/* * 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. */ /* * ip/ip4_forward.c: IP v4 forwarding * * Copyright (c) 2008 Eliot Dresselhaus * * Permission is hereby granted, free of charge, to any person obtaining * a copy of this software and associated documentation files (the * "Software"), to deal in the Software without restriction, including * without limitation the rights to use, copy, modify, merge, publish, * distribute, sublicense, and/or sell copies of the Software, and to * permit persons to whom the Software is furnished to do so, subject to * the following conditions: * * The above copyright notice and this permission notice shall be * included in all copies or substantial portions of the Software. * * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, * EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND * NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE * LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION * OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION * WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. */ #include #include #include #include /* for ethernet_header_t */ #include /* for ethernet_arp_header_t */ #include #include /* for srp_hw_interface_class */ #include /* for API error numbers */ #include /* for FIB table and entry creation */ #include /* for FIB table and entry creation */ #include /* for FIB uRPF check */ #include #include #include #include #include /* for mFIB table and entry creation */ #include #include /** @brief IPv4 lookup node. @node ip4-lookup This is the main IPv4 lookup dispatch node. @param vm vlib_main_t corresponding to the current thread @param node vlib_node_runtime_t @param frame vlib_frame_t whose contents should be dispatched @par Graph mechanics: buffer metadata, next index usage @em Uses: - vnet_buffer(b)->sw_if_index[VLIB_RX] - Indicates the @c sw_if_index value of the interface that the packet was received on. - vnet_buffer(b)->sw_if_index[VLIB_TX] - When the value is @c ~0 then the node performs a longest prefix match (LPM) for the packet destination address in the FIB attached to the receive interface. - Otherwise perform LPM for the packet destination address in the indicated FIB. In this case [VLIB_TX] is a FIB index value (0, 1, ...) and not a VRF id. @em Sets: - vnet_buffer(b)->ip.adj_index[VLIB_TX] - The lookup result adjacency index. Next Index: - Dispatches the packet to the node index found in ip_adjacency_t @c adj->lookup_next_index (where @c adj is the lookup result adjacency). */ VLIB_NODE_FN (ip4_lookup_node) (vlib_main_t * vm, vlib_node_runtime_t * node, vlib_frame_t * frame) { return ip4_lookup_inline (vm, node, frame); } static u8 *format_ip4_lookup_trace (u8 * s, va_list * args); /* *INDENT-OFF* */ VLIB_REGISTER_NODE (ip4_lookup_node) = { .name = "ip4-lookup", .vector_size = sizeof (u32), .format_trace = format_ip4_lookup_trace, .n_next_nodes = IP_LOOKUP_N_NEXT, .next_nodes = IP4_LOOKUP_NEXT_NODES, }; /* *INDENT-ON* */ VLIB_NODE_FN (ip4_load_balance_node) (vlib_main_t * vm, vlib_node_runtime_t * node, vlib_frame_t * frame) { vlib_combined_counter_main_t *cm = &load_balance_main.lbm_via_counters; u32 n_left, *from; u32 thread_index = vm->thread_index; vlib_buffer_t *bufs[VLIB_FRAME_SIZE], **b = bufs; u16 nexts[VLIB_FRAME_SIZE], *next; from = vlib_frame_vector_args (frame); n_left = frame->n_vectors; next = nexts; vlib_get_buffers (vm, from, bufs, n_left); while (n_left >= 4) { const load_balance_t *lb0, *lb1; const ip4_header_t *ip0, *ip1; u32 lbi0, hc0, lbi1, hc1; const dpo_id_t *dpo0, *dpo1; /* Prefetch next iteration. */ { vlib_prefetch_buffer_header (b[2], LOAD); vlib_prefetch_buffer_header (b[3], LOAD); CLIB_PREFETCH (b[2]->data, sizeof (ip0[0]), LOAD); CLIB_PREFETCH (b[3]->data, sizeof (ip0[0]), LOAD); } ip0 = vlib_buffer_get_current (b[0]); ip1 = vlib_buffer_get_current (b[1]); lbi0 = vnet_buffer (b[0])->ip.adj_index[VLIB_TX]; lbi1 = vnet_buffer (b[1])->ip.adj_index[VLIB_TX]; lb0 = load_balance_get (lbi0); lb1 = load_balance_get (lbi1); /* * this node is for via FIBs we can re-use the hash value from the * to node if present. * We don't want to use the same hash value at each level in the recursion * graph as that would lead to polarisation */ hc0 = hc1 = 0; if (PREDICT_FALSE (lb0->lb_n_buckets > 1)) { if (PREDICT_TRUE (vnet_buffer (b[0])->ip.flow_hash)) { hc0 = vnet_buffer (b[0])->ip.flow_hash = vnet_buffer (b[0])->ip.flow_hash >> 1; } else { hc0 = vnet_buffer (b[0])->ip.flow_hash = ip4_compute_flow_hash (ip0, lb0->lb_hash_config); } dpo0 = load_balance_get_fwd_bucket (lb0, (hc0 & (lb0->lb_n_buckets_minus_1))); } else { dpo0 = load_balance_get_bucket_i (lb0, 0); } if (PREDICT_FALSE (lb1->lb_n_buckets > 1)) { if (PREDICT_TRUE (vnet_buffer (b[1])->ip.flow_hash)) { hc1 = vnet_buffer (b[1])->ip.flow_hash = vnet_buffer (b[1])->ip.flow_hash >> 1; } else { hc1 = vnet_buffer (b[1])->ip.flow_hash = ip4_compute_flow_hash (ip1, lb1->lb_hash_config); } dpo1 = load_balance_get_fwd_bucket (lb1, (hc1 & (lb1->lb_n_buckets_minus_1))); } else { dpo1 = load_balance_get_bucket_i (lb1, 0); } next[0] = dpo0->dpoi_next_node; next[1] = dpo1->dpoi_next_node; vnet_buffer (b[0])->ip.adj_index[VLIB_TX] = dpo0->dpoi_index; vnet_buffer (b[1])->ip.adj_index[VLIB_TX] = dpo1->dpoi_index; vlib_increment_combined_counter (cm, thread_index, lbi0, 1, vlib_buffer_length_in_chain (vm, b[0])); vlib_increment_combined_counter (cm, thread_index, lbi1, 1, vlib_buffer_length_in_chain (vm, b[1])); b += 2; next += 2; n_left -= 2; } while (n_left > 0) { const load_balance_t *lb0; const ip4_header_t *ip0; const dpo_id_t *dpo0; u32 lbi0, hc0; ip0 = vlib_buffer_get_current (b[0]); lbi0 = vnet_buffer (b[0])->ip.adj_index[VLIB_TX]; lb0 = load_balance_get (lbi0); hc0 = 0; if (PREDICT_FALSE (lb0->lb_n_buckets > 1)) { if (PREDICT_TRUE (vnet_buffer (b[0])->ip.flow_hash)) { hc0 = vnet_buffer (b[0])->ip.flow_hash = vnet_buffer (b[0])->ip.flow_hash >> 1; } else { hc0 = vnet_buffer (b[0])->ip.flow_hash = ip4_compute_flow_hash (ip0, lb0->lb_hash_config); } dpo0 = load_balance_get_fwd_bucket (lb0, (hc0 & (lb0->lb_n_buckets_minus_1))); } else { dpo0 = load_balance_get_bucket_i (lb0, 0); } next[0] = dpo0->dpoi_next_node; vnet_buffer (b[0])->ip.adj_index[VLIB_TX] = dpo0->dpoi_index; vlib_increment_combined_counter (cm, thread_index, lbi0, 1, vlib_buffer_length_in_chain (vm, b[0])); b += 1; next += 1; n_left -= 1; } vlib_buffer_enqueue_to_next (vm, node, from, nexts, frame->n_vectors); if (node->flags & VLIB_NODE_FLAG_TRACE) ip4_forward_next_trace (vm, node, frame, VLIB_TX); return frame->n_vectors; } /* *INDENT-OFF* */ VLIB_REGISTER_NODE (ip4_load_balance_node) = { .name = "ip4-load-balance", .vector_size = sizeof (u32), .sibling_of = "ip4-lookup", .format_trace = format_ip4_lookup_trace, }; /* *INDENT-ON* */ #ifndef CLIB_MARCH_VARIANT /* get first interface address */ ip4_address_t * ip4_interface_first_address (ip4_main_t * im, u32 sw_if_index, ip_interface_address_t ** result_ia) { ip_lookup_main_t *lm = &im->lookup_main; ip_interface_address_t *ia = 0; ip4_address_t *result = 0; /* *INDENT-OFF* */ foreach_ip_interface_address (lm, ia, sw_if_index, 1 /* honor unnumbered */ , ({ ip4_address_t * a = ip_interface_address_get_address (lm, ia); result = a; break; })); /* *INDENT-OFF* */ if (result_ia) *result_ia = result ? ia : 0; return result; } #endif static void ip4_add_subnet_bcast_route (u32 fib_index, fib_prefix_t *pfx, u32 sw_if_index) { vnet_sw_interface_flags_t iflags; iflags = vnet_sw_interface_get_flags(vnet_get_main(), sw_if_index); fib_table_entry_special_remove(fib_index, pfx, FIB_SOURCE_INTERFACE); if (iflags & VNET_SW_INTERFACE_FLAG_DIRECTED_BCAST) { fib_table_entry_update_one_path (fib_index, pfx, FIB_SOURCE_INTERFACE, FIB_ENTRY_FLAG_NONE, DPO_PROTO_IP4, /* No next-hop address */ &ADJ_BCAST_ADDR, sw_if_index, // invalid FIB index ~0, 1, // no out-label stack NULL, FIB_ROUTE_PATH_FLAG_NONE); } else { fib_table_entry_special_add(fib_index, pfx, FIB_SOURCE_INTERFACE, (FIB_ENTRY_FLAG_DROP | FIB_ENTRY_FLAG_LOOSE_URPF_EXEMPT)); } } static void ip4_add_interface_prefix_routes (ip4_main_t *im, u32 sw_if_index, u32 fib_index, ip_interface_address_t * a) { ip_lookup_main_t *lm = &im->lookup_main; ip_interface_prefix_t *if_prefix; ip4_address_t *address = ip_interface_address_get_address (lm, a); ip_interface_prefix_key_t key = { .prefix = { .fp_len = a->address_length, .fp_proto = FIB_PROTOCOL_IP4, .fp_addr.ip4.as_u32 = address->as_u32 & im->fib_masks[a->address_length], }, .sw_if_index = sw_if_index, }; fib_prefix_t pfx_special = { .fp_proto = FIB_PROTOCOL_IP4, }; /* If prefix already set on interface, just increment ref count & return */ if_prefix = ip_get_interface_prefix (lm, &key); if (if_prefix) { if_prefix->ref_count += 1; return; } /* New prefix - allocate a pool entry, initialize it, add to the hash */ pool_get (lm->if_prefix_pool, if_prefix); if_prefix->ref_count = 1; if_prefix->src_ia_index = a - lm->if_address_pool; clib_memcpy (&if_prefix->key, &key, sizeof (key)); mhash_set (&lm->prefix_to_if_prefix_index, &key, if_prefix - lm->if_prefix_pool, 0 /* old value */); /* length <= 30 - add glean, drop first address, maybe drop bcast address */ if (a->address_length <= 30) { pfx_special.fp_len = a->address_length; pfx_special.fp_addr.ip4.as_u32 = address->as_u32; /* set the glean route for the prefix */ fib_table_entry_update_one_path (fib_index, &pfx_special, FIB_SOURCE_INTERFACE, (FIB_ENTRY_FLAG_CONNECTED | FIB_ENTRY_FLAG_ATTACHED), DPO_PROTO_IP4, /* No next-hop address */ NULL, sw_if_index, /* invalid FIB index */ ~0, 1, /* no out-label stack */ NULL, FIB_ROUTE_PATH_FLAG_NONE); /* set a drop route for the base address of the prefix */ pfx_special.fp_len = 32; pfx_special.fp_addr.ip4.as_u32 = address->as_u32 & im->fib_masks[a->address_length]; if (pfx_special.fp_addr.ip4.as_u32 != address->as_u32) fib_table_entry_special_add (fib_index, &pfx_special, FIB_SOURCE_INTERFACE, (FIB_ENTRY_FLAG_DROP | FIB_ENTRY_FLAG_LOOSE_URPF_EXEMPT)); /* set a route for the broadcast address of the prefix */ pfx_special.fp_len = 32; pfx_special.fp_addr.ip4.as_u32 = address->as_u32 | ~im->fib_masks[a->address_length]; if (pfx_special.fp_addr.ip4.as_u32 != address->as_u32) ip4_add_subnet_bcast_route (fib_index, &pfx_special, sw_if_index); } /* length == 31 - add an attached route for the other address */ else if (a->address_length == 31) { pfx_special.fp_len = 32; pfx_special.fp_addr.ip4.as_u32 = address->as_u32 ^ clib_host_to_net_u32(1); fib_table_entry_update_one_path (fib_index, &pfx_special, FIB_SOURCE_INTERFACE, (FIB_ENTRY_FLAG_ATTACHED), DPO_PROTO_IP4, &pfx_special.fp_addr, sw_if_index, /* invalid FIB index */ ~0, 1, NULL, FIB_ROUTE_PATH_FLAG_NONE); } } static void ip4_add_interface_routes (u32 sw_if_index, ip4_main_t * im, u32 fib_index, ip_interface_address_t * a) { ip_lookup_main_t *lm = &im->lookup_main; ip4_address_t *address = ip_interface_address_get_address (lm, a); fib_prefix_t pfx = { .fp_len = 32, .fp_proto = FIB_PROTOCOL_IP4, .fp_addr.ip4 = *address, }; /* set special routes for the prefix if needed */ ip4_add_interface_prefix_routes (im, sw_if_index, fib_index, a); if (sw_if_index < vec_len (lm->classify_table_index_by_sw_if_index)) { u32 classify_table_index = lm->classify_table_index_by_sw_if_ind
/*
 * Copyright (c) 2011-2016 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.
 */
/**
 * @file
 * @brief LLDP packet generation implementation
 */
#include <vnet/lldp/lldp_node.h>

static void
lldp_build_mgmt_addr_tlv (u8 ** t0p, u8 subtype, u8 addr_len, u8 * addr,
			  u32 if_index, u8 oid_len, u8 * oid)
{
  lldp_tlv_t *t = (lldp_tlv_t *) * t0p;

  lldp_tlv_set_code (t, LLDP_TLV_NAME (mgmt_addr));
  t->v[0] = addr_len + 1;	/* address string length */
  t->v[1] = subtype;		/* address subtype */
  clib_memcpy (&(t->v[2]), addr, addr_len);	/* address */
  t->v[addr_len + 2] = 2;	/* interface numbering subtype: ifIndex */
  t->v[addr_len + 3] = (if_index >> 24) & 0xFF;	/* interface number */
  t->v[addr_len + 4] = (if_index >> 16) & 0xFF;
  t->v[addr_len + 5] = (if_index >> 8) & 0xFF;
  t->v[addr_len + 6] = (if_index >> 0) & 0xFF;
  t->v[addr_len + 7] = oid_len;	/* OID string length */

  if (oid_len > 0)
    clib_memcpy ((u8 *) & (t->v[addr_len + 8]), oid, oid_len);

  lldp_tlv_set_length (t, addr_len + oid_len + 8);
  *t0p += STRUCT_SIZE_OF (lldp_tlv_t, head) + addr_len + oid_len + 8;
}

static void
lldp_add_chassis_id (const vnet_hw_interface_t * hw, u8 ** t0p)
{
  lldp_chassis_id_tlv_t *t = (lldp_chassis_id_tlv_t *) * t0p;

  lldp_tlv_set_code ((lldp_tlv_t *) t, LLDP_TLV_NAME (chassis_id));
  t->subtype = LLDP_CHASS_ID_SUBTYPE_NAME (mac_addr);

  const size_t addr_len = 6;
  clib_memcpy (&t->id, hw->hw_address, addr_len);
  const size_t len =
    STRUCT_SIZE_OF (lldp_chassis_id_tlv_t, subtype) + addr_len;
  lldp_tlv_set_length ((lldp_tlv_t *) t, len);
  *t0p += STRUCT_SIZE_OF (lldp_tlv_t, head) + len;
}

static void
lldp_add_port_id (const vnet_hw_interface_t * hw, u8 ** t0p)
{
  lldp_port_id_tlv_t *t = (lldp_port_id_tlv_t *) * t0p;

  lldp_tlv_set_code ((lldp_tlv_t *) t, LLDP_TLV_NAME (port_id));
  t->subtype = LLDP_PORT_ID_SUBTYPE_NAME (intf_name);

  const size_t name_len = vec_len (hw->name);
  clib_memcpy (&t->id, hw->name, name_len);
  const size_t len = STRUCT_SIZE_OF (lldp_port_id_tlv_t, subtype) + name_len;
  lldp_tlv_set_length ((lldp_tlv_t *) t, len);
  *t0p += STRUCT_SIZE_OF (lldp_tlv_t, head) + len;
}

static void
lldp_add_ttl (const lldp_main_t * lm, u8 ** t0p, int shutdown)
{
  lldp_ttl_tlv_t *t = (lldp_ttl_tlv_t *) * t0p;
  lldp_tlv_set_code ((lldp_tlv_t *) t, LLDP_TLV_NAME (ttl));
  if (shutdown)
    {
      t->ttl = 0;
    }
  else
    {
      if ((size_t) lm->msg_tx_interval * lm->msg_tx_hold + 1 > (1 << 16) - 1)
	{
	  t->ttl = htons ((1 << 16) - 1);
	}
      else
	{
	  t->ttl = htons (lm->msg_tx_hold * lm->msg_tx_interval + 1);
	}
    }
  const size_t len = STRUCT_SIZE_OF (lldp_ttl_tlv_t, ttl);
  lldp_tlv_set_length ((lldp_tlv_t *) t, len);
  *t0p += STRUCT_SIZE_OF (lldp_tlv_t, head) + len;
}

static void
lldp_add_port_desc (const lldp_main_t * lm, lldp_intf_t * n, u8 ** t0p)
{
  const size_t len = vec_len (n->port_desc);
  if (len)
    {
      lldp_tlv_t *t = (lldp_tlv_t *) * t0p;
      lldp_tlv_set_code (t, LLDP_TLV_NAME (port_desc));
      lldp_tlv_set_length (t, len);
      clib_memcpy (t->v, n->port_desc, len);
      *t0p += STRUCT_SIZE_OF (lldp_tlv_t, head) + len;
    }
}

static void
lldp_add_sys_name (const lldp_main_t * lm, u8 ** t0p)
{
  const size_t len = vec_len (lm->sys_name);
  if (len)
    {
      lldp_tlv_t *t = (lldp_tlv_t *) * t0p;
      lldp_tlv_set_code (t, LLDP_TLV_NAME (sys_name));
      lldp_tlv_set_length (t, len);
      clib_memcpy (t->v, lm->sys_name, len);
      *t0p += STRUCT_SIZE_OF (lldp_tlv_t, head) + len;
    }
}

static void
lldp_add_mgmt_addr (const lldp_intf_t * n, const vnet_hw_interface_t * hw,
		    u8 ** t0p)
{
  const size_t len_ip4 = vec_len (n->mgmt_ip4);
  const size_t len_ip6 = vec_len (n->mgmt_ip6);

  if (!(len_ip4 | len_ip6))
    {
      /*
         If no management address is configured, the interface port's MAC
         addressis sent in one TLV.
       */

      lldp_build_mgmt_addr_tlv (t0p, 1,	/* address subtype: Ipv4 */
				6,	/* address string lenth */
				hw->hw_address,	/* address */
				hw->hw_if_index,	/* if index */
				vec_len (n->mgmt_oid),	/* OID length */
				n->mgmt_oid);	/* OID */
      return;
    }

  if (len_ip4)
    {
      lldp_build_mgmt_addr_tlv (t0p, 1,	/* address subtype: Ipv4 */
				len_ip4,	/* address string lenth */
				n->mgmt_ip4,	/* address */
				hw->hw_if_index,	/* if index */
				vec_len (n->mgmt_oid),	/* OID length */
				n->mgmt_oid);	/* OID */
    }

  if (len_ip6)
    {
      lldp_build_mgmt_addr_tlv (t0p, 2,	/* address subtype: Ipv6 */
				len_ip6,	/* address string lenth */
				n->mgmt_ip6,	/* address */
				hw->hw_if_index,	/* if index */
				vec_len (n->mgmt_oid),	/* OID length */
				n->mgmt_oid);	/* OID */
    }
}

static void
lldp_add_pdu_end (u8 ** t0p)
{
  lldp_tlv_t *t = (lldp_tlv_t *) * t0p;
  lldp_tlv_set_code (t, LLDP_TLV_NAME (pdu_end));
  lldp_tlv_set_length (t, 0);
  *t0p += STRUCT_SIZE_OF (lldp_tlv_t, head);
}

static void
lldp_add_tlvs (lldp_main_t * lm, vnet_hw_interface_t * hw, u8 ** t0p,
	       int shutdown, lldp_intf_t * n)
{
  lldp_add_chassis_id (hw, t0p);
  lldp_add_port_id (hw, t0p);
  lldp_add_ttl (lm, t0p, shutdown);
  lldp_add_port_desc (lm, n, t0p);
  lldp_add_sys_name (lm, t0p);
  lldp_add_mgmt_addr (n, hw, t0p);
  lldp_add_pdu_end (t0p);
}

/*
 * send a lldp pkt on an ethernet interface
 */
void
lldp_send_ethernet (lldp_main_t * lm, lldp_intf_t * n, int shutdown)
{
  u32 *to_next;
  ethernet_header_t *h0;
  vnet_hw_interface_t *hw;
  u32 bi0;
  vlib_buffer_t *b0;
  u8 *t0;
  vlib_frame_t *f;
  vlib_main_t *vm = lm->vlib_main;
  vnet_main_t *vnm = lm->vnet_main;

  /*
   * see lldp_template_init() to understand what's already painted
   * into the buffer by the packet template mechanism
   */
  h0 = vlib_packet_template_get_packet (vm, &lm->packet_template, &bi0);

  if (!h0)
    return;

  /* Add the interface's ethernet source address */
  hw = vnet_get_hw_interface (vnm, n->hw_if_index);

  clib_memcpy (h0->src_address, hw->hw_address, vec_len (hw->hw_address));

  u8 *data = ((u8 *) h0) + sizeof (*h0);
  t0 = data;

  /* add TLVs */
  lldp_add_tlvs (lm, hw, &t0, shutdown, n);

  /* Set the outbound packet length */
  b0 = vlib_get_buffer (vm, bi0);
  b0->current_length = sizeof (*h0) + t0 - data;

  /* And the outbound interface */
  vnet_buffer (b0)->sw_if_index[VLIB_TX] = hw->sw_if_index;

  /* And output the packet on the correct interface */
  f = vlib_get_frame_to_node (vm, hw->output_node_index);
  to_next = vlib_frame_vector_args (f);
  to_next[0] = bi0;
  f->n_vectors = 1;

  vlib_put_frame_to_node (vm, hw->output_node_index, f);
  n->last_sent = vlib_time_now (vm);
}

void
lldp_delete_intf (lldp_main_t * lm, lldp_intf_t * n)
{
  if (n)
    {
      lldp_unschedule_intf (lm, n);
      hash_unset (lm->intf_by_hw_if_index, n->hw_if_index);
      vec_free (n->chassis_id);
      vec_free (n->port_id);
      vec_free (n->port_desc);
      vec_free (n->mgmt_ip4);
      vec_free (n->mgmt_ip6);
      vec_free (n->mgmt_oid);
      pool_put (lm->intfs, n);
    }
}

static clib_error_t *
lldp_template_init (vlib_main_t * vm)
{
  lldp_main_t *lm = &lldp_main;

  /* Create the ethernet lldp packet template */
  {
    ethernet_header_t h;

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

    /*
     * Send to 01:80:C2:00:00:0E - propagation constrained to a single
     * physical link - stopped by all type of bridge
     */
    h.dst_address[0] = 0x01;
    h.dst_address[1] = 0x80;
    h.dst_address[2] = 0xC2;
    /* h.dst_address[3] = 0x00; (memset) */
    /* h.dst_address[4] = 0x00; (memset) */
    h.dst_address[5] = 0x0E;

    /* leave src address blank (fill in at send time) */

    h.type = htons (ETHERNET_TYPE_802_1_LLDP);

    vlib_packet_template_init (vm, &lm->packet_template,
			       /* data */ &h, sizeof (h),
			       /* alloc chunk size */ 8, "lldp-ethernet");
  }

  return 0;
}

VLIB_INIT_FUNCTION (lldp_template_init);

/*
 * fd.io coding-style-patch-verification: ON
 *
 * Local Variables:
 * eval: (c-set-style "gnu")
 * End:
 */
st, }; ai = adj_nbr_add_or_lock (FIB_PROTOCOL_IP4, VNET_LINK_IP4, &nh, sw_if_index); adj = adj_get (ai); /* Peer has been previously resolved, retrieve glean adj instead */ if (adj->lookup_next_index == IP_LOOKUP_NEXT_REWRITE) { if (refresh) unicast_rewrite = 1; else { adj_unlock (ai); ai = adj_glean_add_or_lock (FIB_PROTOCOL_IP4, VNET_LINK_IP4, sw_if_index, &nh); adj = adj_get (ai); } } /* Add encapsulation string for software interface (e.g. ethernet header). */ vnet_rewrite_one_header (adj[0], h, sizeof (ethernet_header_t)); if (unicast_rewrite) { u16 *etype = vlib_buffer_get_current (b) - 2; etype[0] = clib_host_to_net_u16 (ETHERNET_TYPE_ARP); } vlib_buffer_advance (b, -adj->rewrite_header.data_bytes); { vlib_frame_t *f = vlib_get_frame_to_node (vm, hi->output_node_index); u32 *to_next = vlib_frame_vector_args (f); to_next[0] = bi; f->n_vectors = 1; vlib_put_frame_to_node (vm, hi->output_node_index, f); } adj_unlock (ai); return /* no error */ 0; } #endif typedef enum { IP4_REWRITE_NEXT_DROP, IP4_REWRITE_NEXT_ICMP_ERROR, IP4_REWRITE_NEXT_FRAGMENT, IP4_REWRITE_N_NEXT /* Last */ } ip4_rewrite_next_t; /** * This bits of an IPv4 address to mask to construct a multicast * MAC address */ #if CLIB_ARCH_IS_BIG_ENDIAN #define IP4_MCAST_ADDR_MASK 0x007fffff #else #define IP4_MCAST_ADDR_MASK 0xffff7f00 #endif always_inline void ip4_mtu_check (vlib_buffer_t * b, u16 packet_len, u16 adj_packet_bytes, bool df, u16 * next, u32 * error) { if (packet_len > adj_packet_bytes) { *error = IP4_ERROR_MTU_EXCEEDED; if (df) { icmp4_error_set_vnet_buffer (b, ICMP4_destination_unreachable, ICMP4_destination_unreachable_fragmentation_needed_and_dont_fragment_set, adj_packet_bytes); *next = IP4_REWRITE_NEXT_ICMP_ERROR; } else { /* IP fragmentation */ ip_frag_set_vnet_buffer (b, adj_packet_bytes, IP4_FRAG_NEXT_IP4_REWRITE, 0); *next = IP4_REWRITE_NEXT_FRAGMENT; } } } /* Decrement TTL & update checksum. Works either endian, so no need for byte swap. */ static_always_inline void ip4_ttl_and_checksum_check (vlib_buffer_t * b, ip4_header_t * ip, u16 * next, u32 * error) { i32 ttl; u32 checksum; if (PREDICT_FALSE (b->flags & VNET_BUFFER_F_LOCALLY_ORIGINATED)) { b->flags &= ~VNET_BUFFER_F_LOCALLY_ORIGINATED; return; } ttl = ip->ttl; /* Input node should have reject packets with ttl 0. */ ASSERT (ip->ttl > 0); checksum = ip->checksum + clib_host_to_net_u16 (0x0100); checksum += checksum >= 0xffff; ip->checksum = checksum; ttl -= 1; ip->ttl = ttl; /* * If the ttl drops below 1 when forwarding, generate * an ICMP response. */ if (PREDICT_FALSE (ttl <= 0)) { *error = IP4_ERROR_TIME_EXPIRED; vnet_buffer (b)->sw_if_index[VLIB_TX] = (u32) ~ 0; icmp4_error_set_vnet_buffer (b, ICMP4_time_exceeded, ICMP4_time_exceeded_ttl_exceeded_in_transit, 0); *next = IP4_REWRITE_NEXT_ICMP_ERROR; } /* Verify checksum. */ ASSERT ((ip->checksum == ip4_header_checksum (ip)) || (b->flags & VNET_BUFFER_F_OFFLOAD_IP_CKSUM)); } always_inline uword ip4_rewrite_inline_with_gso (vlib_main_t * vm, vlib_node_runtime_t * node, vlib_frame_t * frame, int do_counters, int is_midchain, int is_mcast, int do_gso) { ip_lookup_main_t *lm = &ip4_main.lookup_main; u32 *from = vlib_frame_vector_args (frame); vlib_buffer_t *bufs[VLIB_FRAME_SIZE], **b; u16 nexts[VLIB_FRAME_SIZE], *next; u32 n_left_from; vlib_node_runtime_t *error_node = vlib_node_get_runtime (vm, ip4_input_node.index); n_left_from = frame->n_vectors; u32 thread_index = vm->thread_index; vlib_get_buffers (vm, from, bufs, n_left_from); clib_memset_u16 (nexts, IP4_REWRITE_NEXT_DROP, n_left_from); if (n_left_from >= 6) { int i; for (i = 2; i < 6; i++) vlib_prefetch_buffer_header (bufs[i], LOAD); } next = nexts; b = bufs; while (n_left_from >= 8) { ip_adjacency_t *adj0, *adj1; ip4_header_t *ip0, *ip1; u32 rw_len0, error0, adj_index0; u32 rw_len1, error1, adj_index1; u32 tx_sw_if_index0, tx_sw_if_index1; u8 *p; vlib_prefetch_buffer_header (b[6], LOAD); vlib_prefetch_buffer_header (b[7], LOAD); adj_index0 = vnet_buffer (b[0])->ip.adj_index[VLIB_TX]; adj_index1 = vnet_buffer (b[1])->ip.adj_index[VLIB_TX]; /* * pre-fetch the per-adjacency counters */ if (do_counters) { vlib_prefetch_combined_counter (&adjacency_counters, thread_index, adj_index0); vlib_prefetch_combined_counter (&adjacency_counters, thread_index, adj_index1); } ip0 = vlib_buffer_get_current (b[0]); ip1 = vlib_buffer_get_current (b[1]); error0 = error1 = IP4_ERROR_NONE; ip4_ttl_and_checksum_check (b[0], ip0, next + 0, &error0); ip4_ttl_and_checksum_check (b[1], ip1, next + 1, &error1); /* Rewrite packet header and updates lengths. */ adj0 = adj_get (adj_index0); adj1 = adj_get (adj_index1); /* Worth pipelining. No guarantee that adj0,1 are hot... */ rw_len0 = adj0[0].rewrite_header.data_bytes; rw_len1 = adj1[0].rewrite_header.data_bytes; vnet_buffer (b[0])->ip.save_rewrite_length = rw_len0; vnet_buffer (b[1])->ip.save_rewrite_length = rw_len1; p = vlib_buffer_get_current (b[2]); CLIB_PREFETCH (p - CLIB_CACHE_LINE_BYTES, CLIB_CACHE_LINE_BYTES, STORE); CLIB_PREFETCH (p, CLIB_CACHE_LINE_BYTES, LOAD); p = vlib_buffer_get_current (b[3]); CLIB_PREFETCH (p - CLIB_CACHE_LINE_BYTES, CLIB_CACHE_LINE_BYTES, STORE); CLIB_PREFETCH (p, CLIB_CACHE_LINE_BYTES, LOAD); /* Check MTU of outgoing interface. */ u16 ip0_len = clib_net_to_host_u16 (ip0->length); u16 ip1_len = clib_net_to_host_u16 (ip1->length); if (do_gso && (b[0]->flags & VNET_BUFFER_F_GSO)) ip0_len = gso_mtu_sz (b[0]); if (do_gso && (b[1]->flags & VNET_BUFFER_F_GSO)) ip1_len = gso_mtu_sz (b[1]); ip4_mtu_check (b[0], ip0_len, adj0[0].rewrite_header.max_l3_packet_bytes, ip0->flags_and_fragment_offset & clib_host_to_net_u16 (IP4_HEADER_FLAG_DONT_FRAGMENT), next + 0, &error0); ip4_mtu_check (b[1], ip1_len, adj1[0].rewrite_header.max_l3_packet_bytes, ip1->flags_and_fragment_offset & clib_host_to_net_u16 (IP4_HEADER_FLAG_DONT_FRAGMENT), next + 1, &error1); if (is_mcast) { error0 = ((adj0[0].rewrite_header.sw_if_index == vnet_buffer (b[0])->sw_if_index[VLIB_RX]) ? IP4_ERROR_SAME_INTERFACE : error0); error1 = ((adj1[0].rewrite_header.sw_if_index == vnet_buffer (b[1])->sw_if_index[VLIB_RX]) ? IP4_ERROR_SAME_INTERFACE : error1); } /* Don't adjust the buffer for ttl issue; icmp-error node wants * to see the IP header */ if (PREDICT_TRUE (error0 == IP4_ERROR_NONE)) { u32 next_index = adj0[0].rewrite_header.next_index; vlib_buffer_advance (b[0], -(word) rw_len0); tx_sw_if_index0 = adj0[0].rewrite_header.sw_if_index; vnet_buffer (b[0])->sw_if_index[VLIB_TX] = tx_sw_if_index0; if (PREDICT_FALSE (adj0[0].rewrite_header.flags & VNET_REWRITE_HAS_FEATURES)) vnet_feature_arc_start (lm->output_feature_arc_index, tx_sw_if_index0, &next_index, b[0]); next[0] = next_index; } else { b[0]->error = error_node->errors[error0]; } if (PREDICT_TRUE (error1 == IP4_ERROR_NONE)) { u32 next_index = adj1[0].rewrite_header.next_index; vlib_buffer_advance (b[1], -(word) rw_len1); tx_sw_if_index1 = adj1[0].rewrite_header.sw_if_index; vnet_buffer (b[1])->sw_if_index[VLIB_TX] = tx_sw_if_index1; if (PREDICT_FALSE (adj1[0].rewrite_header.flags & VNET_REWRITE_HAS_FEATURES)) vnet_feature_arc_start (lm->output_feature_arc_index, tx_sw_if_index1, &next_index, b[1]); next[1] = next_index; } else { b[1]->error = error_node->errors[error1]; } if (is_midchain) { calc_checksums (vm, b[0]); calc_checksums (vm, b[1]); } /* Guess we are only writing on simple Ethernet header. */ vnet_rewrite_two_headers (adj0[0], adj1[0], ip0, ip1, sizeof (ethernet_header_t)); /* * Bump the per-adjacency counters */ if (do_counters) { vlib_increment_combined_counter (&adjacency_counters, thread_index, adj_index0, 1, vlib_buffer_length_in_chain (vm, b[0]) + rw_len0); vlib_increment_combined_counter (&adjacency_counters, thread_index, adj_index1, 1, vlib_buffer_length_in_chain (vm, b[1]) + rw_len1); } if (is_midchain) { if (adj0->sub_type.midchain.fixup_func) adj0->sub_type.midchain.fixup_func (vm, adj0, b[0], adj0->sub_type.midchain.fixup_data); if (adj1->sub_type.midchain.fixup_func) adj1->sub_type.midchain.fixup_func (vm, adj1, b[1], adj1->sub_type.midchain.fixup_data); } if (is_mcast) { /* * copy bytes from the IP address into the MAC rewrite */ vnet_ip_mcast_fixup_header (IP4_MCAST_ADDR_MASK, adj0->rewrite_header.dst_mcast_offset, &ip0->dst_address.as_u32, (u8 *) ip0); vnet_ip_mcast_fixup_header (IP4_MCAST_ADDR_MASK, adj1->rewrite_header.dst_mcast_offset, &ip1->dst_address.as_u32, (u8 *) ip1); } next += 2; b += 2; n_left_from -= 2; } while (n_left_from > 0) { ip_adjacency_t *adj0; ip4_header_t *ip0; u32 rw_len0, adj_index0, error0; u32 tx_sw_if_index0; adj_index0 = vnet_buffer (b[0])->ip.adj_index[VLIB_TX]; adj0 = adj_get (adj_index0); if (do_counters) vlib_prefetch_combined_counter (&adjacency_counters, thread_index, adj_index0); ip0 = vlib_buffer_get_current (b[0]); error0 = IP4_ERROR_NONE; ip4_ttl_and_checksum_check (b[0], ip0, next + 0, &error0); /* Update packet buffer attributes/set output interface. */ rw_len0 = adj0[0].rewrite_header.data_bytes; vnet_buffer (b[0])->ip.save_rewrite_length = rw_len0; /* Check MTU of outgoing interface. */ u16 ip0_len = clib_net_to_host_u16 (ip0->length); if (do_gso && (b[0]->flags & VNET_BUFFER_F_GSO)) ip0_len = gso_mtu_sz (b[0]); ip4_mtu_check (b[0], ip0_len, adj0[0].rewrite_header.max_l3_packet_bytes, ip0->flags_and_fragment_offset & clib_host_to_net_u16 (IP4_HEADER_FLAG_DONT_FRAGMENT), next + 0, &error0); if (is_mcast) { error0 = ((adj0[0].rewrite_header.sw_if_index == vnet_buffer (b[0])->sw_if_index[VLIB_RX]) ? IP4_ERROR_SAME_INTERFACE : error0); } /* Don't adjust the buffer for ttl issue; icmp-error node wants * to see the IP header */ if (PREDICT_TRUE (error0 == IP4_ERROR_NONE)) { u32 next_index = adj0[0].rewrite_header.next_index; vlib_buffer_advance (b[0], -(word) rw_len0); tx_sw_if_index0 = adj0[0].rewrite_header.sw_if_index; vnet_buffer (b[0])->sw_if_index[VLIB_TX] = tx_sw_if_index0; if (PREDICT_FALSE (adj0[0].rewrite_header.flags & VNET_REWRITE_HAS_FEATURES)) vnet_feature_arc_start (lm->output_feature_arc_index, tx_sw_if_index0, &next_index, b[0]); next[0] = next_index; } else { b[0]->error = error_node->errors[error0]; } if (is_midchain) { calc_checksums (vm, b[0]); } /* Guess we are only writing on simple Ethernet header. */ vnet_rewrite_one_header (adj0[0], ip0, sizeof (ethernet_header_t)); if (do_counters) vlib_increment_combined_counter (&adjacency_counters, thread_index, adj_index0, 1, vlib_buffer_length_in_chain (vm, b[0]) + rw_len0); if (is_midchain) { if (adj0->sub_type.midchain.fixup_func) adj0->sub_type.midchain.fixup_func (vm, adj0, b[0], adj0->sub_type.midchain.fixup_data); } if (is_mcast) { /* * copy bytes from the IP address into the MAC rewrite */ vnet_ip_mcast_fixup_header (IP4_MCAST_ADDR_MASK, adj0->rewrite_header.dst_mcast_offset, &ip0->dst_address.as_u32, (u8 *) ip0); } next += 1; b += 1; n_left_from -= 1; } /* Need to do trace after rewrites to pick up new packet data. */ if (node->flags & VLIB_NODE_FLAG_TRACE) ip4_forward_next_trace (vm, node, frame, VLIB_TX); vlib_buffer_enqueue_to_next (vm, node, from, nexts, frame->n_vectors); return frame->n_vectors; } always_inline uword ip4_rewrite_inline (vlib_main_t * vm, vlib_node_runtime_t * node, vlib_frame_t * frame, int do_counters, int is_midchain, int is_mcast) { vnet_main_t *vnm = vnet_get_main (); if (PREDICT_FALSE (vnm->interface_main.gso_interface_count > 0)) return ip4_rewrite_inline_with_gso (vm, node, frame, do_counters, is_midchain, is_mcast, 1 /* do_gso */ ); else return ip4_rewrite_inline_with_gso (vm, node, frame, do_counters, is_midchain, is_mcast, 0 /* no do_gso */ ); } /** @brief IPv4 rewrite node. @node ip4-rewrite This is the IPv4 transit-rewrite node: decrement TTL, fix the ipv4 header checksum, fetch the ip adjacency, check the outbound mtu, apply the adjacency rewrite, and send pkts to the adjacency rewrite header's rewrite_next_index. @param vm vlib_main_t corresponding to the current thread @param node vlib_node_runtime_t @param frame vlib_frame_t whose contents should be dispatched @par Graph mechanics: buffer metadata, next index usage @em Uses: - vnet_buffer(b)->ip.adj_index[VLIB_TX] - the rewrite adjacency index - adj->lookup_next_index - Must be IP_LOOKUP_NEXT_REWRITE or IP_LOOKUP_NEXT_ARP, otherwise the packet will be dropped. - adj->rewrite_header - Rewrite string length, rewrite string, next_index @em Sets: - b->current_data, b->current_length - Updated net of applying the rewrite string Next Indices: - adj->rewrite_header.next_index or @c ip4-drop */ VLIB_NODE_FN (ip4_rewrite_node) (vlib_main_t * vm, vlib_node_runtime_t * node, vlib_frame_t * frame) { if (adj_are_counters_enabled ()) return ip4_rewrite_inline (vm, node, frame, 1, 0, 0); else return ip4_rewrite_inline (vm, node, frame, 0, 0, 0); } VLIB_NODE_FN (ip4_rewrite_bcast_node) (vlib_main_t * vm, vlib_node_runtime_t * node, vlib_frame_t * frame) { if (adj_are_counters_enabled ()) return ip4_rewrite_inline (vm, node, frame, 1, 0, 0); else return ip4_rewrite_inline (vm, node, frame, 0, 0, 0); } VLIB_NODE_FN (ip4_midchain_node) (vlib_main_t * vm, vlib_node_runtime_t * node, vlib_frame_t * frame) { if (adj_are_counters_enabled ()) return ip4_rewrite_inline (vm, node, frame, 1, 1, 0); else return ip4_rewrite_inline (vm, node, frame, 0, 1, 0); } VLIB_NODE_FN (ip4_rewrite_mcast_node) (vlib_main_t * vm, vlib_node_runtime_t * node, vlib_frame_t * frame) { if (adj_are_counters_enabled ()) return ip4_rewrite_inline (vm, node, frame, 1, 0, 1); else return ip4_rewrite_inline (vm, node, frame, 0, 0, 1); } VLIB_NODE_FN (ip4_mcast_midchain_node) (vlib_main_t * vm, vlib_node_runtime_t * node, vlib_frame_t * frame) { if (adj_are_counters_enabled ()) return ip4_rewrite_inline (vm, node, frame, 1, 1, 1); else return ip4_rewrite_inline (vm, node, frame, 0, 1, 1); } /* *INDENT-OFF* */ VLIB_REGISTER_NODE (ip4_rewrite_node) = { .name = "ip4-rewrite", .vector_size = sizeof (u32), .format_trace = format_ip4_rewrite_trace, .n_next_nodes = IP4_REWRITE_N_NEXT, .next_nodes = { [IP4_REWRITE_NEXT_DROP] = "ip4-drop", [IP4_REWRITE_NEXT_ICMP_ERROR] = "ip4-icmp-error", [IP4_REWRITE_NEXT_FRAGMENT] = "ip4-frag", }, }; VLIB_REGISTER_NODE (ip4_rewrite_bcast_node) = { .name = "ip4-rewrite-bcast", .vector_size = sizeof (u32), .format_trace = format_ip4_rewrite_trace, .sibling_of = "ip4-rewrite", }; VLIB_REGISTER_NODE (ip4_rewrite_mcast_node) = { .name = "ip4-rewrite-mcast", .vector_size = sizeof (u32), .format_trace = format_ip4_rewrite_trace, .sibling_of = "ip4-rewrite", }; VLIB_REGISTER_NODE (ip4_mcast_midchain_node) = { .name = "ip4-mcast-midchain", .vector_size = sizeof (u32), .format_trace = format_ip4_rewrite_trace, .sibling_of = "ip4-rewrite", }; VLIB_REGISTER_NODE (ip4_midchain_node) = { .name = "ip4-midchain", .vector_size = sizeof (u32), .format_trace = format_ip4_forward_next_trace, .sibling_of = "ip4-rewrite", }; /* *INDENT-ON */ static int ip4_lookup_validate (ip4_address_t * a, u32 fib_index0) { ip4_fib_mtrie_t *mtrie0; ip4_fib_mtrie_leaf_t leaf0; u32 lbi0; mtrie0 = &ip4_fib_get (fib_index0)->mtrie; leaf0 = ip4_fib_mtrie_lookup_step_one (mtrie0, a); leaf0 = ip4_fib_mtrie_lookup_step (mtrie0, leaf0, a, 2); leaf0 = ip4_fib_mtrie_lookup_step (mtrie0, leaf0, a, 3); lbi0 = ip4_fib_mtrie_leaf_get_adj_index (leaf0); return lbi0 == ip4_fib_table_lookup_lb (ip4_fib_get (fib_index0), a); } static clib_error_t * test_lookup_command_fn (vlib_main_t * vm, unformat_input_t * input, vlib_cli_command_t * cmd) { ip4_fib_t *fib; u32 table_id = 0; f64 count = 1; u32 n; int i; ip4_address_t ip4_base_address; u64 errors = 0; while (unformat_check_input (input) != UNFORMAT_END_OF_INPUT) { if (unformat (input, "table %d", &table_id)) { /* Make sure the entry exists. */ fib = ip4_fib_get (table_id); if ((fib) && (fib->index != table_id)) return clib_error_return (0, " %d does not exist", table_id); } else if (unformat (input, "count %f", &count)) ; else if (unformat (input, "%U", unformat_ip4_address, &ip4_base_address)) ; else return clib_error_return (0, "unknown input `%U'", format_unformat_error, input); } n = count; for (i = 0; i < n; i++) { if (!ip4_lookup_validate (&ip4_base_address, table_id)) errors++; ip4_base_address.as_u32 = clib_host_to_net_u32 (1 + clib_net_to_host_u32 (ip4_base_address.as_u32)); } if (errors) vlib_cli_output (vm, "%llu errors out of %d lookups\n", errors, n); else vlib_cli_output (vm, "No errors in %d lookups\n", n); return 0; } /*? * Perform a lookup of an IPv4 Address (or range of addresses) in the * given FIB table to determine if there is a conflict with the * adjacency table. The fib-id can be determined by using the * 'show ip fib' command. If fib-id is not entered, default value * of 0 is used. * * @todo This command uses fib-id, other commands use table-id (not * just a name, they are different indexes). Would like to change this * to table-id for consistency. * * @cliexpar * Example of how to run the test lookup command: * @cliexstart{test lookup 172.16.1.1 table 1 count 2} * No errors in 2 lookups * @cliexend ?*/ /* *INDENT-OFF* */ VLIB_CLI_COMMAND (lookup_test_command, static) = { .path = "test lookup", .short_help = "test lookup [table ] [count ]", .function = test_lookup_command_fn, }; /* *INDENT-ON* */ #ifndef CLIB_MARCH_VARIANT int vnet_set_ip4_flow_hash (u32 table_id, u32 flow_hash_config) { u32 fib_index; fib_index = fib_table_find (FIB_PROTOCOL_IP4, table_id); if (~0 == fib_index) return VNET_API_ERROR_NO_SUCH_FIB; fib_table_set_flow_hash_config (fib_index, FIB_PROTOCOL_IP4, flow_hash_config); return 0; } #endif static clib_error_t * set_ip_flow_hash_command_fn (vlib_main_t * vm, unformat_input_t * input, vlib_cli_command_t * cmd) { int matched = 0; u32 table_id = 0; u32 flow_hash_config = 0; int rv; while (unformat_check_input (input) != UNFORMAT_END_OF_INPUT) { if (unformat (input, "table %d", &table_id)) matched = 1; #define _(a,v) \ else if (unformat (input, #a)) { flow_hash_config |= v; matched=1;} foreach_flow_hash_bit #undef _ else break; } if (matched == 0) return clib_error_return (0, "unknown input `%U'", format_unformat_error, input); rv = vnet_set_ip4_flow_hash (table_id, flow_hash_config); switch (rv) { case 0: break; case VNET_API_ERROR_NO_SUCH_FIB: return clib_error_return (0, "no such FIB table %d", table_id); default: clib_warning ("BUG: illegal flow hash config 0x%x", flow_hash_config); break; } return 0; } /*? * Configure the set of IPv4 fields used by the flow hash. * * @cliexpar * Example of how to set the flow hash on a given table: * @cliexcmd{set ip flow-hash table 7 dst sport dport proto} * Example of display the configured flow hash: * @cliexstart{show ip fib} * ipv4-VRF:0, fib_index 0, flow hash: src dst sport dport proto * 0.0.0.0/0 * unicast-ip4-chain * [@0]: dpo-load-balance: [index:0 buckets:1 uRPF:0 to:[0:0]] * [0] [@0]: dpo-drop ip6 * 0.0.0.0/32 * unicast-ip4-chain * [@0]: dpo-load-balance: [index:1 buckets:1 uRPF:1 to:[0:0]] * [0] [@0]: dpo-drop ip6 * 224.0.0.0/8 * unicast-ip4-chain * [@0]: dpo-load-balance: [index:3 buckets:1 uRPF:3 to:[0:0]] * [0] [@0]: dpo-drop ip6 * 6.0.1.2/32 * unicast-ip4-chain * [@0]: dpo-load-balance: [index:30 buckets:1 uRPF:29 to:[0:0]] * [0] [@3]: arp-ipv4: via 6.0.0.1 af_packet0 * 7.0.0.1/32 * unicast-ip4-chain * [@0]: dpo-load-balance: [index:31 buckets:4 uRPF:30 to:[0:0]] * [0] [@3]: arp-ipv4: via 6.0.0.2 af_packet0 * [1] [@3]: arp-ipv4: via 6.0.0.2 af_packet0 * [2] [@3]: arp-ipv4: via 6.0.0.2 af_packet0 * [3] [@3]: arp-ipv4: via 6.0.0.1 af_packet0 * 240.0.0.0/8 * unicast-ip4-chain * [@0]: dpo-load-balance: [index:2 buckets:1 uRPF:2 to:[0:0]] * [0] [@0]: dpo-drop ip6 * 255.255.255.255/32 * unicast-ip4-chain * [@0]: dpo-load-balance: [index:4 buckets:1 uRPF:4 to:[0:0]] * [0] [@0]: dpo-drop ip6 * ipv4-VRF:7, fib_index 1, flow hash: dst sport dport proto * 0.0.0.0/0 * unicast-ip4-chain * [@0]: dpo-load-balance: [index:12 buckets:1 uRPF:11 to:[0:0]] * [0] [@0]: dpo-drop ip6 * 0.0.0.0/32 * unicast-ip4-chain * [@0]: dpo-load-balance: [index:13 buckets:1 uRPF:12 to:[0:0]] * [0] [@0]: dpo-drop ip6 * 172.16.1.0/24 * unicast-ip4-chain * [@0]: dpo-load-balance: [index:17 buckets:1 uRPF:16 to:[0:0]] * [0] [@4]: ipv4-glean: af_packet0 * 172.16.1.1/32 * unicast-ip4-chain * [@0]: dpo-load-balance: [index:18 buckets:1 uRPF:17 to:[1:84]] * [0] [@2]: dpo-receive: 172.16.1.1 on af_packet0 * 172.16.1.2/32 * unicast-ip4-chain * [@0]: dpo-load-balance: [index:21 buckets:1 uRPF:20 to:[0:0]] * [0] [@5]: ipv4 via 172.16.1.2 af_packet0: IP4: 02:fe:9e:70:7a:2b -> 26:a5:f6:9c:3a:36 * 172.16.2.0/24 * unicast-ip4-chain * [@0]: dpo-load-balance: [index:19 buckets:1 uRPF:18 to:[0:0]] * [0] [@4]: ipv4-glean: af_packet1 * 172.16.2.1/32 * unicast-ip4-chain * [@0]: dpo-load-balance: [index:20 buckets:1 uRPF:19 to:[0:0]] * [0] [@2]: dpo-receive: 172.16.2.1 on af_packet1 * 224.0.0.0/8 * unicast-ip4-chain * [@0]: dpo-load-balance: [index:15 buckets:1 uRPF:14 to:[0:0]] * [0] [@0]: dpo-drop ip6 * 240.0.0.0/8 * unicast-ip4-chain * [@0]: dpo-load-balance: [index:14 buckets:1 uRPF:13 to:[0:0]] * [0] [@0]: dpo-drop ip6 * 255.255.255.255/32 * unicast-ip4-chain * [@0]: dpo-load-balance: [index:16 buckets:1 uRPF:15 to:[0:0]] * [0] [@0]: dpo-drop ip6 * @cliexend ?*/ /* *INDENT-OFF* */ VLIB_CLI_COMMAND (set_ip_flow_hash_command, static) = { .path = "set ip flow-hash", .short_help = "set ip flow-hash table [src] [dst] [sport] [dport] [proto] [reverse]", .function = set_ip_flow_hash_command_fn, }; /* *INDENT-ON* */ #ifndef CLIB_MARCH_VARIANT int vnet_set_ip4_classify_intfc (vlib_main_t * vm, u32 sw_if_index, u32 table_index) { vnet_main_t *vnm = vnet_get_main (); vnet_interface_main_t *im = &vnm->interface_main; ip4_main_t *ipm = &ip4_main; ip_lookup_main_t *lm = &ipm->lookup_main; vnet_classify_main_t *cm = &vnet_classify_main; ip4_address_t *if_addr; if (pool_is_free_index (im->sw_interfaces, sw_if_index)) return VNET_API_ERROR_NO_MATCHING_INTERFACE; if (table_index != ~0 && pool_is_free_index (cm->tables, table_index)) return VNET_API_ERROR_NO_SUCH_ENTRY; vec_validate (lm->classify_table_index_by_sw_if_index, sw_if_index); lm->classify_table_index_by_sw_if_index[sw_if_index] = table_index; if_addr = ip4_interface_first_address (ipm, sw_if_index, NULL); if (NULL != if_addr) { fib_prefix_t pfx = { .fp_len = 32, .fp_proto = FIB_PROTOCOL_IP4, .fp_addr.ip4 = *if_addr, }; u32 fib_index; fib_index = fib_table_get_index_for_sw_if_index (FIB_PROTOCOL_IP4, sw_if_index); if (table_index != (u32) ~ 0) { dpo_id_t dpo = DPO_INVALID; dpo_set (&dpo, DPO_CLASSIFY, DPO_PROTO_IP4, classify_dpo_create (DPO_PROTO_IP4, table_index)); fib_table_entry_special_dpo_add (fib_index, &pfx, FIB_SOURCE_CLASSIFY, FIB_ENTRY_FLAG_NONE, &dpo); dpo_reset (&dpo); } else { fib_table_entry_special_remove (fib_index, &pfx, FIB_SOURCE_CLASSIFY); } } return 0; } #endif static clib_error_t * set_ip_classify_command_fn (vlib_main_t * vm, unformat_input_t * input, vlib_cli_command_t * cmd) { u32 table_index = ~0; int table_index_set = 0; u32 sw_if_index = ~0; int rv; while (unformat_check_input (input) != UNFORMAT_END_OF_INPUT) { if (unformat (input, "table-index %d", &table_index)) table_index_set = 1; else if (unformat (input, "intfc %U", unformat_vnet_sw_interface, vnet_get_main (), &sw_if_index)) ; else break; } if (table_index_set == 0) return clib_error_return (0, "classify table-index must be specified"); if (sw_if_index == ~0) return clib_error_return (0, "interface / subif must be specified"); rv = vnet_set_ip4_classify_intfc (vm, sw_if_index, table_index); switch (rv) { case 0: break; case VNET_API_ERROR_NO_MATCHING_INTERFACE: return clib_error_return (0, "No such interface"); case VNET_API_ERROR_NO_SUCH_ENTRY: return clib_error_return (0, "No such classifier table"); } return 0; } /*? * Assign a classification table to an interface. The classification * table is created using the 'classify table' and 'classify session' * commands. Once the table is create, use this command to filter packets * on an interface. * * @cliexpar * Example of how to assign a classification table to an interface: * @cliexcmd{set ip classify intfc GigabitEthernet2/0/0 table-index 1} ?*/ /* *INDENT-OFF* */ VLIB_CLI_COMMAND (set_ip_classify_command, static) = { .path = "set ip classify", .short_help = "set ip classify intfc table-index ", .function = set_ip_classify_command_fn, }; /* *INDENT-ON* */ static clib_error_t * ip4_config (vlib_main_t * vm, unformat_input_t * input) { ip4_main_t *im = &ip4_main; uword heapsize = 0; while (unformat_check_input (input) != UNFORMAT_END_OF_INPUT) { if (unformat (input, "heap-size %U", unformat_memory_size, &heapsize)) ; else return clib_error_return (0, "invalid heap-size parameter `%U'", format_unformat_error, input); } im->mtrie_heap_size = heapsize; return 0; } VLIB_EARLY_CONFIG_FUNCTION (ip4_config, "ip"); /* * fd.io coding-style-patch-verification: ON * * Local Variables: * eval: (c-set-style "gnu") * End: */