPSEC_INTEG_ALG_NONE) &&
      !(itp->itp_flags & IPSEC_PROTECT_ITF))
    next = (VNET_LINK_IP4 == linkt ? im->esp4_no_crypto_tun_node_index :
				     im->esp6_no_crypto_tun_node_index);
  else if (itp->itp_flags & IPSEC_PROTECT_L2)
    next = (VNET_LINK_IP4 == linkt ? im->esp4_encrypt_l2_tun_node_index :
				     im->esp6_encrypt_l2_tun_node_index);
  else
    {
      switch (linkt)
	{
	case VNET_LINK_IP4:
	  next = im->esp4_encrypt_tun_node_index;
	  break;
	case VNET_LINK_IP6:
	  next = im->esp6_encrypt_tun_node_index;
	  break;
	case VNET_LINK_MPLS:
	  next = im->esp_mpls_encrypt_tun_node_index;
	  break;
	case VNET_LINK_ARP:
	case VNET_LINK_NSH:
	case VNET_LINK_ETHERNET:
	  ASSERT (0);
	  break;
	}
    }
  return (next);
}

static void
ipsec_tun_protect_add_adj (adj_index_t ai, const ipsec_tun_protect_t * itp)
{
  vec_validate_init_empty (ipsec_tun_protect_sa_by_adj_index, ai,
			   INDEX_INVALID);

  if (NULL == itp)
    {
      ipsec_tun_protect_sa_by_adj_index[ai] = INDEX_INVALID;
      adj_nbr_midchain_reset_next_node (ai);
    }
  else
    {
      ipsec_tun_protect_sa_by_adj_index[ai] = itp->itp_out_sa;
      adj_nbr_midchain_update_next_node
	(ai, ipsec_tun_protect_get_adj_next (adj_get_link_type (ai), itp));
    }
}

static index_t
ipsec_tun_protect_find (u32 sw_if_index, const ip_address_t * nh)
{
  ipsec_tun_protect_itf_db_t *idi;
  uword *p;

  if (vec_len (itp_db.id_itf) <= sw_if_index)
    return INDEX_INVALID;

  if (vnet_sw_interface_is_p2p (vnet_get_main (), sw_if_index))
    return (itp_db.id_itf[sw_if_index].id_itp);

  idi = &itp_db.id_itf[sw_if_index];
  p = hash_get_mem (idi->id_hash, nh);

  if (NULL == p)
    {
      return INDEX_INVALID;
    }
  return (p[0]);
}

static void
ipsec_tun_protect_rx_db_add (ipsec_main_t * im,
			     const ipsec_tun_protect_t * itp)
{
  const ipsec_sa_t *sa;
  u32 sai;

  if (ip46_address_is_zero (&itp->itp_crypto.dst))
    return;

  /* *INDENT-OFF* */
  FOR_EACH_IPSEC_PROTECT_INPUT_SAI(itp, sai,
  ({
      sa = ipsec_sa_get (sai);

      ipsec_tun_lkup_result_t res = {
        .tun_index = itp - ipsec_tun_protect_pool,
        .sa_index = sai,
        .flags = itp->itp_flags,
        .sw_if_index = itp->itp_sw_if_index,
      };

      /*
       * The key is formed from the tunnel's destination
       * as the packet lookup is done from the packet's source
       */
      if (ip46_address_is_ip4 (&itp->itp_crypto.dst))
        {
          ipsec4_tunnel_kv_t key = {
            .value = res,
          };
          clib_bihash_kv_8_16_t *bkey = (clib_bihash_kv_8_16_t*)&key;

          ipsec4_tunnel_mk_key(&key, &itp->itp_crypto.dst.ip4,
                               clib_host_to_net_u32 (sa->spi));

	  if (!clib_bihash_is_initialised_8_16 (&im->tun4_protect_by_key))
	    clib_bihash_init_8_16 (&im->tun4_protect_by_key,
				   "IPSec IPv4 tunnels",
				   IPSEC_TUN_DEFAULT_HASH_NUM_BUCKETS,
				   IPSEC_TUN_DEFAULT_HASH_MEMORY_SIZE);

	  clib_bihash_add_del_8_16 (&im->tun4_protect_by_key, bkey, 1);
	  ipsec_tun_register_nodes (AF_IP4);
	}
      else
        {
          ipsec6_tunnel_kv_t key = {
            .key = {
              .remote_ip = itp->itp_crypto.dst.ip6,
              .spi = clib_host_to_net_u32 (sa->spi),
            },
            .value = res,
          };
          clib_bihash_kv_24_16_t *bkey = (clib_bihash_kv_24_16_t*)&key;

	  if (!clib_bihash_is_initialised_24_16 (&im->tun6_protect_by_key))
	    clib_bihash_init_24_16 (&im->tun6_protect_by_key,
				    "IPSec IPv6 tunnels",
				    IPSEC_TUN_DEFAULT_HASH_NUM_BUCKETS,
				    IPSEC_TUN_DEFAULT_HASH_MEMORY_SIZE);
	  clib_bihash_add_del_24_16 (&im->tun6_protect_by_key, bkey, 1);
	  ipsec_tun_register_nodes (AF_IP6);
	}
  }))
  /* *INDENT-ON* */
}

static adj_walk_rc_t
ipsec_tun_protect_adj_add (adj_index_t ai, void *arg)
{
  ipsec_tun_protect_t *itp = arg;
  adj_delegate_add (adj_get (ai), ipsec_tun_adj_delegate_type,
		    itp - ipsec_tun_protect_pool);
  ipsec_tun_protect_add_adj (ai, itp);

  if (itp->itp_flags & IPSEC_PROTECT_ITF)
    ipsec_itf_adj_stack (ai, itp->itp_out_sa);

  return (ADJ_WALK_RC_CONTINUE);
}

static void
ipsec_tun_protect_tx_db_add (ipsec_tun_protect_t * itp)
{
  /*
   * add the delegate to the adj
   */
  ipsec_tun_protect_itf_db_t *idi;
  fib_protocol_t nh_proto;
  ip46_address_t nh;

  vec_validate_init_empty (itp_db.id_itf,
			   itp->itp_sw_if_index,
			   IPSEC_TUN_PROTECT_DEFAULT_DB_ENTRY);

  idi = &itp_db.id_itf[itp->itp_sw_if_index];

  if (vnet_sw_interface_is_p2p (vnet_get_main (), itp->itp_sw_if_index))
    {
      if (INDEX_INVALID == idi->id_itp)
	{
	  // ipsec_tun_protect_feature_set (itp, 1);
	}
      idi->id_itp = itp - ipsec_tun_protect_pool;

      FOR_EACH_FIB_IP_PROTOCOL (nh_proto)
	adj_nbr_walk (itp->itp_sw_if_index,
		      nh_proto, ipsec_tun_protect_adj_add, itp);
    }
  else
    {
      if (NULL == idi->id_hash)
	{
	  idi->id_hash =
	    hash_create_mem (0, sizeof (ip_address_t), sizeof (uword));
	  /*
	   * enable the encrypt feature for egress if this is the first addition
	   * on this interface
	   */
	  // ipsec_tun_protect_feature_set (itp, 1);
	}

      hash_set_mem (idi->id_hash, itp->itp_key, itp - ipsec_tun_protect_pool);

      /*
       * walk all the adjs with the same nh on this interface
       * to associate them with this protection
       */
      nh_proto = ip_address_to_46 (itp->itp_key, &nh);

      adj_nbr_walk_nh (itp->itp_sw_if_index,
		       nh_proto, &nh, ipsec_tun_protect_adj_add, itp);

      ipsec_tun_register_nodes (FIB_PROTOCOL_IP6 == nh_proto ?
				AF_IP6 : AF_IP4);
    }
}

static void
ipsec_tun_protect_rx_db_remove (ipsec_main_t * im,
				const ipsec_tun_protect_t * itp)
{
  const ipsec_sa_t *sa;

  /* *INDENT-OFF* */
  FOR_EACH_IPSEC_PROTECT_INPUT_SA(itp, sa,
  ({
    if (ip46_address_is_ip4 (&itp->itp_crypto.dst))
      {
          ipsec4_tunnel_kv_t key;
          clib_bihash_kv_8_16_t res, *bkey = (clib_bihash_kv_8_16_t*)&key;

          ipsec4_tunnel_mk_key(&key, &itp->itp_crypto.dst.ip4,
                               clib_host_to_net_u32 (sa->spi));

          if (!clib_bihash_search_8_16 (&im->tun4_protect_by_key, bkey, &res))
            {
              clib_bihash_add_del_8_16 (&im->tun4_protect_by_key, bkey, 0);
              ipsec_tun_unregister_nodes(AF_IP4);
            }
      }
    else
      {
        ipsec6_tunnel_kv_t key = {
          .key = {
            .remote_ip = itp->itp_crypto.dst.ip6,
            .spi = clib_host_to_net_u32 (sa->spi),
          },
        };
        clib_bihash_kv_24_16_t res, *bkey = (clib_bihash_kv_24_16_t*)&key;

        if (!clib_bihash_search_24_16 (&im->tun6_protect_by_key, bkey, &res))
          {
            clib_bihash_add_del_24_16 (&im->tun6_protect_by_key, bkey, 0);
            ipsec_tun_unregister_nodes(AF_IP6);
          }
      }
  }));
  /* *INDENT-ON* */
}

static adj_walk_rc_t
ipsec_tun_protect_adj_remove (adj_index_t ai, void *arg)
{
  ipsec_tun_protect_t *itp = arg;

  adj_delegate_remove (ai, ipsec_tun_adj_delegate_type);
  ipsec_tun_protect_add_adj (ai, NULL);

  if (itp->itp_flags & IPSEC_PROTECT_ITF)
    ipsec_itf_adj_unstack (ai);

  return (ADJ_WALK_RC_CONTINUE);
}

static void
ipsec_tun_protect_tx_db_remove (ipsec_tun_protect_t * itp)
{
  ipsec_tun_protect_itf_db_t *idi;
  fib_protocol_t nh_proto;
  ip46_address_t nh;

  nh_proto = ip_address_to_46 (itp->itp_key, &nh);
  idi = &itp_db.id_itf[itp->itp_sw_if_index];

  if (vnet_sw_interface_is_p2p (vnet_get_main (), itp->itp_sw_if_index))
    {
      // ipsec_tun_protect_feature_set (itp, 0);
      idi->id_itp = INDEX_INVALID;

      FOR_EACH_FIB_IP_PROTOCOL (nh_proto)
	adj_nbr_walk (itp->itp_sw_if_index,
		      nh_proto, ipsec_tun_protect_adj_remove, itp);
    }
  else
    {
      adj_nbr_walk_nh (itp->itp_sw_if_index,
		       nh_proto, &nh, ipsec_tun_protect_adj_remove, itp);

      hash_unset_mem (idi->id_hash, itp->itp_key);

      if (0 == hash_elts (idi->id_hash))
	{
	  // ipsec_tun_protect_feature_set (itp, 0);
	  hash_free (idi->id_hash);
	  idi->id_hash = NULL;
	}
      ipsec_tun_unregister_nodes (FIB_PROTOCOL_IP6 == nh_proto ?
				  AF_IP6 : AF_IP4);
    }
}

static void
ipsec_tun_protect_set_crypto_addr (ipsec_tun_protect_t * itp)
{
  ipsec_sa_t *sa;

  /* *INDENT-OFF* */
  FOR_EACH_IPSEC_PROTECT_INPUT_SA(itp, sa,
  ({
    if (ipsec_sa_is_set_IS_TUNNEL (sa))
      {
	itp->itp_crypto.src = ip_addr_46 (&sa->tunnel.t_dst);
	itp->itp_crypto.dst = ip_addr_46 (&sa->tunnel.t_src);
	if (!(itp->itp_flags & IPSEC_PROTECT_ITF))
	  {
	    ipsec_sa_set_IS_PROTECT (sa);
	    itp->itp_flags |= IPSEC_PROTECT_ENCAPED;
	  }
      }
    else
      {
        itp->itp_crypto.src = itp->itp_tun.src;
        itp->itp_crypto.dst = itp->itp_tun.dst;
        itp->itp_flags &= ~IPSEC_PROTECT_ENCAPED;
      }
  }));
  /* *INDENT-ON* */
}

static void
ipsec_tun_protect_config (ipsec_main_t * im,
			  ipsec_tun_protect_t * itp, u32 sa_out, u32 * sas_in)
{
  index_t sai;
  u32 ii;

  itp->itp_n_sa_in = vec_len (sas_in);
  for (ii = 0; ii < itp->itp_n_sa_in; ii++)
    itp->itp_in_sas[ii] = sas_in[ii];
  itp->itp_out_sa = sa_out;

  ipsec_sa_lock (itp->itp_out_sa);

  /* *INDENT-OFF* */
  FOR_EACH_IPSEC_PROTECT_INPUT_SAI(itp, sai,
  ({
    ipsec_sa_lock(sai);
  }));
  ipsec_tun_protect_set_crypto_addr(itp);
  /* *INDENT-ON* */

  /*
   * add to the DB against each SA
   */
  ipsec_tun_protect_rx_db_add (im, itp);
  ipsec_tun_protect_tx_db_add (itp);

  ITP_DBG (itp, "configured");
}

static void
ipsec_tun_protect_unconfig (ipsec_main_t * im, ipsec_tun_protect_t * itp)
{
  ipsec_sa_t *sa;
  index_t sai;

  /* *INDENT-OFF* */
  FOR_EACH_IPSEC_PROTECT_INPUT_SA(itp, sa,
  ({
    ipsec_sa_unset_IS_PROTECT (sa);
  }));

  ipsec_tun_protect_rx_db_remove (im, itp);
  ipsec_tun_protect_tx_db_remove (itp);

  ipsec_sa_unlock(itp->itp_out_sa);

  FOR_EACH_IPSEC_PROTECT_INPUT_SAI(itp, sai,
  ({
    ipsec_sa_unlock(sai);
  }));
  /* *INDENT-ON* */
  ITP_DBG (itp, "unconfigured");
}

static void
ipsec_tun_protect_update_from_teib (ipsec_tun_protect_t * itp,
				    const teib_entry_t * ne)
{
  if (NULL != ne)
    {
      const fib_prefix_t *pfx;

      pfx = teib_entry_get_nh (ne);

      ip46_address_copy (&itp->itp_tun.dst, &pfx->fp_addr);
    }
  else
    ip46_address_reset (&itp->itp_tun.dst);
}

int
ipsec_tun_protect_update (u32 sw_if_index,
			  const ip_address_t * nh, u32 sa_out, u32 * sas_in)
{
  ipsec_tun_protect_t *itp;
  u32 itpi, ii, *saip;
  ipsec_main_t *im;
  int rv;

  ITP_DBG2 ("update: %U/%U",
	    format_vnet_sw_if_index_name, vnet_get_main (), sw_if_index,
	    format_ip_address, nh);

  if (vec_len (sas_in) > ITP_MAX_N_SA_IN)
    {
      rv = VNET_API_ERROR_LIMIT_EXCEEDED;
      goto out;
    }

  rv = 0;
  im = &ipsec_main;
  if (NULL == nh)
    nh = &IP_ADDR_ALL_0;
  itpi = ipsec_tun_protect_find (sw_if_index, nh);

  vec_foreach_index (ii, sas_in)
  {
    sas_in[ii] = ipsec_sa_find_and_lock (sas_in[ii]);
    if (~0 == sas_in[ii])
      {
	rv = VNET_API_ERROR_INVALID_VALUE;
	goto out;
      }
  }

  sa_out = ipsec_sa_find_and_lock (sa_out);

  if (~0 == sa_out)
    {
      rv = VNET_API_ERROR_INVALID_VALUE;
      goto out;
    }

  if (INDEX_INVALID == itpi)
    {
      vnet_device_class_t *dev_class;
      vnet_hw_interface_t *hi;
      vnet_main_t *vnm;
      u8 is_l2;

      vnm = vnet_get_main ();
      hi = vnet_get_sup_hw_interface (vnm, sw_if_index);
      dev_class = vnet_get_device_class (vnm, hi->dev_class_index);

      if (NULL == dev_class->ip_tun_desc)
	{
	  rv = VNET_API_ERROR_INVALID_SW_IF_INDEX;
	  goto out;
	}

      pool_get_zero (ipsec_tun_protect_pool, itp);

      itp->itp_sw_if_index = sw_if_index;
      itp->itp_ai = ADJ_INDEX_INVALID;

      itp->itp_n_sa_in = vec_len (sas_in);
      for (ii = 0; ii < itp->itp_n_sa_in; ii++)
	itp->itp_in_sas[ii] = sas_in[ii];
      itp->itp_out_sa = sa_out;

      itp->itp_key = clib_mem_alloc (sizeof (*itp->itp_key));
      ip_address_copy (itp->itp_key, nh);

      rv = dev_class->ip_tun_desc (sw_if_index,
				   &itp->itp_tun.src,
				   &itp->itp_tun.dst, &is_l2);

      if (rv)
	goto out;

      if (ip46_address_is_zero (&itp->itp_tun.src))
	{
	  /*
	   * must be one of those pesky ipsec interfaces that has no encap.
	   * the encap then MUST come from the tunnel mode SA.
	   */
	  ipsec_sa_t *sa;

	  sa = ipsec_sa_get (itp->itp_out_sa);

	  if (!ipsec_sa_is_set_IS_TUNNEL (sa))
	    {
	      rv = VNET_API_ERROR_INVALID_DST_ADDRESS;
	      goto out;
	    }

	  itp->itp_flags |= IPSEC_PROTECT_ITF;
	}
      else if (ip46_address_is_zero (&itp->itp_tun.dst))
	{
	  /* tunnel has no destination address, presumably because it's p2mp
	     in which case we use the nh that this is protection for */
	  ipsec_tun_protect_update_from_teib
	    (itp, teib_entry_find (sw_if_index, nh));
	}

      if (is_l2)
	itp->itp_flags |= IPSEC_PROTECT_L2;

      /*
       * add to the tunnel DB for ingress
       *  - if the SA is in trasnport mode, then the packates will arrive
       *    with the IP src,dst of the protected tunnel, in which case we can
       *    simply strip the IP header and hand the payload to the protocol
       *    appropriate input handler
       *  - if the SA is in tunnel mode then there are two IP headers present
       *    one for the crytpo tunnel endpoints (described in the SA) and one
       *    for the tunnel endpoints. The outer IP headers in the srriving
       *    packets will have the crypto endpoints. So the DB needs to contain
       *    the crpto endpoint. Once the crypto header is stripped, revealing,
       *    the tunnel-IP we have 2 choices:
       *     1) do a tunnel lookup based on the revealed header
       *     2) skip the tunnel lookup and assume that the packet matches the
       *        one that is protected here.
       *    If we did 1) then we would allow our peer to use the SA for tunnel
       *    X to inject traffic onto tunnel Y, this is not good. If we do 2)
       *    then we don't verify that the peer is indeed using SA for tunnel
       *    X and addressing tunnel X. So we take a compromise, once the SA
       *    matches to tunnel X we veriy that the inner IP matches the value
       *    of the tunnel we are protecting, else it's dropped.
       */
      ipsec_tun_protect_config (im, itp, sa_out, sas_in);
    }
  else
    {
      /* updating SAs only */
      itp = pool_elt_at_index (ipsec_tun_protect_pool, itpi);

      ipsec_tun_protect_unconfig (im, itp);
      ipsec_tun_protect_config (im, itp, sa_out, sas_in);
    }

  ipsec_sa_unlock (sa_out);
  vec_foreach (saip, sas_in) ipsec_sa_unlock (*saip);
  vec_free (sas_in);

out:
  return (rv);
}

int
ipsec_tun_protect_del (u32 sw_if_index, const ip_address_t * nh)
{
  ipsec_tun_protect_t *itp;
  ipsec_main_t *im;
  index_t itpi;

  ITP_DBG2 ("delete: %U/%U",
	    format_vnet_sw_if_index_name, vnet_get_main (), sw_if_index,
	    format_ip_address, nh);

  im = &ipsec_main;
  if (NULL == nh)
    nh = &IP_ADDR_ALL_0;

  itpi = ipsec_tun_protect_find (sw_if_index, nh);

  if (INDEX_INVALID == itpi)
    return (VNET_API_ERROR_NO_SUCH_ENTRY);

  itp = ipsec_tun_protect_get (itpi);
  ipsec_tun_protect_unconfig (im, itp);

  if (ADJ_INDEX_INVALID != itp->itp_ai)
    adj_unlock (itp->itp_ai);

  clib_mem_free (itp->itp_key);
  pool_put (ipsec_tun_protect_pool, itp);

  return (0);
}

void
ipsec_tun_protect_walk (ipsec_tun_protect_walk_cb_t fn, void *ctx)
{
  index_t itpi;

  /* *INDENT-OFF* */
  pool_foreach_index (itpi, ipsec_tun_protect_pool)
   {
    fn (itpi, ctx);
  }
  /* *INDENT-ON* */
}

void
ipsec_tun_protect_walk_itf (u32 sw_if_index,
			    ipsec_tun_protect_walk_cb_t fn, void *ctx)
{
  ipsec_tun_protect_itf_db_t *idi;
  ip_address_t *key;
  index_t itpi;

  if (vec_len (itp_db.id_itf) <= sw_if_index)
    return;

  idi = &itp_db.id_itf[sw_if_index];

  /* *INDENT-OFF* */
  hash_foreach(key, itpi, idi->id_hash,
  ({
    fn (itpi, ctx);
  }));
  /* *INDENT-ON* */
  if (INDEX_INVALID != idi->id_itp)
    fn (idi->id_itp, ctx);
}

static void
ipsec_tun_feature_update (u32 sw_if_index, u8 arc_index, u8 is_enable,
			  void *data)
{
  ipsec_tun_protect_t *itp;
  index_t itpi;

  if (arc_index != feature_main.device_input_feature_arc_index)
    return;

  /* Only p2p tunnels supported */
  itpi = ipsec_tun_protect_find (sw_if_index, &IP_ADDR_ALL_0);
  if (itpi == INDEX_INVALID)
    return;

  itp = ipsec_tun_protect_get (itpi);

  if (is_enable)
    {
      u32 decrypt_tun = ip46_address_is_ip4 (&itp->itp_crypto.dst) ?
			  ipsec_main.esp4_decrypt_tun_node_index :
			  ipsec_main.esp6_decrypt_tun_node_index;

      vnet_feature_modify_end_node (
	feature_main.device_input_feature_arc_index, sw_if_index, decrypt_tun);
      itp->itp_flags |= IPSEC_PROTECT_FEAT;
    }
  else
    {
      u32 eth_in =
	vlib_get_node_by_name (vlib_get_main (), (u8 *) "ethernet-input")
	  ->index;

      vnet_feature_modify_end_node (
	feature_main.device_input_feature_arc_index, sw_if_index, eth_in);
      itp->itp_flags &= ~IPSEC_PROTECT_FEAT;
    }

  /* Propagate flag change into lookup entries */
  ipsec_tun_protect_rx_db_remove (&ipsec_main, itp);
  ipsec_tun_protect_rx_db_add (&ipsec_main, itp);
}

static void
ipsec_tun_protect_adj_delegate_adj_deleted (adj_delegate_t * ad)
{
  /* remove our delegate */
  ipsec_tun_protect_add_adj (ad->ad_adj_index, NULL);
  adj_delegate_remove (ad->ad_adj_index, ipsec_tun_adj_delegate_type);
}

static void
ipsec_tun_protect_adj_delegate_adj_modified (adj_delegate_t * ad)
{
  ipsec_tun_protect_add_adj (ad->ad_adj_index,
			     ipsec_tun_protect_get (ad->ad_index));
}

static void
ipsec_tun_protect_adj_delegate_adj_created (adj_index_t ai)
{
  /* add our delegate if there is protection for this neighbour */
  ip_address_t ip = IP_ADDRESS_V4_ALL_0S;
  ip_adjacency_t *adj;
  index_t itpi;

  if (!adj_is_midchain (ai))
    return;

  adj = adj_get (ai);

  ip_address_from_46 (&adj->sub_type.midchain.next_hop,
		      adj->ia_nh_proto, &ip);

  itpi = ipsec_tun_protect_find (adj->rewrite_header.sw_if_index, &ip);

  if (INDEX_INVALID != itpi)
    ipsec_tun_protect_adj_add (ai, ipsec_tun_protect_get (itpi));
}

static u8 *
ipsec_tun_protect_adj_delegate_format (const adj_delegate_t * aed, u8 * s)
{
  const ipsec_tun_protect_t *itp;

  itp = ipsec_tun_protect_from_const_base (aed);
  s = format (s, "ipsec-tun-protect:\n%U", format_ipsec_tun_protect, itp);

  return (s);
}

static void
ipsec_tun_teib_entry_added (const teib_entry_t * ne)
{
  ipsec_tun_protect_t *itp;
  index_t itpi;

  itpi = ipsec_tun_protect_find (teib_entry_get_sw_if_index (ne),
				 teib_entry_get_peer (ne));

  if (INDEX_INVALID == itpi)
    return;

  itp = ipsec_tun_protect_get (itpi);
  ipsec_tun_protect_rx_db_remove (&ipsec_main, itp);
  ipsec_tun_protect_update_from_teib (itp, ne);
  ipsec_tun_protect_set_crypto_addr (itp);
  ipsec_tun_protect_rx_db_add (&ipsec_main, itp);

  ITP_DBG (itp, "teib-added");
}

static void
ipsec_tun_teib_entry_deleted (const teib_entry_t * ne)
{
  ipsec_tun_protect_t *itp;
  index_t itpi;

  itpi = ipsec_tun_protect_find (teib_entry_get_sw_if_index (ne),
				 teib_entry_get_peer (ne));

  if (INDEX_INVALID == itpi)
    return;

  itp = ipsec_tun_protect_get (itpi);
  ipsec_tun_protect_rx_db_remove (&ipsec_main, itp);
  ipsec_tun_protect_update_from_teib (itp, NULL);
  ipsec_tun_protect_set_crypto_addr (itp);

  ITP_DBG (itp, "teib-removed");
}

/**
 * VFT registered with the adjacency delegate
 */
const static adj_delegate_vft_t ipsec_tun_adj_delegate_vft = {
  .adv_adj_deleted = ipsec_tun_protect_adj_delegate_adj_deleted,
  .adv_adj_created = ipsec_tun_protect_adj_delegate_adj_created,
  .adv_adj_modified = ipsec_tun_protect_adj_delegate_adj_modified,
  .adv_format = ipsec_tun_protect_adj_delegate_format,
};

const static teib_vft_t ipsec_tun_teib_vft = {
  .nv_added = ipsec_tun_teib_entry_added,
  .nv_deleted = ipsec_tun_teib_entry_deleted,
};

void
ipsec_tun_table_init (ip_address_family_t af, uword table_size, u32 n_buckets)
{
  ipsec_main_t *im;

  im = &ipsec_main;

  if (AF_IP4 == af)
    clib_bihash_init_8_16 (&im->tun4_protect_by_key,
			   "IPSec IPv4 tunnels", n_buckets, table_size);
  else
    clib_bihash_init_24_16 (&im->tun6_protect_by_key,
			    "IPSec IPv6 tunnels", n_buckets, table_size);
}

static clib_error_t *
ipsec_tunnel_protect_init (vlib_main_t *vm)
{
  ipsec_main_t *im;

  im = &ipsec_main;
  clib_bihash_init_24_16 (&im->tun6_protect_by_key,
			  "IPSec IPv6 tunnels",
			  IPSEC_TUN_DEFAULT_HASH_NUM_BUCKETS,
			  IPSEC_TUN_DEFAULT_HASH_MEMORY_SIZE);
  clib_bihash_init_8_16 (&im->tun4_protect_by_key,
			 "IPSec IPv4 tunnels",
			 IPSEC_TUN_DEFAULT_HASH_NUM_BUCKETS,
			 IPSEC_TUN_DEFAULT_HASH_MEMORY_SIZE);

  /* set up feature nodes to drop outbound packets with no crypto alg set */
  im->esp4_no_crypto_tun_node_index =
    vlib_get_node_by_name (vm, (u8 *) "esp4-no-crypto")->index;
  im->esp6_no_crypto_tun_node_index =
    vlib_get_node_by_name (vm, (u8 *) "esp6-no-crypto")->index;
  im->esp6_encrypt_l2_tun_node_index =
    vlib_get_node_by_name (vm, (u8 *) "esp6-encrypt-tun")->index;
  im->esp4_encrypt_l2_tun_node_index =
    vlib_get_node_by_name (vm, (u8 *) "esp4-encrypt-tun")->index;

  ipsec_tun_adj_delegate_type =
    adj_delegate_register_new_type (&ipsec_tun_adj_delegate_vft);

  ipsec_tun_protect_logger = vlib_log_register_class ("ipsec", "tun");

  teib_register (&ipsec_tun_teib_vft);

  vnet_feature_register (ipsec_tun_feature_update, NULL);

  return 0;
}

VLIB_INIT_FUNCTION (ipsec_tunnel_protect_init);

/*
 * fd.io coding-style-patch-verification: ON
 *
 * Local Variables:
 * eval: (c-set-style "gnu")
 * End:
 */