vxlan.c

Go to the documentation of this file.

/*
 * 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/vxlan/vxlan.h>
#include <vnet/ip/format.h>
#include <vnet/fib/fib_entry.h>
#include <vnet/fib/fib_table.h>
#include <vnet/fib/fib_entry_track.h>
#include <vnet/mfib/mfib_table.h>
#include <vnet/adj/adj_mcast.h>
#include <vnet/adj/rewrite.h>
#include <vnet/dpo/drop_dpo.h>
#include <vnet/interface.h>
#include <vnet/flow/flow.h>
#include <vnet/udp/udp_local.h>
#include <vlib/vlib.h>

/**
 * @file
 * @brief VXLAN.
 *
 * VXLAN provides the features needed to allow L2 bridge domains (BDs)
 * to span multiple servers. This is done by building an L2 overlay on
 * top of an L3 network underlay using VXLAN tunnels.
 *
 * This makes it possible for servers to be co-located in the same data
 * center or be separated geographically as long as they are reachable
 * through the underlay L3 network.
 *
 * You can refer to this kind of L2 overlay bridge domain as a VXLAN
 * (Virtual eXtensible VLAN) segment.
 */


vxlan_main_t vxlan_main;

static u32
vxlan_eth_flag_change (vnet_main_t *vnm, vnet_hw_interface_t *hi, u32 flags)
{
  /* nothing for now */
  return 0;
}

static u8 *
format_decap_next (u8 * s, va_list * args)
{
  u32 next_index = va_arg (*args, u32);

  if (next_index == VXLAN_INPUT_NEXT_DROP)
    return format (s, "drop");
  else
    return format (s, "index %d", next_index);
  return s;
}

u8 *
format_vxlan_tunnel (u8 * s, va_list * args)
{
  vxlan_tunnel_t *t = va_arg (*args, vxlan_tunnel_t *);

  s = format (s,
              "[%d] instance %d src %U dst %U src_port %d dst_port %d vni %d "
              "fib-idx %d sw-if-idx %d ",
              t->dev_instance, t->user_instance, format_ip46_address, &t->src,
              IP46_TYPE_ANY, format_ip46_address, &t->dst, IP46_TYPE_ANY,
              t->src_port, t->dst_port, t->vni, t->encap_fib_index,
              t->sw_if_index);

  s = format (s, "encap-dpo-idx %d ", t->next_dpo.dpoi_index);

  if (PREDICT_FALSE (t->decap_next_index != VXLAN_INPUT_NEXT_L2_INPUT))
    s = format (s, "decap-next-%U ", format_decap_next, t->decap_next_index);

  if (PREDICT_FALSE (ip46_address_is_multicast (&t->dst)))
    s = format (s, "mcast-sw-if-idx %d ", t->mcast_sw_if_index);

  if (t->flow_index != ~0)
    s = format (s, "flow-index %d [%U]", t->flow_index,
                format_flow_enabled_hw, t->flow_index);

  return s;
}

static u8 *
format_vxlan_name (u8 * s, va_list * args)
{
  u32 dev_instance = va_arg (*args, u32);
  vxlan_main_t *vxm = &vxlan_main;
  vxlan_tunnel_t *t;

  if (dev_instance == ~0)
    return format (s, "<cached-unused>");

  if (dev_instance >= vec_len (vxm->tunnels))
    return format (s, "<improperly-referenced>");

  t = pool_elt_at_index (vxm->tunnels, dev_instance);

  return format (s, "vxlan_tunnel%d", t->user_instance);
}

static clib_error_t *
vxlan_interface_admin_up_down (vnet_main_t * vnm, u32 hw_if_index, u32 flags)
{
  u32 hw_flags = (flags & VNET_SW_INTERFACE_FLAG_ADMIN_UP) ?
    VNET_HW_INTERFACE_FLAG_LINK_UP : 0;
  vnet_hw_interface_set_flags (vnm, hw_if_index, hw_flags);

  return /* no error */ 0;
}

/* *INDENT-OFF* */
VNET_DEVICE_CLASS (vxlan_device_class, static) = {
  .name = "VXLAN",
  .format_device_name = format_vxlan_name,
  .format_tx_trace = format_vxlan_encap_trace,
  .admin_up_down_function = vxlan_interface_admin_up_down,
};
/* *INDENT-ON* */

static u8 *
format_vxlan_header_with_length (u8 * s, va_list * args)
{
  u32 dev_instance = va_arg (*args, u32);
  s = format (s, "unimplemented dev %u", dev_instance);
  return s;
}

/* *INDENT-OFF* */
VNET_HW_INTERFACE_CLASS (vxlan_hw_class) = {
  .name = "VXLAN",
  .format_header = format_vxlan_header_with_length,
  .build_rewrite = default_build_rewrite,
};
/* *INDENT-ON* */

static void
vxlan_tunnel_restack_dpo (vxlan_tunnel_t * t)
{
  u8 is_ip4 = ip46_address_is_ip4 (&t->dst);
  dpo_id_t dpo = DPO_INVALID;
  fib_forward_chain_type_t forw_type = is_ip4 ?
    FIB_FORW_CHAIN_TYPE_UNICAST_IP4 : FIB_FORW_CHAIN_TYPE_UNICAST_IP6;

  fib_entry_contribute_forwarding (t->fib_entry_index, forw_type, &dpo);

  /* vxlan uses the payload hash as the udp source port
   * hence the packet's hash is unknown
   * skip single bucket load balance dpo's */
  while (DPO_LOAD_BALANCE == dpo.dpoi_type)
    {
      const load_balance_t *lb;
      const dpo_id_t *choice;

      lb = load_balance_get (dpo.dpoi_index);
      if (lb->lb_n_buckets > 1)
        break;

      choice = load_balance_get_bucket_i (lb, 0);

      if (DPO_RECEIVE == choice->dpoi_type)
        dpo_copy (&dpo, drop_dpo_get (choice->dpoi_proto));
      else
        dpo_copy (&dpo, choice);
    }

  u32 encap_index = is_ip4 ?
    vxlan4_encap_node.index : vxlan6_encap_node.index;
  dpo_stack_from_node (encap_index, &t->next_dpo, &dpo);
  dpo_reset (&dpo);
}

static vxlan_tunnel_t *
vxlan_tunnel_from_fib_node (fib_node_t * node)
{
  ASSERT (FIB_NODE_TYPE_VXLAN_TUNNEL == node->fn_type);
  return ((vxlan_tunnel_t *) (((char *) node) -
                              STRUCT_OFFSET_OF (vxlan_tunnel_t, node)));
}

/**
 * Function definition to backwalk a FIB node -
 * Here we will restack the new dpo of VXLAN DIP to encap node.
 */
static fib_node_back_walk_rc_t
vxlan_tunnel_back_walk (fib_node_t * node, fib_node_back_walk_ctx_t * ctx)
{
  vxlan_tunnel_restack_dpo (vxlan_tunnel_from_fib_node (node));
  return (FIB_NODE_BACK_WALK_CONTINUE);
}

/**
 * Function definition to get a FIB node from its index
 */
static fib_node_t *
vxlan_tunnel_fib_node_get (fib_node_index_t index)
{
  vxlan_tunnel_t *t;
  vxlan_main_t *vxm = &vxlan_main;

  t = pool_elt_at_index (vxm->tunnels, index);

  return (&t->node);
}

/**
 * Function definition to inform the FIB node that its last lock has gone.
 */
static void
vxlan_tunnel_last_lock_gone (fib_node_t * node)
{
  /*
   * The VXLAN tunnel is a root of the graph. As such
   * it never has children and thus is never locked.
   */
  ASSERT (0);
}

/*
 * Virtual function table registered by VXLAN tunnels
 * for participation in the FIB object graph.
 */
const static fib_node_vft_t vxlan_vft = {
  .fnv_get = vxlan_tunnel_fib_node_get,
  .fnv_last_lock = vxlan_tunnel_last_lock_gone,
  .fnv_back_walk = vxlan_tunnel_back_walk,
};

#define foreach_copy_field                                                    \
  _ (vni)                                                                     \
  _ (mcast_sw_if_index)                                                       \
  _ (encap_fib_index)                                                         \
  _ (decap_next_index)                                                        \
  _ (src)                                                                     \
  _ (dst)                                                                     \
  _ (src_port)                                                                \
  _ (dst_port)

static void
vxlan_rewrite (vxlan_tunnel_t * t, bool is_ip6)
{
  union
  {
    ip4_vxlan_header_t h4;
    ip6_vxlan_header_t h6;
  } h;
  int len = is_ip6 ? sizeof h.h6 : sizeof h.h4;

  udp_header_t *udp;
  vxlan_header_t *vxlan;
  /* Fixed portion of the (outer) ip header */

  clib_memset (&h, 0, sizeof (h));
  if (!is_ip6)
    {
      ip4_header_t *ip = &h.h4.ip4;
      udp = &h.h4.udp, vxlan = &h.h4.vxlan;
      ip->ip_version_and_header_length = 0x45;
      ip->ttl = 254;
      ip->protocol = IP_PROTOCOL_UDP;

      ip->src_address = t->src.ip4;
      ip->dst_address = t->dst.ip4;

      /* we fix up the ip4 header length and checksum after-the-fact */
      ip->checksum = ip4_header_checksum (ip);
    }
  else
    {
      ip6_header_t *ip = &h.h6.ip6;
      udp = &h.h6.udp, vxlan = &h.h6.vxlan;
      ip->ip_version_traffic_class_and_flow_label =
        clib_host_to_net_u32 (6 << 28);
      ip->hop_limit = 255;
      ip->protocol = IP_PROTOCOL_UDP;

      ip->src_address = t->src.ip6;
      ip->dst_address = t->dst.ip6;
    }

  /* UDP header, randomize src port on something, maybe? */
  udp->src_port = clib_host_to_net_u16 (t->src_port);
  udp->dst_port = clib_host_to_net_u16 (t->dst_port);

  /* VXLAN header */
  vnet_set_vni_and_flags (vxlan, t->vni);
  vnet_rewrite_set_data (*t, &h, len);
}

static bool
vxlan_decap_next_is_valid (vxlan_main_t * vxm, u32 is_ip6,
                           u32 decap_next_index)
{
  vlib_main_t *vm = vxm->vlib_main;
  u32 input_idx = (!is_ip6) ?
    vxlan4_input_node.index : vxlan6_input_node.index;
  vlib_node_runtime_t *r = vlib_node_get_runtime (vm, input_idx);

  return decap_next_index < r->n_next_nodes;
}

/* *INDENT-OFF* */
typedef CLIB_PACKED(union
{
  struct
  {
    fib_node_index_t mfib_entry_index;
    adj_index_t mcast_adj_index;
  };
  u64 as_u64;
}) mcast_shared_t;
/* *INDENT-ON* */

static inline mcast_shared_t
mcast_shared_get (ip46_address_t * ip)
{
  ASSERT (ip46_address_is_multicast (ip));
  uword *p = hash_get_mem (vxlan_main.mcast_shared, ip);
  ALWAYS_ASSERT (p);
  mcast_shared_t ret = {.as_u64 = *p };
  return ret;
}

static inline void
mcast_shared_add (ip46_address_t * dst, fib_node_index_t mfei, adj_index_t ai)
{
  mcast_shared_t new_ep = {
    .mcast_adj_index = ai,
    .mfib_entry_index = mfei,
  };

  hash_set_mem_alloc (&vxlan_main.mcast_shared, dst, new_ep.as_u64);
}

static inline void
mcast_shared_remove (ip46_address_t * dst)
{
  mcast_shared_t ep = mcast_shared_get (dst);

  adj_unlock (ep.mcast_adj_index);
  mfib_table_entry_delete_index (ep.mfib_entry_index, MFIB_SOURCE_VXLAN);

  hash_unset_mem_free (&vxlan_main.mcast_shared, dst);
}

int vnet_vxlan_add_del_tunnel
  (vnet_vxlan_add_del_tunnel_args_t * a, u32 * sw_if_indexp)
{
  vxlan_main_t *vxm = &vxlan_main;
  vnet_main_t *vnm = vxm->vnet_main;
  vxlan_decap_info_t *p;
  u32 sw_if_index = ~0;
  vxlan4_tunnel_key_t key4;
  vxlan6_tunnel_key_t key6;
  u32 is_ip6 = a->is_ip6;
  vlib_main_t *vm = vlib_get_main ();
  u8 hw_addr[6];

  /* Set udp-ports */
  if (a->src_port == 0)
    a->src_port = is_ip6 ? UDP_DST_PORT_vxlan6 : UDP_DST_PORT_vxlan;

  if (a->dst_port == 0)
    a->dst_port = is_ip6 ? UDP_DST_PORT_vxlan6 : UDP_DST_PORT_vxlan;

  int not_found;
  if (!is_ip6)
    {
      /* ip4 mcast is indexed by mcast addr only */
      key4.key[0] = ip46_address_is_multicast (&a->dst) ?
                      a->dst.ip4.as_u32 :
                      a->dst.ip4.as_u32 | (((u64) a->src.ip4.as_u32) << 32);
      key4.key[1] = ((u64) clib_host_to_net_u16 (a->src_port) << 48) |
                    (((u64) a->encap_fib_index) << 32) |
                    clib_host_to_net_u32 (a->vni << 8);
      not_found =
        clib_bihash_search_inline_16_8 (&vxm->vxlan4_tunnel_by_key, &key4);
      p = (void *) &key4.value;
    }
  else
    {
      key6.key[0] = a->dst.ip6.as_u64[0];
      key6.key[1] = a->dst.ip6.as_u64[1];
      key6.key[2] = (((u64) clib_host_to_net_u16 (a->src_port) << 48) |
                     ((u64) a->encap_fib_index) << 32) |
                    clib_host_to_net_u32 (a->vni << 8);
      not_found =
        clib_bihash_search_inline_24_8 (&vxm->vxlan6_tunnel_by_key, &key6);
      p = (void *) &key6.value;
    }

  if (not_found)
    p = 0;

  if (a->is_add)
    {
      l2input_main_t *l2im = &l2input_main;
      u32 dev_instance;         /* real dev instance tunnel index */
      u32 user_instance;        /* request and actual instance number */

      /* adding a tunnel: tunnel must not already exist */
      if (p)
        return VNET_API_ERROR_TUNNEL_EXIST;

      /*if not set explicitly, default to l2 */
      if (a->decap_next_index == ~0)
        a->decap_next_index = VXLAN_INPUT_NEXT_L2_INPUT;
      if (!vxlan_decap_next_is_valid (vxm, is_ip6, a->decap_next_index))
        return VNET_API_ERROR_INVALID_DECAP_NEXT;

      vxlan_tunnel_t *t;
      pool_get_aligned (vxm->tunnels, t, CLIB_CACHE_LINE_BYTES);
      clib_memset (t, 0, sizeof (*t));
      dev_instance = t - vxm->tunnels;

      /* copy from arg structure */
#define _(x) t->x = a->x;
      foreach_copy_field;
#undef _

      vxlan_rewrite (t, is_ip6);
      /*
       * Reconcile the real dev_instance and a possible requested instance.
       */
      user_instance = a->instance;
      if (user_instance == ~0)
        user_instance = dev_instance;
      if (hash_get (vxm->instance_used, user_instance))
        {
          pool_put (vxm->tunnels, t);
          return VNET_API_ERROR_INSTANCE_IN_USE;
        }

      hash_set (vxm->instance_used, user_instance, 1);

      t->dev_instance = dev_instance;   /* actual */
      t->user_instance = user_instance; /* name */
      t->flow_index = ~0;

      if (a->is_l3)
        t->hw_if_index =
          vnet_register_interface (vnm, vxlan_device_class.index, dev_instance,
                                   vxlan_hw_class.index, dev_instance);
      else
        {
          f64 now = vlib_time_now (vm);
          u32 rnd;
          rnd = (u32) (now * 1e6);
          rnd = random_u32 (&rnd);
          memcpy (hw_addr + 2, &rnd, sizeof (rnd));
          hw_addr[0] = 2;
          hw_addr[1] = 0xfe;
          if (ethernet_register_interface (
                vnm, vxlan_device_class.index, dev_instance, hw_addr,
                &t->hw_if_index, vxlan_eth_flag_change))
            {
              hash_unset (vxm->instance_used, t->user_instance);

              pool_put (vxm->tunnels, t);
              return VNET_API_ERROR_SYSCALL_ERROR_2;
            }
        }

      vnet_hw_interface_t *hi = vnet_get_hw_interface (vnm, t->hw_if_index);

      /* Set vxlan tunnel output node */
      u32 encap_index = !is_ip6 ?
        vxlan4_encap_node.index : vxlan6_encap_node.index;
      vnet_set_interface_output_node (vnm, t->hw_if_index, encap_index);

      t->sw_if_index = sw_if_index = hi->sw_if_index;

      /* copy the key */
      int add_failed;
      if (is_ip6)
        {
          key6.value = (u64) dev_instance;
          add_failed = clib_bihash_add_del_24_8 (&vxm->vxlan6_tunnel_by_key,
                                                 &key6, 1 /*add */ );
        }
      else
        {
          vxlan_decap_info_t di = {.sw_if_index = t->sw_if_index, };
          if (ip46_address_is_multicast (&t->dst))
            di.local_ip = t->src.ip4;
          else
            di.next_index = t->decap_next_index;
          key4.value = di.as_u64;
          add_failed = clib_bihash_add_del_16_8 (&vxm->vxlan4_tunnel_by_key,
                                                 &key4, 1 /*add */ );
        }

      if (add_failed)
        {
          if (a->is_l3)
            vnet_delete_hw_interface (vnm, t->hw_if_index);
          else
            ethernet_delete_interface (vnm, t->hw_if_index);
          hash_unset (vxm->instance_used, t->user_instance);
          pool_put (vxm->tunnels, t);
          return VNET_API_ERROR_INVALID_REGISTRATION;
        }

      vec_validate_init_empty (vxm->tunnel_index_by_sw_if_index, sw_if_index,
                               ~0);
      vxm->tunnel_index_by_sw_if_index[sw_if_index] = dev_instance;

      /* setup l2 input config with l2 feature and bd 0 to drop packet */
      vec_validate (l2im->configs, sw_if_index);
      l2im->configs[sw_if_index].feature_bitmap = L2INPUT_FEAT_DROP;
      l2im->configs[sw_if_index].bd_index = 0;

      vnet_sw_interface_t *si = vnet_get_sw_interface (vnm, sw_if_index);
      si->flags &= ~VNET_SW_INTERFACE_FLAG_HIDDEN;
      vnet_sw_interface_set_flags (vnm, sw_if_index,
                                   VNET_SW_INTERFACE_FLAG_ADMIN_UP);

      fib_node_init (&t->node, FIB_NODE_TYPE_VXLAN_TUNNEL);
      fib_prefix_t tun_dst_pfx;
      vnet_flood_class_t flood_class = VNET_FLOOD_CLASS_TUNNEL_NORMAL;

      fib_prefix_from_ip46_addr (&t->dst, &tun_dst_pfx);
      if (!ip46_address_is_multicast (&t->dst))
        {
          /* Unicast tunnel -
           * source the FIB entry for the tunnel's destination
           * and become a child thereof. The tunnel will then get poked
           * when the forwarding for the entry updates, and the tunnel can
           * re-stack accordingly
           */
          vtep_addr_ref (&vxm->vtep_table, t->encap_fib_index, &t->src);
          t->fib_entry_index = fib_entry_track (t->encap_fib_index,
                                                &tun_dst_pfx,
                                                FIB_NODE_TYPE_VXLAN_TUNNEL,
                                                dev_instance,
                                                &t->sibling_index);
          vxlan_tunnel_restack_dpo (t);
        }
      else
        {
          /* Multicast tunnel -
           * as the same mcast group can be used for multiple mcast tunnels
           * with different VNIs, create the output fib adjacency only if
           * it does not already exist
           */
          fib_protocol_t fp = fib_ip_proto (is_ip6);

          if (vtep_addr_ref (&vxm->vtep_table,
                             t->encap_fib_index, &t->dst) == 1)
            {
              fib_node_index_t mfei;
              adj_index_t ai;
              fib_route_path_t path = {
                .frp_proto = fib_proto_to_dpo (fp),
                .frp_addr = zero_addr,
                .frp_sw_if_index = 0xffffffff,
                .frp_fib_index = ~0,
                .frp_weight = 1,
                .frp_flags = FIB_ROUTE_PATH_LOCAL,
                .frp_mitf_flags = MFIB_ITF_FLAG_FORWARD,
              };
              const mfib_prefix_t mpfx = {
                .fp_proto = fp,
                .fp_len = (is_ip6 ? 128 : 32),
                .fp_grp_addr = tun_dst_pfx.fp_addr,
              };

              /*
               * Setup the (*,G) to receive traffic on the mcast group
               *  - the forwarding interface is for-us
               *  - the accepting interface is that from the API
               */
              mfib_table_entry_path_update (t->encap_fib_index,
                                            &mpfx, MFIB_SOURCE_VXLAN, &path);

              path.frp_sw_if_index = a->mcast_sw_if_index;
              path.frp_flags = FIB_ROUTE_PATH_FLAG_NONE;
              path.frp_mitf_flags = MFIB_ITF_FLAG_ACCEPT;
              mfei = mfib_table_entry_path_update (t->encap_fib_index,
                                                   &mpfx,
                                                   MFIB_SOURCE_VXLAN, &path);

              /*
               * Create the mcast adjacency to send traffic to the group
               */
              ai = adj_mcast_add_or_lock (fp,
                                          fib_proto_to_link (fp),
                                          a->mcast_sw_if_index);

              /*
               * create a new end-point
               */
              mcast_shared_add (&t->dst, mfei, ai);
            }

          dpo_id_t dpo = DPO_INVALID;
          mcast_shared_t ep = mcast_shared_get (&t->dst);

          /* Stack shared mcast dst mac addr rewrite on encap */
          dpo_set (&dpo, DPO_ADJACENCY_MCAST,
                   fib_proto_to_dpo (fp), ep.mcast_adj_index);

          dpo_stack_from_node (encap_index, &t->next_dpo, &dpo);
          dpo_reset (&dpo);
          flood_class = VNET_FLOOD_CLASS_TUNNEL_MASTER;
        }

      vnet_get_sw_interface (vnet_get_main (), sw_if_index)->flood_class =
        flood_class;
    }
  else
    {
      /* deleting a tunnel: tunnel must exist */
      if (!p)
        return VNET_API_ERROR_NO_SUCH_ENTRY;

      u32 instance = is_ip6 ? key6.value :
        vxm->tunnel_index_by_sw_if_index[p->sw_if_index];
      vxlan_tunnel_t *t = pool_elt_at_index (vxm->tunnels, instance);

      sw_if_index = t->sw_if_index;
      vnet_sw_interface_set_flags (vnm, sw_if_index, 0 /* down */ );

      vxm->tunnel_index_by_sw_if_index[sw_if_index] = ~0;

      if (!is_ip6)
        clib_bihash_add_del_16_8 (&vxm->vxlan4_tunnel_by_key, &key4,
                                  0 /*del */ );
      else
        clib_bihash_add_del_24_8 (&vxm->vxlan6_tunnel_by_key, &key6,
                                  0 /*del */ );

      if (!ip46_address_is_multicast (&t->dst))
        {
          if (t->flow_index != ~0)
            vnet_flow_del (vnm, t->flow_index);

          vtep_addr_unref (&vxm->vtep_table, t->encap_fib_index, &t->src);
          fib_entry_untrack (t->fib_entry_index, t->sibling_index);
        }
      else if (vtep_addr_unref (&vxm->vtep_table,
                                t->encap_fib_index, &t->dst) == 0)
        {
          mcast_shared_remove (&t->dst);
        }

      vnet_hw_interface_t *hw = vnet_get_hw_interface (vnm, t->hw_if_index);
      if (hw->dev_class_index == vxlan_device_class.index)
        vnet_delete_hw_interface (vnm, t->hw_if_index);
      else
        ethernet_delete_interface (vnm, t->hw_if_index);
      hash_unset (vxm->instance_used, t->user_instance);

      fib_node_deinit (&t->node);
      pool_put (vxm->tunnels, t);
    }

  if (sw_if_indexp)
    *sw_if_indexp = sw_if_index;

  if (a->is_add)
    {
      /* register udp ports */
      if (!is_ip6 && !udp_is_valid_dst_port (a->src_port, 1))
        udp_register_dst_port (vxm->vlib_main, a->src_port,
                               vxlan4_input_node.index, 1);
      if (is_ip6 && !udp_is_valid_dst_port (a->src_port, 0))
        udp_register_dst_port (vxm->vlib_main, a->src_port,
                               vxlan6_input_node.index, 0);
    }

  return 0;
}

static uword
get_decap_next_for_node (u32 node_index, u32 ipv4_set)
{
  vxlan_main_t *vxm = &vxlan_main;
  vlib_main_t *vm = vxm->vlib_main;
  uword input_node = (ipv4_set) ? vxlan4_input_node.index :
    vxlan6_input_node.index;

  return vlib_node_add_next (vm, input_node, node_index);
}

static uword
unformat_decap_next (unformat_input_t * input, va_list * args)
{
  u32 *result = va_arg (*args, u32 *);
  u32 ipv4_set = va_arg (*args, int);
  vxlan_main_t *vxm = &vxlan_main;
  vlib_main_t *vm = vxm->vlib_main;
  u32 node_index;
  u32 tmp;

  if (unformat (input, "l2"))
    *result = VXLAN_INPUT_NEXT_L2_INPUT;
  else if (unformat (input, "node %U", unformat_vlib_node, vm, &node_index))
    *result = get_decap_next_for_node (node_index, ipv4_set);
  else if (unformat (input, "%d", &tmp))
    *result = tmp;
  else
    return 0;
  return 1;
}

static clib_error_t *
vxlan_add_del_tunnel_command_fn (vlib_main_t * vm,
                                 unformat_input_t * input,
                                 vlib_cli_command_t * cmd)
{
  unformat_input_t _line_input, *line_input = &_line_input;
  ip46_address_t src = ip46_address_initializer, dst =
    ip46_address_initializer;
  u8 is_add = 1;
  u8 src_set = 0;
  u8 dst_set = 0;
  u8 grp_set = 0;
  u8 ipv4_set = 0;
  u8 ipv6_set = 0;
  u8 is_l3 = 0;
  u32 instance = ~0;
  u32 encap_fib_index = 0;
  u32 mcast_sw_if_index = ~0;
  u32 decap_next_index = VXLAN_INPUT_NEXT_L2_INPUT;
  u32 vni = 0;
  u32 src_port = 0;
  u32 dst_port = 0;
  u32 table_id;
  clib_error_t *parse_error = NULL;

  /* Get a line of input. */
  if (!unformat_user (input, unformat_line_input, line_input))
    return 0;

  while (unformat_check_input (line_input) != UNFORMAT_END_OF_INPUT)
    {
      if (unformat (line_input, "del"))
        {
          is_add = 0;
        }
      else if (unformat (line_input, "instance %d", &instance))
        ;
      else if (unformat (line_input, "src %U",
                         unformat_ip46_address, &src, IP46_TYPE_ANY))
        {
          src_set = 1;
          ip46_address_is_ip4 (&src) ? (ipv4_set = 1) : (ipv6_set = 1);
        }
      else if (unformat (line_input, "dst %U",
                         unformat_ip46_address, &dst, IP46_TYPE_ANY))
        {
          dst_set = 1;
          ip46_address_is_ip4 (&dst) ? (ipv4_set = 1) : (ipv6_set = 1);
        }
      else if (unformat (line_input, "group %U %U",
                         unformat_ip46_address, &dst, IP46_TYPE_ANY,
                         unformat_vnet_sw_interface,
                         vnet_get_main (), &mcast_sw_if_index))
        {
          grp_set = dst_set = 1;
          ip46_address_is_ip4 (&dst) ? (ipv4_set = 1) : (ipv6_set = 1);
        }
      else if (unformat (line_input, "encap-vrf-id %d", &table_id))
        {
          encap_fib_index =
            fib_table_find (fib_ip_proto (ipv6_set), table_id);
        }
      else if (unformat (line_input, "l3"))
        is_l3 = 1;
      else if (unformat (line_input, "decap-next %U", unformat_decap_next,
                         &decap_next_index, ipv4_set))
        ;
      else if (unformat (line_input, "vni %d", &vni))
        ;
      else if (unformat (line_input, "src_port %d", &src_port))
        ;
      else if (unformat (line_input, "dst_port %d", &dst_port))
        ;
      else
        {
          parse_error = clib_error_return (0, "parse error: '%U'",
                                           format_unformat_error, line_input);
          break;
        }
    }

  unformat_free (line_input);

  if (parse_error)
    return parse_error;

  if (is_l3 && decap_next_index == VXLAN_INPUT_NEXT_L2_INPUT)
    {
      vlib_node_t *node = vlib_get_node_by_name (
        vm, (u8 *) (ipv4_set ? "ip4-input" : "ip6-input"));
      decap_next_index = get_decap_next_for_node (node->index, ipv4_set);
    }

  if (encap_fib_index == ~0)
    return clib_error_return (0, "nonexistent encap-vrf-id %d", table_id);

  if (src_set == 0)
    return clib_error_return (0, "tunnel src address not specified");

  if (dst_set == 0)
    return clib_error_return (0, "tunnel dst address not specified");

  if (grp_set && !ip46_address_is_multicast (&dst))
    return clib_error_return (0, "tunnel group address not multicast");

  if (grp_set == 0 && ip46_address_is_multicast (&dst))
    return clib_error_return (0, "dst address must be unicast");

  if (grp_set && mcast_sw_if_index == ~0)
    return clib_error_return (0, "tunnel nonexistent multicast device");

  if (ipv4_set && ipv6_set)
    return clib_error_return (0, "both IPv4 and IPv6 addresses specified");

  if (ip46_address_cmp (&src, &dst) == 0)
    return clib_error_return (0, "src and dst addresses are identical");

  if (decap_next_index == ~0)
    return clib_error_return (0, "next node not found");

  if (vni == 0)
    return clib_error_return (0, "vni not specified");

  if (vni >> 24)
    return clib_error_return (0, "vni %d out of range", vni);

  vnet_vxlan_add_del_tunnel_args_t a = { .is_add = is_add,
                                         .is_ip6 = ipv6_set,
                                         .is_l3 = is_l3,
                                         .instance = instance,
#define _(x) .x = x,
                                         foreach_copy_field
#undef _
  };

  u32 tunnel_sw_if_index;
  int rv = vnet_vxlan_add_del_tunnel (&a, &tunnel_sw_if_index);

  switch (rv)
    {
    case 0:
      if (is_add)
        vlib_cli_output (vm, "%U\n", format_vnet_sw_if_index_name,
                         vnet_get_main (), tunnel_sw_if_index);
      break;

    case VNET_API_ERROR_TUNNEL_EXIST:
      return clib_error_return (0, "tunnel already exists...");

    case VNET_API_ERROR_NO_SUCH_ENTRY:
      return clib_error_return (0, "tunnel does not exist...");

    case VNET_API_ERROR_INSTANCE_IN_USE:
      return clib_error_return (0, "Instance is in use");

    default:
      return clib_error_return
        (0, "vnet_vxlan_add_del_tunnel returned %d", rv);
    }

  return 0;
}

/*?
 * Add or delete a VXLAN Tunnel.
 *
 * VXLAN provides the features needed to allow L2 bridge domains (BDs)
 * to span multiple servers. This is done by building an L2 overlay on
 * top of an L3 network underlay using VXLAN tunnels.
 *
 * This makes it possible for servers to be co-located in the same data
 * center or be separated geographically as long as they are reachable
 * through the underlay L3 network.
 *
 * You can refer to this kind of L2 overlay bridge domain as a VXLAN
 * (Virtual eXtensible VLAN) segment.
 *
 * @cliexpar
 * Example of how to create a VXLAN Tunnel:
 * @cliexcmd{create vxlan tunnel src 10.0.3.1 dst 10.0.3.3 vni 13 encap-vrf-id
 7}
 * Example of how to create a VXLAN Tunnel with a known name, vxlan_tunnel42:
 * @cliexcmd{create vxlan tunnel src 10.0.3.1 dst 10.0.3.3 instance 42}
 * Example of how to create a multicast VXLAN Tunnel with a known name,
 vxlan_tunnel23:
 * @cliexcmd{create vxlan tunnel src 10.0.3.1 group 239.1.1.1
 GigabitEthernet0/8/0 instance 23}
 * Example of how to create a VXLAN Tunnel with custom udp-ports:
 * @cliexcmd{create vxlan tunnel src 10.0.3.1 dst 10.0.3.3 vni 13 src_port
 59000 dst_port 59001}
 * Example of how to delete a VXLAN Tunnel:
 * @cliexcmd{create vxlan tunnel src 10.0.3.1 dst 10.0.3.3 vni 13 del}
 ?*/
/* *INDENT-OFF* */
VLIB_CLI_COMMAND (create_vxlan_tunnel_command, static) = {
  .path = "create vxlan tunnel",
  .short_help =
    "create vxlan tunnel src <local-vtep-addr>"
    " {dst <remote-vtep-addr>|group <mcast-vtep-addr> <intf-name>} vni <nn>"
    " [instance <id>]"
    " [encap-vrf-id <nn>] [decap-next [l2|node <name>]] [del] [l3]"
    " [src_port <local-vtep-udp-port>] [dst_port <remote-vtep-udp-port>]",
  .function = vxlan_add_del_tunnel_command_fn,
};
/* *INDENT-ON* */

static clib_error_t *
show_vxlan_tunnel_command_fn (vlib_main_t * vm,
                              unformat_input_t * input,
                              vlib_cli_command_t * cmd)
{
  vxlan_main_t *vxm = &vxlan_main;
  vxlan_tunnel_t *t;
  int raw = 0;

  while (unformat_check_input (input) != UNFORMAT_END_OF_INPUT)
    {
      if (unformat (input, "raw"))
        raw = 1;
      else
        return clib_error_return (0, "parse error: '%U'",
                                  format_unformat_error, input);
    }

  if (pool_elts (vxm->tunnels) == 0)
    vlib_cli_output (vm, "No vxlan tunnels configured...");

/* *INDENT-OFF* */
  pool_foreach (t, vxm->tunnels)
   {
    vlib_cli_output (vm, "%U", format_vxlan_tunnel, t);
  }
/* *INDENT-ON* */

  if (raw)
    {
      vlib_cli_output (vm, "Raw IPv4 Hash Table:\n%U\n",
                       format_bihash_16_8, &vxm->vxlan4_tunnel_by_key,
                       1 /* verbose */ );
      vlib_cli_output (vm, "Raw IPv6 Hash Table:\n%U\n",
                       format_bihash_24_8, &vxm->vxlan6_tunnel_by_key,
                       1 /* verbose */ );
    }

  return 0;
}

/*?
 * Display all the VXLAN Tunnel entries.
 *
 * @cliexpar
 * Example of how to display the VXLAN Tunnel entries:
 * @cliexstart{show vxlan tunnel}
 * [0] src 10.0.3.1 dst 10.0.3.3 src_port 4789 dst_port 4789 vni 13
 encap_fib_index 0 sw_if_index 5 decap_next l2
 * @cliexend
 ?*/
/* *INDENT-OFF* */
VLIB_CLI_COMMAND (show_vxlan_tunnel_command, static) = {
    .path = "show vxlan tunnel",
    .short_help = "show vxlan tunnel [raw]",
    .function = show_vxlan_tunnel_command_fn,
};
/* *INDENT-ON* */


void
vnet_int_vxlan_bypass_mode (u32 sw_if_index, u8 is_ip6, u8 is_enable)
{
  vxlan_main_t *vxm = &vxlan_main;

  if (pool_is_free_index (vxm->vnet_main->interface_main.sw_interfaces,
                          sw_if_index))
    return;

  is_enable = ! !is_enable;

  if (is_ip6)
    {
      if (clib_bitmap_get (vxm->bm_ip6_bypass_enabled_by_sw_if, sw_if_index)
          != is_enable)
        {
          vnet_feature_enable_disable ("ip6-unicast", "ip6-vxlan-bypass",
                                       sw_if_index, is_enable, 0, 0);
          vxm->bm_ip6_bypass_enabled_by_sw_if =
            clib_bitmap_set (vxm->bm_ip6_bypass_enabled_by_sw_if,
                             sw_if_index, is_enable);
        }
    }
  else
    {
      if (clib_bitmap_get (vxm->bm_ip4_bypass_enabled_by_sw_if, sw_if_index)
          != is_enable)
        {
          vnet_feature_enable_disable ("ip4-unicast", "ip4-vxlan-bypass",
                                       sw_if_index, is_enable, 0, 0);
          vxm->bm_ip4_bypass_enabled_by_sw_if =
            clib_bitmap_set (vxm->bm_ip4_bypass_enabled_by_sw_if,
                             sw_if_index, is_enable);
        }
    }
}


static clib_error_t *
set_ip_vxlan_bypass (u32 is_ip6,
                     unformat_input_t * input, vlib_cli_command_t * cmd)
{
  unformat_input_t _line_input, *line_input = &_line_input;
  vnet_main_t *vnm = vnet_get_main ();
  clib_error_t *error = 0;
  u32 sw_if_index, is_enable;

  sw_if_index = ~0;
  is_enable = 1;

  if (!unformat_user (input, unformat_line_input, line_input))
    return 0;

  while (unformat_check_input (line_input) != UNFORMAT_END_OF_INPUT)
    {
      if (unformat_user
          (line_input, unformat_vnet_sw_interface, vnm, &sw_if_index))
        ;
      else if (unformat (line_input, "del"))
        is_enable = 0;
      else
        {
          error = unformat_parse_error (line_input);
          goto done;
        }
    }

  if (~0 == sw_if_index)
    {
      error = clib_error_return (0, "unknown interface `%U'",
                                 format_unformat_error, line_input);
      goto done;
    }

  vnet_int_vxlan_bypass_mode (sw_if_index, is_ip6, is_enable);

done:
  unformat_free (line_input);

  return error;
}

static clib_error_t *
set_ip4_vxlan_bypass (vlib_main_t * vm,
                      unformat_input_t * input, vlib_cli_command_t * cmd)
{
  return set_ip_vxlan_bypass (0, input, cmd);
}

/*?
 * This command adds the 'ip4-vxlan-bypass' graph node for a given interface.
 * By adding the IPv4 vxlan-bypass graph node to an interface, the node checks
 *  for and validate input vxlan packet and bypass ip4-lookup, ip4-local,
 * ip4-udp-lookup nodes to speedup vxlan packet forwarding. This node will
 * cause extra overhead to for non-vxlan packets which is kept at a minimum.
 *
 * @cliexpar
 * @parblock
 * Example of graph node before ip4-vxlan-bypass is enabled:
 * @cliexstart{show vlib graph ip4-vxlan-bypass}
 *            Name                      Next                    Previous
 * ip4-vxlan-bypass                error-drop [0]
 *                                vxlan4-input [1]
 *                                 ip4-lookup [2]
 * @cliexend
 *
 * Example of how to enable ip4-vxlan-bypass on an interface:
 * @cliexcmd{set interface ip vxlan-bypass GigabitEthernet2/0/0}
 *
 * Example of graph node after ip4-vxlan-bypass is enabled:
 * @cliexstart{show vlib graph ip4-vxlan-bypass}
 *            Name                      Next                    Previous
 * ip4-vxlan-bypass                error-drop [0]               ip4-input
 *                                vxlan4-input [1]        ip4-input-no-checksum
 *                                 ip4-lookup [2]
 * @cliexend
 *
 * Example of how to display the feature enabled on an interface:
 * @cliexstart{show ip interface features GigabitEthernet2/0/0}
 * IP feature paths configured on GigabitEthernet2/0/0...
 * ...
 * ipv4 unicast:
 *   ip4-vxlan-bypass
 *   ip4-lookup
 * ...
 * @cliexend
 *
 * Example of how to disable ip4-vxlan-bypass on an interface:
 * @cliexcmd{set interface ip vxlan-bypass GigabitEthernet2/0/0 del}
 * @endparblock
?*/
/* *INDENT-OFF* */
VLIB_CLI_COMMAND (set_interface_ip_vxlan_bypass_command, static) = {
  .path = "set interface ip vxlan-bypass",
  .function = set_ip4_vxlan_bypass,
  .short_help = "set interface ip vxlan-bypass <interface> [del]",
};
/* *INDENT-ON* */

static clib_error_t *
set_ip6_vxlan_bypass (vlib_main_t * vm,
                      unformat_input_t * input, vlib_cli_command_t * cmd)
{
  return set_ip_vxlan_bypass (1, input, cmd);
}

/*?
 * This command adds the 'ip6-vxlan-bypass' graph node for a given interface.
 * By adding the IPv6 vxlan-bypass graph node to an interface, the node checks
 *  for and validate input vxlan packet and bypass ip6-lookup, ip6-local,
 * ip6-udp-lookup nodes to speedup vxlan packet forwarding. This node will
 * cause extra overhead to for non-vxlan packets which is kept at a minimum.
 *
 * @cliexpar
 * @parblock
 * Example of graph node before ip6-vxlan-bypass is enabled:
 * @cliexstart{show vlib graph ip6-vxlan-bypass}
 *            Name                      Next                    Previous
 * ip6-vxlan-bypass                error-drop [0]
 *                                vxlan6-input [1]
 *                                 ip6-lookup [2]
 * @cliexend
 *
 * Example of how to enable ip6-vxlan-bypass on an interface:
 * @cliexcmd{set interface ip6 vxlan-bypass GigabitEthernet2/0/0}
 *
 * Example of graph node after ip6-vxlan-bypass is enabled:
 * @cliexstart{show vlib graph ip6-vxlan-bypass}
 *            Name                      Next                    Previous
 * ip6-vxlan-bypass                error-drop [0]               ip6-input
 *                                vxlan6-input [1]        ip4-input-no-checksum
 *                                 ip6-lookup [2]
 * @cliexend
 *
 * Example of how to display the feature enabled on an interface:
 * @cliexstart{show ip interface features GigabitEthernet2/0/0}
 * IP feature paths configured on GigabitEthernet2/0/0...
 * ...
 * ipv6 unicast:
 *   ip6-vxlan-bypass
 *   ip6-lookup
 * ...
 * @cliexend
 *
 * Example of how to disable ip6-vxlan-bypass on an interface:
 * @cliexcmd{set interface ip6 vxlan-bypass GigabitEthernet2/0/0 del}
 * @endparblock
?*/
/* *INDENT-OFF* */
VLIB_CLI_COMMAND (set_interface_ip6_vxlan_bypass_command, static) = {
  .path = "set interface ip6 vxlan-bypass",
  .function = set_ip6_vxlan_bypass,
  .short_help = "set interface ip6 vxlan-bypass <interface> [del]",
};
/* *INDENT-ON* */

int
vnet_vxlan_add_del_rx_flow (u32 hw_if_index, u32 t_index, int is_add)
{
  vxlan_main_t *vxm = &vxlan_main;
  vxlan_tunnel_t *t = pool_elt_at_index (vxm->tunnels, t_index);
  vnet_main_t *vnm = vnet_get_main ();
  if (is_add)
    {
      if (t->flow_index == ~0)
        {
          vxlan_main_t *vxm = &vxlan_main;
          vnet_flow_t flow = {
            .actions =
              VNET_FLOW_ACTION_REDIRECT_TO_NODE | VNET_FLOW_ACTION_MARK |
              VNET_FLOW_ACTION_BUFFER_ADVANCE,
            .mark_flow_id = t->dev_instance + vxm->flow_id_start,
            .redirect_node_index = vxlan4_flow_input_node.index,
            .buffer_advance = sizeof (ethernet_header_t),
            .type = VNET_FLOW_TYPE_IP4_VXLAN,
            .ip4_vxlan = {
                          .protocol.prot = IP_PROTOCOL_UDP,
                          .src_addr.addr = t->dst.ip4,
                          .dst_addr.addr = t->src.ip4,
                          .src_addr.mask.as_u32 = ~0,
                          .dst_addr.mask.as_u32 = ~0,
                          .dst_port.port = t->src_port,
                          .dst_port.mask = 0xFF,
                          .vni = t->vni,
                          }
            ,
          };
          vnet_flow_add (vnm, &flow, &t->flow_index);
        }
      return vnet_flow_enable (vnm, t->flow_index, hw_if_index);
    }
  /* flow index is removed when the tunnel is deleted */
  return vnet_flow_disable (vnm, t->flow_index, hw_if_index);
}

u32
vnet_vxlan_get_tunnel_index (u32 sw_if_index)
{
  vxlan_main_t *vxm = &vxlan_main;

  if (sw_if_index >= vec_len (vxm->tunnel_index_by_sw_if_index))
    return ~0;
  return vxm->tunnel_index_by_sw_if_index[sw_if_index];
}

static clib_error_t *
vxlan_offload_command_fn (vlib_main_t * vm,
                          unformat_input_t * input, vlib_cli_command_t * cmd)
{
  unformat_input_t _line_input, *line_input = &_line_input;

  /* Get a line of input. */
  if (!unformat_user (input, unformat_line_input, line_input))
    return 0;

  vnet_main_t *vnm = vnet_get_main ();
  u32 rx_sw_if_index = ~0;
  u32 hw_if_index = ~0;
  int is_add = 1;

  while (unformat_check_input (line_input) != UNFORMAT_END_OF_INPUT)
    {
      if (unformat (line_input, "hw %U", unformat_vnet_hw_interface, vnm,
                    &hw_if_index))
        continue;
      if (unformat (line_input, "rx %U", unformat_vnet_sw_interface, vnm,
                    &rx_sw_if_index))
        continue;
      if (unformat (line_input, "del"))
        {
          is_add = 0;
          continue;
        }
      return clib_error_return (0, "unknown input `%U'",
                                format_unformat_error, line_input);
    }

  if (rx_sw_if_index == ~0)
    return clib_error_return (0, "missing rx interface");
  if (hw_if_index == ~0)
    return clib_error_return (0, "missing hw interface");

  u32 t_index = vnet_vxlan_get_tunnel_index (rx_sw_if_index);;
  if (t_index == ~0)
    return clib_error_return (0, "%U is not a vxlan tunnel",
                              format_vnet_sw_if_index_name, vnm,
                              rx_sw_if_index);

  vxlan_main_t *vxm = &vxlan_main;
  vxlan_tunnel_t *t = pool_elt_at_index (vxm->tunnels, t_index);

  if (!ip46_address_is_ip4 (&t->dst))
    return clib_error_return (0, "currently only IPV4 tunnels are supported");

  vnet_hw_interface_t *hw_if = vnet_get_hw_interface (vnm, hw_if_index);
  ip4_main_t *im = &ip4_main;
  u32 rx_fib_index =
    vec_elt (im->fib_index_by_sw_if_index, hw_if->sw_if_index);

  if (t->encap_fib_index != rx_fib_index)
    return clib_error_return (0, "interface/tunnel fib mismatch");

  if (vnet_vxlan_add_del_rx_flow (hw_if_index, t_index, is_add))
    return clib_error_return (0, "error %s flow",
                              is_add ? "enabling" : "disabling");

  return 0;
}

/* *INDENT-OFF* */
VLIB_CLI_COMMAND (vxlan_offload_command, static) = {
    .path = "set flow-offload vxlan",
    .short_help =
    "set flow-offload vxlan hw <interface-name> rx <tunnel-name> [del]",
    .function = vxlan_offload_command_fn,
};
/* *INDENT-ON* */

#define VXLAN_HASH_NUM_BUCKETS (2 * 1024)
#define VXLAN_HASH_MEMORY_SIZE (1 << 20)

clib_error_t *
vxlan_init (vlib_main_t * vm)
{
  vxlan_main_t *vxm = &vxlan_main;

  vxm->vnet_main = vnet_get_main ();
  vxm->vlib_main = vm;

  vnet_flow_get_range (vxm->vnet_main, "vxlan", 1024 * 1024,
                       &vxm->flow_id_start);

  vxm->bm_ip4_bypass_enabled_by_sw_if = 0;
  vxm->bm_ip6_bypass_enabled_by_sw_if = 0;

  /* initialize the ip6 hash */
  clib_bihash_init_16_8 (&vxm->vxlan4_tunnel_by_key, "vxlan4",
                         VXLAN_HASH_NUM_BUCKETS, VXLAN_HASH_MEMORY_SIZE);
  clib_bihash_init_24_8 (&vxm->vxlan6_tunnel_by_key, "vxlan6",
                         VXLAN_HASH_NUM_BUCKETS, VXLAN_HASH_MEMORY_SIZE);
  vxm->vtep_table = vtep_table_create ();
  vxm->mcast_shared = hash_create_mem (0,
                                       sizeof (ip46_address_t),
                                       sizeof (mcast_shared_t));

  fib_node_register_type (FIB_NODE_TYPE_VXLAN_TUNNEL, &vxlan_vft);

  return 0;
}

VLIB_INIT_FUNCTION (vxlan_init);

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