src/vnet/sr/sr_replicate.c - fdio/vpp - Gitiles

 /*
  * sr_replicate.c: ipv6 segment routing replicator for multicast
  *
  * Copyright (c) 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 Functions for replicating packets across SR tunnels.
  *
  *  Leverages rte_pktmbuf_clone() so there is no memcpy for
  *  invariant parts of the packet.
  *
  *  @note Currently requires DPDK
 */

 #if DPDK > 0			/* Cannot run replicate without DPDK */
 #include <vlib/vlib.h>
 #include <vnet/vnet.h>
 #include <vnet/pg/pg.h>
 #include <vnet/sr/sr.h>
 #include <vnet/devices/dpdk/dpdk.h>
 #include <vnet/ip/ip.h>
 #include <vnet/fib/ip6_fib.h>

 #include <vppinfra/hash.h>
 #include <vppinfra/error.h>
 #include <vppinfra/elog.h>

 /**
  *   @brief sr_replicate state.
  *
 */
 typedef struct
 {
   /* convenience */
   vlib_main_t *vlib_main;
   vnet_main_t *vnet_main;
 } sr_replicate_main_t;

 sr_replicate_main_t sr_replicate_main;

 /**
  *    @brief Information to display in packet trace.
  *
 */
 typedef struct
 {
   ip6_address_t src, dst;
   u16 length;
   u32 next_index;
   u32 tunnel_index;
   u8 sr[256];
 } sr_replicate_trace_t;

 /**
  *  @brief packet trace format function.
  *
  *  @param *s u8 used for string output
  *  @param *args va_list  structured input to va_arg to output @ref sr_replicate_trace_t
  *  @return *s u8 - formatted trace output
 */
 static u8 *
 format_sr_replicate_trace (u8 * s, va_list * args)
 {
   CLIB_UNUSED (vlib_main_t * vm) = va_arg (*args, vlib_main_t *);
   CLIB_UNUSED (vlib_node_t * node) = va_arg (*args, vlib_node_t *);
   sr_replicate_trace_t *t = va_arg (*args, sr_replicate_trace_t *);
   ip6_sr_main_t *sm = &sr_main;
   ip6_sr_tunnel_t *tun = pool_elt_at_index (sm->tunnels, t->tunnel_index);
   ip6_fib_t *rx_fib, *tx_fib;

   rx_fib = ip6_fib_get (tun->rx_fib_index);
   tx_fib = ip6_fib_get (tun->tx_fib_index);

   s = format
     (s, "SR-REPLICATE: next %s ip6 src %U dst %U len %u\n"
      "           rx-fib-id %d tx-fib-id %d\n%U",
      "ip6-lookup",
      format_ip6_address, &t->src,
      format_ip6_address, &t->dst, t->length,
      rx_fib->table_id, tx_fib->table_id,
      format_ip6_sr_header, t->sr, 0 /* print_hmac */ );
   return s;

 }

 #define foreach_sr_replicate_error \
 _(REPLICATED, "sr packets replicated") \
 _(NO_BUFFERS, "error allocating buffers for replicas") \
 _(NO_REPLICAS, "no replicas were needed") \
 _(NO_BUFFER_DROPS, "sr no buffer drops")

 /**
  * @brief Struct for SR replicate errors
  */
 typedef enum
 {
 #define _(sym,str) SR_REPLICATE_ERROR_##sym,
   foreach_sr_replicate_error
 #undef _
     SR_REPLICATE_N_ERROR,
 } sr_replicate_error_t;

 /**
  * @brief Error strings for SR replicate
  */
 static char *sr_replicate_error_strings[] = {
 #define _(sym,string) string,
   foreach_sr_replicate_error
 #undef _
 };

 /**
  * @brief Defines next-nodes for packet processing.
  *
 */
 typedef enum
 {
   SR_REPLICATE_NEXT_IP6_LOOKUP,
   SR_REPLICATE_N_NEXT,
 } sr_replicate_next_t;

 /**
  *   @brief Single loop packet replicator.
  *
  *   @node sr-replicate
  *   @param vm vlib_main_t
  *   @return frame->n_vectors uword
 */
 static uword
 sr_replicate_node_fn (vlib_main_t * vm,
 		      vlib_node_runtime_t * node, vlib_frame_t * frame)
 {
   u32 n_left_from, *from, *to_next;
   sr_replicate_next_t next_index;
   int pkts_replicated = 0;
   ip6_sr_main_t *sm = &sr_main;
   int no_buffer_drops = 0;
   vlib_buffer_free_list_t *fl;
   unsigned socket_id = rte_socket_id ();
   vlib_buffer_main_t *bm = vm->buffer_main;

   from = vlib_frame_vector_args (frame);
   n_left_from = frame->n_vectors;
   next_index = node->cached_next_index;

   fl = vlib_buffer_get_free_list (vm, VLIB_BUFFER_DEFAULT_FREE_LIST_INDEX);

   while (n_left_from > 0)
     {
       u32 n_left_to_next;

       vlib_get_next_frame (vm, node, next_index, to_next, n_left_to_next);

       while (n_left_from > 0 && n_left_to_next > 0)
 	{
 	  u32 bi0, hdr_bi0;
 	  vlib_buffer_t *b0, *orig_b0;
 	  struct rte_mbuf *orig_mb0 = 0, *hdr_mb0 = 0, *clone0 = 0;
 	  struct rte_mbuf **hdr_vec = 0, **rte_mbuf_vec = 0;
 	  ip6_sr_policy_t *pol0 = 0;
 	  ip6_sr_tunnel_t *t0 = 0;
 	  ip6_sr_header_t *hdr_sr0 = 0;
 	  ip6_header_t *ip0 = 0, *hdr_ip0 = 0;
 	  int num_replicas = 0;
 	  int i;
 	  u32 len_bytes = sizeof (ip6_header_t);
 	  u8 next_hdr, ip_next_hdr = IPPROTO_IPV6_ROUTE;

 	  bi0 = from[0];

 	  b0 = vlib_get_buffer (vm, bi0);
 	  orig_b0 = b0;

 	  pol0 = pool_elt_at_index (sm->policies,
 				    vnet_buffer (b0)->ip.save_protocol);

 	  ip0 = vlib_buffer_get_current (b0);
 	  /* Skip forward to the punch-in point */
 	  vlib_buffer_advance (b0, sizeof (*ip0));
 	  next_hdr = ip0->protocol;

 	  /* HBH must immediately follow ipv6 header */
 	  if (PREDICT_FALSE
 	      (ip0->protocol == IP_PROTOCOL_IP6_HOP_BY_HOP_OPTIONS))
 	    {
 	      ip6_hop_by_hop_ext_t *ext_hdr =
 		(ip6_hop_by_hop_ext_t *) ip6_next_header (ip0);
 	      u32 ext_hdr_len = 0;
 	      ext_hdr_len = ip6_ext_header_len ((ip6_ext_header_t *) ext_hdr);
 	      len_bytes += ext_hdr_len;
 	      next_hdr = ext_hdr->next_hdr;
 	      ext_hdr->next_hdr = IPPROTO_IPV6_ROUTE;
 	      ip_next_hdr = IP_PROTOCOL_IP6_HOP_BY_HOP_OPTIONS;
 	      /* Skip forward to the punch-in point */
 	      vlib_buffer_advance (b0, ext_hdr_len);

 	    }

 	  orig_mb0 = rte_mbuf_from_vlib_buffer (b0);

 	  i16 delta0 = vlib_buffer_length_in_chain (vm, orig_b0)
 	    - (i16) orig_mb0->pkt_len;

 	  u16 new_data_len0 = (u16) ((i16) orig_mb0->data_len + delta0);
 	  u16 new_pkt_len0 = (u16) ((i16) orig_mb0->pkt_len + delta0);

 	  orig_mb0->data_len = new_data_len0;
 	  orig_mb0->pkt_len = new_pkt_len0;
 	  orig_mb0->data_off += (u16) (b0->current_data);

 	  /*
 	     Before entering loop determine if we can allocate:
 	     - all the new HEADER RTE_MBUFs and assign them to a vector
 	     - all the clones

 	     if successful, then iterate over vectors of resources

 	   */
 	  num_replicas = vec_len (pol0->tunnel_indices);

 	  if (PREDICT_FALSE (num_replicas == 0))
 	    {
 	      b0->error = node->errors[SR_REPLICATE_ERROR_NO_REPLICAS];
 	      goto do_trace0;
 	    }

 	  vec_reset_length (hdr_vec);
 	  vec_reset_length (rte_mbuf_vec);

 	  for (i = 0; i < num_replicas; i++)
 	    {
 	      uint8_t nb_seg;
 	      struct rte_mbuf *clone0i;
 	      vlib_buffer_t *clone0_c, *clone_b0;

 	      t0 = vec_elt_at_index (sm->tunnels, pol0->tunnel_indices[i]);
 	      hdr_mb0 = rte_pktmbuf_alloc (bm->pktmbuf_pools[socket_id]);

 	      if (i < (num_replicas - 1))
 		{
 		  /* Not the last tunnel to process */
 		  clone0 = rte_pktmbuf_clone
 		    (orig_mb0, bm->pktmbuf_pools[socket_id]);
 		  if (clone0 == 0)
 		    goto clone_fail;
 		  nb_seg = 0;
 		  clone0i = clone0;
 		  clone0_c = NULL;
 		  while ((clone0->nb_segs >= 1) && (nb_seg < clone0->nb_segs))
 		    {

 		      clone_b0 = vlib_buffer_from_rte_mbuf (clone0i);
 		      vlib_buffer_init_for_free_list (clone_b0, fl);

 		      ASSERT ((clone_b0->flags & VLIB_BUFFER_NEXT_PRESENT) ==
 			      0);
 		      ASSERT (clone_b0->current_data == 0);

 		      clone_b0->current_data =
 			(clone0i->buf_addr + clone0i->data_off) -
 			(void *) clone_b0->data;

 		      clone_b0->current_length = clone0i->data_len;
 		      if (PREDICT_FALSE (clone0_c != NULL))
 			{
 			  clone0_c->flags |= VLIB_BUFFER_NEXT_PRESENT;
 			  clone0_c->next_buffer =
 			    vlib_get_buffer_index (vm, clone_b0);
 			}
 		      clone0_c = clone_b0;
 		      clone0i = clone0i->next;
 		      nb_seg++;
 		    }
 		}
 	      else
 		/* First tunnel to process, use original MB */
 		clone0 = orig_mb0;


 	      if (PREDICT_FALSE (!clone0 || !hdr_mb0))
 		{
 		clone_fail:
 		  b0->error = node->errors[SR_REPLICATE_ERROR_NO_BUFFERS];

 		  vec_foreach_index (i, rte_mbuf_vec)
 		  {
 		    rte_pktmbuf_free (rte_mbuf_vec[i]);
 		  }
 		  vec_free (rte_mbuf_vec);

 		  vec_foreach_index (i, hdr_vec)
 		  {
 		    rte_pktmbuf_free (hdr_vec[i]);
 		  }
 		  vec_free (hdr_vec);

 		  goto do_trace0;
 		}

 	      vec_add1 (hdr_vec, hdr_mb0);
 	      vec_add1 (rte_mbuf_vec, clone0);

 	    }

 	  for (i = 0; i < num_replicas; i++)
 	    {
 	      vlib_buffer_t *hdr_b0;
 	      u16 new_l0 = 0;

 	      t0 = vec_elt_at_index (sm->tunnels, pol0->tunnel_indices[i]);
 	      /* Our replicas */
 	      hdr_mb0 = hdr_vec[i];
 	      clone0 = rte_mbuf_vec[i];

 	      hdr_mb0->data_len = len_bytes + vec_len (t0->rewrite);
 	      hdr_mb0->pkt_len = hdr_mb0->data_len +
 		vlib_buffer_length_in_chain (vm, orig_b0);

 	      hdr_b0 = vlib_buffer_from_rte_mbuf (hdr_mb0);

 	      vlib_buffer_init_for_free_list (hdr_b0, fl);

 	      memcpy (hdr_b0->data, ip0, len_bytes);
 	      memcpy (hdr_b0->data + len_bytes, t0->rewrite,
 		      vec_len (t0->rewrite));

 	      hdr_b0->current_data = 0;
 	      hdr_b0->current_length = len_bytes + vec_len (t0->rewrite);
 	      hdr_b0->flags = orig_b0->flags | VLIB_BUFFER_NEXT_PRESENT;
 	      hdr_b0->trace_index = orig_b0->trace_index;
 	      vnet_buffer (hdr_b0)->l2_classify.opaque_index = 0;

 	      hdr_b0->total_length_not_including_first_buffer =
 		hdr_mb0->pkt_len - hdr_b0->current_length;
 	      vnet_buffer (hdr_b0)->sw_if_index[VLIB_TX] = t0->tx_fib_index;

 	      hdr_ip0 = (ip6_header_t *) hdr_b0->data;
 	      new_l0 = clib_net_to_host_u16 (ip0->payload_length) +
 		vec_len (t0->rewrite);
 	      hdr_ip0->payload_length = clib_host_to_net_u16 (new_l0);
 	      hdr_sr0 = (ip6_sr_header_t *) ((u8 *) hdr_ip0 + len_bytes);
 	      /* $$$ tune */
 	      clib_memcpy (hdr_sr0, t0->rewrite, vec_len (t0->rewrite));
 	      hdr_sr0->protocol = next_hdr;
 	      hdr_ip0->protocol = ip_next_hdr;

 	      /* Copy dst address into the DA slot in the segment list */
 	      clib_memcpy (hdr_sr0->segments, ip0->dst_address.as_u64,
 			   sizeof (ip6_address_t));

 	      /* Rewrite the ip6 dst address */
 	      hdr_ip0->dst_address.as_u64[0] = t0->first_hop.as_u64[0];
 	      hdr_ip0->dst_address.as_u64[1] = t0->first_hop.as_u64[1];

 	      sr_fix_hmac (sm, hdr_ip0, hdr_sr0);

 	      /* prepend new header to invariant piece */
 	      hdr_mb0->next = clone0;
 	      hdr_b0->next_buffer =
 		vlib_get_buffer_index (vm,
 				       vlib_buffer_from_rte_mbuf (clone0));

 	      /* update header's fields */
 	      hdr_mb0->pkt_len =
 		(uint16_t) (hdr_mb0->data_len + clone0->pkt_len);
 	      hdr_mb0->nb_segs = (uint8_t) (clone0->nb_segs + 1);

 	      /* copy metadata from source packet */
 	      hdr_mb0->port = clone0->port;
 	      hdr_mb0->vlan_tci = clone0->vlan_tci;
 	      hdr_mb0->vlan_tci_outer = clone0->vlan_tci_outer;
 	      hdr_mb0->tx_offload = clone0->tx_offload;
 	      hdr_mb0->hash = clone0->hash;

 	      hdr_mb0->ol_flags = clone0->ol_flags & ~(IND_ATTACHED_MBUF);

 	      __rte_mbuf_sanity_check (hdr_mb0, 1);

 	      hdr_bi0 = vlib_get_buffer_index (vm, hdr_b0);

 	      to_next[0] = hdr_bi0;
 	      to_next += 1;
 	      n_left_to_next -= 1;

 	      if (n_left_to_next == 0)
 		{
 		  vlib_put_next_frame (vm, node, next_index, n_left_to_next);
 		  vlib_get_next_frame (vm, node, next_index,
 				       to_next, n_left_to_next);

 		}
 	      pkts_replicated++;
 	    }

 	  from += 1;
 	  n_left_from -= 1;

 	do_trace0:
 	  if (PREDICT_FALSE (b0->flags & VLIB_BUFFER_IS_TRACED))
 	    {
 	      sr_replicate_trace_t *tr = vlib_add_trace (vm, node,
 							 b0, sizeof (*tr));
 	      tr->tunnel_index = t0 - sm->tunnels;
 	      tr->length = 0;
 	      if (hdr_ip0)
 		{
 		  memcpy (tr->src.as_u8, hdr_ip0->src_address.as_u8,
 			  sizeof (tr->src.as_u8));
 		  memcpy (tr->dst.as_u8, hdr_ip0->dst_address.as_u8,
 			  sizeof (tr->dst.as_u8));
 		  if (hdr_ip0->payload_length)
 		    tr->length = clib_net_to_host_u16
 		      (hdr_ip0->payload_length);
 		}
 	      tr->next_index = next_index;
 	      if (hdr_sr0)
 		memcpy (tr->sr, hdr_sr0, sizeof (tr->sr));
 	    }

 	}

       vlib_put_next_frame (vm, node, next_index, n_left_to_next);
     }

   vlib_node_increment_counter (vm, sr_replicate_node.index,
 			       SR_REPLICATE_ERROR_REPLICATED,
 			       pkts_replicated);

   vlib_node_increment_counter (vm, sr_replicate_node.index,
 			       SR_REPLICATE_ERROR_NO_BUFFER_DROPS,
 			       no_buffer_drops);

   return frame->n_vectors;
 }

 /* *INDENT-OFF* */
 VLIB_REGISTER_NODE (sr_replicate_node) = {
   .function = sr_replicate_node_fn,
   .name = "sr-replicate",
   .vector_size = sizeof (u32),
   .format_trace = format_sr_replicate_trace,
   .type = VLIB_NODE_TYPE_INTERNAL,

   .n_errors = ARRAY_LEN(sr_replicate_error_strings),
   .error_strings = sr_replicate_error_strings,

   .n_next_nodes = SR_REPLICATE_N_NEXT,

   .next_nodes = {
         [SR_REPLICATE_NEXT_IP6_LOOKUP] = "ip6-lookup",
   },
 };

 VLIB_NODE_FUNCTION_MULTIARCH (sr_replicate_node, sr_replicate_node_fn)
 /* *INDENT-ON* */

 clib_error_t *
 sr_replicate_init (vlib_main_t * vm)
 {
   sr_replicate_main_t *msm = &sr_replicate_main;

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

   return 0;
 }

 VLIB_INIT_FUNCTION (sr_replicate_init);

 #endif /* DPDK */

 /*
  * fd.io coding-style-patch-verification: ON
  *
  * Local Variables:
  * eval: (c-set-style "gnu")
  * End:
  */
	/*
	* sr_replicate.c: ipv6 segment routing replicator for multicast
	*
	* Copyright (c) 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 Functions for replicating packets across SR tunnels.
	*
	* Leverages rte_pktmbuf_clone() so there is no memcpy for
	* invariant parts of the packet.
	*
	* @note Currently requires DPDK
	*/

	#if DPDK > 0 /* Cannot run replicate without DPDK */
	#include <vlib/vlib.h>
	#include <vnet/vnet.h>
	#include <vnet/pg/pg.h>
	#include <vnet/sr/sr.h>
	#include <vnet/devices/dpdk/dpdk.h>
	#include <vnet/ip/ip.h>
	#include <vnet/fib/ip6_fib.h>

	#include <vppinfra/hash.h>
	#include <vppinfra/error.h>
	#include <vppinfra/elog.h>

	/**
	* @brief sr_replicate state.
	*
	*/
	typedef struct
	{
	/* convenience */
	vlib_main_t *vlib_main;
	vnet_main_t *vnet_main;
	} sr_replicate_main_t;

	sr_replicate_main_t sr_replicate_main;

	/**
	* @brief Information to display in packet trace.
	*
	*/
	typedef struct
	{
	ip6_address_t src, dst;
	u16 length;
	u32 next_index;
	u32 tunnel_index;
	u8 sr[256];
	} sr_replicate_trace_t;

	/**
	* @brief packet trace format function.
	*
	* @param *s u8 used for string output
	* @param *args va_list structured input to va_arg to output @ref sr_replicate_trace_t
	* @return *s u8 - formatted trace output
	*/
	static u8 *
	format_sr_replicate_trace (u8 * s, va_list * args)
	{
	CLIB_UNUSED (vlib_main_t * vm) = va_arg (args, vlib_main_t );
	CLIB_UNUSED (vlib_node_t * node) = va_arg (args, vlib_node_t );
	sr_replicate_trace_t t = va_arg (args, sr_replicate_trace_t *);
	ip6_sr_main_t *sm = &sr_main;
	ip6_sr_tunnel_t *tun = pool_elt_at_index (sm->tunnels, t->tunnel_index);
	ip6_fib_t rx_fib, tx_fib;

	rx_fib = ip6_fib_get (tun->rx_fib_index);
	tx_fib = ip6_fib_get (tun->tx_fib_index);

	s = format
	(s, "SR-REPLICATE: next %s ip6 src %U dst %U len %u\n"
	" rx-fib-id %d tx-fib-id %d\n%U",
	"ip6-lookup",
	format_ip6_address, &t->src,
	format_ip6_address, &t->dst, t->length,
	rx_fib->table_id, tx_fib->table_id,
	format_ip6_sr_header, t->sr, 0 /* print_hmac */ );
	return s;

	}

	#define foreach_sr_replicate_error \
	_(REPLICATED, "sr packets replicated") \
	_(NO_BUFFERS, "error allocating buffers for replicas") \
	_(NO_REPLICAS, "no replicas were needed") \
	_(NO_BUFFER_DROPS, "sr no buffer drops")

	/**
	* @brief Struct for SR replicate errors
	*/
	typedef enum
	{
	#define _(sym,str) SR_REPLICATE_ERROR_##sym,
	foreach_sr_replicate_error
	#undef _
	SR_REPLICATE_N_ERROR,
	} sr_replicate_error_t;

	/**
	* @brief Error strings for SR replicate
	*/
	static char *sr_replicate_error_strings[] = {
	#define _(sym,string) string,
	foreach_sr_replicate_error
	#undef _
	};

	/**
	* @brief Defines next-nodes for packet processing.
	*
	*/
	typedef enum
	{
	SR_REPLICATE_NEXT_IP6_LOOKUP,
	SR_REPLICATE_N_NEXT,
	} sr_replicate_next_t;

	/**
	* @brief Single loop packet replicator.
	*
	* @node sr-replicate
	* @param vm vlib_main_t
	* @return frame->n_vectors uword
	*/
	static uword
	sr_replicate_node_fn (vlib_main_t * vm,
	vlib_node_runtime_t * node, vlib_frame_t * frame)
	{
	u32 n_left_from, from, to_next;
	sr_replicate_next_t next_index;
	int pkts_replicated = 0;
	ip6_sr_main_t *sm = &sr_main;
	int no_buffer_drops = 0;
	vlib_buffer_free_list_t *fl;
	unsigned socket_id = rte_socket_id ();
	vlib_buffer_main_t *bm = vm->buffer_main;

	from = vlib_frame_vector_args (frame);
	n_left_from = frame->n_vectors;
	next_index = node->cached_next_index;

	fl = vlib_buffer_get_free_list (vm, VLIB_BUFFER_DEFAULT_FREE_LIST_INDEX);

	while (n_left_from > 0)
	{
	u32 n_left_to_next;

	vlib_get_next_frame (vm, node, next_index, to_next, n_left_to_next);

	while (n_left_from > 0 && n_left_to_next > 0)
	{
	u32 bi0, hdr_bi0;
	vlib_buffer_t b0, orig_b0;
	struct rte_mbuf orig_mb0 = 0, hdr_mb0 = 0, *clone0 = 0;
	struct rte_mbuf hdr_vec = 0, rte_mbuf_vec = 0;
	ip6_sr_policy_t *pol0 = 0;
	ip6_sr_tunnel_t *t0 = 0;
	ip6_sr_header_t *hdr_sr0 = 0;
	ip6_header_t ip0 = 0, hdr_ip0 = 0;
	int num_replicas = 0;
	int i;
	u32 len_bytes = sizeof (ip6_header_t);
	u8 next_hdr, ip_next_hdr = IPPROTO_IPV6_ROUTE;

	bi0 = from[0];

	b0 = vlib_get_buffer (vm, bi0);
	orig_b0 = b0;

	pol0 = pool_elt_at_index (sm->policies,
	vnet_buffer (b0)->ip.save_protocol);

	ip0 = vlib_buffer_get_current (b0);
	/* Skip forward to the punch-in point */
	vlib_buffer_advance (b0, sizeof (*ip0));
	next_hdr = ip0->protocol;

	/* HBH must immediately follow ipv6 header */
	if (PREDICT_FALSE
	(ip0->protocol == IP_PROTOCOL_IP6_HOP_BY_HOP_OPTIONS))
	{
	ip6_hop_by_hop_ext_t *ext_hdr =
	(ip6_hop_by_hop_ext_t *) ip6_next_header (ip0);
	u32 ext_hdr_len = 0;
	ext_hdr_len = ip6_ext_header_len ((ip6_ext_header_t *) ext_hdr);
	len_bytes += ext_hdr_len;
	next_hdr = ext_hdr->next_hdr;
	ext_hdr->next_hdr = IPPROTO_IPV6_ROUTE;
	ip_next_hdr = IP_PROTOCOL_IP6_HOP_BY_HOP_OPTIONS;
	/* Skip forward to the punch-in point */
	vlib_buffer_advance (b0, ext_hdr_len);

	}

	orig_mb0 = rte_mbuf_from_vlib_buffer (b0);

	i16 delta0 = vlib_buffer_length_in_chain (vm, orig_b0)
	- (i16) orig_mb0->pkt_len;

	u16 new_data_len0 = (u16) ((i16) orig_mb0->data_len + delta0);
	u16 new_pkt_len0 = (u16) ((i16) orig_mb0->pkt_len + delta0);

	orig_mb0->data_len = new_data_len0;
	orig_mb0->pkt_len = new_pkt_len0;
	orig_mb0->data_off += (u16) (b0->current_data);

	/*
	Before entering loop determine if we can allocate:
	- all the new HEADER RTE_MBUFs and assign them to a vector
	- all the clones

	if successful, then iterate over vectors of resources

	*/
	num_replicas = vec_len (pol0->tunnel_indices);

	if (PREDICT_FALSE (num_replicas == 0))
	{
	b0->error = node->errors[SR_REPLICATE_ERROR_NO_REPLICAS];
	goto do_trace0;
	}

	vec_reset_length (hdr_vec);
	vec_reset_length (rte_mbuf_vec);

	for (i = 0; i < num_replicas; i++)
	{
	uint8_t nb_seg;
	struct rte_mbuf *clone0i;
	vlib_buffer_t clone0_c, clone_b0;

	t0 = vec_elt_at_index (sm->tunnels, pol0->tunnel_indices[i]);
	hdr_mb0 = rte_pktmbuf_alloc (bm->pktmbuf_pools[socket_id]);

	if (i < (num_replicas - 1))
	{
	/* Not the last tunnel to process */
	clone0 = rte_pktmbuf_clone
	(orig_mb0, bm->pktmbuf_pools[socket_id]);
	if (clone0 == 0)
	goto clone_fail;
	nb_seg = 0;
	clone0i = clone0;
	clone0_c = NULL;
	while ((clone0->nb_segs >= 1) && (nb_seg < clone0->nb_segs))
	{

	clone_b0 = vlib_buffer_from_rte_mbuf (clone0i);
	vlib_buffer_init_for_free_list (clone_b0, fl);

	ASSERT ((clone_b0->flags & VLIB_BUFFER_NEXT_PRESENT) ==
	0);
	ASSERT (clone_b0->current_data == 0);

	clone_b0->current_data =
	(clone0i->buf_addr + clone0i->data_off) -
	(void *) clone_b0->data;

	clone_b0->current_length = clone0i->data_len;
	if (PREDICT_FALSE (clone0_c != NULL))
	{
	clone0_c->flags \|= VLIB_BUFFER_NEXT_PRESENT;
	clone0_c->next_buffer =
	vlib_get_buffer_index (vm, clone_b0);
	}
	clone0_c = clone_b0;
	clone0i = clone0i->next;
	nb_seg++;
	}
	}
	else
	/* First tunnel to process, use original MB */
	clone0 = orig_mb0;


	if (PREDICT_FALSE (!clone0 \|\| !hdr_mb0))
	{
	clone_fail:
	b0->error = node->errors[SR_REPLICATE_ERROR_NO_BUFFERS];

	vec_foreach_index (i, rte_mbuf_vec)
	{
	rte_pktmbuf_free (rte_mbuf_vec[i]);
	}
	vec_free (rte_mbuf_vec);

	vec_foreach_index (i, hdr_vec)
	{
	rte_pktmbuf_free (hdr_vec[i]);
	}
	vec_free (hdr_vec);

	goto do_trace0;
	}

	vec_add1 (hdr_vec, hdr_mb0);
	vec_add1 (rte_mbuf_vec, clone0);

	}

	for (i = 0; i < num_replicas; i++)
	{
	vlib_buffer_t *hdr_b0;
	u16 new_l0 = 0;

	t0 = vec_elt_at_index (sm->tunnels, pol0->tunnel_indices[i]);
	/* Our replicas */
	hdr_mb0 = hdr_vec[i];
	clone0 = rte_mbuf_vec[i];

	hdr_mb0->data_len = len_bytes + vec_len (t0->rewrite);
	hdr_mb0->pkt_len = hdr_mb0->data_len +
	vlib_buffer_length_in_chain (vm, orig_b0);

	hdr_b0 = vlib_buffer_from_rte_mbuf (hdr_mb0);

	vlib_buffer_init_for_free_list (hdr_b0, fl);

	memcpy (hdr_b0->data, ip0, len_bytes);
	memcpy (hdr_b0->data + len_bytes, t0->rewrite,
	vec_len (t0->rewrite));

	hdr_b0->current_data = 0;
	hdr_b0->current_length = len_bytes + vec_len (t0->rewrite);
	hdr_b0->flags = orig_b0->flags \| VLIB_BUFFER_NEXT_PRESENT;
	hdr_b0->trace_index = orig_b0->trace_index;
	vnet_buffer (hdr_b0)->l2_classify.opaque_index = 0;

	hdr_b0->total_length_not_including_first_buffer =
	hdr_mb0->pkt_len - hdr_b0->current_length;
	vnet_buffer (hdr_b0)->sw_if_index[VLIB_TX] = t0->tx_fib_index;

	hdr_ip0 = (ip6_header_t *) hdr_b0->data;
	new_l0 = clib_net_to_host_u16 (ip0->payload_length) +
	vec_len (t0->rewrite);
	hdr_ip0->payload_length = clib_host_to_net_u16 (new_l0);
	hdr_sr0 = (ip6_sr_header_t ) ((u8 ) hdr_ip0 + len_bytes);
	/* $$$ tune */
	clib_memcpy (hdr_sr0, t0->rewrite, vec_len (t0->rewrite));
	hdr_sr0->protocol = next_hdr;
	hdr_ip0->protocol = ip_next_hdr;

	/* Copy dst address into the DA slot in the segment list */
	clib_memcpy (hdr_sr0->segments, ip0->dst_address.as_u64,
	sizeof (ip6_address_t));

	/* Rewrite the ip6 dst address */
	hdr_ip0->dst_address.as_u64[0] = t0->first_hop.as_u64[0];
	hdr_ip0->dst_address.as_u64[1] = t0->first_hop.as_u64[1];

	sr_fix_hmac (sm, hdr_ip0, hdr_sr0);

	/* prepend new header to invariant piece */
	hdr_mb0->next = clone0;
	hdr_b0->next_buffer =
	vlib_get_buffer_index (vm,
	vlib_buffer_from_rte_mbuf (clone0));

	/* update header's fields */
	hdr_mb0->pkt_len =
	(uint16_t) (hdr_mb0->data_len + clone0->pkt_len);
	hdr_mb0->nb_segs = (uint8_t) (clone0->nb_segs + 1);

	/* copy metadata from source packet */
	hdr_mb0->port = clone0->port;
	hdr_mb0->vlan_tci = clone0->vlan_tci;
	hdr_mb0->vlan_tci_outer = clone0->vlan_tci_outer;
	hdr_mb0->tx_offload = clone0->tx_offload;
	hdr_mb0->hash = clone0->hash;

	hdr_mb0->ol_flags = clone0->ol_flags & ~(IND_ATTACHED_MBUF);

	__rte_mbuf_sanity_check (hdr_mb0, 1);

	hdr_bi0 = vlib_get_buffer_index (vm, hdr_b0);

	to_next[0] = hdr_bi0;
	to_next += 1;
	n_left_to_next -= 1;

	if (n_left_to_next == 0)
	{
	vlib_put_next_frame (vm, node, next_index, n_left_to_next);
	vlib_get_next_frame (vm, node, next_index,
	to_next, n_left_to_next);

	}
	pkts_replicated++;
	}

	from += 1;
	n_left_from -= 1;

	do_trace0:
	if (PREDICT_FALSE (b0->flags & VLIB_BUFFER_IS_TRACED))
	{
	sr_replicate_trace_t *tr = vlib_add_trace (vm, node,
	b0, sizeof (*tr));
	tr->tunnel_index = t0 - sm->tunnels;
	tr->length = 0;
	if (hdr_ip0)
	{
	memcpy (tr->src.as_u8, hdr_ip0->src_address.as_u8,
	sizeof (tr->src.as_u8));
	memcpy (tr->dst.as_u8, hdr_ip0->dst_address.as_u8,
	sizeof (tr->dst.as_u8));
	if (hdr_ip0->payload_length)
	tr->length = clib_net_to_host_u16
	(hdr_ip0->payload_length);
	}
	tr->next_index = next_index;
	if (hdr_sr0)
	memcpy (tr->sr, hdr_sr0, sizeof (tr->sr));
	}

	}

	vlib_put_next_frame (vm, node, next_index, n_left_to_next);
	}

	vlib_node_increment_counter (vm, sr_replicate_node.index,
	SR_REPLICATE_ERROR_REPLICATED,
	pkts_replicated);

	vlib_node_increment_counter (vm, sr_replicate_node.index,
	SR_REPLICATE_ERROR_NO_BUFFER_DROPS,
	no_buffer_drops);

	return frame->n_vectors;
	}

	/* INDENT-OFF */
	VLIB_REGISTER_NODE (sr_replicate_node) = {
	.function = sr_replicate_node_fn,
	.name = "sr-replicate",
	.vector_size = sizeof (u32),
	.format_trace = format_sr_replicate_trace,
	.type = VLIB_NODE_TYPE_INTERNAL,

	.n_errors = ARRAY_LEN(sr_replicate_error_strings),
	.error_strings = sr_replicate_error_strings,

	.n_next_nodes = SR_REPLICATE_N_NEXT,

	.next_nodes = {
	[SR_REPLICATE_NEXT_IP6_LOOKUP] = "ip6-lookup",
	},
	};

	VLIB_NODE_FUNCTION_MULTIARCH (sr_replicate_node, sr_replicate_node_fn)
	/* INDENT-ON */

	clib_error_t *
	sr_replicate_init (vlib_main_t * vm)
	{
	sr_replicate_main_t *msm = &sr_replicate_main;

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

	return 0;
	}

	VLIB_INIT_FUNCTION (sr_replicate_init);

	#endif /* DPDK */

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